EPA-650/2-74-063
JUNE 1974
Environmental Protection Technology Series
i:!:iiig:ii^
-------
EPA-650/2-74-063
ADSORPTION OF ODOROUS
POLLUTANTS BY ACTIVE
MANGANESE DIOXIDE
by
D.F.S. Natusch, J.L. Hudson, R.L. Solomon,
R. Tanner, and A. Miguel
School of Chemical Sciences
University of Illinois at Urbana-Champaign
Urbana, Illinois 61801
Grant No. R-801603
ROAP No. 21-AFB-010
Program Element No. 1AB015
EPA Project Officer: Belur N. Murthy
Control Systems 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
June 1974
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does rtot 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
-------
ill
TABLE OF CONTENTS
I. Introduction 1
II. Experimental Method 1
A. Apparatus 1
B. Preparation of MnO- Sawdust 2
C. Operating Conditions 3
III. Results 3
A. Hydrogen Sulfide 3
1. Performance of MnO -Sawdust and Activated Carbons 3
2. Concentrations 6
3. Relative Humidity 10
k. Temperature 12
5. Residence Time 12
6. Mass Transfer Effects lU
7. Particle Size lU
8. Pressure Drop 16
9. Optimization 16
10. Summary 16
B. Mercaptans 18
C. Amines 18
IV. Fundamental Studies on MnO /Sawdust 19
A. Standard Preparation Procedure 19
1. Materials 19
2. Procedure 19
B. Hemlock, KMnO, , NaOH Preparation 20
C. Hemlock, HaMnO,, NaOH Preparation 22
D. Hemlock, KMnO, , KOH Preparation 25
E. Method of Addition of KMnO^ 28
F. Reproduction of Catalyst 28
V. Conclusions ^
VI. Recommendations -1
VII. List of References 33
VIII. Appendix ^
A. Conversion Factors ^
B. Tables 35
-------
-------
-1-
I: INTRODUCTION
In this report we discuss the adsorption of pollutants such as
H S, mercaptans, and amines on a solid filter which is a manganese oxide
deposited on the surface of wood sawdust. The work was carried out in the
one year period June 15, 1972, to June 15, 1973.
The filter is prepared by reaction of KMnO, with the sawdust.
For convenience, the oxide is denoted by MnO? although its exact form has
not yet been determined. Its performance is also compared to several
commercially available activated carbons. For H S, both the efficiency and
capacity of the MnO -sawdust are superior to those of all activated carbons
tested. The influence of a number of operating parameters on odor removal
efficiency and capacity was determined. Some data correlation has been
made with a view toward possible use in preliminary system design and/or
scale-up, and fundamental studies on the filter material were made. A standard
preparation of MnO_-sawdust was employed and its performance assessed as a
function of odor concentration (O.U-200 ppm for H S), relative humidity
t-9Q%), flow rate (20-600 ft/min) ,*bed depth (1-3 inches)* and particle
size (U-6, 6-8, 8-10 and 10-lU mesh, Tyler). Hemlock and fir sawdusts were
employed and limited studies were carried out with the ground hard inner ring
of corn cobs.
II. EXPERIMENTAL METHOD
A. Apparatus
The MnO_-sawdust or activated carbon to be tested was held in a bed
through which an air pollutant mixture was drawn by a vacuum pump. The desired
concentration of odor in the inlet to the bed was obtained by mixing the main
air stream with a second stream of high odor concentration; the latter was
contained in a laminated Mylar balloon or a gas cylinder. The desired relative
humidity was obtained by humidifying or dehumidifying the main air stream.
*Although it is EPA policy to use metric units in all its reports, certain non-
metric units are sed herein for convenience. Readers more familiar with
metric units are requested to use the Conversion Factors listed in the Appendix.
-------
-2-
The absorbent filter material was held in position by a stainless steel vire
mesh and was compacted gently until a reproducible standard pressure drop
across the bed was achieved for a given height of the filter bed. The diameter
was usually three inches, although some runs were made with a diameter of one inch.
HpS concentrations in the inlet and outlet streams were determined with a Gelman
Model 23000 paper tape sampler utilizing a silver nitrate impregnated tape
(Natusch et al, 1972). The optical density of the resulting silver sulfide
spot was measured with a Gelman Model 1^101 paper tape densitometer. The
mercaptans were analyzed by means of a gas chromatograph with a flame photometric
detector.
B. Preparation of MnO -Sawdust
MnO -sawdust was prepared as follows:
After grading for size and drying to ambient moisture content, a 200 gram
sample of sawdust was placed in a 1500 ml beaker and soaked in 1000 mis of one
percent reagent grade sodium hydroxide solution for two hours. The sodium
hydroxide solution was then drained off and the sawdust washed with 1000 mis
of water to remove any extracted material.
The sawdust mixture was transferred to a 20 liter Pyrex container,
covered with water and heated to 80 C. 200 grams of crushed technical grade
granular KMnO, (Carus Chemical Co.) were then added. The mixture was stirred
rapidly and the highly exothermic reaction between sawdust and permanganate was
considered complete when it ceased to produce heat. After removal of the liquid,
the MnO -sawdust was placed on a 20 mesh screen to dry under an exhaust hood.
(Note: If this material is taken to complete dryness at a temperature greater
than 100°C over a period of days, it can ignite.) The final material has a
bulk density of about 0.38 g/cm .
No effect of storage life on either the activity or capacity of
MnOp-sawdust for H S has been observed over a period of at least two years. The
majority of the MnO prepared in this way adheres firmly to the surface of the
sawdust (or ground corn cobs) and is not dislodged by passage of air through
the sawdust particles.
-------
-3-
C. Operating Conditions
Unless otherwise stated, all the filter materials were presented in a
one-inch deep and three-inch diameter bed. The particle size was 8 to 10 mesh
and air containing the desired concentration of odor at 60% RH and 75°F was
passed through this bed at 200 feet per minute. These values were varied
individually from the standard conditions given in order to determine the effect
of individual parameters on the efficiency and capacity of each filter material.
III. RESULTS
A. HYDROGEN SULFIDE
1. Performance of MnO?-Sawdust and Activated Carbons
The performance of various adsorbents in removing H S from air, under the
standard conditions given above, is presented in Figure 1. The ratio of the H S
concentration in the air leaving the bed (C ) to the concentration in the inlet
stream (C.) is shown as a function of time.
It is seen that the MnO impregnated fir and hemlock are considerably more
efficient in removing H S from air than any of the activated carbons tested.
(Activated carbons were obtained from Barnebey-Cheney Inc., types CJ, GI, and AC
and Pittsburg Activated Carbon, type BPL; type CJ is impregnated with a metal oxide.
The bulk density range is O.Ul-0.63 g/cm ). The MnO -sawdust filter is at least
an order of magnitude better than the carbons over the course of the eight-hour run.
We also investigated the efficiency of ground corn cobs impregnated with
MnO . As can be seen in Figure 1, this material is less efficient than the MnO
-sawdust but more efficient than the activated carbon. The part of the corn cob
used in this work is very hard, and the lower efficiency compared to the sawdust
substrate may be due to a lower surface area.
In order to run at conditions closer to those which might be
obtained in practice, MnO -sawdust was tested against three activated
carbons at a flow rate of 20 feet per minute and an H S concentration of
80 ppm. The results are shown in Figure 2. These results show the performance
of the activated carbons relative to MnOp-sawdust to be somewhat improved with
respect to the previous testing conditions. However, none of the activated
-------
-4-
Q Hemlock
A Fir
Corncobs
A Carbon (GI)
O Carbon (BPL)
Carbon (AC)
V Carbon (CJ)
0.001
4 6
Time, hours
8
Figure 1. H2S outlet/inlet concentration ratio vs. time for vari-
ous materials
200 ft/min, 5 ppm H2S, 8-10 mesh, 60% relative humidity, 1-in. bed
depth, 75°F
-------
-5-
A Barnebey-Cheney
carbon type GI
O Pittsburgh
carbon type BPL
Barnebey- Cheney
carbon type CJ
8 12 16
Time, hours
Figure 2. H2S outlet/infet concentration ratio vs. time for vari-
ous materials
20 ft/min, 8-10 mesh, 80 ppm H2S, 1-in. bed depth, 60% relative humid-
ity, 75°F
-------
-6-
carbons is comparable to MnO -sawdust in activity.
The foregoing experiments were stopped before the efficiency
of H S removal had dropped to zero; however, we did carry out a few runs
to complete, or nearly complete, bed exhaustion. Some results are shown
in Figure 3 (200 ft/min, 5 ppm H S) for both MnO -hemlock and carbon type
GI. At higher concentrations similar results are obtained although both
the MnO -sawdust and the carbon are exhausted in a shorter time. For
example, at 80 ppm and 200 ft/min, the two curves cross at U-l/2 hours at
which time the efficiency of each has dropped to 15$; the hemlock and
carbon are exhausted in 8 and 10 hours respectively.
The capacities of these filters can be found by integrating
data such as that shown in Figure 3. In this way one obtains the total
mass of H S which is adsorbed before the filter efficiency drops below
some specified value. Capacities in grams of H S removed/gram filter
are shown in Figure h for MnO -hemlock and carbon type GI for H S inlet
concentrations of 5 and 80 ppm as functions of efficiency. The capacities
of the MnOp-sawdust filter are significantly higher than those of the
activated carbon. For example, consider an inlet H S concentration of 80 ppm.
The total H?S adsorbed by the MnO -sawdust before the efficiency drops
below 60% is 0.075 g/g whereas for the carbon, it is 0.00*125 g/g. For an
inlet HpS concentration of 5 ppm, the sawdust and carbon adsorb 0.08U g/g
and 0.0171 g/g, respectively, before the efficiency drops below 60%. The
capacity of the hemlock at exhaustion is roughly 0.12 g/g (either H S level)
while that of the carbon is 0.1 g/g (either H S level).
2. Concentrations
The behavior of the MnO -sawdust and activated carbon type GI
at various H S concentrations may be compared as in Figure 5, where the ratio
of the H S concentrations in the effluent and inlet streams is shown as a
function of the total mass of H S fed to the filter on a per gram of filter
basis. In the case of MnO?-sawdust, the filter efficiency increases somewhat
with the concentration of H S passing through the bed.
-------
-7-
100
10
n Hemlock
A Barnebey- Cheney
carbon type GI
20
60 80 100 120 140 160
Time, hours
Figure 3. Efficiency vs. time
5 ppm H2S, 200 ft/min, 8-10 mesh, 1-in. bed depth, 60% relative humid-
ity, 75°F
-------
-8-
o
o>
CL
O
O
9 Hemlock, 5ppm
A Hemlock, 80 ppm
Carbon (GI), 5 ppm
Carbon (G I), 80 ppm
"8 Q05
100 80 60 40 20
Efficiency, %
Figure 4. Capacity as a function of desired adsorption efficiency
Bed diameter, 3-in.; bed depth, 1-in.
-------
-9-
Hemlock, 0.4 ppm
O Hemlock, 5ppm
A Hemlock , 80 ppm
D Carbon (61), 5ppm
O Carbon (GI), 80 ppm
0.001
0 QO4 O.O8 0.12 Q16
Mass H2S to Filter, g/g Filter
Figure 5. H2S outlet/inlet concentration ratio vs. mass of H2S
feed
200 ft/min, 1-in. bed depth, 60% relative humidity, 8-10 mesh, 75°F,
3-in, bed diameter
-------
-10-
This behavior is consistent with reaction kinetics in which the rate of
reaction between MnO_ and H S is a positive function of the H?S concentration.
The activated carbon efficiency, however, varies inversely with H S concentration.
This is consistent with predictions based on the activated carbon adsorption
»
isotherm (Kipling, 1956).
3. Relative Humidity
The variation of filter activity with relative humidity is shown
in Figure 6 for MnO -sawdust and carbon type BPL. In Figure 6a the ratio of
C /C. at 80% RH to that at any humidity level is shown for both filters
[(C /C.) g0rf RH/ (C /C.)]. This ratio is essentially independent of time
except at the onset of a run where efficiency is nearly 100% at all humidity
levels. The efficiency of MnO -sawdust peaks around Q0% RH. A similar
maximum is often observed for the rate of corrosion of metals by H?S with varying
RH (Backlund, ]966). This maximum is attributed to a reduction in the rate
of H S reaction at high humidity due to formation of a film of water on the
reactive surface. Thus, the rate of mass transfer of H_S to the surface becomes
partially liquid firm controlled. The same effect would also be expected to
occur for activated carbon; however, it should be much less apparent due to
the lower activity of the carbon. In practice no convincing activity peak is
observed for activated carbon.
In Figure 6b is shown the variation of C /C. with time at 'RH =
MnO -sawdust is superior in performance to the activated carbon at all RH levels;
however, neither performs well at low humidities (<20% RH). Figure 6 provides a
concise design basis for efficiency variation with time at all RH levels. For
hemlock:
log [(C /C ) onrf OT/(C /C.)] = 3.19 log (RH) - 6.06 20*
-------
0.5
0.1
* Q05
g
»
6"
\
o° 0.01
0.005
0.001
0
-11-
Hemlock
A Carbon (BPL)
345678
Time, hours
Figure 6s. Dependence of H?S
outlet/inlet concentration on
relative humidity
5 ppm H S, 200 ft/min, 1-in bed
depth, 75°F, 8-10 mesh
1.0
0.9
0.8
0.7
0.6
0.5
a:
5? 04
o
oo
O
O
O
0.3
0.2
0.1
0
0
Hemlock
A Carbon (BPL)
20 40 60 80
Relative Humidity, (%)
100
Figure 6b Dependence of H S
outlet/inlet concentration on
time for RH=t
(Conditions as in 6a)
-------
-12-
h. Temperature
Preliminary studies were conducted to determine whether the
activity and capacity of MnOp-savrdust depend on temperature at constant
relative humidity (kO% RH). No substantial effect was observed over
the temperature range 75-113 F. In this narrow temperature range any
increase in the rate of chemical reaction may be offset by a decrease in the
rate of adsorption of reactant. Control by diffusion or mass transfer
would also explain the experimental temperature behavior. The latter is
suggested by results presented in this paper.
5. Residence Time
The activity dependence of MnO -sawdust on the flow rate of
air passing through the filter bed is shown in Figure 7- No capacity data
are available for Figure 7, although filter capacity is expected to be
essentially independent of flow rate unless autoregenerative processes take
place.
The dependence of the activity of MnO -sawdust on the depth
of the filter bed was also investigated. The ratio of the concentration of
H S in the outlet stream to that of the inlet, of course, decreases as the -
depth of the bed is increased. If there were no mass transfer or channeling
effects and if the activity were independent of H S concentration, the ratio
should initially decrease exponentially with increasing bed height; if the
ratio decreases by some multiplicative factor when the bed depth is increased
from one to two inches, it should decrease by the same factors when the bed
depth is further increased to three inches. In practice, the change from
two to three inches is somewhat less. For example, at the conditions of
Figure 1, the ratio C /C. at the beginning of a run is 0.002, 0.0006, and
O.OOOit for bed depths of one, two, and three inches, respectively.
-------
-13-
Flow fate, ft./min
0.0001
468
Time, hours
10
Figure 7. H2S outlet/inlet concentration ratio vs. time for vary-
ing flow rates
5 ppm H2S, 8-10 mesh, 1-in. bed depth, 60% relative humidity, hemlock,
75°F
-------
-Ik-
6. Mass Transfer Effects
An H_S molecule in the gas stream must be transported to the
surface of the solid MnO -sawdust before it can react. A high free gas
stream velocity reduces the external mass transfer resistance. Thus, by
varying the gas flow rate and the filter bed depth simultaneously so as
to maintain a constant residence time, the importance of mass transfer
limitations in the filter bed can be assessed. The results of such an
experiment are shown for MnO -sawdust in Figure 8.
Clearly under these experimental conditions there is a significant
mass transfer reduction in the efficiency of adsorption of H S-sawdust.
7. Particle Size
The dependence of H S removal on particle size for MnOp-hemlock and
carbon type BPL is given by:
R = 5.^8 x 10 d ~ ' Mn02-hemlock
= 2.1*8 x 10~3d ~1'19 Carbon (BPL)
R is the ratio of the quantity (C /C.) for mesh size 8-10 Tyler to that at
any given mesh size. [R = C /C.)Q , ,/(C /C.)]. This value is independent
o i o-lO mesh o i
of time except initially where conversion is approximately 100% for all mesh
sizes. The characteristic material dimension, d , in feet, was taken as the
average aperture opening for a given mesh size range. The above correlations
were determined at a superficial gas velocity of 200 ft/min, H S concentration
of 5 ppnij 60% RH, 75 F, and one inch bed depth for a mesh size range k-lk Tyler.
These correlations may be used in conjunction with Figure 1 for rough design
purposes. As expected, the rate of H S removal increases monotonically as the
particle size is decreased from h-6 to 10-ll* mesh. It should be noted that
even the largest sawdust tested (k-6 mesh) is a more efficient adsorbent than
the smallest carbon (lO-ll* mesh). For both the MnOp-sawdust and the carbon,
the H_S removal efficiency increases with decreasing particle size because
of the reduction of mass transfer resistance within the particles. For
the MnOp-sawdust adsorbents, there is also the additional factor that the smaller
-------
-15-
0.01
0.005
Co
Ci
0.001
0.0005
0.0001
200 ft/min., 1 inch bed
+ 400 ft./min., 2 inch bed
A 600 ft/min,, 3 inch bed
0123456
Time, hours
Figure 8. H2S outlet/inlet concentration ratio for constant resi-
dence times and varying flow rates
8-10 mesh, 5 ppm H2S, 60% relative humidity, 75°F
-------
-16-
particles probably contain a larger mass of MnO per unit bed volume since
the MnOp is found only in a layer near the outside of the sawdust particles,
8. Pressure Drop
We have measured the pressure drop across the bed of MnOp-sawdust
at velocities up to 300 ft/min for various mesh sizes. The pressure drop
Ap/L (inches H 0/ft bed) is given by
A£ = av"' a = .017 ^-6 mesh Tyler
L 8-10 mesh Tyler
= .05^ 10-llt mesh Tyler
where v is the superficial gas velocity in ft/min. The pressure drops are
similar to those across a bed of carbon (Adsorption Handbook, 1973).
9. Optimization
No attempt was made in this work to optimize the performance of
MnO -sawdust by varying preparative parameters. Preliminary investigations
have, however, indicated that considerable improvement in the performance of.
MnO -sawdust can be achieved by varying a number of preparational parameters.
These studies will be reported later.
10. Summary
The foregoing performance tests have clearly established the
superiority of MnO_-sawdust over some of the better activated carbons for
the absorption of H?S. However, while MnO -sawdust can be readily prepared,
it is not commercially available at present and must be discarded after use.
In practice, activated carbon used in odor control applications is rarely
regenerated. The possibility of regeneration is attractive and is being
investigated for MnO?-sawdust .
MnOp-sawdust has the disadvantage that it will ignite if allowed to
dry out completely over a period of days at a temperature above 100 C.
-------
-17-
Th e material has remained stable vhen operated for a period of weeks at
temperatures over 100 C in the presence of steam and even when ambient air
(TO°P, 55$ RH) is heated to 120°C and passed through the "bed. However, high
temperature (>100 C) operation should be approached with caution. The problem
can be overcome by presenting the active MnO on non-flammable substrates such
as pumice, vermiculite, perlite or coke although there is a marked decrease in
performance (Natusch and Sewell, 1970).
In attempting to make a general correlation for the behavior of the
MnO -sawdust in fixed bed applications for H S removal, one realizes that the
system is transient, that there are substantial mass transfer effects both
within and outside the particles and that the reaction kinetics are unknown.
The task is thus quite difficult and only part of the picture is available.
It is already clear that many physical and chemical properties of the MnO_
present on MnO -sawdust are quite different from those of any other MnO_ yet
reported (Bienstock and Field, I960; Korshunov and Vereshchagin, 1966). The
material appears to be most similar to 5-MnOQ.
The overall reaction of MnO.-sawdust with H0S can be written
Mn02 + 2H2S -> MnS + S + 2H20
indicating that one mole of MnO is capable of adsorbing two moles of H S. -
For the MnO_-s.awdust utilized herein a standard one-inch deep bed (diameter =
3 inches) contains ho g of MnO -sawdust which in turn contains approximately
11.8 g of MnO_. For this bed configuration calculations based on the results
given in Figure 3 show that only about half the total MnO_ reacts with H S
before the bed is exhausted. Additional capacity can, however, be obtained by
agitating the exhausted MnO -sawdust so as to expose fresh MnO?.
Work is presently in progress on the basic nature of the MnO -sawdust
in order to produce a more active filter material and on obtaining and correlating
further data necessary for design of filter beds under a variety of conditions.
-------
-18-
B. MERCAPTANS
N - butanethiol and methanethiol were investigated. Air containing
one to fifty ppm thiol was passed through the MnO~-sawdust for several hours.
Inlet and outlet gas analyses were made with a Tracer Model 550 chromatograph
with a flame photometric detector. As with H S, all runs were made at or
near room temperatures. In contrast to the situation with HpS, however, the
filter did not simply adsorb the mercaptan, but acted catalytically to convert
the n - butanethiol to a disulfide. The efficiency of n - butanethiol removal
was 90% to 99-9$ depending on contact time. All this mercaptan was converted
to the disulfide and no decrease in efficiency of the filter was noted with
time. Since the MnO_-sawdust is acting catalytically rather than as an absorbent,
the filter in its present form is inferior to activated carbon in removing
mercaptan from air. However, no attempt has yet been made to optimize the
filter material for mercaptan control and it is possible that further studies
could find a MnOp-sawdust preparation useful for mercaptans. In its present
form, the MnO -filter could be used along with activated carbon for simultaneous
removal of HJ5 and mercaptans.
When dry nitrogen was used in place of air to carry the mercaptans,
the filter acted as an adsorbent rather than catalytically and the efficiency
of the filter decreased with time.
C. AMINES
Preliminary studies were made on normal hexyl amine. The results,
to date, show that although some amine is removed, the MnO -sawdust filter
presently being used is inferior to carbon BPL. For example, at a concentration
of 3.5 ppm, a flow rate of 50 ft/min, a bed height of 6 inches, a relative
humidity of 60% and a temperature of 76 F, both MnOp-sawdust and carbon BPL
initially remove greater than 95% of the amine. However, the efficiency of the
MnO^-sawdust decreased to zero over a period of five hours whereas that of the
carbon remains essentially constant.
-------
-19-
IV. FUNDAMENTAL STUDIES.ON Mn02~SAWDUST
A large batch of MnO -sawdust (Preparation 20) was prepared in the
standard method described previously. Large quantities of MnO? shaken off of
the impregnated sawdust (20-A), MnO? decanted from the reaction solution (20-B),
and MnO formed by reaction between the 1% NaOH extract of the sawdust and
potassium permanganate (20-C) were also obtained.
An extensive series of tests for these materials was carried out.
When methods of preparation other than the standard method are used, the same
series of tests was carried out on the products. It was then possible to
correlate the activity of the MnO_ and MnO_-sawdust with several of the fundamental
properties and this will serve as a guide for optimization of filter performance.
A. Standard Preparational Procedure
1. Materials (l) Wood 200g sawdust
(II) NaOH Solution lOg NaOH in 1 liter de-ionized water
(III) KMnO, 200g Carus Chemical Co.
2. Procedure
(l) soaked in (II) for two hours. Soaking solution decanted and
saved. Sawdust washed once with 1 liter de-ionized water. Sawdust covered
with de-ionized water and heated to 80°C. Crushed III added and solution
stirred to the end of the reaction. All material flushed down the sides of
the reaction vessel into bulk solution. Mixture is allowed to set 15 minutes.
The liquid is decanted and saved. The impregnated sawdust is air-dried for
two days. After drying, the material is shaken loosely on a screen (20 mesh).
Material shaken off is sieved > 100 mesh (Tyler) and is labelled A sample
(MnO shake-off). The decanted liquid is vacuum filtered and washed 3x. The
filtered solid is air dried and labelled B sample (MnO? decant). The filtered
extract solution is reacted with 20g KMnO, at 80°C and filtered. The resultant
powder is called C sample (extract).
-------
-20-
B. Hemlock, KMnO, , NaOH Preparation
In previous work, MnO_ impregnated sawdust preparations prepared from
hemlock were found to give the greatest activity. Therefore, the first material
prepared was a batch using 8 x 10 mesh (Tyler) hemlock. The relative activities
toward ELS of the activated MnO -sawdust (Preparation 20)y the shakeoff MnO
powder (Preparation 20-A), decanted MnO? powder (Preparation 20-B), and the
extract MnO_ powder (Preparation 20-C) are shown in Figure 9- In order to
obtain measureable activities for 20A and 20B, it was necessary to dilute
the MnO powders with inert glass (170-230 mesh). A weight ratio glass; powder
of 3.5:1-0 was used in all experiments. The high pressure drops across the
reactor "bed did not allow the high gas velocities possible with the sawdust.
Complete data for Figure 9 are given in Table I.*
In general, the MnO -wood samples were run at about 5 ppm H S while
the powder/glass samples were run at about 100 ppm H S. The activity of the
powders was always in the order decant > shakeoff > extract.
In examining the various MnO -sawdust and MnO powder preparations a
series of chemical and physical properties was determined. The most important
of these and the method of determination were:
(l) BET Surface Area Orr Surface Area Analyzer
(2) Cone. K Atomic Emission Spectroscopy
(3) Cone. Na Atomic Emission Spectroscopy
(k) %C C02 Evolution
(5) $Mn EDTA Analysis
(6) Mn Valence State Active Oxygen Method
The data are tabulated in Table II.* Differences between the 20A, 20B, and 20C
powders may be attributed to surface area differences. Large amounts of potassium
are incorporated into the sawdust and powders. The relative amount of Mn in the
(IV) oxidation state is a significant indicator of the activity. If the total Mn
is determined independently by EDTA titration and if the assumption is made that
Tables I and II are in the Appendix.
-------
-21-
1.0
0.1
0.01
o
0.001
0.0001
0.00001
\
I I
Extract (20-C)
Shake-off (20-B)
Decant 20-A)
I I
0
23456
Hours
Figure 9 H^S outlet/inlet concentration vs. time for various Hemlock, KMnO^, NaOH
preparations.
-------
-22-
no Mn(ll) is present, then the relative amounts of Mn(lV) and Mh(lll) can be
determined by reduction of the "MnOp" in acidic iodide solution and titration of
I formed by standard thiosulfate solution. The ratio of Mn(lV)/Mn(lIl) is
usually about 3:1. The shakeoff samples are not homogeneous and thus a range
of composition is indicated for these samples in Table II.
We note that the shakeoff, although more closely related to the
impregnated savdust, is substantially less active than the decanted manganese
dioxide. If one c.ould characterize the decant and how it differs from the
shakeoff, it may be possible to improve the sawdust. X-ray diffraction patterns
at room temperature of both shakeoff and decant indicate no lines or a few
unidentifiable lines. Differential thermal analysis when coupled with X-ray
diffraction of the heated products has given some interesting data from which
the following thermal transformation scheme may be deduced:
Decant ^> 6 - MnO? ^ a - MnO? ^ Y - Mn 0
(20B) 200°C 200-550°C or 550-T50°C d
6 - MnO
< -i. \ » ^ 850-900°c
(HausmanniteJ Mn 0, s
The initial heating drives off water and causes a slight ordering
of the crystal structure. Hence 20B (decant) is most closely related to 6 - MnO .
The shakeoff material has not yet been characterized by the DTA-x-ray diffraction
combination.
C. Hemlock, NaMnO, , NaOH Preparation
In order to determine if the cation in the permanganate affects the
sawdust activity, NaMnO, was substituted for KMnO, in the preparation procedure.
The ionic radii of K and Na are 1.33A° and 0.9TA° respectively. Results are
shown in Figures 10, 11 for 8 x 10 mesh Hemlock wood. Run conditions for the
21 series materials are similar to the series 20 runs in Figure 9 (see Table I)
except in Figure 11 the superficial velocity for all materials was 35 ft/min.
The activity of the MnO -sawdust prepared from NaMnO, is much lower
than that prepared from KMnO, . The activity of the shakeoff from the NaMnO,
-------
-23-
1.0
0°
o.oi
Mn02/Sawdust (21)
(NaMn04)
0
/2/Sawdust (20)
(KMn04)
3 4
Hours
Figure 10 Effect of substituting NaMnO, for KMnO, in standard preparation - activity
vs. time for sawdust
-------
-24-
1.0
0.1
.- 0.01
O
0°
o.ooi
0.0001
0
Shake-off (21-A) (NaMnCU
Shake-off (20A)
Decant
(20-B)
(KMn04)
Decant (21-B)
(NaMn04)
3
Hours
Figure 11 Effect of substituting NaMnOi for KMnOi in standard preparation - activity
vs. time for powders
-------
-25-
preparation is also lower than the shakeoff from the KMnO, preparation. However,
the decanted MnO prepared from NaMnO, has a higher activity than that from
potassium permanganate.
Comparison of the properties of the sawdust prepared form KMnOi with
those of the sawdust prepared from NaMnO, indicates that the higher activity of
the former may be due to its much higher concentration of K . The same holds
for the shakeoff material. However, by this rationale the decant from KMnOi -
(20-B) would be more active than the corresponding 21-B. This is not true.
We hypothesize that K stabilizes the active form of MnO? during formation at
the sawdust-solution interface in such a way that more of the MnO remains
physically attached to the sawdust. Because the Na ion is smaller, it does
not so stabilize the active MnO? "lattice". Some data support this hypothesis
(less MnO_ sawdust formed from NaMnOi + sawdust, lower surface area) but some
do not (same % Mn in the product, approximately the same Mn(lV)/Mn(lIl) ratio).
The MnO_ powder not adhering to the sawdust in the NaMnOi-sawdust reaction
(decanted MnO ) is presumably more active because, in the absence of K ion,
a more amorphorus though less adhesive MnO_ sample is formed.
D. Hemlock, KMnO, , KOH Preparation
Since high concentrations of K may increase the sawdust activity,
a standard preparation of 8x10 mesh Hemlock was made using a 1% KOH soaking
solution instead of a NaOH solution. The results are shown in Figures 12 and 13.
Run conditions for the 2k series materials are the same as for the series 20 and 21
materials in Figures 10 and 11. The new sawdust preparation is inferior to the
standard preparation using NaOH. The decanted and shakeoff material prepared for
KOH have about the same activity as the corresponding powders prepared from NaOH.
Properties of these materials are shown in Table II. Substitution
of KOH for NaOH did not lead to appreciably higher K levels in the material.
The decrease in activity of the MnO? sawdust may be due to reduction of the
ratio of Mn(lV)/Mn(lIl); this ratio is about 1:1 in preparation 2\ (from KOH)
compared to about 3:1 in the standard NaOH preparation. Mn(lV) oxide is the
more active form so formation of relatively more Mn(lll) would reduce the
-------
-26-
1.0
0.1
o
0.01
Mn02/Sawdust (24)
(KOH)
o
Mn02/Sawdust (20)
(NaOH)
i
i
3 4
Hours
Figure 12 Effect of substituting KOH for NaOH in standard preparation - activity
vs. time for sawdust
-------
-27-
1.0
0.1
0.01
0.001
Shake-off (20A)
(NaOH)
Shake-off (24A)
(KOH)
^Decant (24B)
(KOH)
Decant (20B)
(NaOH)
0
3
Hours
Figure 13 Effect of substituting KOH for NaOH in standard preparation - activity
vs. time for powders
-------
-28-
MnO -sawdust activity. Since the MnO? powders (s:hakeoff and decant) do not
show this reduced Mn(lV)/Mn(lV) ratio, it is probable that KOH-treated sawdust
exhibits greater reducing properties toward permanganate than does' NaOH-treated
sawdust. . Thus, the cation used for base treatment is significant and this
treatment may affect the reducing properties of the sawdust as well as expanding
it physical structure.
E. Method of Addition of KMnO,
The manner of addition of the permanganate to the wood is not an
important experimental parameter. The standard preparation was modified in
that the KMnOi was added to the sawdust as an aqueous solution at 80 C. For
these experiments 8x10 mesh pine was used. Results are shown in Figure lk.
Run conditions same as for Preparation 20 in Table I. In addition, it is seen
that the pine is inferior to the hemlock. Several of the properties of the
pine/MnO are shown below.
Preparation
8 x 10 Pine
Std. Prep.
8 x 10 Pine
Std. Prep.
but KMnO^ addi-
tion in solutior
Run No
719
720
Prep. No.
22
23
Surface Area
m /g
U.87
6.73
%c
wt
25.96
27. 3U
%Mn
wt
16.09
17. 6U
F. Reproduction of Catalyst
To determine the reproducibility of catalyst manufacture several
batches of 6x8 mesh Hemlock were used to make activated sawdust by the standard
method. Results shown in Figure 15 indicate little difference between the
preparations. Run conditions were as in Table I for Preparation 20.
-------
-29-
1.0
O 0.1
o
O
0.01
MnCyPine (22)
(Solid KMn04 Addition)
Mn02/Pine (23)
(KMn04 Solution Addition)
o
Mn02/Hemlock (20)
Standard Prep.
(Solid KMn04 Addition)
234
Hours
Figure lU Effect of method of addition of KMnO, on standard preparation - activity
vs. time for sawdust
-------
-30-
1.0
0.1
o
O
0.01
0
,A Batch 1
(RLS-2)
Batch 2
(RLS-3)
3 4
Hours
Figure 15 Reproduction of standard preparation - activity vs. time for sawdust
-------
-31-
V. CONCLUSIONS
1. Both the efficiency and the capacity of a MnO /Sawdust Filter
are greatly superior to those of activated carbons for removal
of Hydrogen Sulfide from air.
2. For Amines and Mercaptans, activated carbons are better than the
form of MnOp/Savdust presently available. However, the MnO /Sawdust
has not yet been optimized for control of these odors.
For control of odors comprised of a number of pollutants, the
combined use of MnOp/Sawdust and activated carbon would be effective.
-------
-32-
'VI. RECOMMENDATIONS
The folloying objectives should be met:
(l) To optimize the efficiency and capacity of
MnO -sawdust for adsorbing E^S, mercaptans and
amines.
(2) To obtain the necessary data to enable prediction
of the performance of full scale Mn02-sawdust
filters in practical air pollution control
situations.
(3) To obtain data on the performance of MnOp-sawdust
filters for adsorbing noxious, odorous and
corrosive gaseous constituents from selected
effluent sources (Viscose, Kraft, Refinery and
Sewage effluents).
-------
-33-
VII. LIST OF REFERENCES
"Adsorption Handbook", Pittsburgh Activated Carbon Division, Calgon Corp.,
Pittsburgh, Pa. (1973).
Backlund, P., Fjellstrom, B. , HammerbHck, S., Maijgren, B., Arkiv fBr
Kemi, 26_, 267 (1966).
Bienstock, D. , Field, F. J., J. Air Poll. Control Assoc., 10, 121 (i960).
Kipling, J. J. , Quart. Rev., 10., 1 (1956).
Korshunov, S. P., .Vereshchagin, L. E. , Russ. Chem. Revs., 35, 9^2 (1966).
Natusch, D. F. S., Klonis, H. B., Axelrod, H. D., Teck, R. J. , Lodge, J. P.
Anal. Chem., ]iU_, 2067 (1972).
Natusch, D. F. S., Sewell, J. R., Proc. Second International Clean Air
Congress, Washington, D. C., Eds., H. M. Englund and W. T. Beery,
Academic Press, New York, 9^8 (1970).
-------
-34-
VIII. APPENDIX
A. Conversion Factors
multiply by
Gas Velocity ft/min .5080 cm/s
Temperature o,, 5/9(°F-32°) on
r 0
Bed Dimensions inches 2^.kO mm
2
Pressure mmHg 1332. dynes/cm
-------
-35-
bD
H
-P
cd
Q
0
o
CM
O
H
-p
03
!H
° o3
ft
0)
PH
PH
0
CM
O
H
-P
J>H
03
ft
0)
PH
^
o
CM
e
o
rl
-P
03
fn
03
ft
0)
£
O
CM
C
O
H
-P
o3
03
ft
OJ
PH
_=]
1 (
t-
_^J-
0
LPS
O
t
H
O
t
0
£j
^
PH
In
X S
00 W
H
ctf
H
£_i
OJ
-p
*>l
OJ
H
-P
O
0
H
. «
LP\
OO
0
H
. .
Lf\
ro
o
.
H
. .
LP\
PO
03
co
cd
H
O
fn
0)
|j
O
^-^^
W
CO
cd
H
O
O
H
-p
Ctf
K
^3-
_jj-
-d"
-j-
_^-
j-
O
O
CVJ
£j
H
e
-p
^>-J
-p
H
O
0
i i
0)
03
0?
O
LT\
OO
O
CM
H
t-
Ch
CM
ir\
's
ft
ft
a
0
H
[ )
03
^
p
C
0)
0
c
0
o
ro
CM
w
^5_
o
VD
"^5^
0
ts^v.
o
VD
1^5.
o
VD
* .
"fe^»
V ^
J>J
-p
H
H
tS
3
w
0)
J>
H
-p
03
H
0)
K
bO
ffi
g
g
CJ\
o
bD
tn
g
t
H
bO
K
g
^|
OO
H
O
CM
K
g
g
00
01
FQ
w
0
O
f^
ft
0
f_n
0)
£-J
Jj
w
to
CD
JLj
PH
t
CM
O
CM
f
H
.
t
OO
*
, ^
bD
V -
d
01
cq
fl
H
C
§
-P
j^j
p
0
5,
ON
t
d~
t
VO
t
O
CO
[~Tt
O
0)
pi
-p
03
^
0)
ft
R
CJ
^-(
^_,
H
^
S
S
OJ
H
p.
H
CM
H
g
g
CVJ
H
^
H
40
x;
bO
H
0)
JT~]
qj
0)
w
Q
g
vo
H
^
g
VD
H
^3
s
VD
H
£3
R
O
CO
^
^
g
s
£H
0)
-P
0)
J3
CO
H
Q
d
0)
FP
-------
-36-
to
0)
H
-P
0)
o
£
S is
"65.
^.
W -p
M
aoj
03 a
!H 03
P 0)
CO to
ffi
0)
H
ft O
fi ^
03
CO
O
H
-P
03
I
0)
PH
0)
CO
'-» H
> M
H H
"!a"!fi
s s
65. VS.
OJ OO
00 H
1 t
t- OO
-3- LA
t OJ
___,.
OO
oo
H
oo
o
o
oo
o
CD
OO
o
ON
LA
LA
^j-
H
0
t-
0
OJ
-P
£j
>
o3
CO
a
o
H
-P
n3
to
03 O
(L)
£H W
PH
co
T3
^t r^l
03 O
^ 0
a -a
-p 0)
CO W
M
H
5" ^
S
65. t
VQ OO
ON
I I
ON H
M
Hi i
li
rt ^
S S
Vt Vi.
oo-t
t OJ
t-
t-
0
VQ
o .
Oj"
£
vo
o
^
oo
H
LA
-d-
LA
t
t
H
t-
pq
o
OJ
40
rt
o
0)
o
:
'-* H
> M
M H
~J3 ^
§ (§
H ON
VQ OO
OO H
LA
CO
[J.
OO
H
H
OJ
OJ
*
0
H
H
CO
LA
--j-
H
t
O
O
OJ
_p
o
-p
- ~ H
> M
Hi i
i i
Jd ^S
S S
Vi VI
_*} \D
I OJ
H
t
LA
i-l
0")
O
ON
OJ
-d-
vo
t
oo
o
[-
t
oo
t--
o
C
H
OJ
p
CQ
^
'S
03
CO
a
0
H
P ^
a3 C
to 1-
O cd
H ft -
0) C
£ 1
oo a
T3 ^
O o5 +
O ^ f
"i i 2
(u p >
M co a
^-~- H
> H
H M
j3 Ja
>§ S
0 &«.
j- vo
H 00
t OJ
__,. '
H
OO
OO
00
LA
_^
ON
OJ
1-1
LTN
O
o
ii -^t
J O
^ f^t
) S
'-^ H
> H
H H
J3 £
g ^
OO t
oovo
CO H
LA
H
ON
oo
0
H
OJ*
ON
ON
ON
OO
O
CO
-
ON
LA
H
H
t
PQ
H
OJ
p
O
0)
P
"^"i-H
> M
H H
J3 £
§ j§
OJ OO
CO rH
CO H
LA
OO
0
H
OJ
,H
t-
O
H
OO
o
03
p
O o3
iz, Q
03
P
O o3
& o
O
H
OJ
P
O
oi
P
W
'-^ M
> M
M M
£ G
!§ ^
OO OJ
Lf»_d-
-d- LA
ON
oo
VD
H
ON
OJ
ON
OJ
OJ
0
o
OJ
-d-
CO
f
OJ
.
^J-
OJ
OJ
t
"^
aj
-P
m
p
d
03
CO
0
H
p
03
to
O 03 f-
H ft C
X to
PH H
00 C
V] y_j
O 03 +
o -d P
H a a
B o3 t
OJ p >
w co a
-~« H
> H
M H
rt. rt
!>| ^
H ON
0 ON
t OJ
OJ
LA
OO
H
OO
LA
H
0
ON
t--
H
rH
H
H
LA
OJ
t
_^J-
OJ
c^
O
(L)
AJ
03
H
)
H
H
^
1|
JS
H
H H
J3 ^
|g| ^
o fc«
oo t--
J- LA
t OJ
t_
H
OO
-*
^
VO
H
oo
_-j-
o
d
OJ
H
H
CO
.
_-f-
ON
_j-
OJ
t-
pq
OJ
P
a
a3
0
&
' I-l
> H
H H
rt c
S r^l
C,
0 0
H ON
ON
vo
OO
OJ
-^
VD
LA
O
03
p
^
03
-p
O 03
S3 P
03
p
K Q
o3
p
O 03
s o
O
OJ
-P
0
03
to
P>
X
-------
-37-
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-650/2-74-063
3. RECIPIENT'S ACCESSION-NO.
4. TITtE AND SUBTITLE '
Adsorption of Odorous Pollutants by Active
Manganese Dioxide
5. REPORT DATE
June 1974
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
D. F.S.Natusch, J.L.Hudson,
R.L.Solomon, R. Tanner, and A. Miguel
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORG \NIZATION NAME AND ADDRESS
School of Chemical Sciences
University of Illinois at Urbana-Champaign
Urbana, Illinois 61801
10. PROORAM ELEMENT NO.
1AB015; ROAP 21AFB-010
11. CONTRACT/GRANT NO.
Grant R-801603
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
NERC-RTP, Control Systems Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; Through June 1973
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
repQrjh gjves results of an investigation of the adsorption of odorous
pollutants by active manganese dioxide (MnO2). Hydrogen sulfide (H2S) , mercaptans ,
and an amine were adsorbed in a fixed bed of active MnO2 (impregnated on sawdust)
and several activated carbons. Pollutant removal efficiencies were measured as
functions of H2S concentration, flow rate, particle size, relative humidity (RH), and
bed depth. The outlet concentration of H2S from the MnO2-sawdust bed is consider-
ably below that from an activated charcoal bed of the same volume and at the same
conditions , often by more than an order of magnitude. The useful capacity of the
MnO2-sawdust bed can be greater than 4 times that of activated carbon. The effic-
iency of MnO2-sawdust is maximum at 80% RH. Both MnO2-sawdust and activated
carbon are inefficient at below 20% RH. The MnO2 -sawdust bed acts catalytically
when mercaptans and air are passed through it at room temperature: the disulfide
is formed and there is no decrease in bed efficiency with time. Amines are adsorbed
by the filter at an efficiency less than with activated carbon. Fundamental studies
of the MnO2 -sawdust were carried out and preliminary correlations were made bet-
ween bed efficiency and capacity and filter characteristics such as surface area,
percent Mn, K, Na, and C, and the Mn valence st?.te.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Air Pollution
Odor Control
Deodorizing
Manganese Oxides
Adsorption
Hydrogen Sulfide
Thiols
Amines
Activated Carbon
b.IDENTIFIERS/OPEN ENDED TERMS
Air Pollution Control
Stationary Sources
Mercaptans
Activated Manganese
Dioxide
c. COSATI Field/Group
13B, 07C
05E
07A, 11G
07B
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report}
Unclassified
21. NO. OF PAGES
40
Unlimited
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |