United States
Environmental Protection
Agency
Robert S Kerr Environmental Research EPA 600 2-79-080
Laboratory April 1979
Ada OK 74820
Research and Development
Development of
Treatment and
Control
Technology for
Refractory
Petrochemical
Wastes
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution-sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service., Springfield, Virginia 22161.
-------
EPA-600/2-79-080
April 1979
DEVELOPMENT OF TREATMENT AND CONTROL TECHNOLOGY
FOR REFRACTORY PETROCHEMICAL WASTES
by
John H. Coco, Ellas Klein, Donna Rowland,
James H. Mayes, William A. Myers, Earl Pratz,
Clyde J. Romero, and Floyd H. Yocum
Gulf South Research Institute
New Orleans, Louisiana 70186
Grant No. S800773
Project Officers
Luther F. Mayhue and Thomas E. Short
Robert S. Kerr Environmental Research Laboratory
Ada, Oklahoma 74820
ROBERT S. KERR ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ADA, OKLAHOMA 74820
-------
DISCLAIMER
This report has been reviewed by the Robert S. Kerr Environmental
Research Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily reflect
the views and policies of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
ii
-------
FOREWORD
The Environmental Protection Agency was established to coordinate
administration of the major Federal programs designed to protect the
quality of our environment.
An important part of the agency's effort involves the search for
information about environmental problems, management techniques and new
technologies through which optimum use of the nation's land and water
resources can be assured and the threat pollution poses to the welfare
of the American people can be minimized.
EPA's Office of Research and Development conducts this search
through a nationwide network of research facilities.
As one of these facilities, the Robert S. Kerr Environmental
Research Laboratory is responsible for the management of programs to:
(a) investigate the nature, transport, fate and management of pollutants
in groundwater; (b) develop and demonstrate methods for treating waste-
waters with soil and other natural systems; (c) develop and demonstrate
pollution control technologies for irrigation return flows, (d) develop
and demonstrate pollution control technologies for animal production
wastes; (e) develop and demonstrate technologies to prevent, control
or abate pollution from the petroleum refining and petrochemical in-
dustries, and (f) develop and demonstrate technologies to manage pol-
lution resulting from combinations of industrial wastewaters or indus-
trial/municipal wastewaters.
This report contributes to the knowledge essential if the EPA
is to meet the requirements of environmental laws that it establish
and enforce pollution control standards which are reasonable, cost
effective and provide adequate protection for the American people.
William C. Galegar, Director
Robert S. Kerr Environmental
Research Laboratory
iii
-------
ABSTRACT
This summary document presents the results of research and development
work pertaining to the treatment of biorefractory organic pollutants emanat-
ing from petrochemical processing plants. Specifically, it covers applica-
tion of the unit operations of (1) carbon adsorption, (2) steam stripping,
(3) solvent extraction, and (4) ozonation to a number of process effluents
from halogenated hydrocarbon, isocyanate, acetylene, and styrene manufacturing
industries.
In the carbon adsorption study, different commercial activated carbons
were evaluated via Freundlich isotherms for streams containing C, and C2
chlorinated hydrocarbons in quantities less than 1000 mg/liter. Subsequent
field trials were conducted to yield breakthrough data.
Steam stripping evaluations were made for process effluents containing
chlorinated hydrocarbons and aromatic hydrocarbons. This unit operation
removed up to 75 percent of the TOC present in the raw process effluent. Data
for a commercial design were obtained from a pilot demonstration unit.
Commercial design specifications were prepared for installation of a unit at
one of the participating companies.
Solvent extraction evaluations were made for process effluents contain-
ing chlorinated hydrocarbons and aromatic hydrocarbons. The best solvent for
maximum removal of organics with minimum TOC residuals in the extracted
effluent was a straight-chain paraffin hydrocarbon in the Clf. to C..- range.
Ozonation evaluations were initiated to determine if chemical oxidation
with ozone was feasible as a pretreatment method for biorefractory industrial
wastewater. Complete oxidation with ozone would be uneconomical, but the
partial oxidation of refractory wastes to a biodegradable form with ozone
could be economically competitive. Ozonation was proved to be an effective
method of pretreating wastewaters from the manufacture of toluene diisocyanate
ethylene glycol, styrene monomer, and ethylene dichloride. Batch biooxidation
studies showed ozonation improved the biotreatability of these industrial
wastes.
In addition to these unit operation evaluations, a study was undertaken
to develop an activated carbon from a by-product soot which results from the
manufacture of acetylene. Quantities of raw soot were dried, pretreated,
pelletized, and activated to yield a product with commercial potential which
had about 80 percent as much adsorption capacity as a similar commercial
product.
-------
CONTENTS
Foreword iii
Abstract iv
Figures • vii
Tables x
Acknowledgment xvi
1. Introduction 1
2. Conclusions 2
3. Recommendations 3
4. Summary 4
Process evaluations and point source identification 4
Source of effluents 14
Individual process effluent characterization 15
Effluent treatment 19
5. Activated Carbon Adsorption 26
Laboratory program 27
Demonstration unit program 34
Process effluent stream 221A 36
Process effluent 041A 41
Process effluent 081D and 081A 50
Economic discussion 55
6. Monochem Soot Carbon 56
Preliminary evaluation for application as a commercial
activated carbon 57
Laboratory and field testing 61
Development program for improving activity and mechanical
strength 65
Laboratory evaluation of binders for soot carbon gran-
ulation and activation 66
Compounding, pelletizing, and activating Monochem soot
carbon 75
Optimization of compounding and activating Monochem soot
carbon 79
Pilot work on activated carbon density, activation and
regeneration 85
Comparative isotherm and breakthrough data 93
7. Solvent Extraction 102
Laboratory development on solvents selected for evaluation
of the process effluents 103
Demonstration unit design 107
Demonstration unit operations at plant site Ill
8. Steam Stripping 119
Laboratory bench-scale stripping column 121
Steam stripping trials 125
-------
9. Ozonation 141
Introduction and principles of ozonation 141
Ob j actives 141
Mathematical model development 142
Laboratory study 146
Mass-transfer study. ' 147
Pilot studies 153
Economic evaluation 180
References 182
Appendix. 184
vi
-------
FIGURES
Number Page
1 Manufacture of ethylene dichloride via direct chlorination of
ethylene 5
2 Ethylene dichloride and vinyl chloride monomer manufactured with
HCl recovery 5
3 Manufacture of ethylene dichloride via oxychlorination of ethylene
utilizing hydrogen chloride 8
4 Manufacture of perchloroethylene and trichloroethylene 8
5 Ethylbenzene and styrene manufacture 12
>
6 Typical breakthrough curves 31
7 Laboratory continuous carbon adsorption apparatus 31
8 EDC adsorption isotherm for monochem activated carbon (stream
221A) 60
9 Adsorption isotherm for phenol on regenerated carbon 8.4
10 Laboratory pellet mill 8.4
11 Iodine number results from large oven optimization 94
12 Continuous-flow apparatus 98
13 Solvent extraction pilot plant (operational concept) 110
14 Model of pilot steam stripper (GSRI - New Orleans) 124
15 Material balance, process and instrument flowsheet 137
16 Liquid phase concentration of the gas-phase reactant 143
17 Comparison of ideal and experimental C-curve for a pulse tracer
input for a stirred tank reactor 147
18 Ozone concentration profile for various turbine speeds at
12'-15°C 149
vii
-------
19 Ozone concentration profile for various turbine speeds at
22°-25°C
20 Ozone concentration profile for various gas feed rates at
12°-15°C 150
21 Volumetric mass-transfer coefficient at various gas feed rates
as a function of power input 150
22 Volumetric mass transfer coefficient corrected for superficial
gas velocity as a function of power input 152
23 Volumetric mass transfer coefficient at various turbine speeds
as a function of superficial gas velocity 152
24 Equipment arrangement in ozone pilot plant building 153
25 Tubular reactor schematic 155
26 Ozone pilot plant schematic 155
27 Standard reactor configuration 156
28 Ozonation of toluene diisocyanate wastewater as a function of
ozone transferred 161
29 Ozonation of toluene diisocyanate as a function of time 161
30 Ozonation of polyol wastewater as a function of ozone transferred 154
31 Ozonation of polyol wastewater as a function of time 154
32 Ozonation of styrene wastewater as a function of ozone
transferred
33 Ozonation of styrene wastewater as a function of time 167
34 Ozonation of stripped ethylene dichloride wastewater as a
function of ozone transferred 168
35 Ozonation of stripped ethylene dichloride wastewater as a function
of t±me 168
36 Determination of stoichiometric ratio for the ozonation of
benzoic acid 2.73
37 Effect of temperature on the ozonation of benzoic acid in aqueous
solution 173
38 Effect of pH on the ozonation of benzoic acid in aqueous solution 174
39 Determination of stoichiometric ratio for the ozonation of
benzaldehyde 174
viii
-------
40 Effect of temperature on the ozonation of benzaldehyde in
aqueous solution 177
41 Effect of pH on the ozonation of benzaldehyde in aqueous
solution 177
'**'',
42 Ozonation in styrene in aqueous solution and the effect of
airstripping on the system 178
43 Application of the model for benzaldehyde from the ozonation
of styrene in aqueous solution 178
ix
-------
TABLES
Number PaSe
1 Manufacturing Processes 4
2 Ethylene Bichloride Production 6
3 Production of Vinyl Chloride 9
4 Trichloroethylene Production 10
5 Styrene Monomer Production. 13
6 Coded Process Wastewater Streams 15
7 Soluble Organic Stream Characterization 16
8 Soluble Organic Stream Characterization 17
9 Stream Characterizations 18
10 Air Stripping and BOB5 Results 20
11 Percent Removal of Volatile Organics by Air Stripping 21
12 Feasibility of Solvent Extraction 23
13 Applicability of Steam Stripping 24
14 Adsorption Capacity of Activated Carbons for Removal of Ethylene
Bichloride 28
15 Adsorption Capacity of Activated Carbons for Removal of TOG 28
16 Regeneration Efficiency - EDC 33
17 Regeneration Efficiency - TC 33
18 Variation of Regeneration Time 34
19 Effects of Flow Rate on Adsorption Capacity for Stream 221A 37
20 Analysis of Adsorption Bemonstration Unit Operation by Individual
Organic Component for Stream 221A 38
x
-------
Number Page
21 Regeneration of Spent Activated Carbon - Run No. 1. . ., 40
22 Regeneration of Activated Carbon - No. 2 40
23 Adsorption Capacity Data from Breakthrough Study of Stream 22LA.. 42
24 Study of the Adsorption Capacity of Regenerated Westvaco-WVG
Comparison of the Effluent and Influent from Stream 221A 43
25 Summary of Carbon Adsorption Field Trials on Stream 041A 45
26 Breakthrough Data with Filtrasorb-400 (Stream 041A) 47
27 Breakthrough Data with Witco 718 (Stream 041A) 47
28 Breakthrough Data with Westvaco (Stream 041A) 48
29 Breakthrough Data with WVG (Stream 041A) 49
\
30 Breakthrough Data with WVG (Stream 041A) 49
31 Summary of Carbon Adsorption Field Trials on Stream 041A 51
32 Adsorption Capacity Data from Breakthrough Study of Stream 081A
(Westvaco WVG #1 Activated Carbon) 51
33 Regeneration of Westvaco WVG //I Activated Carbon 52
34 Comparison of All Successful Runs for the Carbon Adsorption Unit
At Saturation: Stream 081D 54
35 Adsorption Data on the Individual Components for the Carbon
Adsorption Unit: Stream 081D 54
36 Steam Regeneration Data for the Carbon Adsorption Unit Stream
081D 55
37 Activation Studies - I 59
38 Thermogravimetric Measurements (Programmed Rate of Heating:
20°C/min) 60
39 Adsorptivity of Monochem and Commercial Carbons (Stream 221A).... 63
40 Summary, of Carbon Adsorption Field Trials on Stream 041A 63
41 Shirco Dried Soot Carbon 66
xi
-------
Number
42 Material Selection for Binder Studies 67
43 A Tabulation of the Data for the Pellet Pressed Carbon Samples... 69
44 A Tabulation of the Data for the Extruded Carbon Samples Prepared
From Semi-Dry Carbon 71
45 A Tabulation of the Data for the Samples Prepared from Wet
Carbon 73
46 A Description of Samples Prepared from Different Types of
Methocel 74
47 Effect of Inorganic Additives on the Activity of Samples Methocel
HB Binder 75
48 A Comparison of the Activity From Low Temperature Heating with
That from Two-Stage Heating 76
49 Determination of the Best Combination of Low and High Temperature
Heating Conditions 77
50 Comparison of Activity for Methocel HB Samples Activated in
Nitrogen and in Steam 80
51 Activation by Two Different Gases: Nitrogen and Steam 80
52 Activity Measured for Different Lots of Methocel HB 81
53 Activity Determination by the Molasses Number 81
54 Results from Freundlich Adsorption Isotherms: Nitrogen Gas Used
in Activation of Carbon. 82
55 Results for Reactivation of Carbon as Measured by Phenol
Adsorption 83
56 Results from Freundlich Adsorption Isotherms: Steam Used in
Activation of Carbon 85
57 Comparison of the Apparent Density of Various Carbons 86
58 Activation Conditions for Optimization of Small Furnace 89
59 Results from the Analysis of Carbon Weight Loss During
Activation 88
60 Results from the Compression Tests • 91
xii
-------
Number Page
61 Relative Iodine Numbers: Percent of Commercial Carbon
(Westvaco) Activity 92
62 Determination of Reproducibility of Activation Results 92
63 Results from Compression Tests and Iodine Numbers: Samples
Activated in Large Furnace 94
64 Tabulation of Multicomponent Isotherm Results (Steam Activation
Gas) 9f
65 Continuous-Flow Data 99
66 Continuous-Flow Data on Activation and Regeneration Procedure....101
67 Process Effluents 103
68 Initial Evaluation of Stream 161A 104
69 Extraction of Stream 161A with C- -C12 Solvent; and Kerosene 104
70 Initial Evaluation of Stream 221A 105
71 Organic Components in Stream 221A 106
72 Solvent Extraction of Waste Stream 221A 106
73 Initial Evaluation of Stream 231A.1 107
74 Selected Solvent Extraction of Stream 231A 108
75 Properties of Kerosene and C,Q-C.jo Solvent Ill
76 Chlorinated Organics Removed with a Kerosene-Diesel Solvent 112
77 Efficiency of Pilot Steam Stripper for Removal of Chlorinated
Organics from Extract 113
78 Chlorinated Organics Removed with a C10~G12 Paraffin Solvent 114
79 Determination of C10-C12 Paraffin Solvent Solubility 116
80 Coded Process Wastewater Streams 121
81 Stream Characterizations 122
xiii
-------
Number J.aS.e-
82 Soluble Organic Stream Characterization I23
83 Steam Stripping of Stream 221A 127
84 Steam Stripping of Stream 221A Component Analysis of Feed Overhead
and Bottoms
85 Characterization of Stream 011C Process Effluent 128
86 Pilot Plant Evaluation for Steam Stripping of Steam 011C 130
87 Analysis of Stream 011C 132
88 Characterization of Stream 161A Process Effluent 133
89 Tests Run on Stream 161A 134
90 Stream Characterization of Stream 351A 135
91 Steam Stripping of Stream 351A 136
92 Design Parameters Commercial Unit for Stream 011C 138
93 Process Wastestreams Used for the Ozonation Study With Their
Range of TOG Concentrations Found 158
94 Characterization of Industrial Wastewaters Ozonated (Average
Data; Concentrations in mg/liter) 158
95 Ozonation of Toluene Diisocyanate Wastewater in a Tubular Reactor
With Nozzle Dispersion 159
96 Ozonation of Toluene Diisocyanate Wastewater in a Tubular Reactor
With Static Mixers 160
97 Summary of Batch Study Results for Raw and Ozonated TDI
Wastewater, 162
98 Summary of Batch Study Results for Raw and Ozonated Polyol
Wastewater 165
99 Comparison of Raw and Ozonated Polyol Wastewater 163
100 Summary of Batch Study Results for Stripper BCD Wastewater Before
and After Ozonation 170
101 Summary of Ozonation as a Pretreatment for Industrial
Wastewaters 171
xiv
-------
Number Page
102 Summary of the Ozonation Reaction Study for Benzole Acid,
Benzaldehyde and Styrene in Aqueous Solutions 175
103 Summary of Five-Day Biological Oxygen Demand to Total Organic
Carbon Ratio Data for Reaction Solutions 179
104 Capital and Operating Cost for Ozone Pretreatment 180
105 Cost of Pretreatment 181
xv
-------
ACKNOWLEDGMENT
This research was supported by the U.S. Environmental Protection Agency
with guidance from the Robert S. Kerr Environmental Research Laboratory in
Ada, Oklahoma. The principal investigators of this project were Dr. Elias
Klein and Dr. James H. Mayes. Technical and staff personnel who made valuable
contributions toward the completion of this project were Denzel A. Brown,
Dr. John H. Coco, Michael Curtis, Dr. A.J. Englande, Jr., Kevin Hester, Donna
Howland, Kerry LaBauve, Cynthia Littleton, Marlene Mock, Charles Major,
William A. Myers, Earl Pratz, Laurie Rando, Roy Rando, Clyde J. Romero, and
Dr. Floyd H. Yocum. Finally, this research would not have been possible
without the cooperation of the participating process industries and their
associated technical personnel.
xvi
-------
SECTION I
INTRODUCTION
In 1969, the Water Quality Administration authorized a study to evaluate
all point sources of effluents discharging to the Mississippi River from St.
Francisville to Venice, Louisiana. The study period included the last two
quarters of 1969 and all of fiscal 1970 and 1971. This study focused attention
on the need to develop best practicable treatment for the more difficult-to-
treat effluents from several of the major chemical plants in the area.
The initial concept of this study was to utilize the combined efforts of
the Louisiana State Science Foundation, participating industrial companies,
Gulf South Research Institute (GSRI), and university assistance to develop
technology for biological and physical-chemical treatment of industrial
effluents to achieve desired goals of reducing contaminant loadings. The major
objectives of the project were...
...to develop and demonstrate effluent treatment procedures for the reduction
of refractory petrochemical effluents with presently available water treatment
technology.
...to develop an economical source of activated carbon from a waste carbon by-
product stream.
...to analyze present and proj ected loadings, treatment levels, and treatment
costs.
Both laboratory bench and in-plant pilot scale equipment were used. The
major processes investigated were: carbon adsorption with regeneration
evaluation; solvent extraction; biological treatment (aerobic and anaerobic);
ozone oxidation; steam stripping; and compounding and activation of a soot
carbon by-product. Parallel with these efforts was work to develop an econom-
ical and commercial source of activated carbon from a waste soot carbon by-
product. Research was also directed toward development of a better and more
economical method to regenerate spent activated carbon. These regeneration
studies were made on the developed carbon and on commercial carbons. To meet
the objectives of this program, nine project work areas were identified and
described:
1. Process evaluations and point source identification.
2. Waste characterization.
3. Bench-scale process feasibility studies.
4. Carbon preparation, activation, and regeneration procedures.
5. Pilot plant studies.
6. Analytical services.
7. Analysis of loadings, treatment levels, and treatment costs.
8. Supporting services and studies.
9. Evaluation of results and recommendations.
-------
SECTION 2
CONCLUSIONS
Several different unit operations were evaluated on a total of 21 indi-
vidual plant process effluents for their effectiveness in removing contami-
nants present in the wastewater. Steam stripping evaluations were made on
process effluents containing chlorinated hydrocarbons and aromatic hydro-
carbons. This unit operation removed up to 99 percent of the chlorinated
hydrocarbons and up to 75 percent of the total organic carbon (TOG). Residual
TOC was in the form of chloral, which includes aldehyde and hydroxy-type
chlorinated organics resulting from the oxychlor process. This type of
treatment process would cost $0.0005 per liter of wastewater or about $0.05
per kilogram of product.
Solvent extraction evaluations were made on process effluents containing
chlorinated hydrocarbons and aromatic amines. Carbon disulfide, tetrachloro-
ethane, pentane, cyclohexane, and freon had little effect on removing chlori-
nated hydrocarbons. Removal of aromatic amines by tetrachloroethane, pentane,
cyclohexane, octane, dodecane, dichlorobutane and tetrachlorobutane never
exceeded 50 percent. A C1f.-C19 straight-chain paraffin was found to be effec-
tive in removing the chlorinated organics. This solvent also exhibited low
residual solubility (15 mg/liter) in the extracted water. This type of treat-
ment process would cost $0.0005 per liter of water treated and $0.066 per
kilogram of ethylene dichloride produced.
Activated carbon adsorption evaluations were made on process effluents
containing chlorinated hydrocarbons, aromatic amines, and aromatic hydrocarbons.
Four commercial activated carbons (Calgon, Barneby-Cheney, Witco, and Westvaco)
were found to be effective in removing chlorinated organics and aromatics from
wastewater. Spent carbon was regenerated with steam. In the resulting read-
sorption tests, the regenerated carbon exhibited over 90 percent of fresh
carbon activity.
A soot carbon by-product was evaluated as a potential activated carbon.
From this study, a carbon was developed which had about 80 percent of the
activity of commercial samples. Sufficient material was produced to conduct
isotherm and breakthrough tests on several different contaminated effluents.
Ozonation was shown to be an effective method of pretreating biorefrac-
tory wastewaters. The biorefractory compounds in toluene diisocyanate,
ethylene glycol, and ethylene dichloride contaminanted process wastewaters
were oxidized to a biodegradable form. Ozonation was found to improve the
biotreatability of water containing styrene by transforming it to a compound
such as benzoic acid. The ozonation of styrene reaction products, benzaldehyde,
and benzoic acid were affected by both solution pH and temperature.
-------
SECTION 3
RECOMMENDATIONS
Each of the unit operations tested under this grant program showed a
capability for removing wastewater contaminants to a greater or lesser degree.
The work under this grant should be expanded to cover the following aspects
of reducing contaminants in industrial process effluents.
The ozonation of industrial effluents showed that refractive organics
can be converted to a biodegradable state. Further development work should
be carried out to establish dosages and optimize process conditions.
Ozonation changed the basic characteristics of the organic components
and could have an effect on carbon adsorption activities. Further work
should be considered for utilizing ozonation as a pretreatment process.
The project concentrated on evaluating each separate unit operation and
its effect on reducing the contamination in the process effluents. Very
little consideration was given to evaluating the effect of using two of these
processes in series. The possibility of utilizing such a series of unit
operations to treat some of the process effluents should be investigated.
As a result, the operation and capital cost for each treatment step could be
optimized.
Although an activated carbon was developed from the Monochem soot carbon,
neither the resulting material nor the manufacturing procedure was optimized.
Additional studies should be undertaken to follow up this development.
Additional joint federal, state, and industrial projects would be of value
to achieve optimized, cost-effective treatment systems for all the various
industrial process effluents.
-------
SECTION 4
PROGRAM SUMMARY
PROCESS EVALUATIONS AND POINT SOURCE IDENTIFICATION
Among the 20 point source samples studied, several originated from the
same process operation. In addition, several of the companies in the
program manufactured the final product under similar processing conditions.
The manufacturing processes investigated are listed in Table 1.
_ TABLE 1. MANUFACTURING PROCESSES _
Product Process
Ethylene dichloride via direct chloriixation
Ethylene dichloride via oxychlorination
Perchloroethylene and
1,1,1-trichloroethylene via chlorination of ethylene dichloride
Ethyl benzene via alkylation of benzene
Toluene diamine via polymerization of dinitrotoluene
Polyols via polymerization of ethylene oxide
Acetylene via combustion of methane
Methylcellulose via propylene oxide
Ethylene dichloride (1,2-dichloroethane) via chlorination
Ethylene dichloride (1,2-dichloroethane) is the largest volume organic
chlorine derivative. Domestic production in 1974 was over 5 billion kg.
The greatest single outlet for ethylene dichloride (EDC) is for the
production of vinyl chloride. There are 16 major producers of ethylene
dichloride; of these, 8 are located on the Louisiana Gulf Coast. These
8 companies produce slightly more than 55 percent of the total production
(Table 2).
Production of EDC from ethylene and chlorine (Fig. 1) can be represented
by the following equation:
C1CH2CH2C1 (1)
-------
Ethylene-
Chlorine
O
wash
ustic
wash
1
3
— It 1
— »
k
1
Light
ends
C
•H
CO
•H
C
•H
z>
> '
Ethylene
dichloride
Tars
Effluent
Figure 1. Manufacture of ethylene dichloride via direct chlorination of
ethylene.
Light ends HC1 to recovery
and purification '
Light ends
Distill. Residue
Stripper Refiner Storage tank
Figure 2.
Ethylene dichloride and vinyl chloride monomer manufactured
with HC1 recovery.
-------
TABLE 2. ETHYLENE DICHLORIDE PRODUCTION
Capacity
Producer (million kg/yr)
Allied, Baton Rouge, Louisiana 270
American Chemical, Long Beach, California 105
Conoco, Lake Charles, Louisiana 545
Diamond Shamrock, Deer Park, Texas 45
Dow, Freeport, Texas 590
Dow, Oyster Creek, Texas 500
Dow, Plaquemine, Louisiana 525
Ethyl, Baton Rouge, Louisiana 225
Ethyl, Houston, Texas 110
Goodrich, Calvert City, Kentucky 110
PPG, Guayanilla, Puerto Rico 405
Shell, Deer Park, Texas 315
Union Carbide, Taft, Louisiana 635
Union Carbide, Texas City, Texas 65
Vulcan, Geismar, Louisiana 65
Total 4,630
*Chemical marketing Reporter, vol. 202, IMo. 21, Nov. 21), LV12.
At fairly low temperatures (below 25°C), this addition reaction proceeds
to almost 100% conversion. As the temperature is raised, 1,1,2-trichloroethane
formation becomes significant. Chlorine can be added in either liquid or
vapor phase with or without the aid of a catalyst. Catalysts suggested for
this reaction include antimony chloride and ferric chloride. The reaction
is generally carried out in an ethylene dichloride solvent which has a
catalytic effect in conjunction with the metal salts.
Since this reaction is exothermic, recycle cooling is necessary. One
method of cooling is withdrawal of the product in the vapor form, condensing
and refluxing. The by-product side effects are controlled by several
processing techniques. The chlorine-to-ethylene ratio, the temperature,
and the pressure are regulated. In general, the reactor is maintained in
the liquid phase, and ethylene and chlorine are injected via sparge nozzles.
o
By carrying out the reaction at higher pressures (5.25 kg/cm ), the
reaction rate increases perceptably. Side reactions, notably the production
of 1,1,2-trichloroethane, also increase. Since there is a market for this
by-product as a solvent, faster reaction rates with separation and sale of
this product are favored.
Oxychlorination
With the advent of the market development of vinyl chloride came the
problem of disposing of the hydrogen chloride by-product from the dehydro-
chlorination of ethylene dichloride (see Figure 2).
-------
Cl
C + 2 HC1 (2)
A process was developed to recycle the hydrogen chloride to make more
ethylene dichloride. The technique became known as the oxychlor process.
Cl Cl
H-C - C-H + H?0 (3)
H H H
The next logical step in the development of this technology was the incor-
poration of reactions (1) and (3) .
.4- ^
+ 2 HC1 + C
H
H
= r _
9
H
air
Cl
H-C -
H
Cl
C-H + 21
H
Cl- + 2 HC1 + C = C —> H-C - C-H + 2HC1 + H-0 (4)
H H H H
The system to accomplish this process sequence does not vary signifi-
cantly from the initial chlorine and ethylene reactor system. In a typical
reactor, mole ratios^ of 1:3.3:1.6 oxygen:hydrogen:chloride and ethylene are
contacted in a reactor containing aluminum impregnated with copper chloride.
Temperatures are in the 300°C range at atmospheric pressure. The reaction
products contain approximately 16 mole percent EDC, 72 mole percent water,
and 9 mole percent hydrogen chloride. Overall yield of EDC is about 92
percent on ethylene and 98 mole percent on hydrogen chloride. A typical
process flow is shown in Figure 3.
Vinyl Chloride from Acetylene
A traditional route to the manufacture of vinyl chloride is the reaction
of hydrogen chloride with acetylene.
H H H Cl
HC1 + C = C » C = C (5)
H H
The main by-products of this reaction are 2-chloropropene and some
residual tar from the purification operation. The catalyst is a copper
based material. The spent catalyst is changed infrequently, with the
residual going to landfill. The chloropropene can be recycled to an oxychlor
process with chlorine and ethylene as the principal feed materials. This
process is rapidly losing its commercial importance bacause of the reduction
in acetylene capacity. Major producers of vinyl chloride are listed in
Table 3.
-------
Vent
Air
HC1 from viny
chloride plan
Ethylene
^
1
s
t
f
•A
^
X
7
~^
^ \
~n
f \
^ Wafer wash
<— a D
»«Caustic wash
^-a D
i
i — Q
K
^
H
J
s
2k
r
^
^"in
"•
>•
p^
Mm
V
\
K
^•H
f7
r
-------
TABLE 3.. PRODUCTION OF VINYL CHLORIDE*
Capacity
Producer (million kg/yr)
Air Products, Calvert City, Kentucky 65
Air Products, Pensacola, Florida 20
American Chemical, Long Beach, California 65
Borden, Illiopolis, Illinois 60
Borden, Leominster, Massachusetts 80
Conoco, Oklahoma City, Oklahoma 110
Diamond Shamrock, Deer Park, Texas 120
Diamond Shamrock, Delaware City, Delaware 45
Ethyl, Baton Rouge, Louisiana 80
Firestone, Perryville, Maryland 105
Firestone, Pottstown, Pennsylvania 120
General Tire, Ashtabula, Ohio 55
General Tire, Point Pleasant, West Virginia 20
Goodrich, Avon Lake, Ohio 60
Goodrich, Henry, Illinois 60
Goodrich, Long Beach, California 60
Goodrich, Louisville, Kentucky 155
Goodrich, Pedricktown, New Jersey 75
Goodyear, Niagara Falls, New York 45
Goodyear, Plaquemine, Louisiana 45
Great American Chemical, Fitchburg, Massachusetts 20
Hooker, Burlington, New Jersey 80
Hooker, Hicksville, New York 7
Keysor-Century, Saugus, California 16
Monsanto, Springfield, Massachusetts 30
National Starch, Meredosia, Illinois 5
Olin, Assonet, Massachusetts 65
Pantasote, Passaic, New Jersey 25
Pantasote, Point Pleasant, West Virginia 40
Robintech, Plainesville, Ohio 110
Stauffer, Delaware City, Delaware 80
Tenneco, Burlington, New Jersey 75
Tenneco, Flemington, New Jersey 30
Union Carbide, South Charleston, West Virginia 70
Union Carbide, Texas City, Texas 110
Uniroyal, Plainesville, Ohio 60_
Total 2,268
*Chemical Marketing Reporter, Vol. 208, No. 2, July 14, 1975.
-------
Manufacture of Perchloroethylene and 1,1,2-Trichloroethylene
An inexpensive fluid bed catalyst with a copper base is used in the
reaction below to produce perchloroethylene from ethylene dichloride.
2C12 •* C2C14H2 + 2HC1
To produce trichloroethylene, a portion of the chlorinated organic
by-product can be recycled, along with the light ends.
C2H4C12 + 2C12 -»• C2HC13 + 3HC1
Combination production of perchloroethylene and trichloroethylene is
shown in Figure 4. In the simultaneous chlorination and dechlorination of
ethylene dichloride, dry chlorine and liquid EDC are fed to a fluidized
catalyst reactor. Light chlorinated organics are recycled back to the reactor.
The crude perchloroethylene is neutralized with a circulating caustic wash
stream to remove dissolved HCl. The caustic solution is purged periodically
to remove dissolved solids.
_ TABLE 4. TRICHLOROETHYLENE PRODUCTION* _
Capacity
Producer (million kg/yr)
Diamond Shamrock, Deer Park, Texas 27
Dow, Freeport, Texas 68
Ethyl, Baton Rouge, Louisiana 18
Hooker, Tacoma, Washington 14
Hooker, Taft, Louisiana 27
PPG, Lake Charles, Louisiana 90
Total 244
*Chemical Marketing Reporter, Vol. 202, No. 21, November 20, 1972.
After decantation, the crude is dried and the perchloroethylene purified
in a two-step distillation system. Light ends are recycled back to the system
and the perchloroethylene is taken overhead from the second column. The HCl
can be used as a secondary feed to the reactor step with the addition of air
or oxygen. The reactor conditions are carefully controlled to selectively
produce perchloroethylene.
With different reactor conditions and additional distillation equipment,
trichloroethylene can be selectively produced along with the perchloroethylene
product. Steam can be generated by circulating condensate through tubes
in the reactor. The reactor off-gas is easily condensed with water cooling.
Refrigeration loads for secondary recovery are necessary to minimize air
pollution. Reaction temperatures are 300° to 400°C with pressures slightly
above atmospheric. A typical catalyst would be an active carbon impregnated
with cupric chloride.
10
-------
Ethylbenzene Manufacture
Ethylbenzene is the major raw material for the production of styrene
monomer (Fig. 5) . A small portion (about 2%) is made by super-f ractionation
of an aromatics fraction rich in ethylbenzene. The major process for
ethylbenzene manufacture involves the reaction of ethylene with benzene in
a Friedel-Crafts alkylation reaction.
A typical reaction system consists of a ceramic lined reactor operating
at atmospheric conditions. The fresh benzene is introduced in the bottom
of the reactor and the ethylene is sparged into the system. The reactor
is usually operated at the boiling point of benzene. Since the reaction is
exothermic, the benzene vapors boil overhead and are condensed and refluxed
to the reactor. In this manner, the reaction temperature can be controlled.
The reaction system forms a complex with the aluminum chloride. The
product, along with the complex, overflows the reactor and goes to a decanting
vessel. The complex is a heavy, oily material and readily settles out of
the overflow. After settling, the complex is recycled to the reactor. The
product is water washed to remove the aluminum chloride and is then "sweetened"
with a caustic wash. Purification is by distillation.
There are two effluents of concern in the manufacture of ethylbenzene
by alkylation: (1) The overflow from the water washing of the alkylate
reactor, which contains soluble aromatics (benzene, ethylbenzene, diethylben-
zene, etc.) and aluminum chloride. This overflow is acidic with residual
hydrochloric acid (if the pH is adjusted to a basic condition, the aluminum
will appear as the hydroxide floe). (2) The benzene feed to the reactor,
which is usually predried by azeotropic distillation. The water from this
drying operation is decanted.
Styrene
Ethylbenzene is normally made in conjunction with the manufacture of
styrene. After purification, the ethylbenzene is preheated in the presence
of steam and passed over an iron oxide based catalyst where it is partially
converted to styrene. The reaction products are cooled and the oil layer
goes to a distillation section. The water is oil stripped and recycled to
the steam process. In the distillation section, the styrene is purified by
vacuum distillation. A possible effluent source in this area is the vacuum
drain from the steam jets. A process innovation would be to recycle these
to the dehydrogenation intermediate tank.
11
-------
Benzene
Ethylene
Benzene recycle
cn
cd
co
cd
is
/ I
£~~
: \ «^ 1
^T^
f Q
A M (1 b ^
1—1 ^j__ " -^
H *
p
T
t Benzene "^
saturated^^ ^
element 1 _ ,< ,,,.
ii ^ i
Alum:
chloi
arom£
acid
*
1 *l
f \t
.num Spei .t caustic
•ide
itic satu: ated
solution
\t
L_J L^T
To fuel Ethylbenzene
f
_A|
- 'sAx
Ml
Steam
Benzene
toluene
r^'"^
i a
S-I-H
J>^CiJ ^
0)(U 1
QM! ^
Condensate
to stripping
and steam
regeneration
1
t.
J^
•>
Jl
JP "
i->
_J >]r
Tar to fuel
Figure 5. Ethylbenzene and styrene manufacture.
12
-------
TABLE 5. STYRENE MONOMER PRODUCTION*
Producer
Production
(million kg/yr)
Ajnoco, Texas City, Texas
Qos-Mar, Carville, Louisiana
|i)ow, Freeport, Texas
bow, Midland, Michigan
/El Paso, Odessa, Texas
|Foster Grant, Baton Rouge, Louisiana
/ Gulf, Donaldsonville, Louisiana
/ Monsanto, Texas City, Texas
I Shell, Torrance, California
• Sinclair-Koppers, Houston, Texas
\ Sinclair-Koppers, Kobuta, Pennsylvania
\Sun (DX), Corpus Christi, Texas
Union Carbide, Seadrift, Texas
Total
360
225
295
160
55
325
225
360
110
50
195
35
135
2,530
*Chemical Marketing Reporter, Vol. 201, No. 8, February 21, 1972.
Isocyanate
The manufacture of toluene diisocyanate (TDI) has the following sequential
processing steps, starting with dinitrotoluene as the base material.
CH, CH,
Dinitrotoluene
+ 4H,
4H20
Toluene diamine
CO + Cl,
COC1,
CH,
+ 2COC1,
NCO
4HC1
13
-------
Phosgene is usually generated as needed and is fed to the reactors via
a selective solvent where it is reacted with toluene diamine at progressively
higher temperatures. The reaction products are first stripped of residual
phosgene and hydrogen chloride and then the TDI product is purified via
vacuum distillation. The effluent is usually highly colored, but can be
treated biologically.
Polyols
Polyols are used primarily in the production of polyurethanes, although
they have specific applications themselves. Polyols may be made from two
sources:
1. Ethylene oxide
2n
C - 9
H H
H H
9 - 9
H H
0-9-9-0-
H H
n
2. Propylene oxide
2n
H - C - C - C-H
i i T
H H H
H H H H
-C-C-0-C-C
H-C-H H-C-H
- 0 -
H
H
n
The reaction generally takes place under pressure in the presence of a
solvent. Process effluent consists of an aqueous wastewater saturated with
solvent, intermediate and the final product.
Methyl Cellulose
The effluent from a synthetic oxide methyl cellulose process consists
of the wash water from the final product cleaning operation. Because the
finished product is a solid, cellulosic material, the wash water contains
traces of the methyl cellulose. The water will also contain trace amounts
of the unreacted raw materials and certain intermediate or incomplete
reaction products. The final reaction products in aqueous solution are
neutralized, creating a salt solution which contains all of the trace
components. Treatment of an aqueous brine solution saturated with relatively
nonvolatile components is limited. Steam stripping and solvent extraction
can be eliminated from possible consideration, since the suspended solids are
nonvolatile, and the cellulose trace components are not very soluble in any
solvents.
SOURCE OF EFFLUENTS
The point sources of the individual process effluents included in the
program are listed in Table 6. The general nature of each process source is
indicated.
14
-------
TABLE 6. CODED PROCESS WASTEWATER STREAMS
Stream
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Code
Designation
011A001
011B001
011C001
041A001C
041BC01
041C001
041D001
041E001
041F001
041G001B
081A001
081B001
081C001
131A001B
161A001C
161B001
161C001
221A001C
221B001
221C001
231A001B
231B001B
351A001B
Process Origin
Halogenated hydrocarbons
Halogenated hydrocarbons
Halogenated hydrocarbons
Halogenated hydrocarbons
Halogenated hydrocarbons
Halogenated hydrocarbons
Aromatics and organic oxides
Halogenated hydrocarbons
Halogenated hydrocarbons
Methyl cellulose and hydroxyethers
Halogenated hydrocarbons
Halogenated hydrocarbons
Halogenated hydrocarbons
Aromatics (quench)
Halogenated hydrocarbons
Halogenated hydrocarbons
Halogenated hydrocarbons
Halogenated hydrocarbons
Halogenated hydrocarbons
Halogenated hydrocarbons
Aromatic amines
Polyols and oxides
Aromatics
INDIVIDUAL PROCESS EFFLUENT CHARACTERIZATION
Each of the process effluents was completely characterized before
decisions were made as to the type of treatment steps which would be applied
to these streams. Individual samples were collected for each stream which
were representative of the process conditions and specific tests conducted.
Details of the tests performed on each individual process stream are given
in Tables 7, 8 and 9.
Since the project extended over a four-year period, a recharacterization
of the process streams under investigation was made to check for any effluent
variations. A comparison was made of the stream characterizations at the
beginning and end of the period.
Analytical Developments
To investigate the effluent with respect to the effectiveness of certain
unit operations for contaminant removal, it was necessary to identify specific
organic components of the process effluents. Known analytical techniques
were not sufficiently commercialized to utilize set procedures. Consequent:!;',
it was necessary to develop certain test procedures to quantify these specific
organic compounds.
15
-------
TABLE 7. SOLUBLE ORGANIC STREAM CHARACTERIZATION
011A 011B 011C 041A 041B
Soluble Organics (mg/1) (mg/1) (mg/1) (mg/1) (mg/1)
Methyl chloride Trace 1.0 736.0 0-40 2.4
Ethyl chloride 19.7 Trace 0.9 0-30 Trace
Vinylidene chloride 1.0 0.8
Trans 1 , 2-dichloroethylene
Cis 1, 2-dichloroethylene Trace 6.0
Chloroform 2.5 0.1 25.0
Ethylene dichloride 6974.0 2858.0 5755 9122.0 6.0
Carbon tetrachloride
Trichloroethylene
1,1,2-Trichloroethane 17.0
Perchloroethane
Tetrachloroethane
Toluene
Benzene
041C 041D 041E
(mg/1) (mg/1) (mg/1)
2.0
'
0.3 2.0
0.1
0.1
0.2 0.2
6.0 6.0 2.0
5.0
3.0
2.0
11.6
041F 081A
(mg/1) (mg/1)
200.0
Trace
5.7 0.1
10.0
15.6
3.3 6.0
1.0
202.0
7.0
306.0
081B
(mg/1)
3.0
1.0
42.0
2.0
3.0
7.0
081C
(mg/1)
2.0
0.1
0.5
1.0
2.0
8.0
84.0
20.0
-------
TABLE 8. SOLUBLE ORGANIC STREAM CHARACTERIZATION
Soluble Organics
Methyl chloride
Ethyl chloride
Vinylidene chloride
Trans 1,2-dichloroethylene
Cis 1,2-dichloroethylene
Chloroform
Ethylene dichloride
Carbon tetrachloride
Tr ichlor oethy lene
1,1, 2-Trichloroethane
Perchloroethane
Tetrachloroethane
Toluene
Benzene
161A
(mg/1)
3.0
10.0
0.1
7.0
23.0
200.0
6.0
40.0
14.0
161B
(mg/1)
2.0
6.5
6.5
19.7
64.0
410.0
25.5
29.6
10.4
-
161C
(mg/1)
13.0
2.7
56.8
327.4
546.3
1077.6
109.6
221A 221B
(mg/1) (mg/1)
2.0 0.2
3.8
1.3
1.1
683.0 0.96
8331.0 9238
53.6
221C 351A
(mg/1) (mg/1)
21.4
54.8
0.21
1.2
5.65
-------
TABLE 9. STREAM CHARACTERIZATIONS
Stream
011A
011B
one
041A
041B
041C
041D
041E
041F
041G
081A
081B
081C
131A
161A
161B
161C
221A
221B
221C
231A
231B
061A
COD
(mg/1)
3678
1112
615
1764
30
454
325
10752
18
23198
6
100
390
1050
1040
418
6279
16100
11280
6080
10680
2562
463.6
TC
(mg/D
1565
712
1703
1638
250
13
69
2576
54
7409
50
17
57
273
1843
148
4390
9022
10250
5840
1810
821
479
PH
12.2
1.5
11.2
11.7
2.9
2.5
6.9
5.8
10.5
10.6
10.6
1.4
12.1
9.5
7.7
0.9
<0.1
0.1
<0.1
12.7
9.7
11.9
11.9
Flow
(1/min)
76-227
57-151
189-416
341-454
341-454
241-454
1514-1892
341-454
341-454
114-227
76-114
76-114
76-114
114
76-189
57-114
38-95
57-114
38-57
19-38
38-57
57-95
76-151
Alkalinity
as CaCO~
(mg/ir
16000
0
4840
1050
0
0
30
376
325
—
1180
0
12635
220
304
0
0
0
0
31800
1475
1140
2060
Acidity
as CaCO™
(mg/ir
0
29254
0
0
135
241
10
8
0
—
0
2580
0
0
10
0
0
102,312
269,892
0
0
0
0
Chlorides
as Cl
(mg/1)
18113
1988
6564
2662
408
762
7
6
418
28617
1
1909
9685
119
5140
10557
77112
116,127
107,289
46
14
32
147
18
-------
During the study period, there were two methods used for analyzing
chlorinated hydrocarbons: (1) the vapor-liquid partition technique developed
by Dow Research* and (2) a direct liquid injection utilizing a reference
component as base identification.
After careful investigation, the partition coefficient method was selected
since it offered more flexibility in analyzing different effluents with
varying amounts of trace organics.
EFFLUENT TREATMENT
The process effluent characterization data indicated that 17 of the
individual wastewater streams were derived from a chlorinated organic process.
Even though the individual process effluents came from different plants, the
amounts of the chlorinated hydrocarbons in each were similar.
Initially, air stripping trials were intended to indicate which process
effluents would create potential emission problems because of the volatile
nature of the organics present in the wastewater. The need for this informa-
tion was obvious where biological treatment was being considered.
Air Stripping Studies
Wastewaters which contain primarily volatile organic compounds pose a
problem to biological waste treatment systems which utilize open vessel
areation as a means of dissolved oxygen control. As the wastewater enters
the aeration vessel, a major portion of the organic loading is air stripped.
An estimate of the volatile and nonvolatile fractions of the organic contami-
nation must be known to interpret data collected during biological treatment
studies.
Since the majority of the streams selected for the project are waste-
waters from chlorinated solvent processes which contain volatile organics, a
study was conducted to determine the effect of aeration alone on the streams'
pollutant concentrations.
Results of the air stripping study are shown in Tables 10 and 11. The
results indicate that 11 of the 18 streams tested are air strippable in
excess of 50 percent of their total oxygen demands. Stream 011C displayed
the highest air stripping potential (88 percent), and stream 041C the least
(10 percent), based on the TOD analyses of the feed and residual streams.
These results show that an efficiently operated steam or air stripping opera-
tion can reduce the organic concentration of most of the streams appreciably
and in some cases leave relatively biodegradable residual pollutants.
Table 10 shows the results from air stripping of the various industrial
waste streams. Total oxygen demand (TOD) of the initial sample and TOD of
the residual stream were determined. In column six, the ratio of the BOD to
the TOD in the residual is shown. This number is interpreted as the biodegrad-
*Analytical Method No. QA-466, Dow Chemical, U.S.A., Louisiana Division,
Plaquemine, Louisiana 70764.
19
-------
TABLE 10. AIR STRIPPING AND BODr RESULTS*
Stream
041A
041B
041C
041D
041E
041F
011A
011B
one
081A
081B
081C
161B
161C
221A
221B
221C
131A
TOD
Initial
(mg/1)
4260
1389
71
510
10500
275
—
2200
1230
120
260
—
500
7600
5800
5700
4500
1700
TOD
Residual
(mg/D
642
1380
64
235
5000
140
—
1950
130
92
36
—
235
5050
2350
2300
2850
1950
TOD
Removal
(%)
85
0
10
54
52
49
—
11
89
23
86
—
53
34
59
60
37
-15
BOD5
Residual
(mg/D
247
37
726
61
2137
131
38
159
44
124
11
159
61
1800
1750
1450
2450
658
-, :
BOD5/TOD
Residual
(mg/1)
0.38
0.03
—
0.26
0.43
0.94
—
0.08
0.34
—
0.31
—
0.26
0.36
0.74
0.63
0.86
0.34
Total Carbon
Initial
(mg/1)
__
300
40
225
2500
88
1920
830
830
9
124
105
67
3900
3000
2150
2750
600
Total Carbon
Residual
(mg/1)
240
450
22
125
3200
53
215
825
100
34
18
90
62
3450
1450
1400
1800
600
*Aeration time for all samples was 500 min; each sample was approximately 1 liter. Air flow
was not measured. The surface of the reactor was visibly agitated by the air stream entering
at the bottom of the reactor. The air flow was designed to simulate a bio-aeration system.
-------
TABLE 11. PERCENT REMOVAL OF VOLATILE ORGANICS BY AIR STRIPPING
041A 041B 041C 041^ 041E 041F 011A 011B 011C 131A 081A 081B 081C 221A 221B 221C 161B 161C
Soluble Organics
Low molecular weight
volatiles 99.9 100.0 89.9 99.7
Vinylidene Chloride 99.9 100.0
Cis 1,2-Dichloroethylene 99.8 92.2 93.8
Trans 1,2-Dichloroethylene
Chloroform 99.7 74.2
Ethylene Dichloride 89.5 89.5
1,1, 1-Tr ichloroethane
Carbon Tetrachloride
Trichloroethylene 100.0 99.6
1,1,2-Trichloroethane 85.0
Perchloroethylene
1,1,2, 2-Tetrachloroethane
Benzene 99 . 9
Toluene 99 . 9
Xylene 100 . 0
Styrene 100.0
100.0 99.9 50.0 100.0 39.6 32.8 98.5 98.6 99.9 100.0 100.0
91.6 47.5
100.0 100.0 99.9 27.8
100.0 97.1
99.9 98.3 98.8 95.4 99.4 97.5 99.9 98.8 68.2 99.0 100.0 99.0 87.7
99.6 99.7
97.8 99.8 99.9 99.9 80.9
89.6 86.9 100.0 99.0 71.1
99.7 100.0 99.9 100.0 32.6
98.2
-------
ability of the stream. A ratio close to one would indicate that the residual
contaminants after air stripping were easily biodegradable. A very small
number would indicate the reverse. The range of this ratio for the various
industrial streams varies from 1.05 to a minimum of 0.05.
Carbon Adsorption
To determine if the process effluent stream contaminant loading could be
reduced by carbon adsorption with subsequent regeneration, wastewaters from
the representative processes were collected and adsorption isotherms experi-
mentally obtained. The results indicated that for streams 041A, 161A, and
221A the chlorinated hydrocarbons were readily adsorbed. The possibility of
regeneration with steam in situ was good for these streams. Isotherm and
breakthrough data for streams 011A, B, and C, and streams 081A, B, and C were
not determined for the preliminary evaluation since these streams exhibit unit
operation characteristics similar to the ones for which data were obtained.
A preliminary examination of the constituents in streams 041F, 231A, and
231B indicated that these would not be amenable to carbon adsorption.
Stream 041F contained some suspended solids and would require prefiltration.
Batch Biological Treatment Evaluations
Evaluation of the biological degradability of the various process
effluents was conducted to aid in determining the potential process sequence.
These results are shown in column 5 of Table 10. All of the process effluents
exhibited biodegradability. Chloral was the main contaminant being biodegraded
in the chlorinated hydrocarbon contaminated effluents. The process effluents
considered for biotreatment were:
1. Stream 041G - methyl cellulose
2. Stream 131A - quench wastewater blowdown
3. Stream 231A - aromatic amines
4. Stream 231B - polyols
5. Stream 351A - aromatics
Organic Oxide Methyl Cellulose Waste Stream (041G)—
This stream was found to contain high chloride levels averaging 32,339
mg/1, and would require neutralization and supplemental nutrients preceding
biological treatment. The high chloride levels inhibited the biotreat-
ability of this process effluent, although it was possible to acclimate a
microorganism which would function in this environment. Under typical
operating conditions, approximately 60 percent of the BOD- could be removed
from the inlet water.
Aromatic Wastewater (351A)—
This wastewater contains a high concentration of total dissolved
solids (TDS) (70,198 mg/1). The wastewater requires neutralization and
supplemental nutrients (nitrogen and phosphorus) prior to biological treat-
ment. Although this process effluent did not exhibit any toxicity toward
the microorganisms, the refractive nature of the organics did not permit a
total reduction of residual BOD- below 60 percent of the incoming stream.
Similar results were obtained with streams 231A and 231B.
22
-------
Solvent Extraction
Since air stripping of the halogenated hydrocarbon contaminated waste
water indicated a major portion of the organics would be removed, it is
expected most of the organics would be removed by volatilization instead
of biological degradation. Air dispersion would not contribute to total
environmental removal. As a result of this preliminary evaluation, it was
concluded that a unit operation such as solvent extraction might be a
suitable pretreatment step or effective enough to remove the organics to
desired levels. Several solvents showed promise for this in-process method
of reducing the organic contaminants in the process effluent. Feasibility
of solvent extraction for four streams is outlined in Table 12.
TABLE 12. FEASIBILITY OF SOLVENT EXTRACTION
Stream
Number Type of Contaminant Discussion
221A Halogenated hydrocarbon Easy to solvent extract.
Solvent selection dependent on
residual in wastewater and its
resulting need for further
treatment.
351A Aromatics High dissolved solids.
Aromatics could be solvent
selective.
231A Aromatic amines Not amenable to extraction.
231B Polyols Some detergent action which
could inhibit solvent selectivity.
Preliminary work at Texas A & M University on process effluent samples
indicated that a kerosene fraction made from petroleum might selectively
extract the chlorinated hydrocarbons present in process effluent. The results
suggested that a straight chain saturated hydrocarbon fraction in the C.. „
range might also be selective. This latter solvent might be preferable
because of its low solubility in water, as well as its high boiling point,
which could be easier to control in a solvent recovery and recycle.
Steam Stripping
The high volatility of the chlorinated hydrocarbons dissolved in the
process effluents from this industry group make steam stripping of the waste-
water attractive. Because of the low volatility of the residual chlorals,
these related compounds will be left in the stripper bottoms. The two
factors which affect removal of organic contaminants by steam stripping are:
(1) amount of organics present and (2) formation and composition of azeo-
tropic mixtures.
The existence of an azeotrope is not necessarily deleterious to the
application of steam stripping for removal of trace organics. If the organic-
23
-------
water mixture going overhead forms two immiscible layers, it is still possible
to utilize this approach to controlling contamination of process effluents.
The two phases which condense from the overhead can be directed back into the
system at different points. The water layer can be refluxed to the top of
the stripping column while the organic layer is directed back to an intermedi-
ate stage of the process.
An examination of the effect of recycling the organic layer back to
process is necessary. If there are undesirable by-products present, the
injection of this organic layer must be such that these by-products can be
recovered in bottoms or overheads in the subsequent finishing section.
The process effluents under evaluation in this program exhibit the
characteristics listed in Table 13 with regard to steam stripping as a means
of trace organic removal.
TABLE 13. APPLICABILITY OF STEAM STRIPPING
Stream
Discussion
Halogenated hydrocarbons
Aromatic rich effluent from
styrene plants
Aromatic amines and polyols
Very volatile - azeotrope with water,
two-phase overhead system, stripper
bottoms biodegradable. Steam strip-
ping offers possibilities for
application to this process effluent.
Volatile - azeotropic with water, two-
phase overhead system, stripper
bottoms organic free. Steam strip-
ping should be effective for this
process effluent.
Extremely water soluble - The number
of various organics present in this
waste water would not permit success-
ful aromatic steam stripping.
Monochem Soot Carbon
In the manufacture of acetylene utilizing methane as raw material, a by-
product soot results. This soot is in the 25 to 150 micron range and appears
as an emulsified slurry in final form from the process. Preliminary investi-
gations indicated that this material could be dried and activated to yield a
commerically acceptable product. This soot carbon was developed and compared
with the commercial activated carbons for its applicability to removal of
organics from the process effluents.
Ozonation
The original rationale for using ozone as an oxidizing agent was that
the halogenated hydrocarbons and aromatics exhibited increased biological
activity with high residuals of total organic carbon. For example, toluene
24
-------
itself is not very biodegradable; however, its oxidized counterpart, benzoic
acid, is readily biodegradable. Thus, partial oxidation of an aromatic in a
process effluent, prior to a biological treatment, would improve its bio-
degradability. The same technique could be used for halogenated hydrocarbons,
polyols, and aromatic amines.
25
-------
SECTION 5
ACTIVATED CARBON ADSORPTION
Based on the characterizations of the process effluents and inspection
of the individual contaminants, the wastewaters were divided into three
main subgroups.
Number Manufacturing Major Organic
of Streams Wastewater Source Contaminant
17 Production of ethylene Halogenated hydrocarbons
dichloride, perchloroethylene,
and 1,1,1-trichloroethane
2 Polyols and methycellulose Polyhydric alcohols
4 Ethylbenzene, styrene, Aromatics, organic
toluene diisocyanate and oxides, or aromatic
other process quench waters amines
Past experience indicated that halogenated hydrocarbons, aromatics,
and organic amines should be readily adsorbed by activated carbon. Polyols
and polyhydric alcohols should not be very adsorbable due to their high
molecular weights. Previous studies also indicated they were readily
biodegradable, so that best available technology would tend to favor treat-
ment schemes other than carbon adsorption for these two process effluents.
With this in mind, it was decided to investigate process effluents
containing chlorinated hydrocarbons, aromatics, and aromatic amines. Since
wastewaters contaminated with chlorinated hydrocarbons made up the bulk of
the process effluents, it was desirable to include several different
industry sources. Wastewaters from three manufacturing processes for
chlorinated organics were selected.
Stream number Type of contaminant
041-A Chlorinated hydrocarbons
161-A Chlorinated hydrocarbons
221-A Chlorinated hydrocarbons
351-A Aromatic hydrocarbons
231-A Aromatic amines
26
-------
LABORATORY PROGRAM
Introduction
After the process effluents were selected, the laboratory procedure
for implementation of this phase of the program was set up. The ultimate
aim of the project was to field test two adsorption demonstration units at
several selected sites. An indepth laboratory program was devised to yield
the information desired for the subsequent field work.
The process effluents have been selected. Standard commercial activated
carbons were chosen for comparative purposes. The initial step in the
evaluation of an activated carbon was to determine its adsorption capacity
for the contaminants present in the wastewaters. This was done by applying
Freundlich isotherm studies to each carbon tested.
Effluent Isotherms Monitored for Halogenated Hydrocarbon Residuals
Ethylene dichloride (EDC) adsorption isotherms were obtained for
streams 041A, 221A, and 161A. Westvaco (WVG) Calgon, WITCO, and Barneby
Cheney carbons were used for streams 041A and 221A. Only WVG carbon was
used for stream 161A. The other carbons were dropped because WVG carbon
gave the best results in the previous tests. Although other organics were
known to be present^ in these wastewaters, EDC was the chlorinated hydrocarbon
most prevalent. Thus, the adsorption characteristics could be interpreted
with the determination of this compound. Effluent isotherms have been
plotted for streams 041A, 221A, and 161A and are included in the Appendix.
The data are summarized in Tables 14 and 15.
Observations—
1. Stream 041A: Calgon (FS-400) exhibited the highest adsorption
capacity for EDC, followed by Barneby-Cheney (BCNB-9377),
WITCO, and Westvaco.
2. Stream 221A: WITCO, Westvaco, and Calgon were effective in this
order for EDC removal.
3. Stream 161A: Calgon exhibited less adsorption capacity than
Westvaco over a narrower range of EDC residual concentration.
4. A carbon having the highest adsorption capacity in the higher
range of residual EDC concentration did not necessarily exhibit
highest adsorption capacity in the lower range of residual EDC
concentration.
Monochem Activated Soot Carbon Isotherms
Included in this program was a waste stream emanating from an acetylene
manufacturing process, which forms acetylene from the cracking of natural
gas. A very fine carbon black is formed as a by-product and is removed
from the product gas by a water quench. This water carries the carbon
black "soot carbon" to a separator station where it is dewatered and burned.
27
-------
TABLE 14. ADSORPTION CAPACITY OF ACTIVATED CARBONS FOR REMOVAL OF ETHYLENE DICHLORIDE
QO
Filtrasorb Westvaco
400 12 X 14 Witco
Stream
Number
(gms/gm (gms/gm (gms/gm
Carbon) Carbon) Carbon)
Carbon required for a
041A
221A
0.85 1.
0.55 1.
0 0.55
25 0.60
Carbon required for a
041A
221A
161A
0.13 0.
0.055 0.
0.014 0.
0126 0.07
055 0.10
0145
Barneby
NB3377
(gms/gm
Carbon)
residual
0.47
-
residual
0.12
-
—
Mono ch em
Soot Carbon
(gms/gm
Carbon)
of 10 mg/1 EDC
0
0
of 0.1 mg/1
.13
.125
EDC
TABLE 15. ADSORPTION CAPACITY OF ACTIVATED
CARBONS
FOR REMOVAL
OF TOG
Filtersorb Westvaco
Stream
Number
041A
221A
231A
351A
400
(gms/gm
Carbon)
0.35
0.65
100
25
12 X 14
(gms/gm
Carbon)
Carbon required
0.062
0.7
Carbon required
150
20
Witco
(gms/gm
Carbon)
Barneby
Cheney
(gms/gm
Carbon)
for a residual of 0.5
0.052
0.6
0.42
for a residual of 50
—
15
7,0
Mono ch em
Soot Carbon
(gms/gm
Carbon)
mg/1 TOC
—
0.3
mg/ liter
25
7.0
-------
A development program was established to formulate an activated carbon from
this waste material. The soot carbon was activated by Envirotech in Cali-
fornia. The activation temperature used was between 870°C and 925°C with a
steam blanket.
The granular activated carbon sample was received at GSRI in the form
of a very thick slurry in water. It had a considerable amount of carbon
fines in it. These fines were washed off with copious quantities of water.
The granules were oven dried at 200°C for approximately 48 hr. The dried
carbon granules were used to make adsorption isotherm and breakthrough
measurements.
Effluent Isotherms Monitored for Total Carbon Residuals
In an attempt to define the composition of effluent waters after
treatment with activated carbon, adsorption isotherms were generated using
total carbon (TC) as a measure of concentration for streams 041A and 221A.
Filtrasorb-400, WVG, WITCO, Barneby-Cheney BCNB-9377, and varieties of
activated soot carbon were used. Standard experimental procedure was
followed to obtain the adsorption isotherms.
Ob s erva t ions—
1. WVG was the best among the carbons evaluated for total carbon
removal of streams 041A and 221A.
2. Activated carbon that adsorbs maximum EDC does not necessarily
adsorb maximum TC.
3. Stream 041A lent itself to adsorption with commercial carbons
better than stream 221A.
4. TC adsorption capacity of the activated soot carbon was less than
50 percent of the adsorption capacity of commercial carbons.
Effluent Isotherms for Streams 231A and 351A
Two streams from other processes were investigated as part of the
preliminary studies. These two streams were produced by an ethyl benzene-
styrene plant (351A) and a toluene diisocyanate plant (231A).
Stream 351A contained predominantly benzene and ethyl benzene and is
basic. This stream was treated with various activated carbons. Experimental
data indicate that Westvaco Nuchar carbon yielded an isotherm that would
have an adsorption capacity of 0.285 g TOC/kg carbon. All other carbons
tested would leave a residual TOG concentration of 4 to 30 mg/1; only the
Nuchar carbon appears to treat the stream effectively. Stream 231A contained
water-soluble organic compounds from a process making toluene diisocyanate
and toluene diamine compounds. The wastewater was too dilute to permit
compound identification. However, Westvaco and Filtrasorb-400 activated
carbon appeared to offer the best removal characteristics when tested for
the reduction of TOG as the gross parameter.
29
-------
Dynamic Tests for Breakthrough Evaluations
Estimation of mass transfer parameters for column design and comparison
of performance of different carbons under dynamic conditions is accomplished
by breakthrough measurements. A plot showing the relationship between the
cumulative volume of liquid passed through a carbon column and the effluent
concentration of the component being removed is termed "breakthrough"
curve. For most adsorption operations in water and wastewater treatment,
breakthrough curves exhibit a characteristic S shape, but with a varying
degree of steepness. A steep curve indicates a high rate of adsorption; a
short, steep curve indicates a high rate of adsorption, but also a short
column service life. Figure 6 shows typical breakthrough curves. Obviously,
Column A will have a shorter service life than Column B.
A specially fabricated glass apparatus provided a constant liquid head
to the three columns (made of chromatographic tubes with 20 mm ID and 450
mm length). Feed flow rates were varied and accurately controlled by the
precision stopcock on each of the tubes. The schematic of the apparatus is
shown in Figure 7. Waste stream samples were analyzed on a gas chromatograph
for EDC and on a total carbon analyzer for total carbon.
Five combinations of carbons and waste streams were used in establish-
ing the dynamics expected in the field trials. Two characteristic streams
were used. The sample of 041A used for these experiments was high in
chlorinated hydrocarbons (approximately 8,000 mg/1), highly alkaline, and
showed a total carbon by the Beckman method of approximately 1,200 mg/1.
Stream 221A had a total chlorinated hydrocarbon content of approximately
12,000 mg/1 and was acidic. The total carbon level of about 6,000 mg/1
clearly indicated the presence of nonchlorinated hydrocarbons.
Three carbons were used during the trials: (1) WVG carbon, which had
shown a high equilibrium adsorption of EDC and total carbon previously, (2)
Filtrasorb-400, which had also shown a high adsorptive capacity, and (3)
activated soot carbon treated in the laboratory to a medium level of activa-
tion. The limited amount of soot carbon precluded carrying out the complete
test program with this carbon.
For each combination, three breakthrough curves were generated at
three different flow rates denoted by L (low), M (medium) and H (high).
The range of flow rate were expressed as liters per square meter of activated
carbon.
L: 0.57 1/nu to 1.13 1/nu
M: 1.13 1/nu to 1.98 1/nu
H: 1.98 1/m to 2.83 1/m
For stream 041A, the reduction of EDC was accomplished equally well by
Filtrasorb-400 and WVG. However, bed life of the soot carbon sample studied
would be only 40 to 50 percent as long as that of the two commercial carbons.
The Filtrasorb-400 appeared to have less sensitivity to flow rates that the
WVG in the range studied. The soot carbon did not show any flow effects.
30
-------
Effluent volume (liters)
Figure 6. Typical breakthrough curves.
JD
^AF
A - Feed reservoir
B - Feed pump
C - Constant level
device
D - Feed distribution
manifold
E - Carbon columns
F - Sample flasks
G - Stopcocks
Figure 7. Laboratory continuous carbon adsorption apparatus.
31
-------
For stream 221A, the minimum total carbon level obtainable in the
effluent from carbon column is about 300 tag/1 independent of the type of
carbon used. Highest capacity for the solutes in this stream was shown by
WVG, followed by Filtrasorb-400 and soot carbon. Little sensitivity to the
flow rates was found for the soot carbon. The data indicated that a species
other than chlorinated hydrocarbons was present and was poorly adsorbed by
carbons.
The capacity of both commercial carbons for stream 221A was nearly the
same, and was significantly less than for stream 041A. The capacity of the
activated soot carbon again was lower than that of the commercial carbons.
With this stream, higher flow rates bring about an earlier breakthrough.
While high acidity may be expected to reduce the available adsorption sites
of the carbon, this has not been substantiated by reductions in acidity.
The reduction of total carbon in effluent stream 221A was inefficient
with WVG and Filtrasorb-400. The removal of the nonchlorinated hydrocarbons
will have to be achieved by some other method.
Observations—
1. At 6000 mg/1, EDC, adsorption capacity of the activated soot
carbon was no greater than 55 percent of the adsorption capacity
of commercial activated carbons. However, at 1000 mg/1 EDC, the
adsorption capacity of the former was approximately 75 percent of
that of commercial activated carbon.
2. A better EDC adsorption capacity for stream 221A than for stream
041A with activated soot carbon was observed.
3. An improvement of 20 to 25 percent was observed in the adsorption
capacity of EDC over the two previous batches (#79 and #104) of
activated soot carbon.
4. The breakthrough profile for stream 041A indicated a higher
service life for stream 041A than for stream 221A. The concen-
tration gradient of the organics in the carbon bed indicated the
absorption capacity of the carbon would be better utilized before
the concentration of the organics in the outlet exceeded specified
limits.
Regeneration of Spent Activated Carbon with Stream
Regeneration and reuse of spent activated carbon is one of the most
important factors in determining the economic feasibility of the activated
carbon adsorption process. Carbon beds have to be regenerated after their
product exceeds a predetermined level of acceptable impurities.
Regenerability—
Five spent carbon samples were regenerated with 1 atm steam for a
period of 15 min. The adsorptivity of the regenerated samples of carbon
was compared with the fresh carbon samples. For the comparison, accurately
weighed 2 g samples of the fresh and regenerated carbons were added to 100
32
-------
ml waste stream in a glass stoppered flask. After equilibrating these
samples for 24 hours, fractional reduction of EDC for the regenerated and
the fresh carbon was determined. The regeneration efficiency E was computed
as:
for 2 g regenerated carbon
x iUU
FR for 2 g fresh carbon
where FR is the fractional reduction of solute.
Results with EDC and TC as measures of concentration are listed in Tables
16 and 17.
TABLE 16. REGENERATION EFFICIENCY - EDC
!•
Stream
041A
221A
041A
221A
041A
Carbon Fraction Reduced
Name With Fresh Carbon
WVG
WVG
Filtrasorb
Filtrasorb
Monochem
0.945
0.64
0.94
0.65
0.66
Fraction Reduced
with Regenerated
Carbon
0.93
0.60
0.92
0.59
0.57
Regeneration
Efficiency
(Percent E)
98
94
98
91
93
Weight of Carbon:
Volume of Waste Stream:
Solute:
2.00 g
100 ml
EDC (11,000 mg/1)
The data of Table 17 were obtained to check the conclusions based on
EDC measurements. Both sets of data show an excellent potential for low
pressure steam regeneration of the spent activated carbon in situ. Regener-
ation efficiencies were generally lower on stream 221A than on stream 041A.
TABLE 17. REGENERATION EFFICIENCY - TC
Stream
Number
221A
041A
Carbon
Name
Westvaco
Filtrasorb
Fraction Reduced
With Fresh Carbon
0.135
0.092
Fraction Reduced
With Regenerated
Carbon
0.101
0.09
Regeneration
Efficiency
(Percent E)
75
98
Weight of Carbon:
Volume of Waste Stream:
Solute:
Solute Concentration:
2.00 g
100 ml
TC
#041A - 950 mg/1
#221A - 5200 mg/1
33
-------
Regeneration Time—
To determine the effect of regeneration time on the efficiency of
regeneration, approximately 200 g of Filtrasorb-400 was equilibrated with
3 liters of stream 221A for a period of 16 hr. The EDC loading on the
carbon was determined by measuring fractional EDC reduction in the residual
221A solution. The 221A solution was then drained and the EDC loaded
carbon was washed three or four times with 2 liters of tap water. The
carbon was divided into two equal fractions of about 100 g each. At 1 atm
steam pressure, each fraction was regenerated for the time indicated.
Regeneration efficiencies were determined. Results are listed in Table 18.
TABLE 18. VARIATION OF REGENERATION TIME
Fraction Reduced Regeneration
Regeneration Fraction Reduced With Regeneration Efficiency
Time With Fresh Carbon Carbon (Percent E)
2 Minutes
5 Minutes
15 Minutes
0.76
0.76
0.65
0.70
0.71
0.59
92
93
91
Carbon: Filtrasorb-400
Waste Stream: 221A
Solute: EDC
Repeated Regenerations—
Preliminary tests indicate that it is technically feasible to regenerate
the carbon that is subject to loading by streams 041A and 221A. However,
these tests were limited to single regeneration. To study the effect of
repeated regenerations on the adsorption capacity of the spent reactivation
carbon, EDC loading on carbon and the regeneration efficiency were determined
after several regenerations at 1 atm. A quantity of 100 g of Filtrasorb-
400 was equilibrated overnight with 2 liters of 221A solution. The loaded
carbon was washed several times with tap water and was regenerated with 1
atm steam for 5 min. This procedure was repeated four times on the same
carbon sample. After five regenerations, EDC loading on the regenerated
sample was 0.186 g EDC/g carbon, as compared to 0.2 g EDC/g fresh carbon.
The regeneration efficiency was 93 percent.
DEMONSTRATION UNIT PROGRAM
Introduction
After the laboratory evaluation of activated carbon for removing trace
organics from selected process effluents, an on-site field evaluation was
conducted. The two activated carbon adsorption units for use in the field
testing portion had the following items of equipment.
34
-------
1. Four 127 mm I.D. by 1828.8 mm long plastic columns connected in
series with 12.7 mm polypropylene tubing. The initial columns
were made of Plexiglas. These columns were subsequently replaced
with fiberglass and polypropylene because of the solvent effects
of the ethylene dichloride contaminated wastewaters.
2. An A-frame rack on which the columns were suspended. The A-frame
was built of 38 mm nominal pipe with welded assembly.
3. Variable speed pump: Maximum discharge pressure 1.36 atm. Pump
speed and liquid flow rate were controlled by motor drive and
pump pulley sizes. The pulley provided flow rates ranging from
1.48 to 3.69 1/min. The pump was supplied with Viton® tubing
with an expected lifetime of 10 hr at flow rates 3.42 to 3.8
1/min. Calibration curves were provided for the various pulley
sizes.
4. Electric motor: 110 V. A.C. 1/4 HP T.E.F.C. motor for the pump.
5. Temperature and pH monitor: A flow-through chamber containing pH
and temperature sensor provided with a dual channel "Rustrak"
recorder.
The columns were prepared by placing a double layer of fiberglass
window screen in the bottom of each column followed by a 10 cm to 15 cm
layer of pea gravel. A heavy slurry of water and carbon (pre-soaked 24 hr
to deaerate the carbon) was then added to the columns to a desired height.
The columns were backwashed when the gauge pressure on the lead column was
1.3 atm or more. The backwash flow was 15.2 1/min (this corresponds to a
30 to 50 percent bed expansion).
Process Effluent Selections for Demonstration Unit Evaluations
The demonstration units were operated at three process effluent loca-
tions: (a) process effluent stream 221A, (b) process effluent stream 041A,
and (c) process effluent stream 081A. Three commercial carbons and the
Monochem activated soot carbon were evaluated.
Each pilot unit was operated at approximately 1.0 1/min during the
testing. This operation provided a screening for selecting the best of the
commercial carbons to compare with the activated soot carbon. Test duration
depended on the adsorption rate of each carbon relative to each stream.
After the best carbon had been selected, a series of tests was conducted
to determine design data for that carbon, as well as the Monochem carbon.
These tests consisted of breakthrough runs at various flow rates and organic
loading rates. Hydrodynamic data were also collected to ascertain flow and
pressure drop characteristics of the pilot unit.
©Viton is a registered trademark of E.I. du Pont de Nemours and Co., Inc.,
Wilmington, Del.
35
-------
PROCESS EFFLUENT STREAM 221A
Capacity and Breakthrough Data
The first field tests were made on process effluent stream 221A. Four
commercial carbons were evaluated for breakthrough and adsorptive capacity
limitations. However, the process influents for this series of tests
varied considerably during the experimental operations, and breakthrough
points were inconclusive.
The effectiveness of activated carbon for stream 221A was initially
measured by total organics adsorbed. It was recognized that other organics
might be present which were not being adsorbed. Since removal of organics
would be indicated by a reduction in oxygen demand, total carbon and total
oxygen demand (TOD) measurements on the influent to and effluent from the
carbon columns were incorporated as an indication of the carbon needed to
remove total contaminants present.
When a comparison was made between the analyses for organics, TC, and
TOD, the activated carbon appeared to allow some of the organic contaminants
to pass through because some variation in inlet flow occurred during the
field tests. This variable flow caused the loading on the carbon to change,
thus affecting the concentration of organics in the column effluent. Most
of the data obtained for the organic components -in the process effluent
indicated a definite breakthrough for this parameter.
In some instances, the gas chromatograph (GC) detected low minimum
chlorinated hydrocarbons, although the TC and TOD data indicated considerable
contaminants passing through the adsorption unit.
The production unit generating effluent stream 221A utilizes the
oxychlor process. In the reaction of ethylene, chlorine, oxygen, and
recycled higher chlorinated hydrocarbons, side reactions occur which produce
trace components. These trace organics include chloroacetaldehyde (chloral-
trichloroacetaldehyde), aldehydes, hydroxy chlorinated hydrocarbons, and
other related compounds.
Chloral and formic acid are two of the most prevalent side reaction
trace components generated during the oxychlor reaction. Both of these
trace compounds are biodegradable. A central biotreatment system would
remove chloral and formic acid.
Effects of Flow Rate on Adsorption Capacity
As expected, the capacity of the carbon to remove organic contaminants
from the wastewaters varied with the rate of flow for a consistent amount
of contaminants in the feed. Certain organic contaminants, such as chloral,
which are water soluble and form hydrates, affect the adsorption capacity
at different flow rates. These data are summarized in Tables 19 and 20.
36
-------
TABLE 19. EFFECTS OF FLOW RATE OH ADSORPTION CAPACITY FOR STREAM 221A
(Kg Adsorbed/Kg Carbon)
Total Organics
1
Feed
(kg) .
0.071
0.982
2.257
2.958
3.961
5.322
7.042
8.146
9.181
10.036
.64 1/min
Adsorption
(kg/kg carbon)
0.0023
0.0317
0.0728
0.0954
0.1277
0.1716
0.2270
0.2626
0.2924
0.3048
2.
Feed
(kg)
0.277
1.516
2.980
5.057
8.124
12.222
15.241
18.731
by Gas Chromatography
27 1/min
Adsorption
(kg/kg carbon)
0.0097
0.0531
0.1043
0.1771
0.2844
0.4005
0.4235
0.4241
3.
Feed
(kg)
6.029
10.984
12.297
17.706
23.429
31.775
41 1/min
Adsorption
(kg/kg carbon)
0.263
0.425
0.358
0.489
0.521
0.435
1
Feed
(kg) .
0.157
5.142
1.296
1.691
2.245
2.731
3.342
4.043
4.333
4.942
.64 1/min
Adsorption
(kg/kg carbon)
0.0046
0.0156
0.0348
0.0450
0.0596
0.0695
0.0823
0.0989
0.1040
0.1041
Total
2.
Feed
(kg)
0.150
1.109
1.658
2.388
3.208
4.576
5.374
6.884
Organics by TOD
27 1/min
Adsorption
(kg/kg carbon)
0.0052
0.0388
0.0486
0.0663
0.0717
0.0824
0.0778
0.0979
3.
Feed
(kg)
1.791
3.414
4.042
6.222
8.004
10.00
41 1/min
Adsorption
(kg/kg carbon)
0.072
0.122
0.123
0.159
-
-
37
-------
TABLE 20. ANALYSIS OF ADSORPTION DEMONSTRATION UNIT OPERATION BY INDIVIDUAL ORGANIC COMPONENT FOR STREAM 221A
U>
00
Test One
Carbon: Westvaco
Stream: 221A
Organic
Component in
Influent and
Effluent
Ethyl Chloride
Vinyl Chloride
Vinylidene Chloride
Flow: 1.
Loading: 0.
Columns : 15
Volume: 64
Total
Organics
To Columns
Before
Breakthrough
(mg)
4000
224
0
Trans-Dichloroethylene 0
1 , 1-Dichloroethane
Chloroform
Ethylene Dichloride
2944
1255
51100
1,1,1-Trichloroethane 555
1,1,2-Trichloroethane 2530
Chloral Hydrate
Chloral
Propylene Dichloride
51000
1 l/min2
74 I/cm /min
x 150 cm
.8 liters
Organics
Adsorbed
Before
Breakthrough
Test
Carbon :
Stream:
Flow:
Loading :
Columns :
Volume:
Total
Organic
Two
Test Three
Westvaco
221A
1.9 1/min,
1.23 I/cm min
15
64.
To columns
Before
x 150 cm
8 liters
Amount
Adsorbed
Before
Breakthrough
Breakthrough
(mg)
4000
224
0
0
NB
NB
51100
NB
NB
51000
(mg)
585
6829
0
0
13357
13401
155743
1726
2151
18459
102600
4155
(mg)
585
6829
0
0
NB
NB
155743
NB
NB
18459
102600
NB
Carbon:
Stream:
Flow:
Loading :
Columns :
Volume:
Total
Organics
To columns
Before
Westvaco
221A
2.85 1/min
1.84 1/m /min
15 x 150 cm
64.8 liters
Amount
Adsorbed
Before
Breakthrough
Breakthrough
(mg)
992
182
898
0
5250
2254
69369
-
3369
3920
63612
1318
(mg)
992
182
NB
0
5250
NB
69369
-
NB
3920
63612
NB
Test
Carbon:
Stream:
Flow:
Loading :
Columns :
Volume:
Total
Organics
Four
Filtrasorb
300
221A
1.
0.
15
64
1 1/min
71 1/m /min
x 150 cm
.8 liters
Amount
Adsorbed
To columns
Before
Before
Breakthrough
Breakthrough
(»g)
1034
31
19
0
2024
1391
65494
312
463
-
-
—
(mg)
1034
31
NB
0
NB
NB
65494
NB
NB
-
_
—
NB = No Breakthrough
-------
Effectiveness of Activated Carbon on Removing all Chlorinated Hydrocarbons
As discussed in the preceding section, the effectiveness of activated
carbon in removing the refractory chlorinated hydrocarbons was determined
by measuring the removal of total organics, total carbon, and total oxygen
demand in both the influent and effluent of the adsorption system. The
influent and effluent were also monitored to study the selectiveness of the
removal of specific organics from the process effluent. The monitoring for
the specific chlorinated hydrocarbon was done by gas ehromatography using a
flame ionization tester. This GC method determined all of the chlorinated
hydrocarbons quantitatively.
Activated carbon showed an excellent capacity for removing trace
levels of chlorinated hydrocarbons. Chloral and chloral hydrates did not
adsorb as readily as did the chlorinated hydrocarbons. When effluent
samples contained chloral and chloral hydrates in concentrations as great
as EDC, these streams exhibited the least adsorptive characteristics and
showed breakthrough about the same time as the ethylene dichloride.
Regeneration Results with the Shirco Infrared Furnace—
Regeneration of activated carbon utilized in stream 221A involved six
samples of four commercial carbons (Tables 21 and 22). After six sample
runs, operation was shut down because of the condition of the adsorption
columns. The chlorinated organics present in the wastewater attacked the
Plexiglas walls and produced crazed cracking. The tubes were replaced
with polypropylene reinforced with carbon black.
Since the test activated carbon was regenerated off-site, it was
stored in metal drums for transport. Minimum sealing was maintained since
the carbon would be heated for regeneration. Storage and weathering prior
to regeneration led to the removal of some 80 to 90 percent of the volatile
chlorinated hydrocarbons. Data at saturation showed from 0.2 to 0.45 kg of
adsorbed organics/kg activated carbon, while the analysis prior to regenera-
tion indicated only 0.01 to 0.08 kg organics present/kg activated carbon.
Results of the regeneration tests on all four commercial carbons are
shown in Table 21. Several samples had become dry through storage and were
run as received. Initially, a low temperature and long residence time were
utilized for regeneration. One test was made at a short residence time and
a higher temperature to evaluate these variables. The first run indicated
that residence time and condition of the carbon (wet or dry) had little
effect on the regeneration results, but that higher temperatures were
needed for complete regeneration. A second run was made to study the
effects of higher temperature on regeneration. These results are presented
in Table 22.
Since the earlier studies on stream 221A had shown Westvaco WVG to be
the best activated carbon in terms of organics removed per kilogram of
carbon, it was decided to run the iodine numbers for this carbon.
39
-------
TABLE 21. REGENERATION OF SPENT ACTIVATED CARBON - RUN NO. 1
Furnace Temperature °C
Activated
Carbon
Filtrasorb-300
Filtrasorb-400
Witco 718
Westvaco WVG #1
Westvaco WVG #2
Westvaco WVG #3
Residence Time Output
Condition
Wet
Wet
** Wet
** Dry
** Dry
Dry
(Minutes)
18-20
10-12
10-12
10-12
10-12
10-12
(kg/hr)
45.5
90.8
90.8
90.8
90.8
90.8
Lamp Section
1 & 2
371-427
399-454
399-454
399-454
399-454
399-454
3 & 4
677-732
732-788
732-788
732-788
732-788
732-788
*
Kg of Organics/Kg of AC
Initial
0.0137
0.0867
0.0455
0.0312
0.0349
0.0517
First Run
0.0064
0.0436
0.0232
0.0160
0.0214
0.0305
Analyses were made on a hexane extract of the activated carbon granules.
**
Refers to flow rate of Stream 221A while this activated carbon was in the Number 2 unit
Number 1, lowest; Number 2, medium; Number 3, highest.
TABLE 22. REGENERATION OF ACTIVATED CARBON - RUN NO. 1
Furnace Temperature °C
Activated
Carbon Conditions
Filtrasorb-300
Filtrasorb-400
Witco 718
Westvaco WVG #1
Westvaco WVG #2
Westvaco WVG #3
Dry
Dry
Wet
Dry
Wet
Wet
Residence Time Output
(Minutes) (kg/hr)
6-7
6-7
6-7
6-7
6-7
6-7
300
300
300
300
300
300
Lamp Section Kg of Organics/Kg of AC
1 & 2 3 & 4 Initial First Run
816-927
760-899
704-816
760-899
704-899
760-899
816-1066
927-954
760-927
927-982
954-982
899-927
0.0137
0.0867
0.0455
0.0312
0.0349
0.0517
0.00018
0.0041
0.00102
0.00060
0.00029
0.00070
-------
Iodine
Number
WVG Carbon as Received 1165
WVG #1 after regeneration at 927-982°C 1058
WVG #2 after regeneration at 954-982°C 993
WVG #3 after regeneration at 889-954°C 838
The difference in iodine numbers was not reflected in the organics removed.
Retesting of Westvaco Activated Carbon
Stream 221A was re-evaluated on a new batch of WVG activated carbon on
an adsorption-regeneration cycle. The demonstration unit was repacked with
a fresh batch of WVG carbon. After saturation, an adsorption capacity of
0.34 kg organics/kg carbon was observed (Tables 23 and 24). This was
almost identical to the initial test where capacity was 0.33 kg organics/kg
carbon.
The saturated carbon was transported to the site of the Shirco furnace
in polyethylene drums. The regeneration conditions for this batch were
approximately the same as those used for the earlier batches of carbon.
Conditions were as follows:
Carbon Sample Wet Westvaco
Furnace Temperature
Lamp Sections 1 and 2 496°C Average
Lamp Sections 3 and 4 737.5°C Average
Atmosphere Nitrogen
Residence Time 10 Minutes
The iodine number of the fresh WVG was 1150. After regeneration, it
was 1056. The amount of adsorbed organics determined by a GC analysis of a
perchloroethylene extract of the carbon was 0.138 kg organics/kg carbon
before regeneration, and 0.0008 kg organics/kg of carbon after regeneration.
These data indicate much better regeneration characteristics than the
previous run. However, it must be realized that small quantities of carbon
were being handled (22.3 kg/case) and that small losses are easily magnified
in the recycle process.
PROCESS EFFLUENT 041A
Introduction
This process effluent was selected for demonstration unit evaluation
because it was similar in organic contaminants to stream 221A; however, the
pH of stream 041A was 12, compared to pH 1 for stream 221A. The procedure
used to develop the demonstration unit data was much the same as that
followed for stream 221A.
41
-------
TABLE 23. ADSORPTION CAPACITY DATA FROM BREAKTHROUGH STUDY OF STREAM 221A
Ni
Hour
0-2
2-24
24-49
49-59
59-72
Liters
228
2120
5267
6680
8460
Influent Number
Kg Organics
Period Total
.745
4.075
4.731
4.264
4.433
.745
4.819
8.550
13.815
18.248
Effluent Number
Kg Organics
Period Total
0.000
0.000
0.301
0.740
4.380
0.000
0.000
0.301
1.041
5.421
Kg Percent
Adsorbed Adsorbed
.745
4.819
4.430
3.222
0.053
100.000
100.000
93.63
75.56
0.103
Cumulative
Kg
Adsorbed
.745
4.819
9.250
12.482
12.525
Cumulative Number
Kg of Organics/
Kg/Carbon
0.02024
0.1310
0.2515
.3391
.3405
-------
TABLE 24. STUDY OF THE ADSORPTION CAPACITY OF REGENERATED WESTVACO-WVG
COMPARISON OF THE EFFLUENT AND INFLUENT FOR STREAM 221A
Date
7/22
7/22
7/22
7/23
7/23
7/23
7/24
7/24
7/24
Water
Sample
1-01-1
1-11-E
1-21-E
1-02-1
1-12-E
1-22-E
1-03-1
1-13-E
1-23-E
2-01-1
2-11-E
2-21-E
2-02-1
2-12-E
2-22-E
COD
(mg/1)
1,192
446
225
995
562
512
1,554
1,394
1,166
1,110
1,138
1,549
1,525
1,232
898
TOD
(mg/D
1,759
264
136
1,804
624
460
1,716
1,812
1,105
2,392
477
1,198
3,638
585
620
TOG
(mg/1)
724
76
40
627
437
347
567
614
962
663
297
588
640
394
271
olllnics
(mg/1)
2,253
0
0
2,420
0
0
2,229
2,220
1,553
4,180
0
607
2,500
0
263
EDCb
952
0
0
917
0
0
963
756
167
3,019
0
177
1,274
0
78
Major
CHC
1,086
0
0
1,172
0
0
1,037
1,250
1,170
132
0.
191
994
0
122
Organics
ECd
387
0
0
105
0
0
59
152
190
327
0
239
128
0
63
DCEe
40
0
0
79
0
0
78
45
8
305
0
0
193
0
0
Acidity
(mg/1)
10,550
8,750
1,300
10,150
10,400
10,000
25,200
3,600
19,000
4,260
30,500
45,100
4,200
20,400
19,500
1st number designates run; 2nd, column //; 3rd, sample #.
The letters I and E stand for influent and effluent.
3Ethylene Dichloride.
^i
'Chloral Hydrate.
dEthyl Chloride.
B
1-1-Dichloroethane.
-------
1. Review of the activated carbon laboratory equilibrium data.
2. Demonstration unit operation.
3. In-situ steam regeneration.
4. Breakthrough data.
5. Further demonstration unit tests.
Location and Setup of Demonstration Unit
Stream 041A was first evaluated to determine the removal of trace
chlorinated hydrocarbons by carbon adsorption. Based on preliminary labora-
tory data, it was decided to use WVG, Filtrasorb, and Witco activated
carbon for this study. The stream was pumped through three carbon beds in
series at selected flows. Regeneration evaluations were made utilizing
steam passed through the carbon beds in situ.
Demonstration Unit Operation
The demonstration unit was set up adjacent to the production unit, and
the column was packed with the first carbon to be- tested, WVG. These first
test results are represented as run number one of Table 25. The second run
was initiated with Filtrasorb-400 charged to the unit. A high flow rate
(3.4 1/min) was used for this second run and resulted in a relatively quick
breakthrough as indicated by EDC measurements. The first three columns
were saturated in less than eight hours. The breakthrough was not followed
through each column since it occurred before the first set of samples was
completely analyzed. The capacity calculations listed in Table 25 for the
second run were based on three columns and resulted in high values when
compared with subsequent data.
The Witco 718 carbon, run number three (Table 25), appeared to be the
best carbon tested based on adsorption capacity before breakthrough of EDC.
This carbon removed 0.35 kg EDC/kg carbon. Unfortunately, this represents
one set of data with no confirmation.
The remaining capacity runs were made with the WVG carbon. Runs four
and five resulted in adsorption capacities of 123 g EDC/liter carbon, or
0.28 kg EDC/kg carbon.
In Situ Steam Regeneration
An important part of the carbon adsorption process is the regeneration
and neutralization of the carbon after it has become saturated. The conven-
tional regeneration technique is to remove the carbon and regenerate it,
burning off the adsorbed impurities in a multihearth furnace. After this
burnoff, the carbon is cooled and replaced in the original reactor for the
next cycle of operation.
44
-------
TABLE 25. SUMMARY OF CARBON ADSORPTION FIELD TRIALS ON STREAM 041A
Run carbon Flow Rate
Number Type 1/min
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
WVG
Filtrasorb
400
Witco
718
WVG
WVG
WVG
WVG
WVG
WVG
WVG
WVG
WVG
WVG
WVG
WVG
WVG
0.95
0.76
0.95
0.76
0.95
0.95
0.76
0.76
0.76
0.76
0.76
0.76
0.76
0.76
0.76
0.95
Total
Run Time
Hours
16-24
3
100
38
48
72
24
24
26
23
48
48
48
120
24
17.5
No of
Columns
1
3
5
3
3
5
1
1
1
1
1
1
1
1
1
1
Height
(m)
1.82
3.95
6.87
2.01
3.40
5.56
0.14
1.21
1.21
1.21
1.21
1.21
1.21
1.21
1.21
1.21
Weight
(kg)
5.66
16.6
61.4
15.4
27.1
44.4
9.71
9.71
9.71
9.71
9.71
9.71
9.71
9.71
9.71
9.71
Volume
liters
13.44
39.14
124.8
36.1
28.32
103.7
22
22
22
22
22
22
22
22
22
22
EDC Retained
Influent Effluent in Bed
mg/1 mg/1 (kg)
3600
3662
3508
2080
2750
3042
2300
2350
1800
2817
-
_
_
-
-
3067
_
0
<14
<50
<98
<50
_
-
-
249 Composite
-
_
-
-
_
80 Composite
3.45-4.90
5.72
21.5
5.26-6.12
7.76
125.8
_
-
-
2.72
-
_
-
-
_
2.72
Carbon Adsorption Capacity
(kg EDC) (kg EDC)
1 Carbon
0.26-0.36
9.2
0.17
0.11-0.12
0.12
0.12
_
-
-
0.12
-
_
-
-
-
0.12
kg Carbon
0.6-0.8
0.33
0.35
0.25-0.29
0.28
0.28
_
-
-
0.28
-
_
-
_
_
0.28
-------
An undesirable aspect of thermal regeneration of carbons saturated
with chlorinated organics is that the combusting gases contain hydrogen
chloride. The results of solving the effluent problem could be to create
an emission problem via this process. Steam regeneration of activated
carbons saturated with certain organics has distinct advantages. With
steam, it is possible to condense the hot vapors escaping from the reactor
outlet. If the stripped organics are immiscible, they can be decanted and
returned to the process. If they are soluble, it may be possible to separate
them from the condensate by fractionation. In either case, this regeneration
technique permits recovery and recycling of the contaminants.
One of the objectives of this program was to evaluate steam regeneration
of carbon. Toward this end, provisions were made to steam-strip the carbon
after saturation with EDC from the process effluent. The results of these
tests (runs six through sixteen) are presented in Table 26. Each run was
preceded by up to 24 hours of low pressure steam-stripping.
Runs six, ten, and sixteen were analyzed for EDC. Subsequent testing
yielded the same adsorption capacities as did runs four and five, which
used fresh carbon.
This experimental work pointed up several salient aspects of the
utilization of activated carbon to recover EDC from this process effluent.
1. The carbon bed could be steam-stripped of the EDC in situ.
2. The original capacity of the carbon was completely restored.
3. There was no apparent effect of repeated regeneration on the
physical integrity of adsorption capacity of the carbon.
4. After 11 in situ steam regenerations of the same carbon bed,
there were no signs of "bed wear."
Breakthrough Data
Each of the adsorption columns was monitored for EDC while evaluating
the selected commercial activated carbon installed for the test. Tables
26-30 follow the breakthrough of EDC through each column of the demonstration
unit.
Further Demonstration Unit Tests on Stream 041A
Three additional carbon adsorption runs were made on stream 041A. One
involved a recheck of the adsorption capacity of Filtrasorb-400, and the
Other two were studies of the adsorption capacity of new and regenerated
Monochem activated carbon.
The waste in stream 041A was composed of about 99 percent by weight of
ethylene dichloride. The EDC concentration in the influent and effluent
were determined by gas chromatography. Earlier field studies of stream
041A indicated that Filtrasorb-400 and Witco 718 had higher adsorption
46
-------
TABLE 26. BREAKTHROUGH DATA WITH FILTRASORB-400 (STREAM 041A)
Total
Liters
1640
3276
Feed
3662
Flow: 3.42 1/min.
EDC (mg/1)
Column
1234
3446
3819
3766
3472
3409
3598
Feed
Total Carbon (mg/1)
Column
2345
Feed
Loading: 1.61
Pressure Drop (atm)
Column
12345
0 0 1220
3851 3693 -
1530
1330
1270
1360
1470
1470
320
1260
200" 0.612
1240 2.00
0.477
1.80
0.374
1.53
0.136
1.36
0.136
1.34
0
1.24
TABLE 27. BREAKTHROUGH DATA WITH WITCO 718 (STREAM 041A)
Total
Liters
170
795
1389
1643
2188
2612
2953
3407
3861
4315
4770
5224
5678
6132
6587
7052
Flow: 0.95 1/min
EDC (mg/1)
Total Carbon (mg/1)
Column
Feed
3508
-
3409
-
_
3560
_
_
3566
-
-
3377
-
-
3914
-
1
0.8
1780
2588
2999
4135
3251
3685
3283
3819
3693
-
3377
3377
3377
3914
3914
2
4.8
0.9
61.9
7.4
7.6
1570
3756
3472
3724
3598
-
3377
3377
3377
3914
3914
3
4.6
0.7
13.9
7.9
6.5
17.5
1.9
5.4
10.7
3819
-
3377
3377
3377
3914
3914
4
6.8
1.7
7.3
5.7
6.8
12.9
1.7
6.9
16.0
4.2
-
2525
3387
3377
3914
3914
5
1.5
0.4
4.0
10.7
3.6
8.2
3.6
7.3
22.9
5.9
-
3.2
14.3
2765
3787
3787
Feed
2183
-
2183
-
-
2300
-
-
1815
-
-
1554
-
2765
1649
-
1
_
484
1478
1934
1022
1045
-
-
-
-
-
-
-
-
-
-
Loading: 0.45
Pressure Drop
Column
2
_
387
-
497
359
510
-
-
-
-
-
-
-
-
-
-
3
_
-
-
-
553
-
499
-
538
510
-
-
-
-
-
-
4 5 Feed
- - 0
- - 0
0.14
0.14
499
_
_
_
- 0.07
0.07
0.14
475
463
0.14
0.21
_
•1
0
0
0
0
0
0
0
0
0.07
0.07
0.07
-
_
0.14
0.21
_
1/min M2
(atm)
Column
2
0
0
0
0
0
0
0
0
0
0
0
-
_
0
0.07
_
3
0
0
0
0
0
0
0
0
0
0
0
_
_
0
0.07
_
4
0
0
0
0
0
0
0
0
0
0
0
_
_
0
0.07
_
5
0
0
0
0
0
0
0
0
0
0
0
_
_
0
0
_
-------
TABLE 28. BREAKTHROUGH DATA WITH WESTVACO (STREAM 041A)*
-e-
oo
Total
Liters
273
636
999
1363
1726
2090
2553
2816
3180
EDC (mg/1)
Column
Feed 1
128
2194
2398 2241
- -
- -
1768
- —
- -
— —
2
42
15
939
-
-
-
-
—
345
_ _ _
_ _ _
118
_ _ _
1033
1957
_ _ _
1736
Total Carbon (mg/1)
Column
Feed 12345
372
736
440
- 1600 -
_ _ _ _ _ _
3500 _____
_ _____
_ _ _ _ _ _
— _____
Flow: 0.76 1/min
2
Loading: 0.36 1/min-m
-------
TABLE 29. BREAKTHROUGH DATA WITH WVG (STREAM 041A)*
EDC (mg/1)
Total
Liters
363
818
127
1726
2180
2635
3089
Feed
2557
-
-
2730
-
-
2983
-
Column
1 2
35 30
87
63
46
B.T.
- -
- -
3 4
-
- -
24
- -
17
B.T. 98
260
33
5
8
16
2
32
40
48
-
Flow: 0.95 1/m
2
Loading: 4.80 1/min-m
TABLE 30. BREAKTHROUGH DATA WITH WVG (STREAM 041A)*
EDC (mg/1)
Total
Liters
454
908
1362
1817
2271
2726
3180
3634
4111
Feed
-
-
2919
-
-
2768
—
-
3440
1
0
2
962
-
-
-
-
-
-
-
Column
2
0
-
0
14
1578
2815
2615
-
-
-
3
0
-
0
0
36.5
39
2615
-
-
-
4
0
0
0
0
0
0
24
694
-
—
5
0
0
0
0
0
0
0
37
53
300
*
Flow: 0.95 1/min
2
Loading: 4.81 1/min-m
49
-------
capacities for EDC than the WVG. These findings were contrary to lab
studies which showed that WVG was the best for both TC and EDC. To reconcile
this discrepancy, one column was repacked with fresh Filtrasorb 400 and a
run was made to determine the adsorption capacity. The data shown in Table
26 indicated that the Filtrasorb 400 did have a smaller adsorption capacity
for EDC than the WVG. The capacity variance could be explained by lot-to-
lot variation from a single commercial source.
The other two runs were to determine the adsorption capacity of the
Monochem activated carbon (AC). Laboratory studies had shown that its EDC
adsorption capacity was about one-half that of the best commercial AC.
Since there was less than 25 kg of the Monochem activated soot carbon, only
one column was packed. Unfortunately, the column plugged shortly after
start-up. Copious quantities of carbon fines were found blocking the
effluent line. Two hours of backflushing solved the problem, and the run
was restarted. The breakthrough occurred after 23.5 hr. It was found that
the Monochem carbon had an adsorption capacity of approximately 0.16 kg
EDC/kg carbon (Table 31). This capacity confirmed earlier laboratory
studies since the value was around one-half of the capacity of WVG (0.28 kg
EDC/kg carbon).
To determine the effectiveness of steam regeneration on the Monochem
carbon, the column was regenerated for 48 hr with 43.3°C steam (0.2-0.34
atm). These two runs confirm earlier findings on the adsorption capacity
of the Monochem carbon and indicate that it could be steam regenerated just
as readily as the commercial activated carbons.
PROCESS EFFLUENT 081D AND 081A
Introduction
After completing the activated carbon adsorption field tests on stream
041A, it was decided to study a process effluent whose contaminants were
more highly chlorinated and whose boiling points were much higher than that
of ethylene dichloride. Such organic compounds are tetrachloroethane,
perchloroethylene, and trichloroethylene. The higher boiling points of
these compounds would make adsorption and subsequent regeneration more
difficult than the test work done on stream 041A.
Tests on Process Effluent 081A
The test program was divided into five distinct phases.
1. Breakthrough data for WVG carbon. This commercial carbon had
exhibited the overall best adsorptive characteristics for compara-
tive stream service.
2. Furnace regeneration adsorption.
3. Preliminary in situ stream regeneration.
50
-------
TABLE 31. SUMMARY OF CARBON ADSORPTION FIELD TRIALS ON STREAM 041A
Run
No.
17*
18
19
Carbon Flow Rate
Type (1/min)
Monochem
Monochem
Filtrasorb
400
0.84
0.84
0.8
Total
Run Time
(hrs.)
23.5
19.5
19
Carbon Bed
No. of
Columns
1
1
1
Height
(m)
1.40
1.40
1.63
Weight
(kg)
10.2
10.2
11.9
Volume
(m3)
0.025
0.025
0.03
Average EDC Concentration
Influent
(mg/1)
2468
3100
1794
Effluent
(mg/1)
1115
974
37
Carbon Adsorpt
EDC Retained £DC)
X(k ) (m3 Carbon)
1.59
2.09
1.52
65.35
85.85
53.42
''
ion Capacity
(kg EDC)
(kg Carbon)
0.16
0^2-Q
•0..13
Run followed by 48 hours of stream regeneration of carbon (in situ).
TABLE 32. ADSORPTION CAPACITY DATA FROM BREAKTHROUGH STUDY OF STREAM 081A
(Westvaco WVG #1 Activated Carbon)
Hour Liters
0-3 681
3-6 1363
6-9 2044
9-12 2736
12-15 3407
15-18 4088
18-21 4845
21-24 5451
24-27 6132
27-30 5814
30-33 7495
33-36 8177
36-39 8858
39-42 9539
42-45 10,932
45-48 10,932
48-51 11,537
51-54 12,218
54-57 12,787
57-60 13,469
Influent Effluent Cummulative
Number of Organics Number of Organics Number of
Kilograms Kilograms Organics Cummulative Number of
3-hour Cummulative 3-hour Cummulative KilogratiB Percent Adsorbed Organics Adsorbed
Period Total Period Total Adsorbed
0.299
0.273
0.272
0.065
0.249
0.178
0.140
0.138
0.063
0.025
0.146
0.130
0.087
0.095
0.033
0.032
0.035
0.090
0.036
0.19
0.390
0.663
0.934
1.000
1.248
1.426
1.566
1.704
1.767
1.791
1.937
2.067
2.155
2.249
2.282
2.314
2.349
2.439
2.476
2.495
0.047
0.006
0.007
0.034
0.047
0.002
0.134
0.190
0.109
0.128
0.123
0.081
0.105
0.083
0.062
0.081
0.349
0.318
0.039
0.049
0.047
0.053
0.060
0.094
0.141
0.143
0.277
0.467
0.576
0.704
0.827
0.908
1.013
1.096
1.208
1.239
1.274
1.305
1.345
1.393
0.343
0.266
0.266
0.031
0.202
0.176
0.006
0.052
0.046
0.103
0.02
0.045
0.018
0.011
0.029
0.049
0.000
0.058
0.002
0.029
Adsorbed (Kilograms) (Kilograms of Carbon)
88.02 0.343
97.67 0.610
97.49 0.875
47.22 0.905
81.20 1.107
98.97 1.283
4.20 1.289
37. 50 added 1.237
47. 82 added 1.191
422.22 added 1.088
15.52 1.107
36.63 1.159
20. 31 added 1.142
11.96 1.153
86. 30 added 1.124
154. 28 added 1.075
00.00 1.075
64.64 1.134
6. 17 added 1.131
154. 76 added 1.102
0.075
0.134
0.194
0.201
0.245
0.287
0.287
0.273
0.265
0.240
0.245
0.256
0.254
0.256
0.249
0.238
0.238
0.251
0.251
0.245
Kilograms of Organics
Through Unit
0.390
0.663
0.934
1.000
1.249
1.426
1.566
1.704
1.767
1.791
1.937
2.067
2.155
2.249
2.282
2.314
2.349
2.439
2.476
2.495
-------
4. Expanded adsorption field studies.
5. Split sample analytical comparison.
Breakthrough Data
The data gathered from the breakthrough run on stream 081A are shown
in Table 32. The breakthrough curve was determined using a hydraulic
loading of 3.89 1/m /min with a total volume of 28 liters of WVG carbon in
one column, at an upflow rate of 3.8 1/min. This hydraulic loading re-
presented the maximum value to be studied.
2
The EPA "Process Design Manual for Carbon Adsorption" states that in
the range of 0.7-3.5 1/m min, the hydraulic loading per se seems to have
little effect on adsorption capacity. Table 32 indicates an observed
adsorption capacity of approximately 0.12 kg organics/kg activated carbon.
This value was much less than the 0.30 value observed for this same carbon
for streams 221A and 041A. Since the hydraulic loading value was very
high, it is possible that the maximum loading was exceeded, thus affecting
the adsorption capacity. It is also possible that streams containing more
highly chlorinated C-2 hydrocarbons with higher boiling points behave
differently than those containing mainly EDC.
Shirco Furnace Regeneration
After saturation by stream 081D, the activated carbon (WVG) was removed
and transported in drums to the site of the Shirco furnace for regeneration.
The regeneration temperatures and organics removal are shown in Table 33.
It would appear that over 95 percent of the organics were removed at listed
process conditions.
TABLE. 33. REGENERATION OF WESTVACO .WVG //I ACTIVATED CARBON
Stream 081D
Prior No. of Regenerations 0
Mat Condition Wet
Residence Time (min) 3
Output (kg/hr) 181
Furnace temperature, lamp section 1300-1500
kg organics/kg activated carbon
Saturated condition —
Prior to Regeneration 0.0352
Final 0.0009
Furnace Blanket Gas Nitrogen
Preliminary In Situ Steam Regeneration
Initial tests utilizing 0.20 to 0.34 atm steam required over 60 hr to
reduce the TOG and TOD levels in the condensate below 5 mg/1. This long
generation time was attributable to the presence of tetrachloroethane,
52
-------
which boiled at 148.9°C, while the regeneration steam temperature was only
110°C. To accomplish this regeneration, sufficient steam would have to be
present to lower the partial pressure of the tetrachloroethane sufficiently
to cause boiling.
Expanded Adsorption Field Studies
As demonstrated earlier, lower adsorption capacities and longer steam
regeneration times were found for activated carbon saturated by the chloro-
organics in stream 081A. To study the effect of higher temperature steam
regeneration, two Teflon-lined steel columns were installed for further
test work on stream 081A.
Table 34 shows a summary of all successful adsorption runs and indicates
the total organics present in the influent, effluent, as well as the kilo-
grams of organics adsorbed per kilogram of carbon. The average adsorption
capacity of WVG carbon for organics in stream 081A was 0.15 kg/kg carbon.
This was approximately one-half the value found for this carbon on streams
221A and 041A.
EDC was the main impurity in streams 041A and 221A. Stream 041A
exhibited no chloro-organic heavier than EDC. Stream 221A showed heavier
chlorohydrocarbons present in trace concentrations. Stream 081A exhibited
chlorohydrocarbons impurities much heavier in greater concentrations than
those in stream 041A.
Table 34 shows the average concentration of each component in the
influent versus the average percent adsorption for each run. In the previous
field tests involving adsorption, there was little marked contrast in the
adsorptivity of each component. The data in Table 35 indicate the perchloro-
ethylene is most strongly adsorbed, followed by trichloroethylene, tetrachlo-
roethane, and ethylene dichloride.
The adsorption capacity of WVG carbon apparently was greatly affected
by the presence of the larger molecules of perchloroethylene, trichloro-
ethylene, and tetrachloroethane. These three chlorinated hydrocarbons also
had a much higher boiling point. All three compounds were present in
sufficient quantities to prevent EDC from dominating the adsorption process.
Since the new columns were constructed of Teflon-lined steel, it was
possible to regenerate them with an in-plant live steam (Table 36). The
steam for regeneration was saturated at 120.1°C, a temperature which has
been sufficient to regenerate the carbon completely in streams 041A and
221A.
The data indicate that the steam temperature for stream 081A will have
to be higher for complete regeneration. For example, the presence of
tetrachloroethane with a boiling point of 145°C indicates the necessity^of
higher steam temperatures for regeneration, particularly when removal times
are shortened to minimize recycling time and equipment usage.
53
-------
TABLE 34. COMPARISON OF ALL SUCCESSFUL RUNS FOR THE CARBON ADSORPTION UNIT
AT SATURATION: STREAM 081D
Average
Carbon Influent Cone.
Run
1A
1
*2
3
4
5
Characteristics
21.9 kg
15.5 kg
15.5 kg
15.5 kg
11.3 kg
11.3 kg
- New
- New
- Reg. 1
- Reg. 2
- New
- Reg. 1
(mg/D
349
180
148
89
101
100
Average
Effluent Cone
(mg/D
36
52
26
27
35
31
Saturation
Time
18 hrs.
43
26
33
54
39
Kilograms
Adsorbed
1.2
1.2
0.7
0.44
0.8
0.57
kg Adsorbed
kg of Carbon
0.12
0.17
0.095
0.062
0.15
0.11
Percent
Adsorbed
90
71
82
70
66
69
Adsorption and Steam Regeneration Data Do Not Agree.
Ul
TABLE 35. ADSORPTION DATA ON THE INDIVIDUAL COMPONENTS FOR THE
CARBON ADSORPTION UNIT: STREAM 081D
Kg of Component / Kg of Total
Through Column / Organic Through Column
Run
1A
1
2
3
4
5
TETRA
0.52
9,61
0.73
0.73
0.54
0.68
PERC
0.02
0.11
0.05
0.12
0.12
0.08
TRI
0.03
0.09
0.15
0.09
0.15
0.17
EDC
0.34
0.13
0.05
0.04
0.11
0.05
Kg of Component /
Adsorbed /
TETRA
0.88
0.81
0.85
0.70
0.74
0.65
PERC
0.98
0.98
0.96
0.96
0.96
0.99
Kg of Component
Through Column
TRI
0.98
0.78
0.93
0.77
0.79
0..76
EDC
0.91
0.10
-0.05
-0.32
-0.11
0.50
-------
The use of steam stripping as a means of regenerating the fixed acti-
vated carbon beds has significant technical importance. The exit vapors
could be condensed and the organic phase separated and recycled to the
process. The aqueous phase could be recycled to the plant effluent. Thus
a closed-loop effluent treatment process could be effected, with by-product
credit to reduce equipment expenses.
ECONOMIC DISCUSSION
Field trials of the adsorption demonstration units were not very
promising. Overall results on stream 041A indicated that, down to a level
of 25 mg/1 in the treated effluent, approximately 1 kg of activated carbon
was needed to remove 0.1 kg of organics. The spent carbon, however, can be
regenerated in place with a low-pressure steam. The exhaust vapors from
this regeneration technique can be condensed and recycled to the separation
process.
Because of the low adsorption efficiency, the operating cost for the
adsorption system would be $3.50 to $4.50 per kilogram of organics removed.
For average inlet concentrations of 2000 mg/1 EDC and a nominal flow of 380
1/min from a 32-million kg/yr ethylene dichloride plant, this cost would
amount to 4.5 to 7 cents per kilogram of product.
TABLE 36. STEAM REGENERATION DATA FOR THE CARBON ADSORPTION UNIT
STREAM 08ID
Run
1
2
4
5
Steam
Pressure
30 psig
15
15
15
Flow
Rate
(kg/hr)
67
42
41
40
Regeneration
Time3
8.2 hr.
6.3
6.5
4.75
Average
Condensate Cone.
(mg/liter)
599
-
455
-
Kilograms
Recovered
1.14
1.0
0.54
1.24
Percent
Recovered
94
157
70
178
It was observed that 90 percent of the removable organics were recovered in
the first two hours.
55
-------
SECTION 6
MONOCHEM SOOT CARBON
Soot carbon is generated as a by-product in the manufacture of acetylene
via high temperature pyrolysis of air and methane. This material generally
accumulates in the process quench water. The soot carbon saturated quench
water flows to a settling basin. The carbon floats to the top and is
recovered and filtered to a 20 percent carbon and 80 percent water mixture.
The consistency of this mixture is that of a fine mud.
Monochem, Inc., which operates an acetylene plant at Geismar, Louisiana,
was an active participant in the grant program. By-product soot carbon from
this operation approximates 9-million kg/yr. Preliminary indications were
that the Monochem soot carbon, if properly prepared, would make a satisfac-
tory activated carbon. In-plant work confirmed that the soot carbon, when
predried, did show evidence of activation. However, these evaluations were
elementary and not at all conclusive. Therefore, the following program was
established to make a definitive evaluation of the adsorption capability of
the soot carbon.
1. Preliminary evaluation of Monochem soot carbon for application as
a commercial activated carbon.
2. Laboratory and field testing.
3. Developmental program for improving activity and mechanical
strength.
4. Laboratory evaluation of binders for better granulation and
activation.
5. Compounding, pelletizing, and activating of Monochem soot carbon.
6. Optimization of compounding and activating of Monochem soot
carbon.
7. Pilot work on activated carbon density, activation, and regeneration.
8. Isotherm and breakthrough data for Monochem activated soot carbon
compared with several commercial carbons.
56
-------
PRELIMINARY EVALUATION FOR APPLICATION AS A COMMERCIAL ACTIVATED CARBON
The soot carbon discharged from the filter drum had the consistency of
a fine mud and a water content of approximately 80 percent. It was necessary
to have a final granulated product that could be examined and tested for its
adsorptive characteristics. A preliminary experimental program was devised
to determine the feasibility of developing a commercially acceptable product.
The program included the following steps.
1. Drying of sufficient material to granulate for initial tests and
evaluations.
2. Granulating the dried soot into a material suitable for activation
and testing.
3. Laboratory scale infrared furnace activation.
4. Laboratory experimental results by iodine number.
5. Laboratory evaluation by isotherm determinations.
6. Laboratory adsorption breakthrough data.
7. Thermogravimetric measurements.
Test samples of granulated soot were obtained by drying the raw material
in a commercial steam-heated tube dryer. Stack losses were observed during
this process, and may present a problem in commercial practice. Initial
granulation was carried out by K-G Industries, Rosemont, Illinois, with a
50-60 CS press using 4.25-11 cm corrugated rolls. Other attempts at pelletiz-
ing were made with Simpson Mix Muller, with a ribbon blender, and with a
Carver hydraulic bench press.
Batch activation in the laboratory was carried out in metal trays in a
furnace that could operate up to 980°C. Preheating to drive off volatiles
was investigated at lower temperatures. Steam or other inert gas blankets
were furnished to the furnace. The heat transfer was by infrared radiation,
leading to an efficient use of power because of the high absorption coeffi-
cient of the carbon (black body).
The effects of preheating and activation schedules with two atmospheres
during each stage were investigated. The use of a laboratory sized Shirco
oven operating on infrared radiation principles offered several advantages
in operations.
1. Flexibility of Operation—The preheat and activation temperatures
could be obtained quickly because of the application of the heat to
the carbon granules.
2. Flexibility of Runs—Since only the carbon is heated, a second run
could be made quickly all the way from preheat to activation heat
without waiting for the box to cool down.
57
-------
3. Easy Maintenance—Lamps could be replaced readily. Insulation
materials were unaffected by the ups and downs of temperature.
For the preliminary experiments, 50 g of pelletized carbon was placed in
the oven. The temperature was recorded by a probe just above the sample.
Each experimental trial was evaluated by measurement of iodine number. The
results of these analyses, together with the experimental activation condi-
tions are shown in Table 37. Iodine numbers as high as 650 can be obtained
from the soot carbon; this compares to values in the range of 80-1,300
measured by the same method for commercial products.
To provide data for isotherms of ethylene dichloride adsorption on the
experimentally produced activated carbons, it was necessary to combine
batches to obtain adequate sample size. The combinations of trials are
listed below:
Blend A: Run # 17, 18, 19, 24, Iodine No. 200 to 300
Blend B: Run # 27, 28, 29, Iodine No. 500 to 600
Blend C: Run # 30, 31, 33, Iodine No. 400 to 500
Blend D: Run # 39, 43, 46, 47, Iodine No. 600 to 700
The adsorption isotherms are shown in Figure 8 for four blends derived from
the initial experiments. Blends A and D show high capacity for EDC but only
at relatively concentrated aqueous solutions. Blends B and C have a some-
what lower capacity, but operate better at lower equilibrium concentrations.
The EDC adsorptivity of the Monochem carbon was compared with that of
three commercial carbons using stream 221A. At a residual EDC concentration
of 6000 mg/1, the following adsorptivities were observed:
Monochem (Blend C) 0.75 g EDC/g carbon
Westvaco (W?G - 12 X 40) 1.5 g EDC/g carbon
Calgon (FS - 400) 0.45 g EDC/g carbon
Witco (Witco - 18 x 40) 0.40 g EDC/g carbon
Thus, Monochem Blend C showed better adsorptivity than Calgon (FS-400) or
Witco (18 x 40) at 6000 mg/1 EDC concentration. However, the commercial
carbons showed good adsorptivity over a wider range of EDC residual con-
centration (10 to 10,000 mg/1) than the Monochem Blend C.
The loss of mass on activation with the schedules shown in Table 37 is a
cause of concern. To search for the effect of heat and environment on mass
loss, a series of experiments was performed with a Perkin-Elmer thermogravi-
metric analyzer. Dry powdered soot was subjected to a programmed thermal
rise while on the microbalance of the instrument. The atmospheric conditions
in the balance compartments were controlled. The results of this analysis
are summarized in Table 38.
Equilibrium adsorption experiments indicate the limiting capacity of
the carbon as a sorbent. However, in practice, the solution and sorbent do
not operate under equilibrium conditions.
58
-------
TABLE 37. ACTIVATION STUDIES - I
Run No
1
2
3
4
Darco
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Pre-
Activation
Time
Pre-
Activation Activation
Time Gas
Activation
Gas
Pre-
Activation
Temp °C
Activation
Temp °C
S-51, as received
10
15
10
10
15
15
15
15
15
5
5
5
5
5
5
5
5
5
0
5
5
5
5
10
10
5
5
5
5
5-5
10
10
10
15
15
15
5+10
5+10
15
15
20
15
10
5 + 7 1/2+7
5+7 1/2 + 7
5 + 7 1/2 + 7
10
15
10
10
15
15
15
15
15
10
10
10
10
10
10
20
10
10
25
20
20
20
20
20
20
20
20
10
10
5-5
10
10
10
15
15
15
15
15
15
15
15
15
15
1/2 15
1/2 15
1/2 15
Witco Chemical Company
Filtrasorb 400 Calgon
Nitrogen
Nitrogen
Nitrogen
Steam
Nitrogen
Nitrogen
Nitrogen
Steam
Steam
Steam
Steam
Steam
Steam
Nitrogen
Nitrogen
Nitrogen
Nitrogen
Nitrogen
Nitrogen
Nitrogen
Nitrogen
Nitrogen
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
18x40 mesh, as
Corporation, as
Westvaco Chemical Division 12x40, as
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Steam
received
received
received
427
427
393
393
371
371
371
371
371
371
371
371
371
371
371
371
371
371
700
371
371
371
427
427
371
371
482
482
427-538
482-538
482-538
482-538
427
427
427
260+427
260+427
427
427
371+538
-15 - 93
538
260+427 + 538
260 + 427 + 538
260 + 427 + 538
Barneby-Cheney NB-9377
51
15
15
Steam
Steam
538
927
927
927
927
927
927
927
927
927
760
816
871
927
760
760
760
871
927
927
816
871
816
927
871
927
871
927
871
927
871-888
943 (max)
871
871
871
927
871(max)
871
871
927+982
871
871
-9 - 982
-8- 999
888
893
899
938
Iodine,
Number,
< 207
< 177
< 185
< 187
511
410
445
328
416
417
341
429
449
497
< 182
< 204
242
338
227
266
254
405
517
365
275
372
298
595
569
547
457
484
386
457
381
568
390
650
541
651
333
636
581
659
548
539
664
648
507
484
563
1264
1248
1314
817
342
Yield
Percent
69
55
62
64
63
61
73
71
59
81
56
74
7-6
84
79
79
75
64
73
77
74
77
47
49
49
63
50
73
63
72
51
79
39
54
40
85
61
61
38
59
63
32
38
61
65
59
f O
68
59
-------
10
00
13
O
CO
0)
4-1
iH
O
1.0
A Blend A
o Blend B
X Blend C
a Blend D
1 10 100
Residual EDC Concentration (g/liter)
Figure 8. EDC Adsorption isotherm for Monochem
activated carbon (stream 221A).
TABLE 38. THERMOGRAVIMETRIC MEASUREMENTS
(Programmed Rate of Heating: 20°C/min)
Medium
Air
Nitrogen
co2
Temperature
Range °C
0-800
0-800
0-700
Percent
NEX1
87.7
27.5
56.4
Weight Loss
EX2
27.6
9.9
18.9
•'"NEX:
2 EX:
Not Extracted with Methylene Chloride
Extracted with Methylene Chloride
Through use of breakthrough measurements, mass-transfer parameters were
estimated for column design and for performance comparisons of different
carbons under dynamic conditions. Four combinations of carbon and waste-
stream were used in establishing the dynamic behavior expected in the field
trials. Two characteristic streams (streams 041A and 221A) were used. The
sample of 041A used for these experiments was high in chlorinated hydrocarbons
(approximately 8,000 mg/1), was highly alkaline, and showed a total carbon
by the Beckman method of approximately 1200 mg/1.
60
-------
Three carbons were used during the trials: WVG carbon, which previously
had shown a high equilibrium adsorption of EDC and total carbon; Filtrasorb-
400, which had also shown a high adsorptive capacity; and Monochem activated
soot, treated in the laboratory (first activation) to a medium level of
activation. The limited amount of Monochem carbon precluded carrying out the
complete test program with this carbon.
For each combination, three breakthrough curves were generated at flow
rates denoted by L (low) , M (medium) and H (high). The range of flow rate
was:
L: 0.07 1/min m* to 0.14 1/min nu
M: 0.14 1/min nu to 0.25 1/min m^
H: 0.25 1/min m to 0.35 1/min m
Filtrasorb-400 and WVG performed equally in the reduction of EDC for
stream 041A (breakthrough curves are included in the Appendix). However,
the data indicate that bed life of Monochem sample studies would be only 40
to 50 percent as long as for the two commercial carbons. The Filtrasorb-400
appears to be less sensitive to flow rates than is the WVG in the range
studied. The Monochem carbon apparently was not affected by flow rates.
For stream 041A, the minimum TC level obtainable in the carbon column
effluent was about 300 mg/1, independent of the type of carbon used. Highest
capacity for the solutes in this stream is shown by WVG, followed by Filtra-
sorb-400, and finally, Monochem. Again, little sensitivity to the flow
rates was observed for Monochem. The data indicated that a species other
than chlorinated hydrocarbons must contribute to the TC level, and that this
species is poorly adsorbed by carbons.
LABORATORY AND FIELD TESTING
Experiments conducted thus far indicated that the soot carbon had
potential merit as a commercial activated carbon. The following program was
formulated to upgrade the soot carbon to acceptable levels.
1. Drying of sufficient soot carbon for in-depth laboratory and field
tests.
2. Granulation of this material by a commercial specialist.
3. Activation in a Herrschoff furnace.
4. Laboratory isotherm and breakthrough measurements.
5. Field tests with activated Monochem soot carbon.
6. Shirco furnace for field activation and regeneration.
61
-------
Drying, Granulating, and Activating Monochem Soot Carobn
Approximately 1400 kg of the emulsified soot was dried and shipped
to K-G Industries for granulation. Due to the fluffiness of the dried
material, considerable losses as particulate were noted in the stack of the
dryer. This experiment indicated an alternate means of dewatering the raw
by-product would be necessary to achieve a granulated product.
The dried soot received by K-G was mixed with 10 percent pitch and
granulated by passing the mixed powder through rollers. After sizing, the
fines were recycled and the specified material restored in drums. Due to
processing losses, only half of the granulated material was recovered. The
product exhibited some deficiency in hardness characteristics.
Bartlett-Show, Inc., of California, was contracted to activate the
granulated soot in a small Herrschoff furnace over a two-day period. The
activation temperature was between 871°C and 927°C. Activation in a multi-
hearth furnace revealed (1) the granulated material produced could be
activated in this type of furnace and (2) the activated carbon product could
be used for development purposes. However, only one drum of activated
product was produced from the five drums of feed. A program to generate an
additional supply of activated soot carbon was planned after installation of
the field activation and regeneration unit.
Laboratory Isotherm and Breakthrough Measurements
Isotherm and breakthrough measurements were carried out for a batch of
activated Monochem soot carbon. The granular activated carbon sample was in
the form of a very thick slurry in water and contained a considerable amount
of carbon fines which were washed off with copious quantities of water. The
granules were oven dried at 200°C for approximately 48 hr to ensure complete
drying of the carbon. The dried carbon granules were used to make adsorption
isotherm and breakthrough measurements. Standard experimental procedures
were used for these measurements. The plots are included in the Appendix,
and interpretation of these results follows.
Field Tests with Activated Monochem Soot Carbon
As evident from Table 39 at 6000 mg/1 EDC, Monochem adsorptivity was 55
percent or less than the adsorptivity of Filtrasorb-400, Witco, and WVG.
However, at 1000 mg/1 EDC, the adsorptivity of the Monochem carbon was 75
percent or more of the adsorptivity of the commercial activated carbons.
The Monochem activated carbon exhibited a better EDC adsorptivity for
stream 221A than for stream 041A. Weber attributes increased organics
adsorption to the decreased pH of the water (pH=l for stream 221A and pH=ll
for stream 041A). According to Weber, the increased hydrogen ion concentra-
tion neutralizes negative charges at the surface of the carbon, thereby
reducing hindrance to diffusion and making available more of the active
surface of the carbon. The field trials on stream 041A are summarized in
Table 40.
62
-------
TABLE 39. ABSORPTIVITY OF MONOCHEM AND COMMERCIAL CARBONS
(Stream 221A)
OJ
Carbon
Adsorptivity at
6000 mg/1 EDC
g EDC/g C
FILTRASORB-400 0.4
WITCO 0.45
WVG 1.5
MONOCHEM 0.22
TABLE 40. SUMMARY OF CARBON ADSORPTION FIELD
Run
No.
17*
18
19
Total
Carbon Flow Rate Run Time No. of
Type (liters/min) (hours) Columns
Monochem 0.836 23.5 1
Monochem 0.836 19.5 1
Flltrasorb 0.76 19 1
400
Carbon Bed Average EDC
Height Weight Volume Influent
(m) (kg) (n>3) (mg/1)
1.39 10.2 0.083 2468
1.39 10.2 0.083 3100
1.63 11.91 0.10 1794
Adsorptivity at
1000 mg/1
g EDC/g C
0.15
0.18
0.20
0.15
TRIALS ON STREAM 041A
Concentration EDC Retained Carbon Adsorption Capacity
Effluent in Bed (kp
(mg/1) (kg) (m3
1115 1.6 62.
974 2.1 81.
37 _ 1.52 50.
\_ EDC) (kg EDC)
carbon) (kg Carbon)
24 0.16
76 0.20
88 0.13
Run followed by 48 hours 'of steam regeneration of carbon (in situ).
-------
Batch #79 and Batch #104 had adsorptivities of 0.16 g EDC/g carbon and
0.165 g EDC/g carbon, respectively, with stream 221A at 6000 mg/1 EDC concen-
tration. Adsorptivity for Monochem was 0.22 g EDC/g carbon under the identi-
cal conditions. Thus, a significant improvement is observed over the two
previous activated batches.
The breakthrough profiles indicated a higher rate of adsorption and
higher service life for stream 221A than for stream 041A. Two runs were
made to determine the adsorption capacity of the Monochem activated carbon
on the demonstration unit at process effluent 221A. Lab studies had shown
that its EDC adsorption capacity was about one-half that of the best commer-
cially available activated carbon. Since there was less than 25 kg of the
Monochem activated carbon, only one column was packed. The breakthrough
occurred after 23.5 hrs. The Monochem activated carbon had an adsorption
capacity of approximately 0.15 kg of EDC/kg activated carbon (Table 39).
This capacity confirms earlier studies, since this value is around one-half
the capacity of WVG (0.28 kg of EDC/kg activated carbon).
To determine the effectiveness of steam regeneration on the Monochem
activated carbon, the column was regenerated for 48 hrs with 110° steam
(0.20-0.34 atm). In the second run, adsorption capacity was slightly
increased (0.20 kg EDC/kg activated carbon vs 0.16 kg EDC/kg AC for the
first run). These two runs supported earlier findings that the Monochem
activated carbon has about one-half of the adsorption capacity of commercial
activated carbon and can be steam regenerated just as readily.
Shirco Furnace for Field Activation and Regeneration—
For the field testing program, the soot carbon drying, activation, and
regeneration unit had to be large enough to generate sufficient material for
the ongoing demonstration unit tests. A recent development in this field
was the application of an infrared radiant heat source. Since the furnace
size could be regulated by the number of infrared lamps placed in the heating
zone, this technology appeared logical for adaptation to this activated soot
carbon program. The installation of such a unit offered several other
advantages.
1. The furnace could be heated to operating temperature within one
hour without damage to the insulation.
2. Air could be restricted so that only the amount required for
regeneration was used during the process.
3. There was no mechanical particle breakage within the furance.
4. Simple one-story modular design allowed for easy construction as
well as economical expandability.
Shirco, Inc., has applied this technology to the regeneration of acti-
vated spent carbon and has been able to approach operating costs of natural
gas. with an all-electric system utilizing smaller gas scrubber systems, less
air, and lower capital investments than other units presently on the market.
64
-------
The infrared lamps in the system (General Electric) were high-temperature
tungsten filament quartz lamps with the filament operating at approximately
2200°C. These lamps radiated 86 percent of input energy, leaving the remain-
ing 14 percent to be dissipated by convection to the cooling air. At rated
voltage, the lamps had an average life expectancy of 5000 hrs.
In the carbon regeneration system, the carbon is discharged into the
machine onto a conveyor belt. The high temperature belt conveyor carries
the carbon through a drying zone and then into an activation zone. In the
activation zone, a battery of infrared lamps mounted just above the belt
maintains the desired temperature. The belt discharges the carbon into a
quench tank at the end of the furnace. The lamps and end seals are cooled
by drawing outside air through the cooling air ducts. The temperature is
maintained automatically by electronic controllers.
DEVELOPMENT PROGRAM FOR IMPROVING ACTIVITY AND MECHANICAL STRENGTH
Improvements were needed for both the activity and the granulated
mechanical strength of the Monochem soot carbon. When measured against a
cross-section of commercial carbons, the Monochem soot carbon was shown to
have about 50 percent of the adsorption capacity. Further, its crush strength
was not very good and limited its regeneration applicability.
Even at a comparative adsorptive value of 50 percent, the activated
soot carbon offered some commercial potential. Since it had limited by-
product value (heating value over the water content), the only major manufac-
turing costs were the actual production and depreciation charges. Additional
capital equipment would be necessary to accommodate the lower adsorptive
capability of the soot carbon. Thus, for a specific commercial application,
an assessment would have to be made comparing the lower activated soot
carbon cost versus the additional capital equipment costs. However, if a
marked improvement could be made in the adsorption capacity and particle
strength of the soot carbon, then its competitive position would be greatly
improved. Toward this end, a program was initiated to improve the qualities
of the soot carbon.
Because of the unanticipated losses in producing the intitial quantity
of activated soot carbon for laboratory and field tests, an additional
amount had to be dried for granulation and activation. Early in 1974, 450
kg of soot carbon was sent to K-G for granulation. The carbon had been
dried in the Shirco furnace under the conditions listed in Table 41. All
granulation runs made with carbon dried to the 20 percent moisture level and
the 10 percent moisture level using a pitch-binder formulation were unsuccess-
ful.
Several other formulations were tried, and only a 50 percent water/asphalt
emulsion gave granules of sufficient hardness to withstand handling. Approxi-
mately 10 kg of these granules was returned for activation studies, with a
twofold objective: (1) to activate the granules despite the low hardness.
(2) to increase the physical hardness, using the activation conditions as a
heat treatment.
65
-------
TABLE 41. SHIRCO DRIED SOOT CARBON
Entering Soot 40-60 Percent Moisture
Exiting Soot 20-25 Percent Moisture
Drying Rate 5.4 - 6.8 kg/hr
Furnace Temperatures Lamp Section 1 & 2 593-649°C
Lamp Section 3 & 4 704-760°C
Blanket Gas Nitrogen
Total Soot Dried 453.6 kg
The experimental work by K-G Industries to develop a suitable activated
carbon from the soot by-product at Monochem covered a period of over one
year. In general, efforts were centered around adding binders to previously
dried samples of the soot. These binders of conventional character (as
pitch) did not produce a granulated material with suitable compaction strengths.
Therefore, novel solutions were sought to produce a satisfactory granule.
Physical and Chemical Characterization of Dried Soot
K-G personnel were unable to granulate soot dried on the Shirco furnace,
but were able to granulate soot dried in a steam tube furnace. Therefore,
experiments were conducted to determine the chemical and/or physical differ-
ences between these two soots. One sample each of the steam-tube-dried soot
(sample A) and the Shirco dried soot (sample B) was examined by scanning
electron microscopy and transmission electron microscopy to determine if any
physical differences existed. The differences in granulating behavior,
indicated sample A should be radically, different in size and degree of
aggregation from sample B. However, neither micrograph shows an appreciable
difference between the samples in either degree of agglomeration or the
ultimate or discrete particle size. Scanning electron micrographs of the
carbon black samples as they were dispersed in water and air dried on an
aluminum sample holder, and transmission electron micrographs of the samples
obtained by drying dilute water dispersions of the carbon black onto collodion
membranes are included in the Appendix.
LABORATORY EVALUATION OF BINDERS FOR SOOT CARBON GRANULATION AND ACTIVATION
In this phase of the project, organic and inorganic materials were
screened for their ability to bind the soot carbon into a hard, dense form.
The adsorbing activity of the processed carbon, as measured by the relative
iodine number, was studied.
Previous efforts to determine a suitable binder for Monochem carbon
soot concentrated on the use of pitch. The products from various pitch
binders were crumbly and had about 50 percent of the absorbing ability of
commercial carbons. Development of a suitable activated carbon from the
Monochem soot was divided into several research categories.
1. Moisture content of soot carbon and selection of binders.
66
-------
2. Activation of carbon samples.
3. Measurement of adsorption activity.
4. Further soot carbon development.
Moisture Content of Soot Carbon and Selection of Binders
Initially, work was done with the soot carbon at three levels of water
content: (a) soot carbon dried to a water content of 10% (powder dry), (b)
soot carbon dried to a water content of 57% (still had caking tendencies),
and (c) soot carbon, as received, with a water content of 79% (mud-like in
appearance). Each of these samples was evaluated by the addition of different
potential binders. The selection of binders is shown in Table 42.
TABLE 42. MATERIAL SELECTION FOR BINDER STUDIES
Test Conditions of Soot Carbon Additives as Binder
(a) Dried to 10% moisture content Plaster of Paris, Bakelite
resin, phenol formaldehyde
resins, arofene
(b) Dried to 57% moisture content Bakelite, Reichhold 29-353,
methocels, starch, bagasse,
polyvinyl alcohol, filter
cell, cement, calcium
carbonate, phosphoric acid,
calcium hydroxy phosphate,
dental investment, silicone-
latex, sodium aluminum oxide.
(c) As received, 79% moisture Methocels and bagasse
content
The material with 10 percent moisture content was mixed with varying
concentrations of the various binders. The mixture was then pressed in a
Buehler hydraulic press. During the heating process, the compression die
was surrounded by a heating mantle which maintained a temperature of 71-
82°C. The material with a moisture content of 51 percent was too wet to
use in the hydraulic press but could be extruded in a simple hand-operated^
unit. Extrusion was chosen as the method of handling the material with 79%
moisture content; however, the strands formed a puddle-like mass after being
extruded. To eliminate the puddling effect, the carbon was mixed with the
binding agent and then dried to about 66 percent moisture. This procedure
produced a stable extruded material.
Activation of Carbon Samples
All pressed or extruded samples were subjected to activation by heating
in the presence of nitrogen gas. The heating was conducted in a cylindrical
furnace manufactured by Aten, Ltd. The furnace had a maximum temperature of
1204°C and was accurate within +6.7°C. Power was supplied to each half
67
-------
section of the furnace unit by a 10 amp, 140 V rheostat purchased from the
W.H. Curtin Company. For all activation heating, the furnace was flushed
with dry nitrogen. The entire furnace was wrapped in a ceramic insulation
blanket, and additional layers of the blanket were used to protect the
counter top.
The activation times and temperatures were chosen on the basis of the
previous pitch binder study. The highest activity for the pitch binder was
produced by a pre-heating at about 454°C for 15 min, followed by an activation
heating at 982°F for 15 min. These same heating conditions were chosen for
the present work. However, steam was used to flush the furnace during the
heating of the pitch samples in the current study. The use of nitrogen as
the inert medium for laboratory evaluation appeared to be as effective as
steam, although the water content of the carbon could have been a factor in
the activation process.
Laboratory Evaluation of Binders for Granulation and Activation
The evaluation of different binders and their effect on the activity of
the Monochem soot carbon was divided into several phases.
1. Pressed pellet study.
2. Extrusion of semi-dry carbon.
3. Extrusion of wet carbon.
4. Further soot carbon development.
Pressed Pellet Study—
The experimental data from the hard pellet samples are summarized in
Table 43. For all samples the pressure applied to the press was about 544
atm. The percent of binder in the sample is the percent by weight of the
carbon in the sample. The water is tabulated as percent of total weight of
the sample. The description of the samples was based solely on a qualitative
judgment. Those samples described as hard required a strong pressure to
break into two pieces. Those samples listed as brittle were relatively easy
to break apart. The designation of homogeneous or nonhomogeneous refers to
the distribution of resin within the pellet.
Earlier studies had demonstrated that the 20 percent carbon (80 percent
moisture) soot cake coming from the waste soot concentrating process contains
an appreciable quantity of extractable organics. Since the drying temperature
in the Shirco furnace was approximately 538°C higher than in the steam tube
dryer, a greater quantity of organics was expected to be driven off during
drying in the Shirco furnace. To test this hypothesis, 1 g each of Shirco
dried soot, and 20 percent soot cake was extracted in a Soxhlet extraction
apparatus for 24 hrs with 100 ml of chloroform. The chloroform was then
distilled off and the residue dried on a steam bath. The results listed
below indicate that the Shirco furnace drives off an important binder ingredi-
ent. In some way, these extractable organics must aid the dry pitch in
penetrating the surface of the discrete particles.
68
-------
TABLE 43. A TABULATION OF THE DATA FOR THE PRESSED PELLET CARBON SAMPLES
Code
11
12
13
PI
P2
P3
P4
P5
P6
Cl
C2
C3
C4
Composition of Sample
Binder %H
Plaster of Paris-10%
mixed dry
Plaster of Paris-10%
mixed as slurry
Plaster of Paris-10%
Phenol Formaldehyde
Reichhold 29-353-15%
Phenol Formaldehyde
Reichhold 29-389-15%
Reichhold 29-389-10%
Reichhold 29-353-12%
Phenol Formaldehyde
Arofene-10%
Arofene-10%
Bakelite-10%
Bakelite-5%
Bakelite-2%
Bakelite-10%
2°
10
10
51
10
10
51
51
10
51
10
10
10
51
Description of
Prior to
Activation
Very brittle
Hard,
homogeneous
Sample squeezed
out of press
Hard
Poor distribution
of resin
Brittle,
very moist
Very soft
Nonhomgeneous
Very soft
Hard
Hard
Hard
Homogeneous,
not as hard
as dry carbon
Sample
After
Activation
Hard,
nonhomogeneous
Hard,
nonhomgeneous
Hard,
stratified
Brittle
Brittle
Hard
Brittle
Hard
Brittle
Hard
Hard
Brittle
Brittle
69
-------
Percent
Soot Sample Extractables
20 Percent Cake 10.2
Steam-Tube-Dried 8.5
Shirco-Dried 4.98
Since the steam tube dryer was no longer available to test these findings,
4.5 kg of soot was dried under the soot-burning furnaces to less than 30
percent moisture. The temperature at this location was 66-71°C and would not
drive off an appreciable quantity of organics. This soot was shipped to K-G
for granulating.
The pressed pellets all were firmly compacted. However, it was believed
that this compacting would produce a nonporous carbon structure which would
be poorly adsorbent. Therefore, the pellet press method of compacting was
abandoned in favor of an extrusion method. The pellet samples were not
examined for their ability to adsorb iodine.
Extrusion of Semi-Dry Carbon—
The moisture content of the semi-dry carbon was initially 51 percent.
It was found that this carbon was too dry to extrude well. Varying amounts
of water were added to the sample to produce a mix which would give smooth,
uniform strands and leave only a small amount of material in the grinder.
The results from the extruded carbon samples are listed in Table 44. The
binders are tabulated as percent by weight of carbon in the sample.
The iodine number was measured for those samples which were judged to
be hard or moderately hard. Considering only the iodine number, the sample
containing 4 percent methocel HB (hydroxy butyl methyl cellulose) would be
the best product. However, the carbon samples containing 8 percent methocel
HB showed a greater degree of hardness. The water content of the sample
was important when using methocel HB as a binder. For the 8 percent methocel
level, the integrity of the extruded strand changes from a compact, smooth
strand to a fluted strand as the water content varies.
Two other grades of methocel tested (methocel HG and methocel HC) did
not bind the carbon as well as methocel HB. This was probably due to the
difference in viscosity among the various grades. Methocel HB has a viscosity
of 12,000 cps, whereas methocel HG and methocel HC both have viscosities of
4000 cps.
The samples containing bagasse showed relatively good activity with
moderate hardness. It appears that both activity and hardness increase as
the concentration of bagasse decreases to about 5 percent. This result is
encouraging with regard to the use of a substance such as bagasse as a
binder.
Extrusion of Wet Carbon—
The moisture content of the carbon was about 79 percent. Trials with
two samples indicated that the carbon was too wet to extrude into discrete
strands. The remaining samples were prepared by mixing the appropriate
binder with the wet carbon and then drying the mix to about 67 percent
70
-------
TABLE 44. A TABULATION OF THE DATA FOR THE EXTRUDED CARBON SAMPLES
PREPARED FROM SEMI-DRY CARBON
Composition of Sample
Code
C5
C5
C7
C8
14
15
16
17
18
19
110
Ml
M2
M3
M5(B1)
M6
M7
M8
M9
M10
Mil
B2
B3
B4
B5
01
02
03
04
05
06
07
Binder
Bakelite-107
Phenol Formaldehyde
Reichhold 29-353
Reichhold 29-353-27%
Cement-8%
CaC03-340%
CaC03-81%
Phosphoric Acid-81%
Calcium hydroxy phosphate-81%
Dental investment-81%
Silcone-latex binder-
40 ml of liquid
NaAl 02-21%
Methocel HB-8%
Phenol Formaldehyde-8%
Methocel HB-8%
Phenol Formaldehyde
Reichhold 20-353-8%
Bagasse-16.3%
Polyvinyl Alcohol-8%
Filter Cel-
Diatomaceous Earth-8%
Starch-8%+0.05M ZnCl2
Starch-8%+0.10M ZnCl2
Methocel HG-8%
Methocel HC-8%
Bagasse-38%
Bagasse-8%
Bagasse-4.8%
Bagasse-8%+0.05M ZnCl2
Methocel HB-13.6%
Methoeel HB-8%
Methocel HB-8.8%
Methocel HB-4%
Methocel HB-5.7%
Methocel HB-8%
Methocel HB-8%
% H?0
51
Unknown
51
68.9
35.3
57
42.6
57
53
31
49
67
68.9
69
67
69
69
69
69
69
69
73.6
69
69
69
69
70
69
69
69
67
68
Description of Sample
After Activation
Powder
Soft
Brittle
Moderately hard
Powder
Soft, easily broken
Soft, easily broken
Moderately hard
Soft
Hard
Powder
Soft
Hard
Brittle
Soft
Moderately hard
Brittle
Moderately hard
Soft
Soft
Hard
Very soft
Moderately hard
Moderately hard
Moderately hard
Moderately hard
Moderately hard
Moderately hard
Moderately hard
Moderately hard
Relatively hard
Hard
Iodine Number
(Commercial=1084)
—
—
—
384.05
~
—
—
—
—
353.85
—
—
570.93
421.57
377.10
__
468.26
—
—
492.52
381.98
465.11
531.16
446.89
—
—
—
635.17
500.03
—
389.00
Relative Iodine
Number
(Commercial=l. 000)
—
—
—
0.354
—
—
—
—
—
0.326
—
—
0.527
__
0.389
0.348
—
0.432
—
—
0.455
0.352
0.429
0.490
0.412
—
—
—
0.586
0.488
—
0.379
71
-------
moisture content. The results from this phase of the work are given in
Table 45.
Samples Al and A2 were each collected as an amorphous mass which was
dried and activated. The hardness was determined after activation. Note
that the activity of sample A2 is greater than that of Al, indicating that
bagasse can increase the adsorbing activity of the carbon. Sample A5,
containing 4.7 percent methocel HB and 4.7 percent bagasse, showed the
highest activity of all the samples tested. These samples again indicate
that the adsorbing ability of the carbon decreases as the weight percent of
binder increases. The water content of the sample is very important to the
extrudability. The best mix appears to have a moisture content of about 68
percent.
Further Soot Carbon Development
Work to this point consisted primarily of using a wide variety of
organic and inorganic compounds as binders for the soot carbon. The moisture
content of the carbon and the method of compacting the carbon were varied.
All samples prepared during the previous phases of the project were activated
in a nitrogen atmosphere. The results of this effort indicated that wet
carbon of 73 to 79 percent moisture content can be converted to a hard form
by extruding the carbon mixed with 8 percent methocel HB. The activity of
this carbon after activation was about 60 percent that of commercial carbon.
The price of methocel HB as of January 12, 1976 was quoted as $1.80/kg
based on a minimum order of 30,000 kg. At this price, the cost of the
binder for one kg of activated carbon containing 8 percent methocel HB
would be 14.4
-------
TABLE 45. A TABULATION OF THE DATA FOR THE SAMPLES PREPARED FROM WET CARBON
Code
Al
A2
A3
A4
A5
Relative Iodine
Description of Sample Iodine Number Number
Binder %H_0 After Activation (Commercial=1024) (Commercial=l . 000)
Methocel HB-9.5% 79
Methocel HB-9.5% 79
Methocel HB-4.7% 66.8
Methocel HB-4.7% 66.8
Methocel HB-4.7% 68.2
Bagasse-4.7%
Very hard 604.66
Did not extrude
Moderately hard 659.54
Did not extrude
Moderately hard 505.27
Moderately hard 589.86
Moderately hard 626.32
0.590
0.644
0.493
0.576
0.611
-------
order of 30,000 kg. Considering only the cost of the binder, this is a
cost of 10.4£/kg of activated carbon.
As shown in Table 46, the activity of the activated carbon as measured
by the iodine number ranged from 33.5 to 33.7 percent that of commercial
carbon. Experiments indicate that these low activities may be significantly
improved.
TABLE 46. A DESCRIPTION OF SAMPLES PREPARED FROM DIFFERENT TYPES OF METHOCEL
Description Relative
Type of Percent after Iodine Number Iodine Number
Code Methocel* Binder Activation (Commercial=1012) (Commercial=1.00)
010
Oil
012
J12
K15
J20
8
8
8
Brittle
Relatively Hard
Hard
381
370
339
0.377
0.366
0.335
*
The type of Methocel denotes the viscosity of a 2 percent aqueous solution
at 20°C. Type J12 would have a viscosity of 12,000 cps at 20°C whereas K15
would have a viscosity of 15,000 cps.
There has been considerable interest in the potential of bagasse as a
binding material. Samples containing 4.8, 8, 38, and over 50 percent
bagasse were investigated. A sample containing 70 percent bagasse was
prepared from soot carbon of 79 percent water content. The extruded carbon
strands were dried at 82°C and then heated in a nitrogen atmosphere at
427°C for 15 min and finally at 982°C. The activated sample was brittle
:Vhen compressed between the fingers, and its activity was less than 30
percent that of commercial activated carbon. These results indicate that
if bagasse is to be used as a binder it must contribute less than 50 percent
by weight to the sample.
Effect of Inorganic Additives—
The effects of several inorganic salts on the activity-of the carbon
were investigated. The Encyclopedia of Chemical Technology states that
numerous inorganic compounds increase the activity of activated carbon.
Among these compounds are the alkaline earth carbonates, chlorides, sulfates,
and phosphates. One alkaline earth element, calcium, is a constituent of
coconut shells and bones, from which activated carbon has traditionally
been made. The effects of calcium carbonate, sulfate, and phosphate on the
activity of the spot carbon containing methocel HB as a binder were examined.
Samples were prepared from soot carbon containing 73 percent water and
8 percent methocel HB as a binder. The extruded strands were activated
under nitrogen in two heating stages. The first heating was a 538°C for 15
min and the second was at 1038°C for 15 min. The activity of the resulting
carbon was measured via the iodine number. The results from this investi-
gation are shown in Table 47. (All the samples were judged to be hard.)
74
-------
As shown in Table 47, only a small amount of inorganic material was
added to the carbon. The amount of added material was deliberately minimized
as the results from previous samples (Table 46) indicated that a high
concentration of carbonate or phosphate produces a soft product. The
presence of either CaCO-j for CaSO, appeared to decrease the measured activity
of the carbon. The activity of the carbon containing Ca (P0,)9 did not
seem to be very different from that measured for samples containing only
methocel HB.
TABLE 47. EFFECT OF INORGANIC ADDITIVES ON THE ACTIVITY OF SAMPLES
METHOCEL HB BINDER
Description Relative
Inorganic Percent after Iodine Number Iodine Number
Code Additive Additive Activation (Commercial=1012) (Commercial=1.00)
M12 CaC03
Ml 3 CaSO.
M14 Ca3(P04)2
0.5
0.5
0.5
Hard
Hard
Hard
<296
<289
370
<0.292
0.286
0.365
The results from this phase of the project indicate that neither CaCO,
nor CaSO, is a likely candidate for increasing the native activity of the
soot carbon. In higher concentrations, Ca_(PO,)7may be useful in increasing
adsorption by the carbon. However, this material is not readily available
commercially.
COMPOUNDING, PELLETIZING, AND ACTIVATING MONOCHEM SOOT CARBON
The first step in the optimization of the activation process was to
determine the reproducibility of any given activation process. For this
procedure, sample K7, containing 1.8 percent Kelzan and 5.5 percent methocel
HB, was divided into two portions. Each portion was subjected to essentially
the same activation conditions. The measured iodine number of the first
portion was 483, which corresponded to a relative iodine number of 47.1
percent. The second portion was analyzed to have an iodine number of 500 and
a relative iodine number of 48.9 percent. The difference in iodine number
between these two portions is 3.4 percent. This error is about that which
could be expected from the method of analyzing the carbon. Results of this
experiment indicate that any variation in the activity of a uniform batch
of carbon must result from the conditions of activation.
The next investigation was to determine if the carbon could be activated
in one heating stage rather than two. For this phase of the project, the
temperature conditions and the time of heating were varied. The results
from this study, listed in Table 48, clearly indicate that a high temperature
heating stage is necessary to produce higher activities. A comparison of
samples D7 and Kll activated at low temperatures shows that a slight increase
in both time and temperature does not produce a corresponding increase in
the activity of the carbon.
75
-------
TABLE 48. A COMPARISON OF THE ACTIVITY FROM LOW TEMPERATURE HEATING
WITH THAT FROM TWO-STAGE HEATING
Sample
D7
Binder (s)
Methocel HB-8%
Activation
Procedure
Two-Stage Heating:
Relative
Iodine Number Iodine Number
(Commercial=1024) (Commercial=l. 00)
389
0.379
D7 Methocel HB-8%
Kll Kelzan-2%
Bagasse-5%
Kll Kelzan-2%
Bagasse-5%
15 min at 454°C
15 min at 982°C
Low Temperature Heating 200
30 min at 677-760°C
Two-Stage Heating:
15 min at 441°C
15 min at 954°C
398
Low Temperature Heating: 200
15 min at 454°C
0.195
0.388
0.195
The next logical step was to determine if the low temperature heating
stage could be eliminated. All attempts to heat the carbon at temperatures
of 788-954°C without prior low temperature heating resulted in combustion of
the carbon. For one such attempt, the carbon was stored under nitrogen
before being placed in the furnace at 954°C. It was hoped that the nitrogen
would diffuse into the carbon and displace any free oxygen present. This
procedure, however, did not prevent combustion. The two experiments indicated
that a two-stage heating procedure was necessary to achieve optimum activity.
The low temperature heating was needed to drive off the volatile organic
compounds, and the high temperature heating was needed to produce active
sites for adsorption.
Next, a procedure was devised to determine what combination of low and
high temperatures would produce the maximum activity. The procedure consisted
of activating samples at various combinations of temperatures. The low
temperature range extended from 315°C to 538°C, and the high temperature
range from 871°C to 1038°C. The temperature combinations are illustrated
below. Results of the tests are listed in Table 49.
Low
315°C
High
871eC
927°C
982PC
1038°C
Low
370°C
Low
538°C
871°C
927°C
982°C
1038°C
Low
427°C
871°C
927°C
982°C
1038°C
76
-------
TABLE 49. DETERMINATION OF THE BEST COMBINATION OF LOW AND HIGH
TEMPERATURE HEATING CONDITIONS
Activation Temperatures Preheat Temperature
(°C) 315 370 427 538
Relative Iodine Numbers
871
927
982
1038
0.245
0.259
0.283
0.312
0.265
0.280
0.325
0.350
0.272
0.297
0.321
0.358
0.370
The carbon used for this procedure was soot carbon of 73 precent moisture
content; the binder was 8 percent methocel HB. About 400 g of carbon was
used to prepare one large batch of material. All samples were taken from
this batch of material to assure uniformity of composition among the samples.
All samples were activated in an inert nitrogen atmosphere. The activation
process consisted of heating a sample at one of the low temperatures for 15
min and then at one of the high temperatures for another 15 min. The activa-
tion times were held constant for all the samples. The activity of each of
the 13 samples was measured in terms of the relative iodine number. The
relative iodine number was calculated as:
« •, Iodine Number of Test Sample
Relative Iodine Number = Iodine Number of Commercial Carbon
The iodine number of the commercial carbon was determined to be 1012.
One sample was pre-activated for 15 min at 260°-343°C and then activated
at 954-982°C for 15 min. Another sample was pre-activated at 260-427°C for
15 min and then activated for 15 min at 954-982°C. The samples were stored
under nitrogen between the pre-activation and the activation heatings. The
samples were quenched in water immediately after the final heating.
The samples activated in steam were compared for relative hardness with
those activated in nitrogen. The steam activated samples appeared to be more
brittle than those activated under nitrogen. The activity of the steam-
activated samples was measured by the iodine number technique._ The activity
of both steam activated samples was 58 percent that of commercial carbon. A
control sample activated in nitrogen atmosphere exhibited only 32 percent of
the activity of commercial carbon.
These results indicated that some type of mild oxidizing agent must be
present during the activation process to achieve reasonable adsorb^activity-
The results also helped explain some of the variation seen in the activity of
previous samples.
Determination of Compression Strength-- ,.,.„-j» nf the material to
-
77
-------
with 8 percent methocel HB as a binder. Compression was measured by a South-
wark-Emery compression tester. The applied pressure was recorded by a Tate-
Emery load indicator. The compression strength was determined for both the
lengthwise and crosswise directions. For either direction, the pressure
required to break the strand was essentially the same. The maximum applied
pressure before crushing the sample was 5 pounds. The diameter in the cross-
wise direction was 3.175 mm. Using these data, the compression strength of
the sample was calculated to be 28.61 Kg/cm (27.7 atm).
Determination of Molasses Number—
The adsorbing ability of activated carbon can be measured in a variety
of ways. The iodine number gives an indication of the ability of the carbon
to adsorb small molecules. An indication of the ability of an activated
carbon to adsorb large, color-producing molecules is given by the molasses
number.
A method of determining the molasses number of activated carbon has been
described by the Environmental Protection Agency in the Process Design Manual
for Carbon Adsorption . The procedure followed in this study was essentially
the same as that described by the EPA.
The standard carbon used for this analysis was Calgon Filtrasorb-400.
According to a representative of Calgon, the molasses number of this carbon
is about 250. The absorbance of this carbon, corrected for the absorbance
of the 15 percent sucrose solution blank, was 0.516 absorbance units.
The molasses number was determined for two samples of activated carbon
prepared with 8 percent methocel HB as a binder. One sample had been
activated in nitrogen and the other sample had been activated in steam.
The absorbance of the sample activated under nitrogen, corrected for the
sucrose blank, was 0.564, which translates into a molasses number of 228.
Although the molasses number for this carbon compares well with that of the
commercial carbon, a measurement of the absorbance of the stock molasses
solution diluted 1 to 20 revealed that the test carbon had not removed any
color producing compounds. The absorbance of the diluted molasses solution
was 0.565.
The absorbance of the carbon that had been activated in steam, corrected
for the sucrose blank, was 0.469. Using this absorbance, the molasses
number was calculated to be 275. This indicates that the steam-activated
soot carbon is slightly better than commercial carbon in removing color-
producing compounds from molasses.
Large Batch Production—
The furnace assembly previously described had been adequate for the
activation of small amounts of material. The production of large amounts
of activated carbon required the use of a much larger furnace. On-site in
New Orleans was a 15K watt, HG-12 furnace built by Westinghouse which was
capable of processing a large amount of material in a short period of time.
Repairs and modifications were made to the furnace to make it operational.
The furnace had been completely rewired and the standard cell for the
pyrometer replaced. The ceiling above the oven was insulated with asbestos
78
-------
and a flue constructed directly above the door of the oven. Tubing was
connected to the furnace to provide for either a nitrogen or steam inlet.
A stainless steel "basket" 533.4 mm long, 254 mm wide, and 25.4 mm high was
constructed to contain the sample during the activation process.
OPTIMIZATION OF COMPOUNDING AND ACTIVATING MONOCHEM SOOT CARBON
Nitrogen versus Steam Activation
Most of the Monochem soot carbon samples prepared to this point had
been activated in a nitrogen atmosphere. Based on the iodine number, one
sample which was activated in steam was found to have a much higher adsorption
activity than a similar sample activated in nitrogen. Several samples were
activated both in nitrogen and in steam to determine if the increase in
activity was reproducible. Table 50 gives a comparison of the activity of
carbon samples prepared from methocel HB and activated in either a steam or
a nitrogen atmosphere. The results clearly indicate that activation in
steam increases the measured adsorption activity by almost 100 percent over
that measured for samples activated in nitrogen.
To further investigate the effect of activation gases on the carbon
adsorption capacity, a fresh batch of carbon was prepared and activated in
both nitrogen and steam. Two control samples were activated, one in nitrogen
and one in steam. Table 51 lists the results derived from this experiment.
The results indicate that steam increases the activity when used during the
high temperature activation, but does not seem to have much effect during
the pre-activation heating. One disturbing result of this study was the
sharp decrease in the iodine number for the steam activated sample compared
with the steam-activated samples described in Table 50. The only difference
between the samples described in Table 50 and those in Table 51 was the lot
of methocel HB which was used in carbon preparation.
The next step was to compare the adsorption activity for carbon samples
prepared from several different lots of methocel HB. Dow Chemical supplied
a new sample of methocel HB. Three different batches of carbon were prepared:
one from an old batch of methocel (Lot No. 121693), one from the methocel
used for the samples described in Table 52 (Lot No. MM030458B), and one
from the new batch of methocel (Lot No. MM030463B).
As shown, the activity of the carbon as measured by the iodine number
varied with the particular lot of methocel HB used in sample preparation.
It was learned from Dow Chemical that Lot No. MM030458B consisted of methocel
HB which had less methyl substitution than the other two lots of methocel.
Furthermore, this lot of methocel did not meet product specifications in
terms of nonaqueous solution viscosity. However, Dow stated that Lot No.
MM030463B was representative of the typical quality of methocel which the
firm manufactures.
Finally, the carbon activity was measured by the molasses number
procedure. The results from these measurements are given in Table 53.
Carbon activated by steam has a greater ability to decolor molasses than does
79
-------
TABLE 50. COMPARISON OF ACTIVITY FOR METHOCEL HB SAMPLES
ACTIVATED IN NITROGEN AND IN STEAM
Activation
Gas
Nitrogen
Nitrogen
Steam
Steam
Steam
Description of Iodine
Sample After Number
Activation (Commercial=1012)
Hard
Hard
Brittle
Brittle
Moderately Hard
362
322
588
584
471
Relative
Iodine Number
(Commercial=l . 00)
0.358
0.318
0.581
0.577
0.465
Activation of wet strands produced a soft product with a low iodine number:
<281, relative iodine number: <0.278.
TABLE 51. ACTIVATION BY TWO DIFFERENT GASES:
NITROGEN AND STEAM
Pre-Activation Activation Relative Iodine Number
Gas Gas (Commercial=1.00)
Nitrogen
Steam
Nitrogen
Steam
Nitrogen
Steam
Steam
Nitrogen
0.257
0.322
0.388
0.218
ACTIVATION CONDITIONS:
Pre-Activation
Time (Minutes)
Activation
Pre-Activation Time
Temperature C°C) (Minutes)
Activation
Temperature
(°C)
15
427-454
15
954-999
80
-------
TABLE 52. ACTIVITY MEASURED FOR DIFFERENT LOTS OF METHOCEL HB
Lot Number
121693
121693
MM030458B
MM030458B
MM030463B
MM030464B
Activation
Gas
Nitrogen
Steam
Nitrogen
Steam
Nitrogen
Steam
Iodine
Number
Commercial=1012
322
588
<260
<326
357
607
Relative
Iodine Number
Commercial=l. 00
0.318
0.581
<0.257
<0.322
0.353
0.624
TABLE 53. ACTIVITY DETERMINATION BY THE MOLASSES NUMBER
Sample
Calgon
019
020
020
013
Activating
Gas
Nitrogen
Nitrogen
Steam
Steam
Molasses
Absorbance Number
0.430
0.562
0.547
0.512
0.470
250
191
196
209
228
Relative
Molasses
Number
1.00
0.764
0.784
0.836
0.912
Relative
Iodine
Number
1.00
0.353
0.318
0.465
0.577
81
-------
carbon activated by nitrogen. For this series of samples, the activity of
the steam activated carbon was almost 100 percent of that measured for the
commercial carbon.
Benzene, 1,2-dichloroethane (EDC), phenol, and chloral hydrate were
studied by means of Freundlich adsorption isotherms. In addition to these
individual compounds, stream 231-A, containing toluene diisocyanate, was
investigated. The carbon used for the isotherms was prepared by extruding
the Monochem carbon with methocel HB.* One large batch of the carbon was
prepared and activated in a nitrogen atmosphere. The relative iodine
number for this carbon was 28 percent that of commercial carbon.
The results from the Freundlich isotherms generated from nitrogen
activated carbon are given in Table 54. The individual isotherms for each
compound are shown in the Appendix. The results for each of the compounds
show that the Monochem carbon is better than commercial carbon in the
removal of 1,2-dichloroethane, benzene, and acidified phenol. The activated
Monochem carbon is not as good as commercial carbon when applied to alkaline
phenol, chloral hydrate, or stream 231-A.
The results indicate that, of the compounds studied, only phenol in
acid solution could be completely removed from solution by adsorption.
When the solution was made basic, a phenol residual of 32 mg/1 resulted.
The residual in all cases was estimated by extrapolating the adsorption
isotherm curve to the x-axis. It is possible that the actual curve would
intersect the x-axis at higher residual concentration than those estimated
by a straight line extrapolation. If this were so the minimum adsorption
limit would be greater than zero.
TABLE 54. RESULTS FROM FREUNDLICH ADSORPTION ISOTHERMS:
NITROGEN GAS USED IN ACTIVATION OF CARBON
Compound
Ultimate Carbon Loading Factor mg TOG
gm C
pH C (mg/1, TOG) or Residual Compound
Monochem
1,2-dichloroethane 4.80 159
Benzene 6.55
Phenol 5.93
Phenol 11.7
Chloral hydrate 6.45 97
Stream 231A-TDI 11 630
Commercial (Calgon)
21 mg/1 residual 27 mg/1 residual
14.5 mg/1 residual
13.18
4.29
77.4 10 mg/1 residual
810 15.:
gm C
768 32 mg/1 residual
gm C
TOC
35 mg/1 residual
600 mg/1 residual
gm C
6.4 mg/1 residual
400 mg/1 residual
*It was later discovered that the batch of methocel HB which was used
did not conform to specifications. This accounts for the low iodine number
of the carbon.
82
-------
Regeneration Studies
In practice, most of the carbon used for wastewater treatment has been
regenerated one or more times. The regeneration of activated carbon is
usually performed by heating in an oxidizing atmosphere to volatilize adsorbed
organic compounds and to create additional micropores. The process results
in a change in the adsorption characteristics of the activated carbon.
These effects of regeneration on the adsorption capacity of the activated
carbon were investigated.
The regeneration study was carried out by collecting the Monochem
carbon which was used for the alkaline phenol adsorption isotherm. This
carbon was washed with distilled water and dried at 100°C. The carbon was
then regenerated by heating at 1000°C for 15 min in a nitrogen atmosphere.
Fumes were evolved during the heating process. The commercial carbon
(Calgon), which was also used for the alkaline phenol isotherm, was
regenerated in the same manner.
The results from the reactivation study are listed in Table 55 and
illustrated in Figure 9. The data indicate that these reactivation conditions
are not sufficient for complete regeneration of either the Monochem or the
commercial carbon. Higher temperatures and an oxidizing atmosphere apparently
are necessary for complete regeneration.
The final adsorption isotherms were performed with Monochem carbon
which had been activated in steam. Benzene and phenol were studied in this
set of experiments. The activated carbon was prepared by preheating in
steam at 427°C for 15 min followed by a heating at 982°C for 15 min. The
iodine number for this carbon was 41 percent that of commercial activity.
The results from the Freundlich isotherms generated from steam activated
carbon are shown in Table 56. The individual isotherms are illustrated in
the Appendix. A comparison with the nitrogen activated carbon shows that
the steam activated carbon is more efficient in removal of acidic phenol
but less effective in removal of benzene.
TABLE 55. RESULTS FOR REACTIVATION OF CARBON AS MEASURED
BY PHENOL ADSORPTION
Virgin Carbon
Regenerated Carbon
PH
11.7
11.7
TOC
(ppm)
768
902
Ultimate Carboi
or Residual
Monochem
32 mg/1 residual
— — ^JQC
i Loading _• •-
Compound
Commercial
, ™ mg TOC
--3 gc
2.2 mg/1 residual
83
-------
o
bO
g 100
60
6
10
d
Q
o Commercial carbon
O Monochen
^ 10 100
TOG Concentration (mg/1)
Figure 9. Adsorption isotherm for phenol
on regenerated carbon.
Pellet chamber
Feed hopper and
feed mechanism
Figure 10. Laboratory pellet mill.
84
-------
TABLE 56. RESULTS FROM FREUNDLICH ADSORPTION ISOTHERMS:
STEAM USED IN ACTIVATION OF CARBON
Compound
Benzene
Phenol
PH
5.5
8.5
Ultimate Carbon Loading Facto
T°C or Residual Compound
(ppm)
250
1194
Monochem
18.5 ppm residual 14.
04 mg TOG
"4 g C 13'
mg TOC
1 gc
Commercial
5 ppm residual
0 mg TOC
10 lc
PILOT WORK ON ACTIVATED CARBON DENSITY, ACTIVATION AND REGENERATION
Up until this point in the program, the laboratory samples of carbon
had been prepared from bench equipment. It was decided to perform develop-
ment work on pilot-scale equipment.
1. Pelletizing for increased density.
2. Small furnace activation procedure.
3. Large furnace.
Pelletizing for Increased Carbon Density—
Extruded samples prepared during these studies were extruded through a
household-type meat grinder. This "extruder" had a limited ability to
compact the carbon. As a result, the density of the carbon was very low
compared with commercial activated carbons. When activated in steam, the
hand-extruded carbon became very soft and crumbly.
In an attempt to increase the hardness and the density of the carbon,
45 kg of carbon mixed with methocel HB was shipped to the California Pellet
Mill Company for extrusion. An illustration of the pellet mill is shown in
Figure 10. The pelletizing operation is described below.
The carbon shipped for extrusion initially contained 64 percent water.
When some of this material was run through the pellet mill, it was found to
be too moist to give a good compaction. When dried to a moisture content
of 53 percent, the carbon mixture extruded well and gave good, compact
pellets. The extruded pellets were shipped back to GSRI where they were
thoroughly dried.
The apparent density of each of the pelletized samples was measured by
filling a weighed 100 ml graduated cylinder with the carbon in a specified
time period . The pelletized carbons were activated by heating in steam.
The apparent densities of the activated carbons were determined in the same
manner used for the unactivated carbon. The apparent densities of the acti-
vated and unactivated pelletized carbons were compared with the apparent
densities of commercial activated carbons and with carbon extruded by hand
through the meat grinder. The results of this comparison are shown in
Table 57.
85
-------
TABLE 57. COMPARISON OF THE APPARENT DENSITY OF VARIOUS CARBONS
oo
o\
3
Carbon Moisture Content Die Size Apparent Density (kg/m )
Westvaco (Nuchor)
Calgon
Monochem-Nitrogen Activated
Pelletized Monochem-Unactivated
Pelletized Monochem-Unactivated
Pelletized Monochem-Steam Activated
Pelletized Monochem-Unactivated
Pelletized Monochem-Unactivated
Pelletized Monochem-Steam Activated
73%
64%
64%
64%
57%
53%
53%
hand-
extruded
25.4 mm
38.1 mm
38.1 mm
38.1 mm
38.1 mm
38.1 mm
682.92
452.76
297.53
302.74
354.48
350.11
381.53
396.98
389.76
-------
The pelletizing process produces a marked improvement in the apparent
density of the carbon over that achieved by hand-extrusion. The results
also indicate that the apparent density of the carbon increases as the moisture
content decreases. At a 53 percent moisture content, the apparent density of
the carbon approaches that exhibited by the Calgon activated carbon. It
should be noted that the high apparent density measured for the Westvaco
activated carbon was due to an excessive amount of fines in the carbon sample.
The Calgon carbon and the pelletized Monochem carbons contained almost no
fines. The results show that steam activation decreases the apparent density
only slightly.
The increase in carbon density achieved by use of the California Pellet
Mill was encouraging. With the pellet mill, it was possible to: (1) vary
both moisture content and binder concentration and (2) process large amounts
of carbon and thus achieve uniformity in the final activated carbon.
Optimization of the Steam Activation Procedure Using Small Furnace
The optimzation of the activation procedure using nitrogen as an inert
atmosphere had fairly well established its limitations. The results from
this optimization indicated that a high preactivation temperature (482°C) and
high activation temperature (1038°C) were required to achieve maximum adsorp-
tion capacity. The preactivation and activation times were held constant at
15 min each for the ^entire optimization procedure. The flow rate of nitrogen
was not varied.
The use of steam as an activation gas necessitated the repeat of the
optimization procedure. Activation by steam required the optimization of the
following parameters: activation time, activation temperature, and flow rate
of steam. The preactivation heating was held constant at 427°C for 15 min
throughout the optimization procedure. It was necessary to carry out two
separate optimizations—one for the small cylinder furnace, and one for the
large Westinghouse furnace—because the large furnace was not thermostatically
controlled above 938°C. The overall procedure consisted of four separate
phases:
1. Activation of the carbon under varying conditions.
2. Determination of the percent weight loss due to activation for
selected samples.
3. Measurement of the static compression strength of the activated
pellets.
4. Measurement of the adsorption capacity by the iodine number
method.
All samples were taken from the same batch of carbon which contained 53
percent water and 7.4 percent methocel HB. The carbon-binder mix had been
extruded through the California Pellet Mill into uniform pellets about 6 mm
long and 3 mm in diameter. Steam was introduced by pumping water through a
stainless steel tube directly into the furnace. Since the temperature of the
87
-------
furnace was well above the boiling point of water, steam was produced as the
water exited the tube. The amount of steam produced was controlled by regulat-
ing the flow rate of water through the Masterflex pump. Two different flow
rates, 3.4 and 4.6 ml of HLO/min, were chosen for study. Above a flow rate
of 4.6 ml/min, it was difficult to control the furnace-temperature. The
volume of the small furnace was measured to be 3800 cm . At the lower flow
rate, the "water" concentration within the furnace was 5.95 x 10 , ml/cm -min.
At the higher flow rate, the concentration was 12.10 x 10 ml/cm -min.
Samples were activated under the conditions listed in Table 58. The
entire set of samples was preactivated at 427°C for 15 min. The percent of
weight loss was determined for samples 7, 8, 11, 12, 19, 20, and 23-27.
During the heating of sample 28, the steam flow stopped and the carbon burned.
The percent of weight loss was determined by weighing the carbon prior to and
subsequent to activation. After the preactivation heating and after the
activation heating, the samples were stored in a desiccator under dry nitrogen.
This prevented a weight change due to uptake of moisture. The results from
this study are given in Table 59. With the exception of samples 23 and 24,
the results indicated that higher activation temperatures and longer activation
times produce a higher weight loss of carbon. No immediate explanation can
be given for the discrepancy in the results from samples 23 and 24.
Following the activation procedure, the static compressibility of the
samples was measured in a Plexiglas sample holder to quantitate the hardness
of the carbon after activation. The pellet was held in place by carefully
lowering the Plexiglas cover onto the pellet. A 19.1 mm recession and a
smaller indentation in the cover allowed the cover to slide easily onto the
base. Calibrated weights were applied directly to the cover to compress the
pellet. Weights were added until the pellet crushed.
TABLE 59. RESULTS FROM THE ANALYSIS OF CARBON WEIGHT LOSS DURING ACTIVATION
Sample
7
8
11
12
25
26
19
20
23
24
27
Weight loss
Rate of
Water Flow
(ml /min)
3.4
3.4
3.4
3.4
3.4
3.4
4.6
4.6
4.6
4.6
4.6
due 3.2
Activation
Time (min)
15
15
20
20
30
30
15
15
20
20
30
15
Activation
Temp (°C)
982
1038
982
1038
982
1038
982
1038
982
1038
982
427
Percent of
Weight Lost
42
52
51
86
64
85
62
80
43
65
88
13
to pre-activation
88
-------
TABLE 58. ACTIVATION CONDITIONS FOR OPTIMIZATION OF SMALL FURNACE
OD
Activation Time:
Sample
1
2
3
4
Lower
10 min.
Activation
Temperature
871
927
982
1038
Pump Speed: 3.4 ml/min
Sample
5
6
7
8
Higher Pump
Activation Time:
Sample
13
14
15
16
10 min.
Activation
Temperature
871
927
982
1038
Sample
17
18
19
20
15 min.
Activation
Temperature
871
927
982
1038
of water flow
20 min.
Activation
Temperature
Sample (°C) Sample
9 871
10 927
11 982 25
12 1038 26
30 min.
Activation
Temperature
982
1038
Speed: 4.6 ml/min of water flow
15 min.
Activation
Temperature
871
927
982
1038
20 min.
Activation
Temperature
Sample (°C) Sample
21 871
22 927
23 982 27
24 1038 28
30 min.
Activation
Temperature
982
1038
-------
Table 60 lists the results accumulated from the compression tests. Most
tests were performed in duplicate to account for sample variation. The area
was calculated from the cylindrical surface of the pellet.
The results point out several features of the pellets and the activation
process. First, there is a wide variance in the crush pressure for several
of the samples. This variance does not correlate well with the variance in
the external dimensions of the pellets. It is believed that the observed
variation is the result of differences in the distribution of pellet activa-
tion. Second, the average crush pressure does not seem to be a function of
the activation conditions. Only at the longest activation times and highest
activation temperatures was there a significant decrease in the average crush
pressure. The majority of the samples exhibited a very good crush pressure.
These results indicate that a satisfactory sample hardness can be achieved
under typical activation conditions.
The adsorption capacity of the activated carbon was determined by the
iodine number method. Iodine numbers were also determined for samples of
Westvaco and Calgon commercial carbons. The relative iodine number was
calculated for each of the activated samples according to the relationship:
„ -, , T :• • « i Monochem Iodine No. „ ., nn
Relative Iodine Number = 7; . ., _—r T~T- M~ x 100
Commercial Carbon Iodine No.
The calculated relative iodine numbers are given in Table 61.
A remarkable increase in the adsorption capacity of the carbon was
achieved through optimization. Capacities were comparable to, and in some
cases exceeded, commercial activities. However, there was some concern about
reproducibility of the results. To check for reproducibility, the activation
procedures for several of the samples were repeated. These results are
tabulated in Table 62 and indicate that the high adsorption capacities are
reproducible and not artifacts.
The carbon activity data, as measured by the iodine number, correlates
well with the results from the weight loss analysis. The highest activities
were achieved at the expense of a high weight loss. This is a reasonable
result as the activation proceeds by the burn-off of carbon, creating innumer-
able micropores. If a 50 percent weight loss is considered acceptable, then
the optimum conditions are as follows:
Water flow 3.4 ml/min
Activation time 20 minutes
Activation temperature 982°C
Expected percent weight loss 51 percent
Measured range of relative iodine numbers 101 - 114 percent
Optimization of the Steam Activation Procedure Using the Large Furnace
There were two major restrictions on the optimization of the large
furnace. First, the maximum temperature was 927°C. Second, the Masterflex
pump could not deliver sufficient water to achieve the same steam concentra-
tions as were created in the small furnace. With these two limitations in
90
-------
TABLE 60. RESULTS FROM THE COMPRESSION TESTS
Crush Weight
Sample (kg)
I A
2 A
B
3 A
B
4 A
B
5 A
B
6 A
B
7 A
B
8 A
B
9 A
B
10 A
B
11 A
B
12 A
B
13 A
B
14 A
B
15 A
B
16 A
17 A
B
18 A
R
tj
19 A
B
20 A
B
21 A
B
22 A
B
23 A
B
24 A
B
25 A
B
26 A
27 A
B
8.62
5.06
6.44
2.56
4.40
2.39
5.57
3.53
2.53
5.90
7.57
5.42
4.34
7.57
3.69
1.51
6.54
3.70
4.34
3.00
4.43
3.04
6.54
8.44
6.93
9.13
8.98
8.11
8.98
3.95
5.66
4.42
7.91
7.36
7.87
6.50
3.20
2.56
3.86
2.17
4.72
4.21
5.92
4.30
10.99
5.95
1.39
5.62
5.08
0.91
0.83
1.03
Crush
Pressure
(kg/cm2)
131
81
103
46
75
41
96
66
49
106
127
85
72
156
81
29
115
68
96
61
80
69
133
141
108
159
147
142
173
66
88
76
170
128
127
118
55
53
65
37
98
77
99
77
189
104
36
102
91
17
17
20
Average
Crush Pressure
Xg/cm )
1923
1346
888
1020
847
1710
1156
1740
1740
1056
1205
1040
1481
1828
2256
2317
1134
1806
1868
1269
872
995
1301
1954
1025
1414
251
268
Variance About
the Average
—
+167
+210
+413
+122
+155
+ 98
+553
+636
+201
+142
+469
+238
+ 87
+225
+163
+690
+ 6
+461
+ 88
+449
+162
+823
+497
+ 80.9
__
+ 23
91
-------
TABLE 61. RELATIVE IODINE NUMBERS:
PERCENT OF COMMERCIAL CARBON (WESTVACO) ACTIVITY*
Activation Time
Rate of
Water Flow
3.4 ml/min.
Activation Time
Rate of
Water Flow
4.6 ml/min.
10
1
43%
2
52%
3
67%
4
73%
10
13
47%
14
50%
15
60%
16
76%
min.
1600
1700
1800
1900
min.
1600
1700
1800
1900
15 min.
5 1600
46%
6 1700
66%
7 1800
81%
8 1900
84%
15 min.
17 1600
54%
18 1700
65%
19 1800
75%
20 1900
105%
20 min.
9 1600
56%
10 1700
86%
11 1800
114%
12 1900
102%
20 min.
21 1600
55%
22 1700
64%
23 1800
80%
24 1900
121%
30 min.
25 1800
100%
26 1900
130%
30 min.
27 1800
160%
28 1900
Sampled Burned
Key to Table 61: Sample Number
Percent of
Activation Temperature
Commercial Activity
TABLE 62. DETERMINATION OF REPRODUCIBILITY OF ACTIVATION RESULTS
Sample
7
8
11
12
19
20
23
24
Activation
Time (min)
15
15
20
20
15
15
20
20
Activation
Temperature (°C)
982
1038
982
1038
982
1038
982
1038
Relative
1st Run
81%
84%
114%
102%
75%
105%
80%
121%
Iodine Numbers
2nd Run
72%
89%
101%
136%
103%
130%
92%
97%
92
-------
mind, the activation temperature and the water flow were held constant and
only the activation time was varied. The activation temperature was set at
927°C and the Masterflex pump was adjusted to its highest setting at which
it delivered 25 ml of water per minute. (To produce the same steam concen-
tration in the large furnace as was produced in the small furnace, the pump
would have had to deliver 74 ml of water per min.) The activation times were
15, 30, 45, 60, and 120 min.
Steam was introduced into the large furnace by pumping water through
stainless steel tubing which ran first through a small preheater and then
into the large furnace. The preheater was heated to 538°C. The water passing
through the preheater was vaporized into steam which then entered the large
furnace. The samples were activated by heating at 427°C for 15 min and then
at 927°C for one of the selected times. The carbon was stored under nitrogen
between the preactivation and the activation heating stages. Immediately
after activation, the carbon was dumped into water to prevent it from burning.
Samples activated in the large furnace were analyzed to determine the
compression strengths and iodine numbers. The samples were analyzed according
to the procedures described for the small furnace. The results of these
analyses are listed in Table 63 and illustrated in Figure 11. The results
show that the compression strength does not vary greatly with the length of
activation. In addition, the activity of the carbon is a linear function of
activation time up to one hour. Beyond one hour, the linear relationship
ceases; however, the activity of the carbon continues to increase even after
two hours of activation. It was deemed impractical to continue activation
times beyond 2 hrs, so the activation time was chosen to be 2 hrs. This
concluded the optimization procedure.
COMPARATIVE ISOTHERM AND BREAKTHROUGH DATA
It is important to generate isotherm and breakthrough data in direct
application to a wastewater under investigation. Data gathered under labora-
tory conditions with pure water may not be relevant to a particular process
effluent. Therefore, for final evaluation, a set of experiments was developed
using isotherm and breakthrough data for several of the process streams
studied in this program. The experiments compared the best developed activated
Monochem soot carbon with several commercial carbons. Three facets of this
study were:
1. Isotherm determinations and comparisons.
2. Breakthrough data for the Monochem active soot carbon compared to
the best commercial carbon from the isotherm data.
3. Regeneration of readsorption evaluation of Monochem active carbon
and Nuchar carbon.
Four activated carbons were used for each of the above tests: Calgon-
400, Nuchar WVC, Barneby-Cheney, and Monochem activated soot carbon optimum.
The Monochem activated soot carbon represented a recipe with methocel HB
93
-------
TABLE 63. RESULTS FROM COMPRESSION TESTS AND IODINE NUMBERS:
SAMPLES ACTIVATED IN LARGE FURNACE
Sample
1
2
3
4
5
Average Relative
Activation Crush Pressure Iodine Number
Time (min) (kg/cm^) (commercial-100%)
15 117 44%
30 95 52%
45 95 65%
60 56 71%
120 -- 84%
w
•HL
M -U
GJ CJ
^ < 80
B _j
ft
(U U
•S « 60
Relative loc
(Percent of Comj
4>
o
-^
1 1 1 1 1 1 1
- /^ -
A 1 J I.I 1 1 1
30
60
90
120
Activation Time (min)
Figure 11. Iodine number results from large
oven optimization.
94
-------
which had been preactivated for 15 min at 437°C and activated at 927°C in the
presence of steam. The best of the three commercial carbons was selected
from these isotherm tests to make continuous flow determinations comparison
tests with the Monochem carbon.
Three process effluents were selected for isotherm determinations. In
addition to the tests on raw effluents, the effects of various treatment
sequences on the ability of activated carbon to remove trace contaminants
were determined. Consequently, the following tests were planned:
Isotherm Determinations—
Stream 231-A—No pretreatment raw process effluent; ozonated process
effluent; biotreated; ozonated followed by biotreatment; pretreatment by
acidification and filtration; pretreatment by acidification and filtration
followed by biotreatment.
Stream 221-A—After steam stripping; after steam stripping and ozonation.
Stream 351-A—No pretreatment.
Breakthrough Data—
Stream 231-A—No pretreatment raw process effluent; ozonated process
effluent; biotreated; ozonated followed by biotreatment; pretreatment by
acidification and 'filtration; pretreatment by acidification and filtration
followed by biotreatment.
Stream 221-A—After steam stripping; after steam stripping and ozonation.
Stream 351-A—No pretreatment.
Regeneration—
Stream 231-A—No pretreatment raw process effluent; ozonated process
effluent; biotreated; ozonated followed by biotreatment; pretreatment by
acidification and filtration; pretreatment by acidification and filtration
followed by biotreatment.
Stream 221-A—After steam stripping; after steam stripping and ozonation.
Stream 351-A—No pretreatment.
Isotherm Determinations and Comparisons
All three commercial activated carbons were readily available. The
Monochem carbon was preactivated for 15 min at 427°C and activated for 2 hrs
at 927°C, utilizing steam as the activation medium. Tabulation of the influent
and residual conditions for the various experiments are given in Table 64.
Plots of the isotherm data are given in the Appendix. The isotherms show
that Monochem continues to adsorb better than the Calgon carbon; however,
Barneby-Cheney adsorbs better than Calgon on some treated streams, and
Nuchar adsorbs better than Calgon on other streams. By comparing the
residual concentration and/or ultimate loading factor, Nuchar seemed to be
the best of the three commercial activated carbons.
95
-------
TABLE 64. TABULATION OF MULTICOMPONENT ISOTHERM RESULTS
(Steam Activation Gas)
Stream
231A
231A
231A
231A
231A
231A
231A
231A
231A
231A
231A
231A
221A
221A
351A
351A
Treatment
Raw
Acidified
Ozonated
>
Biotreated
Ozonated-
Biotreated
Acidified-
Bio treatment
Steam Stripped
Bottoms
Ozonated-Steam
Stripped Bottoms
Raw
Influent
PH TOC(mg/l)
12.0
12.3
2.5
7.0
11.8
11
8.3
8.9
8.2
9.0
8.3
1.0
0.3
0.5
11.7
12.0
235
419
224
333
147
193
144
143
68
80
209
466
396
350
59
56
Residual Concentration Ultimate Carbon Loading Factor
mg/1 TOG mg/TOC/MC
Monochem Calgon Nuchar B-C Nuchar B-C
17
35
34
0.89 1.64
40
1.3 1.0
Nonequilibrium
23
2
12
60
180
9
29
20 52
15
12 26
18
75
72
75 95
33 — — — —
3.5 0.3
-------
Breakthrough Data
Using data obtained from the multicomponent isotherm study, a continuous-
flow carbon study was initiated using Monochem and Nuehar carbons. A continu-
ous flow system consists of a series of carbon columns through which the
water to be treated is pumped continuously until contaminant "breakthrough"
occurs. Breakthrough is the point at which the contaminant concentration
reaches an undesirable concentration. This method gives an accurate estimation
of the large-scale operating capabilities of the activated carbon.
The laboratory pilot unit constructed was similar to the unit used in
the first bench study. Two glass columns 2.54 cm in diameter, and 91.44 cm
long, were constructed into the continuous-flow apparatus, as seen in Figure
12. One column supported the Monochem activated carbon, the other supported
Nuehar. The process streams utilized in the isotherm experiments were used
for these tests.
A glass wool plug was placed in the bottom of each column to trap any
finely granulated carbon particles. All streams were filtered prior to being
subjected to the continuous-flow system; however, a glass wool plug was
placed at the top of each column to remove any particulate matter present
after filtration. Teflon 0-rings were evenly spaced in the carbon column to
minimize channeling. The filtered feed was pumped to each column using one
double-head, calibrated Masterflex pump and control unit. By using a tight-
fitting rubber stopper at the top of each column, a pressure equivalent to
the flow rate force of the pump was exerted on the feed. Although a free-
flow system was desirable, some pressure was detectable depending on how
tightly packed the column was.
The breakthrough curves generated in the initial phase o| the project
indicated a hydraulic loading of between 0.35 and 3.5 1/min-m to yield
satisfactory results. The lower loading gave the_best adsorption data for
the streams being studied; therefore, 0.7 1/min-m was chosen to be the
initial loading. This hydraulic loading was calculated to be approximately a
40 ml/min flow rate. Time of collection was approximately 11 min for every
500 ml sample. Results of the continuous flow study are listed in Table 65.
Regeneration
Regeneration was carried out on the 231A raw and 351A raw process streams.
The procedure for regeneration was as previously set forth:
1. Wet carbon dried at 103°C and weighed.
2. Regenerated at 760°C for 2 hrs under steam.
3. Quenched, dried at 103°G and reweighed.
4. Weight loss calculated.
5. Continuous-flow study on regenerated carbon.
97
-------
Glass wool _
Carbon _
Teflon 0-ring^
T=*=
Backflush
Jc
¥
»
»
MM
^^
h_
1
;. Glass V Glass -
wool _ ' ..wool
Feed
pi.
9
«
^mi
M^
«^
*
.Glass wool
_Carbon
^-Teflon 0-ring
==X=T
Backflush
Sample Sample
Collection Collection
Figure 12. Continuous-flow apparatus.
98
-------
TABLE 65. CONTINUOUS FLOW DATA
231A TDI 231A TDI 231A TDI 231A TDI 351A Aromatic
Raw Biotreat Acidified Ozonated, Biotreat Raw
Monochem Nuchar Monochem Nuchar Monochem Nuchar Monochem Nuchar Monochem Nuchar
Initial
,0 Carbon Weight (g) 135.6g 154.4g 67.8g 67.8g 25.Og 25.Og 20.Og 20.Og 22.2g 21.8g
VD
Volume at
Breakthrough
(Liters) 9.5 10.0 8.0 27.0 1.0 2.5 .5 1.0 15.0 16.5
Mg TOG
Adsorbed/
g carbon 34.5 33.6 18.9 42.8 5.6 10.7 1.1 2.5 90.9 148.4
-------
6. Analytical results, TOG, COD, BOD_ (mg/1) were plotted versus
volume (liters).
A composite list of data correlating activated and regenerated information
is given in Table 66. Breakthrough curves are given in the Appendix.
Discussion of Results
After running the comparative continuous-flow study on Monochem and Nuchar
carbons, definite conclusions were drawn about the adsorbing and regenerating
ability of the Monochem activated carbon.
1. On all streams, Nuchar adsorbs at least 30 percent better than
Monochem carbon—sometimes as high as 55 percent better.
2. It takes longer for breakthrough to occur for Nuchar than
Monochem carbon.
3. Upon regeneration, Monochem has approximately a 50 percent weight
loss, while that of Nuchar is rarely over 20 percent.
4. Monochem seems to retain its adsorbing ability after regeneration,
while Nuchar tends to lose much of its ability.
100
-------
TABLE 66. CONTINUOUS-FLOW DATA ON ACTIVATION AND REGENERATION PROCEDURE
231A TDI 231A TDI 231A TDI 231A TDI 351A Aromatic 231A Regenerated 351A Regenerated
Raw Biotreated Acidified Ozonated.Biotreat Raw Raw Raw
Stream Monochem Nuchaf Monochem Nuchor Monochem Nuchor Monochem Huchor Monochem Nuchor Monochem Nuchor Monochem Nuchor
Initial Carbon
weight used 135.6g 154.4g 67.8g 67.8g 25.Og 25.Og 20.Og 20.Og 22.2g 21.8g 60.0 60.0 19.7g 19.7g
Volume at
Breakthrough
in Liters 9.5 10.0 8.01 27.01 1.01 2.51 .51 1.01 15.0 16.51 47.51 47.1 8.51 10.1
Mg TOC
Adsorbed
g carbon
34.5mg 33.6mg 18.9mg 42.8mg 5.6mg 10.7mg l.lmg 2.5mg 90.9mg 148.4mg 45.1mg 60.3mg 89.3mg 120.3mg
Weight Loss
in Grams due
to Regeneration 60.4g 28.2g 20.6g 9.9g 10.6g 10.2g 10.2 4.7 11.9 2.2
Percent Weight Loss
from Initial Weight 48.4 18.2 30.3 14.6 42.4 40.8 51 23.5 53.6 10.0
Percent Gain or Loss
Mg TOC ,
—» after
gc
Regeneration
23.5 44.2 1.7 18.9
gain gain loss loss
-------
SECTION 7
SOLVENT EXTRACTION
As an integral part of EPA Grant No. S800773, "Development of Treatment
and Control Technology for Refractory Petrochemical Wastes," solvent extrac-
tion was evaluated as an effective means to reduce organic contaminants in
preselected process effluents. The waste streams were made available by
eight participating industries in the Lower Mississippi River area. Inspec-
tion of the contaminants in the various process effluents led to a grouping
of the wastewaters by manufacturing process:
1. Ethylene dichloride via
2. Ethylene dichloride via
3. Perchloroethylene and via
1,1,1-trichloroethylene
4. Ethyl benzene via
5. Toluene diamine via
6. Polyols via
7. Acetylene via
8. Methylcellulose via
direct chlorination
oxychlorination
chlorination of
ethylene dichloride
alkylation of benzene
hydrogenation of dinitrotoluene
polymerization of ethyleneoxide
combustion of methane
propylene oxide
Four process effluent categories were considered for evaluation.
Extraction process evaluations and subsequent field tests were based on the
following considerations:
102
-------
Stream Type
No. Contaminant Discussion
221A Halogenated hydrocarbon Easy to solvent extract.
Solvent selection dependent
on residual in wastewater
and its resulting need for
further treatment.
351A Aromatics High dissolved solids.
Aromatics could be solvent
selective.
231A Aromatic amines Not too amenable to
extraction.
231B Polyols Some detergent action which
could inhibit solvent
selectivity.
A three-point program was devised and carried out to determine if
these process effluents were amenable to solvent extraction for contaminant
reduction.
1. Laboratory development work on solvents selected for evaluation
of the process effluents.
2. Demonstration unit design
3. Demonstration unit operations.
LABORATORY DEVELOPMENT ON SOLVENTS SELECTED FOR EVALUATION OF THE PROCESS
EFFLUENTS
As previously discussed, the halogenated hydrocarbon contaminated
process effluents were amenable to removal of organics by solvent extraction.
The aromatic contaminated wastewater from the styrene process offered
similar possibilities. Three separate process effluents were investigated
on a laboratory scale for removal of trace organics by selective solvent
extraction (Table 67).
TABLE .67. . PROCESS EFFLUENTS
Stream Number Type of Process Effluent
161A Contaminated with chlorohydrocarbons
from oxychlor process
221A Contaminated with chlorohydrocarbons
from oxychlor process
231A Contaminated with aromatic amines
103
-------
Stream 161A
This wastewater is from an ethylene oxychlorination process producing
chlorinated solvents. The stream was analyzed first for gross parameters
(Table 68). The results of these tests revealed some variation in the con-
taminant loadings of the wastewater from day-to-day operations. The process
effluent was tested for organic compound removal with the kerosene and C..--
GI„ petroleum solvents.
TABLE 68. INITIAL EVALUATION OF STREAM 161A
Parameter Sample 1 Sample 2
PH
TOC (mg/1)
TOD (mg/1)
Cl~ (mg/1)
Acidity as CaCO_ (mg/1)
Acidity as Chlorine (mg/1)
1.75
61.7
176
3240
3055
0.22
1.76
123
480
4400
2431
0.09
Conventional separatory funnel laboratory experiments were set up to
determine the equilibrium conditions for contacting these solvents with
this wastewater. The results of these tests are shown in Table 69.
TABLE 69- EXTRACTION OF STREAM 161A WITH C1Q-C12 SOLVENT AND KEROSENE
(Solvent to Wastewater Ratio / to 1)
Kerosene Cin-C1
-------
These two solvents offer excellent possibilities to utilize in the
field test on the larger continuous demonstration unit. There were a
number of advantages in using these two solvents:
1. Excellent selectivity for the specific organic compounds present.
2. Low solubility in the wastewater after extraction.
3. Availability at a reasonable cost.
4. Post-biodegradability and non-toxic characteristics.
5. The boiling point enables separating the extracted organics by
distillation.
The C,Q-C-- hydrocarbon fraction is produced in a petrochemical opera-
tion. This paraffin-based material is extensively used in detergent manufac-
ture and is used as the starting material conversion to a substituted fatty
acid. Because of this specific use, it is manufactured at a rather wide
temperature range. The initial boiling point could cause some problems in
distilling the organics from the solvent.
The subsequent field test results showed that these two solvents could
be used effectively for removal of chlorinated hydrocarbons from process
effluents.
Stream 221A
This wastewater is also from an ethylene oxychlorination process
producing primarily EDO. The stream was analyzed for gross parameters
(Table 70). Initial evaluations of two samples from this process effluent
indicated that there was little variation between the processes. The
analysis of the organic components averaged between the two samples is
shown in Table 71.
TABLE.70. . INITIAL,EVALUATION OF STREAM.221A
Parameter
pH
TOG (mg/1)
COD (mg/1)
TOD (mg/1)
Cl (mg/1)
Acidity as CaCO, (mg/1)
Suspended Solids (mg/1)
Volatile Suspended Solids (mg/1)
Sample 1
—
—
1383
2253
6000
7565
9.7
0.9
Sample 2
1.29
866
1309
1940
6800
8233
8f\
.0
Of\
.9
105
-------
TABLE 71. ORGANIC COMPONENTS IN STREAM 221A
Component
Concentration (mg/1)
Methyl Chloride
Chloroform
Ethylene Bichloride
2
683
8331
Since this wastewater contained primarily ethylene dichloride, solvents
other than kerosene and C-n-C19 hydrocarbons were evaluated. Examination
of the previous solvent work on stream 161A indicated that the petroleum
fractions were effective in removing EDC from the process effluent. Several
solvents were selected which were considered both selective and easily
separated from the extracted organics by distillation. Table 72 presents
the results of these tests with the solvents selected.
TABLE 72. SOLVENT EXTRACTION OF WASTE
STREAM 221A
Solute
Solvent Volume (ml)
Carbon Bisulfide
1,1,2, 2-Tetrachloroethane
Pentane
Cyclohexane
Freon
25
50
100
25
50
100
25
50
100
25
50
100
25
50
100
Solvent
Volume (ml)
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
TOC
(mg/1)
Before After
Extraction Extraction
357
357
357
357
357
357
357
357
357
357
357
357
357
357
357
370
378
350
485
560
762
406
508
521
430
378
346
378
393
362
As an initial evaluation, the TOC values of the influent and effluent
after extraction tests were used to indicate the potential value of the
selected solvents. Results of this series of extraction tests were disap-
pointing. With the exception of two values, the effluent TOC was higher
than the TOC in the influent. (This is because the low molecular weight
solvents are more soluble in water than the higher, longer chain hydro-
carbons.) The selectivity of these solvents for specific organics in the
wastewater was not checked. It was apparent that since the initial TOC was
not Being reduced, the wastewater would have to be treated by other means
to remove contaminants in order to comply with existing effluent standards.
106
-------
Stream 231A
This stream was a wastewater from the manufacture of toluene diisocyanate.
The primary organic constituent was toluene diamine, the actual organic
composition was complex due to the many side reactions which occurred in
the process sequence. The gross characteristic parameters are presented in
Table 73. It was decided to evaluate the extraction efficiency of several
selected solvents utilizing TOC as the analytical means of indicating
solvent selectivity. These results are shown in Table 74.
TABLE 73. INITIAL EVALUATION OF STREAM 231A
Parameters Concentration
PH 11.6
TOC (mg/1) 687
COD (mg/1) 663
TOD (mg/1) 4388
Phenol (mg/1) 6.6
Cl~ (mg/1) 2500
Alkalinity as CaCO., 2408
Suspended Solids (mg/1) 11.8
Volatile Suspended Solids (mg/1) 3.1
As in the case of process effluent 221A, this wastewater was not amenable
to extraction with the various solvents selected. The only solvent which
showed any selectivity was 1,1,2,2-tetrachloroethane. Removal of organics,
as indicated by TOC measurements, was approximately 50 percent. This reduct-
ion would not be sufficient and further treatment of the wastewater would be
required.
DEMONSTRATION UNIT DESIGN
The initial investigations indicated that the organics in the wastewater
would be in the range of 1,000 to 10,000 mg/1. A concentration of 9,000
mg/1 was used as the basis of calculations. The distribution coefficient
was assumed to be 30 for these determinations.*
Design Calculations for Demonstration Unit
X - 9,000 X = 27
X /Xf = 27/9,000 = 0.003
Solvent to process effluent ratio assumed to be 10
B = 0.10 x 0.75 = 0.075
A = 1.0 x 1.0 = 1.0
E =30 x 0.075 = 2.25
To accomplish such an extraction result would take from 10 to 15 theoret-
ical stages. The amount of removal of trace organics by solvent extraction
*Preliminary designs furnished by Texas A & M.
107
-------
TABLE 74. SELECTED SOLVENT EXTRACTION OF STREAM 231A
Solute
Solvent Volume (ml)
Solvent
Volumn (ml)
TOG (mg/1)
Before After
Extraction Extraction
Carbon Disulfide Partially soluble in wastewater.
Dibutyl Phthalate Did not separate readily from wastewater.
Dodecanol 25 100 4320
50 100 4320
1,1,2, 2-Tetrachloroethane
Pentane
Cyclohexane
Freon (H^Cl^)
Octane
Dodecane
1 , 4-Dichlorobutane
1,2,3, 4-Tetrachlorobutane
C10~C12 Paraffins
25
50
100
25
25
50
25
50
100
25
50
100
25
50
100
25
50
100
40
100
25
50
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
4320
4320
4320
4320
4320
4320
3704
3704
2704
3704
3704
3704
3704
3704
3704
3704
3704
3704
3704
3704
3704
3704
3704
4687
4221
3288
1974
2089
4320
4348
3992
3388
3521
3829
4056
3948
3969
4497
3658
3858
6390
7489
15359
3498
5171
3715
4152
3877
108
-------
determines the number of stages required to achieve the desired level of
organics in the extracted aqueous process effluent. The quantity of through-
put determines the diameter of the extraction equipment.
Final Demonstration Unit Design and Construction
A liquid-liquid extraction pilot unit capable of processing 3.8 liters
of wastewater per minute with a solvent rate of 0.76 1/min was designed.
Specifications for the design were as follows:
Wastewater Specifications
Dissolved inorganics
Dissolved organics
Flow
Acidity
Solvent Specifications
Solvent
Flow
Column Specifications
Extraction column
Stripper
Materials of Construction
Main vessels
Seals, gaskets, etc.
5% hydrochloric acid; others
negligible
Primarily, ethylene dichloride
concentrations from 1000 - 9000
mg/1, boiling point 57°C
3.8 1/min
5% HC1
GI -C1„ paraffin based oil,
boiling range 204 to 315°C
0.76 1/min
0.01 m diameter, 3.5 m tall
0.005 m diameter, 2.25 m tall
Haste|loy-C, gj
Viton , Teflon
ass, etc.
, polypropylene
Chem-Pro Equipment Corporation of Fairfield, New Jersey, was contracted
to fabricate the liquid-liquid extraction and solvent recovery components
of the pilot unit. From preliminary design drawings, Chem-Pro developed
working drawings of the solvent extraction and solvent recovery systems
(Fig. 13).
Equipment—
The liquid-liquid extraction system was a Karr reciprocating plate
extraction column which was 3 m in height and 10 cm in diameter. There
were 65 plates mounted on a central shaft reciprocated by a drive mechanism
located above the column. A 1.25-cm amplitude of reciprocation was set by
adjusting the length of the crank arm. The speed of reciprocation was a
function of the solvent particle size, solvent distribution throughout the
column, and the flow rates of the aqueous feed and solvent. Materials of
construction were a Hastelloy-C shaft, borosilicate glass pipe, and T.F.E.
The stripper was 2.25 m in height, made of 0.05 m schedule 10 Hastelloy-C
pipe with a 0.15 m section of Hyperfill packing. Steam was injected through
^Viton and Teflon are registered trademarks of E.I. Du Pont de Nemours and
Co., Inc., Wilmington, Delaware.
109
-------
Solvent Extractor-
Legend
R - Rotameter
P - Pressure Gauge
T - Temperature Probe (Recorder)
X- Needle Valve
20-liter tank (condenser)
140-liter
Feed Reservoir
Pumps
60-liter epoxy-lined tank
Chloro solvent-water separator
Stripping column
120-liter epoxy lined tank
Rich oil-water separator
Oil layer
Water layer
60-liter epoxy-lined tank (lean oil reservoir
and heat exchanger)
Cold water in
Cold water out
Figure 13. Solvent extraction pilot plant (operational concept).
-------
a 0.02 m nozzle, 0.53 m above the bottom of the column. The variable-flow
self-priming pump, made by Flotect, Inc., can lift water 6-7.5 m and dis-
charge between 0-18 1/min. The body and impeller were made of a chemical-
resistant epoxy plastic and the shaft, cone, and backplate were made of
Hastelloy-C. The power requirements for the 1/3 HP explosion-proof motor
were 230V, single-phase current.
DEMONSTRATION UNIT OPERATIONS AT PLANT SITE
Based on the laboratory evaluations of solvent selectivity, it was
decided to set up the field trials utilizing a petroleum kerosene fraction
as the first solvent and a C-.-C. paraffin based on petroleum fraction as
the second solvent (Table 75J;
TABLE 75. PROPERTIES OF KEROSENE AND C^-C,,, SOLVENT
—" iu — ±z
Solvent
Boiling Range (°C)
Specific Gravity
Viscosity (cP @ 24°C)
Kerosene
149-260
0.862
1.1
C10-C12
174-216
0.75
1.1
>
Solvent characteristics of kerosene-type distillates vary according to
the amounts of paraffins, aromatics or naphthas present. These classes of
compounds vary depending on the source of crude from which they are manufac-
tured. The kerosene base used for these tests contained more paraffin and
naphtha based products than aromatics. Its solvent.properties were more
like the C --C.. - based material. The C10~C12 Petroleum fraction is obtained
by selective fractionation, adsorption or extraction of a paraffin petroleum
fraction. Both solvents were readily available at a reasonable price.
The demonstration unit was moved to the process effluent stream 161A
and set up for the trial runs. It was planned to utilize the kerosene-
based material first, followed by the C^-C^ solvent.
Kerosene-Diesel Oil Trials
Kerosene-diesel solvent was run to find an optimum water-to-solvent
ratio. Fifteen extraction runs were made with kerosene-diesel oil with
water-to-oil ratios varying from 3.7:1 to 18.3:1. Table 76 shows total
chlorinated organics removed. The effects of kerosene-diesel oil solvent
on removal of the three main contaminants (1,1,2-trichloroethane, EDC, and
1,1,2,2-tetrachloroethane) from stream 161A are presented in the Appendix.
An approximate ratio of 5.5:1 (water:solvent) gave the maximum percent
removal of 94 to 96 percent of the chlorinated organics.
The percent of the chlorinated organics removed increased as the
water-to-solvent ratio decreased, except for the runs on October 11 and
December 11. No explanation can be given for the deviation of these two
runs from the norm. Runs on December 10 and 16 were very high in chlorinated
111
-------
TABLE 76. CHLORINATED ORGANICS REMOVED WITH A KEROSENE-DIESEL SOLVENT
Solvent
Flow
Date (1/min)
10/23/74
10/11/74
10/11/74
10/31/74
11/21/74
11/25/74
12/11/74
12/16/74
12/04/74
01/06/75
12/2-3/74*
12/10/74
12/17/74
01/07/75
0.205
0.205
0.205
0.288
0.205
0.205
0.205
0.205
0.205
0.205
0.205
0.205
0.205
0.204
Water Chlorinated
Flow in
(1/min) (mg/1)
3.76
2.81
1.86
2.62
1.86
1.14
1.14
1.14
1.14
1.14
1.1 & 0.8
0.76
0.76
0.76
1,591
907
715
609
334
1,043
701
3,689
382*
654t
771#
2,757
907
413t
Organics Amount
Out Removed
(mg/1) (mg/1)
514
81
101
92
62
58
129
143
224*
33t
63#
122
45
20t
1,077
826
614
517
272
985
572
3,546
158
621
700
2,635
862
393
Ratio
Percent H^O
Removed Solvent
68
91
86
85
81
94
82
96
41
95
92
96
95
95
18.3:1
13.7:1
9.1:1
9.1:1
9.1:1
5.5:1
5.5:1
5.5:1
5.5:1
5.5:1
3.7:1 & 5.5:1
3.7:1
3.7:1
3.7:1
*The solvent was reused before it was transferred to the stripper to deter-
mine if the oil was loaded with chlorinated organics. EDC and three other
components increases in the raffinate while seven components decreased at
varying degrees.
tAn uncontaminated solvent was used in this run to compare results with the
recycled solvent.
#The lab combined samples from 12/2 and 12/3 when H20:0il ratios were 3.3:1
and 6.0:1, respectively.
112
-------
organics (2,757 and 3,689 mg/1, respectively), with both runs having 96
percent removal. These results indicated that further concentration of the
contaminated waste streams would increase the efficiency of removing chlori-
nated organics from water.
On January 6, an uncontaminated solvent was used to compare results
with the recycled solvent. The total chlorinated organics in the raffinate
(effluent water) were 33 mg/1 for a water-to-oil ratio of 5:1. This is 25
mg/1 lower than the best results obtained using the recycled solvent. Any
increase in the solvent or steam flow caused overhead entrainment. A
laboratory analysis of the solvent indicated that there was 1.24 percent to
0.23 percent C12 in the unstripped solvent and 0.02 to 0.03 percent Cl_
after the solvent had been stripped (Table 77).
TABLE 77. EFFICIENCY OF PILOT STEAM STRIPPER FOR REMOVAL OF CHLORINATED
ORGANICS FROM EXTRACT
Tested for Total Cl- in the Kerosene-Diesel Oil Extract
Percent
Date
11/4/74
12/3/74
12/17/74
Before Stripper
0.23% C12
0.74% C12
1.24% C12
After Stripper
0.02% C12
0.02% C12
0.03% C12
Removed
91
97
98
C10~C12 Solvent
A water-to-solvent ratio of 16.4:1 gave the maximum number of theoreti-
cal stages and 93.8% removal of the chlorinated organics (Table 78). The
number of theoretical stages increased as the water-to-solvent ratio increased
up to 16.4:1. At 20.0:1, problems with the column flooding and poor disper-
sion of the solvent caused the percent removed to drop to 72.3 and 76.6.
The C -C paraffin solvent was not recycled and no tests were run to
determine the amount of chlorinated organics in the extract.
To estimate the solubility of the C1Q-C12 solvent in water, a comparison
between the GC analyses was made. The TOC analyses were run four days to
one week after the samples were taken. (With this delay, errors in the
comparison of GC carbon and TOC are possible.) The influent TOC should not
be smaller than the influent GC carbon. For the runs on 3/24, 3/26 and
3/27, the influent GC carbon is larger than the influent TOC. These runs
were not included in determining the solubility of the solvent (C107C12^
in the raffinate. A summary of all runs is shown in Table 79, and is
accompanied by a sample calculation. Comparing the TOC analysis and the
GC analysis, the solubility of the solvent in the raffinate was estimated
to be 12-27 mg/1.
113
-------
TABLE 78.
• ±u — ±£
Date
3/20
3/17
3/25
3/21
3/24
3/26
3/27
Ratio
(HO: Solvent)
5:1
6.5:1
8:1
10:1
16.5:1
20:1
20:1
Water
Flow
(1/min)
1.23
1.71
2.13
2.66
4.37
5.32
5.32
Solvent
Flow
(1/inin)
0.27
0.27
0.27
0.27
0.27
0.27
0.27
•s
Chlorinated Organics
In Out
(mg/1) (mg/1)
148.4
184.6
164.9
297.4
266.6
713.8
672.0
3.2
3.0
1.8
6.6
16.5
197.6
157.3
Percent
Removed
97.8
98.4
98.9
97.8
93.8
72.3
76.6
Total Organic
Carbon
In Out
(mg/1) (mg/1)
58
73
59
76
54
173
75
37
48
38
39
75
86
87
Analyzed by gas chromatograph
Analyzed by TOG analyzer Beckman 915
-------
Economic Evaluation for Contaminant Removal of Stream 161A by Solvent
Extraction
The following criteria were established as a basis to develop the
cost-benefit ratio for utilizing solvent extraction removal of trace organics
in the process effluent.
Basis
1. 331 days per year and/or 8,000 hr per year was the annualized
time the unit would be in operation.
2. Straight-line depreciation of capital equipment over the first 10
years of operation.
3. Interest on capital investment at 8 percent.
4. 532 1/min of process effluent as feed to the extractor.
5. The halogenated hydrocarbon organics which are extracted will be
steam-stripped and recycled back to the the process.
6. The concentration of halogenated hydrocarbon in the process
effluent was 1000 mg/1.
7. Recovered product value was equal to 13.4c/kg.
Capital Cost-
To arrive at a capital cost estimate for installing the equipment
necessary to extract the wastewater and steam strip the resulting organic
rich solvent, conventional techniques of major equipment pricing were employed
with appropriate factoring.
Equipment Cost Estimate
Equipment
1. Feed storage tank $ sn'nnn
2. Extractor column 0.91 x 9.1 m 10 000
3. Stripper column 2.54 cm x 7.62 m 2 000
4. Stripper column condenser 3*000
5. Stripper column exchanger 3*000
6. Stripper column cooler 10'000
7. Pumps 10^000
8. Instruments ^2 QOO
9. Piping, valves and fittings 13*000
10. Intermediate drums and tanks 15*000
11. Miscellaneous equipment !
Subtotal $136,000
Total Installed Cost 315,000
115
-------
TABLE 79. DETERMINATION OF C10-C12 PARAFFIN
SOLVENT SOLUBILITY
Date
3/20
3/17
3/25
3/21
3/24
3/26
3/27
Ratio
H20: Solvent
Influent
TOC
(liter/liter) (mg/1)
5.0:1
0.5:1
8.0:1
10.0:1
16.5:1
20.0:1
20.0:1
Component
VC
EC
VDC
Trans
Cl
ED£
TCP.
Tri
Per
U-Tetra
S-Tetra
Penta
58
73
58
76
54
173
75
SAMPLE
Influent
(mg/1)
0.7
0.1
19.7
28.5
44.4
4.5
30.9
27.7
14.3
4.1
6.8
2.4
Influent Effluent Effluent Solubility of
GC Carbon TOC GC Carbon Solvent in H,0
(mg/1)
32
38
36
62
57
179
148
(mg/1) (mg/1) (mg/1)
3J 1
48 1
38 1
39 2
75 4
86 48
37 36
12
14
18
27
—
—
—
CALCULATION RUN OF 3/17
(Times)
X
X
X
X
X
X
X
X
X
X
X
X
Carbon (Equals)
.38436
.37235
.24780
.24780
.24780
.24274
.18007
.18283
.14486
.14312
.14312
.11875
Influent
GC Carbon
(mg/1)
.27
.04
4.38
7.06
11.00
1.09
5.08
5.06
2.14
.59
.29
38.47
Effluent GC Carbon = 0.34 mg/1
Influent TOC = 73 mg/1
Effluent TOC = 48 mg/1
There possibly? could be compounds in the influent water which are
not shown by the GC analysis that do contain organic carbon.
Assume all the organic carbon not shown by the GC analysis in the
influent is not extracted and remains in the effluent.
Influent TOC
-GC Carbon
73 mg/1
-38 mg/1
Influent Organic Carbon not shown by GC Analysis 35 mg/1
Effluent TOC
-GC Carbon
48 mg/1
-1 mg/1
47 mg/1
-Influent Organic Carbon not shown by GC Analysis -35 mg/1
Organic Carbon from the Solvent 12 mg/1
Organic Carbon from the Solvent
. Percent Carbon in the Solvent (Ci
12 mg/1
.34615
Solubility of the Solvent in Na0Sr
14 mg/1
116
-------
Operating Cost—
The items which go into the operational cost of such a processing unit
are listed below. The manpower allowance was arbitrarily set at a total of
three men for an around-the-clock operation. This is actually less than a
full-man per shift, and it is assumed he could perform other duties associated
with the major process for a portion of his shift duties.
Operating Cost
Steam consumption 350,000///yr @ $3.00/1,000# $ 1,000
Electricity 15 KWH @ 2.5C/KWH 4,000
Labor (3 men @ $15,000/man) 45,000
Supplies and chemicals (1.5% capital) 5,000
Solvent (7600 liters/yr @ $0.26/liter) 2,000
Supervision (2.0% of capital) 6,000
Maintenance and materials (8% of capital) 25,000
Total $88,000
Fixed Cost—
The two significant fixed costs applicable to this operation are depreci-
ation and interest on the capital. The actual capital required would have to
be larger than needed for equipment and installation to cover start-up ex-
penses, working capital, and the general and administrative corporate busi-
ness expenses allocated to each cost center. These costs are outlined
below.
Fixed Costs Per Year
1. Depreciation - 10 years straight line $31,500
2. Interest - 10 years at 8% (Averaged over
a 10-year period) 14,000
3. Start-up expenses 15,000
4. Working capital 15,000
5. General and administrative,
insurance and taxes 12,000
Total $88,000
Recovered Product Credit—
The average concentration of ethylene dichloride in the process efflu-
ent which was evaluated on the bench unit varied from 300 to 3,000 mg/1. An
average value of 1,000 mg/1 was taken for purposes of determining a recovered
product credit. Plant value was assumed to be 13.2c/kg.
Removable Organics By-Product Credit
Flow 254,000,000 kg/yr
Recoverable organics in process effluent 1,000 mg/1
Recoverable organics 254,000 kg/yr
Recoverable organics @ 13c/kg 33,000
117
-------
Cost-benefit Value of Solvent Extraction of Stream 161A—
The amount of ethylene dichloride and chlorinated solvents which would
generate 532 1/min of process effluent would have an annual capacity of
approximately 227,000,000 kg/yr. The cost analysis (below) indicates that
the cost to treat the process effluent would be approximately 0.066c/kg.
Solvent extraction as an initial in-plant treatment represents an amortized
cost of less than 1 percent of the product sales price. It should be recog-
nized that this process effluent would have to be included in the final plant
effluent treatment system. That portion of final treatment cost would have
to be allocated on a proportional basis to this process effluent for a final
cost analysis.
Cost Per Year
Variable costs (operations) $ 88,000
Fixed costs (depreciation, etc.) 88,000
Subtotal $176,000
Recovered Credit 33,000
Total Annual Cost 143,000
Cost/liter of Water Treated $ 0.0007
118
-------
SECTION 8
STEAM STRIPPING
The removal of trace organics from process effluents by steam stripping
offers distinct advantages for a number of organic processes in the chemical
industry. When the trace organics are volatile and make up the bulk of the
contaminant in the process effluent, this unit operations approach is a
viable means of reducing wastewater loadings for a number of reasons.
1. Direct injection of steam eliminates reboiler cost.
2. Azeotropic boiling points of a number of organic compounds with
water permits their removal with a minimum of aqueous boilup.
3. Regulation of the amount of water taken overhead with the organic
contaminants could create a two-phase system, permitting recycle
of the organic layer to the process and refluxing of the water
layer.
4. Wastewater contaminants are controlled in the immediate process
area with maximum product recovery.
5. Final plant effluent treatment load and resulting cost are mini-
mized.
The most frequently used rule for relating the composition of the
vapor phase of an equilibrium liquid mixture to pressure and temperature is
Dalton's law. It states that the total pressure of a system is equal to
the sum of the partial pressures of the individual components present.
Total Pressure = P-+ P« + P, + . . . . PQ
According to Dalton's law, partial pressure is the pressure that would be
exerted by a component if it were alone at the same molal concentration
that it is in the mixture.
The organics present in a process effluent frequently form a constant
boiling concentration with the aqueous phase. Such a mixture can be care-
fully distilled so that the minimum water required for this azeotrope will
Be taken in the overhead. The resulting overhead system could then be
separated into an organic layer and an aqueous layer containing the excess
organics originally present in the contaminated process effluent.
119
-------
to
By utilizing Dalton's law, an aqueous effluent can be tested to see if
steam stripping will effectively remove trace organics. The contaminated
water can be heated to the effective boiling point of the organic in the
presence of the water. Careful regulation of the water taken overhead with
the organic will permit the separation of the overhead into an organic-rich
phase and an aqueous-rich phase. The organic phase can be recycled to the
process and the aqueous phase utilized as reflux.
As in all processes, this unit operations approach is less applicable
certain trace organic contaminated process effluents.
1. If some of the organics in the effluent do not respond to this
unit operations approach, the resulting stripper bottoms may
still contain significant amounts of BOD,, and TOC.
2. The combination of stream stripping secondary treatment may not
be cost effective.
3. Recycle of stripper overhead to the process may not be possible
due to the miscibility characteristics of the organics. Disposal
of the concentrated organic wastewater may present a greater
problem than the dilute stream.
4. Recycle of the stripper overhead may present long-term process
problems even if the organics form an immiscible layer. Buildup
of impurities through recycle could cause the main process to
produce undersirable trace impurities.
The evaluation of stream stripping for the removal of organics from waste-
waters therefore requires a careful analysis of a specific industrial
operation. The 23 process effluents considered for the applicability of
steam stripping are listed in Table 80. There were four major groups of
organic components in the process streams.
1. Wastewater containing chlorinated hydrocarbons—The 17 effluents
in this group contained 2 basic contaminants: (1) soluble organics
which are volatile and steam strip individually or as an azeotrope;
and (2) soluble organics which are not readily volatile and dif-
ficult to steam strip.
2. Wastewaters containing aromatic hydrocarbons—There were two
process effluents in this category: (1) quench waters from an
ethylene cracking unit and (2) wash waters from an ethylbenzene-
styrene plant. The quench water contained too wide a boiling
range of hydrocarbons to be effectively steam stripped. The wash
water had primary aromatics in the benzene, ethylbenzene, and
styrene cateogry, and was selected for evaluation by steam strip-
ping.
3. Wastewaters containing aromatic amines and related organics—
Water solubility precluded these impurities from being concentrated
by stripping or reboiling.
120
-------
TABLE 80. CODED PROCESS WASTEWATER STREAMS
Stream ID
Major Contaminant
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
011A001
011B001
011C001
041A001
041BC01
041C001
041D001
041E001
041F001
041G001
081A001
081B001
081C001
131A001
161A001
161B001
161C001
221A001
221B001
221C001
231A001
231B001
351A001
Halogenated Hydrocarbons
Halogenated Hydrocarbons
Halogenated Hydrocarbons
Halogenated Hydrocarbons
Halogenated Hydrocarbons
Halogenated Hydrocarbons
Aromatics and Organic Oxides
Halogenated Hydrocarbons
Halogenated Hydrocarbons
Methyl Cellulose and Hydroxyethers
Halogenated Hydrocarbons
Halogenated Hydrocarbons
Halogenated Hydrocarbons
Aromatics (Quench)
Halogenated Hydrocarbons
Halogenated Hydrocarbons
Halogenated Hydrocarbons
Halogenated Hydrocarbons
Halogenated Hydrocarbons
Halogenated Hydrocarbons
Aromatics and Amines
Polyols and Oxides
Aromatics
4. Wastewaters containing polyols and polyhydroxy-type organics—The
high-boiling, water-soluble polyols were not amenable to removal
by steam stripping.
Each of the process effluents was completely characterized before any
decisions were made as to the steam stripping tests which would be applied
to the selected streams. Individual samples were collected for each stream
which were representative of the process conditions, and specific tests
were conducted to establish all of the component parameters. Details of
the tests performed on each process stream are given in Tables 81 and 82.
LABORATORY BENCH-SCALE STRIPPING COLUMN
General Design Conditions
After the initial evaluations of the process streams, a pilot stream
stripping system was designed. The demonstration unit (Figure 14) was
based on a feed flow rate of 3.8 1/min. The column was made from a 5.08 cm
piece of Derokane, with an overall length of 3.67 m. Packing for the
column was pall rings made from polypropylene. Feed to the system was
121
-------
TABLE 81. STREAM CHARACTERIZATIONS
Stream
-ID
011A
011B
one
041A
041B
041C
041D
041E
041F
041G
081A
081B
081C
131A
161A
161B
161C
221A
221B
221C
231A
231B
061A
COD
(mg/1)
3,678
1,112
615
1,764
30
454
325
10,752
18
23,198
6
100
390
1,060
1,040
418
6,279
15,100
11,280
6,080
10,680
2,562
463.
TC
(mg/D
1,665
712
1,703
1,638
250
13
69
2,576
54
7,409
60
17
67
273
1,843
148
4,390
9,022
10,250
5,840
1,310
821
6 479
PH
12.2
1.5
11.2
11.7
2.9
2.6
6.9
5.8
10.5
10.6
10.6
1.4
12.1
9.5
7.7
0.9
<0.1
0.1
<0.1
12.7
9.7
11.9
11.9
Flow
(1/min)
76-228
57-152
288-418
342-456
342-456
342-456
1520-2280
342-456
342-456
114-228
76-114
76-114
76-114
114
76-190
57-114
38-95
57-114
38-76
19-38
38-76
57-95
76-152
Alkalinity
CaC03
(ppny
16,600
0 29
4,840
1,060
0
0
30
375
325
—
1,180
0 2
12,635
220
304
0
0
0 102
0 269
31,800
1,475
1,140
2,060
Acidity
CaC03
(ppmj
0
,254
0
0
135
241
10
8
0
—
0
,680
0
0
10
0
0
,312
,392
0
0
0
0
Chlorides
Cl
(ppm)
18,113
1,988
6,664
2,662
408
762
7
6
418
28,617
1
1,909
9,685
119
6,140
10,587
77,112
116,127
170,289
46
14
32
147
122
-------
TABLE 82. SOLUBLE ORGANIC STREAM CHARACTERIZATION
NJ
U>
Soluble Organics
Kethyl Chloride
Ethyl Chloride
Vinylidene Chloride
Trans 1,2-Dichloroethylene
Cis 1,2-Dichloroethylene
Chloroform
EDC
Carbon Tetrachloride
Trichloroethylene
1,1,2-Trichloroethane
Perchloroethylene,
Tetrachloroethane
Toluene
Benzene
011A 011B 011C 041A 041B 041C
(mg/1) (mg/1) (mg/1) (mg/1) (mg/1) (mg/1)
T 1.0 736.0 0-40 2.4 2.0
19.7 T 0.9 0-30 T
1.0 0.8 0.3
0.1
T 6.0 0.1
2.5 0.1 25.0 0.2
6974.0 2858.0 5755 9122 6.0 6.0
5.0
3.0
17.0
041D 041E 041F 081A 081B 081C
(mg/1) (mg/1) (mg/1) (mg/1) (mg/1) (mg/1)
200.0 3.0 2.0
T'
2.0 5.7 0.1 1.0 0.1
10.0 0.5
0.2 15.6. 1.0
2.0 3.3 6.0 42.0 2.0
1.0 2.0 8.0
202.0 3.0 84.0
7.0 7.0 20.0
306.0
2.0
11.6
161A
(mg/1)
3.0
10.0
0.1
7.0
23.0
200.0
6.0
40.0
14.0
161B 161C 221A 221B 221C 231A
(mg/1) (mg/1) (mg/1) (mg/1) (mg/1) (mg/1)
2.0 14.0 2.0 0.2
6.5 2.7 3.8 21.4
6.5 56.8
19.7 327.4 1.3
64.0 546.3 1.1
683.0 0.96
310.0 10V7.6 8331.0 9238 54.8
0.21
25.5 109.6
29.6 53.6
10.4
2
0!
«
O
o
CO
1
nj
QJ
B
I
•O
Cl
c
0)
=>
o
H
231B 351A
(mg/1) (mg/1)
4J
g
O
O)
T3
S
CO
rH
0
,-H
o
ft-
1.2
5.65
-------
Condenser
I
Feed
Overhead
Rotameter
Reflux
Feed
J\
!____•
Packing
Pump
Thermocouple
0
Pressure
Gauge
' \
i \ V—i
t
EZ
Steam
In
Orific
Plate
m
o
Thermocouple
Packing
Reflux
Overhead
Reserved
/ /////
/#///
'•'/;
-C*
Thermocouples
O=
Bottoms Out
Manometer
Aqueous Layer
Organic Layer
(Product, Recycle)
Figure 14. Model of pilot steam stripper (GSRI - New Orleans).
-------
measured through a calibrated rotameter using a peristaltic pump for flow
continuity. Steam was from an on-line boiler which serviced the main
building. Steam feed was preset with a needle valve and measured through a
calibrated orifice.
To minimize process losses, the bottom and overhead products were con-
densed in a refrigerated bath. All flow meters and orifices were precali-
brated under process conditions. The discharge of the steam pipe was
disconnected and the steam condensed by use of the refrigerated bath. The
condensate was collected and weighed over a period of time; the results
were checked against the pressure drop through the orifice system.
Experimental Procedures^
The following procedures were used to collect the samples on which the
data in this report were based.
Overhead—
The overhead samples from the steam stripper were composited hourly
from 6 samples taken every 10 min in a graduated cyclinder. An average
hourly rate was calculated and recorded. A grab sample was then collected
in a 150-ml brown bottle and analyzed for EDC content.
Bottoms—
The bottoms from the steam stripper were monitored through a sight
glass and valve on the column. An average flow rate for the bottoms was
calculated from samples collected for 1 min in a graduated cylinder.
Samples of the bottoms for EDC analysis were collected in the same manner
as for the overhead samples.
Feed—
The feeds for the pilot steam stripper were received in 210-liter
polypropylene-lined drums and fed to the stripper by a Masterflex pump.
The feed rate was measured at 1-min intervals and feed samples for EDC
analysis were taken at the end of each day.
Steam Rate—
Steam rate to the pilot stripper was measured by a U-type manometer
across an orifice plate.
STEAM STRIPPING TRIALS
Experimental Operations and Results
Halogenated Hydrocarbon Contaminated Wastewaters—
There is a great deal of similarity among the organic components
present in the wastewaters from this industry segment. Three representative
effluents were selected. A program was developed to evaluate steam stripping
and its effectiveness in removing trace contaminants.
1 Steam stripping of stream 221A with a detailed analysis of influent
and effluent from the stripper, including individual organic
component analysis.
125
-------
2. Steam stripping of stream 011C with analysis of the influent and
effluent for broad TOC parameters. In addition, the composition
of ethylene dichloride would be measured during the stripping
operation. (Since this EDC usually is the major organic component
in the process effluents, it was felt that if the EDC were removed,
then other organics present would be reduced to a minimum.)
3. Steam stripping of stream 161A with monitoring of the operation
utilizing only gross parameters such as TOC and BOD_.
4. Evaluation of the three pilot unit tests to see if any additional
data are needed to design a commercial-size steam stripping unit.
Steam Stripping of Stream 221A
To ensure adequacy and uniformity of process wastewater, several
drums were collected and delivered to the pilot unit area. The system was
run continuously while varying specified process conditions. Samples were
collected from the feed source over one day's operation to check the unifor-
mity of the feed. The data indicated that the organics could be easily
removed and that minimum overhead would yield the desired results. Further,
the quantity of overhead should be kept to a minimum if the overhead were
to separate into an organic-rich phase and a water-rich phase. The results
of the test program are given in Table 83.
Discussion of Results—
Variations in the data became evident in comparison between the indivi-
dual test runs. Test 2 "allowed 2.8 percent of the feed to be taken overhead.
The amount of 1,2-dichloroethylene removed was only 2.3 percent. The
material balance for this organic was not as good as it should be; the real
conditions for organic removal were much better than the results indicated,
since it was apparent that volatile organics had escaped undetected.
In the third test, the amount of overhead was increased to 5.1 percent.
As expected, the amount of organics removed increased to 90.74 percent. The
amount of 1,2-dichloroethylene in the bottoms decreased to 14.8 mg/1,
indicating a 90 percent removal for this component. The process material
balance for this organic was still not within the expected experimental
range.
In the final tests, the overhead was decreased to a quantity similar
to Tests 1 and 2, and the results of the first run were checked. The
results of this test were promising. With a lower overhead, the organics
removed amounted to 85.2 percent of the organics in the influent. The
amount of 1,2-dichloroethylene in the overhead was much higher, indicating
that this organic is so volatile that it was difficult to collect even with
refrigeration. Special efforts were made during this run to obtain maximum
overhead. An excellent material balance was obtained for several of the
individual components, though the overall material balance was not good.
The last two tests were conducted by holding the overhead fairly
constant and reinjecting the overhead water phase at varying amounts.
126
-------
TABLE 83. STEAM STRIPPING OF STREAM 221A*
Test
Volume (ml/min)
O.H.
Btms
Percent
O.H.
Steam (ml/min)
Temperature (°C)
O.H.
Btms
Column
Pressure
(Btm/Top)
Mass Balance
Efficiency (%)
Stripper Col. TOG
Btm
O.H.
Feed
Reflux
Ratio
TOG
Removal (%)
1
5.76
350
2.3
44.98
103
104
1.0/1.0
+17-08
(mg/1)
256
10446
645
—
37.3
2
7.05
281.25
2.82
53.1
102
103
1.0/1.0
-4.80
292
10462
668
—
55.8
3
12.75
302.5
5.10
50.8
104
104
1.0/1.0
+2.09
593
4766
645
—
37.7
4
4.3
305
2.3
59.7
104
104
1.0/1.0
+3.099
241
4519
785
1.411
32.9
5
13.5
275
2.5
59.7
103
103
1.0/1.0
-6.845
243
9806
636
0.8511
83.3
*O.H. = Distillate or overhead.
Btms = Column effluent or bottoms.
Negative mass balance efficiencies indicate less material into system than
leaves system.
Column feed = 250 ml/min for all tests.
127
-------
Refluxing the overhead water phase increases the efficiency of the stripping
operation but necessitates increasing the amount of steam for stripping.
The removal of organics varied from 80 to 84 percent. These numbers
compare favorably with the previous runs. The data from these five runs
indicated that up to 90 percent of the individual organics can be removed
through the use of steam distillation. It appears that additional height
of column packing is necessary to maximize the reduction of organics in the
stripper bottom.
Contaminant Loading of the Stripper Bottoms—
The feasibility of the steam stripping of the halogenated hydrocarbon
contaminated wastewater had been demonstrated. The next step was to evaluate
the residual impurities of the stripper bottoms. Samples of each of the
previous runs had been retained for TOG analysis. This parameter was felt
to be a good measure of organics removed in the stripping operation and of
the residual contaminants in the bottoms. The results of these tests are
shown in Table 84. The data indicate a consistent stripper performance.
With the exception of one run, the TOG removal varied between 56 and 69
percent.
Since the biological data had indicated that this residual was reducible
by this process, a combination of these two processes would appear to offer
a feasible treatment sequence.
Steam Stripping of Stream 011C
Stream 011C is from an oxychlor operation manufacturing ethylene
dichloride. An analysis of this stream is shown in Table 85. EDC is the
principal organic contaminant in the wastewater.
TABLE 85. CHARACTERIZATION OF STREAM 011C PROCESS EFFLUENT
COD (mg/1) 615
TC (mg/1) 1703
pH 11.2
Flow (1/min) 380
Alkalinity as CaCO (mg/1) 4840
Chlorides Cl (mg/lj 6564
Vinylidene Chloride (mg/1) 2.8
Dichloroethylene (mg/1) 8.5
Chloroform (mg/1) 28.2
Ethylene Dichloride (mg/1) 1913.0
Trichloroethylene (mg/1) 36.3
Tetrachloroethylene (mg/1) 5.2
Total Organics (mg/1) 1999.2
A number of runs were made on the pilot unit, monitoring the total
organic carbon and determining the concentration of EDC in the influent and
effluent streams. A summary of these data is shown in Table 86.
128
-------
TABLE 84. STEAM STRIPPING OF STREAM 221A COMPONENT ANALYSIS OF FEED OVERHEAD AND BOTTOMS
(ppm)
1.4:1 Reflux 0.9:
to OH Ratio
Vinylidene Chloride
Dichlorome thane
1 , 2-Dichloroethylene
Chloroform
1,1, 1-Trichloroethane
Ethylene Bichloride
Trichloroethylene
Chloral
1,1, 2-Tr ichloroethane
Perchloroethylene
1,1,1, 2-Tetrachloroethane
1,1,2, 2-Tetrachloroethane
Feed
61.49
800.89
1583.3
140.3
50.92
1593.0
—
693.18
14.14
14.89
512.8
14.9
2
OH
124.
3511.
350.
1185.
—
4383.
640.
1213.
24.
—
189.
14.
85.
.3% OH*
BTM**
4 32.8
8 114.1
8 373.7
1 0
—
5 42.20
8 34.2
0 171.9
6 0.19
6.8
8 0
9 32.7
24%
Removal in
Bottoms
2.8% OH
OH BTM
_— —
3277.0 89
269.7 1255
882.4 0
—
4105.5 64
567.0 0
1163.6 177
34.0 0
50.2 0
393.8 0
121.7 49
70.13%
Removal
Bottoms
_
.5
.4
-
.5
.1
.84
.5
in
5.1%
OH
111.2
2736.5
465.0
838.3
—
4731.5
627.4
1185.5
76.5
—
22.7
444.4
90.74%
OH
BTM
0
175.6
14.8
0
—
43.1
22.7
172.6
0
—
0
78.4
Removal in
Bottoms
2.3% OH
OH BTM
__ _
1183.0 296
1320.9 16
412.3 0
— _
3654.5 38
640.8 37
2332.3 464
42.4 0
—
25.8 0
8.7 0
84.4%
Removal
Bottoms
_
.3
.1
-
.6
.2
.3
-
.5
in
1 Reflux
to OH Ratio
2.
OH
179.9
5159.9
679.9
1124.3
173.4
5541.3
644.5
2301.6
66.1
9.6
392.5
24.2
79
5% OH
BTM
0
131.7
0
64.7
41.6
436.4
0
434.4
0
0
1.6
0.1
.9%
Removal in
Bottoms
+Feed to column was 250 ml/min for all tests
*OH = distillate as % of feed
**Btm= bottom, effluent from column
-------
TABLE 86. PILOT PLANT EVALUATION FOR STEAM STRIPPING OF STBEAM 011C
Ethylene D'ichloride
Run
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Feed
ml/min
395
300
300
300
390
390
335
335
335
370
370
245
245
245
245
290
290
290
370
400
400
400
400
500
280
380
380
390
390
390
260
290
290
335
335
350
225
225
240
250
250
280
280
Overhead
ml/min
11.4
16.0
10.0
5.2
16.0
11.0
6.4
12.1
20.0
9.8
5.0
55.0
34.0
19.0
52.0
30.0
11.0
36.0
18.5
31.5
16.6
6.0
13.8
12.5
10.9
19.4
12.3
21.7
10.8
33.0
15.7
9.1
35.1
10.3
42.0
18.3
12.0
47.8
14.1
10.8
42.3
14.2
57.1
Bottoms
ml/min
410
340
335
330
420
410
400
425
350
420
410
290
270
255
290
340
345
380
410
425
410
410
475
530
350
450
390
460
490
470
300
350
391
394
418
400
279
324
309
271
285
301
366
Reflux
ml/min
32
25
22
—
27
—
32
—
—
36.3
—
—
33.3
—
42.9
Steam
ml/min
69
69
59
52
75
69
58
65
72
65
55
87
65
46
87
80
65
83
71
76
69
69
77
84
50
77
72
84
80
87
56
65
84
60
84
80
55
77
56
56
77
65
102
Feed
mg/1
1,716
1,612
1,612
1,962
1,962
1,204
1,204
1,204
2,111
2,111
10,734
10,734
12,900
12,900
8,042
8,042
8,042
7,403
7,403
7,403
7,403
5,600
5,600
5,640
5,640
5,640
5,189
5,762
5,762
5,084
5,084
1,463
1,247
1,247
1,426
1,426
1,457
1,457
1,457
Overhead
mg/1
8,835
10,4>46
9,937
7J65
10,470
11,253
7,364
16,184
11,294
8,143
21,595
16,278
15,655
16,698
14,853
13,880
13,377
14,876
14,012
12,433
12,821
13,411
14,800
13,900
19,629
21,419
16,065
16,409
18,100
19,294
16,956
17,180
6,587
6,838
5,230
6,030
5,935
4,925
7,106
5,895
Bottoms
mg/1
65
65
65
65
65
65
65
65
65
65
65
0.3
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
5.3
1.0
1.0
1.0
1.0
43
3.6
4.0
3.7
5.7
1.0
1.0
1.4
6.0
9.2
9.5
7.0
9.3
6.1
8.5
19.4
Feed
mg/1
801
801
801
739
739
710
740
740
732
765
531
531
531
531
531
531
531
512
512
512
512
TOC
BOD5 COD
Bottoms Botoms
mg/1 mg/1
129
102
109
115
120
185
191
166
160
69
67
30
57
121
111
111
125
125
106
91
69
-
—
61.4
74.9
66.6
77.9
72.6
68.1
—
—
—
116.0
110.0
TOD
Feed Bottoms Bottoms
mg/1 mg/1 mg/1
-
—
518.7
518.7
518.7
518.7
518.7
518.7
518.7
409.2
409.2
409.2
409.2
—
—
125.2
146.3
118.1
2b0.3
233.2
199.5
—
—
—
—
—
156
158
76
70
73
64
188
169
161
188
164
195
210
—
—
—
—
—
—
—
—
—
—
—
—
130
-------
During the experimental tests, the feed, overhead reflux and steam
rate were varied sufficiently to establish optimum operating conditions.
An examination of the continuous operation of the steam stripping column
revealed that it was possible to obtain good material balances. Total
input to the column was represented by the feed stream and the injected
steam. Effluent was represented by the overhead and bottoms. Average
deviation from the column material balance was less than 10 percent.
While obtaining these data on stream 011C, the process effluent was
field sampled on several occasions. Sampling at different intervals pointed
up the variation in organic concentrations of this wastewater.
Good material balances were obtained during the experiment, but mass
balances for the measured quantities of EDC in the influent and effluent
were disappointing. Although a refrigerated system was used to collect the
overhead and bottoms samples, it was apparent that some vaporization did
occur.
Considerable effort was made in the initial tests to achieve equili-
brated operating conditions prior to examining the stripper bottoms for
residual organic contamination. Levels of TOC, BOD,., TOD, and COD in the
stripper bottoms were determined. Removal of these parameters is illustrated
in Table 87. From these data, it would appear that the TOC level in the
final effluent can be reduced to an average value of approximately 100
mg/1. The COD values are somewhat higher, and BOD,, somewhat lower.
/
Some minor oxidation products apparently are present in this wastewater
as in stream 221A. However, the products are not as prevalent and should
present no problems.
Discussion of Results—
The EDC content of the wastewater used as feed to the stripper column
varied from 1,247 to 12,900 mg/1. Over this range, it was possible to
reduce the EDC in the stripper bottoms to less than 1 mg/1. The range of
EDC in the bottoms corresponded to a TOC value of 100 mg/1, a BOD value in
the 75 mg/1 range, COD value in the 150 to 200 mg/1 range, and a TOC value
in the 100 to 200 mg/1 range.
As in stream 221A, there was a soluble organic present in stream 011C
which was not steam stripped. This organic has been identified as chloral,
which is a halogenated ethylene compound with a hydroxyl group attached to
one of the carbon compounds. The addition of this hydroxyl group seems to
enhance the Biodegradability of the compound.
Steam stripping of stream Q11C produces a wastewater quite low in the
measurable parameters of TOC and BOD,-.
Steam Stripping of Stream 161A
The principal product of this process was ethylene dichloride.
Wastewater characteristics are listed in Table 88. The analysis ot this
process effluent indicated that it had fewer contaminants at an overall
131
-------
TABLE 87. ANALYSIS OF STREAM 011C
(Concentrations in mg/1)
Bottom R33
Bottom R34
Bottom R35
Bottom R36
Bottom R37
Bottom R38
Bottom R39
Bottom R40
Bottom R41
Bottom R42
Bottom R43
Bottom R44
Bottom R45
Bottom R46
Bottom R47
Bottom R49
Bottom R50
Feed R43-39
Feed R40-43
Feed R44-60
BOD
61
75
66.6
78
73
68
116
110
106
85
COD
125.2
146.3
118.1
200.3
233
199.5
169
186
170
183
170
518.7
409
539
TOG
67
30
55
122
111
111
125
125
106
91
69
116
117
108
110
86
87
531
512
491
TOD
112
135
129
152
181
213
221
251
249
253
221
'274
253
197
217
468
542
5190
4445
Alka-
linity
1310
1450
1350
1350
1430
1200
1390
1635
1400
1684
1425
2514
2324
2426
2518
4918
4950
1915
2247
2250
Chlor-
ides Cu
4370
5280
4930
5440
5330
4700
5220
5310
5310
6660
5480
700
6210
7700
0.24
0.11
0.17
0.96
0.13
0.11
0.07
<0.09
<0.09
<0.09
0.13
Fe
0.23
1.38
0.67
0.31
4.01
2.41
0.63
0.22
0.07
1.42
2.57
PH
12.0
12.0
11.9
11.9
11.9
11.9
11.8
11.8
11.44
11.45
11.6
11.35
11.95
12.1
12.0
132
-------
TABLE 88. CHARACTERIZATION OF STREAM 161A PROCESS EFFLUENT
COD 1040 mg/1
TC 1843 mg/1
PH 7.7
Flow 189.3 1/min
Alkalinity (as CaCO ) 304 mg/1
Chlorides J 5140 mg/l
Vinylidene Chloride 2.6 mg/1
Dichloroethylene 8.1 mg/1
Chloroform 22.0 mg/1
EDC 25.7 mg/1
Trichloroethylene 7.1 mg/1
Trichloroethane 34.2 mg/1
Perchloroethylene 13.8 mg/1
Tetrachloroethane 39.7 mg/1
Total Organics 153.2 mg/1
lower concentration than the two wastewaters previously evaluated. Therefore,
steam stripping of this effluent provided data on unit operation at a lower
concentration range. A minimum of test runs provided the data for this
wastewater. Results for this stream are shown in Table 89.
Discussion of Results—
The EDC content of the wastewater used as feed to the stripper column
varied from 1,247 mg/1 to 12,900 mg/1. Over this range, it was possible to
reduce the EDC in the stripper bottoms to less than 1 mg/1. This range of
EDC in the bottoms corresponded to a TOC value on the order of 100 mg/1, a
BOD value in the 75 mg/1 range, COD value in the 150 to 200 mg/1 range and
a TOD value in the 100 to 200 mg/1 range.
As for wastewater from streams 221A and 011C, chloral was present in
stream 161A and was not steam stripped. The only explanation is that the
addition of the hydroxyl group enhanced the solubility .of this type
of compound. Steam stripping of this process effluent produces a waste-
water quite low in the measurable parameters of TOC and BOD,..
Steam Stripping Aromatic Contaminated Wastewaters for Stream 351A
Two primary process effluents are contaminated with the various organics
utilized in the manufacture of ethylbenzene: the overhead from the benzene
drying column and the wash water from the alkylation intermediate products.
In the production of styrene, sources of process effluent are the chilled
brine tank overflow and the vacuum jet collection water. The process
effluents from the ethylbenzene-manufacturing operations are combined at
battery limits and become the major process effluent from an ethylbenzene-
styrene plant. Other sources of wastewater are minor and intermittent.
133
-------
TABLE 89. TESTS RUN ON STREAM 161A
Run
Number
1
2
3
4
5
Feed
ml /min
276
255
235
245
243
Overhead
ml /min
7.8
13.5
11.4
5.3
9.4
Bottoms
ml /min
388
321
340
290
272
Reflux Steam
ml /min ml /min"
54
50
8 65
51
15.4
TOC (mg/1)
Feed Overhead
150
158
24
16
99
64
115
88
84
132
j
Bottoms
142
139
16
15
76
Mass
Balance
Efficiency
(%)
+19.8
+ 9.7
+17.1
- 0.8
—
Percent
TOC
Removal
1.2
3.9
17.8
11.4
5.2
-------
The stream characteristics of process effluent 351A are listed in
Table 90.
TABLE 90. STREAM CHARACTERIZATION OF STREAM 3514
Item Units Concentration
COD
TOG
pH
Chlorides
Benzene
Toluene
Ethylbenzene
Xylene
Styrene
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
239
49
9.4
315
5.7
1.2
8.5
0.6
0.7
Stream stripping of this wastewater should yield a wastewater with
only the inorganics and suspended solids as impurities. This was verified
with actual stripping tests. A number of runs were made with the aromatic
rich wastewater as feed. The results of these tests are summarized in
Table 91.
Discussion of Results—
Some variation in the feed composition, as measured by TOC, was noted
throughout the series of tests. At the higher feed contaminant loading,
the removal efficiency was better.
Sources of benzene could be a cause of these process contaminants.
Benzene is derived from two major raw material sources: by-product recovery
in the coking industry and solvent extraction of petroleum-based products.
Each source generates different trace contaminants in the final product.
The coke-based product usually includes xylenes, phenolics, and related
substituted compounds boiling in the same range. These compounds can be
present in the benzene feed stack in extremely small quantities and be
magnified several times in the final alkylate wash water.
Design Specifications for Commercial-Size Unit
A commercial-sized unit for installation at the plant site of stream
011C was designed. A site visit was made to get an overview of the physical
layout of the existing operations. Consultation with company participants
established design conditions.
To initiate the design, a complete material balance was established
for the process of each element of the influents and effluents: feed rate,
overhead rate, bottoms rate, and steam rate. All conditions for the final
commercial unit are given in Table 9.2. A material balance flow sheet is
Shown in figure 15 Process and instrument design conditions are included.
135
-------
TABLE 91. STEAM STRIPPING OF STREAM 351A
u>
Run
Number
1
2
3
4
5
6
7
8
9
Feed Overhead
ml/min
250
250
250
258
255
252
250
255
261
ml/min
10
13.8
3.0
11.5
13.4
6.5
8.2
7.5
14.0
Bottoms
ml/min
207.5
290.0
317.5
312.5
343.8
338
342
452
380
Reflux Steam
ml/min ml/min
54
63
59
15.0 53
6.7 36
— _
—
— _
— -
.7
.9
.7
.1
.7
-
-
-
-
TOG (mg/1)
Feed Overhead
315
2416
20.2
67.2
26.5
90.0
80.0
58.0
155.0
65.0
98
193
83.3
93.7
147
280
209
737
Bottoms
24
118
15
45
20
40
46
37
14
.0
.0
.4
.1
.8
.0
.0
.0
.0
Mass
Balance
Efficiency
+4.2
-3.2
+3.5
+4.2
+18.3
—
—
—
—
Percent
TOC
Removal
92.4
—
23.8
32.9
21.5
55.5
79.4
36.2
98.1
-------
Component kg/yr
Water 385,900
Recoverable
organics 385,900
U)
Component
Water
Organics
Solvent
kg/yr
Component
Solvent
Organics
Water
kg/yr
38,582,282
378,-182
2,270
O
to
J-l
w
383,630
7,718
9,080
\f
Component kg/yr
Organics
Water
378,182
229,270
5-1
-------
TABLE 92. DESIGN PARAMETERS COMMERCIAL UNIT FOR STREAM 011C
i
Feed rate (1/m) 530
Feed rate (kg/hr) 32,000
Organics (kg/hr) 135
Overhead product (kg/hr) - 5% of Feed 1,600
Feed Preheat (°C) 93
Packing ring diameter (cm) (Pall Rings) 4
Stripper bottoms (kg/hr) 32,500
Stripper steam (kg/hr) 2,300
Ethylene dichloride in bottoms (kg/hr) 0.45
Ethylene dichloride in bottoms (mg/1) 13
Economic Evaluation of Steam Stripper Operation with Recycle
of Recovered Organics
The following criteria were set as a basis for economic evaluation of
steam stripping of a halogenated hydrocarbon contaminated wastewater from
an economic standpoint.
1. 331 days/yr and/or 8,000 hr/yr.
2. Straight-line 10-yr depreciation.1
3. Interest on capital investment at 8 percent with 10-yr payout.
4. 532 1/min process condensate treatment.
5. Overhead rate based on 5 percent of feed.
Capital Cost—
Conventional techniques of major equipment pricing and factoring were
employed to arrive at a capital cost estimate of installing the equipment.
Operational costs were based on the material balance and applied economic
factors appropriate for this facility.
Equipment Cost Estimate
Equipment Cost
1. Feed storage tank $ 8,000
I 2. Stripping column (packed) 28,000
; 3. Feed pre-heater - 90 m
@ $380/m 2 35,000
4. Condenser ^ 18 m
• @ $380/m 7,000
5. Separator drum 3,500
6. Pumps (feed, oil layer, aqueous layer) 7,800
' 7. Assorted piping 10,000
Total $99,000
Total Installed Cost $250,000
138
-------
Operating Cost
Cost/Year
Steam Consumption ($6.60/1000 kg
-------
Cost Per Year
Variable Costs (operations) $205,000
Fixed Cost (depreciation, etc.) 66,000
Total $271,000
Recovered Credit $177,000
Total Annual Cost $ 94,000
Cost Per Liter $ 0.00041
of Water Treated
The amount of ethylene dichloride which would generate 532 1/min of
process effluent would have an annual capacity of approximately 227,000,000
kg/yr. The above cost analysis indicates that the cost of EDC produced
would be approximately 0.057
-------
SECTION 9
OZONATION
INTRODUCTION AND PRINCIPLES OF OZONATION
Chemical oxidation has the potential for removing from wastewater
organic materials which are resistant to other treatment methods. In many
cases, these refractory materials are toxic to biological systems. The
desirable characteristics of a chemical oxidant are that it be easily and
economically available and that it not contribute secondary pollutants to
the waste stream. Ideally, the oxidative process will result in the forma-
tion of nonpolluting gaseous products; carbon dioxide, water, etc. Practi-
cally, complete oxidation is difficult to achieve for most systems; the cost
of treatment increases exponentially as complete oxidation is approached.
In most cases, fragments of the original species remain and must not contri-
bute to the waste stream contaminant load or must be amenable to further
treatment.
Ozone is one of the strongest and purest oxidants available. Unlike
other oxidants, it can be generated as needed on-site. Ozone has been used
for years to purify, deodorize, and disinfect drinking water in Europe, and
has more recently Been used in the wastetreatment area to oxidize cyanide
and phenolic wa,stewaters. Two types of reactions can occur with ozone as
an oxidant. One is an electrophilic attack by ozone and the other an
ozone-initiated oxidation in which ozone serves as the reaction initiator
and oxygen is the principal reactant.
In a gas-liquid reaction such as the ozonation of industrial wastewaters,
one or more of the reactants must pass through an interface between the
phases to react. The overall relationship between mass transfer and the
reaction kinetics must be studied to identify the controlling regime. Once
the proper regime is identified, an appropriate reactor may be specified.
Reactor types range from a bubble column for use with a slow reaction con-
trolling to a packed column for mass transfer controlled reactions. A
stirred tank reactor was chosen for this study as it can simulate a bubble
column and approach the interfacial areas of a packed column by varying the
agitation,
OBJECTIVES '
The major objective of the ozonation study was to determine if ozonation
would be feasible as a pretreatment for biorefractory wastes prior to bio-
oxidation. (The high costs of ozone generation would probably make complete
141
-------
oxidation impractical.) Ozonation was performed on toluene diisocyanate
(TDI), ethylene glycol, styrene monomer, and ethylene dichloride wastewaters.
The effects of solution pH, temperature, and ozone mass transfer on the
oxidation rate were evaluated in a stirred tank reactor. The reactor was
operated in the semi-batch mode, continuously sparging gas into a constant
volume of wastewater.
MATHEMATICAL MODEL DEVELOPMENT
A mathematical model was developed and used with experimental data to
determine the reaction rate constants, volumetric mass- transfer coefficient,
and the proper reaction regime for the ozone-styrene gas-liquid reaction.
The model accounts for mass transfer of ozone into the liquid phase from
the gas phase, ozone decomposition, and reaction with styrene and its
reaction products. The model was developed from a macroscopic view of the
reaction system promoted by Prengle and Barona (4) . This procedure is
based on a combination of Pick's law of diffusion and a material balance
for each component in both phases over the entire reactor system. This can
be expressed as
dC
where C . is the concentration of component j , C . is the concentration of
component j at equilibrium, It. is the liquid-phase mass-transfer coefficient,
a is the interfacial area, ana r. is the reaction rate of component j.
The macroscopic view is illustrated in Figure 1 with a plot of the
ozone concentration in the liquid phase as a function of time. In the
absence of a liquid reactant, the ozone is absorbed exponentially to its
solubility limit. This concentration profile of ozone in the liquid phase
can be found by equating r. to zero and solving Equation (1):
CB * CBe a-* ') (2)
where C is the ozone concentration in the liquid phase, C is the equilib-
rium ozone concentration in the liquid phase, and is the combination of
the volumetric mass-transfer coefficient and the ozone decomposition rate
constant (5) .
The ozone concentration profile is shown for the three most frequent
gas-liquid reaction regimes in Figure 16. After an injection of liquid
reactant at time zero, the ozone concentration immediately drops to zero if
the reaction £s instantaneous. The result is a truly mass transfer limited
system as depicted by line 1. A quasi-equilibrium concentration is reached
if mass transfer is controlling with a slow reaction (line 2) . The ozone
concentration will remain at the solubility limit if the system is reaction
rate controlled (line 31. The reaction regime was identified for each
compound ozonated By conducting a material balance on each component and
comparing the data to the theoretical plot.
142
-------
Equilibrium Concentration
Be
No Reaction
1-e
4>t
Be
\
\
\\
\ \
\ \
\ \
\
\
M
\
1
\
r Reaction Time, t
1 - Mass Transfer Controlling - Instantaneous Reaction
i - Mass Transfer Controlling - Slow Reaction
3 - Chemical Reaction Controlling - Very Slow Reaction
4 - After Liquid Reactant is Depleted
Figure 16. Liquid phase concentration of the gas-phase reactant
as presented by the macroscopic view of interphase
mass transfer with chemical reaction for semi-batch
operation. Probable reaction regimes are illustrated.
Previous studies have shown that ozone oxidizes styrene to benzaldehyde;
which is easily oxidized to benzoic acid (6,7). The ozone attack on benzoic
acid should be directed to the meta positions, but actually occurs to some
extent at all positions, including the carboxylic acid group. The study (5)
was simplified by not extending the analytical work beyond benzoic acid in
search of the various reaction products. The chemical reactions studied
were:
0
II
HCH
20,
20,
(3)
(4)
143
-------
+ 0
2 (5)
0
II
CQH
Reaction Products
The formaldehyde reaction [Reaction (4)] is instantaneous and was incorpo-
rated into the initial styrene ozonation reaction.
The successive ozone attack on styrene and its reaction products is
similar to certain important industrial reactions, such as the successive
substitutive halogenation or nitration of hydrocarbons, and the addition of
alkene oxides to proton donor compounds such as amines, alcohols, water,
and hydrazine. A general representation of these reactions including the
ozonation of styrene is written:
A + b.B p* R (7)
1 kx
R + b^ _> s (8)
S + b.B T—» T (9)
where A is liquid reactant A, B is ozone, b is stoichiometric mole ratio of
moles of ozone per mole of liquid reactant, and R, S, T are liquid reactants,
and k., k_, k_ are reaction rate constants. This reaction mechanism is
considered to be consecutive with respect to compounds A, R, S, and T and
parallel with respect to compound B, ozone.
The rate expressions for the consecutive-parallel reactions, assuming
the reactions are irreversible, bimolecular, and of constant density, are
dC
Stryene rA = — = -k^ (10)
dCR
Benzaldehyde rR = — . k^ - k^ (11)
dC
Benzoic Acid rg - -^ = k^ - k^ (12)
Levenspiel (8) has discussed the consecutive-parallel reaction type and
presents a general solution to the rate equations.
Mass Transfer Model
The mass-transfer equations were extended to include the chemical
reactions by inserting the rate equations (10), (11), and (12) for the
reaction term in equation (1). The following assumptions were made during
the development to simplify the equations:
144
-------
1. Well-mixed semibatch stirred tank reactor operation of continuously
sparged gas into a constant volume of liquid.
2. Constant temperature and pressure.
3. Constant volume and liquid density.
4. Constant molar gas feed rate.
5 . Constant gas holdup .
6. Ozone decompose via first order reaction.
7. Liquid reactants and products are nonvolatile.
The model was developed for the probable reaction regimes of liquid-
phase mass transfer controlling with an instantaneous reaction and a slow
reaction either mass transfer controlled or reaction rate controlled. The
deciding factor is whether or not ozone penetrates the bulk liquid. If
ozone does penetrate the bulk liquid, and if the reaction is extremely slow
and ozone reaches an equilibrium concentration in the liquid, the reaction
is rate controlled. Both models were applied to the experimental data to
determine the controlling regime.
The initial model was developed for mass transfer controlling with an
instantaneous reaction for the single reaction given in Equation (7). The
liquid reactant A represents the initial reactant which can be styrene,
benzaldehyde, or benzoic acid. The ozone does not penetrate into the bulk
liquid for a truly mass- transfer limiting system. Thus, the molar balances
are simplified. The rate of disappearance of A can be found from the
stoichiometry of the reaction and the limiting form of the ozone balance to
give a solution in terms of the ozone transferred from the gas phase:
<13)
Ao
Deviation of experimental data from the line projected by equation (13)
indicates a slow reaction controlling. The mass-transfer limiting line was
calculated for each test.
If the ozonation reaction is slow, either mass transfer controlling
with a slow reaction or reaction rate controlling, the ozone will penetrate
into the bulk liquid. In the latter case, the ozone reaches its equilibrium
value as predicted by Henry's law. For the model development, an average
ozone Bulk concentration, x , was used. Allowing xg to be a constant
transforms the reaction ints the form of a pseudo first order reaction of
which the solution is
N ' (14)
N.
e
145
-------
where :
L = total moles of liquid.
The model was extended for the second reaction involving the reaction
product R. Substituting equation (14) into the rate equation for R and
integrating, the solution for the moles of R at any instant is expressed as
- -^- (e-V - e-V> (15)
Ao !.
These models were applied to the experimental data obtained in the styrene
ozonation study.
LABORATORY STUDY
A stirred tank reactor was selected for the ozonation study. The
reactor is capable of bringing large quantities of ozone into contact with a
specified volume of wastewater. The stirred tank reactor was operated in the
semibatch mode, and ozone gas was fed continually to a constant volume of
wastewater. The reactor was designed to comply with the standard stirred
tank reactor configuration used by Westerterp and others in mass transfer
studies (9, 10, 11). Prengle used a similar process for the ozonation of five
particular refractory compounds (12). The major dimensions of the reactor are
ratioed to the tank diameter for easy design and scaleup. The plexiglass
reactor used for this study has a diameter of 39.2 cm and dimensions corre-
sponding to the standard configuration. A volume of approximately 20 liters
was achieved with a liquid height of 39.2 cm.
A schematic of the ozone pilot plant is shown in Figure 2. Air was used
as a feed gas to the Ozone Research and Equipment Company (OREC) Model 03B4-
AR ozonator. The ozonator has a rated capacity of 0.8 kg of ozone per day.
The ozone flow to the reactor was controlled by a needle valve. The gas flow
was measured with a rotameter, and excess gas was vented /through a potassium
iodide trap to the atmosphere.
The standard potassium iodide method (13) was used to analyze the ozone
concentration in the feed and exit gases and in the liquid. The temperature
was monitored periodically. An operational immersion heater and temperature
controller were installed to operate at 15° and 25PC.
The gas feed rate was 11.5 1/min, and the feed contained 12-15 mg/1
ozone CL.Q to 1.2 wtg). A turbine speed of 700 rpm provided a sufficient
yolumetric mass-transfer coefficient to consume all of the ozone if the
reaction was instantaneous (5J. The average test was continued for 3 hr.
Liquid samples were analyzed for total organic carbon with a Beckman 915 TOC
analyzer and for composition with a gas chromatograph.
146
-------
MASS-TRANSFER STUDY
Mass transfer of ozone into distilled water and pulse tests were used
to study the model reliability and mixing characteristics of the reactor.
A standard stirred tank reactor configuration was chosen to approach ideal
backmix flow. Deviation from the ideal was determined by pulse testing as
described by Levenspiel (8). Standard operating conditions were used
during-the reaction study; turbine speed was 700 rpm and gas-feed rate was
0.01 m /min. A liquid feed rate of 0.003 m /min gave a residence time of
8.17 min. Sodium chloride solution was used as a tracer; chloride ion
concentration in the exit stream was measured for use in a C curve distrib-
ution test.
The C curve described the concentration-time function of the tracer in
the exit stream of the reactor in response to an idealized pulse tracer
injection. The experimental data for the pulse test, shown as the dotted
line in Figure 17, were compared to that plotted in the solid line repre-
senting idealized backmix flow in a stirred tank reactor. The experimental
data approach the ideal with less than 5 percent error. Though real reactors
are never ideal, the reactor used was considered ideal with negligible
error.
1.0
Turbine Speed = 700 rpm
Gas Feed Rate = 0.41 cfm
Liquid Feed Rate =0.10 cfm
Residence Time =8.17 min
100 ppm Cl pulse
Ideal Stirred Tank
Reactor (Backmix Flow)
— — Experimental Pulse Test
Figure 17.
1.0 2.0
Reduced Time/Residence Time
Comparison of ideal and experimental C-curves for a
pulse tracer input for a stirred tank reactor.
The effects of temperature, power input, and gas velocity on the
volumetric mass-transfer coefficient were evaluated. A total of 83 mass-
transfer tests were conducted with various temperatures, turbine speeds,
147
-------
and gas feed rates. The tests were made with the water either chilled
between 12° and 15°C or at ambient temperatures of 22° to 25°C. The following
turbine speeds were studied: 0, 300, 400, 500, 700, and 900 rpm. These
values were transformed into power inputs of 0, 0.18, 0.40, 0.87, 2,4, and
5.3 hp per 1000 liters, respectively. The gas feed rates were varied
between 0.004, 0.006, 0.008, and 0.011 m /min.
The mass-transfer data were evaluated by plotting the ozone concentra-
tion profiles obtained for the various operating parameters. Figure 18 is
a plot of the ozone concentration profile for the various turbine speeds at
a constant gas feed rate of 0.01 m /min and a temperature range of 12° to
15°C. The data showed that the ozone concentration in the liquid increased
exponentially to the solubility limit.
A constant gas concentration was not maintained for the mass-transfer
studies, giving different equilibrium values for ozone in the liquid for
the various tests. For example, test 14 was performed with an ozone gas
concentration of 1.24 wt percent, giving an equilibrium value of 6.96 mg/1.
Test 3 had a gas concentration of 1.10 wt percent ozone and a corresponding
equilibrium value of 5.75 mg/1. The relative differences are shown in
Figure 18.
The data in Figure 18 show that the ozone absorption rate into the
water was slow without agitation and increased as agitation was applied.
The increase in the ozone absorption line slope with the increased turbine
speed was a result of shearing the gas bubbles by the turbine blades,
thereby increasing the interfacial area. The slopes of the absorption
lines above a turbine speed 400 rpm are similar, indicating the system was
operating in the well-mixed region between 400 and 900 rpm.
A similar set of figures was generated for the various turbine speeds
at constant gas feed rates for the temperature range of 22° to 25 °C.
Figure 19 is for a gas feed rate of 0.01 m /min and is typical of data
obtained at other feed rates. The ozone concentration profiles in Figure
19 are almost duplicates of those in Figure 18, except the ozone equilibrium
concentration was lower at the higher operating temperatures. Again, the
equilibrium concentration levels were dependent on the ozone gas concentra-
tions. The well-mixed region was achieved above a turbine speed of 400
rpm. Careful evaluation of the mass-transfer data showed the small increase
in temperature did not significantly change the volumetric mass-transfer
coefficient. The data were replotted for the various gas feed rates,
holding the turbine speed constant to evaluate the effect of gas feed rate
on the ozone mass transfer. Figure 20 is for a turbine speed of 700 rpm
and a temperature range of 12° to 15°C. The absorption rate of ozone
increased as the gas feed rate was increased.
The volumetric mass-transfer coefficient, k^a, was calculated using
the slope of the absorption lines and the relationship.
V - $
<«>
148
-------
h-
-P-
vo
0
cx
ex
c
o
•H
Ci
0)
o
c
o
o
0)
c
o
N
o
Figure 18.
10 15 20
Time (min)
25
Ozone concentration profile for
various turbine speeds at 12°-15°C.
Figure 19.
0-A
Mass transfer only
Gas feed rate = 0.41-
ft3/min;
0.56 ft/min
RPM Run
D-500 (13)
0-300 (23) A-700 (11)
Q-400 (75) x-900 (16)
10 15 20
Time (min)
25
Ozone concentration profile for
various turbine speeds at 22°-25°C.
-------
Ul
O
I
a
o
§
O
c
o
o
c
o
N
O
Mass Transfer Only
Turbine Speed = 700rpm
Gas Feed Rate Run
DO. 15 cfm (55)
Q0.23 cfm (24)
A 0.31 cf_ (39)
A 0.41 cfm (1)
i i i
50
Figure 20.
10 15 20
Time (min)
Ozone concentration profile
for various gas feed rates at
12°-15eC.
•rl
0
CO
0)
rH
i
10
I ' '"I
I
,|
0.5 1 5 10
Power input (hp/1000 gal)
50
Figure 21. Volumetric mass-transfer coefficient
at various gas feed rates as a
function of power input.
-------
Correlations were made between the volumetric mass-transfer coefficient,
the power input from the turbine, and the gas feed rate. The volumetric
mass-transfer coefficient was plotted on log-log paper as a function of the
power input with the various gas feed rates as parameters (Fig. 21) and
shows the rapid increase in k a as agitation was applied to the system. A
slower increase was achieved in the well-mixed region. This figure is
similar to that published by Oldshue (15) for oxygen transfer into active
milk sludge. The slope of Oldshue' s data in the well-mixed region was
0.48, while the slopes in Figure 20 varied from 0.31 to 0.49, with an
average value of 0.43. Kawecki (10) reported similar curves for the vol-
umetric mass- transfer coefficient and power input.
Cooper, et^ al. (9) investigated the absorption of oxygen from air into
aqueous sodium sulfite solutions. Cooper found the following correlations
between the mass-transfer coefficient and the power input and superficial
gas velocity.
V = ml VG°'6?
kja = m2 (P/V)0'5 (18)
n 67
Cooper's equations suggest a log-log plot of k^a/V * versus (P/V) would
give a straight line with a slope of 0.5. The mass-transfer data were
plotted in such a ^manner in Figure 22 to produce a curve similar to Cooper's.
The slope of the line was 1.15 in the lower power input region and 0.46 in
the well-mixed region. (Cooper reported a maximum slope of 0.98 in his
data). The correlation given by eq. (17) was tested by plotting volumetric
mass-transfer coefficient as a function of the superficial gas velocity
with the turbine speed as a parameter on log-log paper. In the absence of
agitation, a straight-line relationship with a slope of 0.66 was obtained
(Fig. 23). The data show a similar slope at lower superficial gas velocities,
but a rapid increase as the superficial gas velocity and power input were
intensified. The slope varied from 0.98 to 2.10. Yashida (18) found his
mass-transfer data to correlate well with superficial gas velocity raised
to the 0.74 power. Both Cooper and Yashida studied lower superficial gas
velocities than used in the tests. Foust (19) reported a slope of 2.13 for
a superficial gas velocity range of 0.3 to 1.5 m/min and similar power
inputs. The various correlations reported between the volumetric mass-
transfer corfficient and the superficial gas velocity proves there is a
transition range where the exponent varies significantly. Both Calderbank (20)
and Fair (21) report a transitional range in the volumetric mass-transfer
coefficient using bubble diameter instead of superficial gas velocity as a
parameter .
The volumetric mass-transfer coefficient was evaluated by Majumoar
in a bubble column to be 16.18 Kg mole/m min. A value of 9.6 Kg mole/m
min was found under similar operating conditions using Figure 23. The fact
that Majumoar's mass- transfer model did not include a reaction term for the
decomposition of ozone explains the slight deviation between the two values.
The excellent correlation between mass-transfer data obtained using this
model and published data supports the accuracy of the mass-transfer data
and the model developed.
151
-------
Ul
1-0
40
CS|
m
I
>
td
10
' I ' '"I
I I
I
, I , ...I
0.5 1 5 10
Power Input (hp/1000 gallons)
50
Figure 22. Volumetric mass transfer coefficient corrected for
superficial gas velocity as a function of power
input.
100
e
CO
Cfl
-------
PILOT STUDIES
An ozone pilot plant was designed and constructed for the proiect. It
was housed in a portable 2.4 x 3 m building with a major equipment arranged
as shown in Figure 24. The pilot unit was equipped with titrimetric apparatus
for the determination of ozone concentration in the gas and liquid phases.
CD
Ozone Chiller
Stirred-Tank
Reactor
Ozonator
Recorder
Ozone
taalyzer
©
Cooling Water
Pump
Cooling Water
Reservoir
Rotameters
Pump
Zero Grade Air Bottle
Distilled WateY
Work Bench
Q
0000
Ethylene Bottle
•2.4M-
•
c
c
2.4M
3M
-H
Figure 24. Equipment arrangement in ozone pilot plant building.
Ozone is a toxic gas with a threshold limit value (TLV) of 0.1 mg/1.
Masks equipped with activated charcoal filters were used for protection.
Draeger tubes were used to detect ozone concentrations in the work area,
while a Beckman continuous ozone analyzer monitored ozone levels in the
pilot plant building. The analyzer concentration range was 0.1 to 10 mg/1
ozone in air. The sensitivity of the Draeger tubes was 0.05 to 300 mg/1.
Explosion-proof electrical outlets and switch boxes were used. The building
was designed with two 120 x 240 cm flaps and a fan to promote ventilation
through the building.
153
-------
Equipment
An Ozone Research and Equipment Company (OREC) model 0384-AR ozone
generator was used to produce ozone. The ozonator contained five dielectrics
and had a rated capacity of 1 kg of ozone per day. The ozonator included
an air compressor and a chiller.
Reactor Systems—
A tubular reactor was constructed to simulate pipeline flow and provide
an easy adaptation to plant use (Fig. 25). The reactor was made of 18 mm
polypropylene tubing. The ozone gas and wastewater were dispersed with a
nozzle or mixed in a standard tee, with gas-liquid contact being induced by
static mixers. A Masterflex variable speed pump was used to vary the
wastewater flow rate from 1 to 4 liters/min. Rotameters were used to
measure gas and liquid flow rates; excess gas was vented to the atmosphere.
Sample valves were placed at the reactor midpoint and at the end of the 70-m
reactor.
The tubular reactor was used only on the toluene diisocyanate waste-
stream. Analysis of the reaction mechanism showed that 0.5 m of 1.5 wt
percent ozone gas would be needed to treat 4 liters of wastewater. The
volume of air introduced with the ozone makes a tubular reactor impractical
except for low TOG wastestreams.
A schematic of the stirred tank reactor system, including the ozone
generator and distributing network is shown in Figure 26. The reactor was
operated in the semibatch mode to contact large quantities of ozone to a
specified volume of water. The reactor could easily be converted to a
continuous system with the addition of a liquid feed pump. The reactor was
designed to comply with the standard stirred tank reactor configuration
used by Westerterp and others in mass-transfer studies (9-11). Prengle
used a similar process for several ozonation studies (12, 23, 25).
The major dimensions of the reactor were ratioed to the tank diameter
for easy design and scaleup. The Plexiglas reactor had a diameter of 30 cm
and dimensions corresponding to the standard configuration shown in Figure
27. A liquid volume of 20 liters was achieved with the proper ratioed
liquid height of 30 cm. The reactor was well insulated and wrapped with
cooling coils to maintain a constant temperature.
A rotameter was used to measure the gas flow to the reactor. Needle
valves were used to control the gas flow, with excess gas vented to the
atmosphere through a potassium iodide trap to deplete the ozone. Gas
sample ports were located on the ozonator and on the gas exit line from the
reactor. Liquid samples were withdrawn from either the side liquid sample
points or through the bottom sample port.
An optional immersion heater and temperature controller were available
for operating at temperatures between 24° and 121°C. Actual tests were
conducted at 15s and 25°C utilizing ice water circulating through cooling
coils to maintain an isothermal operation. Bayonette thermometers were
used to measure operating temperatures. Various degrees of mixing were
achieved with the variaBle speed mixer.
154
-------
Wastewater
Feed Tank
Vent
Midpoint Sample Valve
Nozzle and/or
Static Mixers
IRotametersI
Vent
4
IJ_
Tubing Collection
Barrel
Variable
Speed Pump
Air
Ozone
Generator
Figure 25. Tubular reactor schematic.
Variable Speed
Mixer
Exit Gas Vent
f"—*Exit Gas Sample Port
I*! Rotameter
«O Pressure Gauge
o
0)
" •
II
i V i o
o
1 « 1 1 i (
f^^^^^^^f ^^^^^^^
1+ 1
Sample
Ports
1
j ,
Rotameter
Vent
h
iquid
Ozone
Generator
Air
Sample Port
Figure 26. Ozone pilot plant schematic.
155
-------
Gas Out
Top
b
h 1
Sparger
H
™13fcsJ
Baffles
Turbine
Impeller
(4 or 6 blad
f'-B
a n _^
1
1 0 £ | 1 1 1 J •
f "
._ _
_
wb
ss)
LiiquiQ
Level
1
Hl
|
*L
1^—
Gas
Drain
I
!
= DT,
DT/3,H1 = T>
2
Dimensions:
Sparger Holds: 3-1/16" holes/in of sparger cross-section
Impellers: b = D.4, h
Figure 27. Standard reactor configuration.
Experimental Procedures
A gas ehromatograph was used to analyze the styrene and benzaldehyde
concentrations. The benzoic acid concentration was measured by monitoring
the ultraviolet light absorption with a spectrophotometer. The total
organic carbon was used to monitor the reaction of the industrial process
wastestreams and was measured with a Beckman TOG analyzer. The standard
potassium iodide method was used to analyze the ozone concentration in the
gas and was adjusted to analyze ozone concentration in the liquid. The
turbine speed was monitored with a phototachometer.
After ozonation, biological batch tests were performed to determine
the enhancement of ozonation on wastewater biodegradability. An acclimated
seed was added to each of the test reactors. Various amounts of wastestream
were. added to the test reactors to obtain a series of organic loadings.
One reactor containing only acclimated seed was used as a baseline. Data
were collected on the reactors during a 3-day test period.
156
-------
The tests were conducted In two separate phases. First, mass-transfer
data were collected on ozone transferring into the distilled water. The
second phase was the gas-liquid reaction study. Toluene diisocyanate
wastewater was studied initially, followed by the the other industrial
wastes. Benzoic acid was the first organic ozonated in aqueous solution
followed by benzaldehyde and styrene. The individual reactions were studied
to evaluate the rate constants for the continuous-parallel reaction system.
Mass-transfer tests were conducted with ozone transferring into distil-
led water. The water was placed in the reactor to give a height-to-diameter
ratio of one. The desired operating temperature was obtained before setting
the turbine speed. Once the turbine speed was established, ozone was
introduced into the bottom of the reactor through the gas sparger. Final
adjustments were made to obtain the desired gas feed rate. The reactor was
operated isothermally during the test and was sampled and analyzed routinely
for ozone content.
The reaction procedure began with the preparation of 20 liters of test
solution with the desired concentration of liquid reactant. The solution
was introduced into the reactor and the pH and temperature were adjusted.
Concentrated sulfuric acid and a sodium hydroxide-sodium carbonate buffering
solution were used to adjust the pH to values of 2 and 11, respectively.
After establishing a turbine speed of 700 rpm, an initial baseline liquid
sample was taken. The reaction was initiated with the addition of ozone
gas. Final adjustments were made to obtain the desired gas feed rate.
Ozone analyses were performed on the liquid and the feed and exit gases
during the test. Liquid samples were taken at pre-determined intervals
through either the side or bottom sample port. Temperature and pH were
recorded with each liquid sample which was sealed in an air-tight sample
vial.
The standard operating parameters chosen for the reaction tests were a
gas feed rate of 11.5 liters/min containing 12-15 mg/liter ozone (1.0 to
1.2 wt%). A turbine speed of 700 rpm provided a sufficient volumetric
mass-transfer coefficient to consume all the ozone for a mass-transfer
controlling.regime with an instantaneous reaction. The tests were conducted
isothermally for 150 min.
Ozonation Reaction Study—
The wastewaters studied were taken from the four processes listed in
Table 93. The streams were not analyzed for individual components, but
were characterized for the significant properties listed in Table 94. The
wastewaters were ozonated at their normal TOC concentrations as received
and after dilution to 100 to 150 mg/1 TOC. .The dilution tests were performed
to reduce the ozbnation time. The ozone studies were conducted with a gas
feed rate of 11.5 liters/min and a turbine speed of 700 rpm. The ozone
transferred into the water and reacted with the organics was calculated
from a material Balance across the reactor. The data obtained in the study
were plotted as the function TOC remaining (TOC/TOC^ versus time and
weight ratio of milligrams ozone transferred (reacted) per milligram TOC.
Extrapolation of the weight ratio line to zero TOC remaining gave the
weight of ozone required for complete oxidation. This was used to determine
157
-------
if the reaction was mass-transfer limited. Deviation from the mass-transfer
limiting line indicated the reaction was ireaction-rate controlled.
t
TABLE 93. PROCESS WASTESTREAMSiUSED FOR THE OZONATION STUDY WITH THEIR
RANGE OF TOG CONCENTRATIONS FOUND
, TOC Range
Major Contaminant (mg/liter)
Toluene Diisocyanate - TDI 250-3500
Ethylene Glycol - Polyol 150-1500
Styrene 30-200
Ethylene Dichloride - EDC (Steam Stripped) 300-1500
TDI Wastewater—
The first wastewater ozonated was a wash stream from a toluene diisocya-
nate (TDI) process. In the process, phosgene was reacted with toluene
diamine (TD) to product TDI, which is used in the manufacture of polyurethane.
The TDI wastewater characterization is given in Table 94. The water is
believed to be contaminated with toluene, amines, and TDA.
The tubular reactor system was tested with the toluene diisocyanate
wastewater. Complete dispersion of the gas and liquid was achieved with a
nozzle, but at a liquid feed rate above 2 liters/min, the nozzle forced
liquid into the ozone gas supply line due to high pack back pressure. The
ozone generator was operated at 1.3 atm pressure compared to the optimum of
1.02 atm to force ozone into the liquid. The gas and liquid separated into
distinct layers 1.5 m downstream of the nozzle. The gas-liquid flow con-
figuration was accurately predicted by the Baker and Govier charts, which
describe gas-liquid flow in horizontal and verticle pipes.
TABLE 94. CHARACTERIZATION OF INDUSTRIAL WASTEWATERS OZONATED
(Averaged Data; Concentrations in mg/liter)
Compound TDI Polyol Styrene EDC
PH
TOC
COD
TOD
Phenol
Cl"
Acidity*
Alkalinity*
Suspended .Solids
Volatile
Suspended Solids
Residual Chlorine
11.6
687
663
439
6.6
2500
—
2408
11.83
3.07
—
11.4
476
1509
1065
—
70 '
—
240
4.80
1.40
1
11.8
85
225
227
1.9
3100
—
2733 ,
3.85
1.04
—
1.3
866
1309
1940
—
6800
8233
—
7.97
, 0.86
, 65.25
* Measured as CaCO with phenolphthalein end point.
158
-------
The residence time in the tubular reactor was controlled by the gas
flow and was recorded by visual timing of the liquid through the reactor.
A constant liquid feed rate of 2 liters/min was chosen, and the gas feed
rate was changed to vary residence times. Samples were taken at both the
midpoint and end sample point to give two residence times per run. All
samples were collected after steady-state operation was achieved. The feed
gas and exit gas were sampled routinely for ozone concentration. On a
majority of the runs, ozone was not detected in the exit gas. A summary of
the ozonation runs with the nozzle is given in Table 95. The mole ratio
was calculated on the basis of the TOC being pure TDA. Since the data are
fairly scattered, the area available for mass-transfer is reduced.
TABLE 95. OZONATION OF TOLUENE DIISOCYANATE WASTEWATER IN A TUBULAR
REACTOR WITH NOZZLE DISPERSION
Liquid Flow =1.75 1/min Original TOC = 560 mg/1
Gas Flow Residence Time TOC Reduction
(1/min) (min) (mg/1) (%)
pH =
PH =
pH =
pH =
11
3.5
3.5
6.0
6.0
8
8
8
4
4
8
8
1
4
4
8
8
6
8
8
1.8
3.7
1.3
2.6
1.0
2.0
1.7
3.4
1.0
2.0
1.7
3.4
1.0
2.0 1
1.0
2.0
586
561
528
549
520
512
891
544
491
481
i
' 538
530
527
541
663
538 |
0
0
5.7
2.0
6.2
8.6
12.3
2.9
12.3
14.1
4.0
5.4
5.9
2.0
0
5.0
Mole Ratio
0.059
0.059
0.102
0.102
0.127
0.127
0.064
0.064
0.127
0.127
0.068
0.068
0.135
0.135
0.135
0.135
Static mixers were added to promote) gas-liquid contact. The pressure
drop was greatly reduced with tl|e static mixers and worked well in their
immediate area. A liquid flow rate of 1.5 1/min was used, with higher gas
flows used to increase the ozone to TDA mole ratio. Table 96 lists the TOC
reductions for the tubular reactor runs using the static mixers. The TOC
reductions were low, with the largest value being 15%. The tubular reactor
system was abandoned, because of the. inefficiency of mixing the gas and the
liquid. I '
159
-------
TABLE 96. OZONATION OF TOLUENE DIISOCYANATE WASTEWATER IN A TUBULAR
REACTOR WITH STATIC MIXERS
Gas Flow
(1/min)
Liquid Flow =1.5
Residence
(min)
1/min
Time
Original
TOC
(mg/1)
TOC = 1068 mg/1
Reduction
(%)
Mole Ratio
pH
10
10
24
24
26
26
1.5
3.0
1.0
2.0
0.7
1.4
970
938
965
933
965
965
9.2
12.2
9.6
12.6
9.6
9.6
0.176
0.176
0.424
0.424
0.451
0.459
pH = 8
10
10
20
20
1.5
3.0
0.8
1.6
1118
1046
946
1026
0
2.1
9.7
3.9
0.220
0.200
0.396
0.396
BOD to TOC Ratios
Sample
Unreacted TDI
0.200 moles 0 /mole
0.396 moles 0^/mole
BOD
(mg/1)
1161
TDI 1341
TDI 1251
TOC
(mg/1)
1068
1048
964
Ratio
1.087
1.239
1.257
The stirred tank reactor was then used to study the ozonation reactions.
The effects of pH, temperature, and ferrous sulfate catalyst were studied
and are shown in Figure 28. The curvature in the TOC line is a result
of the reaction of TDA and amines. Ozone will oxidize an amine group to
the nitro group without a TOC decrease since no organic carbon is released.
BOD,, and biological batch tests were performed because nitro groups are
easier to biodegrade than amines. Extrapolation of the plots indicates a
weight ratio of 7 mg 0 /mg TOC will completely oxidize the organic compounds.
The dark red TDI waste was decolorized to a faint yellow at a weight ratio
of two. The maximum TOC reduction achieved after 3 hr of ozonation was 26
percent. Figures 28 and 29 show that variation in pH, temperature, and
catalyst addition did not alter the reaction significantly. Tests with '
ferrous sulfate catalyst were performed in acidic solutions to prevent
precipitation of the iron. i [
A concentrated waste sample (TOC = 3360 mg/1) was ozonated for 6 hr
before being submitted to biological batch'tests. The TOC reduction was 15.6
percent. The organic loadings and summary data from the batch tests for the
raw TDI and ozonated TDI stream are listed in Table 97. Each reactor was
analyzed periodically for TOC, oxygen utilization rate, BOD and MLVSS over
160
-------
20
10
FeSO, Catalyst, 15°C
pH = 1
o .
§
•rl
4-1
u
"Cl
u
o
H
20
10
20
10
No Catalyst
1.0
0.9
o
H
g 0.7
H.
fi
•H
c
0.6
0.5
8 0.4
H
0.3
0.2
0.1
^
\
Mass Transfer-'*'"'^.
Limiting v
\
\
\
D pH=l
O PH=1
A pH=ll
Cat., 15°C
Cat., 25°C
\
\
\
Figure
0.25 0.50 0.75 1.00 1.25
Weight Ratio (mg 0.,/mg TOG)
28. Ozonation of toluene diisocyanate wastewater
as a function of ozone transferred
30 60 90 120 150 180 210
Time (min)
Figure 29. Ozonation of toluene diiso-
cyanate as a function of time.
-------
TABLE 97. SUMMARY OF BATCH STUDY RESULTS FOR RAW AND OZONATED TDI WASTEWATER
to
Reactor 1
Untreated TDI
F/M
TOG (mg/1)
Time
(hours) 0
0.05
243
TOG (% Reduction)
BOD_ (mg/1)
46
BOD5 (% Reduction)
Ozonated TDI
F/M
TOG (mg/1)
0.05
183
TOG (% Reduction)
BOD5 (mg/1)
39
BOD;? (% Reduction)
24
238
2.1
48*
0
149
18.6
60*
0
Reactor 2
0
0.10
370
50
48.0
0.10
274
125
24
347
6.2
26*
239
12.8
34*
72.8
Reactor 3
0
0.20
652
69
0
0.20
512
215
24
621
4.6
64*
445
19.1
120*
44.2
Reactor 4
0
0.40
1221
83
42.2
0.40
1035
306
24
Reactor 5
0
24
Airstripped
1118
8.4
48*
1136
109
31.2
1315
1.6
75*
Airstripped
962
7.1
86*
71.9
1003
465
1097
0
286*
38.5
Reading taken at 72 hr.
-------
a 3-day period. The BOD5 for the ozonated waste was reduced 30 percent
more than that for the untreated waste. The reductions were associated for
F/M ratios of 0.1 and 0.4. TOC reductions were also up to 4 times greater
for the ozonated waste.
Polyol Wastewater—
The polyol wastewater was taken from an ethylene glycol process plant.
The water was very basic (pH = 11.4) with a grayish tint. It was believed
to contain ethylene glycol and other hydroxy hydrocarbons. Ozonation data
plotted in Figure 30 show a weight ratio of 7.3 mg of ozone would be required
per mg of TOC for complete oxidation as compared to 10 mg for pure ethylene
glycol. The effect of pH is dramatically shown in Figure 31, as the TOC
reduction is negligible with low pH solutions. The polyol waste oxidation
is very pH dependent as is the phenol oxidation with ozone (24). The oxida-
tion proceeded slightly faster with a low operating temperature (15°C) but in
both cases the system is controlled by the reaction. The tapering effect
is a result of the formation of a refractory compound such as acetic. The
ozone concentration reached equilibrium after 150 min.
A batch study was performed on polyol waste which had an initial TOC
of 830 mg/liter. After 5 1/2 hr of ozonation, the TOC batch study consisted
of 3 test reactors at organic loadings of 0.05, 0.10, and 0.15 mg BOD_ per
mg MLVSS-day. A reactor to test air stripping and a baseline reactor
containing only microbial seed were included. Table 98 is a summary of the
test results. The raw polyol wastestream was resistant to biological
degradation; only 7 to 23 percent of the TOC was removed, yielding residual
levels of 142-301 mg/1 TOC. The initial BOD5 levels were low (16 to 28
mg/1) and showed no reduction during the study.
Pretreatment of the polyol wastewater with ozone had a dramatic effect
on the degradability of the waste. As seen in Table 98, 45 to 68 percent
of the TOC and 83 to 89 percent of the BOD were removed in 72 hr. Ozonation
formed volatile organics, as 58 and 72 percent reductions in TOC and BOD5
were achieved with air stripping, respectively.
The polyol oxidation with ozone is pH dependent and proceeds faster at
a pH above 10. A quick TOC reduction is achieved until a refractory compound
is produced to slow the reaction rate. The average TOC reduction within 3
hr of ozonation was 50 percent for an initial TOC of 100 mg/1. The effec-
tiveness of ozonation is clearly illustrated in Table 99; the TOC reduction
after 5-1/2 hr was only 25 percent but the BOD5 was increased from 91.3 to
613.5 mg/1 (671 percent). The ozonation improved the wastewater biodegrad-
ability and could effectively be used as a pretreatment to biooxidation.
However, the resistance of the refractory compound would decrease the feasi-
bility of using ozone for complete oxidation.
TABLE 99. COMPARISON OF RAW AND OZONATED POLYOL WASTEWATER
Pavr
Ozonated
TOC
Cng/U
830
626
BOD
(mg/I)
9J3.1
613.5
BOD5/TOC
0.11
0.98
163
-------
8
u
o
H
toO
C
•H
c
•H
«
e
0)
S-i
o
EH
G
O
O
ca
M
fa
pH=ll, 15°C -
A pH=ll.r
Weight ratio (mg 0 /mg TOC)
Figure 30. Ozonation of polyol wastewater as a
function of ozone transferred
0 30 60 90 120 150 180 210
Time (min)
Figure 31. Ozonation of polyol wastewater
as a function of time.
-------
TABLE 98. SUMMARY OF BATCH STUDY RESULTS FOR RAW AND OZONATED POLYOL WASTEWATER
Reactor 1
Time
Untreated TDI
F/M
TOG (mg/1)
TOG (% Reduction)
BOD, (mg/1)
BOD;: (% Reduction)
Ozonated TDI
F/M
TOG (mg/1)
TOG (% Reduction)
BOD- (mg/1)
BOD;? (% Reduction)
(hours) 0
0.05
162
16
0.05
238
138
72
142
12
20
0
130
45
23
83
Reactor 2
0
0.10
285
11
0.10
354
303
72
264
7
28
0
116
67
32
89
Reactor 3
0
0.15
390
11
0.15
404
395
72
Reactor 4
0
72
Airs tripped
301
23
31
0
385
51
398
0
31
39
Airstripped
130
68
60
87
363
485
151
58
120
75
Reactor 5
0 72
— (Baseline)
38 54
0
16 22
0
(Baseline)
160 135
16
27 20
26
-------
Styrene Wastewater—
Styrene is an important chemical in the production of plastics and
synthetic rubber. The monomer is usually produced by the dehydrogenation
of ethylbenzene. In the process there are two wastewater streams which are
combined in the plant effluent. The styrene wastewater had the lowest TOG
of the 4 waste streams sudied with an average value of 90 mg/1.
Figure 32, the weight-ratio plot, shows that all the oxidation tests
produced sharp decreases in TOG and then leveled off. This was a result of
the air stripping of the organics such as styrene from the water. The high
pH tests show a TOG reduction after air stripping; by contrast, tests with
acidic solutions show a horizontal reduction line. Extrapolation of the
basic data showed a weight ratio of 12.5 is required for complete oxidation
of the remaining organics as compared to theoretical values of 10 and 10.5
for styrene and ethylbenzene, respectively.
Figure 33 illustrates the effectiveness of air stripping. The air
stripped test produced a quicker TOG reduction than the ozonated tests.
The lower air stripping values are accounted for by the volatile reactants
that are oxidized to a soluble product which reenters the liquid phase.
Residual organics were biodegradable, and therefore, batch tests were not
performed on the ozonated waste. Continuous biotreatment tests were made
at both the laboratory and pilot plant levels. The TOG and BOD reductions
were significant in both tests and indicated that ozone could be used as a
final oxidation step after biotreatment.
EDC Wastewater—
The last wastestream studied was from an ethylene dichloride process.
The water-white stream was very acidic and had a high chloride and residual
chlorine content. The stream was highly biorefractory and contained EDC
and other volatiles. These were easily steam stripped from the wastewater
to reduce the average TOG from 1000 to 400 mg/1. The overhead stream from
the stipper was primarily EDC and could be recycled to the process. The
bottom stream from the steam stripper was ozonated and studied for biotreat-
ability. The stripped EDC stream reacted well with ozone after adjustment
to a basic pH. An 82 percent TOG reduction was achieved within 3 hr of
ozonation. Figure 34 indicates a weight ratio of 5.6 mg 0.,/mg of TOG is
required for complete oxidation. Pure EDC would require a ratio of 12.
The initial dip in the TOG fraction is due to air stripping. Air stripping
and ozonation were less effective at low pH values, as shown in Figures 34
and 35. The actual TOC reduction fell below the mass-transfer limiting
line in Figure 35 due to air stripping. The reaction is fairly rapid and
the reaction regime approaches that of a mass-transfer limiting regime.
The EDC wastewater with a pH above 10 had the highest ozone utilization of
those studied. It is possible that the reaction could be carried to complete
oxidation for this waste stream.
Biological batch tests were performed on both stripped and ozonated
stripped EDC wastewater. Preliminary batch tests on stripped EDC wastewater
which was ozonated for 3 hr showed the TOC was reduced below a minimum to
produce significant results. Therefore, a sample was ozonated for 30 min
to maintain a treatable TOC level. The TOC was reduced from 409 to 286
166
-------
APH-12, 15 °C
=12, 25°C
-1, 15°C
-l, 25°C
CJ
o
H
UO
a
o
E-f
a
o
•H
4J
O
td
0123456
Weight ratio (mg 0,/mg TQC)
*0zonation, pH=12,
A15°C
OOzonation, pH=l,
Air strip, pH=12,
\Mass transfer •
\ limiting
2345
Time (min)
Figure 32. Ozonation of styrene wastewater as a
function of ozone transferred.
Figure 33. Ozonation of styrene wastewater
as a function of time.
-------
oo
o
u
8
60
.3
d
•H
cfl
O
o
H
(3
O
•H
4-1
O
C
15°C
ApH=12, 25°C
OpH-1, 15 °C
pH=l, 25°C
0 1234567
Weight ratio (mg 03/mg TOG)
Figure 34. Ozonation of stripped ethylene
dichloride wastewater as a function
of ozone transferred.
2345
Time (min)
Figure 35. Ozonation of stripped ethylene
dichloride wastewater as a
function of time.
-------
mg/1 (30 percent). The ozonated material was significantly more biodegrad-
able than untreated, stripped EDC wastewater, as shown in Table 100. The
TOC reductions averaged 63 percent higher for the ozonated material. An
equivalent reduction in BOD was achieved in 48 hr for the ozonated material
as was achieved in 72 hr for untreated, steam-stripped EDC. These batch
test results were the best obtained during the project.
The combination of steam stripping followed by ozonation produced
excellent results; the TOC was reduced from an original value of 1100 to
400 mg/1 after stripping to less than 100 mg/1 with 3 hr of ozonation.
Ozonation showed potential for either complete oxidation or as a pretreatment
to biooxidation. The quantity of wastewater will determine the process
economics for each ozonation route. A summary of the ozonation study with
potential feasibility is given in Table 101.
Ozonation Studies for Reactions of Benzoic Acid, Benzaldehyde, and Styrene
Aqueous solutions of styrene, benzaldehyde, and benzoic acid were
ozonated in the semi-batch mode. The ozone feed gas concentration and
operating parameters were selected to starve the reaction if it were mass-
transfer limiting with an instantaneous reaction. The effect of temperature
and solution pH on the individual reaction rates were studied. The reaction
rate constant was calculated from the initial 15 min of each test to reduce
the effects of competing side reactions between ozone and reaction products.
The mass-transfer limiting line was calculated for each test using eq. (19).
NA GYBF
^r- = ! - -^r t <19)
N. N. bn
Ao Ao 1
The stoichiometric mole ratio, b.. , was determined from a plot of the fraction
of reactant remaining versus the mole ratio of ozone to the liquid reactant.
The moles of ozone transferred into the liquid for reaction was determined
by a material balance calculation. The proper reaction regime was determined
from comparing the liquid reactant reduction to that predicted with the
truly mass-transfer limiting line and by monitoring the ozone concentration
in the liquid. The reaction rate constant was evaluated with eq. (20) if
the reaction was slow.
N. -o),t
N.
Ao
where:
v \r^
(21)
The 5-day biological oxygen demand (BOD ) was determined on each test
stream before and after ozonation. The ratio of BOD,, to total organic
carbon (TOC) gives the oxygen utilized for oxidizing the organic carbon per
carbon atom. Comparing the experimental BOD to TOC ratio to theoretical
values for complete oxidation gives an indication of the biotreatability of
the stream. An ozonation efficiency was calculated for each compound
169
-------
TABLE 100. SUMMARY OF BATCH STUDY RESULTS FOR STRIPPED EDC WASTEWATER BEFORE AND AFTER OZONATION
-J
O
Reactor 1
Time (hours)
Stripped EDC
F/M
TOC (mg/1)
TOG Reduction (%)
BOD, (mg/1)
BOD;? Reduction (%)
Ozonated Stripped EDC
F/M
TOC (mg/1)
TOC Reduction (%)
BOD, (mg/1)
BOD^ Reduction (%)
0
0.05
75
87
0.05
79
104
72
37
51
17
80
19
76
<30*
>71
Reactor 2
0
0.10
126
158
0.10
138
188
72
38
70
30
81
19
86
33*
82
Reactor 3
0
0.15
170
218
0.15
193
250
72
88
48
84
61
13
94
59
76
Reactor 4
0 72
0.20
175 104
41
298 140
53
0.157
182 44
76
271 68*
75
Reactor 5
0
72
Airstripped
203
302
106
48
163
46
Airstripped
245
408
105
57
204*
50
Reactor 6
0
72
Baseline
27
28
27
0
30
0
Baseline
22
27
16
27
<30*
0
Reading taken at 48 hr.
-------
TABLE 101, SUMMARY OF OZONATIQN. AS. A PRETREATMENT FOR INDUSTRIAL' WASTEWATERS
Stream
% TOG reduction after three hours of
ozonation (initial TOC=100 mg/1)
Weight ratio of ozone to TOG required
complete oxidation, mg O./mg TOG
pH for best TOG reduction
Observed reaction regime
TDI
26%
for 7.0
<3
Reaction rate
controlling
Is ozonation feasible for complete ozidation? No
Did ozonation improve biotreatability
Is ozonation feasible as pretreatment
biooxidation?
? Yes
to Yes
Polyol
48%
7.3
>10
Reaction rate
controlling
(refractory
formed)
No
Yes
Questionable
(Refractory formed)
Styrene
70%*
12.5
>10
Reaction rate
controlling
(slow reaction)
No
No**
No
Stripped EDC
82%
5.6
>10
Mass-transfer
limiting with
a slow reaction
Yes
Yes
Yes
TOG reduction was predominantly from airstripping.
Styrene is biotreatable before ozonation.
-------
ozonated by dividing the time required for complete oxidation based on
mass-transfer limited basis by the actual time required.
Aqueous Benzoic Acid Solution—
Benzoic acid solutions were prepared with an average concentration of
300 mg/1. The stoichiometric ratio for the disappearance of benzoic acid
was found to be approximately 4 moles of ozone per mole of benzoic acid by
extrapolating the data for the first 15 min of reaction in Figure 36. This
value of b, was substituted into eq. (19) to calculate the mass-transfer
limiting line.
The effect of temperature on the benzoic acid oxidation was evaluated
first with Figure 37 showing the benzoic acid reduction as a function of
time at 15° and at 25°C. The benzoic acid reduction at either temperature
deviates from the mass-transfer limiting line^ Rate constants of 1.95
x 10 Kg mole/m min and 3.43 x 10 Kg mole/m min were calculated at 15°
and at 25°C using eq. (22):
-ln(l-C) = u1t (22)
where o> is the reaction rate modulus and £ is the fraction of liquid reactant
converted.
The increase in the reaction rate was offset by a decrease in ozone
solubility at the higher temperature. The mole fraction of ozone in the ,
liquid was calculated from a material balance to be 0.49 x 10 and 0.23 x 10
for 15° and 25°C, respectively. The liquid ozone concentration could not
be measured directly due to interference of the benzoic acid on the potassium
iodide method.
The benzoic acid oxidation was studied at pH values of 2, 4, and 11 at
25°C. The effect of solution pH on the reaction rate was dramatic as the
reaction became mass-transfer limited at pH 11. The data are shown in
Figure 38 for the fraction of benzoic acid remaining as a function of time.
The slowest rate constant was 9-84 x 10 Kg mole/m min for an acid solution
(pH 2). A minimum reaction rate was estimated for the mass-transfer limited
reaction at pH 11 by combining the mass-transfer limited equation with the
slow reaction equation. The estimated minimum reaction rate constant to
maintain a truly mass-transfer limited system was 8.01 x 10 Kg mole/m
min.
The benzoic acid reduction line at pH 11 begins to deviate from the
mass-transfer limiting line after 10 min of ozonation, but the reaction was
starved for ozone a total of 25 to 30 min. This indicates the high pH also
increased the reaction rate of the benzoic acid oxidation products which
were competing with benzoic aetd f°r the ozone. After 30 min of ozonation,
ozone was detected in the exit gas, indicating that some refractor reaction
products were being produced, and that the slower reactions were beginning
to control the system. The benzoic acid reduction was significantly slowed
after 85 percent of the compound was oxidized. This decrease in benzoic
acid reduction at the lower concentrations proves the reaction is dependent
on the benzoic acid concentration.
172
-------
CO
S3
"S
S3
^s
00
•rl
O
O
•H
O
N
d
3
pq
Turbine Speed = 700 rpm
Gas Feed Rate = 0.41 cfm '
Weight Percent Ozone =
1.08%
BA
ppm
N = 108 x 10 6
0 Ib-mole
1.0
0 12345 67
Mole ratio (ozone/benzole acid)
Figure 36.. Determination of stoichiometric
ratio for the ozonation of benzoic
acid.
60
c
•H
§
T)
'H
o
o
N
g
PQ
0.4
0.2
Turbine Speed - 700 rpm
Gas Feed Rate =0.41 cfm
Weight Percent Ozone=108%_
300 ppm
108 x 10 ~6 Ib-mole
pH = 4
Temperature
O 15°C
25°C
" Mass-
Transfer
"Limiting
0 20 40
60 -80 100 120
Time (min)
140 160
Figure 37 . Effect of temperature on the
ozonation of benzoic acid in
aqueous solution.
-------
Turbine Speed = 700 rpm
Gas Feed Rate =0.41 cfm.
Weight Percent Ozone =1.08%
N = 108 x 10~6 Ib-mole .
Temperature = 25°C -
0 20 40 60 80 100 120 140 160
Time (min)
Figure 38. Effect of pH on the ozonation of
benzoic acid in aqueous solution.
1.0
0.9
~ o 0.8
w
f 0.7
*"a
3
-------
The increased oxidation rate with a high pH was attributed to the
formation of hydroxy radicals (OH). The decomposition of ozone in the
liquid phase is greatly increased at high pH values. Previous stud-
ies (14,16,17,22,24) have shown that other ozonation reactions proceed
faster under basic conditions.
Benzoic acid is fairly biodegradable; the experimental BOD to TOG
ratio was 2.46, as compared to a theoretical value of 3.05. The ratio was
increased as the ozonation proceeded to 2.89 after 150 min of ozonation. A
summary of the benzoic acid ozonation data is given in Table 102. A low
efficiency for benzoic acid oxidation was a result of the rapid decrease in
benzoic acid reduction at low residual concentrations.
TABLE 102. SUMMARY OF THE OZONATION REACTION STUDY FOR BENZOIC ACID,
BENZALDEHYDE AND STYRENE IN AQUEOUS SOLUTIONS
Compound
pH=4 pH=2
15°C 25°C 25°C
pH=ll
25°C
Benzoic Acid
Reaction Regime
Reaction Rate Constant,
Kg_mole x 1Q-4
m min
Ozonation Efficiency, %
Benzaldehyde
Reaction Regime
Reaction Rate Constant,
Kg_mole x 1Q-4
m min
Ozonation Efficiency, %
Styrene
Reaction Regime ^,
Reaction Rate Constant,
194.88
47.7
343.04
48.4
98.24
41.8
8000
55.2
213.28
24.6
419.68
22.6
154.08
21.8
3901.60
25.5
Kg mole 0-4
_ X J-U
m min
Operation Efficiency, %
>8000
100
>8000
100
>8000
100
>8000
100
^ass-transfer limiting with a slow reaction.
bMass-transfer limiting with an instantaneous reaction.
CReaction rate controlled.
Aqueous Benzaldehyde Solution—
The benzaldehyde oxidation followed a pattern similar to that of
benzoic acid. The average initial benzaldehyde concentration was 285 mg/1.
The stoichiometric mole ratio for benzaldehyde was calculated to be 1.7
moles of ozone per mole of benzaldehyde by extrapolating the initial ben-
175
-------
zaldehyde reduction data with linear regression in Figure 39. This deviates
from the ratio predicted by the proposed reaction of benzaldehyde to benzoic
acid in eq. (23).
0 0
II II
CH 4- 0 —> COH + 0 (23)
o - o
The data are extremely scattered because benzoic acid has a reaction rate
constant of the same magnitude as the benzaldehyde and competes for the
ozone transferred into the liquid.
The temperature effect for benzaldehyde oxidation was similar to that
for the benzoic acid oxidation; the calculated rate constant was doubled
for the 10-degree increment while the benzaldehyde reduction rate remained
the same. The calculated reaction rates were 13.33 Ib mole/ft min at
15°C and 26.23 Ib mole/ft min at 25°C. The increase in reaction rate wasfi
again offset by the ozone liquid mole fraction being halved from 0.5 x 10
to 0.24 x 10 at the respective temperatures. The system was mass-transfer
controlled with a slow reaction regime; the ozone never reached equilibrium.
The deviation from the mass/transfer limiting line is evident in Figure 40.
Figure 41 shows the benzaldehyde reduction approached the mass-transfer
limiting line at pH 11. The benzaldehyde reaction was truly mass-transfer
limited for the initial 5 min. Unlike the benzoic acid ozonation, a concen-
tration of ozone was detected in the liquid after the initial 5 min, shifting
the reaction to the mass-transfer limited with a slow reaction regime. The
stoichiometric ratio and the early deviation from being mass-transfer
limited suggest an intermediate was formed between the benzaldehyde and the ,
benzoic acid.. The mole fraction of ozone in the liquid dropped from 0.06 x 10
to 0.03 x 10 as the benzoic acid continued to be produced and reacted at
a faster rate than the benzaldehyde. The rate constants were calculated ,.
for the pH extremes using eq. (20). These values were 1.54 x 10 Kg mole/m
min for a solution pH of 2 and 3.91 x 10 Kg mole/m min pH 11. The ben-
zaldehyde relative biodegradability is shown with an experimental BOD,, to
TOG ratio of 2.30 compared to a theoretical value of 2.85. The ozonation
again increased the ratio to a value of 2.66 after 150 min.
Aqueous Styrene Solution—
The ozonation of styrene in aqueous solution was difficult to evaluate
because the styrene was easily air stripped from solution and very insoluble
in water. The average initial concentration achieved was 130 mg/1. The
styrene ozonation data are shown in Figure 42. The ozonated styrene reduction
line fell below the mass-transfer limiting line which is not feasible.
Therefore, air stripping tests were performed which showed the styrene was
air stripped from the solution. The air stripping curve in Figure 42 was
used to correct the ozonation data by compensating for the styrene air stripped
from the solution and not available for oxidation. The styrene reduction
data for the various pH solutions and temperatures were in agreement with
the mass-transfer limiting with an instantaneous reaction line after being
corrected for air stripping as shown in Figure 35. The mass-transfer
176
-------
pa
a
M
toQ
C
•H
c
H
CD
I
r-t
cd
N
g
M
I I I I I
Turbine speed = 700 rpm
Gas feed rate = 0.41 cfm
Weight percent ozone = 1.06%
Ib-mole
pH=5
O- 15°C
- 25°C
• \ Transfer
1 Limiting
20 40 60 80 100 120 140 160
Time (min)
Figure 40. Effect of temperature on the ozonation
of benzaldehyde in aqueous solution.
• i •
Turbine speed = 700 rpm
Gas feed rate = 0.41 cfm
Weight percent ozone = 1.06%
B 119x10
-Temperature = 25 C
°^ O pH=2
A pH=5
D pH=ll
100 120 140 160
Time (min)
Figure 41. Effect of pH on the ozonation
of benzaldehyde in aqueous
solution.
-------
co
CO
&
CO
55
s^^-
00
C
•H
a
a>
CO
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
6 8 10 12
Time (min)
14 16
Figure 42. Ozonation in styrene in aqueous
solution and the effect of air-
stripping on the system.
^-N
O
CO
55
co
53
*^
60
a
•r-l
a
•r^
0)
S-l
a)
a
0)
p*,
4-1
CO
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
n
i i i i i i <
Turbine speed = 700 rpm
Gas feed rate = 0.41 cfm
Weight percent ozone = 1.12%
CgY - 130 ppm
f\ +*f f\
N "- 55.4x10 Ib-mole
SY
o
Temperature = 25 °C
pH = 11^
_
/? o ~ Experimental
"i o Model
fl
-r fc
' i
I C2
| ^
> »x
XX
t lO "**O i i i i
0 2 36 8 10 12 14 .'H
Time (min)
Figure 43. Application of the model for
benzaldehyde from the ozonation
of styrene in aqueous solution.
-------
limiting and reaction rate limiting equations were combined to estimate a
minimum reaction rate constant of 8.01 x 10 Kg mole/in3 min. The method
described by Levnspiel (8) was used to estimate a reaction rate constant of
8.87 x 10 Kg mole/m min.
The styrene rate constant was verified with the use of the model
developed for benzaldehyde. The various rate constants determined for
benzaldehyde at the different operating conditions and the styrene rate
constant were used to calculate the fraction of benzaldehyde as a function
of time. Figure 43 is a comparison of the model prediction, depicted with
the dotted line, with experimental data for a solution pH of 11 and a
temperature of 259C. The model prediction correlated well with the experi-
mental data supporting the rate constants calculated.
Styrene is relatively nonbiodegradable; the experimental BOD_ to TOC
ratio was 0.47 as compared to the theoretical value of 3.33. Ozonation
greatly improved the biotreatability, as the ratio was increased toward the
benzoic acid ratio after 90 min of ozonation and finally 2.69 after 150
min. The total organic carbon content was decreased from 130 to 39 mg/1
for a 70 percent reduction. The ozonation efficiency was 100 percent; all
the tests were mass- transfer limited.
Reaction Study Summary
>
Ozonation was found to improve the biotreatability of water containing
styrene by transforming the styrene to biodegradable compounds such as
Benzoic acid. The BOD,, to TOC ratio was imprpved for all the compounds
studied as shown in Taole 103. The styrene solutuion ratio was significantly
improved after 150 win of ozonation from 0.41 to 2.69.
TABLE 103. SUJfl^Y .OF FIVE-DAY BIOLOGICAL OXYGEN DEMAND TO TOTAL ORGANIC
CARBON RATIO DATA /FOR: REACTION SOLUTIONS
Solution
Benzoic Aeid
Benzaldehyde
S/tyrene
Theoretical
3.05
2.86
3.33
BOD5/TOC
Pre-Ozonation
2.46
2.30
0.47
Post-Ozonation
2.89
2.66
2.69
aOzonation performed at solution pH of 5 and a temperature of 25 °C
for 150 min.
The complete oxidation of styrene -may be costly as some reaction
products react wore slowly, "but the initial oxidation of styrene to benzal-
dehyde was rapid. The ozonation of styrene was truly^ass-transfer limited
with, an estimated reaction rate constant of 8.87 x 10 Kg mole/m min. The
rate constant wass verified, as the model accurately predicted the benzal-
dehyde concentration from the ozonation of styrene using the various rate
constants for benzaldehyde and the rate constant of styrene.
179
-------
The ozonations of benzaldehyde and benzole acid were affected by both
solution pH and temperature. Benzaldehyde and benzole acid ozonations were
mass-transfer limited with a slow reaction under acidic conditions. But at
a pH of 11, the reactions became truly mass-transfer limited. The higher
oxidation rates are related to the increased hydroxy radical formation from
ozone under basic conditions. A 10-degree increase in temperature produced
a higher reaction rate constant which was offset by lower ozone liquid
concentrations. The overall net effect was equivalent reduction rates for
both temperatures. None of the reactions were truly reaction-rate controlled
as the ozone never reached its solubility limit.
The reaction was proved to be bimolecular since the reactions were
dependent on both reactant concentrations. The lower ozone liquid concentra-
tion reduced the reduction in the temperature study while the liquid reactant
reduction was slowed at lower concentrations. A summary of the ozonation
studies including the proper reaction regime, rate constant, and ozonation
efficiency for the oxidation of the compounds was presented in Table 102.
ECONOMIC EVALUATION
Ozone as Pfetfeatment for Styrerie Plant Wastewater
An investigation was made of the economics of treating a process
effluent containing aromatics, principally styrene, as the contaminant.
The capital cost of the installation and manufacturing costs are shown in
Table 104.
TABLE 104. .CAPITAL .AND.OPERATING:COST FOR.OZONE PRETREATMENT
Capital Cost
Ozonator 115-kg/day ozonators $320,000
Reactor Stainless steel tank with agitator
and sparger 70,OdO
Pumps, piping and instrumentation 60,000
Installation x 100,000
Total $550,000
Operating Cost Per Year
Utilities 10 kw/kg of ozone at,$0.03/kwh $150,000
Manpower 5 men at $10/hr 100,000
Maintenance at 6% of capital \ 33,000
Insurance and Taxes at 2% of capital 12,000
G&A at 2% of capital 12,000
Chemicals and Supplies 12,000
Depreciation - 10-year straight line 55,000
Interest at 9% 2,500
Start-up cost \ 24,000
Total Treatment Cost $400,000
_ _____
-------
This ozonator was to be added as a pretreatment step to a system already
biotreating the process effluent. The ozonation equipment was designed to
treat wastewater from a styrene plant containing 300 mg/1 TOC of which 200
mg/1 was styrene. The reactor was sized to treat 600 liters/min wastewater
flow. The ozonators were designed to produce an excess of 20 percent of
the ozone required to oxidize the styrene to benzaldehyde.
Economics of Pretreatment
Table 1Q5 shows the cost of ozone pretreatment of a styrene plant
effluent with the previously indicated contaminant loading. If a manufac-
turing value of 22
-------
REFERENCES
1. Weber, W.J., Physicochemical Processes for Water Quality Control, 1972.
Environmental Science & Technology Series, Wiley Inter-Science, New
York, p. 234.
2. U.S. Environmental Protection Agency, Process Design Manual for Carbon
Adsorption (PB-227 157/5WP), October 1973.
3. Kirk, R.E., and D.F. Othmer, Eds, Encyclopedia of Chemical Technology,
2nd Edn., Vol. 4, 1964. Wiley Inter-science, New York.
4. Barona, N., and H.W. Prengle. Design Reactors This Way for Liquid-
Phase Processes - Part I. Hydrocarbon Processing 52:63, 1973.
5- Yocum, F.H., "Oxidation of Styrene in Aqueous Solution with Ozone,"
D.E. thesis, Tulane University, New Orleans, Louisiana, 1977.
6. Morrison, R.T., and R.N. Boyd. Organic Chemistry. Allyn and Bacon,
Inc., Boston, 1961.
7. Subluskey, L.A., G.C. Harris, A. Maggiolo, and A.L. Tumolo. Improved
Synthesis of Aromatic Aldehydes from Ozonalysis of Olefins. In:
Advances in Chemistry Series No. 21, Ozone Chemistry and Technology,
American Chemical Society, 1959. p. 149-
8. Levenspiel, 0. Chemical Reaction Engineering. John Wiley and Sons,
Inc., New York. 1967.
9. Cooper, C.M., G,A, Fernstrom, and S.A. Miller. Performance of Agitated
Gas-Liquid Contactors. Ind. Eng. Chem. 36:504, 1944.
10. Kawecki, W,, T. Reith, J.W. van Heuven, and W.J. Beck. Bubble Size
Distribution in the Impeller Region of a Stirred Vessel. Chem. Egr.
Sci., 22:1519L, 19L67.
11. Westerterp, K.R., L.L. van Dierendonck, and J.A. de Kraa, Interfacial
Areas in Agitated Gas-Liquid Contactors. Chem. Egr. Sci. 18:157, 1963.
12. Prengle, H.W;,, C.G. Hewes, 111, and C.E. Mauk. Oxidation of Refractory
Materials By Ozone with Ultraviolet Radiation. In: Proceedings of
the First International Symposium on Ozone for Water and Wastewater
Treatment,"R.C, Rice and M.E. Browning, Eds. International Ozone
Institute, Syracuse, N.Y., 1275.
182
-------
13. Rand, M.C., A.E. Greenberg, and M.J. Taras, eds. Standard Methods for
the Examination of Water and Wastewater, 14 ed. American Public Health
Association, Washington, D.C., 1976.
14. Kroop, R.H., "Ozonation of Phenolic Aircraft Paint Stripping Wastewater,"
Proceedings of First International Symposium, International Ozone
Institute, 660, 1975.
15. Oldshue, Ind. Eng. Chem. 48:2194, 1956.
16. Hann, V.A., and T.C. Manley, Encyclopedia of Chemical Technology,
Vol, 9, 1952. Interscience Encyclopedia, Inc., New York, p. 735.
17. Eisenhauer, H.R., "The Ozonation of Phenolic Wastes," Journal WPCF,
40:1887, 1968.
18. Yoshida, F., and Y. Miura, "Gas Absorption in Agitated Gas-Liquid
Contactors," Ind. Eng. Chem. Proc. Des. Dev. 2:263, 1963.- \
19- Foust, C., D.E. Mack, and J.H. Rushton, "Gas-Liquid Contacting by -
Mixers," Ind. Eng. Chem. 36:517, 1944.
/
20. Calderbank, P.H., "Mass Transfer - Chapter 6," in Mixing, Theory and
Practice, Vol, II, V.W. Uhl and J.B. Gray, eds., 1967. Academic Press,
New York.
21. Fair, J.R., "Designing Gas-Sparged Reactors - I," Chem. Eng. 74:67, 1967.
22. Etsenhauer, H.R., Increased Rate and Efficiency of Phenolic Waste
Ozonation," Journal WS>CF. 43:2QQ, 1971.
23. Prengle, R.W., C.E. tiauk, and J.E. Payne, "Ozone/UV Oxidation of
Chlorinated Compounds in Water," paper presented at Forum on Ozone
Disinfection, Chicago, Illinois CJune, 1976).
24. Eisenhauer, H.R., "Dephenolization by Ozonalysis," Water Research
5:467, 1971.
25. Prengle, K.W,, C,E, Mauk, R.W. Legan, and C.G. Hewes, III, "Ozone/UV
Process. - Ef.feet:£ye Ia§tewater Treatment," Hydrocarbon Processing
54s82, 1275.
26. Majumoar, S.B., W.B. Ceckler, and O.J. Spoul, "A physical mathematical
model of isass, transfer and reaction kinetics of ozone," paper presented
a,t ASC6E 68th national meeting, Los Angeles, California, November,
1175.
183
-------
APPENDIX
10
ti
o
,£>
H
cd
o
&o
(30
•d .1.0
-------
_0
1?
..0
G I
O
I
0 0.1
bO
00
o
CO
.01
o Filtrasorb
• Westvaco
o Witco
.001 0.01 0.1 1.0
Residual EDC Concentration (g/1)
Figure 5A. Ethylene dichloride
isotherm for stream 221A (lower
concentration range).
e o.i
a
o
rO
cd
CJ
bo
"bo 0.01
•s
o
ca
Q .001
0.01
0.1
1.0
Residual EDC Concentration (g/1)
Figure 6A. Ethylene dichloride
iostherm for stream 161A (upper
concentration range).
9 i.o
a
o
fj
a
bO
"» 0.1
Adsorbed,
~*i
4J
g 0.01
; 1 1 1 — I i I 1 I | 1 1 1 — I I I I I.
-
~
r__-^— ^ ;r-— ~
.
i I . i I i I il 1 1 1 1 1 1 1 L
i.o
a
o
•s
ctf
a
bO
d bo
(U
f,
O
ca
0.1 1.0 10
Residual EDC Concentration (g/1)
Figure 7A. Ethylene dichloride
isotherm for stream 041A.
0.1
Monochen
10
Residual EDC Concentration (g/1)
Figure 8A. Ethylene dichloride
isotherm for stream 221A.
185
-------
1.1
a
o
,0
00
00
(U
•s
o
.01
o Filtrasorb
• Westvaco WVG
Q Witco
AB-C NB9377
0.1 l.o
Residual TC Concentration (g/1)
Figure 9A. Total carbon isotherm
for stream 041A.
rt
o
•S
to
o
M
-».
00
13
0)
,0
M
O
CO
.01
0.1
Filtrasorb o
Westvaco •
/ Witco a
Soot carbon 79 0
Soot Carbon 104 »
1.0
Residual TC Concentration (g/1)
Figure 10A. Total carbon isotherm
for stream 221A.
100
g
•8
«)
O
60
M
T3
Q)
•s
O
to
'O
10
0.1
O Filtrasorb
0 St»ot carbon
AB-C KE0547
• Nuchar
I 1.0 10 100
Residual TOC Concentration (mg/1)
Figure 11A. Total organic carbon
isotherm for stream 351A.
g
•S
100
00
t>0
13
0)
•s
o
01
4-1
10
O Filtrasorb
0 Soot carbon
A B-C KE0547
• Nuchar
10 100 1000
Residual TOG Concentration (mg/1)
Figure 12A. Total organic carbon
isotherm for stream 231A.
186
-------
Effluent Volume (liters)
*
Figure 13A. Breakthrough curve
for laboratory continuous adsorp-
tion study with Westvaco WVG
carbon.
4J
g
3
t-l
M-l
£6000
§4000
•H
4-)
n)
M
4J
e
g2000
g
o
"~TnFr. BCD 'cone.
• Low
o lledium
o High
2468
Effluent Volume (liters)
Figure 14A. Breakthrough curve
for laboratory continuous adsorp-
tion study with Filtrasorb-400
carbon.
1200
Effluent Volume (liters)
Figure 15A. Breakthrough curve
for laboratory continuous adsorp-
tion study with Monochem soot
carbon.
g
cd
M
4-J
5
§
U
O
H
800
400
-i 1 r
Infl. TC
• Low
o Medium
oHigh
2 4 6
Effluent Volume (liters)
Figure 16A. Breakthrough curve
for laboratory continuous adsorp-
tion study with Westvaco WVC
carbon.
187
-------
bo
a
w
a
S3
o
•H
4-1
td
!-l
•U
C
0)
o
1200
800 -
400 -
• Low
D Medium
QHigh
Effluent Volume (liters)
Figure 17A. Breakthrough curve for
laboratory continuous adsorption
study with Filtrasorb-400 carbon.
12000 -
4-1
§
M-l
W
•S
c
o
•H
4J
cd
n
4-1
fl
0)
a
a
o
u
o
H
.8000 -
4000 -
Effluent Volume /(liters)
Figure 19 A. Breakthrough curve for
laboratory continuous adsorption
s'tudy with Westvaco WVG carbon.
Effluent Volume (liters)
Figure ISA. Breakthrough curve for
laboratory continuous adsorption
study with Monochem soot carbon.
12000r
8000
4000
• Low
a Medium
oHigh
Effluent Volume (liters)
Figure 20A. Breakthrough curve for
laboratory continuous adsorption
study with Filtrasorb-400 carbon.
188
-------
007000
6
i 1 1 1 r
Infl. TC cone.
• Low
D Medium
o High
j_
1357
Effluent Volume (liters)
Figure 21A. Breakthrough curve for
laboratory continuous adsorption
study with Westvaco WVG carbon.
• Low
O Medium
o HiRh
Effluent Volume (liters)
Figure 22A. Breathrough curve for
laboratory continuous adsorption
study with Filtrasorb-400 carbon.
g
ja
M
tfl
O
60
CO
O
•rl
0.5
00
00
0.3
'O
QJ
•8
o
co
0.1
I
3000
«Piltrasorb-300
opiltrasorb-400
• Westvaco WVG
DWitco
1357
Total Organics Through Unit (Ib)
Figure 23A. Field study of var-
ious activated carbons.
oo
n
o
500 1000
Total Gallons Through Unit
Figure 24A. Breakthrough curve for
field study with Filtrasorb-300
carbon, stream 221A.
189
-------
500
§
H
(4-1
.S 30C
g
•H
0)
M
4-1
g 100
o
ti
o
o
Avg.infl.conc.
o
H
500
1000
Total Gallons Through Unit
Figure 25A. Breakthrough curve for
field study with Filtrasorb-300
carbon, stream 221A.
2000
3
•H 1200
g
4-1
g
o
a
o
u
400
500
1000
Total Gallons Through Unit
Figure 26A. Breakthrough curve for
field study with Filtrasorb-300
carbon, stream 221A.
3000
2000
4-1
g
O
a
o
o
o
1
00
1000
Avg. Infl. cone
M
-------
3000
?
2000
ti
o
•rl
4-1
a
n
4J
§
o
a
o
CJ
u
H
iooo
o o
500 1000 1500
Effluent Volume (gallons)
Figure 29A. Breakthrough curve for
laboratory continuous adsorption
study with Filtrasorb-400 carbon.
3000
2000
4J
2
4J
§1000
c
o
u
g
500 1000
Effluent Volume (gallons)
1500
Figure 30A. Breakthrough curve for
laboratory continuous adsorption
study with WITCO carbon.
600
g
400
a
0)
| 200
o
•H
60
500 ' 1000- ' 1500
Effluent Volume (gallons)
Figure 31A. Breakthrough curve for
laboratory continuous adsorption
study with WITCO carbon.
go 2000
1200
oo
fr
400
td
4-1
O
H
500 1000 1500
Effluent Volume (gallons)
Figure 32A. Breakthrough curve for
laboratory continuous adsorption
study with WITCO carbon.
191
-------
3000
1
g 2000
•H
4J
nt
M
•M
G
§
§1000
u
u
•H
c
cti
60
O
1 1 1 1 1
- o
o
~~~- — o A
"~ — "^ /^. "7 r» ft
~~~ ~-~~ _ "X^ *•>
^^^~-o^C-
y
O/
/
/
/
/
A * X yv yv 1 >v >*J y^^^O 1
r-1
^ 800
"fi
B
(U
C-)
Sn 400
00
H
o
H
cd
-j,i
o
H
o
-x. o
- o ^X o
^** o
<^ V \ ®
^. >s
^-^>*
o •{ ^
* O/T^
O
3
O
0 ^ — • o o
s^&Q O ®
O S^
y'
/,,,,,
/. nn 1 9nn onnn
400
1200
2000
Effluent Volume (gallons)
Figure 33A. Breakthrough curve for
laboratory continuous adsorption
study with Westvaco WVG carbon.
3000'
2000
§
•rl
JJ
ni
M
4J
cl
g
g
o
•H
00
M
O
1000
iOO 1600 2400
Effluent Volume (gallons)
Figure 35A. Breakthrough curve for
laboratory continuous adsorption
study with Westvaco WVG carbon.
Effluent Volume (gallons)
Figure 34A. Breakthrough curve for
laboratory continuous adsorption
study with Westvaco WVG carbon.
g
•H
ti 800
g
u
g
o
400
800 1600 2400
Effluent Volume (gallons)
Figure 36A. Breakthrough curve for
laboratory continuous adsorption
study with Westvaco WVG carbon.
192
-------
1200
800 -
a
g
60
fr
o
400 -
o
H
400 1200 2000
Effluent Volume (gallons)
Figure 37A. Breakthrough curve for
laboratory continuous adsorption
study with Westvaco WVG carbon.
^2500
Q
8
1500
500
to
"s
10002000
Effluent Volume (gallons)
Figure 38A. Breakthrough curve for
laboratory continuous adsorption
Study.
'5000-
g
•H
4J
2
4J
g
3000
o
o
•rl
§
60
1000 -
O 11 liters/min
A 19 liters/min
o 29 liters/min
1000 2000
Effluent Volume (gallons)
Figure 39A. Breakthrough curve for
laboratory continuous adsorption
study.
10 30 50
Effluent Volume (gallons)
Figure 40A. Flow rate effects of
best activated carbon.
193
-------
6400
• Low
O Medium
o High
• Low
O Medium
o High
Effluent Volume (liters)
Figure 41A. Breakthrough curve for
stream 041A with WVG carbon.
5 7
Effluent Volume (liters)
Figure 42A. Breakthrough curve for
stream 041A with Filtrasorb-400
carbon (solute: EDC).
7250
•H
4J
to
5000
a
0)
o
c
8 3000
1000
M-l
w
Infl. cone.
1357
Effluent Volume (liters)
Figure 43A. Breakthrough curve for
stream 041A with Monochem carbon,
run #104 (solute: EDC).
1200
c
o
•H
« 800
O
B
O
4-1
rH
M-t
400
—i 1 1 r
~lnfl Tccme.
• Low
O Medium
oHigh
135
Effluent Volume (liters)
Figure 44A. Breakthrough curve for
stream 041A with WVG carbon
(solute: TC).
194
-------
• Low
o Medium
oHigh
2468
Effluent Volume (liters)
Figure 45A. Breakthrough curve for
stream 041A with Filtrasorb-400
carbon (solute: TC) .
Effluent Volume (liters)
Figure 46A. Breakthrough curve for
stream 041A with Monochem carbon,
run #104 (solute: TC).
l.OF
g
"S
o
00
o.i
•B
o
CO
3
O.Ol 071 1.0
Residual EDC Concentration (g/1)
Figure 47A. EDC adsorption isotherm
for stream 041A using Monochem
soot carbon.
00
S
2
4-1
C
o 2000
§
u
1000
4-J
g
3
rH
M-l
Inf1. cone.
o High
o Medium
Effluent Volume (liters)
Figure 48A. EDC breakthrough
curve for stream 041A using
Monochem soot carbon.
195
-------
LOT
o
•8
nj
o
t>o
to
13
0)
•8
o
en
o
oil
1 10
Residual EDC Concentration (g/1)
Figure 49A. EDC adsorption isotherm
for stream 221A using Monochem soot
carbon.
to
e
Effluent Volume (liters)
Figure 50A. EDC breakthrough curve
for stream 221A using Monochem soot
carbon (effluent concentration -
9000 mg/1).
196
-------
Figure 51A. Sample A (500X)
Figure 52A. Sample A (2000X)
Figure 53A. Sample B (2000X) Figure 54A. Sample B (2000X)
Figure 55A. Sample B (5000X)
Figure 56A. Sample B (5000X)
Scanning Electron Micrographs
197
-------
Figure 57A. Sample A (70000X) Figure 58A. Sample A (70000X)
Figure 59A. Sample A (70000X) Figure 60A. Sample B (70000X)
Figure 61A. Sample B (70000X) Figure 62A. Sample B (70000X)
Transmission Electron Micrographs
198
-------
iOOF
00
o
<1>
•8
o
CO
10
i i
i
o Commercial
D Monochem
100
10 ' "
TOG Concentration, (mg/1)
100
Figure 63A. Adsorption isotherm
for EDC.
00
o
H
(U
•e
o
CO
10
o Commercial carbon
O Monochem
10
TOC Concentration (mg/1)
100
Figure 64A. Adsorption isotherm
for benzene.
00
8 100
-------
100
100
u
00
g
•s
VJ
3
TOC Concentration (mg/1) /
Figure 67A. Adsorption isotherm
for chloral hydrate.
00
•0
(U
•p
10
o Commercial carbon
100 1000
TOC Concentration (mg/1)
Figure 68A. Adsorption isotherm
for toluene di-isocyanate
(stream 231A),
1000
o
00
I
O
ca
100
10
1 ' I ' •'••••' ' ' "'I
o Commercial carbon
o Monoehera
10 100
TOC Concentration (mg/1)
Figure 69A. Adsorption isotherm
for phenol on s,team activated
carbon.
•0
(U
•e
o
CO
100
oo
10 _
o Commercial carbon
o Honochem
10
100
TOC Concentration (mg/1)
Figure 7QA, Adsorption isotherm
for benzene on steam activated
carbon.
200
-------
1000
s
-d
(U
,0
M
o
CO
100
10
o B-C
A Calgon
X Monocheta
o Nuchar
10 log
TOC Concentration (mg/1)
Figure ;71A. Adsorption isotherm
for stream 231A (treatment: raw).
IOOOF
100
-a
-------
1000
100
0)
•8
o
co
10
OB-C
ACalgon
xMonocheta
o Nuchar
10 100
TOC Concentration (mg/1)
Figure 75A. Adsorption isotherm
for stream 231A (treatment:
ozonated, biotreated).
100(:
locr
-a
cu
•8
o
CO
§
I
A Calgon
X Monochera
I/
10 100
TOC Concentration (mg/1)
Figure 76A. Adsorption isotherm
for stream 231A (treatment:
acidified, biotreated).
1000
100
TJ
-------
1000
100
13
-------
1 1 1 r-
200
100
o
H
o Monochem
o Nuchar
500
300
p
8 100
o Monochera
a Nuchar
5 ^ 13
Volume (liters)
17
o Fig. 82A. Stream 231A (ozonated, biotreated); TOG.
^ o —
^ *
15 9 13 17
Volume (liters)
Fig. 83A. Stream 231A (ozonated, biotreated); TOG.
M
tft
200
100
Q
O
o Monochem
a Nuchar
-i r
-i r-
500
o Monochem
D Nuchar
1 5 9 13 17
Volume (liters)
Fig. 84A. Stream 231A (ozonated, biotreated); BODg.
10 18 26
Volume (liters)
Fig. 85A. Stream 351A; TOG.
-------
10
O,
Ln
500
300 •
o
8
100
o Monochem
o Nuchar
2 10 18 26
Volume (liters)
Figure 86A. Stream 351A Craw); COD.
500
300
100
o Monochem
Q Nuchar
1 5 9 13
Volume (liters)
Figure 88A. Stream 231A (raw); TOG.
17
o Monochem
o Nuchar
200
60
8
'100
m
Q
2 10 18 26
Volume (liters)
Figure 87A. Stream 35 1A (raw);
o Monochem
o Nuchar
1000
600
m
O
o
pq
200
34
1 5 9 13
Volume (liters)
Figure 89A. Stream 231A (raw); BOD5,
17
-------
c!
s
en
o Monochem
o Nuchar
5 9 13
Volume (liters)
Figure 90A. Stream 231A (raw); COD.
200
100
o Monochem
a Nuchar
(Regenerated)
20 36 52
Volume (liters)
Figure 92A. Stream 231 (raw); COD.
68
200
100
4 20 36 52
Volume (liters)
Figure 91A. Stream 231A (raw); TOC.
i I i 1 1 1 1 1 1 r-
o Monochem
Q Nuchar
(Regenerated)
-J -i 1 i_
365268~
Volume (liters)
Figure 93A, Stream 231A (raw); BODq
-------
K3
O
I
500 I
i1
8
500
300
100
o Monochem
a Nuchar
(Regenerated)
15 9 13 17
Volume (liters)
Figure 94A. Stream 351A (raw); TOG.
o Monochem
a Nuchar
(Regenerated)
300
f.
100
200
o Monochem
D Nuchar
(Regenera fed)
100
in
1 5 9 13
Volume (liters)
Figure 95A. Stream 351A (raw); COD.
17
1 5 9 13
Volume (liters)
Figure 96A. Stream 351A (raw); BOD-,
-------
TABLE 1A. ANALYSIS OF COLUMN INFLUENT AND EFFLUENT BY INDIVIDUAL COMPONENT
Type of Carbon: Westvaco Stream: 221A „ Volume Carbon: 64.8 1
Flow Rate: 1.1 1/min Loading: .74 I/cm min
Carbon Weight: 28.6 1/g
Date: 9/17/73
Sample
Number
Cumulative
Running
Time
(In Hours)
... ,v Component
Concentration
Feed
(fflg/D
Out
6
7
8
9
10
6
7
8
9
10
6
7
8
9
10
7
8
9
10
6
7
8
9
10
6
7
8
9
10
6
7
8
9
10
59
71
83
95
101
69
71
83
95
101
59
71
83
95
101
71
83
9.5
101
59
71
83
95
101
59
71
' 83
' 95
101
- 59
71
, 83
95
101
Ethyl Chloride
Ethyl Chloride
Ethyl Chloride
Ethyl Chloride
Ethyl Chloride
Vinyl Chloride
Vinyl Chloride
Vinyl Chloride
Vinyl Chloride
Vinyl Chloride
Vinylidene Chloride
Vinylidene Chloride
Vinylidene Chloride
Vinylidene Chloride
Vinylidene Chloride
Trans-Dichloroethylene
, Trans-Dichloroethylene
Trans-Dichloroethylene
Trans-Dichloroethylene
1 , 1-Dichloro-Ethane
1 , 1-Dichloro-Ethane
1 , 1-Dichloro-Ethane
1 , 1-Dichloro-Ethane
1 , 1-Dichloro-Ethane
Chloroform
Chloroform
Chloroform
Chloroform
Chloroform
Ethylene Dichloride
Ethylene Dichloride
Ethylene Dichloride
Ethylene Dichloride
Ethylene Dichloride
22.7
66.0
37.6
41.7 /
0.8
5.1
0.5 /
2.7 /
/ '
0 /''
0.4
0 /
0
0 '
0
0
0
0
16.8
34.9
23.6
29.9
'
10.8
16.9
7.6
9.1
f
337.9
539.2
517.1
638.8
0
0
0
18.3
283.9
0
0
0
4.3
14.5
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
o,
0
0
0
0
69.0
j. .
208
(continued)
-------
TABLE 1A (CONTINUED)
Type of Carbon: Westvaco
Flow Rate: 1.1 1/min
Sample
Number
6
7
8
9
10
6
7
8
9
1
2
3
4
5
6
7
8
9
^
10
Cumulative
Running
Time
(In Hours)
59
71
83
95
101
59
71
83
95
2
11
23
35
47
59
71
83
95
101
Stream: ' 221A Volume Carbon: 64.8 1
Loading: .74 1/m -min Carbon Weight: 28.6 1/g
Date: 9.17/73
Component
1,1, 1-Trichloro-Ethane
1,1, 1-Trichloro-Ethane
1,1, 1-Trichloro-Ethane
1,1, 1-Trichloro-Ethane
1,1, 1-Trichloro-Ethane
1,1, 2-Trichloro-Ethane
1,1, 2-Trichloro-Ethane
1,1, 2-Trichloro-Ethane
1,1, 2-Trichloro-Ethane
Chloral Hydrate & Chloral
Chloral Hydrate & Chloral
Chloral Hydrate & Chloral
Chloral Hydrate & Chloral
Chloral Hydrate & Chloral
Chloral Hydrate & Chloral
Chloral Hydrate & Chloral
Chloral Hydrate & Chloral
Chloral Hydrate & Chloral
Chloral Hydrate & Chloral
Concentration
Feed
4.9
7.6
3.9
3.4
18.0
20.5
24.0
30.5
761.5
696.7
643.0
577.1
218.3
(868.1)
314.6
(995.7)
0
(576.3)
0
(397.5)
(mg/D
Out
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
322.0
209
-------
TABLE 2A. ANALYSIS OF COLUMN INFLUENT AND EFFLUENT
BY INDIVIDUAL COMPONENT
Type of Carbon:
Flow Rate:
Westvaco
1.9 1/min
Stream: 221A , Volume Carbon:
Loading: 1.23 1/m min Carbon Weight:
Date:
64.8 1
28.6 1/g
9/27/73
Sample
Number
5
6
7
8
4
5
6
7
8
5
6
7
8
5
6
7
8
6
7
8
5
6
7
8
5
6
7
8
5
6
7
8
Cumulative
Running
Time
(In Hours)
45
56
65
77
33
45
56
65
77
45
56
65
77
45
56
65
77
56
65
77
45
56
65
77
45
56
65
77
45
56
65
77
Concentration
Component Feed
Vinyl Chloride 2.3
Vinyl Chloride 8.1
Vinyl Chloride 4.1
Vinyl Chloride 3.1
Ethyl Chloride 65.7
Ethyl Chloride 46.9
Ethyl Chloride 75.6
Ethyl Chloride 55.6
Ethyl Chloride 55.9
Vinylidene Chloride 0
Vinylidene Chloride 0
Vinylidene Chloride 0
Vinylidene Chloride 0
Trans-1 , 2-Dichloroethylene 0
Trans-1 , 2-Dichloroethylene 0
Trans-1 , 2-Dichloroethylene 0
Trans-1 , 2-Dichloroethylene 0
1,1-Dichloro-Ethane 83.8
1,1-Dichloro-Ethane 139.9
1 , 1-Dichloro-Ethane 50 . 1
Chloroform 35 . 1
Chloroform 171.1
Chloroform 140.0
Chloroform 20.2
Ethylene Dichloride 945.2
Ethylene Dichloride 1146.4
Ethylene Dichloride 1174.8
Ethylene Dichloride 1411.1
1,1,1-Trichloro-Ethane 26.4
1,1,1-Trichloro-Ethane 3.5
1,1,1-Trichloro-Ethane 5.0
1,1, 1-Trichloro-Ethane 12.1
(mg/1)
Out
0
0
0
38.5
0
0
0
83.6
252.7
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
611.6
0
0
0
0
210
(continued)
-------
TABLE 2A. (CONTINUED)
Type of Carbon:
Flow Rate:
Westvaco
1.9 1/min
Stream: 221A
Loading: 1.23 1/m min
Volume Carbon
Carbon Weight
Bate:
: 64.8 1
: 28.6 1/g
9/27/73
Sample
Number
5
6
7
8
3
4
5
6
7
8
3
4
5
6
7
8
5
6
7
8
Cumulative
Running
Time
(In Hours)
45
56
65
77
21
33
45
56
65
77
-21
33
45
56
65
77
45
56
65
77
Component
1,1, 2-Trichloro-Ethane
1,1, 2-Trichloro-Ethane
1,1, 2-Trichloro-Ethane
1,1, 2-Trichloro-Ethane
Chloral Hydrate
Chloral Hydrate
Chloral Hydrate
Chloral Hydrate
Chloral Hydrate
Chloral Hydrate
Chloral
Chloral
Chloral
Chloral
Chloral
Chloral
Propylene Bichloride
Propylene Bichloride
Propylene Bichloride
Propylene Bichloride
Concentration
Feed
12.4
17.4
19.6
9.4
163.4
97.6
321.9
0
0
0
345.2
458.1
1060.2
1735.1
(mg/1)
Out
0
0
0
0
0
0
0
530.7
229.3
125.0
0
0
0
0
1311.0 1955.9
1035.1 1574.7
54.4
9.7
29.4
20.2
0
0
0
0
211
-------
TABLE 3A. ANALYSIS OF COLUMN INFLUENT AND EFFLUENT
BY INDIVIDUAL COMPONENT
Type of Carbon: Westvaco Stream:
Flow Rate: 2.85 1/min Loading:
221A 2
1.84 1/m min
Volume Carbon: 64.8 1
Carbon Weight: 28.6 1/g
Date: 10/1/73
Sample
Number
1
2
3
4
5
6
1
2
3
4
5
6
3
4
5
6
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
Cumulative
Running
Time
(In Hours)
7
19
31
43
55
67
7
19
31
43
55
67
31
43
55
67
31
43
55
67
7
19
31
43
55
67
7
19
31
43
55
67
Concentration (mg/1)
Component
Vinyl Chloride
Vinyl Chloride
Vinyl Chloride
Vinyl Chloride
Vinyl Chloride
Vinyl Chloride
Ethyl Chloride
Ethyl Chloride
Ethyl Chloride
Ethyl Chloride
Ethyl Chloride
Ethyl Chloride
Vinylidene Chloride
Vinylidene Chloride
Vinylidene Chloride
Vinylidene Chloride
Trans-1 , 2-Dichloroethylene
Trans-1 , 2-Dichloroethylene
Trans-1 , 2-Dichloroethylene
Trans-1 , 2-Dichloroethylene
1 , 1-Dichloroethane
1 , 1-Dichloroethane
1 , 1-Dichloroethane
1 , 1-Dichloroethane
1 , 1-Dichloroethane
1 , 1-Dichloroethane
Chloroform
Chloroform
Chloroform
Chloroform
Chloroform
Chloroform
Feed
0
0
3.6
4.4
5.0
6.3
72.6
29.4
11.8
33.4
24.7
36.9
0.5
1.4
2.3
14.6
0
0
0
0
237.9
34.6
19-3
33.7
25.3
33.4
15.8
23.0
1.6
8.7
3.4
18.0
Out
0
0
5.1
10.4
11.1
25.0
0
13.0
202.7
247.6
20.0
108.1
0
0
0
0
0
0
0
0
0
0
0
0
0
52.0
0
0
0
0
0
0
(continued)
212
-------
TABLE 3A. (CONTINUED)
/Type of Carbon: Westvaco
Flow Rate: 2.85 1/min
Cumulative
Running
Sample Time
Number (In Hours)
1
2
3
4
5
6
2
3
4
5
x 6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
gr
5
6
7
19
31
43
55
67
19 ,
31
43
/55 ' ,
' f i 67
! 7
(7
' '19 '
31
43
55
67
7
19
31
43
55
67
7
19
31
43
55
67
Stream: 221A „ Volume Carbon: 64.8 1
Loading: 1.84 1/m min Carbon Weight: 28.6 1/g
Date: 10/1/73
Concentration (mg/1)
Component
Ethylene Dichloride
Ethylene Dichloride
Ethylene Dichloride
Ethylene Dichloride /'
Ethylene Dichloride '
Ethylene Dichloride
1,1, 2-Trichloro-Ethane
1,1, 2-Trichloro-Ethane
1,1, 2-Trichloro-Ethane
1,1, 2-Trichloro-Ethane
1,1, 2-Trichloro-Ethane
Chloral Hydrate
Chloral Hydrate
Chloral Hydrate
Chloral Hydrate /
Chloral Hydrate
Chloral Hydrate
Chloral
Chloral
Chloral
Chloral
Chloral
Chloral
Propylene Dichloride
Propylene Dichloride
Propylene Dichloride
Propylene Dichloride
Propylene Dichloride
Propylene Dichloride
Feed
1338.8
1042.9
448.7
1559.4
1377.0
1532.5
17.4
8.7
29.1
13.1
30.0
83.4
41.6
0
0
0
418.0
3232.7
1231.6
1170.8
862.8
1401.3
1842.2
1.0
3.1
2.9
0
34.3
0
Out
0
0
48.2
93.5
156.7
1699.9
0
0
0
0
0
0
0
0
111.9
555.9
699.2
0
603.1
1090.8
753.5
1787.5
2284.0
0
0
0
0
0
0
213
-------
TABLE 4A. ANALYSIS OF COLUMN INFLUENT AND EFFLUENT
BY INDIVIDUAL COMPONENT
Type of Carbon: Filtrasorb
Flow Rate:
1.1 1/min
300 Stream: 221A Volume Carbon: 64.8 1
Loading: .71 1/m min Carbon Weight: 28.6 1/g
Date:
9/4/73
Sample
Number
4
5
6
7
8
3
4
5
6
7
8
5
6
7
8
5
6
7
8
5
6
7
8
5
6
7
8
4-
5
6
7
8
Cumulative
Running
Time
(In Hours)
38
50
62
74
86
26
38
50
61
74
86
50
62
74
86
50
62
74
86
50
62
74
86
50
62
74
86
38
50
62
74
86
Concentration (mg/1)
Component
Vinyl Chloride
Vinyl Chloride
Vinyl Chloride
Vinyl Chloride
Vinyl Chloride
Ethyl Chloride
Ethyl Chloride
Ethyl Chloride
Ethyl Chloride
Ethyl Chloride
Ethyl Chloride
Vinylidene Chloride
Vinylidene Chloride
Vinylidene Chloride
Vinylidene Chloride
Trans-1 , 2-Dichloroethylene
Trans-1 , 2-Dichloroethylene
Trans-1 , 2-Dichloroethylene
Trans-1 , 2-Dichloroethylene
1 , 1-Dichloroethane
1 , 1-Dichloroethane
1 , 1-Dichloroethane
1 , 1-Dichloroethane
Chloroform
Chloroform
Chloroform
Chloroform
Ethylene Dichloride
Ethylene Dichloride
Ethylene Dichloride
Ethylene Dichloride
Ethylene Dichloride
Feed
0.5
0
0
1.0
0.8
22.0
17.5
17.5
23.8
20.4
27.2
0
0
0.8
0
0
0
0
0
17.9
24.9
23.2
33.5
9.8
1.1
14.0
33.8
1,037.6
814.0
897.3
774.3
1,015.8
Out
0
0
0
0.4
5.2
0
0
Trace
0.9
45.5
66.8
0
0
0
0
0
0
0
0
0
0
1.6
0
0
0
0
0
0
0
0
24.8
47.1
(continued)
214
-------
TABLE 4A. (CONTINUED)
Type of Carbon: Filtrasorb 300 Stream: 221A
Flow Rate: 1.1 1/min Loading: .71 1/m min
Volume Carbon: 64.8 1
Carbon Weight: 28.6 1/g
Date: 9/4/73
Sample
Number
Cumulative
Running
Time
(In Hours)
Component
Concentration (mg/1)
Feed Out
5
6
7
8
4
5
6
7
8
50
62
74
86
38
50
62
74
86
1,1,1-Trichloro Ethane 11.2
1,1,1-Trichloro Ethane 0
1,1,1-Trichloro Ethane 2.1
1,1,1-Trichloro Ethane 0
1,1,2-Trichloro Ethane 10.7
1,1,2-Trichloro Ethane 4.8
1,1,2-Trichloro Ethane 1.4
1,1,2-Trichloro Ethane 7.7
1,1,2-Trichloro Ethane 0
0
0
0
0
0
0
0
0
0
215
-------
TABLE 5A. ANALYSIS OF COLUMN INFLUENT AND EFFLUENT BY INDIVIDUAL COMPONENT FOR STREAM 081A
Type of Carbon:
Carbon Volume:
Date:
Dichloroethane
Westvaco WVG
28.3 liters
9/10/74
1,1,2,2-Tetra-
chloroethane
Loading:
Flow Rate:
Carbon Weight:
41.6 1/min-m
3.79 1/m
11.3 kg
Methylene Chloride Trans-Dichloroethylene
Sample
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Running
Time
(Hours)
3
6
9
12
15
18
21
24
27
30
33
36
39
42
45
48
51
54
57
60
Concentration (mg/1)
Influent Effluent
80.47
45.62
152.46
76.00
250.00
111.48
166.00
172.00
4.96
0.40
188.00
162.00
42.34
42.46
23.55
6.16
5.43
56.67
6.48
2.12
0.12
2.60
0.09
25.10
42.00
0.48
156.00
264.88
144.00
162.00
142.00
94.00
128.00
34.39
62.91
85.04
36.71
33.05
50.00
0.06
Concentration (mg/1)
Influent Effluent
315.39
326.29
191.61
10.75
106.50
138.00
18.00
18.00
50.00
29.50
9.50
10.20
60.38
36.00
3.80
-
42.63
49.80
50.40
20.48
64.50
6.32
7.00
24.00
25.00
0.68
36.00
2.70
10.20
8.50
20.50
3.18
3.95
3.04
-
-
3.96
2.62
3.84
1.56
Concentration (mg/1)
Influent Effluent
0.15
0.13
0.18
0.34
1.32
0.32
0.20
0.36
0.07
0.54
0.32
0.13
0.24
0.13
0.07
0.62
0.40
0.12
0.12
0.32
0.02
0.05
0.04
0.06
0.17
0.07
0.11
0.07
0.06
0.01
0.84
0.09
0.10
—
0.22
0.34
0.23
—
0.42
56.33
Concentration (mg/1)
Influent Effluent
139.10
3.74
7.80
0.94
2.42
6.96
12.40
4.40
0.32
0.06
7.80
11.40
1.77
0.75
0.02
0.03
0.22
17.80
0.11
0.17
0.02
0.10
-
0.50
0.75
1.14
2.62
8.23
0.62
8.60
11.80
17.20
18.54
29.91
0.03
0.03
5.42
7.22
6.80
1.22
(continued)
-------
TABLE 5A (CONTINUED)
Type of Carbon:
Carbon Volume:
Date:
Dichloroethane
Westvaco WVG
28.3 liters
9/10/74
1,1,2,2-Tetra-
chloroethane
Loading:
Flow Rate:
Carbon Weight:
41.6 1/min-m
3.79 1/m
11.3 kg
Methylene Chloride Trans-Dichloroethylene
Sample
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Running
Time
(Hours)
3
6
9
12
15
18
21
24
27
30
33
36
39
42
45
48
51
54
57
60
Concentration (mg/1)
Influent Effluent
2.45
0.73
1.10
0.88
0.56
0.60
1.60
1.84
0.23
1.96
3.28
2.00
2.10
2.65
0.05
0,45
0.86
1.12
0.52
0.25
0.72
0.02
6.18
0.23
0.55
0.56
1.81
2.30
4.20
8.40
4.20
4.20
2.35
4.47
0.05
0.11
3.18
3.12
6.80
11.55
Concentration (mg/1)
Influent Effluent
_
4.58
2.28
-
0.14
0.04
0.35
-
33.20
-
-
0.21
2.05
2.35
2.79
2.00
0.15
0.92
2.21
0.23
2.09
0.02
2.76
-
-
-
-
-
-
0.05
-
-
-
-
10.78
7.87
0.13
0.02
0.02
—
Concentration (mg/1)
Influent Effluent
26.54
12.60
19.38
3.22
1.34
1.37
3.23
2.22
1.31
0.23
2.14
2.10
11.45
20.37
21.84
22.04
0.26
3.93
1.48
2.10
0.65
0.19
0.15
0.33
0.18
0.06
0.23
0.24
0.39
0.31
0.38
0.38
0.39
0.28
22.82
25.20
0.24
0.24
0.30
0.28
Concentration (mg/1)
Influent Effluent
8.75
6.47
14.75
3.46
2.80
2.52
3.88
3.92
1.98
1.42
3.46
2.96
7.48
11.93
0.23
15.05
1.34
1.36
3.21
2.64
0.25
0.23
0.09
0.14
0.20
0.02
0.17
0.09
0.14
0.21
0.24
0.22
0.41
0.24
0.24
0.18
0.09
0.09
0.39
0.14
-------
TABLE 6A. EFFECT OF KEROSENE DIESEL OIL SOLVENT ON REMOVAL OF MAJOR
ORGANICS FROM STREAM 161A - Removal of EDC
*m~i~~maii^mmmmmmm***~~m~*ii~^li~**i~i*~^*mmmimmmm**i^~>^~~^-^^~<~^*~^—~t*~^—^^~'**^^~~~^^~*~~~~*^ • ^»«p«^^^^^»^»»^
Total EDC
Date
10/23/74
10/11/74
10/11/74
10/31/74
11/21/74
11/25/74
12/11/74
12/16/74
12/04/74
1/06/75
12/02/74
12/10/74
12/17/74
1/07/75
In
(mg/1)
920
193
127
208
126
300
297
931
56
316
324
1804
407
23
Out
(mg/1)
356
20
19
47
43
35
81
70
92
16
36
40
13
<1
in !• rt I ill ir —
Amount
Removed
(mg/1)
564
173
108
161
83
265
216
861
(36)
300
288
1764
394
—
— - - _ i
Percent
Removed
61
90
85
77
66
88
73
92
(64)
95
89 3.
98
97
99
—Mill •! '
Ratio
H^O: Solvent
13.3:1
13.7:1
9.1:1
9.1:1
9.1:1
5.5:1
5.5:1 /
5.5:1
5.5:1
5.5:1
7:1 & 5.5:1
3.7:1
3.7:1
3.7:1
TABLE 7A. EFFECT OF KEROSENE DIESEL OIL SOLVENT ON REMOVAL OF MAJOR
ORGANICS FROM STREAM 161A - Removal of 1,1,2-Trichloroethane
Total 1,1,2 TCE
Date
10/23/74
10/11/74
10/11/74
10/31/74
11/21/74
11/25/74
12/11/74
12/16/74
12/04/74
1/06/75
12/02/74
12/10/74
12/17/74
1/07/75
In
(mg/1)
105
355
269
127
57
88
157
155
90
73
180
135
30
177
Out
(mg/1)
16
30
36
21
8
3
6
2
14
2
<1
20
3
3
Amount
Removed
(mg/1)
89
325
233
106
49
85
151
153
76
71
—
115
27
174
Percent
Removed
85
92
87
83
86
97
96
99
84
97
99 3.
85
90
98
Ratio
H-O: Solvent
18.3:1
13.7:1
9.1:1
9.1:1
9.1:1
5.5:1
5.5:1
5.5:1
5.5:1
5.5:1
7:1 & 5.5:1
3.7:1
3.7:1
3.7:1
218
-------
TABLE 8A. EFFECT OF KEROSENE DIESEL OIL SOLVENT ON REMOVAL OF MAJOR
ORGANICS FROM STREAM 161A - Removal of 1.1.2,2-Tetrachloroethane
Date
Total Sym. Tetra. Amount
In Out Removed Percent
(mg/1) (mg/1) (mg/1) Removed
Ratio
H20:Solvent
10/23/74
10/11/74
10/11/74
10/31/74
11/21/74
11/25/74
12/11/74
12/16/74
12/04/74
1/06/75
12/02/74
12/10/74
12/17/74
1/07/75
22
197
147.9
—
22
56
32
5
39.3
59
115
49.4*
162.1
17
95
6
2
19.3
—
2
0
1
0
0
2
3
0.4*
0
0
3
16
195
128.6
—
20
56
31
5
39.3
57
112
49*
162.1
17
92
93
99
87
—
91
100
97
100
100
96.6
97.4
99
100
100
96.8
18.3:1
13.7:1
9.1:1
9.1:1
9.1:1
5.5:1
5.5:1
5.5:1
5.5:1
5.5:1
5.5:1
3.7:1 & 5.5:1
3.7:1
3.7:1
3M ••
.7:1
219
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-79-080
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
DEVELOPMENT OF TREATMENT AND CONTROL TECHNOLOGY FOR
BEFRACTORY PETROCHEMICAL WASTES
5. REPORT DATE
April 1979 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
John H. Coco, Elias Klein, Donna Howland, James
H. Mayes, William A. Myers, Earl Pratz, Clyde J. Romero,
and Floyd H. Yocum
8. PERFORMING ORGANIZATION REPORT NO.
326-429-11
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Gulf South Research Institute
P.O. Box 26518
New Orleans, Louisiana 70186
10. PROGRAM ELEMENT NO.
1BB610
11. CONTRACT/GRANT NO.
S800773
12. SPONSORING AGENCY NAME AND ADDRESS
Robert S. Kerr Environmental Research Laboratory
U.S. Environmental Protection Agency
P.O. Box 1198
Ada, Oklahoma 74820
13. TYPE OF REPORT AND PERIOD COVERED
Final 2/2/72 - 2/28/77
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This summary document presents the results of research/and development work
pertaining to the treatment of biorefractory organic pollutants emanating from petro-
chemical processing plants. Specifically, it covers application of the unit operations
of (1) carbon adsorption, (2) steam stripping, (3) solvent extraction, and (4) ozona-
tion to a number of process effluents from halogenated hydrocarbon, isocyanate, acety-
lene, and styrene manufacturing industries. In addition to these unit operation eval-
uations, a study was undertaken to develop an activated carbon from a by-product soot
which results from the manufacture of acetylene. Quantities of raw soot were dried,
pretreated, pelletized, and activated to yield a product with commercial potential
which had about 80 percent as much adsorption capacity as a similar commercial
product.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Activated Carbon
Activated Carbon Treatment
Activated Sludge Process
Distillation
Ozonation
Solvent Extraction
Steam Distillation
Stripping
carbon adsorption
steam stripping
ozonation
biological treatment
petrochemical wastes
refractory waste
680
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report)
Unclassified
RELEASE TO PUBLIC
21. NO. OF PAGES
236
20. SECURITY CLASS {Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77)
PREVIOUS EDITION IS OBSOLETE---.
U. S. GOVERNMENT PRINTING OFFICE: 1979 — 657-060/5325
------- |