PROCESS
MEASUREMENTS
REVIEW
INDUSTRIAL
ENVIRONMENTAL
RESEARCH
LABORATORY
SEFft
Volume 2, Number 1
Research Triangle Park, N.C. 27711
Summer Edition, 1979
DEVELOPMENTS IN SAMPLING
TECHNIQUES FOR INHALABLE
PARTICIPATE MATTER
In support of a reassessment of the total sus-
pended particulate standard now underway by EPA's
Office of Air Quality Planning and Standards, three
laboratories in EPA's Office of Research and
Development—the Health Effects Research Labora-
tory (HERL-RTP), the Environmental Science Re-
search Laboratory (ESRL), and the Industrial Envi-
ronmental Research Laboratory (lERL-RTP)-are
examining potential sampling requirements. The
HERL-RTP has recommended a 15-ton upper cut size
for inhalable particulate matter and a second division
at 2.5 urn for fine particulate matter. Current par-
ticulate matter sampling techniques do not provide
data at these cut sizes for either ambient or source
samples. At a workshop of leading aerosol scientists
sponsored by the Process Measurements Branch of
IERL-RTP, a measurement development program
was recommended. The program considers short-
term modifications for existing techniques and a
longer term effort to fully investigate the require-
ments for more information, including data on stack
condensable matter.
There have been a number of developments to
date in this program. Extrapolation techniques have
been developed to estimate the 15-^m particulate
loading using existing data on loadings up to 10 /an. A
15-^m cyclone has been designed and is being tested
for use with a Method 5 train. Horizontal elutriators,
being investigated, have shown good laboratory
agreement with theory, and a prototype eiutriator is
being built for use with the Fugitive Air Sampling
Train (FAST) system for fugitive emission measure-
ments. The ESRL is investigating particle losses in
standard nozzles; preliminary data indicate signifi-
cant losses {up to 90 percent) for many particles
below 15 fan.
Bruce Harris
EPA/IERL-RTP
PROCEDURES FOR OBTAINING
INHALABLE PARTICULATE
EMISSION FACTORS
The Process Measurements Branch (PMB),
EPA/IERL-RTP, is developing two procedures
documents for gathering inhalable particulate emis-
sions factor data from stationary and fugitive
sources. The documents will assist IERL-RTP sup-
port being provided to the Office of Air Quality Plan-
ning and Standards (OAQPS). OAQPS is required by
the most recent Clean Air Act Amendments to re-
evaluate total suspended particulate (TSP) stand-
ards. The data gathering effort is scheduled to begin
in September 1979. Southern Research Institute is
drafting the stack manual while TRC and Midwest
Research Institute are coordinating the fugitive
emissions manual. These manuals will be ready when
the first sampling teams are available and will pro-
vide necessary guidance in selection and implementa-
tion of proper methods.
Bruce Harris
EPA/IERL-RTP
ion
The views expressed in the Process Measurements Review do not necessarily reflect the views and policies of the Environmental Protec-
Agency. Mention of trade names or commercial products does not constitute endorsement or recommendation for use by EPA.
-------
Process Measurements Review
Volume 2, Number 1, Summer Edition, 1979
LIMESTONE SCRUBBER SLURRY AUTOMATIC CONTROL INVESTIGATION
An examination of processes for flue gas desulfur-
ization by wet limestone scrubbing has led to con-
sideration of process automation methods. These
methods have the potential for increasing scrubber
reliability, improving economy of operation, and
reducing the variance of controlled variables, in-
cluding S02. Under an EPA grant sponsored by
IERL-ETP with the University of Cincinnati, control
loops crucial to the performance of the slurry circuit
of limestone scrubbers have been identified, mathe-
matically modeled, and computer-simulated to eval-
uate their dynamics. A preliminary analysis of ex-
pected scrubber performance under automatic con-
trol has been completed. Results indicate that main-
taining a high process gain (defined as the ratio of
slurry pH change per unit of limestone and buffer ad-
dition) under varying scrubber operation conditions
is the primary objective of automatic control.
The derived process model is shown in Figure 1
for slurry holding tank residence time td. Results ob-
tained indicate that the dissolution and contribution
of limestone to the neutralization process is a geomet-
ric progression. The pH titration curve is only mildly
nonlinear and does not require a nonlinear controller
for pH control. The scrubber is also a stable process,
but it exhibits a resonant slurry response of period
2ir/t<1 shown in Figure 2. This is a result of the slurry
being controlled in a distributed manner by the lime-
stone dissolution. This phenomenon imposes no diffi-
culty from a control perspective and, in fact, is similar
to the response of a shell and tube heat exchanger. In
addition, the scrubber loop, in combination with the
hold tank time constant, is the equivalent of a pure in-
tegrator function, l/tdS, which provides an increasing
process gain and hence complete neutralization at
lower frequency load disturbances.
For limestone scrubbers it is generally acknowl-
edged that scrubber operating reliability is a signifi-
cant area of concern. Reliability is strongly influ-
enced by internal scaling attributed to two circulat-
ing slurry species—sulfite and sulfate. The solubility
of the sulfite can be increased by maintaining low pH,
which also enhances alkali utilization. The solubility
of the sulfate is controlled by the fraction of slurry
solids recirculated. The objective of automatic pH
control of the scrubber slurry via the limestone addi-
tion rate is to maintain the efficiency of high alkali
utilization while accommodating varying scrubber
S02 loading conditions. This is achieved for a narrow
pH range, which is optimum in the sense that sulfite
scaling can be prevented and a consistent baseline of
SC>2 removal maintained with an adequate response
to scrubber load changes. Based on the modeling and
computer simulations, it was determined that a feed-
back approach to pH control would accommodate the
scrubber geometric limestone dissolution character-
istic. During computer dynamics studies, both feed-
forward and linear predictor compensators were
found to offer only negligible control improvement
over pH feedback because of the inherent damping
effect of the scrubber process. These approaches,
therefore, do not warrant mechanization considering
the additional complexity required. Experimental
proofing of this limestone scrubber pH control
method is planned for the summer of 1979 at the TVA
Shawnee facility.
Geometric
Dissolution
Buffer
DENSITY
GAIN
Figure 1. Elemental limestone scrubber slurry model.
-------
Process Measurements Review
Volume 2, Number 1, Summer Edition, 1979
Control of the solids by weight fraction in the
scrubber slurry is also essential. This provides ade-
quate crystallization sites for calcium sulfate precipi-
tation and simultaneously prevents sulfate scaling
throughout other portions of the system. A conven-
tional proportional-plus-integral controller provides
the correct response for this control loop, which is
simpler than scrubber pH control because the meas-
urement and actuation of process variables is not re-
quired.
These control loops will provide a consistent, effi-
cient, and reliable baseline of scrubber operation.
However, this is achieved at the expense of reduced
S02 removal. Additional 863 removal, to achieve ac-
ceptable control levels, can be effected by the addi-
tion of organic acid buffering additives. A tertiary
control loop that derives a buffer addition ratio based
on limestone addition is the subject of present study.
These control loops constitute the scrubber slurry
circuit shown in Figure 3.
Pat Garrett
University of Cincinnati
20
-40
-60
-90
-180
0.01 0.1 1.0 100
Figure 2. Resonance response of a limestone slurry circuit.
i Frequency Iradians/minutel
Alkali
Flow
Loop
Alkali
Addition
Loop
Reaction
Buffering
Loop
Proportional
Flow
Controller
A M S
Organic
Acid
Percent
Solids
Loop
Batching
pH
Controller
A M S
Ratio
Flow
Controller
A M S
5.5pH
P+I
Density
Controller
M A S
'/i%acid
A actuator
M sensor
S setpoint
15% Solids
fey
Gage
-Gas
From Clarifier
To Clarifier
Scrubber Hold Tank
Figure 3. Automatic control of limestone scrubber slurry system.
-------
Process Measurements Review
Volume 2, Number 1, Summer Edition, 1979
PROBLEMS
WITH
CHROMATOGRAPHIC INTEGRATORS
Quantitation of the total organic content of a com-
plex environmental mixture is critical in the Level 1
screening of process streams. It is influenced by
many factors including sampling, extraction efficien-
cy, and instrumental variance. Recently, various
methods of data collection and subsequent calcula-
tions for the total chromatographic organic (TCO)
portion of the total organic analysis have been in-
vestigated by the Process Measurements Branch
(PMB) of EPA's IERL-RTP. A summary of these
studies is reported here to allow investigators per-
forming environmental assessments and similar anal-
yses to avoid subtle or hidden errors in this "routine"
laboratory technique.
One area often overlooked in determining the ac-
curacy of TCO measurements is the treatment of
chromatographic data by modern microprocessor in-
tegrators. The advent of microprocessor control of
data acquisition and handling has played an impor-
tant part in optimizing the level of effort required to
quantitate gas chromatography (GO data. However,
one negative aspect of processor control is that it
often conceals the mathematical manipulations used
in data processing formerly performed by the ana-
lyst. In PMB studies, modern chromatographic in-
tegrators have been shown to be capable of produc-
ing errors from 30 to 500 percent in TCO values when
used indiscriminately. It is important to note that not
all integrators suffer from the same logic "quirks"
but all can be misused in a way to cause gross errors.
Figure 4. Zero or negative areas.
At the heart of the problem is the difference in ap-
proach between resolved peak integration and TCO
analysis. In the former case, the optimum result is
achieved when the peak in question is well resolved
(chromatographically) and then integrated. The in-
tegration usually includes compensation for drifting
baseline or other interference. For TCO analysis,
complete resolution is seldom possible because of the
complexity of environmental samples and because of
the screening nature of the TCO-GC procedure. For
these analyses, the optimum is achieved when the
complete area within the TCO retention window has
been detected and reported. This type of integration
is often referred to as block integration. Difficulties
arise when the method of resolved peak integration is
used to perform block integration.
The errors caused by misuse of the resolved peak
integral method can be quite subtle. Zero or negative
areas are possible depending on where the baseline is
established by the integrator (Figure 4). Negative
areas can also be added to peaks (Figure 5).
Some of the difficulties in using resolved peak in-
tegration for TCO analysis can be overcome. Judi-
cious choice of integrator area and slope sensitivities
must be made. Also, the way in which the baseline is
established and how baseline points are used in area
calculations are critical factors.
To illustrate the challenges that one may en-
counter with a chromatographic integrator, consider
a complex sample where the total response is the
Figure 5. Addition of negative areas.
-------
Process Measurements Review
Volume 2, Number 1, Summer Edition, 1979
signal above the baseline. It is of primary importance
that the baseline be properly determined by the in-
tegrator. Any manipulation of the baseline to satisfy
data processing constraints can lead to error. For ex-
ample, the worst case occurs when the data system
"draws" the baseline from peak valley-to-valley (a
common default feature with many integrators). The
calculated integral may be far from the TCO value
because the unresolved envelope is not included in
the total area (Figure 6). In relatively simple samples,
such as calibration mixtures, the error is usually
small. However, with complex samples containing
many unresolved components, the valley-to-valley
method can produce reported areas much lower than
the actual value.
Ongoing evaluation of integrators from various
manufacturers has shown that each chromatogram
should be inspected closely and compared to the com-
puted integral to ensure that the report contains no
obvious errors. If the block integration method is not
available with a particular integrator to determine
TCO values, it is evident that the data system must
be forced to construct a horizontal baseline rather
than a valley-to-valley baseline. The approach used
by the integrator for peak recognition and peak areas
must also be reviewed. In addition, planimetry or cut-
Figure 6. Exclusion of unresolved envelope.
and-weigh methods are useful for periodic quality
control of microprocessor output.
The difficulties with modern chromatographic in-
tegrators exemplify the need for constant evaluation
of the performance of Level 1 analysis, data collec-
tion, and data handling. Accuracy in the quantitation
of the total organic content in a complex sample is
dependent on a thorough understanding of the meas-
urement requirements and the individual contribu-
tions to error. As demonstrated here, misuse of mi-
croprocessor integration can be the cause of sig-
nificant error in TCO analysis.
Ray Merrill
Ray Luce II
EPA/IERL-RTP
SPOT TEST FOR THE DETECTION
OF POLYNUCLEAR AROMATIC HYDROCARBONS
Polynuclear aromatic hydrocarbons (PAH) are
among the many polycyclic organic materials (POM)
commonly encountered as trace level environmental
contaminants in effluents associated with combus-
tion, pyrolysis, and other thermal degradation proc-
esses. The PAH category, defined as containing hy-
drocarbon species with three or more fused aromatic
rings, includes some compounds suspected of being
carcinogens as well as many isomeric and other non-
carcinogenic compounds. Determination of emission
levels of PAH is, therefore, important in environmen-
tal assessment.
Procedures such as gas chromatography/mass
spectrometry (GC/MS) are used to obtain compound-
specific information on potential health hazards
associated with PAH-containing effluents. However,
these procedures are necessarily sophisticated (be-
cause of the large number of possible PAH species)
and require state-of-the-art equipment and extensive
investment of expert analysts' time. It is not cost ef-
fective to apply them routinely to samples that may
not contain any detectable levels of PAH.
A rapid inexpensive spot test for preliminary
screening of samples to determine the presence or
absence of PAH has been developed by Arthur D.
Little, Inc., under EPA Contract 68-02-2150. Details of
the method are given in the report Sensitized Fluo-
rescence for the Detection of Polycyclic Aromatic
Hydrocarbons, EPA-600/7-78-182, PB 287-181, Sep-
tember 1978. Basically, the test involves marking
three 0.25-cm-diameter spots on a filter paper, apply-
ing 1 |iL of sample extract to spots 1 and 2, applying
-------
Process Measurements Review
Volume 2, Number 1, Summer Edition, 1979
1 fiL of naphthalene (sensitizer) reagent solution to
spots 2 and 3, and visually observing all three spots
under 254-nm UV light. The following criteria can be
used to estimate the PAH content in the 1 /iL of sam-
ple (diluted if necessary):
Nonfluorescent with sensitizer: < 1 pg
Weakly fluorescent with sensitizer: 1 - 10 pg
Strongly fluorescent with sensitizer,
but not fluorescent alone: > 100 pg
Fluorescent without sensitizer: > 104 pg
From such estimates, the decision to proceed with
further analysis can be made.
In addition to Arthur D. Little, Inc., several other
contractors have applied this sensitized fluorescence
spot test in the course of their ongoing EPA environ-
mental assessment programs. Included are Monsanto
Research Corporation for coal- and wood-burning fur-
naces, Research Triangle Institute for ferroalloy
processes, and TRW for conventional combustion
sources. Their experience has been that the test is
easy to use and is valuable for preliminary screening.
Results could be relied on to identify samples that
contain no PAH and therefore require no GC/MS anal-
ysis and to rank samples by relative abundance of
PAH. The users found that levels of 10400 pg//iL of
PAH, well below the usual GC/MS detection limits,
were readily detectable by the spot test.
Some practical aspects related to implementation
of the test were also noted. Some batches of the
naphthalene sensitizer were found to have excessive
background levels of fluorescent interferences.
Highly colored sample extracts required dilution
prior to spot test analysis for best results.
Judi Harris
Arthur D. Little
LEACHATE GENERATION PROBLEMS
IN SOLID WASTE CHARACTERIZATION
An adequate measure of the inherent toxicity of a
solid waste material can be obtained by relatively
straightforward chemical and biological testing.
However, in contrast to determining the toxicity of
the material itself, any effort to predict the ultimate
effects on the environment after disposal of the
waste is an exceedingly difficult task. Specifically,
characterization of the leaching properties of a waste
material adds a new dimension to environmental
assessment measurement programs. The leachate
from a waste material can be straightforwardly ex-
amined, but generation of the leachate is a complex
problem. The method used to generate leachate be-
comes the central issue because of the desire to
simulate, to the extent practical, the environmental
conditions to which the waste will be subjected. Al-
though some fairly extensive studies have addressed
the leachate generation problem, a single procedure
that satisfies all of the needs of an environmental
assessment (EA) program has not been identified.
The Process Measurements Branch (PMB) of
EPA's IERL-RTP is currently directing research to
identify a leachate generation procedure suitable for
EA programs. As part of this effort, the GCA/Tech-
nology Division (under EPA Contract 68-02-3129) is
evaluating a series of procedures that, based on
previous investigations, have shown the most prom-
ise of meeting EA requirements. The principal
evaluation criteria are general applicability, repro-
ducibility, and EA methods compatibility.
It is essential that any procedure selected for EA
work be applicable to a wide range of waste mate-
rials. In this regard, and with emphasis on energy
systems, about 10 energy process wastes are being
used to evaluate the test procedures. These materials
include conventional, advanced process, and control
device waste; both unprocessed and "fixed" waste
are being used. In order to determine the reproduci-
bility of the procedures, replicate generations are be-
ing analyzed for selected elements by Graphite Fur-
nace Atomic Absorption Spectrometry. Leachate
generated by each procedure will be subjected to
analysis using EA methods to ensure compatibility
with the established protocol. Both chemical and
biological characterization are being performed.
The procedures currently being evaluated by
GCA include:
• EPA/OSW Extraction Procedure (EP): weak
acid
• ASTM Method A (ASTM-A): distilled water
• ASTM Method B (ASTM-B): weak acid
-------
Process Measurements Review
Volume 2, Number 1, Summer Edition, 1979
• Carbonic Acid Extraction (CAE): weak acid
The first three of these procedures are currently be-
ing tested by the ASTM via a round-robin analysis
program. Although the PMB evaluation is primarily
concerned with EA requirements, the data generated
by GCA will also be made available to the ASTM
Committee.
An additional aspect of leachate generation being
considered is the relative extraction efficiencies of
the procedures. It is generally agreed that a weakly
acidic leaching medium is desirable. This simulates
the anticipated disposal environment while avoiding
worst case treatment (e.g., the use of concentrated
acids or the addition of powerful chelating agents).
The EP and the ASTM-B methods specify pH adjust-
ment; the EP uses a 0.5N acetic acid and the ASTM-B
uses a sodium acetate/acetic acid buffer. The CAE
method has been added to the test series primarily
because problems with some biotests and some
chemical analyses have been attributed to the
presence of acetate. The mechanics of the CAE
method are similar to the ASTM procedures with
C02-saturated water as the leaching medium.
Sample extractions and analytical work will be
completed in August and preliminary results
available in September 1979.
Ken McGregor
GCA/Technology Division
COMPARISON OF SPARK
SOURCE MASS
SPECTROMETRY
WITH OTHER LEVEL 1
ANALYSIS METHODS
Spark source mass spectrometry (SSMS) has been
designated by the Process Measurements Branch of
EPA's IERL-RTP as the primary elemental analysis
technique for Level 1 environmental assessments.
The main criterion considered in the choice of the ele-
mental analysis technique was the ability to detect, in
a 50-mg sample, all elements from beryllium to
uranium with a sensitivity consistent with proposed
IERL-RTP multimedia environmental goals. Any
technique chosen for Level 1 had to be capable of
detecting this range of elements because Level 1
analyses must produce a complete characterization of
a source without consideration of prior knowledge of
source species. This philosophy makes possible com-
parisons of all sources because the sampling and anal-
ysis is uniform. Also, it precludes the possibility of
missing species that might not be considered a threat
now but may come under suspicion later. Other prin-
cipal criteria were minimal sample preparation and
minimal cost per element.
(continued on page 8)
REVISION TO EPA's IERL-RTP PROCEDURES MANUAL:
LEVEL 1 ENVIRONMENTAL ASSESSMENT BIOLOGICAL TESTS FOR PILOT
STUDIES, EPA-600/7-77-043
(Changes 1-3 were reported in the Volume 1, Number 4 issue of the PMR.)
Change 4: "Mysid Bioassay"
Chapter 3, beginning on page 71
The old procedure using grass shrimp was effective but required a prohibitively large sample
size. The new procedure uses Mysid shrimp in place of grass shrimp. The new test has been
shown to be quite sensitive to complex samples and requires a much smaller sample size.
NOTE: Revisions appear in condensed form. For complete change notices, contact Ray Merrill,
PMB, EPA/IERL-RTP (919/541-2557), Research Triangle Park, NC 27711.
-------
Process Measurements Review
Volume 2, Number 1, Summer Edition, 1979
Other multielement techniques considered were
inductively coupled argon plasma optical emission
spectroscopy ffCAP), neutron activation analysis
{NAA), and X-ray fluorescence (XRF). Major charac-
teristics of these techniques are summarized in Table
1. XRF was the first technique to be eliminated: sen-
sitivity is poor for many elements and matrix absorp-
tion and enhancement effects are frequent problems.
NAA is an appealing technique because it does not
require any sample preparation and is not sample-de-
structive. One of its main problems is that bismuth,
thallium, yttrium, phosphorous, boron, beryllium, and
lithium are not activated. Also, routine economical
techniques suitable for Level 1 only detect about 25
elements. Of these, several elements (e.g., fluorine,
chromium, and selenium) do not give consistent sensi-
tivity. Another factor against NAA is that samples
must be allowed to "cool down" for about 2 weeks
after irradiation before counting. Also, samples must
be run in batches of about 30 for economical use of the
reactor.
ICAP, a relatively new technique, utilizes an in-
ductively coupled argon plasma, which provides a
much more stable excitation method for optical emis-
sion spectroscopy than former methods such as a
flame or d.c. arc. It provides good sensitivity for ele-
ments other than bismuth, germanium, rhenium, se-
lenium, tungsten, mercury, antimony, and thallium.
One limitation of ICAP is that each element requires
a separate analysis channel. Most multichannel ICAP
instruments are set up for 48 elements or less. There-
fore, to detect the number of elements required for a
complete survey, two instruments would be required
or the balance of other elements would have to be run
by other techniques. This would increase the cost per
sample. However, the chief problem with ICAP is
that the sample must be dissolved before analysis,
which can be a formidable problem with many Level
1 samples. Even if some of the difficult samples could
be dissolved with acids, the chances of sample con-
tamination and handling errors would be increased.
SSMS, because of its ability to detect approxi-
mately 72 elements with sensitivity consistent with
Level 1 requirements, is uniquely qualified for the
complete elemental survey required. The sample
preparation techniques minimize the chances for
sample contamination and handling errors. Matrix ef-
fects are, for all practical purposes, nonexistent.
In conclusion, it can be stated that all of the tech-
niques considered have unique properties that make
them more or less attractive for this particular ap-
plication. The properties of SSMS overwhelmingly
qualify it for use as the primary elemental detection
technique in Level 1 source assessment.
Frank Briden
EPA/IERL-RTP
Table 1. Summary of Multielement Technique Characteristics
NAA
SSMS
XRF
ICAP
Elements not
detectable
Problem elements
Elements per run
Level 1 sensitivity
for elements not
specified as un-
detectable or
problems
Matrix problems
Sample preparation
Analysis time'
Bi,Tl,Y,P,
B, Be, Li
F, Ca, Cr, Mb,
Zr,Se
20
Good
Na,Cl
None
2-3 weeks
Hg
Br,Cl,F,
S,B
72
Good
Organic >50%*
Mix with
graphite
and press
2 days
Be, B, F, Li
Mg,Na,Al,P,
S,C1
60
Poor
Adjacent elements
absorption and
enhancement
Press straight
or with binder
2 days
None
Bi, Ge, Rh, Se,
W, Hg, Sb, TI
48
Good
Alkali and
alkaline
earth metals
Must be
dissolved
V^day
*Sampte must be combusted in Parr bomb.
•jTypical time from start of sample preparation to results available.
8
-------
Process Measurements Review
Volume 2, Number 1, Summer Edition, 1979
PERFORMANCE AUDIT OF LEVEL 1 ENVIRONMENTAL ASSESSMENT
ANALYTICAL SYSTEMS
The current EPA/IERL-RTP program of environ-
mental assessment (EA) is designed to yield data that
will result in the identification of sources of recog-
nized pollutants and other substances of potential en-
vironmental concern. The information gathered from
Level 1 EA studies will normally be used to deter-
mine if further studies are necessary for develop-
ment of control technologies and/or development of
new emission control regulations. The decisions made
regarding these matters can be no better than the
data collected to support them. As a part of the
IERL-RTP quality assurance program for assessing
and assuring data quality, two components of the EA
process —the analytical methods and their applica-
tion—have been evaluated by means of an audit in-
volving several types of samples and eight IERL-
RTP contractors. This audit, which was administered
by Research Triangle Institute (RTI), had the follow-
ing objectives:
• Evaluate several presently prescribed Level 1
analytical procedures.
• Collect objective evaluations of these proce-
dures from participating contractors.
• Collect data that will provide an improved esti-
mate of the accuracy and precision that can be
expected with these procedures.
• Evaluate the analytical capabilities of the par-
ticipating contractors.
• Identify any particular problems that the par-
ticipating contractors might be experiencing
with the Level 1 procedures, and encourage
their communication with EPA, RTI, and other
contractors to solve these problems.
This audit was designed to study four components
of the Level 1 analysis scheme. First, Parr bombing
and spark source mass spectrometry (SSMS) were
used for elemental analysis of an XAD-2 resin sample.
This sample was chosen to check for completeness of
ashing (oxidation) and contamination of the sample
during the ashing process. Second, a modified fly ash
sample was selected to test the SSMS technique.
Third, the Level 1 organic measurement techniques
of infrared spectroscopy (IR), low resolution mass
spectroscopy (LRMS), total chromatographable or-
ganics (TCO), gravimetric analysis (GRAY), and liquid
chromatography (LC) were evaluated using a five-
component organic mixture. Finally, a commercial
dye mixture was selected to test the LC scheme. The
samples prepared for this audit were purposely not
complex to allow easy identification of basic problems
that would be difficult to identify with real-world,
complex samples. For example, the extent of contami-
nation by both organic and inorganic materials would
be difficult to measure if the test samples were not
simple.
The analysis results received from the partic-
ipants have been organized into two types of reports.
Individual internal reports prepared for each par-
ticipating laboratory comparing the laboratory's re-
sults with expected results and the mean of the re-
sults reported by all participating laboratories, and a
general report comparing all results, describing iden-
tified sources of error, and making recommendations
for increasing data quality.
This audit has resulted in a list of identified
sources of error that hopefully can be minimized or
eliminated. One of these involves the quantitation of
nonvolatile organic substances by means of a GRAY
procedure. High results were reported by several
participants apparently due to insufficient drying of
the sample. All participants reported low results for
cadmium (Cd) in the fly ash sample analyzed by
SSMS. It appears that Cd (as cadmium nitrate tet-
rahydrate) was lost during preparation of the elec-
trode when the sample-carbon slurry was heated to
almost 200° C with a heat lamp. As a third example of
an error source, several sets of analysis results clear-
ly indicated the presence of contamination.
Most error sources, as those listed above, are
minor and correctable. Overall, this audit has indi-
cated that the Level 1 environmental assessment pro-
cedures are, for the most part, satisfactory and meet
the accuracy and precision goals of the Level 1 pro-
gram. For example, the absolute mean percent bias*
of the results reported by the participants for 34
elements in the fly ash sample was 20 percent. Like-
wise, the absolute mean percent bias of the values
(TCO plus GRAY) reported by the participants for the
organic sample was 21 percent. The most successful
aspect of this particular audit is that it has clearly
identified a number of correctable analytical prob-
lems and, in that respect, should lead to improved
quality of environmental assessment data.
Bill Gutknecht
Research Triangle Institute
*Percent bias = [(reported value - expected value)/
expected value] x 100.
-------
Process Measurements Review
Volume 2, Number 1, Summer Edition, 1979
FINE PARTICLE STACK SPECTROMETER SYSTEM
Measurement of the particle size distribution in
stacks and other hot emission sources is of fundamen-
tal importance in understanding the nature and quan-
tity of participate matter emitted. By combining up-
stream and downstream measurements, control de-
vice effectiveness can be characterized as a function
of size. Particle size distribution measurements by
conventional methods (e.g., impaction) require a great
deal of effort. Additionally, the low rate of collection
precludes observations of transient phenomena such
as rapping pulses associated with preeipitators.
Sampling periods of tens of minutes to hours are gen-
erally required for gravimetric analysis.
Particle Measuring Systems, Inc. (PMS), under
EPA contract 68-02-2668, has recently developed an
in situ particle size spectrometer using single particle
light scattering from a helium-neon gas laser source.
The Fine Particle Stack Spectrometer System
(FPSSS), in addition to its in situ measuring proper-
ties, provides real-time data acquisition and both size
and time resolutions that are significantly higher
than with other methods.
The FPSSS has four size ranges covering 0.4-
1.15, 0.5-2.0, 1.15-5.65, and 2.0-11.0 /an. Each size
range has 15 size classes. In normal operation, two
size ranges are sampled concurrently (e.g., 0.5-2.0 /un
and 2.0-11.0 /«n) producing 30 classes from 0.5-11.0
fjun. The maximum number density that can be
measured is 5 x 104 cm ~ 3.
Figure 7 depicts the instrument head, heat ex-
changer, and two lateral support bearings (mount to
port flanges). The bearings allow the operator to ex-
tend the head on a segmented boom (not shown) up to
600 cm into the particulate environment. The water-
Lateral Support
Bearings
Instrument Head
Figure 7. FPSSS components.
Figure 8. FPSSS electronics console.
cooled head can operate continuously at tempera-
tures above 250° C. The head contains the laser, con-
densing and imaging optics, and programmable pre-
amplifiers.
Figure 8 shows the FPSSS electronics console,
housing the signal processing electronics, and the
data acquisition and display system. Data acquisition
is accomplished using a microcomputer with firm-
ware programs and random access memory. Both
CRT and hardcopy displays are generated. Sufficient
memory capacity exists to generate size, area, mass,
and accumulative mass distribution for up to ten in-
dividual samples. In addition, numerical listings and
time series plots of selected parameters (e.g., mass
loading and number density) may be generated. Cali-
bration parameters can be manually entered; for in-
stance, the particle density is invariably entered to
compute mass and aerodynamic diameter. The latter
can be chosen as the relevant size parameter for
various outputs.
The prototype FPSSS has undergone testing at
coal-fired generating stations and at PMS and IERL-
RTP laboratory test facilities. It has performed well
in laboratory tests showing good size agreement with
polystyrene spheres of known size and good agree-
ment with mass measurements of fly ash by gravi-
metric methods. Data on size distributions and mass
in the coal-fired boilers are presently being eval-
uated. The instrument's internal velocimeter is not as
accurate as conventional methods, but can provide 10
to 20 percent accuracy over a 1 to 30 m sec"1 range.
Bob Knollenburg
Particle Measuring Systems, Inc.
10
-------
Process Measurements Review
Volume 2, Number 1, Summer Edition, 1979
RECENT EM PUBLICATIONS OF INTEREST
H. Dehne
Design and Construction of a Fluidized-Bed Combus-
tion Sampling and Analytical Test Big, EPA-600/7-78-
166, PB 290-914 (8/78).
G. T. Brookman, J. J. Binder, P. B. Katz, and W. A.
Wade, III
Technical Manual for the Measurement and Modeling
of Non-point Sources at an Industrial Site on a River,
EPA-600/7-79-049, PB 295-028 (2/79).
Phil A. Lawless
Analysis of Cascade Impactor Data for Calculating
Particle Penetration, EPA-600/7-78-189, PB 288-649
(9/78).
J. A. Armstrong, P. A. Russell, and R. E. Williams
Balloon-Borne Particulate Sampling for Monitoring
Power Plant Emissions, EPA-600/7-78-205, PB 290-473
(10/78).
D. G. DeAngelis and R. B. Reznik
Source Assessment: Residential Combustion of Coal,
EPA-600/2-79-019a, PB 295-649 (1/79).
L. E. Ryan, R. G. Beimer, and R. F. Maddalone
Level 2 Chemical Analysis of Fluidized-Bed Combus-
tor Samples, EPA-600/7-79-063b, PB 295-462 (2/79).
W. E. Farthing, D. H. Hussey, W. B. Smith, and R. R.
Wilson, Jr.
Sampling Charged Particles With Cascade Impac-
tors, EPA-600/7-79-027, PB 290-897 (1/79).
G. T. Brookman, B. C. Middlesworth, and J. A. Ripp
Assessment of Surface Runoff from Iron and Steel
Mills, EPA-600/2-79-046, PB 294-981 (2/79).
J. A. Dorsey, L. D. Johnson, and R. G. Merrill
A Phased Approach for Characterization of Multi-
media Discharges from Processes, ACS Symposium
Series No. 94 (11/78).
Copies of these publications are available at cost
from:
National Technical Information Service
U.S. Department of Commerce
5285 Port Royal Road
Springfield, Virginia 22151.
Name
Company
Address
KiCvl
Request for the Process Measurements Review
Position or Title
ISlrrrll
ISuiel iZip Codrl IPtxmrl
Return to: Ann Turner
Research Triangle Institute
P.O. Box 12194, Building 6
Research Triangle Park, N.C. 27709
(919)541-6893
11
------- |