United States
Environmental Protection
Agency
Environmental Monitoring Sy-iems
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S4-83-034 Oct. 1983
<&EBA Project Summary
Development of a Continuous
Monitor for Detection of Toxic
Organic Compounds
T. Hadeishi, R. Mclaughlin, J. Millaud, and M. Pollard
This research and development pro-
gram was undertaken by the Lawrence
Berkeley Laboratory to investigate the
application of a new analytical tech-
nique called tunable atomic line mo-
lecular spectroscopy (TALMS) to the
detection and monitoring of benzene
and other volatile organic molecules of
concern to the Environmental Protec-
tion Agency. Previous studies led to
the design, construction, and delivery
to the Environmental Monitoring Sys-
tems Laboratory at Research Triangle
Park, North Carolina of a relatively
large laboratory TALMS spectrometer.
The goal of the present program was the
design, construction,and delivery of a
smaller continuous monitor for benzene
and other organic compounds.
The most limiting design factor was
found to be the detection limit of the
instrument. To improve this limiting
factor, the following areas were in-
vestigated: intensity of light source
line; location of atomic line relative to
molecular absorption feature; light
source noise; and magnetic field direc-
tion. New developments in TALMS
instrumentation that resulted from the
investigation of these factors include:
a a new light source with high in-
tensity and high stability that can
excite the spectra of many ele-
ments;
b. a negative feedback circuit to
further control light source sta-
bility;
c. instrumentation for double beam
operation;
d. a technique for more rapid loca-
tion of atomic lines that match
molecular absorption.
These developments have improved
the sensitivity of the TALMS benzene
monitor by a factor of 100 to a detec-
tion limit of 3 ppm-v benzene. Cryo-
genic trapping procedures, especially
adapted to the monitor, were also used
to improve sensitivity. Additional con-
centration factors of 100 to 1000 may
be achieved using these procedures.
Estimates of linear dynamic range, pre-
cision, interferences, and detection
limits for benzene are presented. Ap-
proximate detection limits for formal-
dehyde and phenol of 3 ppm-v were
also determined. The effects of tem-
perature and pressure upon the TALMS
signal are also discussed.
The instrument that was delivered is
compact in size (41 inches in length) and
weight (75 IDS.) and requires a modest
amount of power (125 watts). It was
supplied with a permanent magnet of
field strength 7 kG and utilizes the Hg
253.6 nm line in a double beam con-
figuration. It can be equipped with a
monochromator and is designed to
interface with a Hewlett-Packard 85
computer. A cryogenic concentrator
was designed for use with this instru-
ment Block diagrams, photographs of
the instrument, descriptions of its
components, and an operating manual
are included.
This Project Summary was developed
by EPA's Environmental Monitoring
Systems Laboratory, Research Triangle
Park, NC. to announce key findings of
the research project that is fully docu-
mented in a separate report of the
same title (see Project Report ordering
information at back).
-------
Introduction
Without dependable chemical analysis
the environmental protection regulatory
process becomes very difficult. The com-
plexity of environmental analytical prob-
lems is great both in terms of the variety of
compounds that must be detected and in
terms of the complex matrices that must
be analyzed. The monitoring of trace
amounts of organic compounds is espe-
cially difficult. At low concentrations many
ambient air samples contain hundreds- if
not thousands - of different organic com-
pounds. An analytical method that is
dependable when dealing with this type of
sample must include a technique for uni-
quely identifying these many possible
species. Present approaches almost always
involve some type of chromatographic
separation that depends upon small dif-
ferences in solubility or volatility. It follows
that shortcomings in present chromato-
graphic methods impose similar limita-
tions on almost all techniques now used to
determine trace organic compounds. In
addition, this common origin of method-
ologies makes it impossible to confirm an
analysis with a totally independent method,
which is a very important part of quality
assurance.
Previous studies carried out at the
Lawrence Berkeley Laboratory have dem-
onstrated that an optical technique, which
does not depend upon chromatographic
separations, shows promise of being a
successful new method for the determina-
tion of trace organic compounds in am-
bient air (Spectrochimica Acta 37B, 501 -
509, 1982). This technique is called
tunable atomic line molecular spectroscopy
(TALMS). As a part of the previous study a
laboratory TALMS instrument was con-
structed and delivered to the Environ-
mental Protection Agency, Environmental
Monitoring Systems Laboratory in Re-
search Triangle Park, North Carolina.
TALM Spectroscopy consists of splitting
a source atomic emission spectral line by
means of a magnetic field (Zeeman effect)
and making a differential absorption mea-
surement between one Zeeman compo-
nent that has been magnetically tuned to
an analyte rotational-vibrational, absorp-
tion line and the other Zeeman compo-
nent. The difference in polarization be-
tween Zeeman components permits the
matching and nonmatching wavelengths
to be alternately selected and the dif-
ferential absorption measured very rapidly
with an electro-optical device called a vari-
able phase retardation plate. Since the
wavelength separation between Zeeman
components is small, a signal will only be
obtained if the analyte contains a sharp
absorption feature, i.e. less than 3 cm"1
bandwidth. The resolution of this tech-
nique only depends upon the line width of
the atomic emission line and exceeds
500,000.
Thus, the TALMS signal depends upon
a high-resolution (>500,000), differential
ultraviolet visible, absorption measure-
ment and should be free of the limitations
of chromatographic separations. The high
resolution capabilities of the technique
have been demonstrated in a recent pub-
lication on formaldehyde (Journal of Molec-
ular Spectroscopy 92, 272-275, 1982).
One feature of TALMS is its essential
freedom from background interference.
Since the wavelength separation between
the Zeeman components is typically 0.04
nm, any particle scattering or semi-
continuous absorption will affect both
components equally. Therefore, the dif-
ferential absorption measurement will re-
move this interference from the signal.
Hence, this type of interference, which is a
major problem with most spectroscopic
methods, does not effect the TALMS
measurement The limitations on the sen-
sitivity of TALMS in ideal situations are
essentially the same as those on ultraviolet
absorption spectroscopy.
In the previous study TALMS signals
were detected for benzene and chloro-
benzene using the Hg 253.6 nm line. The
objective of the present study is to design
and construct a small continuous monitor
for benzene and other toxic organic com-
pounds in ambient air. In the design of
such an instrument a number of factors
are very important These include instru-
ment size, complexity of operation, sensi-
tivity, interferences, precision and accuracy,
and cost Perhaps the most important of
these factors is sensitivity (lower limit of
detectability) since benzene occurs in am-
bient air at the part per trillion to part per
billion by volume level. This report de-
scribes the development of techniques
that have led to a great improvement in
TALMS instrumental performance over
the previous system. The instrument that
resulted incorporates these techniques
and is designed to be interfaced to a HP-
85 microprocessor.
Conclusions and
Recommendations
This work has been primarily concerned
with the design and construction of a
relatively small monitor for benzene and
other toxic organic compounds that utilizes
the TALMS technique. The most critical
performance parameter in the design and
construction of the monitor has been the
detection limit required for ambient air
analysis. The following areas have been
investigated to improve this limiting
parameter:
a. Light source construction and opera-
tion
b. Effects of magnetic field direction
c. Double beam operation
d. Negative feed back control of the
light source
e. Optimized matching of atomic emis-
sion lines to the molecular absorption
feature
f. Cryogenic trapping.
In order to monitor for benzene with the
mercury 253.6 nm line, it was found that
the configuration employing a magnetic
field perpendicular to the direction of light
emission is required.
As a result of these experiments, it was
possible to incorporate effective modifica-
tions into a prototype benzene monitor
and improve the lower limit of detection by
a factor of 100 over that of the previous
laboratory instrument The resulting de-
tection limit for benzene (3 ppm-v) should
be sufficient for monitoring urban air near
sources or chemical waste sites. An addi-
tional detection limit improvement by a
factor of 100-1000 may be attained using
cryogenic trapping. Any further improve-
ments in the detection limit will have to
come from utilizing more intense absorp-
tion features in the benzene absorption
spectrum, e.g. the cobalt line at 252.9 nm.
In the process of these investigations,
phenol and formaldehyde were also de-
tected and approximate detection limits
(ca. 3 ppm-v) established. The resulting
TALMS monitor is as compact (weight: 7 5
Ibs; length: 41 inches) as possible without
sacrificing performance features. It utilizes
the mercury 253.6 nm line and a double
beam optical system and has no mono-
chromator. The monitor was delivered to
the Environmental Protection Agency in
December, 1982.
The use of this type of instrumentation
for monitoring more organic molecules
should be investigated. Operation of the
light source with a greater variety of
elements should be tested both to allow
the detection of additional molecules and
to improve the sensitivity for molecules
that have already been detected. More
detailed information concerning the oc-
currence of the rotationally sharp absorp-
tion features for compounds of interest to
the Environmental Protection Agency
should be obtained. The TALMS tech-
nique should be compared with other
methods of organic analysis. Efforts should
-------
be expended to further reduce size and
weight of the instrument with the goal of
developing a portable instrument for field
use, particularly near waste sites.
T. Hadeishi, R. McLaughlin, J. Millaud, and M. Pollard are with the University of
California, Berkeley, CA 94720.
D. R. Scott is the EPA Project Officer (see below).
The complete report, entitled "Development of a Continuous Monitor for Detection
of Toxic Organic Compounds, "(Order No. PB 83-234 922; Cost: $11.50, subject
to change) twill 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
AU.S GOVERNMENT PRINTING OfFICE 1983-659-017/7205
-------
United States Center for Environmental Research
Environmental Protection Information
Agency Cincinnati OH 45268
Official Business
Penalty for Private Use $300
PS OGOQ329
U S fciMVIR PKOTtCriUN AbENCY
RfeClOM 5 UtoRAKf
230 S DtAKdURN S
C«ItAi>0 IL 606U4
-------
United States
Environmental Protection
Agency
Environmental Monitoring Systems
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S4-83-033 Aug. 1983
Project Summary
Computer Simulation of the
EPA Provisional Method for
Measuring Airborne Asbestos
Terence Fitz-Simons and Michael E. Beard
A computer simulation program was
developed to reproduce manual count-
ing methods and calculate their accu-
racy in estimating the number of asbes-
tos fibers on a filter surface. A model
arrangement of asbestos fibers on a
filter was generated fora predetermined
number of fibers with lengths and widths
according to lognormal distribution and
uniformly random placement. These
hypothetical fibers were next counted
by computer in a program simulating
manual microscopy estimating proce-
dures. The protocols proved to have a
quantifiable error factor when the com-
puter counting results were compared
with the predetermined, original total of
model fibers. The bias resulted because,
in the counting protocol, fibers on the
sample grid having an aspect ratio less
than 3 were not included. The mass
estimates proved correct at light load-
ings but were biased low at heavy load-
ings. It is suggested that most of the
mass is concentrated in the large fibers;
thus, at light loadings these fibers are
well sampled due to their size. At heavier
loadings they are more likely to extend
past the field of view and their size is
more likely to be underestimated.
This Project Summary was developed
by EPA's Environmental Monitoring
Systems Laboratory, Research Triangle
Park, NC, to announce key findings of
the research project that is fully docu-
mented in a separate report of the same
title (see Project Report ordering infor-
mation at back).
Introduction
This report describes a computer simu-
lation of the EPA provisional method for
measuring airborne asbestos.
The unique physical properties of as-
bestos have encouraged widespread use
of this mineral for centuries in a variety of
applications. A large industry has devel-
oped around usage of asbestos as a
result. Unfortunately, exposure to air-
borne asbestos fibers adversely affects
the respiratory system by reducing lung
capacity, and recent studies have related
various forms of lung cancer to asbestos
exposure. The Environmental Protection
Agency (EPA) and the scientific commu-
nity believe that any level of exposure to
asbestos involves some health risk, al-
though the exact degree of risk cannot be
reliably estimated. EPA has moved to
control emissions of asbestos because of
its widespread use and hazardous nature.1
Because of the difficulty in controlling all
sources of emissions, the need exists to
monitor airborne asbestos.
Airborne asbestos fibers range in length
from a few micrometers down to sub-
micron sizes. Median airborne fiber
lengths reported in the literature range
from about .5 yum to as much as 5.0 j/m2'3
and are best measured using electron
microscopy. Asbestos fibers are identified
by morphology, chemical composition,
and crystal structure. Morphology is
determined by direct observation under
the microscope. Chemical composition is
determined byx-ray fluorescence. Crystal
structure is determined by selected area
electron diffraction.
The EPA provisional methodology for
measurement of airborne asbestos em-
ploys samplers ranging from high volume
using 8" x 10" filters to personal samplers
containing circular filters 37 mm in
diameter. Transmission electron micros-
copy (TEM) is used to identify and measu re
airborne fibers. Because several charac-
-------
teristics of fibers are considered in health
studies, fiber count, length, width, and
mass are reported. Collecting fibers on
filters presents the fewest problems in
the measurement process described in
the provisional method; however, micro-
scopic analysis of the filters presents
many problems. The method is an uneasy
compromise between statistical sampling
and TEM microscopy. The statistician
worries about analyzing a large enough
portion of the filter to make meaningful
estimates, while the microscopist worries
about analyzing a sample that is small
enough to complete the job under time
and budget constraints.
Following collection of fibers, a circular
section 3 mm in diameter is removed
from the filter for TEM analysis. The
section is placed on a TEM grid to identify
positions in the sample, where up to 10
grid openings (75 to 100-/um squares) are
examined according to a strict counting
protocol. Fiber counts on the filter are
estimated from fiber counts in the sample
section multiplied by the ratio of filter
area sampled to total filter area. This ratio
is usually in the neighborhood of 10,000.
The importance of a well conceived
counting protocol is obvious.
Testing the counting protocol has been
accomplished by repeated experimental
observations by TEM which are tedious,
time consuming, and expensive. This
report describes a computer simulation of
the counting protocol. The program was
developed on the UNIVAC at the EPA
National Computer Center. The program
used IMSL4 and TEKTRONIX software
and hardware.
Results
The computer simulation of human
estimation methods for asbestos fibers
on a filter produced a total of 1,161,023.4,
as opposed to the true predetermined
fiber count of 1,570,779 programmed for
the research. The estimated mass on the
filter was 0.002123, as opposed to an
actual figure of 0.00235320815. The
simulation identified error factors in
human counting methods that were 26%
low for total number and 16% low for
mass.
Conclusions and
Recommendations
If the model is a reasonable representa-
tion of the interaction between the provi-
sional method and airborne asbestos,
then the model indicates that the method
provides fiber count data that are biased
low. The most plausible reason for this is
that the method calls all objects with a
ratio of length to width (aspect ratio) less
than 3 to be deleted from further consider-
ation. The simulation model recognizes
the existence of asbestos fibers with
aspect ratios less than 3. It is uncertain
whether or not objects with aspect ratios
below 3 can really be considered asbes-
tos; and even if they are asbestos, there is
uncertainty as to health effects due to
such short fibers.
Mass estimation is also biased low
under higher loadings. A probable cause
for this is that mass is concentrated in
larger fibers. Under the provisional
method, large fibers are not fully meas-
ured when filters are heavily loaded with
fibers. This truncation may result in
biased mass estimates. When there is a
lighter loading, the entire fiber is meas-
ured, thus negating this bias.
Assumptions in the simulation model
targeted for refinement are (1) fibers do
not bend, (2) all objects below an aspect
ratio of 3 are still fibers, and (3) there are
no operator errors in following the proto-
col in measurement and identification.
References
1. Code of Federal Register. Title 40,
Part 61, Subpart B, National Emission
Standard for Asbestos. U.S. Govt.
Printing Office, Washington, DC,
Revised July 1, 1977.
2. Ettinger, H. J., C. I. Fairchild, L. W.
Ortiz, M. I. Tillery. Aerosol Research
and Development Related to Health
Hazard Analysis, LA-5359-PR, Los
Alamos Scientific Laboratory of the
University of California, Los Alamos,
CA, 1973.
3. Johnson, W., A. Berner, G. Smith, J.
Wesolowski. Experimental Determi-
nation of the Number and Size of
Asbestos Fibers in Ambient Air,
Report No. ARB-R-3-68B-76-45, Cali-
fornia Air Resources Board, 1975.
4. IHSL Subroutine Library. Vols. 1 and
2, International Mathematical and
Statistical Library, Inc., Houston, TX,
1975.
The EPA authors Terence Fitz-Simons and Michael E. Beard (also the EPA
contact, see below) are with the Environmental Monitoring Systems Laboratory,
Research Triangle Park, NC 27711.
The complete report, entitled "Computer Simulation of the EPA Provisional
Method for Measuring Airborne Asbestos," (Order No. PB 83-231 852; Cost:
$7.00. subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
Michael £. Beard can be contacted at:
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park. NC27711
,'rU.S. GOVERNMENT PRINTING OFFICE 1983-659-0)7/7)54
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
0000329
------- |