Environmental Protection Technology Series
INFRARED ABSORPTION BY
SULFURIC ACID VAPOR
Environmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-76-191
July 1976
INFRARED ABSORPTION BY SULFURIC ACID VAPOR
by
Darrell E. Burch, Francis J. Gates
and Norman Potter
Aeronutronic Ford Corporation
Aeronutronic Division
Newport Beach, California 92663
Contract No.
68-02-1774
Project Officer
Roosevelt Rollins
Emission Measurement and Characterization Division
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, 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 con-
stitute endorsement or recommendation for use.
ii
-------
ABSTRACT
A sample cell has been designed and built to contain hot l^SO^ vapor for spec-
troscopic analysis. The cell is heated to between 190° and 250°C by an electric
furnace designed specifically for use with the sample cell. A small bulb con-
nected to the main compartment of the sample cell contains liquid l^SO^. The
partial pressure of the t^SO^ vapor is controlled by regulating the temperature
of the liquid 112804, which is at lower temperature than any other part of the
cell that is in contact with the vapor. Transmission curves have been scanned
for a variety of samples over the 7.8 g.m - 12.8 |j,m region with spectral slit-
widths between 0.5 cm'l and 4 cm" . At wavelengths of maximum absorption near
8.2 M,m and 11.4 urn, the absorption coefficient is approximately 0.0004 (ppm -
meters)""^.
iii
-------
iv
-------
CONTENTS
Page
Abstract ±±±
Figures vi
I Introduction 1
II Summary 4
III Conclusions 5
IV Recommendations 7
V Experimental Apparatus and Methods 10
VI Results of Transmission Measurements 18
-------
FIGURES
Number
Optical diagram of the apparatus used to study hot H_SO
2
3
4
5
6
7
8
Two views of the sample cell. .
End view showing the location of
t. "r
the silicon windows and the
Side view of the furnace with the sample cell in place
Spectral curves of transmittance for three samples of H,,SO, . .
Spectra of H.SO, and H_0 between
Spectra of H,,SO, and CO,, between
7 . 85 |im and 8 . 45 urn
11
12
14
15
19
20
21
26
vi
-------
SECTION I
INTRODUCTION
Under a previous contract, Aeronutronic Ford Corporation designed, fabricated,
tested and delivered a cross-stack monitor for either HF, HCl, CO, NO or S02
that is based on the gas-filter correlation (GFC) principle. The instrument was
delivered to EPA in June 1974, and extensive tests performed by EPA personnel
since that time have shown that it meets design specifications for HF, CO, NO,
and S02. Field evaluations have not been performed for HCl, but no problems are
anticipated. During the field tests it was determined that a few minor modifi-
cations to the instrument would simplify the operation and improve the performance.
It was originally intended that the instrument would also measure S03 in addition
to any of the other 5 gas species, but during the design phase it was determined
that the GFC technique is not suitable for S03 because of the difficulties in
maintaining a known constant amount of this gas in a small cell.
The original objectives of the project reported herein were to (1) make recom-
mended modifications to the existing instrument to simplify operation and to
improve its performance, and (2) retrofit the existing instrument with a dis-
persive correlation accessory that would allow either S02 or S03 to be measured.
The gas of primary importance was S03; however, the capability of measuring S02
was to be included because the additional cost would be small. Furthermore, the
expected performance for S02 was better than that obtained with the instrument
when employing the GFC technique.
Objective number 1 listed in the previous paragraph has been completed. During
the early part of the present contract, EPA scientific personnel investigated
the complex chemical reactions involving S03 that occur in the stacks of station-
ary sources. The results of this investigation indicated that the typical ef-
fluent probably contains more H2S04 than S03. Much of the S03 that is formed
probably reacts with the H20 that is present and forms H2S04. Thus, it was de-
cided that the need for an H2S04 monitor was greater that that for an S03 monitor.
1. Burch, D. E., and D. A. Gryvnak. "Infrared Gas Filter Correlation Instrument
for In-Situ Measurement of Gaseous Pollutants." Prepared by Philco-Ford
Corporation for EPA under Contract No. 68-02-0575. EPA Report No. EPA-650/2-
74-094. Also, Burch, D. E., and D. A. Gryvnak. "Cross-Stack Measurement of
Pollutant Concentrations Using Gas-Cell Correlation Spectroscopy." Chapter 10
of Analytical Methods Applied to Air Pollution Measurements, Stevens, R. K.
and W. F. Herget, (eds.). Ann Arbor, Ann Arbor Science Publishers Inc., 1974.
-------
Therefore, the work on the dispersive correlation accessory for 862 and SOo was
stopped, and the effort was diverted to an investigation of a cross-stack method
of monitoring hot HoSO^ vapor. Possible problems in maintaining the chemical
integrity of a sample as it is drawn through a probe to a remote monitor make
this type of instrument much less desirable than a cross-stack instrument that
does not disturb the sample.
2
Alpert has published very high resolution spectra of E2SO^ vapor in the spectral
region from 1312 to 1549 cm"1. The H2S04 vapor appeared as an impurity in
Alpert 's samples of HCN, the gas species of primary interest to him. Alpert 's
spectral curves show approximately 700 separate absorption lines of I^SOA, and
the positions of more than 400 of them have been measured and tabulated to the
nearest 0.001 cm'1. These data are the only data known to the authors that
show individual I^SO^ absorption lines. The intensities of the lines are quite
irregular, and the spacings between adjacent lines vary with an average of ap-
proximately 0.3 cm'1. Because of the strong absorption by H20, S02 and SOo in
the same spectral region, this H2S04 band is of little value for monitoring
purposes.
o
Chackalackal and Stafford have published low resolution spectra of hot H2S04
vapor in the spectral region 400 - 4000 cm"1. The vapor was produced by heat-
ing the azeotropic mixture (98.37o) of liquid l^SO^. Some S03 was also present
in the vapor over the hot liquid; the percentage of S0
-------
re 90 ,^ ?9?2 4-l8?f i* y ^ centere? near 361° cm'1 (2.77 pa), 1450 cnf1
(6 90 jun) 1224 cm'l (8.17 ym), and 882 cm-1 (11.34 p»). The first band listed
overlaps strong absorption by H20 and C02. The second band listed , the one
on which data are given in Reference 2, is also overlapped by an H20 band as
well as by S02 and S03 bands. The 8.17 \m band and the 11.34 urn band have been
emphasized in the present study because they occur where there is much less
absorption by gases likely to be present in samples to be investigated for
H2S04 vapor.
-------
SECTION II
SUMMARY
A sample cell has been designed and built to contain hot H2S04 vapor for spec-
troscopic analysis. The cell is heated to between 190° and 250°C by an electric
furnace designed specifically for use with the sample cell. A small bulb con-
nected to the main compartment of the sample cell contains liquid l^SO^. The
partial pressure of the l^SO^ vapor is controlled by regulating the temperature
of the liquid H2SO^, which is at lower temperature than any other part of the
cell that is in contact with the vapor. Typical temperatures of the liquid
H2S04 and of the vapor being analyzed are approximately 180°C and 210°C, res-
pectively.
The main body of the cell is made of glass, and the gasket material used to
provide a leak-tight seal is resistant to chemical reaction with the E2SO^
vapor. Two silicon windows make it possible to transmit a monitoring beam of
infrared radiant energy through the hot I^SO, vapor. A gold-coated spherical
mirror forms one end of the cell and reflects the monitoring beam so that the
cell is double-passed with a total optical path length of 52 cm through the
vapor.
A small grating spectrometer has been employed to scan spectral curves of
transmittance for several samples over the 7 am to 13 am spectral region. In
addition to the H2SC>4 vapor, some 803 vapor and H20 are also present in the
sample cell. Two strong H2S04 bands occur in this region; wavelengths of maxi-
mum absorption are near 8.2 am and .11.4 urn. Adjacent individual absorption
lines within each band are probably separated by only approximately 0.3 cm"1
as they are in the 6.90 am band (see Reference 2). These closely spaced
individual lines were not resolvable by our spectrometer, which has a minimum
practical spectral slitwidth of approximately 0.4 cm'1 in this portion of the
spectrum.
The absorption by both the 8.2 am band and the 11.4 am band is quite strong.
Approximately 250 ppm-meters of 117804 vapor produces an average absorptance of
10 eal of a few nrl
over an interval of a few cnr near the strongest absorption in either of
these bands.
-------
SECTION III
CONCLUSIONS
Hot H2S04 vapor has four relatively strong absorption bands between 2 urn and
12 M-m. The two centered near 2.77 \ua and 6.90 p,m are probably of little value
for monitoring the I^SO^ concentration because of the strong H20 absorption in
the same spectral regions. The other two bands, centered near 8.2 um and
11.4 p.m, are much more promising for monitoring purposes. Some interference
by H20 is to be expected in the 8.2 \ua region; however, this interference can
probably be accounted for if the H20 concentration is known or is also monitored.
Interference by hot C02 might be a problem for a monitor that makes use of the
11.4 pm band. This problem would probably be most serious when investigating
the exhaust from a coal-burning power plant in which the C0? concentration may
be as high as 10-20%. Most of the error due to interference by C02 can be ac-
counted for by measuring the interference of the monitor to known concentra-
tions of C02 and determining, by other methods, the concentration of this gas
species in the gas being monitored.
Although the absorption by H2S04 is quite strong, the absorption spectrum does
not contain distinct, strong, well-separated features that make infrared de-
tection and discrimination simple for some gases. From other experimental
studies,^ we find that the individual absorption lines have quite irregular
spacings and that the average spacing between adjacent strong lines is usually
less tnan U.3 cm . Thus, any dispersing instrument designed to take advan-
tage of the fine spectral structure must have quite good spectral resolution
particularly for a portable field instrument. The most promising infrared
monitor for H2S04 would take advantage of the coarser, less-sharp, structure
in the infrared spectrum. Such a monitor might involve a dispersing element
or a series of narrow-bandpass filters. Because it employs the coarse structure
in the spectrum, this type of instrument is more subject to interference by
other gas species than is an instrument that makes use of several strong,
sharp absorption features.
Gas-filter correlation techniques are commonly employed to exploit spectral
structure that is not resolved by the instrument. These gas-filter correla-
tion techniques could probably be employed successfully for H2S04 vapor ex-
cept for the difficult problem of confining a fixed and known amount of H2S04
vapor in the gas-filter cell, the key component of such an instrument. The
high reactivity of the vapor makes it difficult to maintain a vacuum tight
seal and to prevent the disappearance of the vapor as it reacts with the walls
of the cell and with the windows. Furthermore, the gas-filter cell would
have to be heated in order to maintain an adequate pressure of the H^SO^ vapor.
-------
Maintaining a constant gas-filter cell temperature in a field instrument is
also a difficult problem. Because of these problems, a gas-filter correlation
instrument does not seem to be practical for I^SO,/, vapor
-------
SECTION IV
RECOMMENDATIONS
For the near future, the most promising type of infrared, cross-stack monitor
for H2S04 vapor would contain a grating as a dispersing element and would make
use of either the 8.2 u-m band or the 11.4 M-m band. Quite good sensitivity can
be obtained by an instrument that compares the transmission in two or more
narrow spectral intervals. If only two intervals are compared, one should be
in a region of strong absorption by l^SO^ vapor, and the other should be in a
nearby region where the absorption is much weaker. It may be desirable to choose
two or more intervals of strong absorption and compare their combined trans-
mission to that of two or more intervals of weak absorption.
The simplest, and probably the most dependable, method of performing the desired
comparison involves using a small grating polychromator with multiple slits in
the focal plane of the instrument. Each slit passes one of the selected narrow
spectral intervals. A chopper located just back of the slits alternately allows
energy to pass through the slits corresponding to the intervals of strong absorp-
tion then through those corresponding to weak absorption. A dispersive instru-
ment with sufficiently high spectral resolution to make use of the fine structure
due to lines separated by approximately 0.3 cm'l does not appear to be practical
for field monitoring. Therefore, it is necessary to use the coarser structure
in the spectrum; there is probably little to be gained by making either spectral
interval narrower than about 4 cm'1. (This corresponds approximately to 0.03 p-m
near 8.2 p,m and to 0.05 p,m near 11.4 p,m.)
The choice of the spectral intervals will depend as much on possible interfering
gases as on the spectral structure of the H2S04 vapor. Minimizing interference
by H20 is essential if the 8.2 |im band is employed. If the 11.4 p-m band is em-
ployed, the interference by H20 is much less, but there may be interference by
C02 if this gas occurs in high concentrations, as it does, for example, in the
exhaust from coal-burning power plants. Other gases that may be present in the
sample and interfere with the measurements must also be considered. Ammonia,
for example, absorbs strongly at many places within the 11.4 gun H2S04 band and
could interfere if it were present. Data shown in Section VI on the absorption
by H20 can be useful in selecting the proper intervals to reduce interference
by this gas.
The instrument built previously by Aeronutronic Ford for EPA (Reference 1) can
be adapted easily to monitor hot l^SO^ vapor by employing two or more narrow
spectral intervals as just discussed. A grating assembly designed and built
specifically for this purpose would fit into the space used for grating assemblies
provided for other gases. As in the assemblies built previously, the grating
-------
and all of the mirrors are fastened permanently into position. Unlike the
previously built assemblies, a chopping mechanism alternately allows the pas-
sage of energy through the various slits in the retroref lector-grid assembly.
The energy transmitted through the slits returns to the grating and is undis-
persed. After leaving the grating assembly, the energy passes on to the de-
tector. The gas-filter cell and the rotating mirror chopper normally used are
not employed with this assembly. The detector signal is processed by the same
two- frequency electronics that have been provided with the instrument.
This instrument with a grid-chopper is recommended primarily as a research in-
strument to obtain approximate values for the concentration of hot HoSO< vapor
in the exhaust of various pollution sources and to investigate the problems
involved in making accurate and reliable measurements on a routine basis. The
instrument can be tested and calibrated in the laboratory by using a heated
cell similar to the one described in Section V to contain the l^SO^ vapor.
The same cell, but with the liquid t^SO^ removed, can contain hot CC^ , 1^0, or
any other gas that might interfere with the measurement.
With little additional effort, the grating assembly can be made so it is capable
of using either the 8.2 p,m band or the 11.4 p.m band. A multiple-slit grid
would be provided for each band, and the grid corresponding to the band not
being used would be blocked while the other was left open. Not more than a
few minutes would be required to change from one band to the other. Compari-
son of the results obtained with the two different bands would provide informa-
tion on the reliability of the instrument and on the interfering gases. A
similar set of slits could also be included near 8 urn to measure the concentra-
tion of H20. By previously measuring the interference by H20 in the I^SO^
channel, the data on the HoO concentration could be used to account for the
H20 interference. A similar set of slits that pass two or more intervals be-
tween 10 and 11 urn might be used to measure the concentration of C02 so its
interference can be accounted for. A single chopping mechanism could serve
for all of the suggested sets of slits that transmit between 8 pm and 12 p,m.
Only one set of slits would be used at a given time so that only one detector
would be required.
On the basis of the data shown in Section VI, we estimate that the minimum
detectable thickness of l^SO^ for an instrument like the one just described
is between 5 ppm-meters and 30 ppm-meters. This value is limited by the un-
certainty in the amount of interference by other gases and by the available
methods of "zeroing" the instrument with no H^SO^ in the sampling path.
For the long term, the best possible infrared monitor for l^SO^ vapor would
probably contain a tunable laser that could make use of the fine structure
in the l^SO^ spectrum. As these lasers are further developed and become more
convenient to use, they can probably be applied to l^SO^ measurements. Before
the optimum wavelength interval for such a scanning laser can be determined,
it is essential that the spectrum of heated HoSO^ vapor in samples at 1 atm
total pressure be investigated with a spectral slitwidth less than 0.1 cnT^.
It would be better if the slitwidth was less than 0.05 cm"* so that most of
the structure could be observed. Such measurements require a high-quality
spectrometer with good wavelength accuracy in addition to the very good reso-
lution. A cell similar to the one described in Section V could be used to
contain the samples.
8
-------
A few minor changes should be made in the design of the cell used to contain
hot H2SO^ for future studies. The most important purpose of these changes is
to reduce as much as possible the amount of reactive material in contact with
the hot H2S04 vapor. By so doing, the samples can be made purer and there
should be much less interfering absorption in the transmission curves. The
continued usage of a glass cell body, gold-coated mirror, teflon gaskets and
silicon windows is recommended.
Better seals between the endplates and the cell body could be obtained if the
teflon gaskets could be compressed more firmly. In order to do this, the
thickness of the glass wall of the cell should be increased to at least 5 mm
to provide additional strength. Teflon gaskets, preferably 0-rings, should be
used between the silicon windows and the glass endplate. A mechanism should
be designed to press the windows firmly against the gasket without breaking
the windows. After the cell has been assembled, any small remaining leaks
around the gaskets could be sealed with RTV, or some other pliable cement that
is relatively inert and can withstand temperatures up to 250°C. Any small
amount of RTV that contacts the l^SO^ vapor will probably not decompose to
cause a leak or to produce a serious amount of impurity in the sample.
The thickness of the silicon windows should not exceed approximately 0.8 mm
so as to minimize the absorption by the silicon. Because the thin windows may
be weak, it is recommended that the diameters be 2.5 cm or less to reduce the
force on the windows when the cell is evacuated. The surface of each window
that contacts the I^SO^ vapor should not be coated with an anti-reflection
coating. The coatings apparently react with the l^SO^ and form a substance
that absorbs strongly in the 8 - 10 |0m region. An anti-reflection coating on
the outside surface of each window will improve the transmission and greatly
reduce the fringes caused by optical interference between the two surfaces of
each window. Wedging the windows slightly can also reduce, or even eliminate,
the fringes.
A teflon valve should replace the glass valve with silicon grease that is used
on the present cell in contact with the vapor. There is noticeable decomposi-
tion of the grease after a few hours use of the cell. The liquid H2S04 reser-
voir connected as an appendage to the main part of the cell has worked satis-
factorily and should be included on any future cell. The capability of heating
the liquid independently of the main body of the cell is quite important and is
recommended.
REGION III LIBRARY
ENVIRONMENTAL PROTECTION AGSHCY
-------
SECTION V
EXPERIMENTAL APPARATUS AND METHODS
The optical arrangement employed to obtain the spectral data on H2SOA vapor is
shown in Figure 1. Radiant energy from a Nernst source is chopped at 450 Hz
and is directed through the sample section to a grating monochromator. Mirror
M2 forms an image of the source near mirror M3, which is adjacent to the en-
trance window of the sample cell. Spherical mirror M4 forms one end of the
sample cell and re-images the source near mirror M5. Spherical mirror M7
forms another image of the source on the entrance slit of the grating mono-
chromator.
The sample cell is enclosed within a custom-made furnace that heats the sample
cell to between approximately 190°C and 250°C. The furnace, which is not shown
in Figure 1, surrounds the sample cell and is placed inside a plywood box. The
box contains insulation that surrounds the furnace. The radiant energy beam
passes through two pieces of glass tubing that hold the insulation away from
the monitoring beam. A portion of the furnace not shown in Figure 1 holds the
insulation away from the monitoring beam in the region near mirrors M2 and M5.
The monochromator is a modified version of a Perkin-Elmer Model 99 Prism Mono-
chromator. The original prism has been removed and replaced by a grating with
75 lines per mm that is used in the first order. A longpass interference fil-
ter placed near the entrance slit eliminates overlapping orders of shorter wave-
lengths that are also passed by the grating. The detector has a copper-doped
germanium element and is cooled by liquid helium. The detector signal is ampli-
fied and demodulated by a synchronous demodulator, the dc output of which is
recorded on a strip-chart recorder. The spectral slit-width varied from approxi-
mately 0.5 to 4 cm"-'-.
The construction of the sample cell is illustrated by the two views shown in
Figure 2. The cell is designed to have a minimum of contact between the hot
H2S04 vapor and any material with which it might react. It is necessary to
heat the sample cell in order to obtain,a sufficiently high pressure of the
^804 vapor for its spectrum to be obtained. Attached to the main body of
the cell by a piece of glass tubing is a small glass reservoir that contains
an azeotropic mixture (98.3%) of liquid ^804 in water. The pressure of the
vapor in the sample cell is controlled by regulating the temperature of the
liquid H2S04. The liquid ^SO^ is maintained a few degrees C cooler than any
other part of the cell with which the vapor comes in contact. This avoids
condensation on the inside of the sample cell. It is particularly important
that condensation be avoided on the cell windows or on the gold-coated mirror.
The vapor in the sample cell contains H2S04, 803, and H20. The relative abund-
ances of the vapors of these three gas species depend upon the temperature
10
-------
Plywood Box
Source
M2
Grating
Monochromator
Figure 1. Optical diagram of the apparatus used to study hot H2S04 vapor.
-------
Glass Endplate
TC #2 Glass Cell Body
Silicon
Windows
TC #4
TC 93
TOP VIEW
Reservoir
Heater for
Reservoir
Removable Block
END VIEW
Figure 2. Two views of the sample cell.
12
-------
and upon the amount of H_0 in the system.
One end of the sample cell is formed by gold-coated mirror M4. The center of
curvature of this mirror is near a point midway between the centers of the two
silicon windows. Thus, radiant energy coming from an image near one of the
windows is refocused to form an image near the other window, as illustrated in
Figure 1. A flat teflon gasket forms the seal between the mirror and the cylin-
drical glass tubing that makes up the main body of the sample cell. A similar
teflon gasket forms a seal between the other end of the cylindrical tubing and
a glass plate that forms the opposite end of the cell. The glass endplate and
the mirror are held in place by two aluminum endplates as illustrated in the
upper portion of Figure 2. Coiled springs on each of four pieces of threaded
rod are compressed to hold the ends firmly against the cylindrical body of the
cell. A flat ring of teflon placed between the aluminum endplate and the mirror
serves as a spacer and places the force against the backside of the mirror ap-
proximately opposite the gasket that forms the seal. Use of this ring makes the
shear on the mirror much less than it would be if the aluminum endplate made
contact with the center portion of the mirror. A similar arrangement is used
between the glass endplate and the corresponding aluminum endplate on the oppo-
site end of the sample cell.
We experienced problems in maintaining a reliable vacuum seal with the teflon
gaskets used as illustrated in Figure 2. This problem was solved by applying
a ring of RTV sealing compound around the gasket after it had been installed.
A continuous ring of RTV completely covered the portion of the teflon gasket
exposed to the outside of the cell. The RTV was bonded to the glass body of
the cell and to the mirror on one end and to the cell body and glass endplate
on the opposite end. The RTV sealed the leaks around the teflon gaskets with
only a very small amount of it in contact with the hot H2SC>4 vapor. After
operating the cell for a few days, we found no evidence of degradation of the
RTV used around the teflon gaskets. More important, there was spectral evi-
dence of only a small amount of an impurity in the sample cell that might have
resulted from a chemical interaction between the H2SC>4 vapor and the RTV.
We originally intended to employ teflon 0-rings to provide the vacuum seal
between the silicon windows and the glass endplate. Because of problems en-
countered in obtaining a sufficiently tight seal, we found a solution that
made use of the RTV in the manner illustrated in Figure 3. Each 25 mm diameter
silicon window was placed over a circular hole in the glass endplate. The
window was held firmly against the glass plate and the RTV seal was applied
around the outer edge of the window as indicated. No RTV was applied to the
flat portion of the window that would be in contact with the hot H2SO^ vapor.
Although a small amount of RTV is necessarily in contact with the H2S(>4 vapor,
we found that the impurities in the sample that result from an interaction
between the vapor and the RTV was not serious when the windows were installed
carefully by the method illustrated. In one instance, one of the windows was
not sealed properly. The RTV was not completely cured when the sample cell
was evacuated, and a small amount of RTV was forced by the higher outside
pressure through the gap between the silicon window and the glass into the
interior of the sample cell. When the cell was heated, the hot H^SO^ reacted
with the RTV and apparently formed an impurity with a spectrum that interferes
with the spectrum of the H2S04. When the cell was dismantled for cleaning, it
was apparent that the H2S04 had reacted strongly with the RTV that it had con-
tacted.
13
-------
Aluminum Endplate
\
Glass End Plate
RTV Seal
Figure 3. End view showing the location of the silicon
windows and the method of sealing them.
\
The silicon windows that were first intended for use are 2 mm thick and contain
an anti-reflection coating on each surface to maximize transmission near 10 p,m.
Although this coating is optimized for 10 u-m, it reduces reflection throughout
the entire 7-12 gun region investigated. Absorption by the silicon was bothersome,
but it did not seriously affect the quality of the transmission data. Fortunately,
the most prominent silicon absorption feature is at approximately 9 |im where it
does not interfere significantly with either of the two strong H^SO^ absorption
bands investigated. A thinner pair of windows, each 0.8 mm thick, were available
in the laboratory and were used in order to reduce the absorption by the silicon.
However, these thinner windows did not contain an anti-reflection coating, and a
problem was created by the interference between the radiant energy reflected from
the two parallel surfaces of each window. This interference produced "fringes"
in the spectrum of the transmitted energy; the fringes were separated by approxi-
mately 2 cm'l and caused more than 20% difference between the maximum and the mini-
mum transmittance. After these tests were completed, the thin windows were removed,
and the original 2 mm thick windows with anti-reflection coating were replaced on
the cell. We then found that the t^SO^ vapor reacted with the anti-reflection coat-
ing and produced a film on the windows that absorbed strongly between approximately
8 urn and 10 p,m. The best data were obtained with the thin windows and with the
monochromator slits wide enough to smooth-out the fringes produced by the windows.
A rectangular hole is cut in the aluminum endplate opposite the windows to allow
passage of the radiant energy.
The method employed to heat the body of the sample cell and the liquid reservoir
is illustrated in the lower panel of Figure 2 and in Figure 4. The stationary
core that makes up the main part of the furnace surrounds the body of the sample
cell and is made from aluminum square-stock approximately 3 mm thick and 100 mm
on each side. Insulated electrical heating wire is wrapped around this alum-
inum as indicated in the figures. The glass tubing that connects the
14
-------
Plywood Box
Hoveable Core of
Furnace
Heating Wire
Stationary Core
of Furnace
Aabeitot Sheet
Figure 4. Side view of the furnace with the sample cell in place.
-------
main body of the sample cell to the liquid HoSOA reservoir protrudes through a
slot in the side of the stationary aluminum core. This allows the temperature
of the reservoir to be regulated almost independently of the temperature of the
main body of the cell. The reservoir is enclosed within two cylindrical pieces
of ceramic tubing with a ceramic cap over the top as indicated in the lower
portion of Figure 2. Heat is supplied to the reservoir by electrical heating
wire wrapped around the lower piece of ceramic. Additional heating wire, not
shown in Figure 2, is wrapped around the glass tubing that connects the reser-
voir to the main body of the cell. The heating wire for the core of the furnace
is illustrated in Figure 4, but not in Figure 2.
A moveable core of the furnace was employed as indicated in Figure 4 to provide
extra heat in the region near the silicon windows. This short, moveable core
is made of aluminum square-stock similar to that of the stationary core. After
the cell has been put in place inside of the stationary core of the furnace, an
optical assembly that contains mirrors M3 and M5 is fastened in place and the
moveable core is installed. Holes cut in the two sides of the moveable core
allow the light to pass through to mirror M3 and out from mirror M5. The bottom
side of the moveable core is also notched out so that it can slide over the
mount for the mirrors M3 and M5.
Pieces of flat asbestos sheet, approximately 5 mm thick, are fastened to the four
sides of both the stationary core and the moveable core of the furnace. In
addition to providing some insulation, these asbestos plates hold the coils of
heating wire firmly against the aluminum square-stock that forms the cores of
the furnace. After the sample cell and the furnace have been installed and
aligned, the plywood box that contains the sample cell and furnace is filled
with loose insulation. The two pieces of glass tubing shown in Figure 1 keep
the insulation from interrupting the monitoring beam as it enters and exits
from the sample cell and furnace assembly. When the reservoir is to be main-
tained at a low temperature in order to reduce the amount of ^SO^ vapor in
the cell, the insulation is removed from around the reservoir. With the insula-
tion removed from around the reservoir and the remainder of the sample heated
to approximately 200°C, the liquid H^O^ in the reservoir is typically at ap-
proximately 45 °C.
The heating wires are connected in five separate circuits, and each is con-
trolled independently in order to maintain the temperature of the different
components as desired. One heater circuit includes the wire around the sta-
tionary core, one includes the wire around the moveable core, and a third is
connected to the moveable block that supports the mounts for mirrors M3 and M5.
The fourth heater circuit is shown in the lower panel of Figure 2 and provides
heat to the liquid ^SO^ reservoir. The fifth heater circuit is not shown in
any of the figures; it includes a few coils of wire around the glass tubing
that connects the reservoir to the sample cell and a few coils around the three
glass valves and the tubing that connects them. The power in each circuit is
manually controlled by a variable transformer. Approximately 130 watts are
required to maintain the sample cell at 210°C.
Figure 2 shows the location of five thermocouples attached to the sample cell.
By carefully adjusting the power to each of the individual heaters, it is pos-
sible to maintain the temperatures indicated by thermocouples 2, 3 and 4 to
within 2°C. Thermocouple No. 1 indicates the temperature of the liquid
16
-------
in the reservoir, which can be controlled almost independently of the other
parts of the cell.
In a typical experiment the sample cell is heated to approximately 210°C with
all of the insulation in place except for that around the reservoir so that its
temperature remains near ambient. Under this condition, the vapor pressure of
the H2S04 is so low that no absorption by the vapor can be observed with the
spectrometer. Therefore, a spectral curve scanned with the cell in this con-
dition can serve as a 100% transmittance curve to be compared with the spectral
curves obtained with vapor in the cell. When the heater for the reservoir and
the insulation have been put in place, the liquid H2S04 can be heated and
stabilized in approximately 30 minutes.
After a series of spectral curves have been obtained at different vapor pres-
sures, the insulation from around the reservoir and the heater for the reservoir
are removed to allow the liquid to cool. As the liquid cools, the vapor in the
cell condenses in the reservoir. The decrease in the absorption is monitored
by recording the output from the synchronous amplifier with the grating mono-
chromator adjusted to pass a wavelength of strong absorption by the vapor.
The total pressures of a few of the samples were measured by using the pressure
transducer shown in Figure 1. The line to the pressure transducer and the
volume bounded by the three glass valves was filled with N2 to a pressure slightly
higher than the estimated pressure in the sample cell. The valve to the cell was
then opened for a few seconds to allow a small amount of N2 to flow into the cell
until the pressure in the cell was the same as that in the transducer. A high
impedance recorder measured the output of the pressure-transducer, which had been
calibrated previously. The volume filled with N2 before it is let into the cell
is approximately 37. of the volume of the cell; therefore, the N2 that flows into
the cell, and thus adds to the pressure, is easily accounted for. By following
the method described above, the N2 in the line served as a buffer and prevented
the flow of H2SC>4 vapor into the room-temperature pressure transducer where it
wouId c ondens e.
17
-------
SECTION VI
RESULTS OF TRANSMISSION MEASUREMENTS
Figures 5 and 6 show representative spectral curves of transmission obtained
for hot H2S04 samples at different pressures. Background curves obtained with
no H2SO^ in the sample cell contained some coarse structure due to absorption
by the silicon windows, spectral variations in the transmission of the inter-
ference filter, and spectral variations in the efficiency of the grating. This
structure also appeared in the original curves for the H9SO, vapor. The curves
in Figures 5 and 6 have been replotted on a percent transmittance scale to elim-
inate the structure due to variations in the background. The physical slitwidth
of the spectrometer was set at 1.2 mm, which produces a spectral slitwidth that
varies from approximately 2 cm-1 at 12 |j,m to 4 cm-1 at 8 |im. The relatively
wide slits were used to smooth out the fringes produced by the 0.8 mm thick
silicon windows. The windows used at the time these data were obtained did not
contain antireflective coatings; thus, the optical interference between the two
surfaces produced fringes that were spaced approximately 2 cm--'- apart.
In order to check for possible spectral structure that was not resolved by the
2 cm" - 4 cm"1 resolution, a few curves were scanned with spectral slitwidths
that varied from approximately 0.5 cm"-'- at 12 |im to 1 cm"1 at 8 p,m. These data
were obtained while using thicker (2 mm), anti-reflection coated windows to
avoid the fringes produced by the thinner windows. Somewhat narrower slits
would have been employed for these data, but the GerCu detector was noisier than
usual at the time so that the detector noise would have been excessive if nar-
rower slits had been used. The spectral curves scanned with the narrow, 0.5 -
1 cm ~L, spectral slits appeared essentially the same as those shown in Figures
5 and 6, which were obtained with wider slits. There was only a small, expected
change in the curves near the Q-branch at approximately 8.2 urn and at a few
other similar places where the slope of the curve changes abruptly. The H2S04
curve in Figure 7 contains the 8.2 |j,m Q-branch and was scanned with a spectral
slitwidth of approximately 1 cm-1. No new structure was observed with the
narrow slits that was not also observed with the wider slits. It is probable
that adjacent lines in much of the spectrum are separated by approximately
0.3 cm" , thus the structure due to individual lines could not be observed with
spectral slitwidths as wide as 0.5 cm" . Known positions of absorption lines
of CH4 and NH3 were used for wavelength calibration of the monochromator.
Scans of the spectra were continued beyond the region represented in Figures 5
and 6 to approximately 7 \im. The region between 7 |j,m and 7.8 p,m has not been
included in the figures because the curves were difficult to interpret quanti-
tatively. The background curve (100% transmittance) sloped steeply downward
from 7 Lj,m to 7.8 u.m because of the decreasing efficiency of the grating toward
shorter wavelengths. In addition, the 7 - 7.8 p,m spectrum shows much structure
due to H20 in the optical path. The extra effort that would have been required
to obtain quantitative data in the 7 - 7.8 U.TO was not considered worthwhile in
the present study because the strong H20 absorption renders this region worthless
for monitoring l^SO^.
18
-------
1200
WAVENUMBER (cm"1)
1000
800
100
VO
w
I
55
o
Figure 5
WAVELENGTH (micrometers)
Spectral curves of transmittance for three samples of ^SO^. Path length
- 52 cm. Vapor temperature - 235°C. The spectral slitwidth varies from
approximately 4 cm-1 at 8 u.m to 2 cm-1 at 12 um. Approximate pressures
of the lSO^ vapor are: A, 3 torr; B, 15 torr; and C, 28 torr.
-------
1200
1100
WAVENUMBER (cm"1)
1000 900
800
100
N>
o
50
w
o
H
H
to
10 11
WAVELENGTH (micrometers)
12
Figure 6.
Spectrum of H2S04 vapor. Path length = 52 cm. Vapor temperature
= 214°C. H2S04 vapor pressure — 7 torr. Spectral slitwidth varies
from approximately 4 cm~l at 8 u-m to 2 cm"* at 12 o,m.
-------
WAVENUMBER (cm" )
1250
1200
100
Figure
WAVELENGTH (micrometers)
7 Spectra of HoSO* and H20 between 7.85 urn and 8.45 urn. The spectral slitwidth is approxi-
mately 1 cm\ The optical path length is 52 cm. The approximate pressures of the gases
are: H?S04, 2 torr; and H20, 600 torr.
-------
The 7 - 7.8 p,m region contains a very strong absorption band of 803 centered
near 7.4 ^m in addition to a portion of the 6.90 (j,m H2SO, band discussed in
Reference 2. Strong absorption by S02 also occurs between 7.2 ^m and 7.5 ^m.
Although this region is difficult to interpret accurately, it does provide
information on the maximun amount of 803 in the samples. For example, sample
C, which is represented in Figure 5, produced approximately 80% absorptance
near 7.3 (j,m. From unpublished S03 data obtained previously in our laboratory,
we estimated that less than 1 torr of S03 would be required to produce this
absorptance. Some of this absorption is undoubtedly due to H2S04, and some may
be due to S02. Therefore, we conclude that the partial pressure of S03 was much
lower than the H2SO^ partial pressure in the samples studied.
The H20 in the approximately 4-meter optical path outside of the sample cell
produced enough absorption in the background curve that the concentration of
H20 in the sample cell could not be determined from the spectra. Little or no
difference in the H20 absorption was observed between the sample spectra and
the corresponding background curves.
Using data from the JANNAF Thermochemical Tables^ on the equilibrium constants
for the gases involved, we estimated the partial pressures of H20, S03, and H2S04
vapors over 98.3% H2SO^ liquid at the experimental temperatures. The calculated
total pressures were typically 2-4 times as high as those measured with the pres-
sure transducer. Although the large discrepancy between the measured and cal-
culated pressures has not been completely resolved, we believe that most of it
is due to erroneously high temperature measurements of the coolest place in the
cell, which determines the vapor pressure. It is possible that the thermocouple
on the liquid H2S04 reservoir was not in good thermal contact with the liquid
and was maintained hotter than the liquid because of heat from reservoir heater.
The top portion of the reservoir may also have been even cooler than the liquid.
The comparison between the calculated and measured pressures suggest that the
temperature indicated by the thermocouple attached to the liquid reservoir may
have been as much as 20°C too high.
We assumed a liquid H-jSO^ temperature that corresponded to a calculated total
vapor pressure equal to that measured with the pressure transducer. The measured
temperature of the main body of the sample cell was then used to calculate the
partial pressures of each of the three gas species for this total pressure. At
the sample cell pressures used, H2SO^ accounted for approximately 50% of the total
pressure. Thus, the H2S04 pressures listed for the samples represented in Figures
5 and 6 are equal to half of the measured total pressure. The calculated pressure
of SO^ was only a few percent of the total pressure, a. result that agrees with the
spectroscopic measurements near 7.3 p,m.
The primary objective has been to obtain spectral data on the absorption by an
amount of H2SO^ vapor that is reasonably well known. Therefore, little effort
has been devoted to investigating the equilibrium conditions. Insufficient time
was available to investigate fully the discrepancy between the measured and cal-
culated pressures. Impurities, including excess H^O, in the sample cell may have
JANNAF Thermochemical Tables, Dow Chemical Co. 1967, plus Addenda to date.
22
-------
contributed slightly to the discrepancy by shifting the S03-H2S04 equilibrium.
The thermal data given in Reference 5 are believed to be reliable.
The build-up and decay of impurities in the sample cell was monitored by observ-
ing the infrared absorption. The impurities appeared in two different forms.
Some of the impurities appeared as a deposit on the windows and on the gold-
coated mirror and remained when the cell was evacuated. Other impurities were
mixed with the l^SO^ vapor and disappeared with the vapor when the cell was
evacuated. The spectra of the impurities varied in shape from day to day, indi-
cating that the same species of gases were not present all of the time. Certain
identifiable spectral features were present on some days but not on others.
After the sample cell had been cleaned, reassembled, and heated to approximately
210 - 220°C, there was little or no spectral evidence of impurities until the
liquid l^SO^ reservoir was heated to produce I^SO. vapor. Most of the impurities
that absorbed significantly were apparently formed by a reaction with the hot
H2SO^ vapor.
On several occasions just after spectra of the I^SO^ vapor had been scanned, the
liquid I^SO/ reservoir was cooled so that the vapor would condense. Spectra
scanned after the vapor had condensed showed little, or no, absorption that
would be attributed to the vapor. Some of the absorption attributed to gaseous
impurities remained after the cooling, but this disappeared when the cell was
evacuated. Part of the absorption attributed to impurities also disappeared
when the reservoir was cooled, indicating that some of the impurities condensed
or that their presence was dependent on the presence of HoSO* vapor.
The absorption feature apparent in Figures 5 and 6 that occurs between approxi-
mately 9.4 |im and 10.1 |j,m is due primarily to a gaseous impurity that was present
in even the cleanest samples. The amount of this absorption varied relative to
the amount of t^SO^. vapor from one sample to another. For example, curves B and
C of Figure 5 both show approximately the same amount of absorption at 9.7 |j,m
although there is much more H2S04 vapor in sample C than in sample B. The shape
of the curve in Figure 6 is also seen to be different from the curves in Figure 5
between 9.4 (j,m and 10.1 p,m. Thus, the impurities producing this absorption were
apparently different on the two different days when the spectra were scanned.
The broken lines in Figures 5 and 6 indicate the estimated spectrum for l^SO^
vapor with no impurities.
The feature with the transmission minimum near 10.4 p,m is probably due to t^SO^
and to some impurity. The absorption in this region appeared to not be com-
pletely correlated with the H^SO^ pressure, but more so than the absorption near
9.7 |j,m. Little special effort was directed toward isolating and identifying the
weak absorption by impurities. The emphasis was on the investigation of strong
absorption in spectral regions that might be useful for monitoring the
concentration.
The hot l^SO^ vapor apparently reacted with the anti-reflective coatings and
produced a substance that absorbs strongly between approximately 8 p,m and 10 ^m.
This absorbing substance could not be wiped or washed from the windows, and the
original anti-reflective coatings appeared to be decomposed. After only a few
minutes of operation with hot l^SO^ vapor in the cell, the absorption between
23
-------
8 (im and LO |j,ra could be seen to increase. This changing absorption made it
impractical to obtain quantitative absorption data for the 8.2 pm band with
the anti-reflection-coated windows in place. The substance deposited on the
windows also absorbed slightly in the region of the 11.4 n,m band, but much less
than in the region of the shorter wavelength band. This substance accounts for
some of the apparerit absorption by I^SO^ illustrated by the curve in Figure 7.
The contribution by this substance is slight for wavelengths shorter than 8.1 p,m,
but it increases for longer wavelengths.
Figure 7 contains a spectrum of hot H2S04 vapor superimposed on a spectrum of
H20 vapor and illustrates the problem of t^O interference if the 8.2 u-tn band
of H2S04 is used. Good sensitivity to H2S04 could be obtained by an instrument
that compares the transmission in two 4 cm~l wide intervals, one centered near
1223 cm-1 and the other centered near 1233 cm-1. The lower wavenumber inter-
val contains strong absorption and is within the Q-branch. The 1233 cm-1
interval contains weaker absorption, yet it is close enough to the other inter-
val that the instrument zero-setting would be relatively insensitive to changes
in source temperature, spectral responsivity of the detector, etc. Unfortunately,
strong absorption by H20 coincides with the Q-branch of H2S04. Therefore, an
instrument that compares these two intervals would also be sensitive to H20 vapor
because the H20 absorption is much stronger in one interval than in the other.
For best results, the H2S04 absorption should be quite different in the two
intervals being compared, and the 1^0 absorption should be weak and approxi-
mately the same in both intervals. If the interval centered near 1233 cm"l
were to be used, the amount of 1^0 interference could be reduced by selecting
the other interval such that the H20 absorption is nearly the same in both
intervals. A 4 cm-1 wide interval centered near 1253 cm-1 might satisfy this
requirement for the interval of weak I^SO^. absorption. The accuracy with which
H20 interference can be accounted for in this manner is limited because the
relative amounts of absorption in the two intervals are usually dependent on
the amount of 1^0 and on the gas temperature. Thus, the intervals may be chosen
so that the H^O absorption is matched for one condition, but it may not be
matched for other temperatures or other amounts of H^O vapor. If an H2SC>4
monitor is to be designed for gases in which the H20 vapor concentration is
quite high, it would probably be best to choose two intervals containing weak
H20 absorption, even if the difference in the absorption by l^SO^ in the two
intervals is somewhat less than in other pairs of intervals that contain strong
H20 absorption.
Of course, the maximum sensitivity can be achieved by one interval where the
H9S04 absorption is the strongest and the other outside of the band where the
absorption is negligible. Such a wide separation of the two intervals generally
increases interference by other gases, by dirt on windows, or by particulate
matter in the sample. The transmittance of particulate matter suspended in the
gas or dirt that accumulates on windows usually changes slowly with changing
wavelength. Therefore, interference by either of these two types of matter is
usually minimized by employing two closely-spaced intervals.
24
-------
A few preliminary data on C02 absorption were obtained in the spectral region
between 10 p,m and 12 urn. Absorption by C02 is negligible near the 8.2 gun band.
The sample cell used for H2S04 was filled to 1 atm with pure C02 while the cell
was heated to 205°C. At the absorption minimum near 11.3 \im, this COj sample
absorbed very slightly, but less than 0.5%. Near 11.0 |im and 11.9 M-rn on
either side of the minimum, the absorption had increased to approximately 2%
and continued to increase toward both longer and shorter wavelengths. These
data on C02 absorption have not been included because the absorption in the
region of primary interest by the largest sample is too small to illustrate the
potential interference. In order to show the nature of the C0£ absorption, we
have included in Figure 8 a spectrum from a report previously published by us.
Superimposed on the C02 spectrum for comparison is a portion of the H2S04 curve A
from Figure 5. The C02 sample represented in Figure 8 was at room temperature and
consisted of 0.30 atm of pure C02 with an optical path length of 1185 meters.
This amount of C02 is much greater than would probably ever be encountered in
the effluent from a stationary source. However, the absorption in this spectral
region by a given amount of C02 increases rapidly with increasing temperature.
As a result, a few atm meters of C02 in a 200°C stack might produce enough ab-
sorption to interfere with an H2S04 measurement that made use of the 11.4 |j,m
band. By measuring the gas temperature and the C02 concentration, the inter-
ference could probably be accounted for.
Other gases that might interfere with an instrument using the 11.4 urn band, de-
pending on the type of pollution source being monitored, are ammonia and ethy-
lene. The absorption by l^SO^ does not change rapidly with wavelength at any
place in the 11.4 p-m band. Therefore, any two narrow intervals within the band
that are compared to measure the I^SO^ concentration must be well-separated and
be subject to the above mentioned disadvantages of instability and interference.
6. Burch, D. E. "Semi-Annual Technical Report Investigation of the Absorption
of Infrared Radiation by Atmospheric Gases." Prepared by Philco-Ford
Corporation, Report No. U-4784, for ARPA, Air Force Cambridge Research
Laboratories under Contract No. F19628-69-C-0263, 31 January 1970.
25
-------
WAVENUMBER (cm"1)
900
850
800
w
50
11
12
13
WAVELENGTH
Figure 8. Spectra of
and C02 between 11 yjn and 12.5 p,m.
26
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-76-191
3. RECIPIENT'S ACCESSION" NO.
4.TITLE ANDSUBTITLE
INFRARED ABSORPTION BY SULFURIC ACID VAPOR
5. REPORT DATE
July 1976
6. PERFORMING ORGANIZATION CODE
7. AUTHORIS)
Darrell E. Burch, Francis J. Gates, and Norman Potter
8. PERFORMING ORGANIZATION REPORT NO,
U-6200
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Aeronutronic Division
Aeronutronic Ford Corporation
Ford Road
Newport Beach, California 92663
10. PROGRAM ELEMENT NO.
1AA010
11. CONTRACT/GRANT
NO.
68-02-1774
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
Environmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Final: R/75-1/76
14. SPONSORING AGENCY CODE
EPA-ORD
15. SUPPLEMENTARY NOTES
16. ABSTRACT
A sample cell has been designed and built to contain hot ^04 vapor for spectro-
scopic analysis. The cell is heated to between 190° and 250°C by an electric
furnace designed specifically for use with the sample cell. A small bulb connected
to the main compartment of the sample cell contains liquid H2S04- The partial
pressure of the H2S04 vapor is controlled by regulating the temperature of the
liquid H2S04, which is at lower temperature than any other part of the cell that
is in contact with the vapor. Transmission curves have been scanned for a variety
of samples over the 7.8 ym - 12.8 ym region with spectral slitwidths between 0.5
cm"1 and 4 cnr"1. At wavelengths of maximum absorption near 8.2 urn and 11.4
the absorption coefficient is approximately 0.0004 (ppm - meters)-!.
ym.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COSATI I ield/Group
*Air pollution
*Sulfuric acid
*Sulfur trioxide
*Sulfur dioxide
*Electromagnetic absorption
*Infrared spectroscopy
13B
07B
20C
14B
13. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNr.l ASSTFTFH
21. NO. OF PAGES
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
27
------- |