United States
Environmental Protection
Agency
Environmental Monitoring
Systems Laboratory
Research Triangle Park NC 27711
Research and Development
EPA/600/S4-86/048 Apr. 1987
Project Summary
Laboratory and Field
Evaluation of Methodology for
Measurement of Cadmium in
Stationary Source Stack Gases
R. F. Moseman, D. B. Bath, J. R. McReynolds, D. J. Holder, A. L. Sykes,
and T. E. Ward
An initial laboratory and field evalua-
tion study was done to assess the use-
fulness of a Modified EPA Method 5
(MM5) sampling train and flame atomic
absorption spectrometry for measure-
ing cadmium in stationary source stack
emissions. Field evaluations were per-
formed at a municipal solid waste incin-
erator and a sewage sludge incinerator.
These industrial sources are currently
being evaluated by EPA/OAQPS for
multiple pollutants including cadmium
emissions. Also, this methodology is
being developed for application, sub-
ject to verification, at other sources of
cadmium emissions at or above the
method detection limit. A formulation
of the methodology was tested
through the laboratory and field sam-
pling validation phases to evaluate pre-
cision and accuracy of the proposed
method. Collocated, quadruplicate flue
gas samples of nominally 30 and 60 dscf
in one- and two-hours sampling time
were collected to assure an adequate
cadmium content, a representative
sample (including volume of stack gas
and duration of sampling time), and the
production of data to validate the
method in terms of between-train pre-
cision. The overall accuracy and preci-
sion of the analysis procedure were
89.2% and 1.7%, respectively. The
method detection limit for a 30 to 60
dscf (0.85 to 1.7 dscm) stack gas sample
was found to be 0.05 to 0.025 |j,g Cd,
respectively per dscf (1.7 to 0.88 |ig Cd
per dscm). The detection limit of the
atomic absorption instrument was 0.03
H.g/ml. The percent coefficient of varia-
tion (precision) of between-train cad-
mium concentrations averaged 13.5%
for the six sampling runs conducted at
the municipal solid waste incinerator
(Field Test #1). The precision of
between-train cadmium concentra-
tions averaged 3.8% for the four sam-
pling runs conducted at the sewage
sludge incinerator (Field Test #2). Sep-
arate analyses of the front half (probe
and filter) and back half (impingers) of
the field samples revealed that all of the
cadmium was collected in the front
half, with the exception of one sample.
In this sample, a faulty filter leaked and
a small percentage of the cadmium was
captured in the impingers. In all of the
other impinger samples, the cadmium
concentration was below the detection
limit. Precision of the cadmium results
was not affected by varying the sample
size from 30 to 60 dscf.
This Project Summary was devel-
oped by EPA's Environmental Monitor-
ing Systems Laboratory, Research Tri-
angle Park, NC, to announce key
findings of the research project that is
fully documented in a separate report
of the same title (see Project Report
ordering information at back).
Introduction
The U.S. Environmental Protection
Agency (EPA) is currently investigating
cadmium emissions from stationary
sources as a potentially hazardous air
pollutant. If EPA makes a determination
to regulate cadmium emissions, appro-
-------
priate methods of sampling and analy-
sis must be available to quantify accu-
rately the emission of cadmium in stack
gases from stationary sources.
EPA's Environmental Monitoring Sys-
tems Laboratory (EMSL) in Research
Triangle Park, North Carolina, is devel-
oping and validating a methodology for
sampling and analysis of cadmium
emissions. This report presents the re-
sults of field and laboratory studies
funded by EPA and conducted by Ra-
dian Corporation. The objectives of the
study were as follows:
• Determine the applicability of a
Modified EPA Method 5 (MM5)
train and flame atomic absorption
spectrometry for measuring sta-
tionary source stack gas cadmium.
• Evaluate the precision and accu-
racy of the proposed laboratory an-
alytical technique. The technique
consists of sample preparation fol-
lowed by analysis for cadmium
using atomic absorption spec-
trometry.
• Assure that the method has a de-
tection limit sufficient to measure
expected cadmium in municipal
solid waste and sewage sludge in-
cinerator flue gas samples of nomi-
nally 30 to 60 dry standard cubic
feet.
• Combine the results of these deter-
minations to validate the propsed
sampling and analytical method-
ologies.
The method validation was con-
ducted in several stages. The initial ef-
fort focused on defining appropriate
sampling and analytical procedures.
The chosen procedures are discussed in
Section 4 of the full report. A laboratory
study was then conducted to determine
overall precision and accuracy of the an-
alytical portion of the methodology in-
cluding sample preparation and analy-
sis. The next stage of this program
involved a field evaluation conducted at
a large municipal solid waste incinera-
tor and a sewage sludge incinerator.
The field evaluations were conducted
according to the procedures outlined in
the site-specific Quality Assurance and
Project Plans, hereafter referred to as
the QA Plans.
Procedures
Method Formulation and
Design of Evaluation Studies
The sampling and analytical methods
evaluated in this field and laboratory
study were proposed after a thorough
literature search. Sampling methods,
sample preparation techniques, and
methods for analysis of cadmium emis-
sions from stationary sources were re-
viewed. The Source Assessment Sam-
pling System (SASS) and EPA Method 5
sampling train were reviewed as possi-
ble sampling methods. The analytical
methods under consideration included
amperometric titration, voltammetry/
polarography, colorimetry, x-ray tech-
niques, gamma and neutron activation,
emission spectrography, mass spec-
trometry, inductively coupled argon
plasma spectroscopy, and flame and
flameless atomic absorption spec-
trometry. Possible sample preparation
techniques included ultrasonic extrac-
tion and variations of acid digestion.
The various methods of sampling,
sample preparation and analyses were
then compared according to method
detection limit, sensitivity, precision,
speed, complexity, availability, cost,
and overall practicality. After consider-
ing these factors and reviewing the liter-
ature, it was determined that sample
collection using a MM5 sampling train,
followed by acid digestion in a Parr
Bomb and flame atomic absorption
spectrometry would be used for the
sampling and analysis of cadmium
emissions from stationary sources.
The laboratory validation phase of
this work was designed to assess the
combined precision and accuracy of the
sample preparation and analysis steps
of the method using a National Bureau
of Standards (NBS) urban paniculate
sample. Samples were digested using a
mixture of nitric and hydrofluoric acids
in a Parr bomb and analyzed in quadru-
plicate using three different methods:
flame atomic absorption spectro-
photometry (AAS), inductively-coupled
argon plasma spectroscopy (ICAP) and
neutron activation analysis (NAA).
A weighed amount of each particulate
sample was digested along with a glass
fiber filter to provide a background ma-
trix consistent with the field samples.
Following digestion, each sample was
split for analysis by AAS, NAA and
ICAP. The two additional analytical
techniques were used to confirm the re-
sults obtained by AAS. Filter blanks
were also digested and analyzed. The
filter blanks were analyzed to determine
the amount of cadmium present in the
glass fiber filter and acid reagents.
The sampling portion of the study
was designed to assess the efficiency of
the glass fiber filter for collecting cad-
mium from stack gases. In previous
work it was common to use 0.1 N to 0.8
N (5%) nitric acid in water to attempt to
capture and hold certain trace elements
in solutions through which stack aero-
sols were passed. The concentration of
5% nitric acid in water in the impingers
was chosen for this research because it
is reasonable to handle and because the
level necessary could be determined
best in field sampling. Impingers con-
taining 5% nitric acid or distilled water
(depending on the test run) were used
in the MM5 trains downstream from the
filter to trap any cadmium that passed
through the filter. A quad-probe was
used to permit the simultaneous collec-
tion of four samples from the same
nominal point in the stack (collocated
samples) by a single probe. The preci-
sion of the sampling method was evalu-
ated by determining the concentration
of cadmium in each of the four\ samples.
Samples were collected isokinetically
for periods of roughly one or two hours
to yield total sample volumes ranging
from 30 to 79 dscf. The sample volumes
collected were chosen to demonstrate
that: 1) cadmium emissions from a mu-
nicipal solid waste incinerator and a
sewage sludge incinerator could be
measured in samples of 30 dscf and
2) increasing the size of the sample vol-
ume (above 30 dscf) would not affect
the analytical precision and accuracy of
cadmium concentration results.
To recover all of the cadmium from
the sampling train components, the
front half rinses were combined with
the glass fiber filter for digestion and
analysis. The impinger solutions and
back half rinses were combined and an-
alyzed to determine if cadmium was
captured in the back half of the MM5
train. At the direction of the EPA Techni-
cal Project Manager, the impingers in
one of the six tests of Field Test #1 were
filled with distilled water in place of 5%
nitric acid to assess how well distilled
water could catch breakthrough cad-
mium and to avoid the use of acid
whenever possible. (All Test #2 runs
used 5% nitric acid.)
In addition to the analysis of field
samples, several other analyses were
performed as part of the laboratory val-
idation procedures. Sample train com-
ponent blanks (probes and nozzles) and
reagent container blanks were analyzed
using AAS to evaluate the quality con-
trol associated with the field sample
preparation procedures.
-------
Results and Discussion
Prior to any field sampling activities,
the precision and accuracy of the pro-
posed analytical procedures were deter-
mined. The sample preparation and
AAS procedures used to analyze field
samples for cadmium had an accuracy
of 89.2% for analyzing known concen-
trations of cadmium. The precision of
these procedures was 1.7%.
The precision and accuracy were de-
termined by Parr bomb digestion of four
aliquots of a National Bureau of Stand-
ards Standard Reference Material and
three independent analysis techniques:
AAS, ICAP spectroscopy, and NAA. In-
strument detection limits for the three
techniques were 0.03, 0.03, and 0.12 p.g/
ml, respectively.
The accuracy (% recovery) and preci-
sion (% CV) for all three analytical meth-
ods are given in Table 1. Overall aver-
age percent recoveries for each
analytical method were as follows:
89.2% for AAS, 99.3% for ICAP, and
94.2% for NAA. The mean percent dif-
ferences for the duplicates were 1.0%
for AAS, 6.2% for ICAP, and 1.7% for
NAA. In terms of standard deviation, the
values were 0.84, 4.49, 1.09 |j,g/g for
AAS, ICAP, and NAA, respectively. The
duplicate means and the standard devi-
ation are measures of precision for
three analytical methods and serve as a
basis for assessing the relative preci-
sion of the analytical methods.
The field testing portion of this study
was designed to evaluate each of the
following:
• The cadmium collection efficiency
of the front and back halves of the
MM5 train,
• The precision of the sampling and
sample recovery procedures, and
• The effect of the sample volume on
the ability of the method to detect
cadmium.
The first field test was conducted at a
municipal waste incinerator that had
previously been tested for cadmium
emissions. Cadmium concentrations re-
ported during the previous testing
ranged from 23 to 230 ji.g/dscm. Cad-
mium found during this study ranged
from 32 to 115 n>9/dscm. The second
field test was performed at a municipal
sewage sludge multiple hearth incinera-
tor. The range of cadmium concentra-
tion reported for this study was 836 to
1,137 (xg/dscm. Except for one run dur-
ing the first field test, the sampling
method employed for both tests was a
Table 1.
Accuracy and Precision of Cadmium in NBS Standard Reference Material # 1648,
Urban Paniculate3
AAS Accuracy0
Sample IDb Individual
Number % Recovery Mean
A1
A2
B1
B2
Cl
C2
D1
D2
Mean Accuracy
% CV (Precision)
Duplicate Differences:
Mean %:d
Std. Dev. of
Random Error, d-e
97.4
97.4
97.4
84.0
84.7
85.4
81.3
82.0
82.7
92.0
92.7
93.3
89.2
1.7
1.0
0.84
ICAP Accuracy0
Individual
% Recovery Mean
98.3
102.5
106.6
107.5
104.5
101.5
91.3
88.2
85.0
100.1
102.2
104.2
99.3
7.8
6.2
4.49
NAA Accuracy0
Individual
% Recovery
105.5
107.4
92.2
91.0
69.6
71.6
107.5
108.4
94.2
17.1
1.7
1.09
Mean
106.5
91.6
70.6
108.0
aBased on an NBS cadmium value of 75 \>.g Cd per gram of paniculate.
bSample A was Parr bombed once; Samples B, C, and D were bombed twice.
°AAS, Atomic Absorption Spectrophotometry; ICAP, Inductively Coupled Argon Plasma; and
NAA, Neutron Activation Analysis.
dA measure of precision of duplicate determinations on four samples.
Youden, W.J., and E.H. Steiner, "Statistical Manual of the Association of Official Analytical
Chemists, "AOAC, Arlington, Virginia 22209 (1975), p. 18.
MM5 train which used nitric acid in the
first two impingers instead of water.
Separate analyses were performed on
the front and back halves of the trains to
determine the collection efficiency of
each half.
To assess sampling and sample re-
covery precision, four "identical" sam-
ples were collected simultaneously
using a quad-probe. Four simultaneous
samples were collected for a total of 24
and 16 samples for Tests 1 and 2, re-
spectively. Sampling was conducted
isokinetically for all test runs.
Sample volumes collected during the
field tests were nominally 30 and 60
dscf (0.9 to 1.8 dscm). The correspond-
ing method detection limit for these
sample volumes is 0.050 and 0.025 |xg
Cd/dscf (1.7 and 0.84 ^.g Cd/dscm).
Based on the previous cadmium emis-
sions testing, cadmium detection was
not expected to be a problem; however,
the total volume was varied to deter-
mine the resulting effect on the cad-
mium analysis.
Tables 2 and 3 present the cadmium
concentrations and precision assess-
ments for the quad-train field studies.
The between train pooled standard de-
viations were 12.39 (xg/dscm (Field Test
#1) and 38.0 (Field Test #2) and repre-
sent the overall precision for the field
studies. In terms of percent coefficient
of variation, the pooled precision was
13.54 and 3.79, which is a measure of
the precision of sampling and analysis
for Tests 1 and 2, respectively.
The within-run or between-train pre-
cision shown in the tables is assessed in
terms of a standard deviation and per-
cent coefficient of variation. Contribu-
tions to these variables result from
(1) differences in the sampling trains,
(2) variations between trains in the sam-
ple preparation and recovery steps, and
(3) analytical variability. Between-test
pooled variability includes all of the
above in addition to: (1) the day-to-day
variability of the cadmium concentra-
tion in the feed, (2) effects of Different
plant operating conditions, and (3) the
potential effect of within-run variability
on the cadmium collection.
A second variable that was addressed
was the collection efficiency of the sam-
pling train. Except for Run 4-B in Test
#2, no cadmium was detected in the
back half, indicating that the filter was
very efficient in preventing break-
through. Cadmium (about 55 n,g) was
found in the first two impingers of the
back half of Run 4-B of the second field
-------
Table 2. Cadmium Concentrations and Within-Run Precision Assessments for Field Test # 1
Test •
No.
1
2
3
4
5
6
A
32.2
96.5
105.5
76.2
80.0
101.4
Sample
B
32.8
85.6
115.5
88.0
111.5
328.0b
Train
C
30.4
87.2
73.5
101.2
103.3
93.8
D
32.6
104.1
74.9
76.7
88.3
95.6
Average
X
32.0
93.4
92.4
84.8
95.8
96.93
pooled
Standard3
Deviation
1.10
8.63
21.36
11.67
14.24
3.97
72.39"
Percent
Coefficient
of Variation
(%CV)
3.4
9.2
23.1
13.8
14.9
4.1
13.54"
aAII sample train results reported in micrograms cadmium per dry standard cubic meter,
\t.g/dscm.
f probe rinse + filter + impingers ing) "1
\_total volume of stack gas sampled (m3)\
bThis value was excluded from the data analysis, because when subjected to a Dixon Outlier
test0 for Test No. 6, it did not meet the acceptance criterion of 126.1 [ig/dscm, at the 5% level
of significance (i.e., 95% probability).
The Dixon Outlier test may be found in Dixon, Wilford J., and Frank J. Massey, Jr.,
"Introduction to Statistical Analysis," McGraw-Hill Book Company, New York (1957).
f" 4
2 I-
1/2
Table 3. Cadmium Concentrations and Within-Run Precision Assessments for Field Test #2
Test •
No.
1
2
3
4
A
886.2
1, 137. 1
989.1
1,097.8
Sample
B
927.8
1,133.4
1,088.0
1,041. 1C
Train
C
835.5
1,081.9
993.6 *
1,030.9
D
886.6
1,081.4
1,057.5
1,081.9
- Average
X
884.0
1,108.5
1,032. 1
1,062.9
pooled
Standard3
Deviation
37.8
31.0
48.7
32.0
38.0"
Percent
Coefficient
of Variation
(% CV)
4.27
2.80
4.71
3.01
3.79"
aAII sample train results reported in micrograms cadmium per dry standard cubic meter,
M-g/dscm.
I" probe rinse + filter + impingers (^g) ~
\_total volume of stack gas sampled (m3)\
"pooled
i=1
1/2
The filter in this run was faulty and it leaked some cadmium to the impingers. However, only
3% of the total cadmium catch for this train was in the impingers, and 97% was in the front half.
test because gas bypassed a faulty filter.
Particulate could be seen in the glass
connections between the filter holder
and the first impinger.
Field Test #1
A third variable evaluated in the data
set of Table 2 was the length of the sam-
pling period for each run. Tests 2,3, and
6 were each conducted for roughly one
hour while Runs 1, 4, and 5 were con-
ducted for about two hours. An analysis
of variance confirmed that there was
not a significant difference in the vari-
abilities of the cadmium concentrations
of the one-hour compared to the two-
hour runs at the 5% level of significance
(95% probability). Thus, sampling times
of about one hour will yield concentra-
tion data equivalent to that for double
the sampling time.
There were several anomalies in the
cadmium concentration data set for the
first field test. The average cadmium
concentration level in Test #1 was
much lower than for Tests 2 through 6.
The cadmium concentrations reported
for Tests 2 through 6 in Table 2 reflect a
reasonably constant average cadmium
concentration in the stack gas. How-
ever, there is no reason to suspect that
the Test #1 samples were collected any
differently than other test run samples.
Therefore, these data were included in
the data analysis. The concentration dif-
ference between Run No. 1 and the
other runs is expected to be due to plant
feed materials or operating conditions.
Also, Sample B in Test #6 shows an
extremely high value in terms of cad-
mium per dscm (Table 2). Table 4 shows
a comparison of cadmium in terms of
gas volume sampled and the amount of
cadmium per gram of particulate. Sam-
ple B, Test #6, is suspected to have
been contaminated either by a large
amount of cadmium after the sample
was collected or by a few particles ex-
tremely rich in cadmium which may
have been pulled into this particular
train. The Dixon outlier test indicates
that this value is an outlier (i.e., it should
not be considered in the statistical anal-
ysis).
Table 4 shows a particulate value for
Sample A in Test #4 which is nearly
seven times higher than other values of
the data set. For this particular sample,
the glass probe liner broke during or
just prior to sampling. A large amount
of milky material was noted in the two
nitric acid impingers, and the filter ap-
peared brownish rather than white.
Even though the particulate value for
this sample was high, the concentration
of cadmium appeared to agree with the
other three runs of the test and was
therefore included in the statistical anal-
ysis. Finally, the comparison of the cad-
mium collection efficiency of impingers
containing nitric acid versus impingers
containing distilled water was inconclu-
sive. Cadmium was not detected in the
train back half for any run of Field Test
#1.
Field Test #2
In Field Test #2, there were no appar
ent outliers in the cadmium concentra
tions, either on the basis of mg/dscm oi
mg Cd/g particulate, as seen in Table 5
-------
Table 4. Comparison of Field Test # 1 Cadmium Concentrations, Paniculate Concentration,
Cadmium to Paniculate Concentration, and Isokinetic Rates for Each Run3
Test No. A B C D
1
Cadmium (mg/dscm)
Paniculate (mg/dscm)
Cadmium to Paniculate (mg/gram)
Isokinetic Value 1%)
0.0322
b
b
102.2
0.0328
b
b
104.1
0.0304
b
b
W2.6
0.0326
b
b
107.6
Cadmium (mg/dscm)
Paniculate (mg/dscm)
Cadmium to Paniculate (mg/gram)
Isokinetic Value (%)
Cadmium (mg/dscm)
Paniculate (mg/dscm)
Cadmium to Paniculate (mg/gram)
Isokinetic Value {%)
Cadmium (mg/dscm)
Paniculate (mg/dscm)
Cadmium to Paniculate (mg/gram)
Isokinetic Value (%)
Cadmium (mg/dscm)
Paniculate (mg/dscm)
Cadmium to Paniculate (mg/gram)
Isokinetic Value (%)
Cadmium {mg/dscm)
Paniculate (mg/dscm)
Cadmium to Paniculate (mg/gram)
Isokinetic Value (%)
0.0965
b
b
100.8
0.0856
b
b
104.2
0.0872
b
b
99.9
0.1041
b
b
100.4
0.1055 0.1155
38.77 38.78
2.72 2.98
102.8 100.0
0.0762
224.12C
c
101.4
0.0850
34.25
2.48
101.2
0.0800 0.1115
40.73 44.21
1.96 2.52
102.6 101.1
0.1014 0.3280d
37.87 38.93
2.68 8.42d
101.0 100.0
0.0735 0.0749
49.48 36.05
1.48 2.08
103.5 101.0
0.1012 0.0767
33.77 32.59
3.00 2.35
103.1 102.9
0.1033 0.0883
37.19 34.54
2.78 2.56
102.2 101.5
0.0938 0.0956
35.37 32.42
2.65 2.95
97.7 101.2
"The paniculate emission and cadmium-to-paniculate ratio values presented in this table are
for information only. In any municipal waste incineration process, significant variation might
be expected of paniculate or cadmium emissions between tests.
bPaniculate weight was not determined.
cNot valid for paniculate concentration or cadmium-to-paniculate ratio because of broken
probe liner which resulted in excessive paniculate weight values.
dNot valid for cadmium, Dixon Outlier test.
Average run concentrations in mg/dscm
ranged from 0.8866 mg/dscm to 1.1085
mg/dscm. An analysis of variance indi-
cated that there was a significant differ-
ence between runs, which was not un-
expected. Within-run precision,
measured in percent coefficient of vari-
ation (% CV), ranged from 2.80 to 4.71,
indicating consistent and precise sam-
pling and analysis capability under the
test conditions encountered in the sec-
ond field test. The average cadmium
concentration in Field Test #1 ranged
from 32.0 |o.g/dscm to 96.93 (xg/dscm for
six runs, while in the second field test,
cadmium concentrations ranged from
886.6 to 1,108.5 (ig/dscm for four runs.
As noted above, the precision in Field
Test #2 was consistent, exhibiting a
narrow range of variability between
runs. Table 2 shows the precision or be-
tween run variability for Field Test #1 to
range from 3.4 to 23.1 in terms of per-
cent CV. Runs 1 and 6 of Field Test #1
show percent CV of 3.4 and 4.1, respec-
tively, which is in the same range of pre-
cision as the four runs of Field Test #2.
These results suggest that the overall
method precision (sampling and analy-
sis) is probably not a function of the av-
erage concentration (e.g., in ^.g/dscm)
over the concentration range from
about 30 to 1100 fig/dscm. Such factors
as source variability over the sampling
period and quality control by the sam-
pling and analysis personnel are proba-
bly more important than the average
cadmium concentration level in deter-
mining the overall precision.
During Field Test #2, one NBS SRM
urban particulate sample was analyzed
for cadmium. The results showed a 5%
bias (or 95% recovery) which is indica-
tive of the accuracy of the analytical re-
sults for Field Test #2. A second in-
house audit sample was prepared from
a secondary cadmium standard and
showed a bias of -16.6% or 83.4% re-
covery.
The method of addition was used to
assess matrix effects for Train B of each
sample run. Reanalysis of all four sam-
ples after spiking showed results within
10% of the expected values, indicating
no adverse matrix effects.
For Field Test #1, the impinger solu-
tions were adjusted to a known volume
and analyzed directly. For Field Test #2,
the impinger solutions were adjusted to
a known volume and one aliquot was
boiled to near dryness. The digested
sample was then reconstituted to one-
half the original volume and analyzed.
The impinger solutions were also ana-
lyzed directly. Because of a faulty filter
in one train of one run, cadmium
reached the impingers in the amount of
3% of the total cadmium catch for that
train. The analysis of the digested sam-
ple of the impingers from this train indi-
cated that it contained about 50 p.g of
cadmium, while the undigested sample
indicated about 60 (ig.
A greenish residue was noted in most
of the impingers of Field Test #2 after
the nitric acid solutions had been trans-
ferred. As a check for cadmium which
may not have been recovered from the
impinger, the four impingers from run
#2 were rinsed a final time with acetone
which removed the greenish residue
from the impinger. The acetone was
evaporated to dryness and the residue
digested with 5% nitric acid. These solu-
tions were analyzed and found to con-
tain no measurable cadmium.
Conclusions and
Recommendations
The quantification of cadmium in sta-
tionary source stack gases consists of
two portions: sample collection and
laboratory analysis. The precision of the
sample collection procedures was eval-
uated by comparing the cadmium con-
centrations of samples collected using
concurrent quad MM5 sampling trains.
The accuracy and precision of the ana-
lytical procedures, including the sample
-------
Table S. Comparison of Field Test #2 Cadmium Concentrations, Particulate Concentration,
Cadmium to Particulate Concentration, and Isokinetic Rates for Each Run*
Test No. A B C D
1
Cadmium (mg/dscm) 0.8862 0.9278 0.8355 0.8866
Particulate (mg/dscm) 120.98 121.27 120.81 120.88
Cadmium to Particulate (mg/gram) 7.33 7.60 6.91 7.04
Isokinetic Value (%) 93.7 92.5 100.4 97.0
2
Cadmium (mg/dscm) 1.1371 1.1334 1.0819 1.0814
Particulate (mg/dscm) 189.91 194.87 193.16 204.38
Cadmium to Particulate (mg/gram) 5.99 5.85 5.60 5.29
Isokinetic Value (%) 99.3 96.7 105.5 100.6
3
Cadmium (mg/dscm) 0.9891 1.0880 0.9936 1.0575
Particulate (mg/dscm) 170.60 175.74 181.88 173.23
Cadmium to Particulate (mg/gram) 5.79 6.19 5.46 6.11
Isokinetic Value (%) 97.8 96.5 104.6 100.0
4
Cadmium (mg/dscm) 1.0978 1.0411b 1.0309 1.0817
Particulate (mg/dscm) 205.96 208.82 203.08 206.61
Cadmium to Particulate (mg/gram) 5.33 5.11 5.07 5.24
Isokinetic Value (%) 97.4 95.1 105.4 100.3
"The particulate emission and cadmium-to-particulate ratio values presented in this table are
for information only.
bThe filter in this run was faulty and it leaked some cadmium to the impingers. However, only
3% of the total cadmium catch for this train was in the impingers, and 97% was in the front
half.
mium was collected in the front half, in
these type tests analysis for cadmium in
the impinger solutions (back half) at
municipal and sludge incinerators may
be omitted when testing conditions are
not significantly different from those
described in the report. When impinger
solution (back half) analysis for cad-
mium is not performed, water as in
Method 5 may be used in the impingers
instead of dilute nitric acid. At least one
sample from each source should be
checked using the method of additions
to ascertain that the chemical composi-
tion and physical properties of the sam-
ple do not cause erroneous analytical
results. Finally, if a purchased stock so-
lution of cadmium is used during analy-
ses for preparing working standards,
the concentration should be verified
against an independently prepared cad-
mium standard.
preparation steps, were evaluated by
performing replicate analyses on
known amounts of cadmium to deter-
mine the percent recovery and re-
peatability of the method. Based on this
work, the following conclusions can be
made:
• A Modified EPA Method 5 sampling
train and flame atomic absorption
spectrometry were found to be ap-
plicable for the measurement of
cadmium in stationary stack gas
samples.
• The overall accuracy and precision
of the analytical steps were 89.2%
and 1.7%, respectively. The detec-
tion limit of the analytical instru-
ment was 0.03 M-g Cd/ml of pre-
pared sample.
• During analysis of the samples
from Field Test #2, an NBS SRM
urban particulate sample was ana-
lyzed and recovery was found to be
95%.
• The corresponding method detec-
tion limit for a 30 to 60 dscf (0.85 to
1.7 dscm) stack gas sample is 0.05
and 0.025 y,g Cd/dscf (1.7 and 0.88
|j,g Cd/dscm). The determined lev-
els of cadmium at the source tested
ranged from 32 to over 1000 ^g/
dscm.
The percent coefficient of variation
(CV) of between-train cadmium concen-
trations ranged from 3.4 to 23.1 % for the
six sampling runs conducted during
Field Test #1 and ranged from 2.8 to
4.7% for Field Test #2. The pooled CVs
were 13.5% and 3.8% for Field Tests 1
and 2, respectively. The method bias
was not affected by total sample vol-
ume. Stack gas samples of approxi-
mately 30 to 60 dscf were collected, and
the cadmium results for the two sample
sizes did not differ significantly for
either field test. Greater than 99.9% of
the cadmium was collected in the front
half of the sampling train for each run,
except for the previously noted single
exception caused by a faulty filter.
Therefore, based on the combined re-
sults of the method evaluation, a MM5
sampling train and atomic absorption
spectrometry are recommended for
measuring cadmium in stack gases.
Since greater than 99.9% of the cad-
-------
R. f. Moseman, D. B. Bath, R. J McReynolds, D H. Holder, and A. L Syke
are with Radian Corporation, Research Triangle Park, NC 27709; the EP*~
author T. E. Ward (also the EPA Project Officer, see below) is with thi
Environmental Monitoring Systems Laboratory, Research Triangle Park, NC
27711.
The complete report, entitled "Field Evaluation of Methodology for Measurement
of Cadmium in Stationary Source Stack Gases," (Order No. PB 87-145 355/
AS; Cost: $18.95, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use S300
EPA/600/S4-86/048
0000329 PS
STREET
IL
------- |