EPA-650/3-74-007
THE ROLE OF SOLID-GAS INTERACTIONS
IN AIR POLLUTION
by
S. Siegel, H. S. Judeikis and C. C. Badcock
The Aerospace Corporation
2350 E. El Segundo Boulevard
El Segundo, California 90245
Grant No. 801340
ROAP No. 21 AJX-3
Program Element No. 1A1008
EPA Project Officer: Jack L. Durham
Chemistry and Physics Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
August 1974
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
ABSTRACT
This study was undertaken to evaluate the potential importance of gas-
solid interactions in polluted atmospheres. Model calculations that employed
collision theory, transition state theory, and data from the catalysis literature
were used to determine the conditions under which the heterogeneous processes
could compete with homogeneous gas phase reactions known to be important.
Laboratory experiments were conducted with simulated atmospheres to
determine whether or not the theoretically derived criteria could be met under
ambient conditions. Among the gases studied were NO7, NO, O,, and CO.
^ .j
The selection of the solids used in these studies was based on their abundance
in polluted atmospheres, as well as on their known catalytic activity.
Evaluation of the experimental results leads us to conclude that hetero-
geneous decomposition of NO-, and O,, as well as heterogeneous oxidation of
CO at ground level, can be important atmospheric processes. Results
from experiments conducted in the presence of moisture indicate that water
does not poison catalytic activity. In fact, in the case of NO?, activity
is significantly increased in the presence of moisture.
In contrast to the results for NO-,, O.,, and CO, it was found that hetero-
geneous processes that involve NO are not likely to compete with gas phase
reactions.
This report was submitted in fulfillment of Project No. 21AJX-03, Grant
No. B801340, by The Aerospace Corporation, under the sponsorship of the
Environmental Protection Agency. Work was completed as of October 1973.
111
-------
CONTENTS
Page
Abstract iii
List of Figures v
List of Tables vi
Acknowledgments viii
Sections
I Conclusions 1
II Recommendations 2
III Introduction 3
NO and O 3
X j
Carbon Monoxide 4
IV Design and Equipment Fabrication 6
General 6
CFA Reactor 6
CFF Reactor 14
PMC Reactor 17
Bulk Reactor 22
V Experimental Procedures 24
CFA Reactor 24
CFF Reactor 24
Slug Reactor 25
Bulk Reactor 26
VI Results and Discussion 27
Results and Conclusions for NO 27
x
Results and Conclusions for O_ 48
Results and Conclusions for CO 48
VII References 66
VIII List of Inventions and Publications ' 70
IX Appendixes 72
IV
-------
FIGURES
No. Page
1 Experimental Light Intensities and Filter Transmission
Compared to the Sea Level Solar Spectrum 8
2 Gas Flow System 9
3 Chemilumine scent Monitor for NO 11
4 Optical System (CFA Reactor) 13
5 Absorption Spectrum of NO? and Filter Transmission 15
6 Diagram of CFF Reactor 16
7 Movable Platform used in CFF Reactor 18
8 Slug Reactor System. 20
9 Average NO£ Pressure vs Time for an NO-O2-N2
Mixture in the Absence of Catalysts. The flow rate was
1.4 cm.3/sec and input pressures of NO, O2, and No at
equilibrium were 0.26, 86 and 102 Torr, respectively 29
10 Decomposition of NO2 over A^Og. Reaction mixture:
0.2% NO2 in N2, total pressure was 700 Torr 33
11 Decomposition of NO2 over MnO2- Reaction mixture:
0. 2% NO2 in N2, total pressure was 700 Torr 39
12 Surface Nitrogen Compounds (as NO^) from the
Decomposition of NO2 over A^Os. Reaction mixture:
0. 2% NO2 in N2, total pressure was 700 Torr 43
13 Decomposition of N©2 over Charcoal. Relative concentra-
tions of NO2 (equal to relative fluorescence intensities)
vs distance for a 0. 2% NO2 in a mixture at 10 Torr 44
14 Decomposition of NO2 over MnO2« Relative pressures of
NO2 (equal to relative fluorescence intensities) vs distance
for a 0.2% NO2 in a mixture at 10 Torr 47
15 Effects of Moisture on NO2 Decomposition 46
16 Pressure of CO2 vs Time. 11 Torr CO, 200 mg MnO2 50
-------
FIGURES (Continued)
No. Page
17 Depletive Oxidation of CO 51
18 Oxidation of CO in the Presence of O? 52
19 Slug Reactor Data: Manganese Dioxide 57
20 Slug Reactor Data: Zinc Oxide 58
21 Slug Reactor Data: Cupric Oxide 59
22 Slug Reactor Data: Ferric Oxide 60
vi
-------
TABLES
No. Page
1 Effective Values of k for the Oxidation of NO in
NO-O-N Mixtures 30
L, C-i
2 Physical Adsorption Surface Areas 32
3 Fraction of Gas-Solid Collisions Leading to Reaction 36
4 Initial Optical Densities in NO~-N? Experiments 37
5 Analysis of Nitrogen Compounds on Catalyst Surface 42
6 Results for NO--A Mixtures over Charcoal 45
7 Conditions Used in Metal Oxide Oxidations: Slug Micro-
Reactor 56
8 Slug Reactor Data Summary: Rate Constants and
Concentration of Active Sites 62
9 Estimated Rates of CO Oxidation by Metal Oxides on
Soils 65
vn
-------
ACKNOWLEDGMENTS
The authors gratefully acknowledge the efforts of T. B. Stewart,
H. R. Hedgpeth, and J. K. Allen in the conduct of laboratory work associ-
ated with this program. Appreciation is also extended to M. Birnbaum for
use of laboratory equipment, and M. Masaki for performing some of the
calculations. We also acknowledge a number of helpful discussions -with
Jack T. Durham, Chemistry and Physics Laboratory, National Environ-
mental Research Center, Research Triangle Park, North Carolina.
Vlll
-------
SECTION I
CONCLUSIONS
In the Aerospace investigation of heterogeneous reactions of potential
importance in polluted urban atmospheres, significant reaction rates for gas-
solid interactions of NO-,, O.,, and CO have been found. In the case of NO,
only minimal activities were observed, and we conclude heterogeneous
reactions of this species are not likely to be competitive with homogeneous
gas phase reactions.
For NO7, we find significant reaction rates, especially in the presence
L-t
of moisture, for decomposition on a variety of materials that are likely to be
found in polluted atmospheres. These include metal oxides, salts, charcoal,
cement, fly ash, and sand. Projection of the laboratory results to the
atmosphere, based on the assumption of a total particle loading of 100(j,g/m ,
leads us to conclude that the particle limited lifetime for removal of NO,,
would be ^j 1 hr. Although this lifetime is somewhat long compared with the
photolytic lifetime (^ 2-3 min), the homogeneous photo-process does not
reduce the total oxidant concentration (NO? + O.-,), as does the heterogeneous
process. Thus, NO,-, -solid interactions can serve to significantly reduce
total oxidant concentrations during daylight hours, as well as lead to an
important nighttime sink for NO? .
Similar reactivities were observed in the case of O,, as in the case of
NO-;, , and similar conclusions were reached. However, experiments with
O, were not nearly as extensive as those with NO? .
Experiments with CO concentrated on CO-metal oxide interactions. It
was found that extrapolation of laboratory results to conservative atmospheric
conditions results in significant rates of CO removal at ground level provided
that certain minimum rates of regeneration of the metal oxide active sites by
atmospheric O7 are met. The oxides studied were MnO?, ZnO, CuO, and
Fe?O . Using the initial metal oxide activity and atmospheric loadings of
1 g/m^ (metal oxides only), the lifetime of CO in the atmosphere as a result of
the heterogeneous oxidation by atmospheric particles ranged from approximately
606 yr for MnO_ to 2.6 X 10 yr for Fe?O . Extrapolation of these rates to
LJ LJ J
soils indicate that oxidation by metal compounds on soils could be a significant
sink.
-------
SECTION II
RECOMMENDATIONS
It is recommended that these studies be extended to other pollutants,
especially SO~. In the case of CO and O.,, more extensive studies on a wider
range of solids are required in order to more accurately estimate the mag-
nitude of possible effects in the atmosphere. Studies should be carried out
on solids collected from the environment.
More extensive quantification of the results is needed. This should
include estimation of reactivities of the individual components of aerosols
and summation of these rates along with those estimated for ground level
surfaces. Poisoning and regeneration of the activity of poisoned surfaces
should be taken into account. AJ,eo, atmospheric mixing and transport
phenomena should be considered, along with reactivities to determine whether
or not gaseous pollutant gradients exist near ground level surfaces.
-------
SECTION III
INTRODUCTION
NO AND O,
x 3
Studies of gas-solid interactions of these species were undertaken to
quantitatively determine the extent to which reactions of these air pollutants
can be catalyzed on airborne particle or ground level surfaces. [NO and
O., are lumped together here since many of the important reactions leading
to disappearance of NO, NO?, and O, originate from the chain initiated by
1-3
NO,, photolysis. In addition, similar experimental techniques were used
in the study of these gases. J
It has been shown by model calculations that under certain favorable
conditions, these surface-catalyzed reactions can realistically compete and
in some cases dominate the homogeneous gas phase reactions that arc usually
4
considered to be important in the atmosphere. This model used a simplified
collision model for gas-particle interactions along with transition state theory
and experimental activation energies to estimate absorption, desorption, and
reaction rates. In general, data from the catalyst literature had to be used.
The latter data were usually obtained from experiments conducted at considera-
bly higher temperatures and pressures than are of interest in air pollution
studies. Nonetheless, model calculations based on these data suggested the
potential importance of gas-solid interactions of NO and O, in polluted
.X -J
atmospheres. Consequently, a series of experiments conducted under condi-
tions more closely approaching those of polluted atmospheres was deemed
desirable. This report describes the results of that experimental program
for these gases as well as CO (see below). Experiments were conducted with
NO, NO?, and, to a lesser extent, O,, using solids representative of those
likely to be found in polluted atmospheres. Both dry and moist simulated
atmospheres were used. Experimental results were analyzed using appropri-
ate analytical models. Projection to actual environmental conditions was
4
made using the heterogeneous reaction model described elsewhere. The end
result of applying these procedures leads us to conclude that particle.catalyzed
-------
decomposition of NCL and probably CL arc likely to be important atmospheric
processes. Catalytic oxidation of NO to NO? is probably unimportant.
CARBON MONOXIDE
The observed lifetime of carbon monoxide (CO) in the atmosphere has
been estimated by Wcinstock to be 0, 1 yr. Estimates based on the homo-
geneous gas phase mechanism (Equation 1) yield values of 0.3 to 0.4 yr for
the lifetime of CO. 6' ?
OH + CO*CO2 + H (1)
Therefore, a deficit in the observed rate of loss from the atmosphere exists.
Several other mechanisms have been proposed ranging from sea water
adsorption and bacterial degradation in soils to stratospheric reactions. Almost
without exception the oxidation of CO by airborne metal oxide particles has
been excluded. This is probably due to the lack of studies in the area and to
Q
the conclusion of Nagarjunian and Calvert that metal oxide oxidation of CO
was an unimportant sink. Their studies were based on a single metal oxide
(ZnO) under photo-oxidation conditions.
A review of the literature shows that very few heterogeneous oxidation
studies of CO over metal oxides have been done under conditions that can be
extrapolated to the atmosphere. Most studies have been performed at high
temperatures under high CO loadings. Another problem with most studies is
that they have been done under equilibrium conditions, that is, the rates of
reaction are those determined at high CO concentrations some time after the
8 9
reaction has been initiated. ' First, since ambient CO concentrations sel-
dom exceed 10 ppm even in polluted atmospheres, these conditions may not
apply. In addition, the ratio of CO to oxygen is generally less than 5x10 ,
thus the regeneration of reactive sites could be favored by such concentration
differences in the atmosphere.
For the heterogeneous oxidation of carbon monoxide to be an important
contributor to the CO sink in the atmosphere, the rate of reaction must exceed
certain values. Generally, the equilibrium rates given in the literature are too
Q
slow, however consideration of a simple heterogeneous mechanism can
demonstrate how an initial rate can be much greater. Consider a metal
-------
oxide (M O ) with an initial concentration of active sites of different
x y
types and reactivities A, B, C . . . . Then the oxidation of CO at the various
sites and the regeneration of the sites can be described by the following
reactions.
CO + M O (A) CO., + M O (B) (2)
x y 2 x y
CO + MxO (B) CO2 + M O (C) (3)
O., + M O (C) M O (B) (4)
2 x y x y
O., f M C (B) -~ M O [A) (5)
2 x yv x y1 ' v '
If the concentration of the A type sites is low but their reactivity is high and
the opposite is true for the B type sites, an experimental study of the oxidation
would yield data only for the B type sites at high CO pressures. Furthermore,
the regeneration reactions (which would result in catalytic activity) could be
too slow to form enough A type sites to provide more than a small effect on
the overall rate of production of CO^ as long as the ratio of CO to O? was
large. Kobayashi and Kobayashi have published a series of papers describing
a mechanism for the CO oxidation over manganese dioxide (MnO.-,) which is
M , ! J , 10-12
similar to that outlined above.
Nearly any experimental method chosen to study the initial behavior of a
metal oxide during heterogeneous oxidations runs the risk of integrating out
that data which might be most pertinent to an atmospheric study. First the
minimum detectable limits may require collection of such a relatively large
sample of product that the initial rate is not detected. Second, metal oxides
can return CO,, either irreversibly or reversibly with a slow rate of dcsorp-
9-12
tion,, Tliis again can result in failure to determine the initial rate of CO?
13
production. We have chosen the slug-type micro-reactor with product
collection to study the heterogeneous oxidation of CO over several metal
oxides. This technique results in high sensitivity and can overcome the
problem of slow desorption.
The results from experiments conducted with this system leads us to
conclude that heterogeneous oxidation of CO on metal oxide particles in the
atmosphere is probably not an important process, however, a similar process
on ground level surfaces would be significant.
-------
SECTION IV
DESIGN AND EQUIPMENT FABRICATION
GENERAL
Four different reactor systems were used in these studies. Two of these
systems were cylindrical flow reactors that differed from each other primari-
ly in their method of rcactant/product detection. Optical absorption using a
chopped dual-hcam system was utilized in one reactor, whereas, detection of
laser-induced fluorescence was used in the other. These will be referred to
as the CFA (cylindrical flow-absorption) and CFF (cylindrical flow-fluorescence)
reactors, respectively. The CFA reactor was used in studies on NO ^ and O_,
and the CFF reactor was used to investigate heterogeneous decomposition of
NO?. The third system was a pulsed microcatalytic reactor (designated the
PMC reactor) that was used in the study of catalytic oxidation of CO. The
fourth was a bulk reactor used for preliminary experiments with CO.
CFA REACTOR
Reaction Chamber
The reaction chamber was made from 50-mm-i. d. pyrex tubing with
30-mm-dia pyrex windows fused on the ends that were tapered to accommodate
the pyrex windows. Single 6. 6-mm tubes were fused onto the cylinder walls
at either end to provide for input and exit of the reacting gas mixture. The
ends of the reactor were tapered over a distance of about 30 to 35 mm to
reduce flow related problems. The reaction chamber -was constructed in two
parts with an O-ring joint near the exit to provide access for loading the
catalysts.
Surrounding the reaction chamber were eight F15T8D daylitc fluorescent
lamps. The chamber and lamps were enclosed in a 13-cm-dia chrome plated
cylinder, which acted as a reflector. The lamps were independent from one
another and each couid be operated at the normal current of 0.300 A (15 W) or
in a high power mode at 0. 600 A by switching in either a single ballast or two
ballasts in parallel. The daylite fluorescent lamps have a spectral distribution
-------
nearly equivalent to sea level sunlight (Figure 1) with a maximum intensity of
3.2 ± 0.5 times the average sea level intensity.
With all fluorescent lamps operating at high power, a black catalyst will
o
absorb enough energy to raise its temperature only 5 C. Cooling for the
lamps themselves is provided by a 2800-1/min squirrel cage blower pulling
air through the 13-cm-dia chrome cylinder.
o
In some experiments a 1-mil-thick layer of DuPont Kapton was wrapped
around the outside of the reaction chamber in order to prevent photo-
dissociation of NO.,. The transmission characteristics of this filter are also
illustrated in Figure 1,. At 410 nm (worst case) this filter transmits about
14
0. 15% and the quantum yield for photodissociation is about 0.2. With UK-
use of this filter, gas phase NCL dissociation was not detected.
Catalysts were introduced into the chamber by supporting them on the
surface of either a glass helix or a solid glass cylinder. The helices were
made- from 3-mm-o.d. pyrex rod by -winding 77 turns on a 32-mm-dia
mandrel. Thus the helix was shaped like a cylinder with an inside diameter
of 3 cm and a length of 43 cm. The spacing between the coils was about
3 mm. The helix was used for photolytic experiments. The solid glass
cylinder had an inside diameter of 40 mm and a length of 400 mm. The
cylinder and helix were placed in the chamber such that they were coaxial
with the reaction chamber.
Gas Handling System
Gas flows are illustrated schematically in Figure 2. There are two
Matheson Company model 660 gas mixers for mixing Matheson nitrogen
oxide-nitrogen or argon mixtures with diluent nitrogen or argon and for
mixing this resultant mixture with Matheson ultra-high purity oxygen. At
the ends of the chamber, there are Nupro cross pattern needle valves for
sampling entrance and exit gas streams. The outputs from these valves are
fed to a four-way teflon stopcock such that one stream is directed to the
chemiluminescent monitor for NO (CLEM) and the other stream goes to the
mass spectrometer sampling valve and then to the gas chromatograph
sampling valve. The Brooks R-215-D flow meter at the chamber exit
measures the total flow through the chamber and is the only flow meter which
-------
10V
oo
oo
oo
O
c/T
oo
10'
OO
DAYLITE LAMP
SEA LEVEL SUNLIGHT
KAPTON FILTER
(two layers)
300 350 400
450 500
X (nm)
550 600 650
Figure 1. Experimental light intensities and filter transmission
compared to the sea level solar spectrum
-------
VACUUM PUMP No. 1
THROUGH GAS
CHROMATOGRAPH
SAMPLING VALVE
TO VACUUM
PUMP No. 1
MASS-SPECTROMETER
INLET
NO-N2 MIXTURE
MATHESON No. 3500
REGULATOR
NITROGEN OR ARGON
ZERO GAS 25-50 kg m
MATHESON No. 19
REGULATORS
E
GUAGE
BROOKS
R-215-D
ALIBRATED AT
760 & 140 Torr
ULTRA-HIGH
PURITY
OXYGEN ,
25-50 kg m
| -- NEEDLE VALVE
UNLESS NOTED OTHERWISE
M
WATER-4 1!
HUMIDIFIER
MATHESON No. 665
GAS MIXERS
ON-OFF
TOGGLE VALVE
Figure 2. Gas flow.system
-------
has been calibrated against a wet-test meter.
Humidifier
For experiments conducted as a function of relative humidity, a separate
gas stream must be humidified and mixed with the NO - ]XL or argon mixture
X. i-t
prior to entering the catalytic chamber. The fact that NO is soluble in and
X
somewhat reactive with H-, O prevents the direct humidification of the gas
mixture. Humidification was accomplished by bubbling an inert gas such as
argon or nitrogen through a gas washing bottle and mixing with the NO7. The
LJ
humidity of the NO gas mixture can be regulated by controlling the water
temperature in the humidifier and the amount of humidified gas that is mixed
with the NO gas mixture. The moisture content of the reacting mixture was
measured by the temperature difference between a dry and wet bulb copper-
constantan thermocouple combination.
Gas Analyzing System
Chemiluminescent Monitor (CLEM)- The chemiluminescent monitor for NO illus-
trated in Figure 3 utilizes the chemilumLnescent reaction between ozone and nitric
oxide to monitor either nitric oxide or ozone concentrations. For these
experiments, however, only the nitric oxide concentrations were monitored.
The ozone used for the reactions was made from oxygen at 200 Torr in a
silent discharge. The oxygen flows between concentric thin wall (1 mm)
pyrcx cylinders spaced 1 mm apart. Copper sulfate solution is used as
electrodes, and a potential of 4000 V ac is applied between the inside and
outside cylinders. At the given operating conditions, <^ 4 Torr of ozone is
generated or about 2% conversion.
The analysis takes place when the ozone-enriched oxygen is mixed with a
sample of entrance or exit gas from the main reaction chamber. The mixing
takes place inside a 1-liter spherical pyrcx flask that has been painted white
on the outside. A United Technology PIN-25 photodiode with a red filter is
mounted near the top of the flask to monitor the chemiluminescent reaction.
Calibration of the CL/EM is accomplished by supplying either 1, 000 or
10, 000 ppm (± 2%) nitric oxide diluted with nitrogen in place of the chamber
gas. Because of the dc drift in the photodiode-electrometer system, the
10
-------
4000 VAC FROM
NEON SIGN TXFMR
CuSO.
FLOWMETER
SAMPLE
INLET
FROM 3-WAY
STOPCOCK -
Hg MANOMETER
0-150 Torr
MOL SIEVE &
STEEL WOOL TRAP
Hg MANOMETER
0 - 200 Torr
FLOWMETER
VACUUM
PUMP No. 2
Figure 3. Chemiluminescent monitor for NO
11
-------
minimum sensitivity is about 50 ppm NO; the response is linear to at least
10, 000 ppm NO. A 2, 000 ppm nitrogen dioxide-nitrogen mixture gives no
detectable signal.
Gas Chromatograph and Mass Spectrometer - The gas chromatograph is an
F and M (Hewlett-Packard) model 810 equipped with a Carle 2114 sampling
valve, thermal conductivity detectors, and a stable dc amplifier. The
column used for nitrogen-oxygen ratio determinations is a 6.6-mm-dia
aluminum column of Lindc molecular sieve 5A, 60 to 80 mesh (with fine
particles removed), 2-m in length, and is activated periodically at about
o
300 C. For low noise, the bridge current was supplied by two 6-V dry-cell
batteries connected in series.
The quadrupole mass spectrometer is an E. A. I. Quad 200 pumped by
a 100 I/sec ion pump. The gas can be analysed by inletting the gas through
a leak valve modified to sample a flowing gas stream with little stagnation.
This instrument was not used on oxides of nitrogen because the nitrogen
dioxide dissociates. It was included in the system for use on other gases.
Optical System - The average NO,., concentration in the reaction chamber is
determined by photometric absorption. The dual-beam system is illustrated
in Figure 4. The light source is a fan-cooled 450-W tungsten filament sun
gun lamp with quartz envelope. A 5-cm water filter with a 475-nm narrow
band interference filter and Corning 3-72 yellow filter is placed in the
optical path.
A fraction of the light is passed through a pinhole and lens to a 90 deg
mirrored face, chopper blade that rotates at 11.3 Hz on an inclined axis of
45 deg to the incident light beam. At one phase of the chopper, the light
passes through the open sections of the chopper blade and through another
lens into the reaction chamber, through the chamber and is reflected off an
(adjustable) first surface mirror into the integrating sphere. At the other
chopper phase, the light is made to by-pass the chamber by reflecting off
one of the two front surface mirrors on the chopper blade through a series
of lenses and mirrors into the integrating sphere. The two beams are com-
bined orthogonally at the integrating sphere, which is coupled (orthogonally)
through another 475-nm narrow band interference filter to a 1 P2 1 phoito-
12"
-------
in
O
-P
O
rt
>
en
i i
nJ
o
'43
OH
O
bo
H
13
-------
multiplier tube operated at -1000 V.
The photomultiplier tube output is fed to a Brower model 129 thermo-
couple amplifier (lock-in voltmeter). The baseline drift of the system is
about one part per thousand short term (10 min) and about five parts per
thousand overnight. An intensity change of one part per thousand (O. D.
0.0004) corresponds to about 1.5-m Torr NO? in the chamber.
16
The absorption spectrum of NO? is quite structured. This fact,
coupled with a filter bandpass of about 10-nm, required that an average
extinction coefficient (e) be determined. This average was determined by
use of a Gary-15 uv/visible spectrophotometer and a 2. 5-cm gas celL The
absorption spectrum of NO? and the filter are illustrated in Figure 5. The
LJ
filter and empty gas cell were measured in the spectrophotometer and the
spectrum was integrated to give I . Similarly the filter and gas cell with a
low pressure of NO,, was measured and integrated to give I. The values of
LJ
1 , I, path length, and NO? pressure (corrected for N7OA content) were
O <- o ~t
substituted into the Beer-Lambert law; the average value of e calculated was
98 1-mol -cm
CFF REACTOR
The fluorescence apparatus used for the experiment is shown in Figure 6.
This system, which was constructed under the direction of Dr. M. Birnbaum,
was initially used to measure, on a real-time basis, the atmospheric con-
centration of the oxides of nitrogen. The apparatus is a vacuum chamber
_4
capable of obtaining pressures of 10 Torr. The laser light from a Spectra
Physics model 140 argon ion laser enters the vacuum chamber through a 4880
interface fiber, iris diaphragm, and a series of light baffles. The light then
enters the observation chamber where the NO? fluorescence excited by the
laser beam is transmitted, at right angles to the exciting light, through a
series of optical filters and focused by a 75-mm focal length lens on an
EMI 9659 QAM photomultiplier tube. The signal is monitored on a pulse
counter and displayed numerically. The system was slightly modified to
accommodate our flow reactor experiments. The air sampling inlet was
replaced by a gas bottle containing 2000 ppm NO., in argon and a Matheson 603
14
-------
900
600
o
300
MEASURED
HERE
FILTER
TRANSMISSION
60
40
GO
on
20
450
460
470
480
X |nm)
490
500
Figure 5. Absorption spectrum of NO2 and filter transmission
15
-------
PULSE -
COUNTER
t
N02/A INLET
.PHOTOTUBE
REFRIGERATED HOUSING AND
PHOTOMULTIPLIER TUBE
0.5-mm SLIT ON
FACE OF PHOTOTUBE
75-mm FL LENS
CoSO4 AND CS2-63 FILTERS (red pass)
2-m FL LENS
OUTPUT
PRESSURE
TRANSDUCER 'CYLINDER
WITH SECOND
HALF COATED
\RIS
DIAPHRAM
ARGON IN
LASER
(4880 A)
NARROW BAND
FILTER
TOP VIEW
Figure 6. Diagram of CFF reactor
16
-------
flow meter which was calibrated over a pressure range of 1 to 100 Torr. A
Dynasciencc model P70 pressure transducer was added to monitor the total
pressure with a sensitivity of lO^m. Light baffles were removed so that the
chamber could accommodate the cylinder. Baffles were placed on the
cylinder ends to reduce scattered light. The sensitivity for detecting NO,,
-2
was 10 Torr.
The spatial resolution of 0. 5 mm was achieved by using the 75-mm lens
and 0. 5-mm slit in combination. The slit was placed on the face of the photo-
multiplier tube which was located at an image distance of 150 mm from the
lens. This image distance was exactly the same as the object distance of the
laser beam from the lens. Thus, by placing the slit on the photomultiplier
and using the image and object distance relationship with a thin lens, one gets
a magnification of one, thus a spatial distance on the beam of 0. 5 mm.
For most of the catalytic measurements, a 1-cm radius pyrex cylinder
was placed in the chamber on a movable platform in such a way that the
cylinder axis coincided with the laser beam. The movable platform, shown
in Figure 7, was constructed from a precision dovetail slide to which was
attached a 32 pitch rack. The rack was used in combination with a 32 pitch
spur gear mounted on a rotating shaft. This gear combination gave no back-
lash motion when the platform was moved back and forth. Linear motion was
imparted to the slide from outside the fluorescence chamber by rotating the
shaft that passed through an ultra-Torr vacuum fitting. The vacuum fitting
was attached to the vacuum flange and a leak-tight seal was made by pressing
_4
a viton O-ring against the shaft. No leaks -were detected at 10 Torr
pressure when the shaft was rotated. A pointer affixed to the shaft indicates
the position of slide by virtue of it pointing to a scale in the flange. The
position of the slide could be measured to 0.25 mm. By using larger pitch
gears and larger radius scales, better accuracy in position could be measured.
The total linear displacement of the platform was 45 mm.
PMC REACTOR
General Description
The experimental design of the system is a modified version of the micro-
17
-------
POINTER
SCALE
DOVETAIL SLIDE
HANDLE
POINTER
MOUNT
ULTRA TORR
FITTING
VACUUM
FLANGE
-SHAFT
RACK
SPUR GEAR
BASE
(aluminum)
Figure 7. Movable platform used in CFF reactor
18
-------
13 18
catalytic slug reactor originally described by Hall and Emmett. ' It
consists of a vacuum line for measuring and mixing gases, a gas sampling
valve that is used to introduce slugs of the gas mixture into a reactor via
the helium carrier gas, a CO collection loop, and a gas chromatograph for
qualitative and quantitative analysis of the reactants and products. A flow
diagram of the experimental apparatus is shown in Figure 8.
Gas Mixing System
The gas mixing system is a glass high vacuum line using Kontes teflon
O-ring stopcocks, a hand operated piston type pump for mixing gases, two
precalibrated bulbs of different volumes for calibration and measuring, a
Matheson 0 - 760 Torr vacuum gauge, and a Validync pressure transducer
with exciter-demodulator and VOM for pressure measurements in the low
pressure region. The pressure transducer, vacuum gauges, and parts of
the gas chromatograph are connected with metal Swagelok connectors. The
connectors between the gas chromatograph and vacuum system are via
flexible metal to glass seals.
Slug Reactor and Gas Chromatograph
The1 reactor and chromatograph consist of a series of gas sampling
valves and calibrated loops that are followed by two chromatographic
columns and a thermal conductivity detector. A small sample of the gas
mixture from the gas mixing system is drawn into a previously evacuated
sample loop (0. 5 cc) in a Carle, eight port, two-position gas sampling valve
(GSV)» The Carle GSV was fitted with plates top and bottom and purged with
helium to prevent air leakage. When this sample is switched into the helium
carrier stream, it passes through the reference side of a Carle micro volume
thermistor thermal conductivity detector and then into a Liocnco six port GSV.
The Loenco GSV is arranged such that the sample may by-pass or be admitted
to the reactor loop containing the suspended metal oxide. This valve also
allows connection of the reactor loop to the vacuum system to allow for
activation (when in the by-pass position). The gas flow then enters a Whitey
four-way ball valve where the flow may either be directed through a stainless
steel loop or by-pass it. The loop may be cooled with liquid N? for collection
19
-------
TO HIGH VACUUM
SAMPLE
LOOP
GAS TO
REACTOR
TO HIGH VACUUM
AND PRESSURE
MEASURE
SAMPLE
LOOP
GAS
CIRCULATING
PUMP
REACTOR
LOOP
4-WAY
BALL VALVE
TO DETECTOR
BRIDGE CONTROL
AND RECORDER
t
CARLE
MICRODETECTOR
GAS
MIXING
SYSTEM
GAS INLET
C02 COLLECTION
LOOP
TO FLOW
MONITOR
5A MOLE-
CULAR SIEVES
CW)
F AND M
4-WAY
VALVE
PARAPAK Q
GAS CHROMATOGRAPHY OVEN
Figure 8. Slug reactor system
ZO
-------
of the product CO?' The gas stream then enters the gas chromatographic
section.
The gas chromatograph is a modified Loenco blood gas analyzer
(AD 2000) oven coupled with the detector described above. The chromato-
graphic columns are a 1.8-m by 0. 32-cm (O. D. ) Porapak Q column that is
connected to an F and M four-way valve that switches in or by-passes a
7. 6-cm by 0.64-cm (O. D. ) molecular sieve 5A column. The- effluent then
enters the- sample side of the detector and exits through a ball-type flow
meter. The Porapak Q column resolves CO and O7 from CO the molecular
LJ LJ
sieve column resolves the former two components and N.,. The detector is
powered by a Carle 100 micro detector control, and the data is displayed on
a Leeds and Northrup strip chart recorder running at 2 in min
Materials
Manganese dioxide and cupric oxide were obtained 99% pure from
ROC/RIC Research Chemical. Ferric oxide used was commerical jewelers
rouge, and zinc oxide was obtained from the New Jersey Zinc Company. All
were used as received. Surface areas by the BET method ' for the metal
oxides were MnO? - 97 m /g, ZnO - 22.2 m /g, Fe? O, - 27.3 m /g, and
CuO - ( 5 ) m /g. The support material for the metal oxide in the catalyst
loops was 90/600 mesh Tec-Six, a Teflon Six powder from Aiialabs. Helium
O
from Air Products was passed through a molecular sieve 5A trap at -196 C
(L-N,,) to eliminate CO,-, and water from the carrier gas. Helium to purge
the Carle CSV was used as supplied. Matheson C. P. carbon monoxide and
O
extra dry oxygen were passed through a -78 C silica gel trap to remove CO,,
and H,>O. Carbon dioxide used in peak-area calibrations -was obtained from
O O
dry ice and pumped on at -78 and -196 C to remove air.
Reactor Loops
The reactor loops were U-shaped, 21 - 24-cm long pyrex glass tubes,
6-cm O. D. , and were attached to the Loenco GS V by Swagelok nuts and
Viton O-rings. The metal oxide and Tec-Six were weighed out to contain a
certain amount of the metal oxide when put into a glass loop. They were
shaken together to distribute the metal oxide throughout the Tec-Six.. The
21
-------
mixture was packed into the glass loop by using vacuum and gentle tapping.
Glass wool plugs at each end prevented loss of the packing during switching
of the Locnco valve and when the loop was activated. The metal oxide was
Q
activated by pumping out the loop while it was heated to 130 C, and then by
pressurizing with 1 atm CX,. The temperature and the' CX atmosphere were
maintained for 10 hr. The excess CX, was pumped off, and the loop switched
into the helium carrier gas flow. A 10 min blank for CO.-, was run before
the experiments began.
The volume of the reactor was measured and the free volume determined
by subtracting the volume occupied by the Tee-Six and the metal oxide.
Gas Mixtures
For each run, the- pressure of CO was the same 0.680 Torr. This
pressure was reached by first measuring a high pressure into a small volume,
pressurising the gas mixing system with 755 Torr He, and mixing thoroughly
with the hand operated circulating pump. When oxygen was used, the same
procedure was followed except that 157 Torr O,, was added to the system
before1 the helium was used. Since the total number of slugs reduced the
pressure only slightly, a total pressure of 750 Torr was used in subsequent
calculations. In some preliminary experiments, water was also added to the
gas mixture. Generally, 100% relative humidity was used. This was
accomplished by extended recirculation over liquid water in a U-loop section
of the system.
HULK REACTOR
Soine preliminary experiments were performed using a bulk reactor
with MnO-,. A reactor was constructed of quartz and consisted of a chamber
3 cm by 7 cm with an O-ring joint in the center of the reactor. Tubes from
each of the longitudinal ends of the chamber are connected to a gas sampling
valve (GSV) and a hand operated circulating pump. The total volume of the
system was 246 cc, and the pump displaced -- 80 cc so that the total volume
could be circulated in three strokes of the pump. Samples of metal oxide
(MnO? in this case) on a fritted glass disk were held in the center of the main
22
-------
chamber by a holder in the O-riag joint. Analysis of samples extracted by
the CSV were analyzed gas chromatographically on a Porapak Q column for
C02°
The reactor was enclosed in a furnace to allow for activation of the;
MnCL. Activation was achieved by adding 100 Torr O to a de-gassed sample
of MnCX and holding the temperature constant for varying periods of time.
o o
In different experiments the temperature was varied from 23 to 110 C.
23
-------
SECTION V
EXPERIMENTAL PROCEDURES
CFA REACTOR
The helix or cylinder, whichever used, was coaled by applying an aqueous
suspension of the catalytic material to the supports. The coated supports were
placed in the chamber in such a way that the leading edge of the helix or
cylinder was about 9 cm from the gas inlet. The chamber was closed and
_4
pumped to 1 0 for about 4 hr. This ensured complete drying of the sample
before experiments took place.
For NO? experiments, the conditions were varied between pressures at
L-,
700 to 150 Torr with a flow range of 2 to 20 cc/scc. For most of the moisture
experiments, 300 mm of argon saturated with H-^O was mixed with 400-mm
NO-,-argon mixture. This gave an operational relative humidity at 44%.
NO-O? experiments were performed at 190 Torr with an O? pressure
range between 50 - 150 Torr with the remainder NO-N? mixture. The flow
was varied between 2-20 cc/sec. The photo experiments were performed
during the NO-O? experiments with lights turned on after the reaction reached
a steady state.
CFF REACTOR
A typical experiment began by coating one-half the length of a 20-cm-long
2-cm-dia glass cylinder. One-half of the cylinder was coated because we
wanted the gas flow patterns to be: established in the cylinder before the gas
encountered the catalyst. This procedure eliminated some possible errors
in measurements taken near the cylinder edge where gas flow perturbations
were taking place. To allow the NO? fluorescence to reach the photomultiplier
lube:, the cylinder was coated such that a narrow strip (6-mm wide) of glass
on the cylinder wall was not coated. This window was on the same plane as
the laser beam and the optical detection system.
Once the cylinder was coated, it was placed on the movable platform such
that the leading edge of the catalyst was displaced 1 cm from the sightline of
24
-------
the lens-slit system. This was done so that a reference fluorescence
intensity could be measured before the NO7 encountered the catalytic material.
-4
The chamber was closed and pumped to 10 Torr for about 3 hr. Experiments
were performed with a 2000 ppm NO? in argon gas mixture at a pressure range
between 1- and 100-mm and with a flow range of 15 cc/sec to 120 cc/sec. The
fluorescence intensity at each point on the cylinder was measured as the'
point was moved past the lens-slit line of sight. Thus, the fluorescence
intensity from the NO should decrease as more catalytic material is moved
upstream from the detection sightline. Preliminary experiments were per-
formed with measurements at every 1-mm interval. However, it was found
that the necessary information could be obtained by taking fluorescence read-
ings at 5-mm intervals with a 5-sec counting time. It took approximately
1 min to measure the fluorescence intensity over the 45-mm length. The
short measuring time was used to reduce the effect of erroneous reading
caused by catalytic poisoning.
SLUG REACTOR
For an experimental run with a metal oxide the flow rate was adjusted to
10 cc/min with a needle valve in the helium supply, a GSV loop loaded with
Q
the gas mixture (1.81 x 10 moles CO/ slug), the Loenco valve switched so
that the catalyst loop was in the carrier gas, and, then, the CSV switched
to send the slug of CO or CO/O? through the catalyst bed. Collection was
LJ
timed for 10 min after the Carle GSV was switched. If more than one slug
was used per collection of product, then collection time was started after the
last slug, although the L-N? was around the collection loop from the first
injection. At the end of 10 min, the collection loop was closed off (also the
catalyst switched out) and the CO9 loop heated up to room temperature or
LJ
above. The carrier flow rate was retimed to 29 cc/min and the CO,, was
"injected" into the carrier gas and analyzed. All experiments were performed
o
at room temperature (^ 298 K).
Carbon dioxide calibrations were determined at 29 cc/min, where CO,
had a retention time of about 1 min (from the CO? collection loop) on the
L*
o
Porapak Q column at 50 C. The peak area (half width times height) was
25
-------
found to be linear with the amount of CO7 for the range used in the experiments,
L*
BULK REACTOR
For an experiment, 11 Torr CO was admitted to the reactor, the gas
volume circulated once (3 pump strokes), and a sample analyzed. Circulation
was continued and samples extracted at varying times. Experiments were
also performed with 5 Torr of O, added with the CO in the same manner. All
LJ
o
experiments were performed at room temperature (~ 298 K).
26
-------
SECTION VI
RESULTS AND DISCUSSION
RESULTS AND CONCLUSIONS FOR NO
x
General
A number of experiments were carried out in the CFA reactor with solids
representative of those materials likely to be found in polluted environments.
Data from atmospheric particle loading measurements as well as the
catalyst literature were used to aid in the selection of these materials. In
addition to low concentrations ol NO r reaction mixtures included O7, N7 or
2\. b LJ
A, and H?O in various combinations. These experiments helped to establish
which of the NO -solid reactions might be important in polluted atmospheres.
Subsequently, more quantitative experiments were conducted using selected
NO -solid pairs in the CFF reactor.
These experiments led to the conclusion that NO7 could react with a
number of solids likely to be in the environment with rates that could be
important in comparison to gas phase reactions. Reaction products included
surface nitrates and nitrites as well as gaseous NO.
In the case of NO itself, the experimental results led us to conclude that
adsorption or heterogeneous reaction (oxidation or decomposition) does not
occur with rates likely to have atmospheric significance.
Moisture was found to have a significant effect on the NO7-solid reactions.
In many cases, materials that were unreactive in "dry" simulated atmospheres
were very effective in decomposing NO7 in humid atmospheres. The results
from the "dry" experiments will be discussed first, to be followed by a
discussion of the effects of moisture.
Preliminary Experiments (CFA Reactor)
Homogeneous Reactions - When a mixture of NO-O7-N7 is passed through the
CFA reactor in the absence of any catalytically active solid, NO7, is formed
by a termolecular homogeneous gas phase reaction between NO and O7. The
NO7 concentration increases with time (distance) as the reaction mixture
flows through the chamber. Experimentally (vide ante), the total optical
27
-------
absorption by NCX, in the chamber is measured. This can be related to the
16
average NCX> pressure by using measured absorption coefficients. Results
from a typical experiment carried out in the absence of any solids (other than
the glass walls of the chamber) arc shown in Figure 9. During the first
30 min of the experiment, gas flows were adjusted to the desired values and
the system allowed to reach a steady state. From the average NCL pressure,
and the input gas pressures and flow, we can calculate the rate constant (k~)
for the gas phase reaction by using Equation (20), which is derived in
Appendix A.
At t = 50 min, the fluorescent lights were illuminated at half power and
the average NO,-, pressure decreased because of its photodecomposition by
ultraviolet light passing through the pyrex walls of the reactor. Similar effects
were noted when the fluorescent lamps were turned on at full power at t = 63
min. In both illuminated conditions, the steady-state average NO7 pressures
h
could be used to calculate an effective rate constant for NCL production using
Equation (20) (e.g., assuming no photodecomposition was occurring). At
t = 70 min, the lamps were extinguished and the average NO-, pressure
approached the previous dark steady-state value. At t = 74 min, gas flows
were terminated and the system allowed to pump out-
The values for kp obtained in the absence of any solids are given in the
top of Table 1. Generally, good agreement was obtained between the values
of k~ for the dark reaction and values reported in the literature, although
the agreement was slightly dependent on flow conditions. Values of k,-, for
3
the dark reaction obtained at low flow ('- 3-cm /sec) were in better agreement
with the literature values than the rate constants calculated from the high
flow experiments. The discrepancy at the higher flow rates is not serious
and is considered to be the result of minor mixing or flow problems or both.
[In earlier reactor designs, discrepancies of a factor of two or more were
not uncommon. Slight modifications of the mixing system and reactor design
to avoid areas of retarded gas flow (where more reaction would take place
leading to larger effective k 's) greatly reduced this problem, j
The results from experiments in which the glass helix, coated with
selected solid materials, was inserted into the reaction chamber are also
28
-------
2NO +
2NO,
0. 10
0.08
0.06
cT'
QC
LU
«X
0.04
0.02
30
40
1 1/2 "SUNS
NO CATALYST
NO-02-N2
TERMINATE
FLOW
50
60
70
80
TIME, min
Figure 9. Average NC>2 pressure vs time for an NO-O2-N2 mixture
in the absence of catalysts. The flow rate was 1.4 cm3/sec
and the input pressures of NO, G>2, and N2 at equilibrium were
0.26, 86, and 102 Torr, respectively.
29
-------
Table 1. EFFECTIVE VALUES OF
FOR THE OXIDATION
OF NO IN NO-O2-N2 MIXTURES
Catalyst
None.
None-
Average
Charcoal
A12°3
Mn02f
Fc2°3
Cu2O
ZnO
V2°5
(NH4)2S04
PbCl2
Sand
Low Flow
F,
cm /sec
1.4C
2.8
1.9
1.2
1.8
2.6
2.2
2.2
2. 0
, -3 ,2 , -2 -1
k~xlO »1 mole sec
Darka
7. 7
8.3
8.0
-0-5. 7d
6.6
7.8
11.2
9.9
14. 0
8.7
1-1/2 Sun sb
3.0
4.7
3.8
3. 5
3 Suns
2.2
2.2
2.6-10.0e
3.0
7.2
2.9
3.7
2.9
High Flow
F,
cm /sec
8.8°
9. 8C
12. 9C
14. 8
9.4
9.6
9.7
10. 0
10. 0
9.3
9.4
9.6
i in-3 i2 i -2 -1
kpxlO ,1 mole sec
Darka
10.9
12.9
10. 0
9.4
10.8
-0-8. Od
-0-9. 7 d
-0-8. Od
8. 0
13.4
11. 0
10. 6
10. 8
1-1/2 Sunsb
9.9
8. 1
7.7
8.6
5.8
9.6
3 Suns
7.9
6.9
7.4
4.4-9.6e
6.7-13.3e
aO-15.5e
4. 5
6. 1
7.3
6.8
6.8
Compare to literature value of 7. 5 ± 0. 75 x 10 1 mole sec" (Reference 3)
Approximate equivalent solar intensity with all lamps on at half and full power,
respectively
,Here only, no helix
d o
' See text for discussion
Here only, Kapton filter
30
-------
summarized in Table 1. The materials used were selected on the basis of
their known photo- or thermal-catalytic activity for various related reactions
(especially oxidation) and/or their likelihood of occurrance in polluted
atmospheres (cither as airborne particles or ground level exposed surfaces).
The various materials used, along with their measured BET surface areas
are given in Table 2.
The results for charcoal, Al^O,,, and MnCX were particularly interesting.
Initial yields of NO? were quite low (actually zero to within experimental error
when corrections were made for reaction taking place between the gas inlet and
the leading edge of the coated cylinder), compared to the NCL produced in the
absence of any solid. These results were indicative of surface-catalyzed
decomposition of NCL, (Alternatively, NO could be adsorbed or react at the
surface to form products incapable of yielding gas phase NO.-,. This possibility
was eliminated by the fact that quantitative measurements of NO in the exit
gas stream indicated concentrations equal to those in the input gas mixture,
except for the small fraction reacting to form NO?. When O? was absent in
the' reaction mixture, the NO concentrations in the input and exit streams
were equal to within experimental error. )
In view of these results, several experiments were conducted with
NO^-N^ mixtures. A representative example of data from such an experiment
with A17O, is shown in Figure 10. The initial optical density (O. D.) at 480-nm
*-« j
due to absorption by NO? for the example illustrated in the figure (as was
found in the cases of charcoal and MnO,,) was quite low. To within experimental
error, most of this absorption could be accounted for by NO? filling the area
between the gas inlet (or exit) tube and the leading (or trailing) edge of the
coated cylinder. The remaining optical absorption is probably due to the
small concentration of NOp present along the cylinder axis and absorbing
light before it diffuses to the wall where it is removed by adsorption and/or
surface reaction.
After several hours, the O. D. begins to rise slowly. This appears to be
due to a diminuation or poisoning of the solid reactivity, which permits the
NO., mixture to penetrate further into the chamber before reaching fresh
surface areas where NO? can be removed. After many hours (~ 7 - 8 for the
31
-------
Table 2. PHYSICAL ADSORPTION SURFACE AREAS
Material
Source
Preparation
Area, M /g
A1Z°3
Wood Charcoal
Powder
Cement
Fc2°3
PbO
Sand
Fly Ash
Cu 0
L-i
ZnO
V2°5
Activated
alumina 214-77
J. T. Baker
#0537
Allied Chemical
#1567
Commercial
jewelers rouge
J. T. Baker
#2338 yellow powder
ROC/RIC MN-37
General Chemical
ROC/RIC CU-33
red 99%
ROC/RIC V-15
ROC/RIC MN-37
(CO2/-78°C)
Ground up
Ground up
Ground to fine
powder
196
233 ±7
40. 7
17.9
27.3
20. 1
87. 5
98. 5 ± 2.9'
107
7.3
15.2
5, 5 ±3. Oa
22.2
14. 1
85.9
119
Errors derived from estimates of error in reading pressure devices, i.e.,
Matheson 0-760 and voltmeter used with pressure transducer plus a 0. 01 V/min
drift noted.
32
-------
z
Q.
liJ
ce.
0.20
0. 16
0. 12
0.08
0.04
N02-N2
NO CATALYST
10 100
200 300 400
TIME, min
AI2°3
DARK
3 "SUNS"
500
600
Figure 10. Decomposition of NC>2 over A^C^. Reaction mixture:
0.2% NO2 in N , total pressure was 700 Torr
33
-------
the experiment in Figure 10 ), the solid apparently loses all reactivity toward
NO2 and the reaction mixture rapidly fills the area between the end of the
helix and the gas outlet. (The magnitude of the final rapid rise in optical
density increases as the end of the helix is moved away from the gas outlet
and vice versa. ) The O. D. obtained at this point in the experiment is
approximately equal to that observed in the absence of catalyst, for the same
experimental conditions.
The results illustrated in Figure 10 also indicate some interesting photo-
effects. Exposure to light (near uv and visible) at ~ 525 min leads initially
to a large increase in O. D. , probably the result of photodcsorption of NCL,
L-i
(about 30% of the increase is attributable to scattered light from the
fluorescent tubes reaching the photomultiplier). Termination of the light
exposure indicates that the O. D. (i. e. , NCX, concentration) has dropped to a
level below the value observed at the time the photolysis was initiated. This
could be due to photolytic depletion of NCX-, in the gas phase while the lights
LJ
were on, followed by a gradual return to the "dark" steady-state, and/or
photolytic restoration of some catalytic activity. The experiment with MnCL
LJ
(Table 1) suggests the latter expJanation as the more probable since the
Kapton filter used in that experiment prevents the photolytic dissociation of
NCX, in the gas phase. (Attempts to restore the catalytic activity of Al? O
by subjecting the system to overnight vacuum were unsuccessful. ) Exposure
to light at ~ 250 min (Figure 10) leads to dramatic effects. Photodesorption
is probably occurring here as well; however, the desorbed NCX-, quickly
encounters fresh catalyst as it flows downstream, and it again adsorbs or
reacts.
Several additional materials listed in Table 1 gave effective values of k,-,
smaller than (Fc^CX,) or greater than (ZnO, V^CX) those observed in the
absence of any added solids when NO-O^-N-, mixtures were used. Still others
gave effective rate constants equal to those obtained in the absence of added
solids. In none of these cases, however, were the effects as dramatic as
those obtained for charcoal, Al?O,, and MnCX-,.
Correlation with the Atmosphere - In order to assess what these data from the
CFA reactor experiments mean in terms of potential reactions in urban
34
-------
atmospheres, we have derived several additional models of our experimental
system in Appendix A to include the homogeneous gas phase reaction and a
heterogeneous process (either NO? decomposition or NO oxidation to NO,,).
^ 2
Using these models, we can then determine the fraction of NO?-solid (cp ) or
L, C
NO-solid (cp' ) collisions leading to reaction. The results of these calculations
are given in Table 3.
In a number of cases, values of
-------
Table 3. FRACTION OF GAS-SOLID COLLISIONS
LEADING TO REACTION3-
Catalyst
Charcoal
A12°3
Mn02
FC2°3
Cu20
ZnO
V2°5
(NH4)2SO^
PbCl2
Sand
Low Flow
F,
cm /sec
1.9
1.2
1,8
2.6
2.2
2.2
2.0
7
cp (or cp' ) x 1 0
Dark
>2000b
0.5
0.2
(0.4)
(0.2)
(0.6)
(0.1)
1-1/2 Suns
0. 3
3 Suns
(0.1) -(1. if
(0.1)
(0.7)
(0. 1)
(0.2)
(0. 1)
High Flow
F,
cm /sec
9.4
9.6
9.7
10.0
10. 0
9.3
9.4
9.6
cp ( o r cp ', ) x 1 0
Dark
>2000b
>2000b
>2000b
8.0
(0.6)
(0.1)
0. 0
0. 0
1-1/2 Suns
11. 1
(0.0)
3 Suns
14.4-(0.5f
2.6-(1.7)C
(0.lH2.lf
14. 6
5. 2
0.3
2.2
3. 1
aUnbracketed values refer to NO£ decomposition (cpc), while values in parentheses
are for NO oxidation to NO£ (cp^)- Data were taken from the same experiments
used to generate Table ].
bThese values were calculated from initial optical densities (e.g. , t x; 10-100 min
in Figure 9). They essentially represent a lower limit for
-------
Table 4. INITIAL OPTICAL DENSITIES IN
NO2 - N2 EXPERIMENTS
CATALYST
AI203
AI203
CHARCOAL
Mn02
PbO
CEMENT
Pt
ITorr)
688
312
204
688
689
691
688
687
688
691
F
(cm /sec)
13.5
10.3
13.3
5.2
12.0
12.3
14.3
14.0
12.8
12.2
NO2
(%)
0.2
0 2
0.2
0.2
0.05
0.05
0 2
0.2
0.2
0 2
AVERAGE PNQ (Torr)
UNCORRECTED
0.448
0. 152
0. 104
0.309
0.097
0.096
0.472
0.484
0.496
0.432
CORRECTED
0. 180
0.030
0.024
0.037
0.029
0 028
0.204
0.212
0.228
0. 168
CALCULATED*
0. 176
0.031
0 018
0 075
0 039
0.040
0 184
0. 180
0. 164
0 160
*Assuming 4> 5 1 x 10
-4
37
-------
-4
model, namely, cp> 10 . Lead oxide and cement were also found to be
reactive. However, other materials, including fly ash, soil, and crushed
oleander leaves, were unreactive (cp < 1 0 ).
A number of experiments were conducted to identify products of the
heterogeneous NO,, decomposition reaction. The results from one such
experiment are plotted in Figure 11. In that experiment, the average NO?
pressure and NO concentration were measured as a function of time. The
average NO? pressure increased as before. Gaseous NO was also detected
in the effluent gas, but only after a substantial induction period. The results
illustrated in Figure 11 are representative for most of the reactive materials.
In general, the NO concentration in the effluent was - 40 - 60% of the NO.,
in the influent reaction mixture. (The NO,, was completely destroyed during
transit through the reactor except at the end of the experiment, where the
solid became totally unreactive. ) As discussed earlier, in separate experi-
ments with NO-N? or NO-A mixtures, NO did not adsorb or react on these
C*
surfaces.
Additional experiments were conducted to determine the composition of
nitrogen compounds deposited on the solid surfaces. Wet chemical methods
- 24
were used for NO.-, and NO, determinations. In general, these experiments
were run only for a short period of time to avoid poisoning of the solid surface.
In some cases, after the experiment, the solid was removed in segments,
washed, and the washings analyzed for NO . Results from one such experiment
with NO.J-N,-, over Al O, are illustrated in Figure 12. The prominent surface
species after reaction was NO,,, with a small amount of NO.,, although, for
convenience, we have plotted the total NO^ + NO, as NO, in Figure 12. The
points in the figure were calculated using the model in Reference 23 and Appendix
B , and were normalized for absolute comparison with the experimental data using the
total NO,, flux into the reactor during the experiment. For the most part, the
L-i
calculated and experimental results are in good agreement. The biggest
discrepancy is at Z< ~ 1-cm, where expected edge effects would give dis-
crepancies in the directions observed. Quantification of the latter effects
would be difficult in this system, although they could be reduced or eliminated
by leaving the first few cm of the cylinder uncoated. Additional qualitative
38
-------
Figure 11
200
TIME, min
300
400
Decomposition of NO-> over MnOp. Reaction mixture:
0.2% NO2 in N2, total pressure was 700 Torr
39
-------
12.5
10.0
D)
N
7.5
5.0
2.5
119J
| [ EXPERIMENTAL
CALCULATED
2.5
5.0
Z, cm
7.5
10.0
Figure 12. Surface nitrogen compounds (as NC>3) from the decomposition
of NC>2 over Al2C>3. Reaction mixture: 0.2% NO2 in N2, total
pressure was 700 Torr
40
-------
and quantitative results are given in Table 5.
Several comments should be made regarding the NO results. First of
!X
all, gaseous NO,, bubbled through water gave a positive test for NCL,
suggesting that both adsorbed NO and NO,, would give positive tests for NO,, .
In addition, the mere act of "washing" the solids might result in chemical
reaction. For example, adsorbed NO? might dissolve in water and react
with surface OH groups to form NO,. Thus, the NO results should be
.J X.
considered cautiously. The latter comments especially also apply to results
obtained when water was added to the reaction mixture and greatly enhanced
reactivities were observed (see below).
CFF Reactor Experiments
Gaseous mixtures in the CFF reactor were generally limited to NO,-,
plus N? or A. The NO,, concentration (detected by laser excited fluorescence)
was measured as a function of distance down the cylinder axis. Results from
a typical experiment are illustrated in Figure 13. The solid curve in the
figure was calculated using the model for diffusion, flow, and wall reactions
described in Reference Z3 and Appendix B. The latter solution isaBesselseries that
reduces to a single exponential term at large distances. An example of the
exponential decay with distance at large distances is illustrated for data
from another experiment in Figure 14.
Data from experiments conducted in the CFF were analyzed by comparing
measured NO? concentration profiles to those obtained from the analytical
LJ
model for various values of cp. Results from experiments with charcoal
conducted in cylinders with radii of 0. 50- and 0. 95-cm are given in Table 6
_3
for various experimental conditions. The average cp of 1.6 x 10 -would give
a particle limited lifetime for NO,, in polluted atmospheres of ~ 1 hr, using
the model in Reference 1. Results for MnO_, exhibited somewhat more
-3 -4
variation with cp ~ 0.3 - 3 x 10 , whereas, Al^O^ gave cp « 3 x 10
Effects of Moisture - Moisture greatly extended the reactivity of those materials
discussed above that were found to decompose NO,,. Alternatively, the activity
of a solid destroyed via the "dry" reaction, could be restored and significantly
extended by adding water to the reaction mixture. Examples of the latter
41
-------
Table 5 . ANALYSIS OF NITROGEN COMPOUNDS
ON CATALYST SURFACE
SURFACE NITROGEN COMPOUNDS
TOTAL (mg) AS N02
CATALYST
AI203a
AI203b
CHARCOAL
Mn02
PbO
N03
SMALL
LARGE
LARGE
LARGE
SMALL
N02 EXPT
LARGE 24
19
0.8
SMALL
SMALL
SMALL
LARGE
CALCULATED
24
14
1.0
-
-
-
_
aAnalabs chromatographic grade activated alumina
bj. T. Baker "Reagent" AI2O3
42
-------
0.75 -
°0.50
o
o MEASURED
CALCULATED
0.25
Figure 13. Decomposition of NC>2 over charcoal. Relative concentrations
of NC>2 (equal to relative fluorescence intensities) vs distance
for a 0. 2% NCU in a mixture at 10 Torr
43
-------
1.000-
0.50
CVI
O
0
z
Q.
0. 10
0.05
\
\
6,
\
\
\
\°
\
\
\
\
\
\
0 EXPT
CALC (0 = 0.002)
(limiting slope)
\
>
\
\
2 3
X, cm
Figure 14. Decomposition of NC>2 over MnC>2. Relative pressures of NO2
(equal to relative fluorescence intensities) vs distance for a
mixture at 10 Torr
44
-------
Table 6. RESULTS FOR NOZ - A MIXTURES OVER CHARCOAL
r,
cm
0_ 50
0.95
P,
Torr
1.0
9.9
10.2
1.0
1.0
9.8
10.3
F,
c c / s e c
17. 5
25. 0
16. 5
0.4
0.4
1.7
1.7
103 rp
2.6
1.2
1.4
1. 5
1. 0
1, 6
1.6
_3
cp (average) = 1.6 x 10
i (atmosphere) :--. 66 min
45
-------
experiments arc shown in Figure 15. An NO?-A mixture was passed through
the cylinder that was coated with a minimal amount of solid in a band ~ 1 in
wide. After ~ 20 - 60 min, depending on the solid, the reactivity toward NO?
was destroyed. Subsequent addition of moisture restored and greatly extended
the activity observed under dry conditions.
More important, experiments in the CFA reactor indicated that virtually
all of the materials listed in Tables 1 and 2., even those that were unrcactive
_4
under dry conditions, effectively decomposed NCX, (cp > 10" ) when moisture
was present (43 - 44% RH). Of the materials listed, PbCl? was not
_5 ^
tested; (NH.)-SO. exhibited minimal activity (cp < ~' 10 ). We also examined
CaO and particulate matter collected from the environment on glass fiber filter
_4
paper. Both materials gave cp > 10 ; however, the results for the particulate
matter from the environment are equivocal since the filter paper itself was
found to decompose NCL. The latter result could have important implications
in particle collection measurements if surface NO is formed.
Summary of NO Reactions on Solid Surfaces
The results for NO,., suggest that this pollutant could be heterogeneously
destroyed in the atmosphere with a lifetime of ~ 1 hr, if actual airborne
particles had reactivities similar to those of charcoal and MnO,-, in our labora-
LJ
tory experiments. This reaction is greatly facilitated by moisture in the
presence of many solids likely to be found in polluted environments. Quantita-
tively, the reaction transforms NO0 to NO and NO , -. ,, and leads to a
y 2 x (ads)
substantial reduction of NO in the gas phase.
_?C
The 1 hr lifetime suggests that particles could be an important nighttime
sink for NO . In addition, although this rate is slow compared with the
noontime photolytic lifetime of NO in the atmosphere (2-3 min), the hetero-
geneous process leads to a net redxiction in oxidant concentration (NO? + O~).
The photolytic process does not. Thus, the overall effect of the heterogeneous
process could be more important than the light-initiated decomposition.
We have considered only airborne particulate matter (~ 100^g/m ) in our
modeling. Consideration of ground level surfaces could greatly enhance total
reactivity. Moreover, reactions taking place on the latter surfaces could
46
-------
0 151
,- 447 R H added
0. 10
0. 05 h
TIME (sec)
Figure 15. Effects of moisture on NO decomposition
47
-------
lead to concentration gradients near the ground, depending on the amount of
transport mixing taking place. These gradients could be significant in
accurately assessing pollution related health problems.
Finally, NO exhibits only minimal activity toward solids examined in this
study. Based on quantitative rate measurements, we conclude heterogeneous
atmospheric processes of NO arc probably unimportant.
RESULTS AND CONCLUSIONS FOR O
Several exploratory experiments were conducted with O~ and various
solids. Gas mixtures containing ozone prepared for chcmiluminescent detection
of NO (Figure 3) were instead fed into the CFA reactor. The average O
pressure in the reactor was determined by measuring optical absorption at
313-nm and experimental absorption coefficients. Here, as in the case of
_4
NO^,, we found that cp > 10 for charcoal and MnO7, whereas (NH.)7SO. was
* LJ T: L* T:
unreactive. In the case of Al^O,,, cp <- 10 was determined. Except for
(NH.)?SO., the effects of moisture were not examined. Water did appear to
lead to some decomposition of O in the presence of (NH.^SO,. However,
reactivities were low and of the order of experimental uncertainties.
For O,, as in the case of NO,,, we conclude that the heterogeneous
reaction of ozone in polluted atmospheres could be important and lead to an
overall reduction of the total oxiclant concentration.
RESULTS AND CONCLUSIONS FOR CO
The results of the study may be divided into two parts: (1) the bulk
reactor and development of the slug reactor; (2) the slug reactor. In the first
part, some valuable observations and kinetic analyses were obtained, but the
initial rates of reaction were not determined. These will be discussed first,
then the slug reactor results will follow. In all preliminary studies, MnO?
L*
was the only metal oxide used since it is known to oxidize CO to CO,, at
10
ambient temperatures. Subsequent to the early treatment of the data,
Kobayashi and Kobayashi ' ' showed that MnO~, is quite reactive even at
lower temperatures.
48
-------
Bulk Reactor
Results from the bulk reactor do not yield initial rates since the large
volume of the reactor and the time required for circulation integrates the
product CO? over fairly large time intervals. Also, problems arising from
the failure to regenerate the MnO? to the same activity level resulted in
extreme data reduction problems in the depletive oxidation studies (without
oxygen present). The data with oxygen were even more intractable. Correc-
tions were made for the varying initial metal oxide reactivity, and the data
were reduced as described below.
To treat the data, the form of the rate equations must first be considered.
Equations (2) and (3) yield a rate expression for CO loss.
-d(CO)/dt - kb(A)"(CO)m + kc(B)p(CO)r + . . . (6)
where A and B are the concentration or number of active sites of each type.
The data from the bulk reactor experiments were reduced by assuming a firsL
order dependence in CO (m - r = 1) and plotting the log of the interval average
rate i ln( Vp.^,-, / it x p^^)] vs time. This allows each experiment to be
*^'^S~* O Vy
normalized to the same MnO,, activity by adjustment along the time axis.
Figure 16 shows the raw data for two separate depletive oxidation runs that
emphasize the difference in metal oxide activity. Figures 17 and 18 show the
results of the data reduction procedure for all experiments for the depictive
studies and studies with added oxygen respectively. Figure 17 shows that,
although the scatter is great, a general trend is obvious, and a line can be
drawn to fit the data. These data best fit the rate law,
-d(CO)/dt - k^CO) AQ e"klt + k2(CO) BQ e"kzt. (7)
It docs not fit a single first or second order dependence in active sites. A
and B arc the initial activity of the metal oxide samples (number or moles of
o
active sites). Two tangent lines can be drawn in Figure 17 associated -with the
two parts of (7) each having the form
In (Aprr. / \t x p ) = k.t - lnk7(CO)A^ (8)
Lvw, (^{J I 1 O
From this equation, the initial rate constant for oxidation k, and the number
49
-------
CM
O
O
Q_
Mn02-11
TIME ( sec x 10 -3 )
Figure 16. Pressure of CC>2 vs time. 11 Torr CO,
200 mg MnO2-
50
-------
o
u
CM
8
o
i
DEPLETIVE OXIDATION
MnO
CO
334
TIME, sec x 1CT3
Figure 17. Depletive oxidation of CO
51
-------
2.4
5.0
TIME, sec x 10
-3
Figure 18. Oxidation of CO in the presence of
52
-------
(or concentration) of active sites may be derived. It was found that
k, -- 351 moles sec and A - 9.5 x 10" moles/g. The rate constant is
i o °
a minimum due to the integration of the raw data; however, extrapolating this
value to the atmosphe re with a MnCL loading of 10ug/m and the measured
activity yields a CO lifetime of 0. 1 yr.
The; data with oxygen present are more complex because of the regenera-
tion of sites by reactions (4) and (5). The inaccuracies of the data preclude
a similar analysis. It was obserxed, however, that addition of even 5 Torr
of O did increase the initial rates by factors of three or so. If a mechanism
i-i
similar to that proposed (reactions Z - 5) is operating, the observation
indicates that the initial rate: constant for oxidation is even higher. Kobayashi
and Kobayashi performed a series of experiments using the transient response
method on the MnO? + CO (+ O?) system at -26° and proposed a mechanism
101112
that agrees quite well with ours. ' '
Slug Reactor
Preliminary Experiments - The originally constructed slug reactor did not
have the sensitivity to allow improvement in the results from the- bulk reactor.
Several experiments were performed to establish the parameters necessary
for later experiments (such as flow rates and metal oxide loadings). Although
quantitative data were not recovered, several observations were important.
It was found that the rate of loss of CO from the slug was unaffected by large
pressures of CO.. This confirmed the earlier conclusion from the kinetic
analysis of the bulk reactor results that CO., does not affect the activity or the
capacity of MriO^. Experiments were also performed in which the slugs were
saturated with water vapor. Again, no change; in the: activity or capacity could
be detected.
Detailed Experiments - Several different processes involving metal oxides
could result in oxidation of CO in the atmosphere. Overall any process
must be catalytic since the metal oxide loadings in the atmosphere are
generally low. Using the generalized equations this means that,
k2(CO) or k3 (C0)
-------
At least three types of catalytic behavior could be included in this requirement:
1) oxidation of CO by a stable active site, A, followed by regeneration of the
site by Oy, 2) oxidation of CO by a labile, active site followed by regeneration;
3) the activation of O by the site which would then oxidize CO. The purpose
LJ
of the detailed experiments was to determine if any of these processes were
possible contributors to the atmospheric CO sink. The study was done in two
parts: depletive oxidation where the metal oxide was activated and then
depleted by CO; and oxidation with oxygen present.
In a slug-type reactor, the solid reactant is placed in a section of a
flowing non-reactive gas stream. The other reactants are introduced in slugs
or pulses by some dosing device (here a gas-sampling valve) and plug flow is
assumed. The slug of reactant then reacts as it passes through the solid
reactant section and is analyzed on exit from the reactor. The data available
are the initial and final reactant product concentrations in the slug, the flow
rate (which is related to the residence time and free volume over the reactant),
and the slug volume. Assuming that we have only one type of site with a rate
of oxidation that is large enough to be detected, the rate of CO loss is,
-d(CO)/dt - k A(CO). (10)
r
A is the effective concentration of active sites and is directly related to the
metal oxide activity, k is actually the product of the rate constant (k ) times
the equilibrium constant K, , ,
C°(g) ^ C°(ads).
This equilibrium was found to be rapid under our conditions since no broadening
of the CO slug was observed and the combination of the rate constant and the
equilibrium constant is justified.
In these slug type experiments each slug is considered separately and
is in fact a separate experiment. Under depletive oxidation conditions where
54
-------
the activity of the metal oxide bed, that is, the effective concentration of A,
docs not change appreciably during the passage of single pulse through
the bed, equation (10)may be integrated with respect to CO. (CO) and
(CO), are outflow and inflow
-In [(CO)o/(CO) ] =
concentrations respectively, A is the effective active site concentration for
, l fi
the n slug, and4 t is the slug transit time?. Equation (12) may bo written,
-[ln(COo)/(CO).)n]/4t = k'rAQ - kr(C02)T, (13)
i
where A is the initial concentration of active sites, and k is the effective
rate constant under the assumptions. (CO ) is the total CO? produced
LJ i LJ
through slug number n. The form of 12 will yield a linear relationship between
the left-hand side and (CO ) when either (CO). - (CO) « CO. or AA« A
however, when neither of these conditions are met, curvature will result. All
experiments were performed under one, or the othei of the proper conditions.
12 is useful for two reasons, first it provides a convenient method of data
presentation, and second, it allows one to determine A . The data for both
o
depletive oxidation and the oxidation with O? present and the linear relationships
are shown in Figures 19-22. The experimental conditions are shown in
Table 7.
The complete equation that must be used to determine the rate constants
is,
d(C02)/dt - kr[A - (C02)] [(CO) - (C02) J . (14)
This equation treats the instantaneous concentrations of reactants and the
gradients that develop during the passage of the slug through the reactor bed.
The integrated form of 13 was used to determine k for each slug. The
approach was to divide the metal oxide bed into j5 arbitrary sections of equal
length and the CO pulse into ,// sections of the same length. Square pulses
55
-------
Table 7. CONDITIONS USED IN METAL OXIDE OXIDATIONS:
SLUG MICRO-REACTOR
Metal Oxide
MnO2
CuO
ZnO
Fe203
Wto Metal Oxide, g
0. 0117
0. 0117
0. 157
00 157
0, 0915
0,, 0915
0.296
0.296
Concentrations in^Each Slug (at 1 atm)a
CO, x 108
3. 63
3» 63
3.63
3.63
3.63
3.63
3.63
3.63
O2, x 106
0
8.28
0
8.28
0
8.28
0
8.28
02/CO = 228
56
-------
0.30r~
0.25
0.20
O
r 0.15
o
o
o 0.10
0.05
0
WITH 02
D
01234567
L(C02) (mo|x107j
Figure 19. Slug reactor data: manganese dioxide
57
-------
6.0|-
5.0
4.0
V)
^ 3.0
o
o
o
o
2.0
1.0
0
WITHOUT 02
I I
D
012345678
E(C02) (memo9)
Figure 20. Slug reactor data: zinc oxide
58
-------
12
x 10
'cj
OJ
zr s
o
O
O
S 4
D
WITHOUT 02
Z(C02)
Figure 21. Slug reactor data: cupric oxide
59
-------
'o 4
o>
CO
o
o
0
WITHOUT 02
D
D
WITH 02
0123
Z(C02) (mol*109)
Figure 22. Slug reactor data: ferric oxide
60
-------
and homogeneous M O distributions were assumed. The integrated
x y
form of 13 was then solved for each H and,//and the resultant concentrations
of CO,,, CO, and A used for each subsequent calculation. The (jfx //) moles
of CO produced in each slug was then compared with the experimental value
and the value of k^ found by iteration to converge the two values. Each slug
should yield an identical value ol k if the mechanism is correct and if A
r o
has been determined correctly.
The results of the analysis of the depletive oxidation of CO by metal
oxides are .shown in Table 8. Using (he initial activity, A , the maximum
rate constant for oxidation in the atmosphere, k A , may be used to
r o
estimate the possible effect of metal oxide oxidation on the lifetime in CO of
the: atmosphere. This condition would correspond to site re-oxidation by
atmospheric: oxygen being faster than the oxidation of CO. Front the last
column of Table 8, it can be seen that, at a reasonable atmospheric loading
of 1 /yg/m for a metal oxide, the half-life of CO is far too long for these
processes to contribute to the CO sink. This is true for the stable site
process no matter what the value of the re-oxidation rate. The particle loading
may actually be higher than this for some metal oxides or the surface area
greater, but the conclusion still remains. With the exception ol MnO., the rate
^ 4
constants or active site concentrations must be in error by more than 10
before the effect would b<^ significant in the atmosphere, k may have significant
errors but not of this magnitude. Kobayashi and Kobayashi's value of k for
MnO^, at lower temperatures is 11 times less than our value for the most
active site. This lends some credence to our k values.
r
The other mechanisms to test are those involving labile oxygen.
Experiments on freshly activated metal oxide with over 200 times the oxygen
concentration in each slug as CO are listed in Table 7 and the data presented
in Figures 18- 21. The data is quite scattered but the effect was in general to
increaise the CO production by only small factors. The rate would need to
increase by at least three orders of magnitude for the labile site or labile O
mechanism to be important in the atmosphere. Since the intercepts of the
lines in the figures are all less than, those without O present, the eftect may
£
be interpreted as slow re-oxidation of the stable active sites lost to the
depletive mechanism. A steady-state appears to have been reached for MnO?
61
-------
Table 8. SLUG REACTOR DATA SUMMARY: RATE CONSTANTS AND
CONCENTRATION OF ACTIVE SITES
Metal
Oxide
MnO
ZnO
CuO
Fe2°3
Rate Constant, kr,
cc/mole secx 10~6,
m^/mole sec.
0. 52±0. 09
2. 5±1. 2
2. 9±0. 6
10. 2±0. 3
Active Sites
AO,
moles/g
6. 97x 10"5
-8
3. 44 x 10
-8
3. 82x 10
-9
7. 48x 10
Ao,
mole sites /mole M O
x y
6. lx 10~3
-6
2.8x 10
-6
3. Ox 10 °
-6
1.2.x 10
)-,
kar >
^/Ug yr.
1. lx 10"3
-6
2. 7x 10
-6
3.5x10
-6
2. 4x 10
CO half -li
rl/2, yr.
606
5
2. 6x 10
5
2. Ox 10
5
2. 9x 10
a.
b.
c.
Only two slugs completely depleted maximum loading of sample in the loop,
error is range between the two values.
kar is the atmospheric rate constant assuming maximum activity (k A - k A ).
T ,7 is computed assuming 1 //g/m atmospheric loading of M O .
62
-------
where dA/dt = 0. At oxygen concentrations outside the range of the apparatus,
this would presumably occur for the other oxides as well. Treating the MnO
cata as a steady state of 1 5 for the slugs where the rate of CO-, production is
4 -1 - \
constant (see Figure 18) yields a value of(l.l±0.4)xlO cc mole sec
for k (k = K, ^ , .k,).
o o (g^ads) 4
-dA/dt = k (CO)(A) - k (O )(A - A) (15)
37 O ^ O
It would appear that the effect of oxidation by metal oxide particles in
the atmosphere on the rate of CO loss from the atmosphere is negligible.
This conclusion is based on limited data on only four metal oxides. However,
manganese dioxide is known to be one of the more efficient oxidizers of CO
at room temperature (due to its large site concentration and not to a high rate
constant) and its reaction is too slow to be important by at least three orders
of magnitude. One exception possibly exists and that is lead oxide, PbO.
We attempted to study this oxide, however, the material could not be rid of
labile CO? and no quantitative measurements could be made. When slugs of CO
were passed through our reactor, no CO effluent could be detected. The
capacity to consume CO was great and the k A product must be at least that
of MnO7. Lead oxide-carbonate equilibria are known to be facile and one
£j
might speculate that a mechanism akin to that observed for NO,, on metal
oxides discussed elsewhere in this document might be operating.
While this work was primarily directed toward atmospheric particle
processes, surface reactions at ground level should also occur by similar
34
processes. Hidy has discussed the rate of loss of species at surfaces
frorTi a purely mass transport standpoint. Using a global production rate of
4 ~) o C -J A ""*
~l'x 10 kg/yr of CO and a world land mass area of 5. 12 x 10 m ,
the concentration of CO in the atmosphere is estimated to be 5 ppb using a
34
deposition velocity at 1 cm/sec." and assuming that each collision results
35
in loss of CO from the atmosphere. Since the global CO average is . 1 - 2 ppm
this calculation only shows that delivery of CO to the earth's surface is not
the limiting factor. To check the viability of heterogeneous process occuring
on surface layers of soil by purely chemical means, a test calculation was
63
-------
performed using our rate constants and maximum activities for the four
metal oxides under the following assumptions:
1. The average crystal content of the various metals -was used
in soil with 1 m^/g34 surface area and a density of 1. 25
g/cc including 50% void volume.
o /
2. The penetration depth of gases into the soil was 2 cm.
3. The composition of the surface of the soil particles is that
given in 1 above and 10% of the surface metal concentration
exists as the active metal oxide form.
Table 9 shows the results of the computation. The rate constant,
k , has been converted into units that may be compared with the deposition
sr . 5
velocity. Hidy's value for CO of 1 cm/sec translates to 3 x 10 m/yr.
compared with the largest rate constant of 59 m/yr. This shows that under
the assumptions the chemical process is rate limiting. The rate of CO
consumption by these oxidation processes is shown in the last column. The
total for the four metals is approximately 1% of the total CO input rate.
While this would be a small contributor to the CO sink the inclusion
of other metals may well increase this contribution significantly. Of course
the assumptions used in the calculation may be overly optimistic and
the rate may be very small. In conclusion, further work examining the
oxidative abilities of other metal oxides and of natural surfaces are necessary
to ascertain the importance of metal oxides as ground level CO sinks.
64
-------
Table 9. ESTIMATED RATES OF CO OXIDATION BY
METAL OXIDES ON SOILS
Metal
Fe
Mn
Zn
Cu
Crustal
Composition, a
w/w
5 x 10~2
1 x 10"3
1. 3 x 10"4
7 x 10~5
A ,b
OS
moles/cc ( soil]
1. 7 x 10~12
9 x 10 "H
Z. 5 x 10~14
6. 7 x 10"14
K ,C
sr
m/yr
21. 9
59. 0
0. 079
0. 245
Total
Rates CO lossd
kg/yr
2. 5 x 109
6. 9 x 109
9. 2 x 107
2. 9 x 107
9. 5 x 10V
a. Ref. 37
b. Moles of active sites per cc of soil, see text for assumptions.
c. (2 k A d), where 2 corrects for the volume occupied by the soil
O1
and the d is the depth of the active layer (2 cm).
14 2
d. Based on a world land area oi 5. 12 x 10 m and 0. 2 ppm.
65
-------
SECTION VII
REFERENCES
1. Ford, H. W. and N. Endow. Rate Constant at Low Concentrations. III.
O
Atomic Oxygen Reaction in the Photolysis of NO7 at 3660A. J. Chem.
4Li
Phys. 27: 1156 - 1160, November 1957.
2. Altshullcr, A. P. and J. J. Bufanlini. Photochemical Aspects of Air
Pollution: A Review. Photochemistry and Photobiology (London). 4:
97 - 146, 1965. Environ. Sci. Tech. _5: 39 - 64, January 1971.
3. Schuck, E. A. and E. R. Stephens, Oxide of Nitrogen. In: Advances in
Environmental Science, Pitts, J. N. and R. L, Metcalf (ed). New York,
Wiley-Interscience. 1969. Vol. 1. p. 73 - 118.
4. Judcikis, H. S. and S. SiegeJ. Particle-Catalyzed Oxidation of Atmospheric
Pollutants. Atmospheric Environment. 7: 619 - 631, June 1973.
5. Weinstock, B. and H. Niki. Carbon Monoxide Balance in Nature. Science.
1_7JS: 290 - 292, April 1972.
6. Bortner, M. H. , R. H. Kummler, and L. S. Jaffe. A Review of Carbon
Monoxide Sources, Sinks and Concentrations in the Earth's Atmosphere.
General Electric Co. Philadelphia, Pa. NASA CR-2081. National
Aeronautics and Space Administration. June 1972. 51p.
8. Nagarjunian, T. S. and J. G. Calvcrt. The Photooxidation of Carbon
Monoxide on Zinc Oxide. J. Phys. Chem. 68: 17 - 26, January 1964.
9. Brooks, C. S. The Kinetics of Hydrogen and Carbon Monoxide Oxidation
over a Manganese Oxide. J. Catalysis. 8: 272 - 282, 1967.
10. Kobayashi, M and H. Kobayashi. Application of Transient Response Method
to the Study of Heterogeneous Catalysis. I. Nature of Catalytically Active
Oxygen on Manganese Dioxide for the Oxidation of Carbon Monoxide at Low
Temperatures. J. Catalysis. 2T\ 100 - 107, October 1972.
11. Kobayashi, M. and H. Kobayashi. Application of Transient Response Method
to the Study of Heterogeneous Catalysis. II. Mechanism of Catalytic
66
-------
Oxidation of Carbon Monoxide on Manganese Dioxide. J. Catalysis. 27:
108 - 113, October 1972.
12. Kobayashi, M. and H. Koba^ashi. Application of Transient Response Method
to the Study of Heterogeneous Catalysis. III. Simulation of Carbon Monoxide
Oxidation Under an Unsteady State. J. Catalysis. 27: 114 - 119, October
1972.
13. Hall, W. K. and P. H. Emmett. An Improved Microcatalytic Reactor.
J. Am. Chem. Soc. 7j?: 2091, 1957.
14. Calvert, J. G. and J. N. Pitts. Effects of Wavelength and Temperature
on Primary Processes in the Photolysis of Nitrogen Dioxide and a Spectro-
scopic Photochemical Determination of the Dissociation Energy. J. Chem.
Phys. 42(12): 3655 - 3662, June 1964.
15. Fontijn, A., A. J. Sabadell, and R. J. Ronco. Feasibility Study for the
Development of a Multifunctional Emission Detector for Air Pollutants
Based on Homogeneous Chemiluminescent Gas Phase Reactions. AeroChem
Research Labs, Princeton, N0 J. Report Number TP-217. September 1969.
16. Hall, T. C. and F. E. Blacet. Separation of the Absorption Spectrum of
NO7 and N.,O, in the Range of 2450 - 5000A. J. Chem. Phys. 20(11):
Li C* T:
1745 - 1749, November 1952.
17, Gelbwachs, J. A., M. Birnbaum, A. W. Tucker, and C. L. Fincher.
Fluorescence Determination of Atmospheric NO7. Opto-electronics
LJ
(London). 4: 155 - 160, May 1972.
18. Hall, W. K. , D. S. Maclver, and H. P. Weber. Semiautomatic Micro-
reactor for Catalytic Research. Inc. Eng. Chem. 52: 421 - 426, May I960.
19. Stewart, T. B. A Positive Displacement Gas Circulating Pump. Rev. Sci.
Inst. 44: 1144, August 1973.
20. Shoemaker, D. P. and C. W. Garland. Experiments in Physical Chemistry.
New York, McGraw-Hill, Inc., 1967. p. 262 - 271.
21. Brunaur, S. , P. H. Emmett, and Teller. Adsorption of Gases in Multi-
molecular Layers. J. Am. Chem. Soc. 60: 309 - 319, February 1938.
67
-------
22. Cheng, R. T. , J. O. Frohlinger, and M. Corn. Aerosol Stabilization
for Laboratory Studies of Aerosol-Gas Interactions. J. Air Poll. Control
Assoc. 21: 138 - 142, March 1971.
23. Judeikis, H. S. and S. Siegel. Efficiency of Gas-Wall Reactions in a
Cylindrical Reactor. The Aerospace Corporation. El Segundo, Calif.
Report Number ATR-73(7256)-2. 23p. See also Appendix B.
24. Feigl, F. Spot Test in Inorganic Analysis. 5th ed. El Sevier Pub. Co. ,
New York, 1958. p. 326 - 332.
25. Mullcr, P. K. , S. Twiss, and G. Sanders. Selection of Filler Media: An
Annotated Outline. Air and Industrial Hygiene Laboratory, State of California,
Department of Public Health. (Presented at the 13th Conference on Methods
in Air Pollution and Industrial Hygiene Studies. Berkeley. October 30-31,
1972) 12p.
26. Calvert, J. G. and J. N. Pitts. The Interaction of Light with Simple
Molecules. In: Photochemistry. New York, John Wiley & Sons, Inc. ,
1966. p.208.
27. Denbigh, K. Tubular Reactors. In: Chemical Reactor Theory. Cambridge,
The University Press. 1961. p. 46.
28. Moelwyn-Hughes, E. A. Mathematical Formulation of the Kinetic-Molecular
Theory. In: Physical Chemistry. 2nd ed. New York, Pergamon Press.
1961. p.46.
29- Paneth,F. and K. Herzfeld, Z. Elektrochem. , 37, 577 (1931).
30. Wise, H. and C. M. Ablow, J. Chem. Phys.,_29, 634 (1958).
31. Jost, W. Diffusion in Solids, Liquids, Gases, Academic Press Inc. ,
N. Y., 1960, pp 51-54
32. Carslaw, H. S. and J. C. Jaeger, Conduction of Heat in Solids,
2nd ed, Clarendon Press, Oxford, 1959, pp 188-213.
33. Present, R. D. , Kinetic Theory of Gases, McGraw-Hill Book Co. , Inc. ,
N. Y., 1958, pp 52-55.
68
-------
34. Hidy, G. M., Removal Processes of Gaseous and Particulate
Pollutants. In: Chemistry of the Lower Atmosphere, ed. by
S. I. Rasool. New York, Plenum Press, 1973. pp. 121-173.
35. Jaffe, L». S., Carbon Monoxide in the Biosphere: Sources,
Distribution, and Concentrations. J. Geophys. Res. 78: 5293 - 5305,
August 1973.
36. Seim, E. C., Sulfur Dioxide Absorption by Soil. Diss. Abs. Intl.
B31: 5111-B, 1971.
37. Weast, R. C., cd. , Handbook of Chemistry and Physics, 52 ed.
Cleveland, The Chemical Rubber Co, 1971, pF-163.
69
-------
SECTION VIII
LIST OF INVENTIONS AND PUBLICATIONS
A. Invention:
Stewart, T. B. Positive Displacement Gas Circulating Pump. The
Aerospace Corp. Disclosure Number 72-30 'Piston Pump1. Reported
separately to E.P. A. on Form 3340.
B. Publications:
1. Judcikis, H. S. and S. Siegel. Particle-Catalyzed Oxidation of
Atmospheric Pollutants. Atmospheric Environment (London). 7:
617 - 631, June 1973.
2. Stewart, T. B. Positive Displacement of Gas Circulating Pump.
Rev. Sci. Instrum. 44: 1144, August 1973.
3. Judeikis, H. S. and S. Siegel,, Efficiency of Gas-Wall Reactions in a
Cylindrical Flow Reactor. Submitted to J. Phys. Chem.
4. Hedgpeth, H. , S. Siegel. T, Stewart, and H. Judeikis,, Cylindrical
Flow Reactor for the Study of Heterogeneous Reactions of Possible
Importance in Polluted Atmospheres. Submitted to Rev. Sci.
Instrum.
5. Badcock, C. C. , S. Siegel, T. B. Stewart, and J. K. Allen. On the
Possibility of the Heterogeneous Oxidation of CO in the Atmosphere.
In preparation.
60 Stewart, T. B. and H. S. Jtideikis. Spatial Reactant/Product Concen-
tration in Flow Reactors Using Non-Probe Laser Induced Fluorescence,
In preparation.
7. Ho S. Judeikis, S. Siegel, To Stewart, and H. Hedgpeth. The Role of
Gas-Solid Interactions in Polluted Atmospheres, II. Reactions of NO?
In preparation.
8. H. S. Judeikis, S. Siegel, To Stewart, and H. Hedgpeth. The Role of
Gas-Solid Interactions in Polluted Atmospheres. III. Reactions of NO.
In preparation.
70
-------
C. Reports:
1. Judeikis, H. S. and S. Siegcl. Particle-Catalyzed Oxidation of
Atmospheric Pollutants The Aerospace Corporation. El Scgundo,
Calif., Report Number \TR-73 (7256)- 1, November 1972.
2. Judeikis, H. S. and S. Siegel. Efficiency of Gas-Wall Reactions in
a Cylindrical Flow Rca :tor. The Aerospace Corporation. El Scgundo,
Calif., Report Number ATR-73 (7256)-2, January 1973. 23p.
:> Stewart, T. B. Positi\e Displacement Gas Circulating Pump. The
Aerospace Corporation. El Segundo, Calif., Report Number
ATR-73(7256)-3, June 1973. 7p.
4. Hedgpeth, H. , S. Siegel, T. Stewart, and H. Judeikis. Cylindrical
Flow Reactor for the Study of Heterogeneous Reactions of Possible
Importance in Polluted Atmospheres. The Aerospace Corporation.
El Segundo, Calif., Report Number ATR-73 (7256)-4 , August
1973. Zip.
71
-------
SECTION IX
APPENDIXES
Page
A. Measurements of Homogeneous and Heterogeneous
Reaction Kinetics of NO by Optical Absorption 73
B. General Solution for Flow, Diffusion, and Wall
Reactions in a Cylindrical Reactor 76
72
-------
APPENDIX A
MEASUREMENTS OF HOMOGENEOUS AND HETEROGENEOUS
REACTION KINETICS OF NO BY OPTICAL ABSORPTION
HOMOGENEOUS GAS PHASE OXIDATION OF NO
Oxidation of NO by O~, in the gas phase is described by the following
3 ^
relation
d[N07 J/clt = 2k_ [N0]2[07j. (15)
£.' Cj i.
In a tubular-flow reactor, such as that described in this paper, the rate of
change of '. NO? j with time is related to its rate of change with distance by the
expre s s nn
d[N02J/dt = dV/dl dZ/dV d[NO2 ] /dZ = F/ A d[ NO2'J /dZ (16)
where A = the cross sectional area of the tube
F - gas flow rate (cm /sec)
Z = length down the tube.
Equation (16) is based on the plug-flow assumption, that is, over any cross
section normal to the gas flow, the mass flow rate and gas properties (pressure,
27
temperature, composition) are uniform. By combining Equations (15) and
(16), using the material balance relation
[NO] -- [NO] + [NO?J (17)
O LJ
and integrating, we obtain the following:
[N02] =[ .*Z/(1 + aZ)][NO]Q (18)
where NO,-,] = the NO? concentration down the tube
a = 2(A/F) kG[02 i;NO]0
In deriving Equation (18), we have assumed [ O? ] ^. constant since, in our
experiments, [O?]/[NOJO< 100 and oxygen is not depleted to any appreciable
extent during the course of reaction.
Assuming that the absorption of light by NO? in the tubular reactor is
described by the Beer-Lambert law,
d log I/dZ = -e[NO ] . (19)
73
-------
where e = the molar extinction coefficient of NO .
L-,
[We shall use log for common logarithms (base 10) and In to designate
Naperian logarithms. Also the association reaction 2NO = IsLO. could cause
(-1 C-j TT
deviations from the Beer-Lambert lav/. However, in our experiments the
pressure of NO? was kept sufficiently low so that this association reaction was
negligible. ] Substitution of Equation (18) into Equation (19) and integration
yields the result
log(I0/I) - e[NO]0J* Ll -(l/ctf)ln(l + ^)J (20)
where H = length of the tubular reactor.
Experimental values for I /I, [NO] , [O? ], H, A, and F were substituted into
the expression for ot and Equation (20). Using the appropriate value for e
(Figure 5), the resulting expression "was solved iteratively to determine k«.
CATALYTIC REACTIONS IN THE NO-O2-N2 SYSTEM
To some extent, catalytic activity in this system could be determined by
use of Equation (20). Thus, if catalytic decomposition of product NO? (from
the homogeneous gas phase oxidation of NO by O?) occurred, the observed
optical density would decrease and solution of Equation (20) would give a
smaller effective k~. Conversely, if NO were catalytically oxidized to NO?
at the walls, a larger effective value of k~ would be obtained.
Alternatively, let us assume the overall surface reaction (decomposition
of NO,,) can be described by a term such as
-cp k (S/V)[NO ] (21)
c c\ / £
where k = the rate of collision of NO-, molecules with a surface
c £
S = surface
cp = the fraction of collision that lead to reaction.
[Term (21) represents a slight modification of the relation given in Reference
28, to include cp and V. Note that this form of the equation assumes that the
active surface is homogeneously distributed throughout the reactor. This
assumption, although an oversimplification, is sufficient for the screening
purposes of these experiments,, ] Adding this term to Equation (1), with
kc= (kT/2aMNQ )1/2 ' (22)
LJ
74
-------
where
k = Boltzman constant
T = absolute temperature
M,_. = mass of NO molecule.
Proceeding as before, we obtain the following expression
logl
e[No] a
4- [ 1 + exp(wjf)
(23!
where <> = has same meaning as before
ti Ml/2o)cp (A/F) k [SJ
, nt.?/..l/2c
w - 2cc p (1 + 2 / p )
Given the appropriate values from experimental measurements or the litera-
ture for the quantities appearing in these equations, they can be solved
iteratively for cp .
Alternatively, we may consider the possibility of catalytic oxidation of
NO to NO? at the surface. In this case, we may substitute the term
/VJ[ NO] (24;
where cp' = the fraction of NO-wall collision leading to reaction
1/2
for Equation (21).
(25)
where p ' - (1 /2a)l4>1 (A/F) k'^ [ Sj .
As in the precceding case, these equations may be solved iteratively for cf
T
c
using the appropriate experimental and literature values for the various
parameters.
75
-------
APPENDIX B
GENERAL SOLUTION FOR FLOW, DIFFUSION, AND WALL REACTIONS
IN A CYLINDRICAL REACTOR
We consider a system in -which the cylindrical walls (but not the ends)
of a tubular flow reactor are coated with catalyst. We assume that C, the
reacting gas, disappears via a heterogeneous first-order process (adsorption
or reaction) that occurs on the walls with an efficiency 0. If the reacting
gas is introduced into the system as a dilute component in a large concentration
of inert gas, we can assume that the total pressure P and therefore D, the
diffusion coefficient of C in the mixture, are approximately constant during
the course of reaction. Fick's second law of diffusion with an additional
term for flow in the direction of the cylinder axis (under steady-state conditions)
. 29-31
=
r 8 r ~2 / o 8 x
ox/
where r and x refer to radial and longitudinal cylindrical coordinates, and
v is the linear flow rate in the x direction (equal to F/A, the volume flow
rate divided by the cross-sectional area of the cylinder).
The following boundary conditions are applicable to the solution of
Eqxiation (26) for the case of interest here
C(r, x = 0) = C
!£. (r = 0, x) - 0 (27)
D - (r = R, x) = 0 krC '
where C is the initial concentration of C, R is the cylinder radius, k is
o r
the average velocity of the reacting molecule in the radial direction
76
-------
Tk - (kT/2jTrri) , where k is the Boltzmann constant, T is the absolute
^* ? m '
temperature, and m is the mass of the reacting gas molecule ] and C is
the concentration of C one mean free path away from the wall.
Paneth and Herzfeld and later Wise and Ablow assumed C(R) ~ C
and obtained the solution of Equation (26) on the basis of the boundary
conditions in Equations (27). This assumption, however, is valid only for
0«1 . In general, C(R) = (1 - 0) C1 . If we apply this assumption and the
boundary condictions in Equations (27), the solution of Equation (26), follow-
29-32
ing the methods described by other workers, is
oo
_
C
[exp(o«. x)]
where
(29)
J and J, are Bessel functions of the first kind, p. is the ith root of the
o 1 L
equation
(30)
and
or
F
2 AD
2 AD
1/2
(31)
The application of the limiting cases for 0«1 or 0 = 1 to Equation (28) gives
29-31
the results previously preported. (Alternatively, the application of
these conditions to the boundary conditions in Equations (27) gives the same
boundary conditions specified by the previous authors for «! or 0 - 1. )
77
-------
P and D are related by
D, 3000kT /*kT) (32)
87Td2P ^ 2// '
where // is the reduced mass of the reacting and inert gas molecules, and
d = (d 4- d )/2, where d and d are the molecular diameters of the
m n m n
reacting and inert gas molecules treated as hard spheres. The other terms
have the same representation as before.
In some experiments, we determined the average value of the NO
concentration at r « 0 by measuring the decrease in intensity (due to optical
absorption by NO? at 475 nm) of a narrow beam of light passing through the
center of the cylinder. Calculated values corresponding to the average NC>
L*
concentration at r = 0 may be obtained by setting r = 0 in Equation (28) and
integrating. In the latter case,
r
Jo
r = 0 => o i ' r "-*: *- ' (33)
where / is the length of the cylinder.
78
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-650/3-74-007
3. RECIPIENT'S ACCESSION NO.
4. TITLE ANDSUBTITLE
The Role of Solid Gas Interactions in Air Pollution
5. REPORT DATE
August, 1974
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
S. Siegel, H. S. Judeikis, and C. C. Badcock
ATR-75(7441)-1
9. PERFORMING ORGANIZATION NAME AND ADDRESS
The Aerospace Corporation
2350 E. El Segundo Blvd.
El Segundo, California 90245
10. PROGRAM ELEMENT NO.
1AA008
11. CONTRACT/GRANT NO.
Grant No. 801340
12. SPONSORING AGENCY NAME AND ADDRESS
Office of Research and Development
U. S. Environmental Protection Agency
Washington, D. C. 20460
13. TYPE OF REPORT AND PERIOD COVERED
Final Report. 7/1/71-1 0/31/
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
is. ABSTRACT
study was undertaken to evaluate the potential importance of gas -solid
interactions in polluted atmospheres. Model calculations that employed collision the-
ory, transition state theory, and data from the catalysis literature were used to deter
mine the conditions under which the heterogeneous processes could compete with
homogeneous gas phase reactions known to be important.
Laboratory experiments were conducted with simulated atmospheres to determine
whether or not the theoretically derived criteria could be met under ambient conditioq
Among the gases studied were NC^, NO, 03, and CO. The selection of the solids
used in these studies was based on their abundance in polluted atmospheres, as well
as on their known catalytic activity.
Evaluation of the experimental results leads us to conclude that heterogeneous de-
composition of NO;? and O^, as well as heterogeneous oxidation of CO, can be impor-
tant atmospheric processes. Results from experiments conducted in the presence of
moisture indicate that water does not poison catalytic activity. In fact, in the case of
NO;?, activity is significantly increased in the presence of moisture.
In contrast to the results for NO?, Oo , and CO, it was found that heterogeneous
processes that involve NO are not likely to compete with gas phase reactions.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Aerosols
Chemical Reactions
Catalysis
NO, NO2, O3, CO
Chemistry
Physical
Chemistry
3. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report)
Unclassified
21 NO. OF PAGES
84
Release Unlimited
20 SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
79
-------
D
« <
8 -.3
Sfr,o°
------- |