EPA-600/4-82-001
Hay 1982
APPLICATION AMD EVALUATION
OF
ANALYTICAL PROCEDURES
FOR
TRACE METALS, TOTAL CYANIDES AND PHENOLICS
by
Gary J. Gottfried
Biospherics Incorporated
4928 Wyaconda Road
Rockville, Maryland 20852
Contract No. 68-03-2788
Project Officer
Gerald McKee
Environmental Monitoring and Support Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
NOTE TO THE READER
Copies of the appendices cited in the text of this report were not
delivered to the project officer in a condition considered to be acceptable
for submission to the National Technical Information Service. Consequently,
they were not included in the report package.
Further, they do not include any new information, but consist of approxi
mately 450 pages of raw data generated during the study and presented in tab-
ular form. In view of this and the impact of the appendices' length on the
price of the report package available from the National Technical Information
Service, it was decided to not include them.
We regret any confusion or misunderstanding resulting from this action.
Persons interested in the raw data, presented in 450 pages of tables, should
address their requests for copies of the appendices to Gerald D. McKee,
Inorganic Analyses Section, Environmental Monitoring and Support Laboratory,
U.S. Environmental Protection Agency, 26 W. St. Clair, Cincinnati, Ohio 45268
-------
DISCLAIMER
This report has been reviewed by Environmental Monitoring
and Support Laboratory-Cincinnati, and approved for publication.
Approval does not signify that the contents necessarily reflect
the views and policies of the U.S. Environmental Protection
Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
i i
-------
FOREWORD
Environmental measurements are required to determine the
quality of ambient waters and the character of waste effluents.
The Environmental Monitoring and Support Laboratory - Cincinnati,
conducts research to:
o Develop and evaluate methods to measure the presence and
concentration of physical, chemical, and radiological
pollutants in water, wastewater, bottom sediments, and
solid waste.
o Investigate methods for the concentration, recovery, and
identification of viruses, bacteria and other micro-
biological organisms in water, and, to determine the
responses of aquatic organisms to water quality.
o Develop and operate an Agencv-wide quality assurance
program to assure standardization and quality control of
systems for monitoring water and wastewater.
o Develop and operate a computerized system for instrument
automation leading to improved data collection, analysis,
and quality control.
This report evaluates analytical procedures used to
determine trace metals, total cyanide and phenolics when applied
to samples representative of industry-wide matrices. Inter-
ferences, method equivalency and analytical precision were
investigated.
Robert L. Booth, Acting Director
Environmental Monitoring and
Support Laboratory - Cincinnati
-------
ABSTRACT
Analytical procedures for the determination of trace metals,
total cyanide and phenolics were systematically evaluated for
their applicability, industry-wide. Matrix interferences, methods
equivalency, and analytical precision were investigated through a
series of duplicate and spiked analyses on non-diluted and
diluted samples. Validation of the methodologies and identifica-
tion of their limitations were, thus, established both within
specific industrial classifications and across multiple indus-
trial processes.
Each analytical technique was affected to some degree by
certain matrix interferences. Direct flame aspiration produced
the most reliable results of any atomic absorption technique.
Mild and vigorous digestions preceding flame aspiration were
equivalent for most samples and elements. Graphite furnace and
chelation/extraction procedures generally produced reliable
results but were more susceptible to matrix effects than direct
flame aspiration. Graphite furnace and sodium borohydride
hydride generation methods for arsenic and selenium proved
acceptable, whereas zinc slurry hydride generation suffered from
losses during digestion. Graphite furnace and chelation extrac-
tion methods for hexavalent chromium were affected by specific
interferences but each produced valid results for most samples.
The three phenolic methods tested were not equivalent and
only 4-AAP colorimetric produced reasonable results. The MBTH
colorimetric and instrumental methods were highly limited by
severe matrix interferences or high background signals.
The pyridine pyrazolone and pyridine barbituric acid methods
for total cyanide produced equivalent results and generally
reliable data.
Holding times for phenolics and cyanides can be increased
from 24 hours to 10 days based on preservation studies.
This report was submited under Contract No. 68-03-2788 by
Biospherics Incorporated under the sponsorship of the U.S.
Environmental Protection Agency. The sampling and analytical
program was conducted over a 12-month period, June, 1979 - June,
1980.
IV
-------
PB 82-218512
CONTENTS
Page
Foreword iii
Abstract iv
Figures vi
Tables vii
Acknowledgements viii
1. Introduction 1
2. Conclusions and Recommendations 5
Trace Metals 5
Cyanides 14
Phenolics 15
3. Analytical Program 17
Trace Metals 17
Cyanides 20
Phenol ics 21
4. Sampling Program 2 7
5. Data Presentation 33
6. Quality Control 34
Trace Metals 34
Cyanides 34
Phenolics 35
Results 35
Data Handling and Management 35
7. Discussion 43
Trace Metals 43
Cyanides 60
Phenolics 63
References 82
v
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
36
37
39
64
65
70
71
72
73
74
75
76
77
78
FIGURES
Title
Trace Metals Data Coding Form
Phenolics/Cyanide Data Coding Form
Data Handling and Management System
Preservation Study Cyanides
Preservation Study
Preservation Study
Preservation Study
Preservation Study
Preservation Study
Preservation Study
Preservation Study
Preservation Study
Preservation Study
Preservation Study
Cyanides
Phenolics
Phenolics
Phenolics
Phenolics
Phenolics
Phenolics
Phenolics
Phenolics
Phenolics
VI
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
2
18
19
22
22
23
24
24
25
25
25
26
26
28
29
30
31
40
42
44
62
69
TABLES
Title
Description of SIC Codes
Trace Metals Procedures
Cyanide and Phenolic Procedures
Trace Metals Protocol 1
Trace Metals Protocol 2
Trace Metals Protocol 3
Trace Metals Protocol 4
Trace Metals Protocol 5
Trace Metals Protocol 6
Trace Metals Protocol 7
Trace Metals Protocol 8
Total Cyanide Protocol
Phenolics Protocol
Trace Metals Sampling Strategy
Total Cyanide. Sampling Strategy
Phenolics Sampling Strategy
Description of SIC Codes
Edit 1 Codes
Edit 2 Error Codes
Silver Loss During Mild Digestion
Precision Study Cyanides
Precision Study Phenolics
VII
-------
ACKNOWLEDGEMENTS
This research effort was conducted under the sponsorship of
the U.S. Environmental Protection Agency, Environmental Monitor-
ing and Support Laboratory (EMSL), Cincinnati, Ohio. Gerald
McKee served as Project Director. Significant contributions were
also made by Ted Martin and John Pfaff.
This effort was conducted for the Environmental Protection
Agency by Biospherics Incorporated under the direction of Mr.
Gary J. Gottfried and Margaret Federline. The following members
of Biospherics' scientific staff were responsible for the per-
formance of the analytical program;
H. Markus Gudnuson, Bench Supervisor, Trace Metals
Frank Fitzpatrick, Project Leader
. John Cooper, Analytical Chemist
Elizabeth DeBelius, Analytical Chemist
William Henson, Analytical Chemist
Dale Watson, Analytical Chemist
Joyce Burgess, Analytical Chemist
Michael Dodds, Quality Control Officer
SCS Engineers, under subcontract to Biospherics, performed
the sampling and data management phases of this effort under the
direction of Mr. E. T. Conrod, SCS Project Director. The follow-
ing individuals of SCS staff made significant contributions to
this effort:
Michael McLaughlin, Project Manager
Diana Fraser, Project Scientist
Steven Stinger, Field Leader
Miles Haven, Data Processing Director
Special acknowledgement and gratitude is extended to the
participating trade associations and industrial facilities whose
gracious cooperation helped assure the success of this project.
viii
-------
SECTION 1
INTRODUCTION
Section 304(h) of the Federal Water Pollution Control Act of
1972 requires that the Administrator of the Environmental Protec-
tion Agency promulgate guidelines establishing test procedures
for the quantification of pollutants in industrial effluents. In
this project, state-of-the-art analytical procedures which have
been approved for trace metals, total cyanide, and phenolics
effluent compliance monitoring were systematically evaluated in a
wide variety of industrial effluents to assess their industry-
wide applicability. Several procedures for these parameters
which have not been officially approved but which show promise
were also evaluated.
The state-of-the-art analytical procedures under study have
generally been found to provide reliable data. However, all ana-
lytical procedures can be rendered ineffective by interferences
which may be present in effluent matrices. The relative presence
of these matrix interferences can vary markedly across industry
in general and are affected by various wastewater treatment
processes. In order to investigate the applicability of these
procedures to industry in general, a wide variety of effluents
representing numerous industries and industrial processes were
sampled and analyzed. Procedures for metals, total cyanide and
phenolics were evaluated for effluents from 24, 8 and 10 Standard
Industrial Classifications (SIC), respectively, and within each
of these classifications, multiple processes were sampled when-
ever possible. SIC Codes investigated are identified in Table 1.
Samples were collected at sites which were permitted or antici-
pated permits to discharge the contaminant of interest. Thus, an
evaluation of the analytical procedures on representative compli-
ance samples was achieved.
Atomic absorption analytical methods were evaluated for
trace metal determinations. Sample preparation, direct flame
aspiration, chelation/extraction, graphite furnace, hydride gen-
eration, and mercury cold vapor techniques of atomic absorption
spectroscopy were included in this investigation. Two colori-
metric methods as well as an instrumental method were evaluated
for phenolic determinations. A titrimetric and two colorimetric
methods for total cyanide were also evaluated.
-1-
-------
TABLE 1. DESCRIPTION OF STANDARD INDUSTRIAL CLASSIFICATION CODES
Standard Industrial
Classification Codes
1061
1094
1099
2231
2262
2269
2621
2812
2819
2821
2822
2823
2879
2911
3229
3312
3313
3315
3331
Descr iption
Ferroalloy Ores, Except Vanadium
Uranium-Radium-Vanadium Ores
Metal Ores, Not Elsewhere Classified
Broad Woven Fabric Mills, Wool (Including
Dying and Finishing)
Finishers of Broad Woven Fabrics of
Man-Made Fiber and Silk
Finishers of Textiles, Not Elsewhere
Classified
Paper Mills, Except Building Paper Mills
Alkalies and Chlorine
Industrial Inorganic Chemicals, Not
Elsewhere Classified
Plastics Materials, Synthetic Resins, and
Non-vulcanizable Elastomers
Synthetic Rubber {Vulcanizable Elastomers)
Cellulosic Man-Made Fibers
Pesticides and Agricultural Chemicals, Not
Elsewhere Classified
Petroluem Refining
Pressed and Blown Glass and Glassware, Not
Elsewhere Classified
Blast Furnaces (Including Coke Ovens),
Steel Works, and Rolling Mills
Electrometallurgical Products
Steel Wire Drawing and Steel Nail Spikes
Primary Smelting and Refining of Copper
-2-
-------
TABLE 1. DESCRIPTION OF STANDARD INDUSTRIAL CLASSIFICATION CODES
(continued)
Standard Industrial
Classification Codes
3332
3333
3341
3471
3861
4911
9999
Descr iption
Primary Smelting and Refining of Lead
Primary Smelting and Refining of Zinc
Secondary Smelting and Refining of
Nonferrous Metals
Electroplating, Plating, Polishing,
Anodizing and Coloring
Photographic Equipment and Supplies
Electric Services
Mixed Domestic and Industrial Sewage
Effluent (MDISE)
-3-
-------
Interferences, both positive and negative, as well as other
procedural limitations, were investigated through a series of
sample dilutions, standard additions, and multiple determina-
tions. Preservation techniques for total cyanide and phenolics
were also evaluated.
-4-
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
TRACE METALS
Silver
Effluents from SIC 2819 and 3861 as well as mixed domestic
and industrial sewage effluent {MDISE) were sampled to evaluate
applicability of atomic absorption methodologies for total
silver.
Silver was lost during the mild acid digestion (MAD)(1) in
all samples including quality control samples prepared in deion-
ized water. Apparently, silver precipitated or plated out on the
walls of digestion beakers. This problem was overcome by dilut-
ing up the mild digest with cyanogen iodide (CNI) and ammonium
hydroxide (NH^0H)(2) which redissolved the precipitated silver
resulting in good recoveries. The vigorous acid digestion
(VAD){3) was effective without requiring CNI. Equivalent results
were achieved for the MAD using CNI and the VAD. Care should be
taken to use the appropriate digestion protocol when analyzing
for silver along with other metals. The effect of using CNI
although positive for silver is unknown for other metals. Thus,
if a common digest is to be analyzed for additional metals, the
VAD should be used. Alternately, separate digestions are recom-
mended. A digestion step could be eliminated for total silver
determinatins by simply adding CNI/NH^OH to a sample aliquot
previously preserved in nitric acid. The entire sample could be
preserved with CNI if only silver determinations were
required(2). Testing to determine the applicability of substi-
tuting CNI for an acid preservative and digestion should be
performed prior to the adoption of this technique.
No significant interferences were detected for direct aspi-
ration analyses in SIC 3861 and MDISE. However, severe negative
interferences were encountered in all SIC 2819 effluents result-
ing in silver recoveries generally below 30%. This loss of
silver appeared to occur during the MAD (even though CNI was
used) since spikes of the digest were recovered to nearly 100%.
SIC 2819 effluents contained higher levels of dissolved salts,
suspended solids, and organic matter than most other samples
investigated.
-5-
-------
Positive interferences were identified for graphite furnace
analyses of samples in SIC 3861. These interferences were effec-
tively eliminated by diluting the digested matrix. No signifi-
cant interferences were detected for graphite furnace analysis of
SIC 2819 or MDISE samples.
Two chelation/extration procedures produced recoveries
between 70% and 130% for 21 of 28 sample spikes. The ammonium
pyrrolidine dithiocarbamate/methyl isobutyl ketone ( APDC./MIBK) ( 4 )
procedure generally showed lower silver recoveries than the pyr-
rolidine dithiocarbamic acid/chloroform (PDCA/CHC1,)(5)
procedure.
Aluminum
Effluents from SIC 3341 and MDISE were analyzed to evaluate
direct aspiration and graphite furnace atomic absorption proce-
dures for total aluminum. The MAD and VAD produced equivalent
results for all samples analyzed. No significant interferences
were identified for direct aspiration or graphite furnace
techniques.
Bar ium
Effluents for SIC 2819 and MDISE were analyzed for total
barium by direct aspiration and graphite furnace atomic absorp-
tion procedures.
Equivalent results were achieved for the MAD and the VAD.
Direct aspiration analyses for MDISE samples produced good
recoveries for all sample spikes. Severe negative interferences
were encountered for five of nine effluent sites in SIC 2819. As
stated previously, SIC 2819 samples were relatively high in dis-
solved salts, solids, and organic matter.
SIC 2819 samples also exhibited interferences, both positive
and negative, for graphite furnace analyses. No inerferences
were noted in MDISE samples.
Cadmium
Samples from SIC 1061, 1094, 1099, 3331, 3332, 3333 and 3471
as well as MDISE were analyzed for total cadmium by atomic
absorption methods to evaluate their applicability.
Equivalent results were achieved for the MAD and the VAD.
Of 54 samples analyzed by direct aspiration only two in SIC
3471 exhibited interference effects. These positive interfer-
ences were largely reduced by 1:1 dilutions.
-6-
-------
The graphite furnace technique yielded good recoveries for
all samples with the exception of those from SIC 3333 (minor
positive interferences) and MDISE (moderate negative interfer-
ences). These interferences were partially or totally eliminated
by 1:2 dilutions.
The PDCA/CHC1^ chelation/extraction technique exhibited no
interferences in arry SIC code. The APDC/MIBK technique exhibited
a positive interference in samples from SIC 3333. This interfer-
ence was largely eliminated by a 1:2 dilution. The APDC/MIBK
technique was successful for all other samples.
Chromium
Samples from SIC 2231, 2262, 2269, 2819, 2821, 2911, 3312,
3313, 3471 and MDISE were analyzed for total chromium by atomic
absorption techniques. Samples from SIC 3312 and 3315 were
analyzed for dissolved chromium. Samples from SIC 2819, 2911,
3312, 3313 and 3471 were analyzed for hexavalent chromium by the
two chelation/extraction procedures and precipitation/graphite
furnace technique.(6)
In preparing samples for total chromium analysis, signifi-
cant chromium was lost (>50%) during the VAD from samples con-
taining <0.5 mg/1 in SIC 2911 and MDISE. The percentage loss
relative to the MAD decreased for higher concentrations. The
apparent loss of chromium could have been due to a negative
interference enhanced by the VAD rather than actual chromium
loss. Equivalent chromium results for the VAD and MAD were
achieved at all concentrations for other SIC Codes.
No significant interferences were noted in any sample from
any SIC Code analyzed for total chromium by direct aspiration.
The graphite furnace technique for total chromium was
successful on 22 of 25 samples analyzed. Negative interferences
were identified in one sample from SIC 2231 (6% recovery) and two
samples from SIC 2269 (68% and 54% recoveries). Sample dilutions
(1:2) improved these recoveries to acceptable limits.
The 30-minute permanganate digestion time preceding chela-
tion/extraction techniques for total chromium was found insuffi-
cient to convert trivalent chromium to the hexavalent state. The
digestion time was increased to 4 hours resulting in complete
oxidation and thus good extraction efficiencies. Many samples
contained matrices which destroyed permanganate resulting in a
heavy brown precipitate. Large amounts of permanganate were thus
required for some digestions. Often it was difficult to deter-
mine if an excess of permanganate was present due to the dark
precipitate. A 4-hour digestion time probably incorporated a
large safety margin and can probably be significantly reduced for
-7-
-------
most samples. The digestion efficiency should be checked with
trivalent chromium sample spikes and reference analyses to verify
that the digestion time used is sufficient for oxidation to
occur.
The APDC/MIBK chelation/extraction method for total chromium
yielded recoveries between 70% and 130% for 39 of 46 sample
spikes. Only occasional positive interferences were encountered.
The PDCA/CHC1., method produced 70% - 130% recoveries for 43 of 46
sample spikeS. A 4-hour digestion time was used for all
analyses.
All methods produced satisfactory results for all samples
analyzed for dissolved chromium.
Interferences affecting hexavalent chromium procedures were
present in a small percentage of samples in SIC 2819, 2911, 3312,
3313 and 3471. These interferences were specific to particular
sites or samples and not widespread throughout the industries.
Interferences in SIC 2819 affected all three methods (APDC/MIBK,
PDCA/CHC1.., furnace), however, in all other cases their effects
were limfted to either both chelation/extraction procedures or
the furnace method. For samples where interferences were not
encountered, the furnace technique produced better recoveries
than either chelation/extraction method. In addition, the
detection limit of the furnace method is 5X better than the
chelation/extraction methods.
Interferences occurring in both chelation/extraction pro-
cedures were often visually noted by the formation of a pro-
nounced black precipitate upon chelate addition. Formation of
this extractable precipitate can be used to predict interferring
effects for hexavalent chromium and other trace metals. Abnor-
malities in the normal hexavalent chromium precipitation pattern
used in conjunction with the furnace procedure were also noticed
in some samples producing interferences. For instance, a heavy
precipitate was formed in some SIC 2819 samples when lead nitrate
was added to precipitate the hexavalent chromium. This precipi-
tation could indicate that the lead nitrate was consumed before
hexavalent chromium could be precipitated. Thus, these observa-
tions can be used to flag potential methodology problems.
The furnace procedure for hexavalent chromium is the most
cost-effective method from a labor and materials basis.
Copper
Effluents from SIC Codes 1061, 1099, 2819, 3331, 3341, 4911
and MDISE were analyzed by atomic absorption procedures for total
copper. Samples from SIC 3315 were analyzed for dissolved
copper.
-8-
-------
The VAD produced up to 40% higher total copper results than
the MAD for samples from SIC 1061 and 2819. The digestion pro-
cedures correlated to within + 5% for all other samples contain-
ing significant amounts of copper.
No significant interferences were identified for any sample
analyzed for total copper by direct aspiration techniques.
Of 24 samples analyzed for total copper by graphite furnace,
23 produced copper spike recoveries between 70% and 130%.
Each chelation extraction procedure yielded total copper
recoveries between 70% and 130% for 43 of 48 sample spikes.
Good recoveries were achieved for SIC 3315 samples analyzed
for dissolved copper by direct aspiration and chelation/ extrac-
tion procedures. Positive interferences affected graphite
furnace analyses resulting in 150% recoveries.
Iron
Effluents from SIC 1099, 2819, 4911 and MDISE were analyzed
for total iron by atomic absorption methods. SIC 3312 and 3315
samples were analyzed for dissolved iron.
The VAD often gave 10% - 40% higher total iron results than
the MAD, particularly for samples high in suspended solids such
as those from SIC 4911.
Direct aspiration analyses for total iron produced recov-
eries between 70% and 130% for 180 of 190 sample spikes. SIC
2819 samples which contained higher levels of particulate matter
showed relatively poorer recoveries than other samples.
In all SIC Codes 26 of 36 spikes were recovered between 70%
and 130% when analyzed for total iron by the graphite furnace
technique. Poor recoveries appeared to be due to variable
efficiencies of the MAD used in conjunction with the furnace
procedure (See Section 7, Discussion).
The APDC/MIBK chelation/extraction procedure for total iron
produced 24 of 36 spike recoveries in the 70% - 130% range. The
PDCA/CHCl^ procedure yielded 27 of 36 spikes recovered in the
same range.
No interferences were identified in any sample analyzed for
dissolved iron by the direct aspiration and chelation/extraction
procedures as 65 of 66 spikes were recovered between 85% and
115%. Success of these dissolved iron determinations confirmed
the digestion step to be the deciding factor for total iron
analytical success.
-9-
-------
Graphite furnace analyses for dissolved iron were affected
by negative interferences in some SIC 3312 and 3315 samples.
These interferences were eliminated by 1:2 dilutons.
Iron is a common element of particulate matter in many
industrial effluents. Consequently, the precision of sample ali-
quotting and MAD efficiencies is a greater limitation for iron
determinations than for other elements typically in the soluble
state. Many apparent interferences affecting total iron were due
to these limiting factors rather than actual matrix
interferences.
Mercury
Samples from SIC 1099 and 2812 were analyzed by the cold
vapor atomic absorption technique for total mercury.
Recoveries between 70% and 130% were achieved for 23 of 36
sample spikes. SIC 2812 samples from 3 of 8 sites produced
recoveries below 55% (see Section 7, Discussion).
Manganese
Samples from SIC 3312, 3313 and MDISE were analyzed by
atomic absorption methods for total manganese.
The VAD gave up to 27% higher results than the MAD, however,
most results agreed to within + 10% or better.
No significant interferences were identified for direct
aspiration, graphite furnace, or PDCA/CHCl- chelation/extraction
techniques.
A standard curve could not be achieved for manganese deter-
minations by the APDC/MIBK chelation/extraction method. Appar-
ently the manganese complex was not sufficiently stable. Samples
and standards were run within 30-60 minutes of chelation, thus,
the complex, if formed, breaks quickly as no absorption was noted
during aspiration.
Nickel
Effluents from SIC 1099, 2819, 3471 and MDISE were analyzed
for total nickel by atomic absorption methods. Samples from SIC
3312 and 3315 were analyzed for dissolved nickel.
The VAD and MAD produced equivalent results.
Flame conditions were critical for total nickel determina-
tions in the direct aspiration technique. An oxidizing flame
must be maintained as recoveries of 170% were noted for quality
-10-
-------
control samples when using even a slightly rich flame. With an
oxidizing flame all 72 spikes were recovered between 70% and 130%
with 70 of 72 recovered between 85% and 115%. Nickel was the
only metal which was consistently and successfully determined on
SIC 2819 samples.
Graphite furnace analyses for total nickel produced 16 of 18
spike recoveries between 70% and 130%.
No interferences were identified for either chelation/
extraction technique for total nickel.
No significant interferences affecting dissolved nickel
results were identified for direct aspiration or the PDCA/CHC1,
chelation/extraction technique. One of five samples exhibited a
moderate negative interference by the graphite furnace procedure.
The APDC/ MIBK procedure was successful on nine of ten samples
with one sample showing a moderate negative interference effect.
Lead
Effluents from SIC 1061, 1099, 2812, 3229, 3312, 3332 and
MDISE were analyzed for total lead by atomic absorption
procedures.
The VAD and MAD produced equivalent results.
All 144 spikes in the seven industrial categories analyzed
by direct aspiration were recovered between 85% and 115%. The
furnace procedure produced 26 of 32 recoveries between 70% and
130%. The PDCA/CHCl- and APDC/MIBK chelation/extraction proce-
dures produced 31 of 32 and 28 of 32 recoveries in the 70% to
130% range, respectively.
Only SIC 1099 samples were problematic for furnace and
chelation/extraction procedures as negative interferences pro-
duced low recoveries. Dilutions remedied difficulties for the
chelation/extraction methods, however, furnace recoveries
remained below 30%.
Tin
Effluents from SIC 3312, 3471 and MDISE were analyzed for
total tin by direct aspiration and graphite furnace atomic
absorption procedures.
The VAD and the MAD produced equivalent results at flame
levels (>0.5 mg/1) when hydrochloric acid (HC1) was used.
Without HC1, precipitation of tin occurred in both reference and
effluent samples resulting in low recoveries. Thus, HCl must be
-11-
-------
utilized in digestions for tin determinations. Tin, when present
at furnace levels (<0.1 mg/1), was lost during the MAD even with
2% HCl. The VAD was not studied at furnace analytical levels.
By using HCl in the MAD all 64 sample spikes were recovered
between 70% and 130% with 62 of 64 recovered between 85% and
115%.
Furnace analyses produced consistently low recoveries due to
tin precipitation in the MAD. Unacceptable recoveries resulted
for both reference and effluent samples. Thus, losses could not
be attributed to matrix effects. The applicability of the VAD
should be studied at furnace levels, and if problems exist an
alternate digestion procedure should be developed for tin.
Z inc
Effluents from SIC 1061, 1094, 2621, 2823, 3312, 3331, 3332,
3333, 3341 and MDISE were analyzed for total zinc by atomic
absorption procedures.
The VAD yielded results lower than the MAD for SIC 1094
samples. The methods were equivalent for all other samples.
Direct aspiration analyses were quite successful, producing
recoveries between 70% and 130% for 164 of 172 spikes.
Graphite furnace determinations for zinc were susceptable to
severe laboratory contamination problems. The extreme sensitiv-
ity of the method and the prevalence of zinc in the laboratory
environment necessitated the use of extraordinarily meticulous
analytical technique to achieve reliable results. The relative
abundance of zinc in reagents, glassware and other experimental
materials created the potential for severe zinc contamination
problems by the furnace method. Meticulous cleaning procedures
and reagent purification was applied for low level zinc concen-
trations with limited success. The potential for introduction of
gross contamination during sample injection into the furnace was
high. Automated sample injection introduced unacceptable zinc
contamination into the furnace. Manual injection when carefully
controlled produced adequate results for most samples, however,
occasional contamination during injection was unavoidable.
Contamination was also a problem for zinc by the chelation/
extraction methods.
It is doubtful that zinc would ever have to be detected in
effluents at che1 ation/extraction and furnace levels (0.5-5
ug/1). Therefore, direct aspiration which is not severely
-12-
-------
affected by zinc contamination problems is recommended for
routine analyses. If the furnace technique is used, for special-
ized analyses when low detection limits are required, manual
furnace injection should be used instead of auto-injection which
was a high contributor to contamination. The analyst's technique
should be refined and investigated to ensure proper furnace
injection procedures.
Arsenic
Effluents from SIC 1061, 1094, 2819, 3331, 3333 and MDISE
were analyzed for total arsenic by graphite furnace as well as
sodium borohydride and zinc slurry hydride generation techniques
of atomic absorption spectrophotometry.
The graphite furnace technique using nickel as a matrix
modifier produced 89 of 108 spike recoveries in the 70% to 130%
range. Moderate positive interferences affected SIC 3333 sample
analyses and severe negative interferences were encountered in
SIC 2819 samples. Despite these interferences, the furnace
method produced superior results to either hydride generation
method.
The sodium borohydride hydride generation technique produced
recoveries between 70% and 130% for SIC 1061, 3331, 3333 and
MDISE samples. Positive interferences were encountered in SIC
1094 and severe negative interferences were again encountered in
SIC 2819.
The zinc slurry method of hydride generation was limited by
losses in the digestion procedure, inconsistencies of the stand-
ard curves, and contamination of zinc powder. Poor resutls were
obtained for all SIC Codes.
Selenium
Effluents from SIC 2819, 3331, 3333 and MDISE were analyzed
for total selenium by graphite furnace as well as sodium boro-
hydride and zinc slurry hydride generation techniques of atomic
absorption spectrophotometry.
The graphite furnace technique proved quite successful for
SIC 3331, 3333 and MDISE samples yielding 33 of 36 spike recov-
eries in the 70% to 130% range. However, as with most other
metals, severe negative interferences were encountered for SIC
2819 samples.
The sodium borohydride hydride generation technique also
produced acceptable results for all samples except those of SIC
2819.
-13-
-------
Similarly as for arsenic, the zinc slurry method produced
poor recoveries for the majority of samples.
General
Effluents from SIC 2819 were the most problematic of any
industrial category sampled. Interferences affected chromium,
hexavalent chromium, iron, silver, copper, barium, arsenic and
selenium determinations. Only nickel was determined with con-
sistent reliability. Relatively high levels of dissolved and
suspended solids as well as organic matter in samples from this
category were probably responsible for method problems.
The MAD or the VAD can be used interchangeably for most
samples and elements. The VAD should be more effective for
samples high in particulate iron. The MAD is quicker, less labor
intensive, and not susceptible to VAD errors introduced by sample
spattering or baking when approaching dryness.
The formation of a black precipitate upon chelate addition
should be used as an indicator of possible negative interferences
for the chelation/extraction procedures. Sample spikes should be
taken through the procedure to investigate the interference
effect. If the interference is significant a sample dilution
should be made and the chelation/extraction performed again with
and without spikes. An alternate method (direct aspiration or
furnace) should be used if problems cannot be overcome by reason-
able dilutions.
With the exception of a few techniques (zinc slurry for
arsenic and selenium, APDC/MIBK for manganese, MAD for tin) each
of the analytical procedures is capable of delivering reliable
results for industrial effluent monitoring. However, each method
also has certain limitations and can be adversely affected by
specific matrix interferences. Several analytical options are
available for most elements. The method of choice will depend on
required detection limits, susceptibility to interferences,
cost-effectiveness, complexity, and other factors. Preferably,
the method chosen should be the simplest one available which
produces reliable results and meets detection limit requirements.
The method, therefore, will be cost-effective and reliable. For
example, the direct aspiration technique is least susceptible to
interferences and is also the quickest and most cost-effective
method available. Its levels of detection are generally poorer
than other methods, but if they are adequate to meet permit
guidelines, direct aspiration would be the method of choice.
Evaluating analytical methods as performed by use of refer-
ence samples and spikes should be used as a routine quality
control check for analytical programs.
-14-
-------
CYANIDES
Total cyanide was determined by the pyridine barbituric acid
(PBA) colorimetric method, the pyridine pyrazalone colorimetric
(PPC)(7) method and the silver nitrate titrimetric (TTM)(7)
method on samples from SIC 1099, 2819, 3312, 3313, 3315, 3471,
3861 and MDISE. All analyses were preceded by the sulfuric
acid/magnesium chloride distillation procedure.(7)
The PBA procedure produced recoveries between 70% and 130%
for 136 of 164 sample spikes. Interferences were identified in a
small percentage of samples from SIC 2819, 3312, 3313 and 3315.
The PBA and the PPC procedures produced equivalent results
for sample concentrations and precision. Both methods were
affected similarly by the few interferences which were encoun-
tered. These methods can thus be used interchangeably.
The TTM procedure also proved effective showing good pre-
cision and spike recoveries. Negative interferences were iden-
tified in one SIC 3312 sample.
Samples preserved with sodium hydroxide (NaOH) and stored at
4 C showed no significant loss of cyanide over a 10-day study
period. However, only a limited data base from two SIC Codes was
accumulated.
PHENOLICS
Phenolics were determined by the 4-amino antipyrine (4-AAP)
colorimetric method, the 3-methyl-2-benzothiazolinone hydrazone
hydrochloride (MBTH)(9) colorimetric method and the PH-3 instru-
mental method (PHI)(10) on samples from SIC 2231, 2262, 2491,
2821, 2879, 2911, 3312, 3313, 3861 and MDISE. All methods were
preceded by a distillation.(8)
The 4-AAP colorimetric method produced recoveries in the 70%
- 130% range for 186 of 256 sample spikes. The 4-AAP method was
less susceptible to interferences than the MBTH colorimetric or
the PHI instrumental method.
Significant interference effects were encountered in all SIC
Codes for MBTH as only 112 of 256 spikes were recovered in the
70% - 130% range. A green or blue colored complex possibly due
to aldehyde reactions often formed instead of the characteristic
pink phenolic complex.
The PHI method produced only 64 of 256 spike recoveries in
the 70% to 130% range. This method was severely affected by
-15-
-------
large background readings relative to actual phenolic response.
These background levels were not reproducible between distilla-
tions which further impacted reliability. An alternate sample
preparation method should be developed for use with the phenol
instrument. Solvent extraction followed by back extraction into
water may be a possibility.
The three methods did not produce equivalent results. PHI
phenolic determinations were often an order of magnitude greater
than the colorimetric methods due to PHI positive interferences
as well as limitations on the phenolic reactivity of 4-AAP and
MBTH. When interferences were absent the MBTH method generally
produced higher results than 4-AAP, thus MBTH is probably reac-
tive with a greater variety of phenolic compounds than 4-AAP.
Method precision and interference effects also were not equiva-
lent for the three methods.
No significant loss of phenolics occurred over a 10-day
preservation period for eight of nine samples from different SIC
Codes preserved with phosphoric acid or sulfuric acid and copper
sulfate. The remaining sample produced variable results due to
an interfering matrix. Small phenolic increases with time
occurred for some samples possibly due to changes in phenolic
substitution increasing reactivity with 4-AAP. Based on these
results from eight industrial effluents, phenolic samples which
have been preserved with phosphoric or sulfuric acid and copper
sulfate and refrigerated to 4 C can be stored for at least 10
days prior to analysis.
Ferrous ammonium sulfate used to destroy oxidizing agents
was ineffective in removing the final traces from the effluent
samples. Excesses also produced a yellow tint to some distil-
lates which posed a potential interference with the 4-AAP yellow
complex. An alternate substance, possibly ascorbic acid, should
be investigated as a substitute for ferrous ammonium sulfate.
-1.6-
-------
SECTION 3
ANALYTICAL PROGRAM
Tables 2 and 3 summarize analytical procedures which were
evaluated in this project. Each procedure was systematically
evaluated through a series of spikes, dilutions and replicate
analyses to investigate precision, accuracy, equivalency, and
susceptability to positive and negative interferences. The
evaluation protocols for all parameters are described below.
TRACE METALS
Methods of Atomic Absorption Spectroscopy (AAS) were evalu-
ated for the determination of trace metals. The analytical
protocols shown in Tables 4-11 at the end of this section were
completed for each AAS method. The basic analytical protocol for
AAS procedures is summarized as follows: (a) initial analyses
were performed in duplicate, (b) duplicate sample aliquots were
spiked prior to processing (digesting, etc.) processed and
analyzed, (c) duplicates from the original processed sample were
combined, diluted in duplicate, and analyzed, (d) duplicate
sample aliquots from the combined duplicates of the original
processed sample were diluted, spiked and analyzed.
The basic analytical protocol was designed to:
o identify samples with interferring matrices and identify
techniques susceptible to interferences
o evaluate recoveries from sample preparation techniques
o evaluate precision of the analytical technique
Matrix interferences can be identified from spike recoveries
and confirmed by dilution analyses. A low or high recovery of
undiluted sample spikes may indicate the presence of a negative
or positive interference. The possible presence of an interfer-
ence can be confirmed by comparing correlation of the diluted
analysis to the original undiluted analysis since interferences
seldom reflect the actual dilution factor. The spike of the
dilution can further confirm the presence of an interference
-17-
-------
Table 2. Trace Metal* Procedures
Atonic Absorption
Techniques
A1 A£ As Ba Cd Cr V£ Cu Pe Pb
Cr
VI
Nn
Ni
Se
Sn
riaaw Aaplration X X
Vi9orous Digestion X X
Hi Id Digestion X X
Filtration
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
CD
I
Chelation Extraction X
Vigorous Digestion X
Fi1tration
Permanqanate Digestion
PDCA/CHCl- X
APDC/MIBK 3 X
Graphite Furnace X X
Nild Digestion X X
Peroxide Digestion
Chromate Precipitation
Hydride Generation
Zn Slurry/Sulfuric Digestion
NaBH^/Permanganate Digestion
Cold Vapor
Permanganate Digestion
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Digestions - Total Metals
Filtration - Dissolved Metals
Reference - Methods for Chemical Analyses of Water and Wastes, F.PA-600 4-79-020, 1979.
-------
CYANIDE AND PHENOLIC PROCEDURES
Analytical Method
Sample Prep - H2S04, MgCl2 Reflux
Distillation, NaOH trapping (7)
Analysis by Pyridine Pyrazolone Colorimetrid7
Pyridine Barbituric Acid (7)
Colorimetric
Silver Nitrate Titration (7)
Sample Prep - Distillation
Analysis by 4 - Amino Antipyrine Colorimetric
(4-AAP) (8)
3 - Methyl-2-Benzothiazolinor.e
Hydrazone Hydrochloride
Colorimetric (MBTH) (9)
PH-3 Instrumental (10)
-19-
-------
since the diluted spike should be recovered to the same degree or
closer to 100% than the original undiluted spike. An example of
this logic is expressed below:
o original analysis yields 10 mg/1
o 10 mg/1 spike into the undiluted sample yields 15 mg/1 or
only a 50% recovery of the spike, thus a possible nega-
tive interference exists
o 1:1 dilution of original processed sample yields 7 mg/1
equivalent to 14 mg/1 when the dilution factor is
applied. Negative interference has been confirmed and at
least partially diluted out
o 15 mg/1 spike of the diluted aliquot yields 20 mg/1 for a
recovery of 87%: 20-7 = 13 as recovery of the spike
(13/15)(100) = 87% recovery. Matrix interference was
diluted out to yield higher recovery.
The sample processing or preparation technique can be evalu-
ated from spikes of the original undiluted sample. A low recov-
ery from spiked original samples may indicate loss in the sample
preparation process particularly if the correlation of diluted
sample values to original sample values is good and the spike
recovery of the diluted processed sample shows no sign of a
negative interference.
The precision of the analytical technique can be estimated
based on duplicate ranges. Precision as a function of SIC Code,
industrial process, and concentration can be evaluated.
The effectiveness of a mild (MAD) versus a vigorous (VAD)
acid digestion were compared by flame aspiration. Different
analytical techniques used to measure common analytes were evalu-
ated for equivalency based on the original analysis, relative
precision and susceptability to interferences.
Each method was evaluated for precision by analyzing seven
replicates from samples in a detectable analytical range.
CYANIDES
The analytical protocol for total cyanide is shown in Table
12. The pyridine barbituric acid procedure(7) was employed for
samples having less than 1 mg/1 cyanide and the titrimetric
method(7) on samples greater than 1 mg/1. The pyridine pyrazo-
lone method(7) was used on all spiked 1:1 dilutions under 1 mg/1
to determine method equivalency. Duplicate distillates were run
-20-
-------
for all samples throughout the analytical scheme. Samples were
analyzed by the appropriate procedure as received with and with-
out spike, and diluted 1:1 with and without spike. Interferences
were detected by computing spiked recoveries of original and
diluted samples and by comparing results from sample dilutions
with undiluted samples.
Precision studies were performed on several samples by
multiple determinations. The possibility of increasing holding
times was evaluated by performing a series of 10-day preserva-
tions studies.
PHENOLICS
Analytical protocols for phenolics are shown in Table 13.
The 4-AAP(8) and instrumental methods(lO) were run on common
distillates from samples preserved with phosphoric acid. The
MBTH method(9) was run on samples preserved with sulfuric acid.
Duplicate analyses were performed on common distillates rather
than single analyses performed on duplicate distillates. When
extraction procedures were performed, 200 ml aliquots diluted to
500 ml were taken for analysis. Samples were analyzed as
received with and without spikes, and diluted 1:1 with and with-
out spikes.
Precisison and preservation studies were run similarly as
for cyanides.
-21-
-------
TABLE 4. TRACE METALS PROTOCOL 1
Flame Aspiration/MAD with HC1
1. Digest original preserved samples in duplicate. Analyze
each duplicate.
2. Spike original preserved samples in duplicate for each metal
to 1.5X or 2X original concentration, digest and analyze.
3. Dilute combined digests from Step 1 in duplicate 1:1 with
deionized water acidified to same concentration as sample
aliquot and analyze.
4. Dilute combined digest as in Step 3 in duplicate adding a
spike for each metal to achieve final concentration as in
Step 2 and analyze.
NOTES: Follow similar protocol for dissolved metals substituting
filtrations for digestions.
If metals not detected spike to 10-15X 1% absorption.
TABLE 5. TRACE METALS PROTOCOL 2
Flame Aspiration/VAD
1. Based on results of Protocol 1, rank sites within SIC from
highest to lowest concentration for each metal using the
highest sample concentration at each site. For SIC's with
only 1 site, rank samples from high to low concentration for
each metal.
2. Based on the ranking in Step 1, perform VAD in duplicate on
the higher one-third of samples (a minimum of 2).
3. Analyze by flame aspiration.
NOTES: Eliminate this protocol for dissolved metals.
-22-
-------
TABLE 6. TRACE METALS PROTOCOL 3
Graphite Furnace/MAD without HC1
1. Rank sites within each SIC similarly as in Protocol 2 but
based on lowest sample concentrations at each site.
2. Based on ranking in Step 1, perform MAD in duplicate on the
lower one-third of samples {a minimum of 2) and analyze.
Digest and analyze all arsenic and selenium samples using
peroxide digestion.
3. Spike the preserved samples to 1.5X or 2X the original con-
centration for each metal. Digest in duplicate and analyze.
4. Combine the duplicate digests in Step 2 and dilute in dupli-
cate 1:2 with 0.5% nitric acid in deionized water and
analyze.
5. Dilute the combined digests as in Step 4 in duplicate adding
a spike for each metal to achieve a final concentration as
in Step 3 and analyze.
NOTES: Follow similar protocol for dissolved metals substituting
filtrations for digestions.
Follow similar protocol for hexavalent chromium using the
chromate precipitation pretreatment. Analyze all
samples.
If metals not detected spike to 10-15X 1% absorption.
HC1 was used for tin in the MAD.
-23-
-------
TABLE 7. TRACE METALS PROTOCOL 4
Chelation/Extraction/VAD
1. Use the VAD and run in duplicate the original preserved
samples chosen in Protocol 3. For total chromium use the
permanganate digestion. For hexavalent chromium, filter
only and analyze all samples.
2. Extract each duplicate by APDC/MIBK and PDCA/CHCl^ and
analyze extracts.
3. Spike the preserved samples to 1.5X or 2X the original con-
centration of each metal. Digest in duplicate, extract by
both methods and analyze.
4. Dilute combined digests from Step 1 in duplicate 1:2 with
deionized water acidified to same concentration as sample
aliquot. Extract by both methods and analyze.
5. Dilute combined digests as in Step 4 spiking each metal to
final concentration in Step 3. Extract by both methods and
analyze.
NOTES: Follow similar protocol for dissolved metals substituting
filtration for digestions.
TABLE 8. TRACE METALS PROTOCOL 5
Zinc Slurry Hydride Generation
Digest original preserved samples in duplicate with
sulfuric/nitric acids and analyze by zinc slurry hydride
generation.
Spike original preserved samples to 2X or 1.5X original
concentration. Digest in duplicate and analyze.
Dilute combined digests in Step 1 in duplicate 1:1 with acid
diluent and analyze.
Dilute as in Step 3 in duplicate adding spike to achieve a
final concentration as in Step 2, and analyze.
-24-
-------
1
2
1
2
3
4
1
2
3
TABLE 9. TRACE METALS PROTOCOL 6
NaBH^ Hydride Generation
Digest one preserved sample in duplicate from each site
requiring As or Se by permanganate digestion and analyze.
Digest in duplicate a 1.5X or 2X spike of that sample and
analyze. If originally undetected spike to 0.01 mg/1.
TABLE 10. TRACE METALS PROTOCOL 7
Mercury Cold Vapor
Digest preserved samples in duplicate and analyze.
Spike samples to 1.5X or 2X the original concentration,
digest in duplicate, and analyze. Spike to 0.002 mg/1 if
undetected.
Dilute combined digests in Step 1 in duplicate 1:1 with
deionized water acidified similarly as the sample, and
analyze.
Dilute combined digest in duplicate as in Step 3 adding a
spike to achieve a final concentration as in Step 2.
TABLE 11. TRACE METALS PROTOCOL 8
Precision Statement
Flame Aspiration/MAD-Analyze seven replicates for three
samples for each element at different concentrations.
Zinc Slurry Hydride Generation; Graphite Furnace; Cold
Vapor; Hexavalent Chromium Extractions-Analyze seven
replicates for two samples at different concentrations.
Extraction Methods for other metals. Analyze seven
replicates from one sample.
-25-
-------
TABLE 12. TOTAL CYANIDE PROTOCOL 9
1. Distill samples in duplicate. Analyze by titration if over
1 mg/1. Analyze by pyridine barbituric acid (PBA) colori-
metric method if under 1 mg/1.
2. Spike sample with 100 ug/1 if original concentration is <1
mg/1 or with 1 mg/1 if original concentration is >1 mg/1.
Distill in duplicate and analyze by PBA or titration.
3. Dilute sample l:lf add spike as in step 2, distill in
duplicate and analyze by PBA or titration.
4. Dilute sample 1:1, add spike as in step 2, distill in
duplicate and analyze by PBA and pyridine pyrazolone
colorimetric method (PPC) or titration.
5. Determine precision by analyzing seven replicates from each
of two samples from SIC 3312, 2819, 3313, 3315 and 3861.
6. Analyze one sample in duplicate on ten consecutive days from
SIC 3312, 3313, 3315, 3861 and 1099 to evaluate preservation
techniques.
TABLE 13. PHENOLICS PROTOCOL 10
1. Distill samples preserved with phosphoric or sulfuric acids
and CuSO.. Analyze phosphoric distillate by 4-AAP and PHI
in duplicate. Analyze sulfuric distillate by MBTH in
duplicate.
2. Spike samples with 25 ug/1 if original concentration is less
than 50 ug/1 or spike with 100 ug/1 if higher than 50 ug/1.
Distill and analyze in duplicate by MBTH or 4-AAP and PHI as
applicable.
3. Dilute samples 1:1, distill and analyze in duplicate.
4. Dilute samples 1:1, add appropriate spike, distill and
analyze in duplicate.
5. Determine precision by analyzing six samples from SIC 2911,
3312, 2821, 2231 and 3313 seven times each.
6. Analyze by 4-AAP one sample from each SIC in duplicate on
ten consecutive days to evaluate the preservation technique.
-26-
-------
SECTION 4
SAMPLING PROGRAM
A listing of industries identified for sampling, as well as
the number of sites and samples taken are provided in Tables 14,
15, and 16. SIC Code identification is provided in Table 17.
Identification of permit holders, gaining permission to sample
eligible sites, sample collection, preservation and delivery to
the laboratory for analysis were major functions of the sampling
program.
Grab samples were collected according to "Sampling Water and
Wastewater," EPA, 1977 and preserved according to the 1979 EPA
"Manual for Chemical Analysis of Water and wastes."
Grab samples were collected directly from end of pipe out-
falls via hand operated pumps, impellor pumps, direct bottle
collection, or submersible grab sampling equipment. Samples from
common sites were collected at least 3 hours apart.
Trace metal samples were collected in 4-liter polyethylene
"cubitainers" and preserved with nitric acid to achieve a final
concentration of 0.5% resulting in a pH <2. These samples were
stored at room temperature.
Cyanide samples were collected in 10 liter polyethylene
"cubitainers" and preserved with 10 ml of 10 N sodium hydroxide
to a pH >12. Prior to preservation oxidizing ^gents were
destroyed with ascorbic acid. Samples were kept at 4 C.
Phenolic samples were collected in 4-liter amber glass
containers. Oxidizing agents were destoryed with an excess of
ferrous ammonium sulfate. The sample was then split in two
parts. One part was preserved with phosphoric acid and copper
sulfate and the second was preserved with sulfuric acid and
copper sulfate. Each sample was stored at 4 C.
Sampling sites were categorized by SIC Code and a process
description is presented in Table 17 and Appendix III. Samples
were usually collected from sites which had, or were in antici-
pation of, National Pollutant Discharge Elimination System
(NPDES) permits to discharge the analyte under study. Discharge
permit parameters for each site are also tabulated in Appendix
III.
-27-
-------
1
6
1
6
4
4
5
6
9
7
4
14
8
10
4
4
1
1
2
2
4
33
1
2
TABLE 14 . TRACE METALS
SAMPLING STRATEGY
No. of Samples Collected
At Each Site For Total Number Required Analyses fori
Total Hexavalent Dissolved of Samples TotalDissolved
Metals Chromium Metals Collected Metals Metals
6
6
As,
Cd,
Cu,
Pb,
Zn
2
--
12
As,
Cd,
Zn
6
6
Cd,
Cu,
Fe,
Pb,
Hg, Ni
2
12
Cr
2
--
B
Cr
2
8
Cr
2
10
Zn
2
12
Pb,
Hq
2
2
36
As,
Ba,
Cr,
Cu,
Pe, Nl, Se, Ag
--
2
14
Cr
2
B
Zn
2
2
56
Cr
2
16
Pb
2
2
2
60
Cr,
Pb,
Mn,
Sn,
Zn Cr,
Fe,
2
2
16
Cr,
Mn
2
8
Cr,
Cu,
6
6
As,
Cd,
Cu,
Se,
Zn
6
--
--
6
Cd.
Pb,
Zn
3
--
6
As ,
Cd,
Se,
Zn
3
--
6
Al.
Cu,
Zn
2
--
B
Ag
--
2
66
Cu,
Fe
6
6
Al,
Aij,
As,
Ba,
Cd, Cr, Cu $
--
Fe,
Pb,
Mn,
Ni,
So, Sn, Zn
3
3
12
Cd,
Sn,
Cr,
Ni
--
TOTAL 404
-------
TABLE 15. TOTAL CYANIDE
SAMPLING STRATEGY
SIC Number of Sampling
Numbers Sites Sampled
3312 10
2819 10
3313 6
3315 5
3861 3
1099 1
MDISE 1
3471 2
Total Number
Number of Samples Samples
Collected at Each Site Per SIC
2 20
2 20
2 12
2 10
2 6
6 6
6 6
3 6
Total Number Samples 86
-29-
-------
SIC
Numb
2911
3312
2821
2231
3313
2262
2879
2491
3861
TABLE 16. PHENOLICS
SAMPLING STRATEGY
Total Number
Number of Sites Number of of Samples
Sampled Samples Per Site Per SIC Code
14 2 28
10 2 20
7 2 14
6 2 12
6 2 12
6 2 12
5 2 10
4 2 8
16 6
16 6
Total Number of Samples 128
-30-
-------
TABLE 17. DESCRIPTION OF STANDARD INDUSTRIAL CLASSIFICATION
CODES
Standard Industrial
Classification Codes
1061
1094
1099
2231
2262
2269
2621
2812
2819
2821
2822
2823
2879
2911
3229
3312
3313
3315
Descr iption
Ferroalloy Ores, Except Vanadium
Uranium-Radium-Vanadium Ores
Metal Ores, Not Elsewhere Classified
Broad Woven Fabric Mills, Wool (Including
Dying and Finishing)
Finishers of Broad Woven Fabrics of
Man-Made Fiber and Silk
Finishers of Textiles, Not Elsewhere
Classified
Paper Mills, Except Building Paper Mills
Alkalies and Chlorine
Industrial Inorganic Chemicals, Not
Elsewhere Classified
Plastics Materials, Synthetic Resins, and
Non-vulcanizable Elastomers
Synthetic Rubber (Vulcanizable Elastomers)
Cellulosic Man-Made Fibers
Pesticides and Agricultural Chemicals, Not
Elsewhere Classified
Petroluem Refining
Pressed and Blown Glass and Glassware, Not
Elsewhere Classified
Blast Furnaces (Including Coke Ovens),
Steel Works, and Rolling Mills
Electrometallurgical Products
Steel Wire Drawing and Steel Nail Spikes
-31-
-------
TABLE 17. DESCRIPTION OF STANDARD INDUSTRIAL CLASSIFICATION
CODES
(continued)
Standard Industrial
Classification Codes
3331
3332
3333
3341
3471
3861
4911
9999
Descr iption
Primary Smelting and Refining of Copper
Primary Smelting and Refining of Lead
Primary Smelting and Refining of Zinc
Secondary Smelting and Refining of
Nonferrous Metals
Electroplating, Plating, Polishing,
Anodizing and Coloring
Photographic Equipment and Supplies
Electric Services
Mixed Domestic and Industrial Sewage
Effluent (MDISE)
-32-
-------
SECTION 5
DATA PRESENTATION
Results for duplicate analyses of the original, diluted, and
spiked samples along with calculated mean, range, and spike
recovery are presented in Appendix II, Raw Data Tables.
Manipulation of these raw data to evaluate a) spike recovery
of original and diluted sample aliquots, b) diluted sample analy-
sis correlation to original analysis, c) equivalency of digestion
techniques, and d) technique equivalency for various metals,
phenolics and cyanide determinations are presented as Appendix I.
Recoveries are totalled as a distribution over three spe-
cific ranges which are multiples of +15%. Good recoveries repre-
sent those between 85% and 115%. Fa i r recoveries are defined as
those between 70% and 84% as well as 116% and 130%. Poor recov-
eries are those between 55% and 69% as well as 131% and 145%.
Recoveires in two outlier ranges (<55% and >145%) were also
totalled. The recovery distribution is segregated by a) tech-
nique, b) original spike and c) dilution spike. The recovery
distribution is summed over all SIC Codes for each procedure.
Dilution versus original analyses are evaluated by comparing
agreement over four correlation ranges, a) +15%, b) +30%, c)
+45%, and d) > + 45%. Results for each analyte are compared for
each SIC Code and then summed.
Both range and actual concentration are compared for the two
digestion procedures for applicable trace metals. Actual mean
concentration of original analyses are listed side by side for
each metal. The range of the means are computed and the ratio
(VAD/MAD) of the results are also listed. The range between
duplicates within each digestion procedure are also listed in
this table to compare precision.
Methods comparison tables for total cyanide, metals and
phenolics list actual concentration for each method with ratios
for cyanide and phenolics.
A detailed key and description of each data set is provided
as an introduction to the Appendices. Methods of calculation are
also provided.
-33-
-------
SECTION 6
QUALITY ASSURANCE
Comprehensive programs for quality assurance were followed
throughout this project. Analytical methodologies were standard-
ized at project initiation. Programs to ensure the integrity of
instrumentation, facilities, labware, reagents, and data handling
were maintained. Samples were collected, received, analyzed, and
stored under chain of custody guidelines and standardized labora-
tory operating procedures. Quality assurance monitoring for the
analytical and data handling programs are described below.
TRACE METALS
With the receipt of samples from each SIC code, one internal
quality control sample per 10 samples received from that SIC Code
was prepared containing all analytes of interest. Quality con-
trol samples were analyzed according to the same protocols as the
samples including all dilutions, spikes, and processing. In
addition, a reagent blank was analyzed with all sample sets.
An unprocessed (undigested) internal quality control sample
was analyzed with a frequency of one per twenty analyses for all
elements. External samples from EPA EMSL QA Branch, Cincinnati
and EPA Region III were analyzed periodically during the project.
A five-point standard curve was prepared with each set of
analyses. A high or low standard was run alternately every seven
analyses. If more than +5% variation occurred, the entire stand-
ard curve was rerun.
CYANIDES
A high and low standard was distilled and analyzed with
every set of 10 analyses. An internal quality control sample was
also run with every 20 determinations. External samples from EPA
EMSL QA Branch, Cincinnati were also analyzed.
A five point standard curve was run with each set of samples
received. This curve was verified with two standards on each set
of analyses. A new standard curve was run if variation of more
than +5% occured with any two check standards.
-34-
-------
PHENOLICS
Two internal quality' control samples were distilled and
analyzed with every 20 determinations. External samples from EPA
EMSL QA Branch, Cincinnati were also run. A five point standard
curve was run with each set of samples and verified with all sets
of analyses by two standards. If these standards varied by more
than +5% from the five point curve, a new curve was run.
RESULTS
Data for quality control samples are presented with the raw
data in Appendix II. Analyses of these samples proved invaluable
in discovering procedural difficulties such as a) lack of silver
recovery without CNI in the digest, b) low tin recovery without
HC1 in the digest, c) loss of recovery in digestion but not in
spikes of predigested samples for arsenic and selenium (Zinc
Slurry Hydride), d) phosphoric acid interferences with MBTH
phenolic determinations. Matrix interferences and other diffi-
culties in sample analyses were also confirmed by the success of
matrix-free quality control analyses for deionized water control
samples.
DATA HANDLING AND MANAGEMENT
Data handling and management were automated for this proj-
ect, with significant attention paid to quality assurance meas-
ures. Concern for quality assurance is reflected in the forms
used for transmitting data, the mechanism (punch cards) for data
entry to the computer, and in the design of the software package
used in handling data.
A single laboratory data recording form was utilized for
each of the two basic sample types, trace metals and cyanides/
phenolics. Examples of these forms are shown in Figures 1 and 2,
respectively. Calculations of mean, range, sample concentration
and spike recovery were performed by laboratory personnel. Codes
for sample type, analyte, sample preparation and technique as
identified in Appendix II were entered on the forms. Use of a
single form for transmitting data from the laboratory to a key-
punch service minimized the opportunity for transcription error.
Completed sets of data forms were reviewed for completeness and
accuracy, and then delivered to a keypunch service. The keypunch
service accuracy proved to be better than 99.5%. The software
developed to process and retrieve the data was designed and coded
to operate within the IBM 370MVS systems currently being used by
the Washington Computer Center (WCC) COMNET system. The programs
for editing the raw data and subsequent formatting were written
-35-
-------
FIGURE 1
ANALYTICAL TECHNIQUE
CONCENTRATION UNITS 119/1 mg/l
BIOSPHERICS INCORPORATED
LABORATORY DATA COOING FORM - EPA/MCP
TRACE METALS
Sample Identification
m
-*
>>
<
m
A A
E«t
Dilution Fa
dor
Spikt
Dilution C
oncentration
Mean
X
Rang*
Spikt
Recovery
Samplt
Concentration
9
CT
C
j:
V
h-
Sic No.
Sit*
No.
Sample
No.
r*pe
Samplt
.0"
O
Dup A
Dup B
r~
"i
r"
1
n
ri
r~
1
\
r
n
lil
l
1
T
¦
V
1
f
1
1
r
r
1
--
--
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
...
-
1
1 w
-------
FIGURE 2
BIOSPHERICS INCORPORATED
LABORATORY CODING FORM - EPAfMCP
PARAMETER PHENOLICS AND CYANIDES
CONCENTRATION UNITS jig/ I mg/ I
Simple Identification
B
>
m
c
<
¦o
o
c.
X
Stmplt
Dilution
Factor
si
« a>
o o
£
Spike
Dilution Concentration
Meen
T
Rengo
R
Spike
Recovery
Sample
Concentration
a
tr
'c
K
Sic No.
Silt
No.
Simple
No.
m
a.
>
Dup A
Dup B
i
r-
1 1
:i
r
1
1
-
1
i
1
]
-
#
i
+
1
l
t-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
_
-
-
-
-
-
-
-4
k-
4
<
-------
in Fortran IV GI. Data manipulation and output formats were
conducted using Pansophic's EASYTRIEVE (1979), an information
retrieval and data management system.
A flow chart for the data management developed for this
project is presented in Figure 3. The programs provided a series
of multiple checks for the data.
Data read from punch cards were entered into a temporary
data base. Two editing programs, Edit 1 and Edit 2 were run on
this data in succession. Edit 1 primarily checked for valid
entry codes while Edit 2 checked for computation errors and the
proper sequence of data. Lists of the error codes and their
description produced by Edit 1 and Edit 2 are presented in Tables
18 & 19, respectively. Edit 1 was run and rerun on the data set
until all entry errors were corrected. After the Edit 2 error
report showed complete absence of computational errors in the
data set, the data were entered into the permanent data, base
(DATAC). Data were added to the permanent data in a cumulative
manner throughout the course of the project. The final data were
manipulated into report tables using the EASYTRIEVE system. As a
precaution against loss of the data through computer malfunction,
a tape copy of the permanent data base was made each time a
significant amount of data was added.
-38-
-------
FIGURE 3
i
u>
ID
I
/ \
£
/
TtMPA
-w
tfCTlM
Cards
oio5cs
V X
wnnim
El) T I
LIST 1
roiT 1
ERROR REPffllT
ERROR REPORT
lifting
of
TESTIH
UATAC
APPtW)
tASTTMEVE
ntPonis
OATAf)
WYLBUR
DATA HANDLING AND MANAGEMENT SYSTEM
EPA Contract No. 60-03-2788
-------
TABLE IB. - EDIT 1 CODES
Description
Not a valid SIC code
Site field has a non-numeric character
Sample number is blank or contains a
non-numeric (other than blank)
Columns 5 or 11 is/are not blank
The type code is not valid for the type
of form (metals or phenolics/cyanides) ¦
The analyte code is not valid for the type
of form
The sample code is not valid for the type
of form
The sample dilution is partially blank
or has a non-numeric character
The digest dilution or days preserved is
partially blank or has a non-numeric
character
Spike contains embedded blanks, non-numeric
character or contains numerics, but none in
columns 28 and 29
Dup a is blank or contains embedded blanks,
non-numeric character(s) (other than -),
or contains numerics but none in columns 34
and 35
Dup b is blank or contains embedded blanks,
non-numeric character(s) (other than -),
or contains numerics but none in columns 4 0
and 41. No error if field is blank and
record is for AC, blank, or precision
Mean is blank or contains embedded blanks,
non-numeric character(s) (other than -) or
contains numerics but none in columns 46
and 47. No error if field is blank and
record is for precision study
-40-
-------
TABLE 18. EDIT 1 CODES
Description
Range is blank or contains embedded blanks,
non-numeric character(s) or contains numerics
but none in columns 52 and 53. Error if
there was no dup b value and range contains
characters
Spike recovery has non-numeric character(s)
or embedded blanks
Concentration contains non-numeric character(s)
embedded blanks, or contains numerics but
none in columns 64 and 65
Technique code is not valid for form type
Units code is not M or U
Spike has value but recovery is blank or
visa versa - also error if recovery has
value for precision records. For QC, error
if spike or concentration is blank
Both spike and concentration have values
or both are blank. For precision error
if concentration is blank. For QC, error
if spike or concentrations is blank
Site number is not 999 for QC and Blank record
-41-
-------
TABLE 19 - EDIT 2 ERROR CODES
Code Description
1 Indicates record has a spike when it should
not or visa versa - out of sequence
2 Mean is incorrect
3 Range is incorrect
4 Error in sample dilution or digest dilution
5 Concentration is incorrect
6 Spike recovery is incorrect
99 Record is out of order
-42-
-------
SECTION 7
DISCUSSION
TRACE METALS
Data tables containing results for duplicate analyses with
calculated mean, range, and spike recoveries are presented in
Appendix II, along with the quality control results. Reduced
data tables are presented in Appendix I. Data have been manipu-
lated to present the following relationships:
o spike recovery by range (Tables 1-20)
o diluted processed sample result in comparison to the
original analysis (Tables 21-40)
o digestion comparison {Tables 41-52)
o method equivalency (Tables 53-71)
o method precision (Tables 72-110)
The following sections describe results for each metal.
Silver
Loss of silver during the MAD was immediately noted from the
analysis of quality control samples prepared in deionized water.
To verify this loss, a standard curve was put through the MAD and
compared to undigested standards. As seen from results in Table
20 silver was lost from standards. Apparently, precipitation or
plating of the silver on beaker walls was responsible for this
loss. The use of CNI and NH^OH(2) was implemented following the
MAD during final sample workup (dilution to volume) resulting in
acceptable recoveries. This procedure was standardized through-
out the project and all MAD data presented was derived using CNI.
The VAD was successful for quality control samples without use of
Care should be taken to use the appropriate digestion
protocol when analyzing for silver along with other metals. The
effect of CNI addition although positive for silver is unknown
for other metals. Thus, if a common digest is to be analyzed for
silver and other metals, the VAD should be used. Alternately,
separate digestions are recommended.
-43-
-------
TABLE 20. SILVER LOSS DURING MAD WITHOUT CNI
Absorbance
Silver
Standard Digested Undigested
(mg/1)
4
0.017
0.253
3
0.004
0.197
2
0.005
0.136
1
_ _ _
0.069
Good recoveries were achieved for the direct aspiration
technique for 20 of 28 spike analyses in SIC 3861 and MDISE. The
remaining eight spikes for these categories were recovered in the
fair range. No significant recovery differentials were observed
between spiked samples and spiked digested dilutions (1:1).
Twelve of fourteen 1:2 dilution analyses agreed to within +15% of
the original sample analysis. Only one dilution result for SIC
3861 varied by more than +30% of the original sample analysis.
Since this sample contained silver at a level near the flame
detection limit, a +30% variation is not significant.
A severe negative interference was encountered in SIC 2819
for the direct aspiration technique. All 18 spiked samples put
through the MAD procedure were recovered below 55%. Conversely,
good recoveries were achieved on 17 of 18 spikes of the diluted
digest, with the remaining spike recovered in the fair range.
Thus, silver was lost for SIC 2819 samples during the MAD even
with CNI/NH.OH addition. These samples were quite high in
dissolved sa\ts as noted from flame color and shape distortion
during aspiration. In addition, several sites contained high
levels of organic matter or sludge.
As seen in Appendix I similar results were generally
achieved by both the VAD and MAD followed by direct aspiration.
In three samples, silver was detected at low levels using the MAD
with CNI while no detectable silver was observed by the VAD
without CNI. However, since the silver concentrations were near
the detection limit no definite conclusions can be drawn. Good
digestion correlation was achieved with quality control samples.
Recoveries in excess of 145% resulted from SIC 3861 sample
spikes analyzed by the graphite furnace technique. Recoveries
-44-
-------
for spiked digest dilutions did not exceed 117%. In addition,
diluted analyses showed proportionately lower results than the
original analyses. Thus, a positive interference appears to have
been identified in this effluent category. However, it should be
noted that as seen in Appendix II (Raw Data Tables), the preci-
sion of the original analysis was quite poor (+50%), yet the pre-
cision of the spiked sample, the diluted digest, and the diluted
digest spike was excellent. Thus, calculations of spike recovery
on the original sample may suffer from lack of precision in the
original duplicate analyses. Two explanations are possible for
the poor precision of the original silver analyses. First, since
silver tends to precipitate as the halides, homogeneous aliquot-
ting of particulate silver may not have been achieved. Secondly,
CNI workup of samples may not be as efficient or consistent for
low level silver samples. In either case, precision of the 1:2
dilutions (spiked or unspiked) would not have been affected since
they were prepared from the combined original digested dupli-
cates. Since the precision of these duplicates was good, the
instrumental contribution to variation was small, whereas, the
sample preparation contribution could have been significant.
Results for MDISS graphite furnace analyses showed all 12
spike recoveries in the good or fair ranges (6 of 12 in the good
range) with good correlation between diluted and original sample
values.
Good recoveries were achieved for all 13 original sample
spikes in SIC 2819. However, only 50% - 60% recoveries were
achieved for spikes of the 1:2 digested dilutions.
The chelation/extract ion techniques were performed on the
same VAD digests. Fair recoveries were achieved for both .MDISE
samples by the PDCA/CHCl^ technique, and good correlation
resulted between diluted and original analyses. One iMDISE
original sample spike analyzed by APDC/MIBK showed a recovery of
<55%. The diluted digest spike was recovered to a significantly
greater degree than the original sample spike, thus a definite
negative interference existed in this sample affecting the APDC/
MIBK technique. A similar negative interference for APDC/MIBK
was identified in one SIC 3861 sample. A recovery of only 2% was
achieved for the original sample spike, whereas a 57% recovery
resulted from the dilution spike. The unspiked diluted sample
result confirmed the negative interference by yielding a con-
centration of 2.5X higher than the original analysis. The
PDCA/CHC1- extraction technique yielded recoveries of 62.6% and
136.5% for the original and dilution spikes of this same SIC 3861
sample. The second SIC 3861 sample yielded nearly 100% recov-
eries by APDC/MIBK and fair recoveries (118% - 125%) by PDCA/
CHC1 . Good or fair recoveries were achieved for all SIC 2819
spikes by both extraction methods which indicates that the VAD
was successful where the MAD with CNI failed.
-45-
-------
Precision for silver determinations by direct aspiration
were +1.0% Cor all three samples at 0.5, 1.0, and 2.0 mg/1.
Furnace precisions of +33% and +4% were achieved for samples "at
50 and 60 ug/1. Precision for a single sample at 45 ug/1 was
+51% and +8% for APDC/MIBK and PDCA/CHC13 chelation/extraction
procedures, respectively.
Aluminum
Good recoveries were achieved for all direct aspiration
analyses of MDISE samples. Dilution analyses were all with +15%
of the original analyses. Good recoveries were achieved for
graphite furnace analyses as well. The furnace analyses of the
1:2 digest dilutions were only half of expected values based on
the original analyses. However, the dilutions resulted in
concentrations near the detection limit and decreased accuracy is
to be expected at low level concentrations.
The direct aspiration technique produced good recoveries for
all three samples from site 002 in SIC 3341. However, a negative
interference was identified in site 001 as noted by poor recov-
eries (between 65% and 70%) for original sample and diluted
digest spikes. Marginal increases (^5%) in spike recoveries were
achieved in diluted digests over the original sample. Graphite
furnace determinations produced two good and two poor (1381)
recover ies.
No aluminum was detectable by direct aspiration in MDISE and
SIC 3341 site 001 using either digestion procedure. SIC 3341,
site 002 was found to contain 1300 ug/1 by the MAD and 1800 ug/1
by the VAD which is within 500 ug/1, the detection limit. MAD
and VAD results for quality control samples correlated to within
+9%.
Direct aspiration precisions
samples at 18, 50, and 75 mg/1.
+9.6% and +7.3% were achieved for
were +2.4%, +2.6%, and +15% for
Graphite furnace precisions of
samples at 43 and 185 ug/1.
Bar ium
Direct aspiration analyses of MDISE samples resulted in good
recoveries for all sample spikes. No detectable barium was found
at flame levels in either the original or dilution analyses. The
graphite furnace technique produced good or fair recoveries for
all sample spikes, and dilution to original analysis correlations
were at least within +30%.
-46-
-------
Severe negative interferences were encountered for direct
aspiration analyses in sites 001, 002, 006, 007,and 008 of SIC
2819. Original sample spike recoveries approaching 0% were
achieved for all 10 samples from those five sites. Good diluted
digest spike recoveries were achieved for sites 002 and 007,
however, diluted digest spikes for sites 001, 006, and 008
remained low. Sites 006 and 008 were high in organics, salts
and/or sludges. Good recoveries were achieved for the other four
sites (eight samples).
Positive interferences affected graphite furnace analyses in
sites 007 and 009 of SIC 2819 and a negative interference was
identified in site 002. The original sample spikes were recov-
ered in excess of 145% or below 55%. The interferences were
reduced in the 1:2 dilutions with good recoveries achieved in two
of the three samples. The interferences were confirmed by the
unspiked diluted digests yielding lower or higher results than
the original analyses.
No barium was detected by direct aspiration in MDISE and SIC
2819, sites 006 and 008 using either digestion procedure. The
sample from site 002, SIC 2819, was found to contain 800 ug/1 by
the MAD and 600 ug/1 by the VAD which is a good correlation
considering that the detection limit is 100 ug/1. Quality
control samples agreed to within +7% and +25% for two samples.
The VAD suffered from lower precision which may be associated
with varying final acid concentrations.
Precisions of +20%, +57% and +71% were achieved for direct
aspiration on three samples in SIC 2819. These analyses suffered
from the severe interferences associated with SIC 2819 samples
and, therefore, the precision is not indicative of that obtain-
able from samples without interferring matrices. Furnace pre-
cisions were +21% and +27% on SIC 2819 samples containing 150 and
350 ug/1.
Cadmium
All spike recoveries for SIC 1061, 1094, 1099, 3331, 3332,
3333, and MDISE samples were in the good range for the direct
aspiration technique. Excellent correlation also existed between
dilution and original analyses. Good recoveries were also
achieved for four of six samples in SIC 3471. One SIC 3471
sample from site 002 exhibited a positive interference with a
186% recovery for the original sample spike and a 144% recovery
for the diluted digest spike. The remaining sample from site 002
had a 137% recovery for the original sample spike and a 54%
recovery for the diluted digest spike. This sample was found to
contain 5 ug/1 by the original analysis, however, the 1:1 dilu-
tion was found to contain 862 ug/1.
-47-
-------
The graphite furnace technique yielded good recoveries and
good dilution correlation for SIC 1061, 1094, 1099, 3331, 3332,
and 3471. However, furnace recoveries of only 50% - 65% were
achieved for MDISE for both original and dilution spikes.
Somewhat higher results were achieved for MDISE dilutions than
for original analyses, thus, a negative interference was identi-
fied in MDISE for furnace. Original analyses of two SIC 3333
samples resulted in spike.recoveries of 118% and 138%. Dilution
spikes were recovered to nearly 100%.
A definite positive interference was identified in SIC 3333
for the APDC/MIBK technique. Original sample spike recoveries of
332% and 214% were achieved relative to recoveries of 150% and
136%' for dilution spikes. The same sample digest analyzed by
PDCA/CHC1, yielded virtually 100% spike recoveries. No other
interferences were identified in any SIC Code for either extrac-
tion/chelation technique.
The digestion procedures produced results agreeable to
within at least 20% for all samples.
Direct aspiration precisions were +3.4%, +2.1%, and +1.6%
for samples of 0.06, 1.0, and 0.6 mg/1. Furnace precisions of
+ 13% and +11% were achieved on samples of 3 and 9 ug/1. Pre-
cision of +2.8% was achieved for both chelation/extraction
methods.
Total Chromium
Of 256 spikes in 10 industrial categories 213 were recovered
in the good range with all but one of the remaining spikes
recovered in the fair range for the direct aspiration technique.
Thus, no severe interferences were encountered. A small inter-
ference was noted in one sample from site 006, SIC 2819 where the
original sample spike was recovered to 63% with the diluted spike
recovered to 95%. The second sample at the same site exhibited
no interference. Good correlation between original and diluted
determinations was achieved for all analyses except those mar-
ginally above detection limits.
Determinations by the furnace technique yielded 38 of 50
good spike recoveries with all but four of the remaining spikes
recovered in the fair range. A negative interference was iden-
tified in SIC 2231, site 006 where the original spike was recov-
ered to only 6% and the diluted spike to 114%. A milder negative
interference was also identified in SIC 2269 where original
spikes were recovered at 68% and 54% with marginally higher
recoveries for diluted digest spikes.
The 30-minute permanganate digestion time preceeding chela-
tion/extract ion techniques was found insufficient to convert
-48-
-------
trivalent chromium to the hexavalent state. Quality control
samples and digested standards exposed to permanganate for 30
minutes yielded no absorption when extracted and analyzed. The
digestion time was increased to four hours resulting in complete
oxidation and good analytical recoveries. Insufficient sample
was available to repeat MDISE determinations, thus, no data is
available.
The APDC/MIBK chelation/extraction technique produced good
recoveries for 24 of 46 spikes and fair recoveries for an addi-
tional 15 spikes. The remaining seven spikes were all recovered
over 130%, thus occasional positive interferences were encoun-
tered. However, these apparent interferences may be due to
random error rather than an interfering sample matrix since a)
often the diluted digest spike was recovered higher than the
original sample spike, and b) the diluted analysis correlated
with the original analysis.
The PDCA/CHC1, chelation/extraction technique yielded good
recoveries for 29 of 46 spikes and fair recoveries for an addi-
tional 14 spikes. One sample in SIC 2821 exhibited a negative
interference as noted by original and diluted digest spike
recoveries of 30%. No significant interferences were noted on
any of the other eight industrial categories.
Significant variations resulted between the VAD and MAD for
MDISE and SIC 2911 samples. For all samples in these categories
containing <0.5 mg/1 chromium by MAD, the VAD resulted in con-
centrations of only 11% to 60% of MAD, thus significant chromium
was lost in the VAD. As concentrations increased correlation
improved. One sample (SIC 2911 009 02) containing 3.5 mg/1
chromium by MAD yielded 2.9 mg/1 by VAD, thus, only a 17% rela-
tive loss. Equivalent results were achieved for other SIC Codes
on samples containing <0.5 mg/1 except for those marginally above
the detection limit (0.05 mg/1). Quality control samples pre-
pared in deionized water at 3 and 5 mg/1 yielded VAD to MAD
correlation within +2%.
Precisions by direct aspiration of 1.6%, 1.9% and 3.0% were
achieved for three samples from different SIC Codes at 1.3, 4.6,
and 7.8 mg/1. Graphite furnace precisions were +26.2% and +12.5%
for samples at 37 and 68 mg/1. The APDC/MIBK technique yielded
precision of +17.6% at 36 mg/1 while the PDCA/CHC1., technique
produced +16.6% at 34 mg/1.
Dissolved Chromium
Dissolved chromium determinations were performed on samples
from SIC 3312 and 3315 by filtration through 0.45u membrane
filters. Direct aspiration yielded good recoveries for 52 of
-49-
-------
56 spikes and fair recoveries for the remaining four. Similar
results were achieved for all other techniques as no interfer-
ences were identified. For all four techniques, 77 of 86 spikes
were recovered in the good range with the remaining 9 spikes
recovered in the fair range. Good dilution to original analysis
correlation was also achieved.
Hexavalent Chromium
The graphite furnace method for hexavalent chromium yielded
good recoveries for 105 of 160 determinations. Twenty-two of the
remaining 55 spikes were recovered between 70% and 130%. Samples
from sites 001, 006, and 008 in SIC 2819 contained a negative
interference with recoveries ranging from 0% - 70%. Site 006
contained a large amount of sludge which was removed by filtra-
tion as well as high levels of dissolved salts. Site 008 con-
tained a viscous organic material orange in color. Normally a
precipitate fell with the addition of acetic acid and ammonium
sulfate. No precipitation occurred for site 008 samples. Site
006 samples formed a heavy precipitate after the addition of lead
nitrate. Normally little or no precipitation occurred during
this step.
Negative interferences were also encountered in the follow-
ing samples:
o SIC 2911, Site 004; 58% - 65% recoveries
o SIC 3312, Site 006; Original sample spike recovery
of 14%, diluted digest spike recoveries of
36% and 54%
o SIC 3313, Site 005 - 0% recoveries
o SIC 3471, Site 002 - 0% recoveries
In each case, all samples collected at the above sites were
affected by the interference.
Severe negative interferences were encountered for both
chelation/extraction techniques for sites 006 and 008 in SIC 2819
and sites 002, 006, and 007 (sample 1) in SIC 3313, and sites 009
and 013 in SIC 2911. Without exception these samples turned deep
black with the addition of the chelating agent. The black
substance was extracted into the chloroform layer. The chelating
agent was obviously affected by the sample matrix inhibiting its
ability to chelate chromium. However, good recoveries were
achieved on two samples which also blackened. The samples with
interfering matrices from SIC 2819 contained high amounts of
organic materials as described above. Separated oil was present
in sites 009 and 013 of SIC 2911. Thus, the presence of organic
-50-
-------
substances may be responsible for the noted interferences. In
any case, the formation of a black chelate represented a flag
indicating possible negative interferences.
The interferences in SIC 2819 (sites 006 and 0081 and SIC
3312 (site 006) affected all three methods. However, the remain-
ing interference effects were limited to either the furnace or
the chelation/extraction methods. Thus, for most samples at
least one of the methods proved reliable. It is difficult to
assess the reducing potential of the sample matrix which could
have resulted in formation of trivalent chromium causing low
recoveries. This effect may be a factor for those samples whose
matrices interfered with all three procedures. However, the
critical step most susceptible to matrix interferences appeared
to be the formation of the chelate (extractions) or the precipi-
tate (furnacel. Matrices which inhibit these processes will
cause low recoveries, whereas interferences during instrumental
detection are not key factors. Unusual reactions during these
steps should be noted and spike recoveries examined before data
is accepted as valid.
For samples where severe interferences were not encountered,
the furnace technique produced a larger percentage of good
recoveries than the chelation/extraction methods. The furnace
technique had roughly a 5X better detection limit. Precision was
between +2.5% and +7.5% for all three methods when interferences
were absent.
The furnace procedure is the most cost effective method from
a labor and materials basis. The chelation/extraction procedures
require bulky volumetric glassware or separatory funnels. The use
of solvents mandates that the extraction procedures be conducted
in hoodspace. Conversely, the furnace method requires only small
15 ml centrifuge tubes and a centrifuge. The extraction proce-
dures, thus, occupy several feet of hood space, whereas, the
furnace method occupies the space for a centrifuge and test tube
rack. The PDCA/CHCl- method is particularly time consuming with
several tedious pH adjusments and several 2-minute separatory
funnel extractions which become exhausting in a large scale
analytical production-type situation.
Total Copper
All 190 spikes for samples in SIC 1061, 1099, 3331, 3341 ,
4911, and MDISE were recovered in the good range by direct
aspiration. Some very minor positive and negative interferences
were noted in SIC 2819, however, even in this effluent 34 of 36
spikes were recovered in the good or fair range.
Graphite furnace analyses produced 38 of 48 spikes recovered
in the good range and eight of the remaining 10 recovered in the
-51-
-------
fair range. A positive interference was identified in one sample
from site 007, SIC 2819 showing an original sample spike recovery
of 170% and a diluted digest spike recovery of 150%.
The extraction procedures also produced generally good
results. Good recoveries were achieved for 37 of 48 spikes by
APDC/MIBK, with six of the remaining 11 spikes recovered in the
fair range. Samples from SIC 4911 produced most of the poor
recoveries and occasional minor positive or negative interfer-
ences were encountered.
The VAD produced about 40% higher results than the MAD for
both SIC 1061 samples analyzed at a concentration roughly 5X
higher than the 0.020 mg/1 detection limit. Higher VAD results
were also achieved for two SIC 2819 samples from sites 006 and
008. These samples contained large amounts of sludge and organic
matter. All other samples correlated to at least within +5%
except those approaching the detection limit.
Flame precisions of +3.0%, +18.7%, and +2.1% were achieved
for samples containing 0.5, 1.65, and 2.1 mg/1. Furnace pre-
cisions of +4.3% and +1.6% were obtained for samples of 0.02 and
0.7 mg/1. Precision for APDC/MIBK of +2.9% was achieved at a
concentration of 240 mg/1 while PDCA/CHC1., produced +2.2% for the
same sample.
Dissolved Copper
Samples from SIC 3315 were analyzed for dissolved copper.
Direct aspiration produced good recoveries for all spikes. A
positive interference producing recoveries of about 150% affected
graphite furnace determinations for both samples analyzed. Good
recoveries were achieved for both chelation/extraction
procedures.
Total Iron
No interferences affecting direct aspiration were identified
in SIC 1099 or MDISE as 24 of 25 spikes were recovered in the
good range. A negative interference was encountered in SIC 2819,
site 007, sample 1 resulting in recoveries of <50%. A higher
percentage of fair or poor recoveries were achieved for SIC 2819
samples than for the other SIC codes. Recoveries on samples from
SIC 4911 were in the good range for 117 of 130 spikes with an
additional eight in the fair range. High recoveries were noted
in site 016 sample 1 and site 019 samples 1 and 2 for the origi-
nal sample spikes. Spiked diluted digests were recovered near
100%, however, a positive interference cannot be confirmed
because the diluted digest analysis correlated well with the
original analysis. Iron is often present as particulate matter,
thus, some of the poor or fair recoveries noted in SIC 2819 and
-52-
-------
SIC 4911 may have been due to problems associated with homogene-
ous aliquotting of sample, or particulate iron digestion effi-
ciencies rather than positive or negative matrix interferences.
Apparent positive interferences were identified in SIC 1099
samples analyzed by graphite furnace. Original sample spikes
were recovered to 140% with diluted digest spikes recovered at
90% - 95%. However, diluted unspiked analyses were roughly 30%
higher than the original analyses which would indicate a negative
interference. Thus, the aliquotting of particulate iron and pre-
cision in the digestion efficiency of particulate iron may be the
overriding factors for analytical precision and accuracy. Simi-
larly, spikes of the original MDISE samples were recovered at
125% and 135% by graphite furnace. Dilution spikes showed 95%
recovery. However, again a positive interference was not con-
firmed since good correlation existed between original and dilu-
tion analyses. Graphite furnace analyses of SIC 4911 samples
followed the same pattern. In all SIC Codes a total of 26 of 36
original sample spikes were recovered in the good or fair range
with another six in the poor range. Spiked and original analyses
were performed on separate sample aliquots digested on different
days. Spikes of the solubilized iron in the sample digest
yielded 16 of 18 recoveries in the good or fair range. Since
good correlation existed between original and diluted samples
these results confirm that variable recoveries were largely due
to digestion efficiencies of particulate iron or non-homogeneous
sample aliquotting.
The APDC/MIBK chelation/extraction procedure yielded good or
fair recoveries for 24 of 36 spikes, poor recoveries for seven of
the remaining 11 and recoveries <55% or >145% in five cases.
Again, apparent positive or negative interferences could not be
confirmed by diluted digest correlation to the original sample
result. Similar results were obtained for the PDCA/CHCl^ method.
The VAD often produced results 10% - 40% higher than the MAD
particularly for samples from SIC 4911. Sample 4911 016 01 con-
tained a 10X higher iron concentration by VAD than MAD. These
results confirm that the efficiency of the digestion impacted
observed recoveries. In the latter case where VAD yielded 10X
more iron than MAD a non-representative "chunk" of particulate
iron may have been aliquotted and digested. This result repre-
sents a severe example of problems inherent in withdrawing a
representative aliquot from a sample containing high particulate
metals.
Precisions for direct
achieved for total iron at
aspiration of +6.0% and
concentrations of 0.08
+ 2.1% were
and 2 mg/1.
-------
Good recoveries were achieved for these samples by direct
aspiration. A filtered sample was digested and a precision of
+1.3% was achieved by flame. This sample represents precision
Tndependent of particulate aliquotting and digestion. Graphite
furnace precisions of +29.4% and +10.5% were obtained on samples
of 18 and 63 ug/1. APDC/MIBK precision was +18.9% at 115 ug/1
while PDCA/CHC13 was +8.5% at 90 ug/1.
Dissolved Iron
Samples from SIC 3312 and 3315 were analyzed for dissolved
iron. Good recoveries were achieved for all 56 spikes analyzed
by direct aspiration.
A negative interference was encountered in SIC 3312 sites
005 and 009 for furnace analyses. Original sample spike recov-
eries were <50% while diluted spikes were in the good range. A
negative interference was also present in one of two SIC 3315
samples. The interference noted by low recoveries was confirmed
by the unspiked diluted analysis yielding higher results than the
original analysis.
PDCA/CHCl~ chelation/extraction yielded good spike recov-
eries in 9 of 10 cases with one fair recovery, thus no interfer-
ences were identified. A positive interference affected the
APDC/MIBK analysis of the site 004 sample of SIC 3315. An
original spike recovery of 164% was reduced to 114% for the
diluted digest spike.
Mercury
Samples for SIC 1099 and 2812 were analyzed for mercury by
the cold vapor technique. Good recoveries were achieved for
seven of 12 spikes in SIC 1099 with three of the remaining five
recovered in the fair range. One diluted digest sample was
recovered to 145% with another at 150%. These two high recov-
eries appeared to be due to random error rather than interfer-
ences since the corresponding original sample spike was recovered
nearly to 100%.
Samples from sites 001, 002, and 004 in SIC Code 2812 showed
good or fair recoveries, thus, no significant interferences were
noted. Samples from sites 006, 008, and 009 showed original
sample spike recoveries of only 30% while the diluted digest
spikes were recovered between 125% and 175%. Diluted analyses
were roughly 25% higher than original analyses indicating the
possibility of a small negative interference, but not to the
magnitude indicated based on spike recoveries. Random error,
drift in standard curve, and possible mercury loss between
original digestion and spiked original digestion may magnify the
apparent negative interference effects.
-54-
-------
Precision of +7.6% and +5.5% was achieved on samples con-
taining 6.6 and 10.7 mg/1.
Manganese
No interferences were identified for direct aspiration,
furnace, or PDCA/CHC1- chelation/extraction for the three indus-
trial categories studied. Direct aspiration produced 62 of 68
good recoveries, two fair recoveries, and four poor recoveries.
The furnace procedure yielded nine of 14 good recoveries, four
fair and two poor recoveries. PDCA/CHCl^ showed good recoveries
for 11 of 14 spikes, fair recoveries for two spikes and poor
recoveries for one spike.
A standard curve for manganese could not be achieved by the
APDC/MIBK chelation/extraction procedure. The EPA methods man-
ual (4) states that the manganese APDC complex breaks quickly and
samples should be analyzed as soon as possible once chelated.
Approximately one hour passed from time of chelation to time of
aspiration. No absorption resulted for standards or samples from
the same digests as those 'used for PDCA/CHC13. Numerous attempts
were made to produce a standard curve without success. Even when
standards were aspirated within 30 minutes of chelation no
absorption resulted.
The VAD produced up to 27% higher results than the MAD,
however, most results agreed to within +10% or less.
Precisions for direct aspiration were +3.3%, +2.4%, and
+1.1% for samples at 1.1, 1.1, and 2.0 mg/1. Precisions of +8.5%
and +2.9% were achieved for furnace analysis of samples at 100
and 159 ug/1. PDCA/CHCl^ precision was +58.5% at 60.9 mg/1. The
poor precision obtained on this single sample was not indicative
of that obtained in duplicate analyses during the course of the
project.
Total Nickel
Flame conditions were extremely critical for nickel deter-
minations by direct aspiration. Recoveries of 170% were noted
with a slightly rich flame during the analysis of quality control
samples. This effect was eliminated when the flame was adjusted
to a leaner burn. It is apparent that an oxidizing flame must be
maintained for nickel. Since standards and samples are run under
the same flame conditions, this source of error should be mini-
mal. However, acid type and concentration may become increas-
ingly important when using a richer flame, and slight variations
between samples and standards may cause inaccuracies.
No interferences were detected with direct aspiration or
chelation/extraction techniques in the four industrial categories
-55-
-------
studied. A spike recovery of only 26% was achieved for one
sample analyzed by PDCA/CHCl.,. However, this recovery was
affected by an apparently erroneous result for one duplicate
original analysis. Flame analyses produced 70 of 72 good spike
recoveries with the remaining two recovered in the fair range.
A small positive interference was detected in MDISE samples
analyzed by graphite furnace. Recoveries of 130% were achieved
for original sample spikes while recoveries of 87% and 103% were
achieved for spiked 1:2 digest dilutions. The positive inter-
ference was confirmed by the 30% - 40% decrease in the dilution
analysis. A negative interference was noted in one SIC 2819
sample. Despite these interferences, furnace analyses produced
good results with 14 of 18 spikes recovered in the good range.
The VAD produced essentially equivalent results as MAD.
Poor correlation ratios for SIC 1099 and MDISE were due to the
samples near detection limits.
Precisions of +1.2% were achieved for direct aspiration and
10% - 11% for graphite furnace. APDC/MIBK and PDCA/CHC1, pro-
duced precisions of +5.2% and +1.0%, respectively.
Dissolved Nickel
Samples from SIC 3312 and 3315 were analyzed for dissolved
nickel. Direct aspiration produced 55 of 56 spike recoveries in
the good range with the remaining one in the fair range. Good
results were also generally achieved for the chelation/extraction
procedures with 15 of 20 good recoveries, four fair recoveries,
and only one poor recovery. The poor recovery (56%> was obtained
on one SIC 3315 sample by APDC/MIBK indicating a negative
interference.
Seven of 10 recoveries were in the good range for the
graphite furnace method with two others in the fair range. A
negative interference in one SIC 3312 sample was responsible for
one poor recovery (52%I . The diluted digest spike was recovered
to 74%.
Lead
All 144 spikes in seven industrial categories were recovered
in the good range. For SIC 1061, 1099, 2812, 3229, 3312, 3332,
and MDISE, the furnace procedure showed 22 of 28 good recoveries
and four fair recoveries; the PDCA/CHCl. method showed 24 of 23
good recoveries; and the APDC/MIBK methoa produced 21 of 28 good
recoveries and three fair recoveries.
SIC 1099 contained a negative matrix interference affecting
both chelation/extraction procedures and the furnace procedure.
-56-
-------
PDCA/CHC1, produced recoveries of 62% and 80% for original sample
spikes, and 103% and 84% for diluted digest spikes. The APDC/
MIBK method yielded 54% recoveries for original sample spikes and
91% - 97% recoveries for diluted digest spikes. The furnace
procedure was more severely affected showing recoveries no higher
than 28%.
A small positive interference affected the APDC/MIBK results
for SIC 2812. Original recoveries of about 130% were achieved in
both SIC 2812 samples, whereas diluted recoveries were nearly
100%.
Considering that samples only marginally above the detection
limit of 100 ug/1 were available for VAD to MAD comparisons,
similar results were obtained. Two samples at 1400 and 600 ug/1
agreed exactly. Quality control samples compared to within +2%
at concentrations of 10 mg/1.
Direct aspiration precisions of +1.2%, +3.5%, and +0.8% were
achieved for samples spiked to 5, 10, and 15 ug/1. Furnace
precisions of +181% and +10.3% were achieved for samples at 9
ug/1 and 34 ug/1. An interference was responsible for poor
precision on the former sample as it was spiked to 20 ug/1 and
results varied from <1 ug/1 to 45 ug/1 for an average of 8.9
ug/1. A 500 ug/1 sample yielded precision of +6.6% for APDC/MIBK
and +5.0% for PDCA/CHC13.
Tin
Good results were obtained for tin determinations by direct
aspiration as 62 of 64 spikes were recovered in the good range
for the three industrial effluents studied, SIC 3312, 3471, and
MDISE. The other two recoveries were in the fair range.
However, little tin was recovered in graphite furnace
analyses. Even quality control samples produced low recoveries
for furnace analyses. These low recoveries were not due to the
analytical detection technique, but rather were due to the diges-
tion technique utilized prior to analysis. The MAD was employed
for both flame and furnace techniques, however, hydrochloric acid
was used with nitric acid for flame analyses while only nitric
acid was used in the furnace technique for MDISE. After low
recoveries were noted hydrochloric acid was added to prevent tin
precipitation for SIC 3312 and 3471. Hydrochloric acid was main-
tained at 2% in an attempt to keep tin in solution without caus-
ing significant furnace interferences. However, at low level
furnace concentrations tin was still lost even with HCl.
MDISE and SIC 3471 samples analyzed for VAD/MAD comparisons
were below detectable levels. Good correlation (+8%) was
achieved for a 7 mg/1 sample in SIC 3312.
-57-
-------
Precisions of +12.5%, +15.4%, and +0.8% were achieved for
direct aspiration. Furnace precisions were +36.2% and +14.5% for
samples at 20 and 78 ug/1. It should be noted that the 20 ug/1
sample was spiked to 100 ug/1, thus, again poor tin recovery was
achieved.
Z inc
The direct aspiration technique for zinc proved to be quite
successful in the 10 industrial categories studied. Only eight
of 172 spikes were recovered outside of the fair range with 147
of 172 recovered in the good range. Good dilution correlation
was also achieved. Only the two samples from site 010 of SIC
3312 exhibited a positive interference as evidenced by recoveries
of 160%.
Serious problems affected graphite furnace determinations.
Zinc is the most sensitive element by the atomic absorption
technique. Contamination in this extremely sensitive technique
is the major problem. Analyses of MDISE and SIC 3331 samples
were performed utilizing the Perkin-Slmer Model 500 Graphite
Furnace and the Model 5000 Atomic Absorption Spectrophotometer.
Contamination was so severe during both manual and auto-injection
that essentially meaningless data were generated. For all other
samples in SIC 1061, 1094, 2621, 2823, 3312, 3332, 3333, and 3341
the Perkin-Elmer HGA 2100 furnace was used in conjunction with
the Perkin-Elmer Model 370 AAS. The HGA 2100 was 2 - 3X less
sensitive than the 500 furnace and contamination from manual
injection was therefore relatively less severe. For the latter
SIC Codes, 19 of 36 spikes were recovered in the good range, 11
of the remaining 15 were recovered in the fair range, and an
additional three in the poor range. Thus, by utilizing a less
sensitive furnace, contamination effects during injection were
minimized and more reasonable results were achieved. The purpose
of using furnace to achieve greater sensitivity was, however,
partially defeated by using a less sensitive instrument. Injec-
tion into the furnace must be carefully controlled. The slight-
est contact of the pipette tip with the exterior of the graphite
tube added gross contamination. Auto-injection produced even
greater contamination than manual injection.
The chelation/extract ion procedures produced occasional high
spike recoveries (as much as 200%) for quality control samples,
an indication of contamination. Sources of contamination from
glassware and reagents were increased for the 3-step digestion/
extraction/detection procedures over the 2-step direct aspiration
and furnace procedures. The PDCA/CHCl^ procedure produced 10 of
38 good spike recoveries, an additional eight fair recoveries,
and another six poor recoveries. Twelve spikes were recovered
below 55%. The APDC/MIBK method produced somewhat better results
for zinc than the PDCA/CHCl^ with 18 of 42 spikes recovered in
-58-
-------
the good range, an additional 10 in the fair range, six in the
fair range, and five below 55%. Interferences were difficult to
identify since contamination caused high or low recoveries
depending on whether it was greater in the unspiked or spiked
analyses.
Thus, for furnace and chelation/extraction methods contami-
nation present in glassware, reagents, and instrumentation was
the greatest limitation to successful zinc determinations. Zinc
is a common element in the laboratory, and contamination effects
are magnified by the furnace and chelation/extract ion methods due
to their greater sensitivities and more complex sample prepara-
tion or instrumental detection systems.
The VAD produced lower results than MAD for SIC 1094 samples
5X to 10X above the 5 ug/1 detection limit. Good correlation was
achieved for other samples.
Precisions of +4.8%, ;+2.1%, and +0.8% were achieved for
direct aspiration at sample concentrations of 77, 630, 2000 ug/1.
Poor furnace precisions of +93.3% and +174% were achieved.
Chelat ion/extract ion precision of +6.3% an"3 +19.7% was achieved
for a 35 ug/1 sample by APDC/MIBK and PDCA/CHCl^
Arsenic
The graphite furnace technique for arsenic produced good
recoveries for 75 of 108 spikes. Three samples collected from
site 001 in SIC 3333 had a definite positive interference.
Recoveries of 130% - 160% were achieved for the spiked original
and diluted digests. The diluted digest spike recoveries were
approximately equal to or less than the recovery of the original
spike. Severe negative interferences were encountered in site
001, 002, 006 and 008 of SIC 2819 resulting in essentially 0%
recoveries for original spiked samples and marginally higher
recoveries for diluted digest spikes.
The sodium borohydride hydride generation method produced
good or fair recoveries for SIC 1061, 3331, 3333, and MDISE
samples with one exception in SIC 3333 where one member of a
spiked duplicate pair exhibited a recovery of 10% with the other
member showing an 80% recovery. This discrepancy appeared to be
an isolated occurance. Four of six samples in SIC 1094 exhibited
positive interferences resulting in 180% to 220% recoveries.
Five of nine samples in SIC 2819 were affected by negative inter-
ferences showing recoveries between 0% and 40%. One SIC 28.19
sample exhibited a positive interference as noted by a 500%
recovery.
The zinc slurry method of hydride generation yielded rela-
tively poor results. Most (40 of 54) original spikes were
-59-
-------
recovered to 55% or less. However, 28 of 54 diluted digest
spikes were recovered in the good range with 16 of the remaining
26 recovered in the fair range, and only four recovered below 55%
(all in SIC 2819). Thus, the digestion appeared to be the source
of arsenic loss. This method is also limited by poor standard
curve linearity and precision, as well as arsenic contamination
in some lots of zinc powder.
Good precision was achieved for graphite furnace determina-
tions, +4.7% and +2.4% for 34 and 54 ug/1. Zinc slurry hydride
generation precisions were +17.7% and +6.7% for concentrations of
14 and 18 ug/1.
Selen ium
The graphite furnace technique for selenium proved quite
successful for SIC 3331, 3333, and MDISE with only three of 36
spikes recovered outside of the fair or good ranges. Sixteen of
18 diluted digests correlated with the original determination
within +30%. The arsenic interference present in SIC 3333 had no
effect on selenium determinations. However, similar negative
interferences as encountered with arsenic affected SIC 2819
determinations. Varying degrees of negative interferences were
present in site 001 , 002, 004, 005, 006, 007, and 008 of SIC
2819.
The sodium borohydride hydride generation method yielded
good or fair recoveries with no apparent interferences for SIC
3331, 3333, and MDISE. Again as with arsenic negative inter-
ferences affected SIC 2819 results in sites 001, 004, 006, 007,
and 008.
The zinc slurry hydride generation method for selenium
proved even less effective than for arsenic. Selenium was also
apparently lost in the digestion. However, poor recoveries were
achieved even for spikes of previously digested sample aliquots.
Poor standard curve precision and linearity greatly limited the
accuracy of data calculations.
Graphite furnace precisions of +20.8% and +5.8% were
achieved for samples at 5 and 23 ug/1. Zinc slurry hydride gen-
eration precision was +0* and +39% for determined concentrations
of 2 and 4 ug/1. Samples were spiked at 10 ug/1 for the zinc
slurry method, thus, loss of selenium was evident during
d igest ion.
General
Trace metal methods correlation tables are provided in
Appendix III. Results for samples analyzed by more than one AAS
method are listed in columns by technique. In evaluating these
-60-
-------
data for method equivalency several factors should be considered
including a) method precision, b) analytical value with respect
to analytical range and detection limit, c) presence or absence
of matrix interferences, and d) other method limitations such as
sources of contamination.
Samples from SIC 2819 caused interferences for most metals
and techniques. Interferences in sites 006 and 008 were particu-
larly severe. These sites contained high levels of organic
materials as noted by the presence of sludge (site 006} or high
viscosity (site 008}. All SIC 2819 samples were also high in
dissolved salts.
CYANIDES
Samples were analyzed for total cyanide from eight indus-
trial categories. Raw data tables with all duplicate determina-
tions and calculated mean, range, spike recovery, and sample
concentrations are presented in Appendix II. Manipulated data
appears in Appendix I showing the following relationsihps:
o Spike recoveries by SIC Code (Table 116)
o Comparison of diluted versus concentrated sample
analyses (Table 117}
o Evaluation of method equivalency (Tables 118)
o Preservation studies (Tables 120)
o Precision studies (Tables 119)
Analytical results are described below.
Pyridine Barbituric Acid Colorimetric Method (PBA)
The pyridine barbituric acid method (PBA) was utilized for
all distillations containing <1 mg/1 cyanide. Over the eight
industrial categories 89 of 164 spikes were recovered in the good
range, an additional 47 spikes were recovered in the fair range,
and eight in the poor range. Low recoveries of <55% were
obtained for 18 spikes and two were recovered over 145%. Quality
control samples were generally recovered between 85% and 100%
indicating a slight loss in the distillation step.
Samples from SIC 1099 and 3471 showed no PBA interferences
as all spikes were recovered in the good range. Good or fair
recoveries were achieved for 22 of 24 spikes for MDISE and SIC
-61-
-------
3861 with the remaining two spikes only marginally lower. Most
spikes were recovered in the 70% - 100% range rather than 100% -
130%, probably a function of the distillation efficiency. Spike
recoveries were marginally lower for SIC 3861 samples than for
simultaneously run quality control samples, however, no signifi-
cant negative interferences were apparent. Diluted sample
analyses correlated to within +15% of the original analyses for
11 of 12 MDISE and SIC 3861 data points.
Severe to moderate negative interferences- affecting the PBA
method were present in 10 of 62 samples from SIC 2819, 3312,
3313, and 3315. Samples from sites 002 and 006 of SIC 2819
exhibited no spike recovery. Site 006 contained a black organic
and inorganic sludge with an ammonia odor. Site 002 samples
displayed no visual peculiarities. Sample 2 from site 009 in SIC
2819 varied markedly in PBA response. Several distillates had
low absorbances while others were off-scale in absorbance read-
ings. Apparently an interference distilled over to varying
degrees. The rate of distillation was probably the factor
determining interference concentration in the distillate. Some
distillates from this sample formed a red precipitate with
reagent addition.
Both site 003 samples and sample 2 of site 007 in SIC 3312
exhibited negative interferences as noted by recoveries in the
14% - 55% range. Sample 3313 005 2 exhibited a negative inter-
ference as observed from a 9% original sample spike recovery and
a 47% diluted sample recovery. A 268% recovery was achieved for
sample 2 of site 003 in SIC 3313, however, this recovery appears
to be random error as one of the duplicates was excessively high.
Site 003 sample 1 in SIC 3315 showed a moderate negative inter-
ference. All sample distillates were checked for the sulfide
interference, but none was found for PBA samples.
Good or fair recoveries were generally achieved for the
remaining 72 samples analyzed by the PBA colorimetric method.
Silver Nitrate Titrimetric Method
Four samples containing greater than 1 mg/1 cyanide were
analyzed by the titrimetric method (TTM). One good and one fair
recovery was achieved for the SIC 3471 sample, and no interfer-
ences were noted.
Three samples were analyzed in SIC 3312 by the titrimetric
method. The two samples from site 001 contained high levels of
sulfide in the distillate, and cadmium carbonate was used to
eliminate this known interference. Spike recoveries of 112% and
157% were achieved for one of these samples with only 47% and 0%
recovered for the other. However, the 1 mg/1 spike accounted for
only 10% - 20% of the total cyanide in this sample. Good recov-
eries were achieved for the third SIC 3312 sample.
-62-
-------
Colorimetric Methods Comparison
Diluted spiked samples with cyanide below 1 mg/1 were anal-
yzed by both the PBA and the PPC colorimetric methods. Analyses
were conducted on common distillates. Results are presented in
Appendix I. The methods produced equivalent results in all cases
with the largest variation in 82 comparable analyses being only
20%. Interferences affected both procedures similarly.
Method precision
Precision for the titrimetric and colorimetric methods was
determined by analyzing seven replicates put through the entire
analytical procedure including digestion. Precision for the
colorimetric methods was determined on common distillates.
Precision replicates are presented in Appendix I and the data is
compiled in Table 21.
TABLE 21. PRECISION DATA, CYANIDES
SIC
Code PBA PPC TTM
X
s
cv
X
s
cv
2819
279.9
+9.4
(3.4%)
272.4
+ 11.9
(4.4%)
45.0
+ 1.0
(2.2%)
45.3
+ 2.1
(4.6%)
3312
161.7
+ 3.3
(2.0%)
161.4
+ 3.5
(2.2%)
3313
241.7
39.3
+ 6.5
+ 2.5
(2.7%)
(6.4%)
251.4
38.0
+ 4.1
+ 1. 9
(1.6%)
(5.0%)
3315
47.0
69.4
+ 5.3
+ 1. 3
(11.3%)
(1.9%)
46.9
67.1
+ 5.3
+ 1.3
(11.3%)
(1.9%)
3861
40.7
279.6
+ 3.4
+9.2
(8.4%)
(3.3%)
39.0
231.7
+ 2.1
+ 7.9
(5.4%)
(3.4%)
cv
1295 +78 (6.0%)
Preservation Studies
One 10-day preservation study was performed for each of five
industrial classifications. Three of the samples collected for
these studies (SIC 1099, 3313, 3315) were found to contain low
level cyanide concentrations near the detection limit, thus
-6 3-
-------
cyanide loss was difficult to monitor. Samples from SIC 3861 and
SIC 3312 contained approximately 80 ug/1 and 30 ug/1, respec-
tively. No significant cyanide loss occurred for these samples
over the 10-day preservation period as seen in Figures 4 and 5.
Samples were preserved with sodium hydroxide to a pH >12 and kept
refrigerated at 4 C.
PHENOLICS
Raw data tables for phenolics are presented in Appendix II
and manipulated data is presented in Appendix I as follows:
o Spike recovery distribution (Table 111)
o Diluted versus concentrated sample analysis
correlation (Table 112)
o Methods comparison (Table 113)
o Precision studies (Tables 114)
o Preservation studies (Tables 115)
Results for these studies are discussed below.
4-AAP Colorimetric Method
Good spike recoveries by the 4-AAP method were achieved in
112 of 256 determinations and fair recoveries were achieved in 74
determinations in 10 industrial categories. Only 35 recoveries
were <55% or >145%. Occasionally a diluted sample exhibited
significantly lower or higher recoveries than the corresponding
original concentrated analysis indicating an interference not
noted in the original sample. This phenomenan could indicate
that the distillation of interferences was not always consistent.
The rate of distillation, homogeneity of interfering substances
in the sample and other factors could impact on the level of
interferences in any distillate. Varying degrees of interfer-
ences between distillates would not be seen within duplicate
pairs because the duplicates were run on common distillates.
These effects would impact recoveries since spikes were made
prior to distillation.
A negative interference affected recoveries in two of 12 SIC
2231 samples. Sample 1 from site 005 showed no recovery for the
original sample spike and a 25% recovery for the spiked dilution.
The diluted sample was determined to contain twice the original
sample analysis (342 ug/1 compared to 198 ug/1), thus, the nega-
tive interference was confirmed. Sample 2 from site 006 also
-64-
-------
Figure 4
Cyanide Preservation Study
Sample 3861 003 7
100
¦o
20.
10 -
2 3 4 5
8 9 10
6
0
7
1
Days Preserved
-6 5-
-------
Figure 5
Cyanide Preservation Study
Sample 3312 010 7
100
90
80
70
.1 60
50
40
30
20
10
0
3 4 5 6
Days Preserved
8
10
-66-
-------
contained a negative interference showing a 16% original sample
spike recovery and 61% dilution spike recovery. Again the inter-
ference was confirmed by the dilution reading 35% higher than
the original analysis. The other 10 samples in SIC 2231 showed
good to fair recoveries.
Four samples from sites 005 and 006 in SIC 2262 exhibited
positive interferences with spikes recovered between 150% and
250%. Generally good or fair recoveries were achieved for the
other eight samples of sites 001-004.
Good recoveries were achieved in 13 of 16 spikes for SIC
2491 samples with the other three recovered in the fair range.
No interferences were identified in this SIC Code.
SIC 2821 samples exhibited poorer results than most other
SIC Codes with only seven of 28 good recoveries and 10 fair
recoveries. Interferences were not as severe as in other SIC
Codes, but a larger percentage of samples were affected. Moder-
ate negative interferences were identified in six samples from
sites 004, 006, and 007 of SIC 2821, where recoveries of 45% -75%
were achieved.
A moderate negative interference affected sample 1 in site
003 of SIC 2879 resulting in 46% - 65% recoveries. Recoveries
appeared poor for samples in site 001, however, these samples
contained high levels of phenolics and the resultant spikes
accounted for only a small percentage of the total. With these
exceptions, generally good or fair results were achieved for the
remaining samples.
A severe negative interference in SIC 3312 was identified in
site 009 sample 2 where essentially 0% recoveries were achieved.
A more moderate negative interference affected sample 1 in site
009 and samples in site 010. Negative interferences also
affected results in SIC 3313 for four samples in sites 002 and
006.
Samples from SIC 2911, 3861, and MDISE generally showed good
or fair recoveries with 41 of 80 spikes recovered in the good
range, 21 in the fair range, and only three below 55%.
PH-3 Instrumental Method (PHI)
The PH-3 instrumental (PHI) method was evaluated using the
same distillates as the 4-AAP method. Spikes were added to the
samples based on the original concentrations as determined by
4-AAP. A large discrepancy often existed between 4-AAP and
instrumental results (PHI yielded much higher values), thus, the
spikes in many cases represented only small percentages of the
phenolics detected by the instrument, and were an insignificant
-67-
-------
addition relative to precision. This factor must be considered
when evaluating spike recoveries which were <55% or >145% for 171
of 256 determinations. However, the overriding factor for poor
recoveries appeared to be method related rather than spike
related. Duplicate analyses performed on common distillates
usually yielded precise results. However, precision between
separate distillates was suspect, as evidenced by significantly
higher or lower results obtained for spiked samples than for
unspiked samples. In cases where the spikes were insignificant
compared to total phenol concentration, results should have been
similar for spiked and unspiked samples, but reproducibility
between distillates was not achieved.
High background levels were partially responsible for the
lack of reproducible results between distillates. In many cases,
the background levels contributed 50% - 80% of the total signal.
Detection limits were often sacrificed by the necessity of using
smaller path length cells or dilutions to reduce background
levels to achieve on-scale responses. In compensating for the
background, a large potential error was introduced. The back-
ground level was not reproducible between distillates, thus, the
degree of error in compensating for the background signal varied
for each distillate. The rate of distillation may be a determin-
ing factor for background levels.
Distillation as a means of sample preparation is unaccept-
able for instrumental detection. An alternate sample preparation
step such as solvent extraction followed by a back extraction
into water may produce better results.
MBTH Colorimetric Method
Separate samples preserved in sulfuric acid were distilled
and analyzed by the MBTH colorimetric method since phosphoric
acid is an interference with MBTH. Spikes were recovered in the
good range for 64 of 256 analyses and in the fair range for 48
additional spikes. Seventy-eight spikes were recovered at <55%
and 31 at >145%.
As evident from these recoveries significant interferences
were encountered in all SIC Codes. Poor recoveries were obtained
on samples from SIC 3861 with nine of 12 spikes recovered at
levels <55% or >145%. However, all diluted SIC 3861 sample
analyses correlated to within 30% of the original analyses.
Several samples in SIC 2911 exhibited positive or negative
interferences with 13 of 56 spikes recovered <55% or >145%.
Severe negative interferences affected samples from site 004 in
SIC 2821 where all determinations including spikes gave no
absorbance readings thus, 0% recoveries. Additional negative
interferences in SIC 2821 although not as severe significantly
affected the MBTH determinations in samples from sites 001, 005,
-68-
-------
006, and 007. Negative interferences were present in SIC 2231,
sites 001 and 002. Severe negative interferences resulted in 0%
recoveries for site 006 samples in SIC 2262 and also affected
samples in sites 001, 002, and 003. Generally poor results or
worse were achieved for SIC 3313 samples with severe negative
interferences resulting in 14 of 24 spikes recovered below 55%.
Only eight recoveries for this SIC Code were in the good or fair
range. positive interferences were prevalent in SIC 2491 with
half of the eight samples affected. Similar positive and nega-
tive interferences were encountered in SIC 2879 and 3312. Only
MDISE samples were free of interferring matrices.
Two indicators of interferences were observed throughout the
study. First, samples often turned cloudy upon addition of
reagents causing positive interferences. Second, formation of a
deep green or blue color (possibly due to aldehyde reactions) was
often formed rather than the normal pink resulting in negative
interferences. Another problem associated with the MBTH method
was the generation of high zeroes and blanks which were often
higher than or near the low level standards. Zeroes could not be
subtracted from samples because the presence of even low levels
of phenolics nullified the zero response producing standards
below zero levels.
Methods Comparison
Sample concentrations as determined by the three methods are
presented in Appendix I. MBTH and instrumental results (PHI) are
ratioed with the 4-AAP response to show correlation. The methods
did not produce equivalent results. The PHI method yielded sample
concentrations consistently higher than MBTH and 4-AAP, with MBTH
averaging roughly 2X higher than 4-AAP. Samples containing high
phenolic concentrations (1 mg/1 or more) usually produced fairly
equivalent results for the three methods. These samples were
diluted to achieve on-scale readings by the colorimetric method,
thus, interferences were minimized. It is difficult to assess if
the high PHI results were due to interferences or substituted
phenolics not reactive with 4-AAP.
Method Precision
Precision for each method was determined from seven repli-
cate analyses put through the entire analytical protocol includ-
ing distillation. All three procedures were run on common
replicate distillates from samples preserved with sulfuric acid.
Replicate results are presented in Appendix I and data is com-
piled in Table 22. Precision varied within SIC Codes probably
dependent on interference effects. Precisions were not equiva-
lent for the three methods, partially a function of interference
effects specific to each method.
-69-
-------
Preservation Studies
Two 10-day preservation studies were performed on each of
the nine SIC Codes, one for sulfuric acid preservation and one
for phosphoric acid. All analyses were conducted by the 4-AAP
colorimetric method. Results are tabulated in Appendix I and are
presented graphically in Figures fi-14. No significant loss of
phenolics was observed for sulfuric or phosphoric acid preserved
samples in SIC 2231, 2491, 2262, 2821, 2911, 3312, or 3313. Some
loss over the last 2-3 days occurred in samples of SIC 3861.
Interferences described earlier caused erratic data in SIC 2879.
Small increases in phenolic concentrations occurred with time in
some samples perhaps due to changes in phenolic substitution
increasing reactivity with 4-AAP.
-70-
-------
dP 6P dP Ofi dP dP
h r- oo oo cm
dP dP dP efP dP dP
O O C\ H ro
dp tfp ctf> tfP dP cfP
o> it vc r- cm in
dP dP dP dP dP cfP dP dP
iTHno^moo
^ O VO ^ CM 00
LD CM ^ KC
ID o Gs 00 CM
cm o cm rn m
m m (N h o> ^
»-H »H i-H ^ ^ i-H
ivcr^rnmmm
oOHH^rn
ho^h^cj
h in rn h h ^
o (\ oo ^ n
cm o> m r* in m
in m cm cm
oo o in o ^ oo
ON o> r- CM
r-H CM
OOVD^TCMTOO
VDO^T^CMVDCNCN
CM m I-H CM CM
CM
+I+I+I+I+I+I +I+I+I+I+I+I +I+I+I+I+I+I
00 n l/1 CO H o
r- cm r cm
oo CM CM CN ro vo
ONfflS'HH
00 C\ M Ifl O T
O rH VD
^ on h c oo
o r- .h r- vc o
r- in ro in ro cm
+I+I+I+I+I+I+I+I
oooo^r-ONi-^r-
oo^miDOMn^c
roinr^NcMvcifikD
ON i-t
cm ro
dP dP dP dP dP dP
ON CM 00 |H ON CN
« t
rn ui h h m o\
CM r-H CM ro CM CM
dP dP dP dP-dP
ON ON CM - CO O
i dP * *
t 00 O [ O
cm ' ¦ *r
dP dP dP dP dp dp
>h o m rH o
00 00 f ON (Ti H
dP dP dP dP dP OP OP
«g>oooovoinoovoi-H
r-oo*3*'-irocM^Dr-
CM tN H n TT
VO CN O ON I CN
h h ic Oi m in
00 -H 00 o O 00
in >-h vo oo "=r
vo
r» ro rH *r ** <-h
cn in CM
OOOOOVCOOrH
"sr^r^oooN^^T
rH o T
+ 1+1+1+1+1+I +I+I+I+I+I+I
r- on ro oo oo oo
r» ro >i ro r»
r^* ro cm »i
o oo ro in on o
o ro rH v vo cm
rO l£ rH rH
+I+I+I+I+I+I +I+I+I+I+I+I+I+I
m <3- on cn ro
ro oo cm o
ro cn in rH
OOCN-HONCNrnr^
OOODl^W^rOOO
on *r cm ro vo r- >h
00
00 *3>
cn ro
dP dP dP dP dP dP
in m vo vo rr
dP dP dP dP dP dP
*3" r~ in in rH
dP dP dP dP dP dP
CM (Ti O O ON
dP dP dP dP dP dP dP dp
oom*rr-oNr--ooaN
ON o CN t a\ cc
HHHIfOIN
>h o t oo r» in
- cm ' ro *.i
in in cm cm vo in
ro
in on *3"
rH ^ *J«
CN
CM
r~ oo on cm cn h
rH r~ r- cm
vo m
ro ro ro rH
KO *-H »H
0
-------
TABLE 22. PRECISION STUDIES, PHENOLICS
(continued)
AAP MBT PHI
X
s
cv
X
s
cv
X
s
cv
129
+
19
(14.7%)
229
+
92
(40.2%)
200
+
28
(39.0%)
110
+
23
(20.9%)
85
+
8
(9.4%)
153
+
58
(37.9%)
450
+
22
(4.9%)
495
+
59
(11.9%)
558
+
113
(20.3%)
171
+
32
(18.7%)
58
+
16
(27.6%)
338
+
41
(12.1%)
56
+
6
(10.7%)
60
+
6
(10.0%)
234
+
118
(50.4%)
58
+
5
(8.6%)
42
+
11
(26.2%)
185
+
172
(93.0%)
-------
100 -
90 -
80 -
70 -
60 -
50 -
40 -
30 -
20 -
10 -
0 -
0
Figure 6
Phenolic Preservation Study
Sample 2231 003 7
Phosphoric Acid Preservative
Sulfuric Acid Preservative
2 3 4 5 6
1
7 8 9 10
Days Preserved
-73-
-------
Figure 7
Phenolic Preservation Study
Sample 2262 003 7
O)
3.
c
o
~3
£
C
8
c
o
(J
O
"5
c
0)
100
90
80
70
60
50
40
30
20
10
Phosphoric Acid Preservative
&Sulfuric Acid Preservative
2 3 4 5 6 7
Days Preserved
8
10
-74-
-------
Figure 8
Phenolic Preservation Study
Sample 2491 004 7
1,000
900
800
H 700
v
.2 600
400 -
Phosphoric Acid Preservative
* 300
Sulfuric Acid Preservative
200
100
2 3 4
0
1
6 7 8 9 10
Days Preserved
-75-
-------
Figure 9
Phenolic Preservation Study
Sample 2821 004 7
U)
3.
c
.2
100
90
80
70
60
50
c
V
I 40
o
o
§ 30
.c
Q.
20
10
Phosphoric Acid Preservative
Sulfuric Acid Preservative
4 5
Days Preserved
8
10
-76-
-------
Figure 10
Phenolic Preservation Study
. Sample 2879 005 7 #
100
90 -
# # Phosphoric Acid Preservative
80 -
^ ^ Sulfuric Acid Preservative
J 70-
! 60 -
E*
+*
§ 50 -
c
o
o
o 40 -
O
| 30-
20 -
10 -
0
01 23456789 10
- Days Preserved
Phosphoric Acid Preservative
^ Sulfuric Acid Preservative
* Interference Present in Sample, see Text
-77-
-------
Figure 11 _
Phenolic Preservation Study
Sample 2911 004 7
100
90
Phosphoric Acid Preservative
80
o> 70
^ Sulfuric Acid Preservative
60
50
40
30
Q.
20
10
0
0 1
23456789 10
Days Preserved
-78-
-------
120
110
100
90
80
70
60 ¦
50
40 ¦
30 ¦
20 ¦
10 ¦
0 .
Figure 12
Preservation Study
Sample 3312 010 7
\ yv /
»" V
A*
/ \
Phosphoric Acid Preservative
Sulfuric Acid Preservative
J 1 1 1 1 1 L » »
1 2 3 4 5 6 7 8 9 10
Days Preserved
-79-
-------
Figure 13
Phenolic Preservation Study
Sample 3313 001 7
Phosphoric Acid Preservative
Sulfuric Acid Preservative
a>
a.
c
o
0
1
93
f
10 -
0 1
23456789 10
i
Days Preserved
-bO-
-------
Figure 14
Phenolic Preservation Study
Sample 3861 003 7
100
90
80
70
60
50
40
30
20
10
0
Phosphoric Acid Preservative
Sulfuric Acid Preservative
I
1
I
3 4 5 6 7
Days Preserved
8
10
-81 -
-------
REFERENCES
1. "Methods for Chemical Analysis of Water and Wastes," EPA
600/4-79-020, U.S. Environmental Protection Agency,
Cincinnati, Ohio, 1979, Metals 4.1.4.
2. Owerback, D., "The Use of Cyanogen Iodide (CNI) as a Stabil-
izing Agent for Silver in Photographic Processing Effluent
Sample," Photographic Technology Division, Eastman Kodak
Company, Rochester, New York.
3. "Methods for Chemical Analysis of Water and Wastes," EPA
600/4-79-020, U.S. Environmental Protection Agency,
Cincinnati, Ohio, 1979, Metals 4.1.3.
4. "Methods for Chemical Analysis of Water and Wastes," EPA
600/4-79-020, U.S. Environmental Protection Agency,
Cincinnati, Ohio, 1979, Metals 9.2.
5. "Methods for Chemical Analysis of Water and Wastes," EPA
600/4-79-020, U.S. Environmental protection Agency,
Cincinnati, Ohio, 1979, Metals 9.2.1.
1 6. "Chromium Hexavalent, Atomic Absorption Furnace Technique,"
Storet 01032, Method 218.5.
7. "Methods for Chemical Analysis of Water and Wastes," EPA
600/4-79-020, U.S. Environmental Protection Agency,
Cincinnati, Ohio, 1979, Method 335.2.
8. "Methods for Chemical Analysis of Water and wastes," EPA
600/4-79-020, U.S. Environmental Protection Agency,
Cincinnati, Ohio, 1979, Method 420.1.
9. "Methods for Chemical Analysis of Water and Wastes," EPA
600/4-79-0 20, U.S. Environmental Protection Agency,
Cincinnati, Ohio, 1979, Method 420.3.
10. Fisher Scientific Company, "Instruction Manual, Fisher
Phenol Analyzer".
-82-
-------
TECHNICAL REPORT DATA
/Please read Instructions on the reverse before completing)
1. REPORT NO. 2.
EPA-600/4-82-001 ORD Report
3. RECIPIENT'S ACCESSION NO.
PS32 2 IS? 1 2
4. TITLE ANDSUBTITLE
Application and Evaluation of Analytical Procedures
for Trace Metals, Total Cyanide^and Phenolics
S. REPORT OATE
May 1982
6. PERFORMING ORGANIZATION CODE
7. AUTHORISI
Gary J. Gottfried
a. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME ANO AOORESS
Biospherics Incorporated
4928 Vyacortda Road
Bockville, Maryland 20852
1Q. PROGRAM ELEMENT NO.
It. CONTRACT/GRANT NO.
68-03-2788
12. SPONSORING AGENCY NAME £ND ADORERS
Environmental Monitoring and Support Laboratory
Office of Research and .Development
U.S. Environmental P wtection Agency
Cincinnati , Ohio 45268
13. TYPE OF REPORT ANO PERIOD COVEREO
Project Summary 6/79 - 6/80
14. SPONSORING AGENCY COOE
EPA-600/06
13. SUPPLEMENTARY NOTES
16. ABSTRACT
Analytical procedures for the determination of trace metals, total
cyanide and phenolics were systematically evaluated for their applicability
industry-wide. Matrix interferences, methods equivalency, and ana^tical
precision were investigated through a series of duplicate and spiked
analyses on non-diluted and diluted samples. Validation of the method-
ologies and identification of their limitations were thus established both
within specific industrial classifications and across multiple industrial
processes.
17. K£V WOROS ANO OOCUM6NT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. cCSati f-ield/Gioup
-
-
18. DISTRIBUTION STATEMENT
Release to public
19. SECURITY CLASS {Thit Report)
Unclassi fi ed
2i. no. of pages
C>4,
30. SECURITY CLASS iTHii page/
Unclassified
22. PRICE
£Pa Foim 2220I 4*77) paCvmqus coition is oBiowsre
t
------- |