REACTIONS OF REDUCED SULFUR COMPOUNDS WITH OZONE
/MICHAEL LARRY TUGGLE
DISSERTATION PRESENTED TO THIE GRADUATE COUNiCIL OF
THE UNIVERSiTY OF FLORIDA IN PARTIAL
FULFILLMENT OF THE RLQUIREM.E1TS FOR TiiE DEGREE OF
DOCTOR OF PHILOSCrHY
UNIVERSITY OF FLORIDA
The author wishes to acknowledge and express his appreciation and
thanks to Dr. R. S. Sholtes, his committee chairman, and to the other
members of the committee -- Dr. C. C. Curtis, Prof. T. DeS. Furman, Dr.
J. E. Singley, and Mr. Herbert Berger -- for their assistance in many
areas associated with the project, and for their patience and understanding
throughout the course of the research program.
He is also indebted to Mr. C. H. Wilson of the National Council
for Air and Stream Improvement for his assistance in assembling the test
stand used in the study, to Mr. H. R. King and Mr. E. C. Logsdon for
fabricating the reaction tube components, to Mrs. Irene Lehman for typing
the correspondence and providing other secretarial services required
during the study, and to Mrs. A. L. Hammond for typing this manuscript.
The author wishes to particularly thank Mr. Thomas Tucker, formerly
with the National Council for Air and Stream Improvement, for his assistance
with the laboratory and field experimentation and analysis during the
The author expresses his appreciation to the National Council for
Air and Stream improvement for financing his graduate studies, and
providing the facilities and equipment needed for the research project.
He also wishes to thank the faculty1 and staff of the Environrnental
Engineering Department for their assistance in courses, laboratories,
and other related areas during the research period.
And finally, the author would like to again express his appreciation
to his wife for her continued understanding and assistance throughout
the graduate program covering the past three years.
TABLE OF CONTENTS
LIST OF TABLES.................................................. vii
LIST OF FIGURES....................................... ........ ... x
I. SCOPE OF RESEARCH PROJECT ................................
Research Objective .................................... 2
Research Outline..................................... 3
!1. ODORS AND THE SULFATE PROCESS............................ 5
The Odor Problem........................................ 5
Basis of Problem................................ 5
Effects of Odorous Emissions.................... 8
Source and Quantity of Odorous Emissions........ 9
Industrial Odor Control................................ 11
Process Modifications or Substitution........... 11
Absorption................................... ... 14
Oxidation Systems............................... 16
Combustion Oxidation........................... 16
Chemical Oxidation............................ .. 17
Odor Masking. ................................... 20
Odor Counteract on............ ..... ............... 21
Odor Control in the Kraft Pulping Industry............ 23
Black Liquor Oxidation............... ............ 23
Elimination of Direct Contact Evaporation...... 25
Proper Operation, Maintenance, and Housekeeping. 26
Combustion Oxidation............................ 27
Chemical Oxidation................................ 28
Absorption............................. ...... 30
Oder ;.asking Iand Countracion. .................. 32
Condensation .................................... 33
TABLE OF CONTENTS (Continued)
Ill. OZONE CHEMISTRY AND APPLICATIONS.......................... 36
Characteristics, Formation, and Toxicity.............. 36
Characteristics ................................ 36
Ozone Formation................................. 38
Ozone Toxicity.................................. 42
Ozone Appl ications .................................... 47
Waste Treatment ....................... ...... . 47
Odor Control.................................... 48
Water Treatment............................... 49
Ozone for Kraft Odor Control............ ... .......... 51
Previous Applications and Investigations........ 51
Basis of Present Study .......................... 60
IV. EXPERIMENTAL APPARATUS.................................. 63
Description of Test Stand............................. 63
Basic Flow Pattern.............................. 63
System Details.................................. 66.
NCASI Mobile Laboratory Equipment ..................... 71
General Description of Mobile Laboratory........ 7!
Electrolytic Titrator and Combustion Furnace.... 72
Gas Conditioning Oven........................... 73
Gas-Liquid Chronatography System................ 76
V. EXPERIMENTS AND RESULTS WITH BOTTLED SULFUR GASES......... 78
Experimental Procedures............................... 78
Hydrogen Sulfide...................................... 80
Methyl ilercaptan......................... ............. 85
Dimethyl Sulfide...................................... 88
Dimethyl Disuifide.................................... 93
VI. EXPERIMENTS AND RESULTS WITH KRAFT MILL EFFLUENT GASES.... 96
System Modifications.................................. 96
Efflucnt Cas Source................................... 98
Experimental Procedures............................... 98
Results of Ozcne Oxidation............................ 99
Th.-rmal Dcc:y of Ozcn .............................
Experimental Error.................................... 106
VII. SUI-iARY AND CONCLUSIONS................................... 10..
A. OZC0 0 ANALYSIS, E "'',,-T CALRATION, AND TEST STAND
OPERATING PROCEDURES ............................. ....... 118
TABLE OF CONTENTS (Continued)
APPENDICES (Continued) Page
B. EXPERIMENTAL DATA.................................... .... 1:32
Bi BL I OGRAPHY ................................................... 173
LIST OF TABLES
1. THRESHOLD ODOR VALUES ........................................ 7
2.. SUMMARY OF KRAFT EMISSION DATA............................. 10
3. BIOLOGICAL EFFECTS OF OZONE................................. 45
4. 070NE CONCENTRATIONS FOR ODOR CONTROL....................... 49
5. REACTION DATA OF CADLE AND LEDFORD FOR THE GAS PHASE
REACTION OF OZONE AND HYDROGEN SULFIDE................. 57
6. GLC SAMPLING LOOPS ........ ......... .......... .... ....... ... 74
7. CAS CHROrMTOGRAPH COLUMN ELUTION TIMES...................... 76
8. TOTAL REACTION TIME FOR OZONE OXIDATION ..................... 79
9. HYDROGEN SULFIDE EXPERIMENTAL CONDITIONS.................... 81
10. METHYL MERCAPTAN EXPERIMENTAL CONDITIONS.................... 85
11. DIMETHYL SULFIDE EXPERIMENTAL CONDITIONS.................... 90
12. DIMETHYL DISULFIDE EXPERIMENTAL CONDITIONS.................. 93
13. SUMMARY OF OZONE-SMELT TANK VENT GAS DATA................... 101
14. SUMMARY OF REGRESSION ANALYSIS DATA FOR BOTTLED SULFUR
GASES .................................................. 112
15. RtACTION TU8E SPACE TIME.................................... 126
16. REACTION TUBE TEMPERATURE CONTROL........................... 127
17. SULFUR GAS TANK ANALYSIS ............................ .............. .. 127
18. OZONE-HYDROGEN SULFIDE REACTION DATA AT 330 C............... 133
19. OZONE-HYDROGEN SULFIDE REACTION DATA AT 550 C............... 134
20. OZONE-HYDROGEN SULFIDE REACTION DATA AT 760 C.............. 135.
LIST OF TABLES (Continued)
21. OZONE-HYDROGEN SULFIDE REACTION DATA AT 1250 C............
22. DATA FOR REGRESSION ANALYSIS OF OZONE-HYDROGEN SULFIDE
REACT I ONS............................................
OZONE-HYDROGEN SULFIDE REACTION
OZONE-HYDROGEN SULFIDE REACTION
OZONE-HYDROGEN SULFIDE REACTION
OZONE-METHYL MERCAPTAN REACTION
OZONE-METHYL MERCAPTAN REACTION
DATA FOR REGRESSION ANALYSIS OF
OZONE-METHYL MERCAPTAN REACTION
OZONE-METHYL MERCAPTAN REACTION
OZONE-METHYL MERCAPTAN REACTION
OZONE-DIMETHYL SULFIDE REACTION
DATA AT 380 C.............
DATA AT 1250 C............
DATA AT 380 C.............
OZONE,-DIMETHYL SULFIDE REACTION DATA AT 1250 C............
DATA FOR REGRESSION ANALYSIS OF OZONE-DIMETHYL SULFIDE
OZONE-DIMETHYL SULFIDE REACTION EQUATION 5.7..............
OZONE-DIMETHYL SULFIDE REACTION EQUATION 5.8..............
070NE-DIMETHYL SULFIDE REACTION EQUATION 5.9.............
OZONE-DIMETHYL OISULFIDE REACTION DATA AT 380 C...........
OZONE-DIMETHYL DISULFIDE REACTION DATA AT 1250 C..........
DATA FOR REGRESSION ANALYSIS OF OZONE-DIMETHYL DISULFIDE
OZONE-DINETHYL DISULFIDE REACTION EQUATION 5.10............
OZONE-DIMFTHYL DISULFIDE REACTION EQUATION 5.11...........
LIST OF TABLES (Continued)
43. OZONE-DIMETHYL DISULFIDE REACTION EQUATION 5.12............. 169
44. OZONE-SMELT TANK VENT GAS REACTION DATA.................... 170
45. THERMAL DECAY OF OZONE IN REACTION TUBE.................... 172
LIST OF FIGURES
1. Conversion of Initial H2S to SO2 and H2SO4 (Gregor and
Martin) ............................................. 55
2. Conversion of Oxidized H2S to SO2 and H2SO4 (Gregor
and Martin) ......... ..... .......................... 56
3. Front View of the Test Panel.............................. 64
4. Rear View of Test Panel................................... 65
5. System Schematic for Oxidation of Bottled Gases............ 67
6. Injection of Sulfur Gases with Spinning Syringe............ 69
7. Diagram of Reaction Tube .................................. 70
8. Mobile Laboratory Flow Diagram............................ 75
9. Ratio and Time Effects on H2S Oxidation ................... 83
10. Temperature and Time Effects on H2S Oxidation............. 83
11. Ratio and Time Effects on RSH Oxidation.................. 89
12. Temperature and Time Effects on RSH Oxidation............. 89
13. Temperature and Time Effects on RSR Oxidation............. 92
14. Ratio and Temperature Effects on RSR Oxidation..,......... 92
15. Temperature and Time Effects on RSSR Oxidation ............ 95
16. Ratio and Temperature Effects on RSSR Oxidation............ 95
17. System Schematic for Oxidation of Stack Gases.............. 97
18. Ozone Oxidation of Smelt Tank Vent Gas................... 102
T9. Thermal Decay of Ozone in Reaction Tube ................... 105
LIST OF FIGURES (Continued)
20. Ratio Effects on Oxidation of Reduced Sulfur Compounds.....
21. Ozone Generator Calibration Curve .........................
22. Spinning Syringe Calibration Device .................. .....
Abstract of Dissertation Presented to the
Graduate Council of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of Doctor of Philosophy
REACTIONS OF REDUCED SULFUR COMPOUNDS WITH OZONE
Michael Larry Tuggle
Chairman: Dr. Robert S. Sholtes
Major Department: Environmental Engineering
Due to increasing emphasis on the elimination of industrial odors,
odor control is a major concern of the kraft pulping industry, and
continuous research is needed to evaluate any feasible method that will
aid in eliminating this problem. Ozone is one of the most powerful
oxidants known, and has been suggested for kraft odor control by gas-
phase oxidation of the sulfur compounds. The application of ozone in
the kraft industry is near non-existent, and laboratory investigations
are very limited and concern primarily only oxidation of hydrogen
The purpose of the present study was to evaluate the effectiveness
of ozone ;n gas.-phase oxidation of the four malodorous sulfur compounds
primarily responsible for the kraft odor problem; namely, hydrogen
sulfide, methyl mercaptan, dimethyl sulfide, and dimethyl disulfide.
The initial phase of the study involved the use of bottled gases, whereas
the last phase evaluated the ozone oxidation of an actual kraft mill
source -- smelt tank vent gas.
Experimental data were obtained with the use of a heated, stainless
steel reaction tube. Six sampling ports spaced dovn the length of the
tube provided reaction times ranging from approximately 5 to 60 seconds.
For the bottled gases, ozone to sulfur gas ratios of 0.5 to 8.0 were
evaluated at reaction tube temperatures of 380 C to 1250 C. The
effectiveness of ozone oxidation was noted by monitoring the decrease
in initial sulfur gas concentrations at various conditions of reaction
time, temperature, and concentration ratios. Multiple regression
analysis was used to evaluate the data points obtained for each gas,
and provide several regression equations representative of the oxidation
process for the conditions studied.
Results of the data on bottled gases indicated that concentration
ratio and reaction temperature had a much more significant effect on
oxidation than did increased reaction time, and to achieve complete
oxidation of the malodorous sulfur compounds, ozone concentrations of
more than twice the stoichiometric amount would be required for the
reaction times and temperatures studied. The percent oxidation
achieved in this study for hydrogen sulfide was significantly higher
than that reported by other authors, and attributed primarily to the
difference in experimental conditions and procedures between the
For experiments with actual kraft emissions, total reduced
sulfur (TRS) of the smelt tank vent gas stream was continuously
monitored with cn cloctrolytic Litrator. Ozone was effective in
eliminating all of the methyl mercaptan, dimethyl sulfide, and dimethyl
disulfide from this source at an ozone to TRS ratio of approximately
2.50. These three sulfur compounds accounted for about 52 percent
of the TRS as determined by gas-liquid chromatography. In view of the
oxidatio' results :.it[l, otil(ed gases, the concentration ratio of 2.5
for a stack source was considered acceptable, recognizing the vast
difference in a stack effluent as opposed to a bottled gas source.
Additional studies are needed to further evaluate the ozone
oxidation of various kraft mill sources, covering a wide range of stack
conditions and reaction times. The current study indicated that complete
oxidation of the malodorous compounds would only be achieved with
reaction times far exceeding those normally experienced in an emission
stack, thus requiring the possible need of holding chambers to achieve
adequate mixing. Costs are also of major concern, both for ozone
production and additional facilities for increased mixing time, but
were not evaluated as part of this study.
SCOPE OF RESEARCH PROJECT
Ozone is one of the most powerful oxidants known, and is finding
increasing use in the field of odor control. Due to its high electro-
negative potential (second only to fluorine) the potential applications
for ozone as an oxidant are considerable, but have been somewhat limited
in the past due to the practical requirements that it be used in its
gaseous state, the low concentrations of ozone to parent gas (air or
oxygen) delivered by ozone generators, and the cost of ozone in pounds/
kilcowatthours. Improvements in ozone generator design and efficiency,
the availability of tonnage oxygen at decreased costs, and the emergence
of cylinder ozone have expanded the economical applications of this
oxidant for various commercial uses.
Gas-phase oxidation of malodorous compounds with ozone for industrial
odor control is mentioned by Yocoin and Duffee (1) and Turk (2: 3) and is
used today for controlling odors from sewage treatment plants, lift
stations, and various commercial operations (kitchens, rendering plants,
food processing, rubber compounding mill, chemical exhausts, etc.). Its
use specifically for controlling malodorous sulfur emissions from sulfate
pulping operations has been suggested by Sarkanen, et al. (4), Douglass (5),
and Cromwel (6) and a patent has been issued for the process- (7), but
application of this method has been very limited. The article by
Crormwell (6) describing the experimental use of ozone to treat effluent
gases from a sulfate recovery system at a pulp and paper mill in Canada
was the only reference found on actual mill effluent application.
Ricca (8) reported on the use of ozone to oxidize exhaust gases from a
bench-scale, black liquor oxidation system, and Akamatsu (9) studied the
use of ozone in deodorizing kraft mill blow gases produced from an
autoclave. Laboratory studies on the reaction of ozone with hydrogen
sulfide (from cylinders) have been reported by Gregor and Martin (10),
Cadle and Ledford (11), Hales (12) and Finlayson, et al. (13).
Due to the decreased costs in tonnage oxygen over the past fe, years,
improved ozone generation, the possibility of on-site production of
oxygen for waste treatment systems and oxygen bleaching, and the increasing
emphasis and regulations directed toward eliminating industrial odors,
additional dataare needed on the feasibility of using ozone for odor
control in the sulfate pulping industry.
The objective of this study was to evaluate the effectiveness of
ozone in gas-phase oxidation of the four major components of kraft mill
odors; namely, hydrogen sulfide (H2S), methyl mercaptan (RSH), dimethyl
sulfide (RSR), and dimethyl disulfide (RSSR). The study involved mixing
ozone with each of the individual gases (from cylinder sources) at a variety
of concentration ratios, temperatures, and reaction times, followed by
analysis of the reaction products. Tests were also performed to evaluate
the effectiveness of various ozone concentrations (and reaction times) on
actual emissions from a smelt dissolving tank by analyzing the effluent
gas prior to and follaoing ozone injection. The range of experimental
conditions (sulfur concentrations, reaction temperatures, etc) for the
study included those normally found in the kraft pulping industry.
Research Outli i
The research was accomplished according to the follaozing outline:
(1) Revie.- the air pollution problem facing the pulp and paper
industry in the elimination of odorous emissions from sulfate
pulping operations, the basic approaches to industrial odor control,
and the methods currently used by the industry for control of
(2) Review the general characteristics and applications of ozone,
such as its physical and chemical properties, methods of generation,
toxicity, reactions with various compounds, and uses in the treat-
ment of dcrnestic and industrial waste, drinking water purification,
(3) Pei form a literature survey For previous studies and applications
of ozone for odor control in the pulp and paper industry, and in
other industries or operations with odor problems (sewage treatment
plants, food processing, rendering plants, etc.).
(1i) Select and obtain ozone generating equipment and other experi-
mental apparatus required for studying ozone reactions with bottled
sulfur gases and kraft mill effluent gases, both at various experi-
(5) Perform experiments designed to evaluate the effectiveness of
ozone in oxidizing each of the four odorous sulfur gases (from
cylinder sources) at various concentration ratios, reaction
temperatures, and reaction times.
(6) Perform experiments designed to evaluate the effectiveness of
ozone at various concentrations and reaction times in oxidizing sulfur
compounds in an actual kraft mill effluent gas.
ODORS AND THE SULFATE PROCESS
The Odor Problem
Basis of Problem
Since the development of the sulfate process by Dahl in 1879 as a
method of making pulp and paper from wood, the operation of the first
sulfate (kraft) mill in 1891, and the appearance in the literature of
the first papers concerning kraft mill odor around 1900, there has been
little doubt that odor control is a major problem of the kraft pulping
Recent evidence of this continuing problem is the number of papers
and reviews published in the last few years concerning atmospheric
emissions from kraft pulping operations, with the major portion concerning
odorous emissions. Literature published prior to 1963 on air pollution
and the kraft pulping industry is reviewed in an annotated bibliography
by Kenline and Hales (14). Hendrickson (15) edited the proceedings of an
international conference on sulfate pulping emissions held in 1964 at
Sanibel Island, Florida. Environmental Engineering, Inc. and J. E.
Sirrine Co. published a three-volume study in 1970 for the National Air
Pollution Control Administration (16) on control of atmospheric emissions
in the wood pulping industry, and a 1970 conference in Stockholm (17) on
methods for measuring and evaluating odors had several presentations
on odors from the kraft pulping industry. Annual reviews (18-23) of pub-
lished literature relative to air pollution control in the pulp and paper
industry are prepared by the National Council of the Paper Industry for
Air and Stream Improvement. Other excellent reviews are by Sarkanen,
et al. (4), Adams (24), and Wright (25).
The major components responsible for this characteristic kraft odor
are all sulfur compounds (16, 26-28); namely, hydrogen sulfide (H S),
methyl mercaptan (RSH), dimethyl sulfide (RSR) and dimethyl disulfide
(RSSR), and are referred to as reduced sulfur compounds. These four
gases were first identified by Klason during a series of studies from
1908-1924 (4), which is another indication of hoq long the kraft odor
problem has been under investigation and hao difficult it is to eliminate.
Although some of the odor-producing components of the effluent may be
associated with non-sulfurous compounds (29), most attempts to control
odors are directed to the four sulfur gases mentioned above.
The elimination of these odors is difficult primarily because of
their extremely lo. odor threshold, or minimum detectable concentration.
Table I is a summary of odor threshold values reported in the literature
for these four gases of interest. The values are for the pure compound,
not for the gas in combination with other gases (sulfur and non-sulfur)
as would be under actual conditions. Odor thresholds for a combination
of these gases may be completely different than those reported, but would
not be a simple addition of the threshold values of each component (37).
Due to different tests used to determine threshold values, data treatment,
purity of the compounds studies, and interpretations of the word
"threshold," a range of values are reported for each compound, but
THRESHOLD ODOR VALUES
Compound Threshold Odor (ppb)a ... Reference
Hydrogen Sulfide 4.1 30
Methyl Mercaptan 1.0 30
Dimethyl Sulfide 2.0 30
Dimethyl Disulfide, 5.6 30
500.0 .. 34-
Parts per billion by volume.
serve to indicate that the odor threshold of each gas
In addition to the problem of lo.w odor threshold values, the fact
that kraft odors are released from so many different points throughout a
mill is another obstacle in odor control for th' industry. To reduce
ambient concentrations of these odorous emissions to threshold values in
mill localities, each source will have to be investigated and controlled.
Although the same basic process is used in each mill and the emission
sources are similar, each mill presents its own particular set of problems,
and what may be effective at one installation may not be at another with-
Effects of Odorous Emissions
Why are odors objectionable, and why all the concern about them?
Odors are intangible, but may be generally described by their odor
threshold, intensity of levels above this threshold, acceptability, and
their characteristic properties which distinguish them from other odors
regardless of intensity or acceptability (38). For air pollution studies,
the most important dimension of an odor is its acceptability, or how
many people are inconvenienced by the smell and to what extent. Numerous
articles (17, 31, 32, 39-48) are found in the literature describing the
evaluation of odors for their objectionability, measurement by test
panels and community observer corps, description of nuisance effects,
differences in human response to odors, measurement by chemical analysis,
and needs for future research in the odor field. Since a qi.antative measure
of inconvenience, objectionability, or acceptability is sometimes hard to
determine, the odor threshold value is normally used for evaluation.
As shown previously, even this value is subject to large variations.
Odors in themselves have not been shown to be the direct cause of
any disease, but various ways in which they may affect a person's well-
being are discussed by McCord and Witheridge (49) and Stockman and
Anderson (48). These include reduced appetite, nausea, headaches,
interference with sleep, and mental stress. Economic effects of odors,
such as the lowering of property values, reduced sales and rentals in
resort areas, and reduced desirability of a community in which to live
are also mentioned. In sufficient concentrations, sulfur emissions
common with kraft mills (particularly hydrogen sulfide) are highly toxic
and can be fatal (50), but health hazards resulting from the presence of
these gases in the atmosphere around sulfate pulping operations are not
considered to be of any concern due to their diluted concentrations (51,
52). Even though not considered a health hazard, discharges from kraft
mills have reportedly caused nausea, headaches, eye irritation, and upper
respiratory irritation (48). Other detrimental effects attributed to
sulfur emissions are corrosion (35, 48, 53) and darkening of paint by
hydrogen sulfide (48, 52, 54).
Regardless of the various economic and health effects that have been
attributed to odorous sulfur compounds, their nuisance aspect is of primary
concern. This is evidenced by the volumes of literature previously
mentioned concerning odors and the sulfate pulping industry, and leaves
no doubt that the odors are an annoyance to the public, and therefore
Source and Quantity of Odorous Emissions
The various process reactions and factors responsible for kraft odor
generation are beyond the scope of this study, but are well described in
the literature (4, 16, 26, 27, 51, 55-59). As previously mentioned,
odorous gases are generated and released at several points in a kraft
mill complex, and at varying temperatures and concentrations. Table 2
is a summary of data from the study for NAPCA (16) listing major sources
of odor release in a kraft mill, and ranges of temperature, gas flows, and
odorous emissions at each source. Similar data are also listed by Hough
and Gross (60), Adams (61), and Sarkanen, et al. (4). Considering the
amount of total reduced sulfur cQopounds that may be generated at various
points in a 500-1000 ton per day kraft mill complex, the low odor
threshold for each of these compounds (Table 1), and increasing emphasis
and regulations (62) directed toward eliminating these emissions, the
magnitude of the kraft pulping industry's odor control problem becomes
SUMMARY OF KRAFT EMISSION DATA
Temp. Gas Flow Range of Sulfur Gas Concentrations (Ib/ADT)b
Source (oC) (CF/ADT)a H2S RSH RSR RSSR
Digester Relief 49 35 0-.01 .01-1.00 .10-.40 .10-.20
Digester Blow 49 300 .10-.12 .40-.47 1.20-1.40 1.30-1.50
Brown Stock Washers 52 80,000 ..01-.12 .10-.25 .01-.02 .01-.02
Oxidation Towers 57 13,000 .01-.02 .05-.10 .02-.08 .05-.15
Multieffect Evap.-Ox. 49 35 .01-.02 .10-.30 .05-.15 .05-.15
Multieffect Evap.-Unox. 55 35 .10-3.00 .10-1.50 .05-.08 .01-.02
Recovery Furnace & DCE-Ox. 163 400,000 1.00-5.00 .01-.25 .01-.10 .01-.20
Recovery Furnace & DCE-Unox. 163 400,000 5.00-30.00 .50-2.50 .10-.30 .10-.40
Smelt Tank 88 44,500 .02-.05 .02-.05 .01-.02 0-.01
Lime Kiln 82 50,000 .20-.52 .10-.50 .01-.28 0-.08
Cubic feet per air dried ton of unbleached pulp.
Pounds per air dried ton of unbleached pulp.
Industrial Odor Control
The control of odorous gases from industrial sources can normally
be accomplished with one or more of the following methods (1-3, 63):
(1) Reducing or eliminating the initial formation of odorous
compounds through process modifications or use of a completely
(2) Use of various techniques, such as adsorption, absorption,
condensation, combustion and chemical oxidation, to remove or
reduce odorous gases in the effluent stream prior to their emmission
into the atmosphere,
(3) dispersion of the odorous gases (usually through tall stacks)
to a greater extent so that they are less concentrated upon
reaching any point where they may be detected,
(4) Addition of other odorant gases to the effluent stream so that
the resultant odor becomes less objectionable through either odor
masking or odor counteraction.
The use of any of the above methods, or their combinations, depends
upon several factors, such as volume of the effluent stream, characteristics
and concentrations of particulates and odorous gases in the effluent,
location of the particular industry, normal atmospheric conditions, and
the chemical reactions taking place in the process. The folloa.ing sections
are brief descriptions of each method mentioned above.
Process Modifications or Substitution
Reduction, or elimination, of the initial formation of odorous gases
through process modifications or the use of a completely different process
would be a major step in any odor control program, and,in many cases,
could be more effective and cheaper than complete abatement procedures
at the stack. Just by reducing the initial odor formation, less load
would then be imposed on any control equipment downstream.
Odor control through process changes may include an adjustment of
process temperature, residence time, or pressure in various operations,
substitution of lai odor (or odorless) solvents or reactants for highly
odorous ones, a change in flo.w diagrams to prevent contact of odor
producing gases and/or liquids, and proper maintenance and housekeeping
Odor control by adsorption results from an interaction between the
malodorous gas and a solid due to the phenomena of surface attractions
universal with all substances. The forces that hold atoms, molecules,
and ions in the solid state exist throughout the solid and also at its
surface, and thus are available for binding other molecules which contact
the surface or come in close proximity to it. As a result of these
forces, any gas, vapor, or liquid will adhere to any solid surface to some
degree (64). This phenomenon is called adsorption (or sorption), the
adsorbing solid is called the adsorbent (or sorbent), and the adsorbed
material is the adsorbate (or sorbate). This process is useful in air
pollution control since it provides a means of concentrating malodorous
emissions for ultimate disposal or recovery.
Depending upon the adsorbent, the gases of concern (or adsorbate),
temperature, etc., various adsorption processes are available. In
chemical adsorption chemisorptionn), chemical bonds are formed between the
gas and the adsorbent, such as the formation of CO and CO2 when oxygen is
adsorbed on activated carbon at ambient temperatures. In physical
adsorption, the adsorbate is chemically unaltered, and the gas-solid bond
may be broken by elevation of the temperature. In some cases, the adsorbates
are reactive with each other, and,by concentrating them, the adsorbent
acts as a catalyst by speeding up their reaction rate. Adsorbents may
also be treated with a specially selected catalyst or reactant prior to
use in order to increase its capacity, rate, or selectivity for odor
removal. Various factors influence the quanity of material that can be
adsorbed by a given weight of adsorbent, and are discussed in detail by
Turk (64) and Summer (65). For all practical purposes, the control of
atmospheric odors by adsorption is limited to the use of activated carbon
as the adsorbent (1, 3). It adsorbs all types of odors under almost any
condition, performance is not weakened by the presence of moisture, and it
can be used without making a careful analysis of odor content. Von
Bergen (63) and Lee (66, 67) discuss the applications of activated
charcoal in air pollution control and the design of adsorption systems.
Other adsorbents are available, but show greater selectivity than
activated carbon due to their polar nature. For this reason, they are
more useful than carbon when separations are to be made among different
pollutants in an effluent stream, but are not as good for overall odor
control applications. Typical of these adsorbents are silica gel, fuller's,
diatomaceous, and other siliceous earths, and synthetic zeolites. Metallic
oxides are also used, but are even more polar than the siliceous
adsorbents listed above. For this reason they are never used directly
for source control of airborne pollutants by physical adsorption, but are
used as desiccants, catalyst carriers, or catalysts. Activated alumina
(aluminum oxide) is typical of this group.
Once the adsorbent is saturated with the odorous material, several
means of disposal are available depending upon the various materials
involved. The adsorbate may be desorbed and either discarded or recovered
(if valuable), the adsorbent and the adsorbate may both be discarded, or
the adsorbate may be oxidized on the adsorbent surface.
Where odorous gases are soluble in a liquid, with or without chemical
reactions, absorption methods may prove useful for odor control. The
transfer of an odorous gas from the effluent stream into a scrubbing
liquid requires initial contact of the two mediums in some type of wet
scrubber, and then separation into a clean gas and contaminated liquid
stream. Gaseous transfer is basically through diffusion, moving from a
region of high concentration in the gas stream to one of low concentration
in the scrubbing medium. The odorous molecules may also be attached (or
adsorbed) to particulate matter in which case both would be removed from
the gas stream through various particulate collection mechanisms
(inertial, gravitational, etc.).
Various methods are used for gas-liquid contact in wet scrubbers,
but are divided into two basic categories -- lao energy types and high
energy type. Low energy scrubbers achieve absorption by floa-ing gas
through spray chambers, or through restricted passages in a plate or
packing on which a liquid head is maintained. Examples of this type
include open spray towers, packed towers, wet centrifigal or cyclonic,
flooded bed, orifice, and wet dynamic (68). The high energy types, or
venturi scrubbers, use a venturi for gas-liquid interaction by impacting
high velocity gas on the injected liquid streams, and include the ejector
venturi, flooded-disc scrubber, and the dry or wet venturi. Additional
details, design information, etc.,on liquid scrubbers are given by
Imperato (68), Sargent (69) Calvert (70), and Summer (65).
In all types of scrubbers, it is possible to use an absorbing medium
other than water to achieve better odor control through chemical reactions
between the gas and liquid rather than just physical solution. Of course,
this would depend upon the odorants to be removed, origin and disposition
of the scrubbing liquid, corrosive effects, etc. As with any odor
control system, basic procedures should be follc.ed when considering or
evaluating absorption for odorant removal, and are listed by Von Bergen
(63) in addition to the above mentioned references (65, 68-70).
Many odorous gases can exist as liquids under ambient conditions,
and through simple condensation, cooling such vapors can remove much of
the odor frnii an effluent stream. If the odorant gases have an appreciable
solubility in water and moisture is present in the exhaust stream,
condensation of the water can also reduce the concentration of malodors
released to the atmosphere (71, 72).
The two basic types of condensers are surface and contact. Surface
condensers rely upon cooled surfaces where the coolant does not come in
contact with the odorous vapors. Contact condensers use a mixture of
coolant, vapors, condensate, and non condensable gases, and are applicable
to a wider range of odor-control problems and less expensive than surface
condensers. However, they produce large volumes of contaminated water
which can create a water pollution problem, and care must be taken to
prevent re-release of the odorous gases further downstream (1).
Condensation is basically the transfer of an air pollution problem to
one of water pollution, but can be very beneficial if an efficient
utilization or treatment of the aqueous effluent is available.
Complete oxidation of odorous gases is an effective means of odor
control since the final products are either odorless (H20, C02) or have
comparatively very high odor threshold values. If oxidation is incomplete,
however, the odors may increase as evidenced by incinerators, incomplete
combustion in diesel and internal combustion engines, conversion of
alcohols to carboxylic acids, etc. (3, 63). Oxidation systems may be
classified as combustion (high temperature) or chemical (loa temperature).
The three basic methods used to incinerate waste gases are flame,
thermal, and catalytic combustion. Each method requires that the odorous
effluent be heated to the point where the combustible contaminants will
burn. The only basic difference is the temperature range at which the
In flame (or direct-combustion) incineration, if the emission
concentrations are in the flammable range, they are destroyed by burning
as in a flare. For fluctuating process conditions, or for concentrations
at the lo-er flammable limit, auxiliary fuel is used to maintain a burning
mixture. Although flame incineration can be used over a wide range of
concentrations, it is most economical at high concentrations since the
contaminants could be utilized as fuel. Operating temperature of these
units is approximately 25000 F.
Normally, the concentration of combustible contaminants in an
effluent stream will be considerably below the laoer limit of flammability,
and thermal incineration is much more economical than flame incineration.
In this system a residence chamber is used, and the contaminants are
destroyed by exposure to temperatures of 900-14000 F in the presence of
a flame. Although electric heat energy can be used, the presence of a
flame is an important factor for contaminant removal. With electric
heat, temperatures of 1500-18000 F are required to obtain the same
efficiency achieved with a flame system at 900-14000 F. Continuous
operation at efficiencies of 90 to 99 percent have been achieved with
flame systems (73).
In a catalytic system, the oxidation of odorous effluent occurs
directly on the surface of a catalyst, usually composed of a precious
metal (platinum), and in the absence of a flame. Temperatures out of
the catalyst bed range from 600-10000 F. The basic parts of this system
are a preheat burner, catalyst bed, heat exchanger (for temperatures of
800-10000 F), exhaust fan, and control and safety equipment. Even though
some oxidation may take place in the presence of the preheat burner
flame, the primary function of this component is only to raise the gas
stream to the desired operating temperature for entry into the catalyst
bed. Efficiencies in the range of 85-92 percent are reported for properly
maintained catalyst systems.
Additional details on each of the above combustion systems, such as
initial costs, opcrction and maintenance costs, design variables, operating
characteristics, and Factors influencing the selection of each method,
are reported by Brewer (74) and Pauletta (73).
Chemical conversion of many organi c gases and vapors to odorless
compounds is generally accomplished with oxidizing agents such as ozone,
chlorine, chlorine dioxide, and potassium permanganate. An important
factor in this method of odor control is the fact that these agents do
not always convert organic substances to their most highly oxidized products
(CO2 and H20), so the odor of the intermediates may be of concern.
Depending upon the emission process and the character of the odorous
effluent (moisture content, temperature, etc.) contact between the oxidizer
and odorant may be gas-gas, gas-liquid, or gas-solid.
The only significant examples of gas-phase oxidation have been the
use of chlorine and ozone. Due to its possible corrosive effects, residual
odor, and toxicity, chlorine is seldom used in this method. Although
also toxic, ozone is used in many odor control applications through
release into a stack or vent containing odorous effluent, and is
discussed in Chapter III.
For gas-liquid contact, chlorine, chlorine dioxide, and potassium
permanganate are the principal oxidants and are normally added either to
the scrubbing liquid or to the gases entering the scrubber.
Gas-solid contact is achieved in many applications with the use of
activated carbon impregnated with various compounds which react with the
odorous effluent after initial adsorption by the carbon. Activated
alumina-and an aqueous potassium permanganate solution are also used in
a similar manner (1).
Reducing the concentration of a malodorous effluent to belcw the
sensory threshold by dilution is still one of the most common methods
of odor control. Such dilution is normally achieved by collecting and
discharging all emissions through a tall stack and/or increasing the
temperature and velocity of the effluent, or by locating the source of
emissions (such as a plant) at increased distances from any receptor.
Formulas are available (75) to estimate the average downwind, ground-
level concentration of a plume discharged from an elevated continuous
source (such as a tall stack), and are a function of emission rates, stack
height, effluent temperature, and various atmospheric parameters that
describe the meteorological situation. Up to distances of a few miles,
and for durations up to ten hours, these formulas will predict concentrations
that are within a factor of two or three of the observed values most of
the time. At greater distances and longer time intervals, the predicted
concentration is much more subject to being in error, but generally
should not exceed a factor of ten within the plume trajectory (1, 2).
The formulas predict average concentrations over a period of time,
but most odor complaints usually result from peak levels. Since even a
brief exposure to a peak concentration may be unacceptable to the populace,
the degree of dilution needed for adequate deodorization is normally
much greater than that predicted by the calculations. Some studies (1, 76)
have sho.,in that the distance from a source at which an odor was detected
was not even approximately predicted by dilution and dispersion calculations.
Large discrepancies were found by Wohlers (76) in comparing odor travel
from different plants based on stack-gas dilution calculations. His
most extreme sample was from a kraft pulp and paper mill for which the
average dilution needed to reach threshold was 32:1, and the maximum
64:1. The minimum dilution of actual stack effluent was calculated to be
990:1; thus predicting that the odors would not be detected at ground
level. Field surveys over a six-month period showed that the kraft odor
could be detected at distances up to eight miles where the calculated
dilution of stack effluent was 840,000:1, and there were some reports of
the odor at a distance of 40 miles.
In addition to errors which can be due to various factors used in
meteorological dispersion formulas, the variability of odor threshold in
humans, and the possible role of cumulative secondary sources, Turk (2, 3)
suggests the role of particulate matter in the transfer and/or perception
of odors by humans and as a possible explanation to some of the large
discrepancies found in previous studies.
Various design factors, cost information, etc., on the construction
of tall chimneys are given in an article by Carlton-Jones and Schneider
(77). Meterological influences on mill effluent dispersion are discussed
by Wright (78) and Cramer (79).
Odor masking is a process in which the perception of an offensive
odorous gas is obscured by the addition of other odorous material under
conditions that do not involve chemical change to either agent. Hopefully,
a more pleasant and acceptable odor sensation is created. The masking
agent does not affect the composition of the original malodorous gas, and
if the two odors are of near equal strength, a blend of the two is
observed and both can be identified. If one is considerably stronger than
the other, generally it alone is perceived (2, 63).
Various chemicals are used for odor masking, such as vanillin,
methyl ionones, eugenols, benzyl acetate, phenylethyl alcohol, and
heliotropin, and must be carefully selected for each particular odor
problem. Factors which must be investigated include an analysis for (or
a lack of) toxic materials in the effluent stream, addition of the masking
agent into the exhaust stream, surrounding area, or as an additive to the
process, odor compatibility between the masking agent and malodorous
compounds so a more offensive odor will not result, amount of malodorous
material in the exhaust stream, temperature and other conditions of the
process and stack effluent, and amount of masking agent required. Pilot
plant or laboratory experiments should be run to establish ratios that
produce a desirable result.
Advantages of using masking agents are minimum or no capital invest-
ment for equipment, ease of application, relatively loa operation and
maintenance costs as compared with other odor control equipment, and
immediate availability for common odor problems. The main disadvantage
of this method is that the original malodorous gases are not eliminated
from the atmosphere, and may be of such magnitude that odor masking is
not economically feasible or desirable. Also, additional odors are
being injected into the atmosphere, and,although may be initially more
acceptable than the original malodors, they are still subject to
calnplaints from the population since individual reactions to odors are
quite variable and change from time to time.
In addition to the two references (2, 63) previously cited, a
thorough review. of masking is also given by Summer (65).
Odor counteraction (neutralization) is a reduction in the intensity
of an offensive odor due to the fact that certain pairs of odors in
appropriate concentrations are antagonistic, and when sniffed together,
both are diminished. This principle of odor control is distinct from
that of odor masking in which strong odors tend to mask weaker ones and
the perception of the offensive odor is obscured.
Various pairs of counteracting odorous substances were first published
in 1895, and,since that time, similar studies have been made where it was
possible to compensate the olfactory effects of various chemicals to a
point of total odor disappearance, or close to it. From those studies,
it is knacn that in the group benzene, toluene, xylene, pseudocumene
and durene, combinations from this group in the correct proportions can
be produced which are almost odorless. Thus, many materials are available
for odor counteraction, including the essential oils (63).
Odor counteraction is particularly effective against multiple odor
sources (typical of many industrial plants) which form an odorous "pool"
that emanates from a plant site through molecular dispersion and general
air movement. An atomizing nozzle is used to vaporize the counteractants
and inject them into the air stream where they are designed to mix with
and follow the odorous pool. The vaporizing points are located either
in the exhaust stacks or in the general vicinity of the major odorous
sources. Odor counteraction formulas are designed for specific odor
groups, and the installations should be made by a specialist who can
discriminate between various odors and estimate the intensity of
concentrations involved in the overall problem.
As with odor masking, various procedures must be followed to obtain
the best results, and are covered in detail by Von Bergen (63) and
Summer (65). Advantages and disadvantages of the two'techniques are
similar also. Effectiveness of the installation must be determined by
actual odor perception of the populace, which is actually the criteria for
evaluating the success of any odor control program.
Odor Control in the Kraft Pulping Industry
The major odor control methods used by the kraft pulping industry
include process modifications (black liquor oxidation, elimination of
direct contact evaporator, etc.), proper operation and maintenance of
equipment, combustion and chemical oxidation, absorption, and dispersion.
Odor masking is used to some extent. Condensation, primarily for heat
recovery, does remove malodorous sulfur compounds from the digester and
evaporator gas streams, but presents additional problems in odor control
and waste treatment.
Brief descriptions of each method and literature references are
presented in the following sections.
Dlack Liquor Oxidation
Typical of a process change developed to prevent initial formation
of malodorous sulfur gases is the practice of black liquor oxidation
which is extensively employed throughout the world as a partial solution
to the kraft odor problem. The main purpose of black liquor oxidation
is to oxidize sodium sulfide in the liquor to sodium thiosulfate to
prevent liberation of hydrogen sulfide in the direct contact evaporator.
In the direct contact evaporator, black liquor (at approximately 50
percent solids) is brought into direct contact with hot gases from the
recovery furnace to increase its solids content to a firing concentration
of 65-70 percent solids. With unoxidized black liquor, carbon dioxide
and sulfur dioxide in the furnace gases will react with sodium sulfide
in the liquor to generate hydrogen sulfide as follows (16):
NaS + CO + H20~ Na2CO3 + H2S
NaS + SO2 + H 0 ---- 14 Na SO0 + HS .
2 2 23 2
With oxidized black liquor, ha'ever, the sodium sulfide has been converted
to sodium thiosulfate which is relatively stable, and will not break dawn
in passing through the direct contact evaporator. Thus, neither of the
above reactions will take place and the sulfur will be retained in the
Black liquor oxidation can take place prior to the multiple effect
evaporators (weak black liquor oxidation) or immediately before the
direct contact evaporator (strong black liquor oxidation). Depending on
which system is used, oxidation is normally accomplished by reacting the
sodium sulfide with oxygen from the air in either a packed taoer, bubble
tray tp er, or air sparged reactor. Various factors affect the choice of
a weak or strong black liquor oxidation system and each has distinct
advantages and disadvantages. Thorough reviews of black liquor oxidation
are by Collins (80), Guest (81), Landry (82), Hendrickson and Harding (83),
and Blosser and Cooper (84). The kinetics of oxidation are discussed by
Wright (85) and Murray (86), and effects of operating variables on strong
black liquor oxidation by Morgan, et al. (87). The use of pure oxygen
instead of air for oxidation was studied by Ricca (8), Galeano and
Amsden (88) and Fones and Sapp (89). Significance of the direct contact
evaporator as a major source of odor and the importance of practically
complete black liquor oxidation to prevent these odors is reported by
Murray and Rayner (58) and Blosser, et al. (90).
With na% recovery furnaces and unit processes designed to eliminate
the direct contact evaporator (as discussed in the following section),
the requirement for black liquor oxidation would be virtually eliminated
in the construction of new mills since it was originally developed to
prevent odor generation in this unit. Although it has been shotn that
weak black liquor oxidation will reduce sulfur emissions from the
multiple-effect evaporators, the disadvantages and difficulties with the
system would outweigh this minor contribution. The small volume of the
evaporator effluent stream (generally less than 100 cfm) makes collection
and burning in the lime kiln a relatively simple matter, and in addition,
elimination of black liquor oxidation would also eliminate the need for
treating the oxidizer off-gas (91).
Elimination of Direct Contact Evaporation
The direct contact evaporator has been shown to be a major source
of malodorous sulfur emissions (51, 58, 90, 92), and elimination of this
unit from the kraft recovery process is a major contribution toward
odor control in the kraft pulping industry. Operation of recovery
furnaces without direct contact evaporators has been practiced in
Scandanavion countries for scme time (93, 94) and is showing increased
interest and application in the United States.
Two basic methods have been employed to increase the black liquor
concentration to firing conditions and avoid direct contact with recovery
flue gases. Hochmuth (95) describes the Air Contact Evaporator (ACE)
system by Combustion Engineering, Inc. (96) in which combustion air for
the furnace is preheated by flue gas to approximately 6000 F in an
air heater. The flue gas is thus cooled from approximately 8000 F to
3250 F, and discharged to the atmosphere without coming in contact with
the black liquor. The stream of preheated air then passes to a direct
contact evaporator where it loses sensible heat while evaporating water
from the black liquor. Air leaving the direct contact evaporator is
conveyed to the primary and secondary air admission ports of .the furnace
as combustion air. Any malodorous sulfur gases that may have been
formed during the evaporation process are destroyed by incineration in
the high temperature combustion zone of the Furnace.
Clement and Elliott (97) describe the development of a new recovery
furnace by The Babcock and 'Jilcox Co. (98) to operate directly an unoxidized
high-solids liquor from multiple-effect evaporators, thus eliminating
direct and indirect contact between the furnace flue gases and black
liquor for evaporation purposes. American Can Company's new mill at
Halsey, Oregon, utilizes a three-section, forced circulation concentrator
in addition to a sextuple effect evaporator system to obtain the desired
black liquor concentration for firing, and is discussed in detail by
Canovali and Suda (99).
Proper Operation, Maintenance, and Housekeeping
Regardless of what process changes or additions are made, adequate
odor control will not be achieved unless operations are performed within
design capabilities of the equipment, and continued efforts are made to
prevent leaks and the accumulation of odor-producing materials. Proper
maintenance of control equipment to prevent breakdowns is an absolute
Even though the recovery furnace has long been regarded as a major
source of kraft mill odor, studies (51, 90) have been made on the relation-
ship between furnace operation and odorous emissions which shao that
malodorous sulfur compounds at the furnace outlet can be reduced to
negligible amounts if the furnace is operated within design capability
with enough excess air to provide about 3 percent oxygen in the flue gas,
and if sufficient turbulence is obtained in the furnace oxidation zone.
Furnace overlodaing by excessive black liquor firing rates is the normal
cause of high malodorous emissions since the air handling capacity of
the furnace is limited by the installation (4).
Along with black liquor oxidation, combustion oxidation has proven
to be one of the most effective means of odor control used in the kraft
pulping industry today. Whereas black liquor oxidation prevents the
initial formation of odors, combustion oxidation is used for the
destruction of malodorous gases generated in various kraft processes.
At the present time, this method of odor control is used primarily for
the destruction of digester and evaporator noncondensable gases in the
lime kiln, and has shown to be virtually 100 percent effective (100).
These two gas streams provide a law volume, high concentration source
of malodorous sulfur gases, and can be collected and piped to the kiln
relatively easily. The lime kiln has proven particularly useful for
this method since the temperature (approximately 20000 F) is sufficient
for complete oxidation, air volume is sufficient for dilution, and sulfur
dioxide (formed during oxidation) emissions are negligible (91, 100, 101).
Early work on the development and successful operation of burning
these gases was done by DeHaas and Hansen (102) using an auxiliary
furnace. Adaptation of the system for burning in the lime kiln is
reported by Coleman (103). Numerous other articles (91, 99, 101, 104-106)
are found in the literature describing the successful use of the lime kiln
for destruction of nonccndensable gases. Reports on he use of specially
designed boilers and incinerators (107-109), recovery furnaces (99, 107, 110),
waste wood incineration units (24) and natural gas-fired power boilers (111)
for burning the noncondensable gases are also found in the literature.
The report by Canovali and Suda (99) describes the system at American
Can Company's nve mill in Halsey, Oregon, in which the noncondensable
gases can be alternatively admitted to the recovery furnace in case the
lime kiln is not in operation. Incineration of brown stock washer hood
gases in the recovery furnace is also discussed in this article (99) and.
by Hisey (110).
Catalytic oxidation of kraft mill odors has been considered and
used to some extent. The major problem appears to be related to the
particulate loading of the gases and subsequent poisoning of the catalyst
by these materials (61). Blosser and Cooper (100) mention the experience
of two mills using catalytic furnaces containing a series of porcelain
rods coated with a thin film of catalytic alumina and platinum alloy.
Complete removal of mercaptan and dimethyl sulfide and more than 85 percent
removal of dimethyl disulfide has been observed in one of these units.
Maintenance of automatic controls has been a problem, and frequent
replacement of catalyst cells may be encountered where removal of
condensables is poor. A description of this system is also reported by
Landry and Longwell (112) with a cost breakdown.
Catalytic oxidation has been suggested (113) for controlling
emissions of organic solvent vapors (from printing and coating operations)
but the presence of inorganic materials or catalytic poisons in the gas
stream would have to be investigated.
Gas-liquid chlorination is widely used throughout the kraft industry
for the chemical oxidation of malodorous sulfur gases, particularly the
digester and evaporator noncondensables. Common practice is to expose
these gases to chlorination stage washer effluent from the bleach plant
in either a scrubber or the washer dropleg. Supplemental chlorine may
be required if sufficient amount is not available in the washer effluent.
In an unbleached pulp mill, chlorine gas would have to be supplied From
an outside source. Typical reactions in this process are discussed by
Douglass (5) and DeHaas and Clark (34), and indicate that excessive
amounts of chlorine are consumed. Ruus (114) reports the chlorine
demand to be 2-9 pounds per pound of malodorous compound, and discusses
the subsequent high cost of elemental chlorine oxidation if the bleach
residues are insufficient.
Thomas, et al. (115) discuss chlorination at the S. D. Warren
KraFt Mill in Westbrook, Maine, in which digester blow gases are passed
through primary and secondary deodorizing scrubbers utilizing the
chlorination stage effluent. As a backup, weak hypo bleach can be added
to the secondary unit. Ghisoni (1ll) discusses chlorine oxidation of
recovery boiler effluent gases, and Morrison (106) reports on the use of
chlorination as a backup needed when the lime kiln is not in operation
for burning the noncondensables. Archibald and Von Donkelaar (116),
Lindberg (107), Adams (61), and Trobeck, et al. (117) also discuss
Gas-phase chlorination of kraft pulp mill gases was studied by Koppe
and Adams (118) in laboratory experiments. Gas samples from the recovery
furnace, batch digester, multiple-effect evaporators, and lime kiln of a
mill were reacted with known volumes of chlorine gas and the reactions
recorded by gas chromatography. Concentrations of hydrogen sulfide and
dimethyl sulfide did not change, and the methyl mercaptan was oxidized
to dimethyl disulfide. The process appears to be of limited value as a
means of kraft odor control since the overall odor reduction of the total
gaseous effluent would probably not be sufficient to justify its use on
a plant scale. Nevertheless, gas phase chlorination of noncondensables
has been used for several years at the Peia Pobre mill in Mexico City.
Excellent control of malodorous sulfur gases are claimed for the TLT (117)
system, and obviously at reasonable cost. Tirado and Gonzalez (119)
explain that the total results at the Pena Pobre mill are a combination
of partial results obtained in each of various stages or processes, and
that the experiments by Koppe and Adams (118) may have been conducted
under too mild conditions. Air oxidation of the noncondensables in
the presence of water is used prior to chlorination to reduce chlorine
consumption. Following chlorination, a final wash of the gases with
white water completes the TLT system.
Gas scrubbing is used throughout the industry for odor control, but,
in the interrelationship of air and water pollution, using scrubbing to
reduce the odor in one gas stream may simply be transferring the problem
to some other point in the mill. Unless an effective oxidant is used
in the scrubber to convert the malodorous compounds to innocuous products,
care must be taken to prevent re-release of the odors. Keeping the
absorbing liquid in the process stream is another solution, and most
Ruus (114) discusses the use of white liquor for absorbing hydrogen
sulfide and methyl mercaptan from evaporator noncondensables which keeps
them in the liquor cycle and eliminates the need for aqueous effluent
disposal. Secondary scrubbing of recovery stack gas with weak wash is
reported by Buxton and La Pointe (120) to give hydrogen sulfide recovery
of 90 percent, and Hawkins (121) discusses the successful use of weak
wash for scrubbing evaporator gases. At U.S. Plywood-Champion Paper
Inc., Pasadena, Texas, scrubbers utilizing weak wash from the causticizing
department are used to absorb evaporator noncondensable gases, while
gases from the blow tank, tall oil vent, chlorination tacer vents, and
chlorination stage washer seal box vents are scrubbed with water (122).
Adams (24) discusses the use of an absorption-adsorption combination
at a mill in Sweden. The digester noncondensable gases are scrubbed
with warm (950 F) white liquor to absorb hydrogen sulfide and methyl
mercaptan. The unabsorbed gases then pass through activated charcoal to
remove methyl sulfides and methyl disulfides. The charcoal is then steam
stripped to recover the organic sulfides.
Pilot plant studies are reported by Oloman, et al. (123) for selective
absorption of hydrogen sulfide from kraft stack gases using a solution of
sodium carbonate--sodium bicarbonate.
As a sole means of odor control, dilution of malodorous sulfur
emissions by dispersion from tall stacks is not always effective for a
kraft mill due to the lowI odor threshold of the gases, changing meteorological
conditions, and the varying concentrations of effluent gases. Average
stack height for most mills is 200-300 feet, and their effect on odor
control is most noticeable in the vicinity of the installation. It is
believed that the tallest stack presently in the paper industry is 475
Tall stacks are very effective for kraft odor control when used in
conjunction with other control methods. Ghisoni (111) reports on the use
of a 260-Foot stack located on the rim of a 140-foot-deep valley in Milan.
Remaining odorous gases from a pulp mill located in the valley are piped
to the stack for release. The mill uses chlorine oxidation and incineration
for primary odor control. ':alther and Amberg (101) discuss the use of a
310-foot recovery stack to discharge gases from the lime kiln, poaver
boiler, recovery furnace, and black liquor oxidizer at Crown Simpson
Pulp Co. in California. Malodorous gases from the continuous digester
and multiple-effect evaporators are burned in the lime kiln and scrubbed
in a Venturi scrubber prior to release through the recovery stack.
Odor Masking and Counteraction
Odor masking and counteraction are not used as extensively in the
kraft pulping industry for odor control as some of the methods previously
discussed, and reports concerning their effectiveness are quite varied
Tremaine (124) describes the application of Alamask (125) for kraft
odor control by either (1) spraying, vaporizing, or atomizing it into
the odorous gas stream, (2) adding directly to the process, (3) adding
to scrubbing liquors, or (4) spreading or floating on contaminated
surfaces, and states that addition directly to the odor-producing process
is the best application for the pulp and paper industry. DeHaas and
Clark (34) report on the use of a masking agent added directly on the
chips in the digester (3-4 pints per digester) prior to closing the lid.
Experiences at the mill indicated that the masking compounds would control
digester odors under ideal conditions (warm temperatures, unlimited ceiling,
and strong winds), but effects were not so pronounced under cold, foggy
conditions. Costs were also relatively high.
Meuly and Tremaine (126) discuss a joint study between Gulf States
Paper Co., Tuscaloosa, Alabama, and DuPont on the use of Alamask as a
masking agent. Results were favorable, but greatly depended-upon
meterological conditions, type of wood being pulped, and liquor sulfidity.
The effect of meteorological factors on the amount of masking agent
required is covered by Zirm (127). Tremaine (128) suggests that a masking
agent should be applied to the blow-down heat recovery and concentrated
black liquor at the cascade evaporator in addition to the digesters, and
lists favorable comments made at five kraft mills using odor masking.
U.S. Plywood-Champion Paper, Inc., Pasadena, Texas, has added masking
agents since 1952, and presently uses them in the digester and recovery
areas at a cost of $75-$100 per day (122).
Von Bergen (129) reviews the phenomenon of odor counteraction, and
states that a majority of observers in an evaluation test, covering over
30 days at a paper mill, were enthusiastically impressed by the absence of
mal dors while ccunteractznts weir being used.
Although not usually considered as part of an odor control program
at a kraft mill, condensation has been practiced in the kraft industry
probably ever since the operation of the first mill in 1891, and is
actually the first step in removing malodorous sulfur compounds from digester
and multiple-effect evaporator gas streams.
Condensation of the digester relief (or steaming vessel relief For
continuous digesters) and blow gas is primarily for heat recovery, and is
accacplished with either surface or spray (barometric) condensers. Surface
condensers are normally used to reduce the volume of liquid to be handled.
Turpentine recovery is another advantage for condensation of the relief
gases. In the multiple-effect evaporators, condensation of the steam
from the black liquor in the last body is used to maintain a high vacuum
in the system for circulation.
For the purposes of odor control, the result of condensation is two
effluent streams --. a condensate stream and a stream of noncondensable
gases. Both contain malodorous sulfur caopounds, and must be controlled
for an effective odor program. Various methods are discussed in
previous sections of this paper for odor control of the noncondensable
gas stream, and include incineration, chemical oxidation, absorption,
and adsorption. For the condensate stream, in addition to possible release
of the malodorous sulfur gases, contamination of water courses may
present another problem since the organic load of the combined condensates
can amount to approximately 50 percent of the total waste load in a mill
effluent going to the river or waste treatment plant (34, 53).
Reuse of the condensate within the mill as water makeup, wai water, etc.
has proven to be an effective way of keeping the sulfur compounds in the
process and alleviating some of the load on liquid waste treatment
facilities, but can result in the release of malodorous gases in open
systems such as holding tanks and washers (4, 53). Studies by 1/alther
and Amberg (91) show that whenever digester and evaporator condensates
are reused for pulp washing or scrubbing, the odor problem is intensified,
and that these streams should be completely treated prior to reuse in
':alther and Amberg (101) discuss the use of air oxidation to
reduce the sulfur content of the condensate stream, and also report a
20 percent reduction in BOD and 50 percent reduction of COD in the
effluent. Air oxidation is also used in the TLT (Trobeck--Lenz-Tirado)
system (117) at the Pena Pobre mill in Mexico City, where both contaminated
condensate and noncondensables are reacted with air and water prior to
chlorination. As reported by Tirado and Gonzalez (119) the condensates
lose nearly all their H S and mercaptan content due to air stripping.
Hisey (110) discusses air oxio'ation at a South African mill in which
digester condensates are blan with air into a packed tacer, and the
effluent is used as make-up water in a spray cooling pond. Odors created
are negligible. Stripping of condensate streams with steam (53, 130)
and recovery furnace flue gas (61) are also effective for removing
malodorous sulfur compounds and producing reusable water. Gaseous
effluents from stripping operations must be disposed of similar to other
noncondensable gascs for effective odor control.
Land disposal of condensates and other mill effluent through spray
irrigation has been investigated (131, 132), and could offer an effective
and economical method of disposal for some locations. Odor release
could still present a problem near populated areas.
OZONE CHEMISTRY AND APPLICATIONS
Characteristics, Formation, and Toxicity
Ozone (03) is a triatomic allotrope of oxygen with a molecular
weight of 48, and exists as a pale blue gas under normal conditions.
The molecule is unstable, and slowly decomposes to ordinary oxygen at
normal temperatures. Decomposition is accelerated by heat and moisture,
and is almost instantaneous at 300-4000 F. Due to its high electro-
negative potential of -2.07 V, which is second only to fluorine at
-2.1 V (133), ozone is one of the most powerful oxidants known, and its
reaction is rapid with most materials..
The scientific discovery (in 1840) of ozone and scme of the earliest
investigations of its characteristics are attributed to Christian
Schonbien of Switzerland. However, ozone was first detected in 1785 by
the Dutch chemist Van Maroom who noticed a previously unknown odor
appearing in the air during an electrical discharge. He considered this
phenomenon to be inherent to electricity. In 1801, Cruikshank of England
detected the same odor in oxygen which was liberated duringthe electrolysis
of water, and in 1840, Schonbien reported that the odor obtained during
an electrical discharge in oxygen was caused by the appearance of a new
gas which he named ozone. During the period 1845-1860, a number of
investigators (Marinyak and Delariv, Fremi and Bekerel, Andrews)
attempted to convert pure oxygen into ozone, and made the first
characterization of ozone as an allotropic modification of oxygen. In
1867, Soret used measurements of density during diffusion of ozone to
introduce the formula of 0 (134). The word ozone is derived from the
Greek "ozon" (smelling), from "ozein" (to smell), and from the base
"od" (to smell) as in Latin "odor."
Gas density of ozone at 00 C and 1 atmosphere is 2.144 g/liter.
At a temperature of -111.90 C ozone thickens and is transformed into a
highly unstable fluid of dark blue color. Density of ozone in the
liquid state at -1830 C is 1.571 g/cc, while in the solid state at
-195.20 C is 1.738 g/cc. The melting temperature of ozone is -192.50 C.
Ozone is absorbed by many liquids (although unstable in most) and its
solubility is inversely proportional to temperature, Between 00 and
200 C ozone is absorbed in water some 35 times as much as is air. At
500 C, the absorption is about five times as much as air, and at 600 C,
air absorption is still approximately 16 mg/lO000 g H2/atom whereas
ozone absorption is zero (65).
Although there is some difference in the literature concerning bond
dimensions and angles, there is agreement that the ozone molecule is
triangular in form with an obtuse apex angle, and a resonance hybrid of
at least two structures. According to Mortimer (135), both oxygen-to-
oxygen bonds have the same length (1.26 A) which is intermediate between
the double bond distance (1.10 A) and the single bond distance (1.48 A),
and the ozone molecule is a hybrid of the following two structures:
U.0- .- 0. .0..
A reference by Summer (65) indicates an interatomic distance of 1.2 A
with an obtuse apex angle of 110 degrees. Bailey, et al.(136) and
Kozhinov (134) suggest that ozone is a hybrid of the following four
** we ee CC 9 ** d Qe e
:o= -0- 0: -- :0o -- --- 0: o -- o -- 0:
+ -- + +
-0 --0 -- 0:
+ ** t
Either way, the ozone molecule is extremely unstable and spontaneously
decomposes into an atom and a molecule of oxygen. The reaction is strongly
exothermic (34.5 kcal per mole of ozone) which explains the explosive
nature of ozone under certain conditions. Chemically pure ozone (100%)
explodes violently at the very slightest impulse, and pure liquid ozone
at temperatures near its normal boiling point is particularly dangerous
(137). In practice, explosion of ozone will not occur if its concentration
in the ozone-oxygen mixture or ozone-air mixture does not exceed 10 percent.
Such mixtures are safe at pressures of several atmospheres and under any
conditions -- during heating or shock and in reactions with traces of
organic contaminants (134).
Ozone in the earth's atmosphere is generated by the interaction of
solar ultraviolet rays and free oxygen atoms. The wavelcengths capable of
producing ozone naturally fall within the absorption spectrum of oxygen,
or approximately 1100-2400 A. Ozone is also decomposed by solar radiation,
particularly in the Hartley Band (2200-3000 A), and even at longer wave
lengths since the dissociation energy is only 24.6 kcal. As a consequence
of this generation and decomposition due to solar radiation, an ozone
concentration equilibrium occurs (subject to various meteorological parameters)
which results in an ozone profile as a function of altitude (138-140).
Although the distribution of ozone in the atmosphere varies from one
author to the next, several research projects indicate maximum concen-
trations at altitudes of 13-16 miles (139, 141). Dependent upon the
season and latitude, concentrations of 5-10 ppm can occur in this upper
atmosphere. Ozone concentrations at or near sea level are normally in
the range of 0.01 to 0.04 ppm, although a concentration of 0.99 ppm
was detected during Los Angeles smog conditions in 1956 (142). Processes
involved in the generation of atmospheric ozone, both in the upper and
lower atmosphere (photochemical oxidation), are well covered in the
literature (65, 141, 143).
For laboratory, municipal, and industrial applications, ozone can
be produced by three different methods: electrolytic, silent arc discharge,
and ultraviolet radiation. Of these, only the latter two have proved
practical.since, in the electrolysis of water, ozone yield is severely
influenced by thermal destruction resulting from high curireii densities
required at the anode (138).
For high ozone output, silent arc discharge generators are normally
used, and was the type used for this study. The generating unit consists
df a pair of large area electrodes, either flat or cylindrical, separated
by an air gap and a dielectric layer. In many generators, the electrodes
are stainless steel and aluminum, and the dielectric is borasilicate
glass. One electrode is contiguous to the glass dielectric, and the
gas gap occurs between the dielectric and the other electrode. A
common gap distance for voltages belao 15000 volts is 1-3 mm (138).
In the presence of high, alternating voltage discharges, ozone is
generated from the oxygen in the gap between the two electrodes. Key to
the process is the presence (in the air gap) of stray electrons from a
previous discharge or background radiation. In a field of sufficient
intensity, the electrons move to the positive electrode, thus setting
up a series of molecular collisions. Some achieve sufficient velocity
to penetrate oxygen molecules, producing additional free electrons or
high energy unstable molecules which break daon into free radicals --
single oxygen atoms. Most of these single atoms combine with oxygen
molecules (02) to form ozone (03 ). The process is repeated when polarity
is reversed in the second half-cycle, and the electron's movement to
the opposite electrode creates new molecular collisions. Thus, there
are two productive instants in each cycle, occupying a relatively small
portion of the total cycle time. The production efficiency is therefore
increased when the cycle is shortened; i.e., when the frequency is
raised. Most generating units are designed to operate in the 5000-25,000
volt range (through suitable transformers)) nd at frequencies up to
1000 Hz, although some designers are exploring the possibilities of higher
frequencies for increased production efficiency (144). Normal generator
output is 1-3 percent ozone by weight.
The ozone yield of the silent arc discharge generator is a function
of the following factors: oxygen purity of the parent gas, oxygen temperature,
pressure, and flo'i, peak voltage, Frequency, capacitance of discharge gap,
and capacity of dielectric. For a given configuration generator, ozone
output is a function of current density with all other conditions held
constant, and the output is approximately doubled when the parent gas
is pure oxygen rather than air.
Using oxygen as a feed, various effects have been noted in generator
output with different gas diluents. As discussed by Cromwell and Manley
(145), the greatest loss in ozone yield is noticed with the presence of
less than 1 percent of hydrogen, water, or Freon 12. Up to 50 percent
of carbon dioxide or argon reduces the yield to about 85 percent of that
expected from pure oxygen. Carbon monoxide or nitrogen, in amounts of
less than 10 percent, apparently increase ozone output, but in greater
amounts a gradual decrease is noted. Inoue and Sugino (146) studied
the inhibiting action of several hydrocarbons on ozone yield by silent
arc discharge, and report that the rate of ozone formation varies
linearly with the hydrocarbon concentration, regardless of the type of
hydrocarbon. Water vapor remains the most common contaminant to be
guarded against, however, and can be eliminated by adequate drying of
the parent gas (usually specified to -600 C dcpoint) prior to ozone
Additional information concerning ozone generation by electrical
discharge is discussed by Lunt (14/), Summer (65), Suzuki, et al. (148),
Ogden (144), and Fuji and Takemura (149).
Ozone generation by ultraviolet radiation is most commonly used
where la ozone concentrations are required, such as laboratory work, air
purification, elimination of molds and bacteria, etc. As with the silent
arc discharge generator, air or oxygen can be used as the parent gas.
Ozone yield is a function of total effective radiation cmitted in the
range of 1100-2200 A, which is in turn dependent upon lamp design,
emission area, current density, and oxygen fla-, pressure, and temperature
(138). The reactions involved are similar to those causing the formation
of ozone in the upper atmosphere, and are discussed along with generator
design information in the literature (65).
In an excessive dosage, most any substance is toxic to the human
organism. Substances assume the title "toxic," however, when the
dosage determined to be excessive is a very lo. unit quantity. Within
the framework of this definition ozone is definitely toxic, since the
maximum average concentration for an eight-hour exposure has been set
at 0.10 ppm by the American Conference of Governmental Industrial
Hygienists (150). With the increasing use of ozone in municipal,
commercial, and industrial installations, the occurrence of ozone in
smog conditions, and the possible hazards of high altitude flight from
atmospheric ozone above 50,000 feet, the toxicity of this gas is an
Ozone was first recognized as a toxic substance by Andrews in 1874
when he exposed small laboratory animals to high concentrations of ozone,
resulting in their death even after a short exposure (151). Since then,
numerous toxicity studies have been performed on both animal and human
subjects. Naturally, a majority of these investigations have dealt
with animals (rats, mice, rabbits, dogs, cats, guinea pigs), and cover a
multitude of conditions, ozone concentrations, and exposure times (152-170).
Most of these studies have established an LD50 rate for the various
animals, which is the dose of ozone that will, on the average, kill 50
percent of a significant number of animals exposed under specified
conditions. Since there are several experimentally simple and statistically
sound methods of obtaining LD50 values and various procedures for determining
ozone concentration, and due to the multitude of test conditions for the
above referenced studies, there is some confusion in the literature on
the actual toxicity of ozone. Regardless of the exact toxicity, the
studies have left no doubt that ozone in acute exposures is a highly
toxic and lethal substance, and that its impact is directed primarily
on the soft tissues of the respiratory tract.
Gilgen and Wanner (151) review many of the ozone toxicity studies
on animals, and noting the diversity of values, provide a brief summary
of the literature. In the case of exposures between 3 and 24 hours, the
LD50 for rats and mice varies between 3.8 and 22 ppm. Hamsters, rabbits,
and guinea pigs tolerate higher concentrations. Death is caused by
pulmonary edema. Small concentrations lead to usually reversible
changes of the respiratory organs, to disturbances of respiration, and to
changes in the activity of various enzymes. Exposure to low concentrations
can also develop a tolerance against subsequent exposure to lethal
concentrations. Chronic exposure to ozone causes damage to the
respiratory organs (bronchitis and lung emphysema), inhibition of the
weight increase of younger animals, and decreased acidity of the urine.
The article by Stokinger (165) is also an excellent review of ozone
toxicity to animals.
Numerous articles are found in the literature describing controlled
experiments on the toxicity of ozone to humans (171-176), and also
some effects of ozone from actual working conditions (177-182). A brief
review of some of this literature to illustrate the toxicity of ozone
includes a decrease in vital capacity of the lungs after a single exposure
to 1.5 ppm for a half hour and a subsequent exposure to 2 ppm for 1.5
hours (176), decrease of the CO diffusion capacity after exposures to
1.2 to 6.0 ppm for 1-2 hours (171), significant reduction in diffusion
capacity, vital capacity, and maximum flow in the middle of the expiration
phase after inhalation of 0.6 to 0.8 ppm for 2 hours (175), Impairment of
the upper respiratory passages in welders working with the shielded arc
welding process who were exposed to 0.8 1.7 ppm ozone every day (180),
accelerated pulse, sleepiness, lasting headaches following brief
inhalation of 5-10 ppm and impulse to cough and great lassitude after
1.5 hour exposure to 1.0 ppm (182), severe ozone intoxication (dry cough,
weak pulse, locered blood pressure, nearly unconscious) of a crane
operator who operated a crane approximately 14 feet above a tank
containing 1 percent ozone (181) and lung edema following exposure to
ozone concentrations above 4-5 ppm for an hour (171).
The study by Clamann and Bancroft (171) involved five subjects, ranging
in age from 19 to 54 years. The highest ozone concentration applied was
6 ppm for 1 hour, and the longest exposure time was 2.5 hours at 1.2 ppm.
They concluded that ozone acts only superficially on the wet, soft
tissues of the respiratory tract, and is not able to penetrate deeply.
A definite impairment on the sense of smell was found at exposures of
2.41 ppm for 1.5 hours and 4.16 ppm for 22 minutes. No effects on
blood pressure, pulse rate, and blood itself were observed. Long exposure
times and higher concentrations tend to carry injury to the respiratory
tract deeper toward the alveoli. in two of their experiments involving
inhalation of air with 4.8 ppm ozone, the exhaled air contained practically
no ozone, suggesting that ozone decomposes completely in the respiratory
tissues; however, studies by ballet: (174) showed that the expired air
of his experimental subjects contained 25-75 percent of the inhaled
ozone. The effects of ozone are not limited to the respiratory system
according to some authors. Brinkman and Lamberts (183) have described
a lowered rate in the release of oxygen from oxyhemoglobin in the skin
capillaries, and Lagerwerff (172) reported an impairment of the visual
A summary of the biological effects of ozone for concentrations
ranging from 0.01 10 ppm is shown in Table 3 (65). Similar values are
reported by Biget, et al. (141).
BIOLOGICAL EFFECTS OF OZONE
Concentration (ppm) Effects
0.01 Threshold odor of pure ozone
0.05 Irritation of the lungs and respiratory mucosae
after prolonged inhalation
0.10 Respiratory discomfort, headache
0.15 Eye irritation threshold
0.50 Considerable discomfort, disordered breathing
1-10 Coughing, fatigue, headache, possible coma after
5-10 Stupefaction, body pain, accelerated pulse,
Tolerance to ozone has been demonstrated in laboratory animals
(153, 155, 156, 165-168), but it is still unclear whether a tolerance
develops in humans. Gilgen and Wanner (151) and Stokinger (155) suggest
the possibility that the population of Los Angeles has developed an ozone
tolerance resulting from the often repeated exposures to low ozone
concentrations, and due to the fact that the inhabitants have reported
no acute pulmonary damages even when the ozone concentration exceeds 0.5
ppm. In the case of welders, ozone concentrations of more than 0.5 ppm
led to the development of acute lung damages in every case, but their
re-exposures occur after short intervals, thus they are not able to
develop a tolerance. Animals always shqoed the best tolerance when their
exposures were separated by an interval of several days.
Various explanations are found in the literature concerning the
mechanism of ozone's action on the body (156, 167, 183-189). Fetner
(188) and Brinkman and Lamberts (183) suggested the possibility of radio-
mimetic effects of ozone. Fetner (185) showed that human cell cultures,
which had been exposed for 5-10 minutes to 8 ppm of ozone, exhibited
chromosomal cleavages similar to those resulting from an irradiation with
200 roentgens. Zelac (187) reported on chromosome aberrations in
circulating blood lymphocytes of Chinese hamsters due to ozone inhalation,
and stated that if the results of the study were directly extended to
the human case, presently permitted human ozone exposures wbuld be
expected to result in break frequencies that are orders of magnitude
greater than those resulting from permitted human radiation exposures.
An excellent review of possible mechanisms of ozone action is given by
Fenn and Rahn (190), and also covers other areas of ozone toxicity (tolerance,
factors affecting toxicity, effects on man and animals, etc.).
As stated by Gilgen and Wanner (151) and Fenn and Rahn (190), so
Far it has not been possible to determine the meaning of the immuno-
chemical changes taking place during the appearance of the ozone effect;
therefore, it is impossible to give a meaningful explanation for the
pulmonary damages resulting after chronic exposure to ozone. Regardless
of its mode of action, controlled studies on animal and human subjects,
and actual cases of ozone exposure in working environments, have
definitely established that ozone is a highly toxic gas.
The market for ozone application in the waste treatment field is
normally as a polishing agent in conjunction with other accepted waste
treatment systems. Studies have shown ozone to be very effective in
treating industrial waste streams, particularly those containing cyanide
and phenols. The Boeing Company plant at Wichita, Kansas, uses more than
350 pounds of ozone per day as a secondary waste treatment process to
cope with cyanide, phenols, oils, detergents, sulphides, and sulphites.
The waste is ozonated at approximately 20 ppm and discharged into a
lagoon where fish have survived since the beginning of the project in
1957 (144, 191, 192). Reports on the kinetics of ozone-cyanide reactions
(193, 194), comparisons of chlorine and ozone treatment of cyanide wastes
(195, 196), laboratory and pilot plant studies of the ozone oxidation of
phenols (197-201) and others (202-205) describe the feasibility of using
ozone for treating industrial waste waters.
Air Reduction Co. (Airco) was awarded a contract by FWlPCA to construct
a pilot plant at the Washington, D.C. Blue Plains treatment facility to
study tertiary treatment of municipal waste with ozone. Previous work
by Airco has demonstrated that ozone treatment of sewage plant effluent
destroys organic contaminants, surfactants, and bacteria, and produces
a clear, odorless effluent that meets PHS standards for drinking water.
Economic evaluations indicate that the ozone treatment will be competitive
with activated carbon (206, 207). Similar results were obtained in
pilot plant work using effluent frcm trickling filter waste treatment plants
at Totawa and Berkeley Heights, New Jersey (144). Additional studies
on ozone treatment of municipal waste effluent are covered in the literature
The most prominent use of ozone for odor control is in the treatment
of sewage exhaust gases, primarily hydrogen sulfide. Installations have
included odor control for sludge storage and supernatant tanks (213),
treatment of ventilating and exhaust gases from fully enclosed aeration and
settling tanks (6), lift station and wet well exhaust (214), complete
underground sewage treatment plant in Atami City, Japan (215), grit
removal, bar screens, settling and chlorination buildings in Mamaroneck,
New York (191), and treatment plants in Florida (216, 217) and Michigan
(218). One of the earliest installations was the Ward's Island Treatment
Plant in New York City, installed about 1930 (6). The plant treats an
exhaust gas stream of 30,000 cfm with an ozone concentration of 1 ppm,
and reports complete elimination of the odor problem.
Other installations using ozone for odor control include a pharmaceutical
plant (219), rendering plants (214), Fish-processing operations (6), rubber
compounding plant exhaust, phenolic odors from felts, and commercial
Table 4 lists typical ozone concentrations For various installations
established from experience by the Welsbach Corp. (220), although exact
dosages would depend upon retention times, temperature, humidity, and
concentration of the exhaust stream, and nature of the malodorous gases
being treated. These figures are based on a contact time of at least
OZONE CONCENTRATIONS FOR ODOR CONTROL
Appl ication Ozone Concentration (ppm)
Se.vage plants, general 1
Phenol plants 3-10
Rubber plants 3-10
Fish-processing plants 10
Sludge storage and vacuum filter 10
Rendering plants 10
Paper mills 10-50
The most widespread application of ozone is in the treatment of drinking
water, and the first recorded commercial installation for this purpose began
operation at Nice, France, in 1906. Over 20 million gallons of water are
steril ized daily with ozone at the water works in this city (221). There
are currently more than 500 municipalities using ozone for water treatment
in over 50 countries. Its use is definitely more widespread in European
countries who use ozone as the disinfectant of choice, while in the
United States, a chlorine residual is required in order to provide a
disinfectant at the consumer's tap. In many of these installations, ozone
is used in conjunction with other water treatment techniques (chlorination,
microstraining, filtration, flocculation, carbon filtration, etc.), and
depends upon the raw water and the water quality requirement (211, 222,
The use of ozone for water disinfection and the removal of tastes,
odors, and color is well covered in the literature. Nobert (224) discusses
the use of ozone at five water works plants in Quebec to produce good
quality water free from tastes and odors, which had not been previously
achieved with chlorination and activated carbon. Studies comparing the
effectiveness of ozone and chlorine (156, 212, 225-228) for water treatment
stress that ozone leaves no residual taste or odor, and reacts more
rapidly than chlorine in killing most bacteria. In addition to municipal
water treatment, ozone has been used to obtain high quality water for
breweries, distilleries, water bottle operations, food processors,
carbonated beverage producers, hospitals, laboratories, etc. (144, 222).
Additional reports on ozone installations (204, 205, 229), production
(230), and methods of mixing for more efficient disinfection (231) are
available in the literature.
The use of ozone for sterilization of food containers (232), and
control of mold and bacteria (65, 156, 233, 234) demonstrate other uses
of this oxidant, although in some cases the required concentrations
exceed the recommended limit for humans and would limit these ozone
applications to certain situations.
Ozone for Kraft Odor Control
Previous A ppi ications and Investigat ions
As a method of eliminating odors in kraft pulp mill operations, a
patent was issued to Limerick (7) concerning the procedure of conducting
ozone to a line discharging odorous stack gases, mixing the ozone with
such gases in a proportion from about 10 to 100 ppm of the total gas
volume, retaining the ozone and odorous gases in intimate mixture for
not less than about two seconds, and thereafter discharging the mixture
to the atmosphere.
As mentioned in the introduction of this paper (Chapter 1), literature
concerning the application of ozone for kraft odor control is very limited.
The only reference found that mentions an actual mill application was by
Cromviell (6) who briefly describes the experimental use of ozone to
treat effluent stack gases from the sulfate recovery system of a pulp
and paper mill in Canada. Ozone was applied to a 100-foot stack at a
point 10 feet above the ground, giving a retention time of approximately
two seconds for mixing with the effluent gases. Maximum ozone output
was 300 pounds per 24 hours, which provided a concentration of 80 ppm
in an effluent flow of approximately 50,000 cfm. Concentration of sulfur
compounds in the effluent stream was not given, and there was no indication
of how effective the ozone application was. A thorough literature
search and correspondence (235) failed to provide any additional
information on these tests.
The literature is also somewhat limited on laboratory studies
concerning ozonation of the individual sulfur compounds primarily responsible
for kraft mill odors; i.e., hydrogen sulfide, methyl mercaptan, dimethyl
sulfide, and dimethyl disulfide. Barnard (236), discussing the reaction
of ozone with organic sulfur compounds, showed that the oxidation of
monosulfides (R.S-R) occurs in two well-defined stages to give, first,
the sulfoxide (R-SO-R) and then the sulfone (R.S02*R), and require
somewhat less than the theoretical amount of ozone for their oxidation.
With this analogy, disulfides would be expected to yield successively
thiosulfinate (R-SO.S-R), thiosulphonate (R-SO2S.R) or disulfoxide
(R-SO.SO-R), and disulfone (R.SO2-SO2 R). Ozone-uptake curves did not
reveal such stelpise reaction and indicated, for "normal" disulfides, the
absorption of 2.5-3.0 moles of ozone. The major product of the reactions
was a sulfonic anhydride (R-SO2.0-SO 2R) with small amounts of thio-
sulfonate and disulfone, although ozonation of dimethyl disulfide
showed the major product to be the thiosulfonate (50 percent) rather than
the sulfonic anhydride (39 percent).
Similar results are reported by Douglass (5) using nuclear magnetic
resonance (NMR) analytical techniques. Ozone was found to react rapidly
with dimethyl sulfide to form the sulfoxide and then the sulfone without
evidence of other products being formed (Equation 3.1).
0 0 0 0
CH SCH -3-- CH SCH --3 CH SCH (3.1)
With methyl mercaptan, a stel:p.ise reaction was noted, first forming
the disulfide which is oxidized to the thicsulfinate, and then to the
thiosulfonate (Equation 3.2).
0 0 0 0 0
CH SH --- CH SSCH -- CH SSCH ----3 CH SSCH (3.2)
Bubbling ozone through liquid dimethyl disulfide appeared to convert it
entirely to the thiosulfonate (R-SO *S-R), and attempting to distill
the thiosulfonate thus obtained gave a mixture of thiosulfonate and
sulfonic anhydride (R-SO2-0SO2R).
Ozone oxidation of kraft pulp mill blow gases (from an autoclave)
was studied by Akamatsu (9) and indicated that most of the odor could be
eliminated; however, concentrations of sulfur compounds in the blaw gas
or ozone concentrations used for odor control were not mentioned.
Results showed that ozone converts the methyl disulfide and methyl
mercaptan of the blaw gas to a "mild odored" oxide. Akamatsu, et al.
(237) also studied the ozone oxidation of dimethyl sulfide, and showed
that 75 percent of this malodorous gas could be oxidized to dimethyl
sulfoxide and the remainder to dimethyl sulfone, similar to the studies
by Barnard (236) and Douglass (5).
During a study on the oxidation of kraft black liquor with oxygen,
Ricca (8) evaluated the effects of oxygen enriched with 1450 ppm ozone.
The reaction rate of dissolved hydrosulfide was not increased but the
mercaptan odor of the black liquor was eliminated, thus indicating
that ozone was more effective in oxidizing gaseous odors than black
liquor. Additional tests showed that the black liquor oxidation exhaust
gases could be effectively deodorized with 100-150 ppm ozone (at 780 C).
Reactions between ozone and hydrogen sulfide have been reported
on by several authors. Gregor and Martin (10) studied the reaction at'
ambient temperature and illumination by injecting known quantities of
each gas into glass reaction vessels, and allca.ing them to stand for
16 hours to insure complete reaction. Results of their work indicated
that when hydrogen sulfide and ozonized oxygen are mixed under initially
anhydrous conditions, reaction occurs with water, sulfur dioxide and
sulfuric acid being formed as end products.
Under their reaction conditions, the percentage of hydrogen sulfide
actually oxidized steadily increased with an increase in the ozone to
hydrogen sulfide ratio, and complete oxidation only occurred at ratios in
excess of 9 to 1 (Figure 1). At ozone to hydrogen sulfide ratios below
9 to 1, sulfur dioxide was the principal product, and at ratios above
this, sulfuric acid became the predominant product. At a ratio of 18 to
1, oxidation of all the initial hydrogen sulfide to sulfuric acid was
considered complete (Figure 2). The overall stoichiometry of the
reaction (under the test conditions) was represented by the equation
H2S -- (1 x) H2SO4 + x H20 + x SO2 (3.3)
where 0 x 1 and x --:- 0 as the initial ratio of 0 /H2S -- 18
Cadle and Ledford (11) also studied the gas-phase reaction of ozone
with hydrogen sulfide, and when the reaction was allowed to go to completion
in the cell of an infrared spectrometer, the only products found were
sulfur dioxide and water. Sulfuric acid was not detected as reported by
Gregor and Martin (10). The stoichiometry corresponded approximately to
H S + 03 -- H 0 + SO2 (3.4)
A summary of Cadle and Ledford's data is given in Table 5, indicating a
maximum ozone to hydrogen sulfide ratio of approximately two, which is
somewhat less than the maximum studied by Gregor and Martin.
At the low concentration ratios, the data of Gregor and Martin
(Figures 1 & 2) also show the major reaction product to be sulfur
Initial H S Oxidi
Initial 0 3/H2S Ratio
Figure 1. -- Conversion of Initial H2S to SO2 and H2SO, (Gregor
0 Converted to SO2
ConeInitial H 2 S
Converted to H 2SO
idized H2S Converted to SO
0 Oxidized H2S Converted
tLo H 2S 04
Initial 0 /H2S Ratio
Figure 2. -- Conversion of Oxidized H2S to SO2 and H (Gregor
REACTION DATA OF CADLE AND LEDFORD FOR THE GAS PHASE REACTION
OF OZONE AND HYDROGEN SULFIDE
Concentrations (moles x 10 /cm3
Initial H S Initial 03 Final SO
78 30 19
31 30 23
78 30 18
16 30 18
dioxide, with only small amounts of sulfuric acid. Also, since the
oxidation of sulfur dioxide to sulfur trioxide is knawn to be very slow
(238), it is possible that sulfuric acid did not have sufficient time
to form in the study by Cadle and Ledford since reaction time was not
mentioned, whereas 16 hours were allowed in the work by Gregor and
Considering data from both studies (10, 11), it would appear that
the stoichiometry represented by Equation 3.4 is valid for lao ozone to
hydrogen sulfide rations, and for all ratios at minimum reaction times.
Cadle and Ledford (11) also studied the kinetics of the ozone-
hydrogen sulfide reaction using a I-meter long reaction tube constructed
of 16 mm Pyrex tubing. Maximum residence time in the reactor was about
3 seconds. Tests were run at 260, 650, and 1000 C with ozone concentrations
ranging from about 245-7350 ppm and hydrogen sulfide concentrations of
245-19,600 ppm. It was determined that the reaction had orders near
zero and 1.5 in hydrogen sulfide and ozone respectively, and followed
the rate expression
c = 2.5 x 108 exp (-8300/RT) c 15 mls (3.5)
dt 0 ml sec
The homogeneity of the reaction was examined by placing several
pieces of Pyrex tubing in the reactor which increased the surface to
volume ratio four-fold. The overall reaction rate was about double
that of the original system, indicating significant wall effects. No
attempt was made to quantify these effects. It was concluded that the
reactions in the system were at least partially heterogeneous, and the
rate law (based on a purely homogeneous reaction) could be used to
estimate upper limits for the rate of reaction between ozone and hydrogen
sulfide in the atmosphere.
Hales (12), considering the significance of simultaneous bulk and
wall reactions, and based upon the information and experiences of Cadle
and Ledford (11), performed additional studies on ozone-hydrogen sulfide
reactions in laminar fl.w reactors. Measurements were made at 28.50
and 480 C in the absence of light, and involved a number of reactant
concentration ratios. The reaction was observed to be almost totally
homogeneous, and described by the expression
dc_2 0.5 1.5 micromoles
dt 22.8 exp (-6500/RT) c0.5 1.5 micromoles (3.6)
dH2S 03 liter min
for the range of variables studied. All evidence obtained during the
experiments indicated that Equation 3.4 represents the true stoichiometry
of the reaction for the conditions studied.
Although the rate equations (3.5 and 3.6) are similar in that they
indicate a reaction order of 1.5 in ozone, they differ in the reaction
order of hydrogen sulfide and in the activation energy. Also, reactions
observed by Hales were thought to be almost totally homogeneous, whereas
Cadle and Ledford observed significant wall effects in their system.
Various possibilities for these differences were investigated by Hales,
and it was concluded that the disagreements stem mainly from a nitric
oxide quench used by Cadle and Ledford which, among other things, could
have conditioned the reactor wall in such a way that it became catalytically
active. This would also offer some explanation to the fact that the
absolute rates predicted by Cadle and Ledford (Equation 3.5) are about
two orders of magnitude higher than those of Hales (Equation 3.6) over
the range of hydrogen sulfide concentrations studied (12).
Noting that Equation 3.6 applies to rather specialized conditions
and that care should be exercised in its application to practical
situations, Hales considers a hypothetical odor control situation
involving the emission of hydrogen sulfide at 500 ppm from a 150 C
stack. Assuming no spontaneous ozone decay and a stack residence time
of 10 seconds, Equation 3.6 predicts that only about 2 percent of the
hydrogen sulfide will be oxidized upon applying a stoichiometric
quantity of ozone. Although recognizing the limits of his rate equation,
and the complexity of the ozone-hydrogen sulfide reaction mechanism and
the stack environment itself, Hales concludes that the effect of ozone for
odor control is due more to its odor-masking properties than to actual
oxidation of the hydrogen sulfide.
Because ozone is a powerful oxidizing agent, its classification as
a masking agent is subject for argument (213), since a true masking
compound does not enter into chemical reactions with the malodorous
substance being masked. Naturally, if excessive amounts of ozone are
injected into an effluent gas stream, and/or adequate reaction time is
not provided, ozone could be emitted into the atmosphere and act as a
masking agent due to its o.'in "sweet" odor. The inhibitory effect of
ozone on the sense of smell would also tend to negate the effect of a
malodorous compound, but its use in this manner, or as a pure masking
agent, could result in serious problems due to the knaon toxic effects
Basis of Present Study
Various methods used for odor control in the kraft pulping industry
were reviewed in Chapter 2. The referenced literature indicated that
proper application of these methods would permit a mill to comply with
most any current or proposed emission standard. What the future regula-
tions will be concerning odor compounds. is not known. Additional odor
control steps may be required to completely eliminate the malodorous
emissions, for whether these gases are released in concentrations of
5, 10, or 100 ppm, if they can be detected by the public and create a
nuisance, they are still too great.
The application of ozone as a polishing agent following a major piece
of odor control equipment is a possibility. Older mills may not find
it feasible to make major process changes and/or additions (black liquor
oxidation, elimination of DCE, etc.) to prevent initial odor formation,
but find it easier to treat the malodorous compounds at the emission
points by injecting ozone into the gas stream. If a mill has effective
recovery furnace control and incineration of the noncondensables, some
of the emissions which were once considered unimportant now become major
odor sources and may be suitable for ozone oxidation, such as the brown-
stock washer hood vents, knotter hood vents, and washer seal tank
Is ozone effective for controlling odors in the kraft pulping
industry, for which effluent gases, and under what conditions? A
relatively few laboratory studies investigating a limited range of
conditions, no reported results of ozone installations in kraFt mills,
and conflicting reports on the effectiveness of ozone for odor control
have not tended to provide suitable answers for these questions, although
ozone has been shown to be effective in other applications (water and
wastewater treatment) and for other malodorous installations.
In almost every investigation regarding either the sole use of ozone,
or comparing it with other oxidizing agents (chlorine, chlorine dioxide,
etc.) for water treatment, wastewater treatment, or odor control, the
cost of ozone production was invariably mentioned as its main disadvantage.
In many instances, it has been sho.n to be more effective and faster
acting than chlorine, and without the residual effects sometimes
experienced with chlorine. With the anticipated decrease in cost of
ozone production (as discussed in Chapter 1), its effectiveness for
kraft odor control needs to be established.
The end products resulting from ozone oxidation of the four malodorous
sulfur compounds of interest have been established by previous investigators
(5, 9-12, 236) and shown to be relatively odorless. Kinetics of ozone-
hydrogen sulfide reactions were studied by Cadle and Ledford (11), resulting
in a rate expression based on the depletion of ozone. Hales (12) also
studied the ozone-hydrogen sulfide reaction kinetics to resolve some of
the difficulties experienced by Cadle and Ledford, and derived a rate
expression based on the generation of sulfur dioxide.
For the present study, ozone oxidation of all four of the major
malodorous sulfur compounds was studied, and the effectiveness of the
ozone treatment was based on the depletion of the malodorous compounds
themselves rather than measurement of the end products or ozone uptake.
Although partial oxidation of some compounds, such as butanol (mild odor)
to butyric acid (very bad odor), can result in an odor increase (63), this
does not pose a problem in ozone oxidation of these four sulfur gases of
interest since their reaction products have a much higher odor threshold
than the original compound. Of course, the first step in the ozone
oxidation of methyl mercaptan is dimethyl disulfide, but the formation
and subsequent depletion of this compound was monitored along with the
destruction of the original methyl mercaptan.
Description of Test Stand
Basic Flaw Pattern
As mentioned in Chapter I, the initial phase of this study was to
evaluate the effectiveness of ozone in oxidizing bottled sulfur gases at
various temperatures, reaction times, and concentration ratios. The
equipment, flow diagrams, and techniques for this part of the study were
slightly different than those required for oxidation of actual stack
effluent. Descriptions in this chapter are directed primarily toward the
initial phase of the study, and any variations required for the stack
gases will be described in Chapter VI. Both phases of the study were
conducted from the NCASI Mobile Laboratory since a majority of the analytical
equipment used for the experiments is permanently mounted in this trailer.
A brief description of these units is given in a later section of this
With the exception of a Few components (primarily the analytical
equipment mentioned above), all the experimental apparatus used for the
study was mounted on a test panel as shown by Figures 3 and 4. This panel
was designed and constructed to (1) allow for changes in reaction tube
temperature, sulfur gas concentrations, ozone concentrations, and
residence time with a minimum time delay, (2) provide ease of handling for
all the components with subsequent minimum set-up and take-do.-n time,
(3) provide ease of operation by concentrating all the necessary controls
in a relatively small area, and (4) require a minimum area for operation
due to the limited space in the NCASI Mobile Laboratory.
A schematic of the complete system used for the initial phase is
shawn by Figure 5 and indicates the components attached to the test
stand. As shoin by the schematic, pure oxygen was released from a
cylinder source, passed through a moisture trap of silica gel, through a
flacmeter, and into an Ozonator Corporation (239) Model 4000 Ozone
Generator where ozone was produced by silent arc discharge. The ozone-
oxygen flay from the ozone generator was then directed to the heated
reaction tube or wasted to the atmosphere as required.
Nitrogen was used as a carrier gas for the sulfur compounds and was
admitted to the system from a cylinder also. Sulfur compounds were
injected into the nitrogen stream using a spinning syringe, and the
nitrogen-sulfur gas mixture combined with the ozone-oxygen flcw in a
mixing chamber attached to the end of the heated reaction tube. The
extent of ozone oxidation was monitored with a gas chromatograph through
sampling ports spaced do.n the reaction tube to provide various mixing
times. Waste gases were vented to the outside of the mobile laboratory.
Ozone production was controlled by varying the voltage to the ozone
generator with a voltage regulator (Variac). As sho.-n by the schematic
(Figure 5), a constant voltage transformer was installed prior to the
variac to dampen out any line fluctuations, and provide a steady 118 volts
to the variac. By regulating vol tge with the variac, output of the 6000
Limits of Test Panel-- Orifice
Nitrogen Flocmeter Control.ler oZ .-
Source Controller Atmosphere
11O V '
Line Control Waste Control
_--- Va sWaste Controlo
""--- ....... V, lv 've Waste 03 & 0O
Constant Variac 6000 V To Atmosphere
I VC olst a g e v r i c T r a n s f o rm e r O z o n e
/ Vol age
T o __ Generator Ozone Sampling Train I
i l .
Oxygen IMoisture Control Oxygen Midget Moisture Sampling Control Vacuum
Source ; Trap Valve Flowmeter Manometer Impinger Trap Flcwmeter Valve ; Pump
Sore:Ta migr Ta
Figure 5. -- System Schematic for Oxidation of Bottled Gases.
volt transformer (supplied with the Model 4000 Ozone Generator) was
controlled, thus controlling ozone production in the ozone generator.
Oxygen flow to the ozone generator, which can also affect ozone production,
was maintained at a constant 350 cc/min, and pressure in the generator was
held at 34.0 in Hg. Voltage settings for desired ozone production were
determined from a calibration curve as described in Appendix A.
From the ozone generator, 80 cc/min of ozone-oxygen mixture was
directed to the heated reaction tube. The remaining 270 cc/min was discharged
to the atmosphere through a waste line. An ozone sampling train, set to
pull 150 cc/min, was connected to the waste line. This allaoed the
ozone output of the generator to be checked simultaneously with an
experimental run to verify the actual ozone concentration entering the
reaction tube. The neutral potassium iodide method was used for ozone
sampling, and is described in Appendix A.
Nitrogen flow to the reaction tube was 320 cc/min, and was used as
a carrier gas for the sulfur compounds. The spinning syringe used for
injecting sulfur compounds into the nitrogen stream is shown in Figure 6,
and flow rates from the syringe could be varied by changing orifice sizes
at the syringe outlet. Additional details and calibration data for this
apparatus are given in Appendix A. Sulfur concentrations in the syringe
were varied by diluting known bottled gases to different sulfur/air ratios,
and were checked periodically with the electrolytic titrator and gas
The reaction tube and mixing chamber used for the study were
fabricated from 316 stainless steel. A diagram of this apparatus is
shown by Figure 7. The mixing chamber was designed to provide thorough
interaction between the ozone and sulfur gases prior to the first sampling
1 1/2" Projection of 1/4"
Note: All material 316 S.S.
Figure 7. -- Diagram of Reaction Tube.
port. Volume of the reactor to the first port was 39.4 cc which provided
a residence time of 5.9 seconds based on a total gas flow of 400 cc/min
(80 cc/min ozone-oxygen and 320 cc/min nitrogen) at 250 C. Residence
time to the last port was 61.5 seconds for the same conditions.
For heating purposes, the tube was wrapped with 1 x 1/8-inch
asbestos tape followed by evenly spaced (at 1/4 inch), 18 gauge ni-chrome
wire. This was covered with two more layers of asbestos tape. With a
voltage regulator, a temperature range of ambient to 1250 C was achieved,
and was monitored with a thermometer inserted through a glass fitting (with
a ground-glass joint) at the reactor outlet. The glass piece allowed
visual observation of any deposits formed during the ozone-sulfur gas
Oxidation of the sulfur compounds was monitored with a gas chromato-
graph. Samples were pulled through a heated 1/8-inch Teflon line from the
various sampling ports, depending upon the particular residence time
desired. Exhaust from the reactor was discharged to the atmosphere through
the previously mentioned glass fitting into 1/2-inch polyethylene tubing
which extended outside the mobile laboratory.
NCASI Mobile Laboratory Equipment
General Description of Mobile Laboratory
The NCASI Mobile Laboratory is a modified, 19-foot travel trailer
containing the necessary sampling and analytical equipment needed to
investigate the emission sources in a kraft mill complex. The trailer
facilitates the transportation of equipment from one mill to another and
the location of it in close proximity to the emission source under
investigation. It also provides a suitable working environment while
making a study. The basic equipment includes an all Teflon-stainless
steel heated gas sampling system, combustion furnace, electrolytic
titrators, and a gas-liquid chromatography unit (GLC). These components
are supplemented with gas-fla measuring equipment, calibration systems,
and miscellaneous laboratory supplies, chemicals, tools, etc.
All the equipment in the mobile laboratory was not utilized for the
initial phase of the study since sampling and oxidation of actual stack
gases were not required. As shown by the schematic (Figure 5), only the
electrolytic titrator, gas conditioning box, and GLC unit were used in
the initial phase. Extra bottles of nitrogen and oxygen were available
in the trailer for connection to the test panel also.
Electrolytic Titrator and Combustion Furnace
As mentioned in the previous section of this chapter, a Barton Model
286 Electrolytic Titrator (240) was used to check concentrations of the
sulfur gases prior to entering the reaction tube. The use of this
instrument in the kraft pulping industry is well documented. Thoen,
et al. (51) and Thoen and DeHaas (241) report on the use of a Barton
Titrator to monitor the operation of kraft recovery furnaces. Blosser
and Cooper (242) evaluate it as a method for the continuous monitoring of
various process emission sources in the sulfate pulping operations, and
Thoen, et al. (243) describe its use in the analysis of sulfur compounds
in samples drawn from various points in the kraft process.
The operation of the Barton Titrator, using bromide coulometry,
depends on reactions between oxidizable sulfur gases from the syringe (or
from an emission source) and brocine in the titrating cell. These reactions,
and additional details on the operation of a Barton Titrator, are thoroughly
described in the literature (51, 243). The major interfering compound in
the bromine titration is alpha-pinene, usually emitted at some points in
the sulfate process, but definitely not of concern using knaon bottled
gases. Haoever, interference has not been noticeable in the analysis of
kraft effluent gases (243).
Sensitivities of the Barton Titrator as reported by Thoen, et al.
(243) are 7-10 ppb for hydrogen sulfide, 10-12 ppb for methyl mercaptan,
and 40-50 ppb for dimethyl sulfide. Previous applications of this instrument,
and its use in NCASI investigations, have sho.n it to be dependable, simple
to operate, and requiring minimum maintenance.
A combustion furnace, operating at 17500 F, was used in conjunction
with the titrator for measuring the sulfur gas concentrations. The
purpose of the furnace was to oxidize all forms of reduced sulfur to sulfur
dioxide, and oxidize organic compounds (such as terpenes) that interfere
with the titrator's response to carbon dioxide and water. Although not
necessary while using bottled gases, the furnace was used since it is
included in the normal gas flow pattern of the mobile laboratory, and only
one set of calibration data (that of sulfur dioxide) was needed to determine
sulfur gas concentrations from the syringe.
Gas Conditioning Oven
The gas conditioning oven, or dilution box, is primarily for use in
the sampling and analysis of moisture laden, kraft mill effluent gases.
It is constructed of 3/4-inch asbestos board and houses an electric
heater, thermostat, and air circulation fan. In addition, it contains
valves, flcimeters, and tubing and fittings necessary to control, monitor,
dilute, and distribute the gas floa. For a clearer understanding of the
function and location of the gas conditioning oven in the mobile laboratory,
Figure 8 shcas a simplified flow diagram of the system while sampling stack
gases. By maintaining a temperature of 1000-1100 C in the box, condensation
is prevented in the lines, valves, flaimeters, etc., and by diluting the
flow to the electrolytic titrator, the dewpoint can be reduced below
ambient temperature and condensation prevented in those lines.
For the initial phase of this study with the bottled gases, the
conditioning oven was used primarily to transfer a sample from the reaction
tube, or spinning syringe, to the gas chromatograph through the GLC sampling
valve. The sampling valve has eight ports, and is pneumatically operated
with nitrogen by a solenoid valve. Sample gases from either the reaction
tube, or spinning syringe, continually pass through the valve as does
the GLC carrier gas (nitrogen). Also connected to the sampling valve are
two sectionsof Teflon tubing (sampling loops) having identical volumes.
At any given time, one sampling loop is in series with the sample gas
flow and the other with the carrier gas flcw. When the solenoid valve is
activated, the relative positions of the sampling loops are changed with
respect to the two gas flows. Thus, the GLC carrier gas sweeps the
sample trapped in one loop into the GLC system for analysis.
Sampling loops of three different volumes were used depending upon
concentrations in the sample gas and are listed in Table 6.
GLC SAMPLING LOOPS ..
Loop Length (in) 0.0. (in) Range (ppm)'
Small 2 1/2 1/16 100-2000
Medium 3 1/4 1/8 50-500
Large 22 . .... .. 1/8 . .1-60
Gas Conditioning Oven
Figure 8. -- Mobile Laboratory Flao Diagram.
Gas-Liquid Chromatography System
The use of gas chromatography for the analysis of sulfur gases
from kraft mills is also covered extensively in the literature (12, 39,
244-251). The unit contained in the mobile laboratory is a F & M Model
720 Gas Chromatograph (252) with a Melpar Model 65-34A Flame Photometric
Detector (253). Two analytical columns are utilized in the separation and
analysis of gaseous sulfur compounds. One is a 33-feet x 1/8-inch Teflon
column packed with 30/60 mesh Haloport F (Teflon), and coated with 5
percent polyphenyl ether (5-ring). The other column is constructed of
identical materials, but is only five feet long. The longer column is for
analyzing hydrogen sulfide, sulfur dioxide, methyl mercaptan and dimethyl
sulfide, while the five-foot column is for dimethyl sulfide and disulfide.
A valve allows the selection of either column, and samples are introduced
to the column selected through the GLC sampling valve previously
described in the gas conditioning oven. Both columns operate at a
temperature of 500 C with the carrier gas (nitrogen) set at 50 cc/min.
Under these conditions the elution time of each column is as shown in
GAS CHROMATOGRAPH COLUMN ELUTION TIMES
Sulfur Compound Long 'Cdl'umn Short Column
Hydrogen Sulfide 2 minutes
Sulfur Dioxide 3 minutes
Methyl Mercaptan 4-5 minutes
Dimethyl Sulfide 7-8 minutes 48 seconds
Dimethyl Dis.ulfi.de. -.- ..... .. 4-5 minutes.......
From either column, the sample enters the Melpar Flame Photometric
Detector where it is mixed with oxygen, hydrogen and nitrogen and burned.
The burning sulfur emits a blue light which is detected by a photomultiplier
tube. The current output is proportional to the light produced, and is
displayed on a strip chart recorder. With the use of calibration curves,
the sulfur gas concentration is determined. Additional details on the
flame photometric detector are covered by Brody and Chaney (254).
Calibration of the GLC unit is discussed in Appendix A.
EXPERIMENTS AND RESULTS WITH BOTTLED SULFUR GASES
To study the effectiveness of ozone in oxidizing the malodorous
sulfur gases of interest, measured amounts of ozone and individual sulfur
gases were injected into the heated reaction tube and allowed to react
for a specified period of time. The products of the reaction were
analyzed by the GLC unit, and the percent reduction in the malodorous
sulfur gas noted. Specific details on operation of the test panel are
covered in Appendix A, and describe the procedures used to obtain.
specific ozone to sulfur gas ratios, to change operating conditions
(ratios, reaction times, temperature), and other necessary operations to
run the experiments.
As discussed in Chapter IV (and Appendix A), reaction times in the
heated reaction tube were controlled by using different sampling ports
along the tube to withdraw samples for the GLC unit. Space times, or
residence times, to each sampling port along the reactor are listed in
Table 15 (Appendix A) for the various temperatures used during the study.
Since a quench was not used at the sampling port to stop the oxidation
process, any residual ozone in the sample would continue to oxidize the
sulfur gas while in the transfer line to the GLC unit, thus resulting in
an actual reaction time longer than those listed in Table 15.
Due to the fact that the transfer line was maintained at the same
temperature as the reaction tube, the assumption was made that any reaction
in the line would proceed similar to that in the reactor, and that the
additional reaction time would be based only on flow and volume of the
transfer line. At a flow of 200 cc/min through the transfer line,
approximately four seconds of additional reaction time would have to be
included. Table 8 reflects this additional transfer time, and lists the
total reaction time associated with each sampling port.
TOTAL REACTION TIME FOR OZONE OXIDATION
Sampling Total Reaction Time (sec)
Port 250C 380C 55uC 76C 1 25TC
1 9.9 9.7 9.4 9.0 8.4
2 21.1 20.4 19.5 18.6 16.8
3 32.4 31.2 29.8 24.3 25.3
4 43.4 41.8 39.9 37.7 33.6
5 54.5 52.4 49.8 47.1 41.8
6 65.5 63.0 59.9 56.5 50.0
The experimental procedures used for each of the four gases were
basically the same, the only difference being the concentration ratios
evaluated for some of the experiments. As discussed in Chapter III, a
majority of the laboratory investigations concerning ozone oxidation of
malodorous sulfur compounds have dealt only with hydrogen sulfide. The
range of conditions studied has been somewhat limited, thus restricting
the practical application of the investigation's results. Based upon
the data of Table 2 (Chapter II), which summarize k-aft emission sources,
stack temperatures, and sulfur gas concentrations, a series of experi-
mental tests were established for this study to cover a wide range of
conditions representative of the kraft pulping industry. The following
sections describe the experimental conditions established for each of
the four malodorous sulfur gases, and the resulting data obtained at
Cadle and Ledford (11) and Hales (12) agreed that the true
stoichiometry of the ozone oxidation of hydrogen sulfide is represented
by Equation 3.4, shown here again for convenience.
H2S + 0 ---- H20 + SO2 (3.4)
Theoretically then, an ozone to hydrogen sulfide ratio of 1.0 would
achieve complete oxidation to sulfur dioxide. However, studies by Gregor
and Martin (10) indicated that a ratio of 9.0 was necessary for complete
oxidation of all the initial hydrogen sulfide, even for reaction times cf
16 hours. Regardless of this, experimental conditions were established
to study low concentration ratios and reduced reaction times since
extremely high ratios and large holding chambers would be impractical from
a cost viewpoint. Table 9 lists the conditions established to study the
ozone oxidation of hydrogen sulfide.
Experimental data obtained on hydrogen sulfide oxidation are listed
in Tables 18, 19, 20 and 21 (Appendix B), and show the remaining hydrogen
sulfide at each sampling port for specific concentration ratios and
temperatures. As noted in the data tables, the exact concentration ratios
HYDROGEN SULFIDE EXPERIMENTAL CONDITIONS
Initial H2S 03/H2S Temperatures
Concentrations (ppm) Ratios (0C)
10 0.5, 1.0, 2.0, 5.0 38, 76, 125
50 11" "I1
250 11 n 11 II II II
500 ii i" "I
1000 I" "
2000 1" "I
varied slightly from those proposed in Table 9 due to variations in ozone
generator output and sulfur concentrations in the spinning syringe (as
discussed in Appendix A -- Test Stand Operating Procedures).
Percent reduction of the initial hydrogen sulfide concentration for
each temperature, sampling point, and concentration ratio (236 data points)
is listed in Table 22 (Appendix B). Data from this table were used for
a multiple regression analysis of the hydrogen sulfide oxidation. The
data were fit to six different models (equations), using combinations of
the variables ozone-hydrogen sulfide ratio (03), temperature (Te),
residence time (Time), ratio squared (03 2), ratio-Te (03 x Te), ratio-
time (0 x Time), and temperature-time (Te x Time). The object was to
find a relatively simple equation with a high coefficient of determination
(R2), which indicates the percent of data variance accounted for by the
particular equation, thus giving a high multiple correlation coefficient (R).
The three equations resulting in a high R2 value and with minimum
variables are as follows:
y = B0 + B103 + B203 (5.1)
y = BO + B 0 + B2Te + B303 (5.2)
and y = B0 + B 10 + B2Te + B Time + B4032 (5.3)
where y = percent reduction of initial hydrogen sulfide,
BO, B, B2, . o = regression coefficients
0 = ozone-hydrogen sulfide ratio,
Te = temperature (oK),
Time = reaction time (seconds).
The R2 values for the above equations are 0.7958, 0.8261, and 0.8422
respectively. As expected with these high coefficients, the "F"
statistic for each equation was highly significant (*.) indicating a
relatively small amount of variance in the data due to error.
Using each of the above equations, estimated values for percent
reduction of hydrogen sulfide were calculated for concentration ratios
of 0.5, 1.0, 2.0, and 3.0, reaction times of 10, 30, and 60 seconds, and
for a temperature of 820 C. The temperature of 820 C (1800 F) is
typical of various kraft emission sources, and residence times of 10, 30,
or 60 seconds could be obtained for many sources to insure good oxidation.
Estimates for the above conditions, their standard error, regression
coefficients, and other data on each equation are listed in Tables 23, 24,
and 25 of Appendix B.
Using the data of Table 25, which estimate percent reduction using
Equation 5.3, the effects of increased concentration ratios and reaction
times are shown by Figure 9 for 820 C. This equation estimates a reduction
of approximately 44 percent for 60 seconds'reaction at a concentration
ratio of 1.0, whereas the data of Gregor and Martin, (Figure 1, Chapter 3)