United States
Environmental Protection
Robert S Kerr Environmental Research EPA-600/2-78-103
Laboratory May 1978
Ada OK 74820
Research and Development
vvEPA
Thermal
Regeneration Of
Activated Carbon
-------
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-600/2-78-103
May 1978
THERMAL REGENERATION OF ACTIVATED CARBON
by
Louis Hemphill
Department of Civil Engineering
University of Missouri-Columbia
Columbia, Missouri 65201
Grant No. S800554
Project Officer
Jack Hale
Treatment and Control Technology Branch
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 endorse-
ment or recommendation for use.
-------
FOREWORD
The Environmental Protection Agency was established to coordinate ad-
ministration 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 in-
formation about environmental problems, management techniques and new tech-
nologies 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) investi-
gate the nature, transport, fate and management of pollutants in ground-
water; (b) develop and demonstrate methods for treating wastewaters with
soil and other natural systems; (c) develop and demonstrate pollution
control technologies for irrigation return flows, (d) develop and demon-
strate pollution control technologies for animal production wastes;
(e) develop and demonstrate technologies to prevent, control or abate
pollution from the petroleum refining and petrochemical industries, and
(f) develop and demonstrate technologies to manage pollution resulting
from combinations of industrial wastewaters or industrial/municipal
wastewaters.
m
-------
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 pro-
vide adequate protection for the American public.
(Signed) W. C. Galegar
legar
IV
-------
ABSTRACT
A three part experimental study of activated carbon sorption and ther-
mal regeneration has been completed. The initial plan of this study pre-
dicated that two 1 gallon-per-minute flow capacity pilot-plants would be
fabricated and used to collect samples of petrochemical waste stream or-
ganic materials in the Baton Rouge, Louisiana area. Samples of the pilot-
plant sorbent (activated carbon saturated with petrochemical waste) were
to be used in an experimental thermal regeneration study. The purpose of
the thermal regeneration experimental study was to determine the specific
thermal regeneration characteristics of the petrochemical waste saturated
activated carbon sorbent. Time requirements requisite to fabrication,
operation, and maintenance of the sorption pilot-plants prompted revision
of the initial study plan. Appropriate to the revised plan, two 1 gallon-
per-minute pilot plants were fabricated and provided for operation. In
addition, a laboratory study of the specific sorption characteristics of
selected petrochemical waste materials, via batch isotherm and flow columns,
was completed along with a low pressure thermal regeneration investigation.
Results of these studies showed:
1) Small polar molecular species, or species highly soluble in water,
are resistant to carbon sorption.
2) Extreme acidic or basic waste streams may require pH adjustment
to promote carbon sorption.
3) Batch isotherm values provide basic information relative to
activated carbon-petrochemical waste column design.
4) Vacuum regeneration of petrochemical saturated activated carbon is
effective and efficient. With most sorbent-sorbate combinations
tested, the carbon sorbent could be regenerated to 95 percent of
original sorption capacity.
5) Temperature required for carbon regeneration was a function of
waste type (composition) and sorption capacity.
-------
The report describes experimental methods of differential thermal
analysis, vacuum thermal regeneration, isotherm and column derived sorp-
tion values, and quantitative relationships of temperature and thermal
regeneration response.
VI
-------
CONTENTS
Page
Foreword iii
Abstract v
List of Figures viii
List of Tables x
Acknowledgements xi
Section 1 - Introduction 1
Section 2 - Conclusions 4
Section 3 - Recommendations '. 6
Section 4 - Design and Construction of Sorption Pilot Plants ... 7
Section 5 - Literature Survey 16
Section 6 - Activated Carbon - Petrochemical Waste Sorption ... 25
Section 7 - Vacuum - Thermal Regeneration of Activated Carbon . . 62
Section 8 - Differential Thermal Analysis of Activated Carbon
Regeneration 83
Section 9 - Methods, Materials, and Procedures 86
Section 10 - Results . 93
Section 11 - Summary 100
References 101
vn
-------
LIST OF FIGURES
Number Page
1. Block Diagram etc. of Activated Carbon Sorption Pilot Plant
Showing Flow Directions and Valve Location 9
2. Pressure vs. Flow Rate Plot for Experimental Pilot Plant ... 10
3. Pressure vs. Flow Rate Plot for Experimental Pilot Plant ... 11
4. Photograph of Completed Pilot Plant Ready for Shipment to
Gulf South Research Institute - Baton Rouge, Louisiana .... 12
5. Types of Sorption Separations - After Weber 38
6. The Influence of Petrochemical Waste on Adsorption
Characteristics 40
7. The Influence of Activated Carbon on Adsorption
Characteristics 42
8. Results of Isotherm Study for Waste I and Carbons A, B and
D in the Linearized Form of Freundlich Isotherm Model .... 43
9. Results of Isotherm Study for Waste II and Carbons A, B, C
and D in the Linearized Form of Freundlich Isotherm Model . . 44
10. Results of Isotherm Study for Waste III and Carbon A, B, C
and D in the Linearized Form of Freundlich Isotherm Model . . 45
11. Results of Isotherm Study for Waste IV and Carbons A, B, C
and D in the Linearized Form Freundlich Isotherm Model .... 46
12. Linearized Form of Freundlich Isotherm Model for Wastes
Exhibiting Favorable Adsorption, Waste II and Unfavorable
Adsorption, Waste II and Carbon B 47
13. Linearized Form of Langmuir Isotherm Model for Wastes
Exhibiting Favorable Adsorption, Waste II and Unfavorable
Adsorption, Waste III with Carbon B 48
14. Linearized Form of BET Isotherm Model for Wastes Exhibiting
Favorable Adsorption, Waste II and Unfavorable Adsorption,
Waste II with Carbon B 49
15. Results of Packed Column Study for Waste II and Carbon B . . . 51
16. Results of Packed Column Study with Waste IV and Carbon A . . 53
17. Effect of Temperature on Regeneration for Waste Il-a and
Carbon B 57
viii
-------
Figure Page
18. Effect of Temperature on Regeneration of Waste IV and
Carbon B 60
19. Experimental Carbon Regeneration Equipment 67
20. Effect of Specific Sorption Capacity on Regeneration 70
21. Effect of Specific Sorption Capacity on Regeneration 72
22. Temperature Rate Coefficient versus Temperature 74
23. Comparison of MT and Regeneration Response 75
24. Methylene Blue Regeneration Efficiency 77
25. Iodine Regeneration Efficiency 79
26. Petrochemical Waste Regeneration Efficiency ... 80
27. Chlorophenol Regeneration Efficiency 81
28. Diagram of DTA Apparatus 88
29. Effect of Low Pressure (10 mm Hg) Degassing Period on High
Temperature Exothermic Reaction of Activated Carbon 89
30. Quantitative Response of DTA - Based on Benzoic Acid
Standard 92
31. Phenol Sorbed Activated Carbon DTA Thermogram 94
32. Regenerated Sorption Capacity Developed by DTA Study of
Phenol-Activated Carbon System 96
33. Methylene Blue Sorbed Activated Carbon DTA Thermogram .... 98
34. Regenerated Sorption Capacity Developed by DTA Study of
Methylene-Blue Activated Carbon System 99
IX
-------
LIST OF TABLES
Number Page
1. Flow Rate vs. Pressure Values for Experimental Sorption
Pilot Plant 8
2. Physical Properties of Activated Carbon Used in the Study ... 26
3. Results of Analysis of Raw Waste for Organic and Total
Carbon 29
4. Results of Isotherm Study 35
5. Sorption Characteristics of Wastes Studied 39
6. Summary of Results of Three Packed Column Studies for Waste
IV, Carbon A 54
7. Results of Regeneration Study; Waste II A, Carbon B 56
8. Coefficients and Exponents of Freundlich Isotherm Model
Equation 58
9. Results of Regeneration Study - Waste IV, Carbon B 59
-------
ACKNOWLEDGMENTS
In recognition of services which directly contributed to this project,
the following individuals merit acknowledgment.
Dr. John T. Novak for assistance in planning the study.
Mr. Mark Valentine and V. Ramaiah for assistance with the sorption
and vacuum thermal regeneration study.
Mr. Dayne Howard for assembly and calibration of the sorption pilot
plants.
Mr. Delbert Morton and Mr. William Ballard for fabrication of the
sorption pilot plants.
xi
-------
SECTION I
INTRODUCTION
The hydrocarbon based petrochemical industry in the United States is
relatively young. Essentially this industry was initiated during the World
War I era (ca. 1918-1920) and showed a relatively stable growth and devel-
opment during the 1920-1940 period. The petrochemical industry expanded
dramatically during World War II for the production of synthetic rubber and
plastic materials. Following World War II the industry expanded rapidly,
a five fold increase in production during the period from 1945 to 1960. The
factors contributing to rapid expansion of the petrochemical industry were:
favorable economic conditions, availability of economical and abundant sup-
plies of petroleum raw material, and popular demand for the finished pro-
ducts, particularly synthetic fabric and plastic materials.
Generally, petrochemical production plants were located in geographi-
cal areas near petroleum production. Consequently, the highest density of
petrochemical plants are found in the coastal regions of southern California
and the gulf coast portion of Texas and Louisiana. Approximately 80 percent
of the United States petrochemical production plants are located in the
gulf coast between New Orleans, Louisiana, and Brownsville, Texas.
One of the problems common to many rapidly developing industries is
waste treatment and disposal. This problem is not unique to the petro-
chemical industry; however, the dominant characteristic of the petrochemical
industry - the continuous expansion and increasing variety of products and
wastes produces - is unique. The basic characteristics of petrochemical
wastes have been established; however, recent findings, relative to the po-
tential biological hazards of petrochemical waste products, shows that more
comprehensive knowledge about production, nature, and specific composition
-------
of wastes must be developed. Concurrent with this demand is the associated
application - developing effective methods of waste treatment. Petrochemi-
cal wastes, a common category, includes an ever expanding list of compounds,
materials and mixtures. Because of this variety, it is necessary to consi-
der each waste component on a molecular basis. Many of the components in
petrochemical waste streams are present in low concentration - yet the chem-
ical and physiological significance of these low concentration waste compo-
nents must be considered. Many of the low concentration organic components
in petrochemical waste exhibit low dose toxicity effects and are subject to
biomagnification in the environment.
Ecologically, petrochemical wastes constitute a major hazard since
waste materials contain relatively large amounts of non-biodegradable and
toxic materials which may be discharged continuously. The latter factor
is particularly important. Slow continuous discharge of a pollutant may
produce a subtle change in the environmental structure, or response, which
predisposes significant alteration of biota. The most direct evidence for
this situation in man would be increased incidence of cancer and genetic
anomalies.
One method of waste treatment which has direct application to petro-
chemical waste treatment is activated carbon sorption. Activated carbon
is an established waste water treatment reagent which has been used exten-
sively in the water treatment and waste water treatment industries. Acti-
vated carbon is a scavenging sorbent which exhibits broad range sorption
affinity and high capacity for many organic and inorganic materials. In
addition, activated carbon is readily availabe in granular form which can
be incorporated into sorption columns for continuous on stream utilization.
The basic interest in this study was directed toward determination of
-------
activated carbon sorption affinity and capacity for petrochemical wastes
and evaluation of thermal regeneration response. These factors are rele-
vant to the effectiveness and efficiency of the activated carbon sorption
process and are requisite to design of waste treatment facilities.
Tentatively, the study program planned for this investigation envi-
sioned a three-step sequence: design and fabrication of two 1 gallon-per-
minute activated carbon pilot plants; application of the pilot plant to
the Baton Rouge, Louisiana area petrochemical plant waste streams; and
laboratory evaluation of activated carbon thermal regeneration. Time re-
quired for construction and maintenance of the sorption pilot plants was
greater than expected and a revised schedule of study was developed. In
the revised program, the pilot plant operation was conducted as a separate
independent study. The balance of the investigation was conducted in the
laboratory.
The basic objectives of this Investigation were:
1) To design and construct two one-gallon-per-minute activated
carbon sorption pilot plant units for application to petro-
chemical waste streams in the Baton Rouge, Louisiana area.
2) To determine the sorption affinity and capacity of various
commercially available activated carbons for representative
samples of petrochemical waste materials, and
3) To determine the thermal regeneration characteristics of
spent activated carbon - petrochemical waste mixtures.
4) To determine the effectiveness and application of differen-
tial thermal analysis (DTA) for detailing thermal reactions
of activated carbon experienced during low-pressure regen-
eration.
-------
SECTION 2
CONCLUSIONS
Results of the vacuum thermal regeneration study show that vacuum ther-
mal regeneration of activated carbon-waste sorbate samples proceeds as an
exponential regeneration response. Generally, the minimum temperature re-
quired for regeneration threshold temperature is characteristic of the waste
sorbate-sorbent combination.
The dominant factors influencing vacuum thermal regeneration response
were identified experimentally as sorbate nature, sorption loading and tem-
perature. The specific influence of each factor was determined separately
and collectively. Results of these studies showed that:
1) Vacuum thermal regeneration of activated carbon-sorbate samples
is effective for a wide variety of sorbates.
2) The nature of the sorbate, in terms of thermal degradation and
volatility, determines the rate of regeneration response.
3) The specific differences in carbon sorbents do not signifi-
cantly influence regeneration response.
4) The dominant factors which influence regeneration response
are specific sorption capacity and temperature.
5) The general response pattern can be described by Re = KTn.
The above relationships are significant because the specific influence
of temperautre and loading are identified as major factors. Since the basic
experimental values used in this study were derived from static experiments,
using equilibrium temperature data and vacuum conditions, cognate factors
such as heat transfer rate and mass transfer rates were not relevant.
Activated carbon can be regenerated under vacuum conditions. Experi-
mental results developed in this study showed that vacuum regeneration, per
se, is effective in restoring activated carbon sorption properties. This
finding indicates that the carrier gases used in conventional thermal
4
-------
regeneration are probably not requisite to thermal regeneration but serve
to implement heat and mass transfer and control oxidation.
Results of the petrochemical waste-activated carbon sorption study
showed that:
1) The process of adsorption on activated carbon is not uniformly
effective for all petrochemical waste components particularly
those small short chain highly oxygenated organic compounds.
2) Adsorption depends on the nature of the components of the waste,
and in general, as the average molecular weight increases better
adsorption occurs.
3) No one carbon exhibited the best results, and the choice of car-
bon depends on the waste as well as the waste concentration range.
4) The Freundlich Isotherm Model described the process of adsorption
in petrochemical waste more accurately than the Langmuir or BET.
5) In the column study the characteristic breakthrough curve was ex-
hibited by only the waste which exhibited a favorable adsorption
pattern.
6) For a waste that exhibits a favorable adsorption pattern the op-
timum temperature of regeneration is in excess of 1000°F; whereas,
for a waste that exhibits an unfavorable adsorption pattern it is
much lower, in the order of 500°F.
7) Optimum temperature of regeneration must be established for each
carbon-waste mixture.
8) Independent investigation of activated carbon by differential ther-
mal analysis (DTA) established that DTA is a practical and useful
analytical tool. Specifically, the results of the DTA study showed
that: special techniques are required for development of realistic
therniograms; low pressure degassing is necessary for removal of
sorbed oxygen; and each activated carbon-sorbate system exhibits a
unique thermal regeneration response.
-------
SECTION 3
RECOMMENDATIONS
Regeneration of waste water sorbents will become an important part of
the cost of waste treatment. Sorbent removal of dissolved organic and in-
organic pollutants will probably become a popular or standard design prac-
tice and the cost of recharging the sorbent may well determine the appli-
cability of sorption removal. Activated carbon is the most popular waste
water treatment sorbent, yet its application is limited and the cost of
regeneration is relatively high. Results presented in this study show that
vacuum regeneration of activated carbon is promising. Potentially, vacuum
regeneration of activated carbon could be achieved in place. This feature
would probably reduce the cost of handling and shipping. In addition,
there is some experimental evidence that shows that vacuum thermal regener-
ation might be less expensive than conventional multiple hearth furnace or
rotary kiln regeneration.
Further study of the basic factors involved in carbon regeneration are
required.
-------
SECTION 4
DESIGN AND CONSTRUCTION OF SORPTION PILOT PLANTS
Determination of sorption affinity and capacity is the first step ef-
fort requisite to design of full scale activated carbon wastewater treat-
ment facilities. Usually the activated carbon-waste affinity and sorption
capacity are determined by small-scale laboratory tests using batch iso-
therm and column study techniques; however, the nature of petrochemical
waste streams usually warrant direct onstream evaluation. Many of the
petrochemical waste streams, common to production facilities, contain
volatile compounds, are aggressively corrosive, and change composition
rapidly. In an attempt to minimize these problems and provide a realis-
tic evaluation of the activated carbon system, two one-gall on-per-minute
activated sorption pilot plants were designed, constructed, and readied
for field application.
The activated carbon sorption pilot plants designed and constructed
for this study, Figure 1, consisted of four "Lucite" columns connected
with plastic tubing. A pH monitoring flowthrough chamber and temperature
probe were located in series with the 1/4 H.P. centrifugal pump. The
pilot plant was designed to be operated in upflow or downflow condition.
A picture of the assembled pilot-plant ready for operation is shown in
Figure 4.
The assembled pilot-plants were operated in the laboratory at various
flow rates, and activated carbon loadings, to determine hydrualic charac-
teristics. Results of this study are presented in Table 1 and Figures 2
and 3. A description of the pilot plant components are listed below.
-------
TABLE 1
FLOW RATE VS. PRESSURE VALUES FOR EXPERIMENTAL SORPTION PILOT PLANT
Column No.
0
1
Pressure Gauge
Sheave
3.0
7.0
12.5
18.5
Sheave
1.0
9.5
17.5
Sheave
1.0
4.0
8.2
15.0
18.5
Sheave
0.2
4.7
7.9
14.0
size 6"
0.0
5.0
11.0
17.0
2
Reading (psi)
x 2 1/2" Belt
0.0
3.8
10.0
16.0
size 6" x 1 1/2" Belt
0.0
7.8
16.2
size 12"
0.0
1.2
6.5
13.2
17.5
size 12"
0.0
2.4
6.6
13.0
0.0
7.2
15.6
x 2 1/2" Belt
0.0
0.6
6.0
13.0
16.7
x 1 1/2" Belt
0.0
1.9
6.2
12.7
3
20"
0.0
2.5
9.0
15.0
19"
0.0
6.5
15.0
43"
0.0
0.2
5.2
12.2
16.2
43"
0.0
1.3
5.8
12.2
4
0.0
1.2
8.0
14.0
0.0
6.2
14.5
0.0
0.0
5.0
11.9
15.8
0.0
1.2
5.6
12.0
Flowrate
(gpm)
0.97
0.94
0.94
0.92
0.70
0.69
0.69
0.73
0.70
0.69
0.64
0.60
0.39
0.28
0.23
0.13
8
-------
Backwash
To 1
Drain ..
t
Sample „ f-y-
In £Z\
Pump
Backwash-*-
t
WM>1
Tt
Carbon
,
?
Carbon
i^viK
^SJrJJ^J
-rf
i 1X1
A_
Carbon
E^^^^S
u
•Xi
T
Out-To- Drain
Carbon
^^R
gg
S
4" Height
Shot Gravel
Filter Meth
Strainer
J. .»«* Output
t^ .5 Gal. Per Mm
ample
Valve
Figure 1. Block diagram of activated carbon sorption pilot-plant showing flow directions
and valve locations.
-------
1.0
0>
05
U.3
0.0
Sheave Size 6" x 2-1/2"
I
I
10 15
Pressure, psi
20
1.0
^
u_
0.5
0.0
Sheave Size 6% 1-1/2"
I
10 15
Pressure, psi
20
25
Figure 2. Pressure vs. flow-rate for experimental pilot-
plant.
10
-------
1.0
0.5
0.0
Sheave Size \Z" x 2-1/2"
5 10 15
Pressure (psi)
20
1.0
6
°-
0>
0.5
0.0
Sheave Size 12" x 1-1/2"
10 15
Pressure (psi)
20
Figure 3. Pressure vs. flow-rate for experimental
pilot-plant.
11
-------
Figure 4. Photograph of completed pilot plant ready for shipment
to Gulf South Research Institute, Baton Rouge, Louisiana,
12
-------
1) Variable speed pump
Pump speed and liquid flow rate are controlled by motor drive and pump
pulley sizes. Pulleys supplied with the pump provide flow rates ranging
from 0.39 to 0.97 gallons per minute. The tubing provided with this unit
is "Viton" tubing.
2) Electric motor
A 110 volt A.C. 1/4 H.P. (explosion proof) totally enclosed motor is
provided with the pump.
3) Filter column
4) Adsorption columns
The assembly consists of four 5" I.D. x 72" long "Lucite" plastic columns,
in series array, interconnected with 1/2" diameter poly-proprophylene tubing.
5) Temperature and pH monitor
A flow through chamber containing pH and temperature sensors is provided.
Flow-through chamber provides opportunity to locate the pH and temperature
sensor at any point in the pilot plant. A dual channel "Rustrak" recorder is
provided with the instrument.
The activated carbon sorptlon pilot plant was pre-packaged and delivered
1n a ready-to-be-used condition. All of the major units such as, pump and
drive motor assembly, columns, valves, pressure gauges, and interconnecting
tubing, were assembled. The location of the pH-temperature sensor and orien-
tation of flow, i.e., upflow or downflow, were variable and optional. Spe-
cific instructions about the pilot plant were enclosed with each unit along
with a notice that the packing case was designed and constructed as a reusable
shipping container.
Preparation for on-stream operation of activated carbon sorptlon pilot
plant involved the following tasks:
1) selection of flow regime (upflow or downflow) and flow rate, and
2) preparation of carbon sorbent
Instructions for flow rate selection were Included along with instructions
for valve and tubing connections for upflow and downflow conditions.
13
-------
Enclosed instruction suggested that activated carbon sorbent is pre-
pared for column application by wetting and soaking. This operation could
be accomplished in small batches or completely in one large mixing vessel.
It was recommended that the activated carbon mix be permitted to soak for
24 hours prior to charging the columns.
Columns were prepared for application by placing a double layer of
fiberglass window screen material (about 1/16 mesh) in the bottom of the
column followed by a 4 to 6 inch layer of pea gravel.
Adsorption column preparation is completed by adding the carbon water
slurry to the columns. Approximately 3 1/2 feet of carbon is used in each
column. In case of very turbid water or water with high solids content it
is recommended that the depth of carbon in the first column be reduced to
2 3/4 feet. This condition prevents rapid plugging of the first column.
Each of the adsorption columns was equipped with a pressure gauge.
This device is particularly useful when a high turbidity, or high solids,
effluent waste stream is being investigated. Under these conditions, it
is probable that one column may become partially plugged or saturated with
debris. This condition will generally produce a relatively high pressure
drop across one of the columns which will be indicated by a high pressure
gauge reading.
A Dole flow control valve limits the backwash flow to approximately
4 gpm. This should be sufficient for cleaning the carbon and sand in the
five columns. A 30 to 50 percent bed expansion indicating that the back-
wash flow quantity is sufficient.
Whenever the carbon has been allowed to drain and has been exposed
to air for ten minutes or more, it should be backwashed.
14
-------
Whenever the pressure In Column 1 is 18 psi or more, the columns
should be backwashed.
The three way valves have no positive stop, thus, they must have
plugs inserted during backwash or they will leak.
15
-------
SECTION 5
LITERATURE SURVEY
Advanced waste treatment studies, using activated carbon sorptlon 1n
large scale study, were Initiated In 1960. Since then, the application
of activated carbon adsorption to the reclamation of waste water has been
subjected to continuous Investigation. Prior to 1960, activated carbon
sorption was considered to be a final clean-up polishing operation fol-
lowing secondary biological treatment. However, by employing this com-
mon treatment approach, some investigators (1,2) have reported that leak-
age difficulties were experienced with the removal of refractory organic
materials from the secondary effluent by carbon adsorption process. The
nature of organic leakage 1s thought to be comprised partially of non-
adsorbable cell fragments produced during the secondary biological stage
and small organic molecules which have been extensively hydrolyzed in the
biological treatment system. Because of the leakage problem, Weber et al.
(3), attempted a new treatment scheme which consists of conventional pri-
mary treatment, chemical coagulation and clarification, filtration, and
carbon adsorption. Weber, and his co-workers reported that an exception-
al high degree of removal of BOD and suspended solids were achieved using
this scheme. Phosphorous and organic nitrogen were removed by chemical
coagulation, and nitrates were removed in the carbon adsorber. This pro-
cess 1s usually considered to be tertiary treatment and the effluent qual-
ity is high, perhaps fit for Human consumption. The advantages of such a
process, over the conventional biological processes, are: it is capable
of intermittent operation, responds rapidly to shock load variation, easy
to operate, amenable to instrumentation and automation, provides removal
of high molecular weight biologically resistant substances and toxic
16
-------
organic waste materials, and requires small space for operation.
The carbon-sorption tertiary treatment has been successfully employed,
in large scale operation, in South Lake Tahoe (4). The plant efficiency, in
terms of removal values for various components, was: BOD-99.4 percent, COD-
96.4 percent, MBAS-99.9 percent, P-99.1 percent, suspended solids, color,
coliform bacteria-99 percent, turbidity-99.9 percent. Obviously the quality
of the effluent influences cost. This tertiary treatment scheme (5) is re-
ported to cost more than twice the conventional biological treatment. How-
ever, it represents an improvement in treatment; it is a step ahead of bio-
logical treatment.
Many investigators have experimented with process schemes similar to
that of Weber et j»l_., with and without modifications. Cohen and Kugelman
(6) designed an almost identical process scheme with packed column carbon
adsorber operating at a hydraulic loading of 2 to 8 gpm per square foot
and a contact time of 30 to 60 minutes. Their investigations, with muni-
cipal waste, revealed that as the contact time increased from 0 to 30 min-
utes the rate of absorption increased; between 30 and 60 minutes of con-
tact time the rate of adsorption remained almost constant; but higher than
60 minutes contact time resulted in a decreased adsorption rate. The re-
moval of total organic carbon (TOC) and COD were reported as 0.15 to 0.13
pounds and 0.4 to 0.6 pounds, respectively, per pound of activated carbon.
They also estimated the cost of treating 1000 gallons of wastewater as 24
to 36£ for a 6 MGD plant, 18 to 27$ for a 10 MGD plant, and 9 to 15$ for
a 100 MGD plant.
Shuckrow et aj_., (7) modified their process scheme for treating muni-
cipal waste by incorporating a two stage carbonation process following
clarification. The carbonation step is followed by mixed media filtration,
17
-------
carbon adsorption, and disinfection. The hydraulic loading and superficial
detention time, for the carbon column, were 20 gpm per square foot and 30
minutes respectively. They have reported that pH influences the rate of
adsorption.
Friedman eit a\_., (8) designed a scheme with a sequence of operation of
coagulation, clarification, and carbon adsorption for 75 percent domestic
and 25 percent industrial waste. Villiers (9) recommended a series of lime
clarifiers, dual media filter, clarifiers, and carbon adsorption to treat
municipal waste. Bishop (10) also investigated the application of physioco-
chemical treatment process, for municipal wastewater, with an automated pilot
system consisting of cyclone degritting, two-stage high pH lime precipitation
with intermediate recarbonation, dual media filtration, pH control, selective
ion-exchange, and downflow granular carbon adsorption. Peoples et jil_., (11)
used a combination of sand filter and activated carbon adsorption for refin-
ery waste and reported the following removal values; suspended solids-62 per-
cent, BOD-85 percent, TOC-65 percent, oil-85 percent, and phenol-99 percent.
Huang and Hardie (12) evaluated the use of carbon adsorption process for the
treatment of acidic and strongly organic chemical wastes. Using both batch
and column techniques, they found that the process could effectively remove
COD along with organic nitrogen and turbidity. However, the process did not
remove phosphorus, and the pH had no influence on adsorption.
Activated carbon 1s available for use in two forms, granular and powder.
The unit operations and the process characteristics are different for the two
forms of carbon, and each has its advantages and disadvantages. The granular
carbon has high absorption capacity (depending on application) and can be re-
generated by the thermal method, but it is more expensive and requires more
sophisticated technology. On the other hand, powdered carbon is much less
18
-------
expensive than granular form. The small particle size reduces the depen-
dence of adsorption on intraparticular transport and requires a simple unit
operation, but regeneration methods are not well developed.
Kropp and Gitchell (13) designed a two stage counter current powdered
carbon adsorption scheme with clarifiers after each stage for flocculation
and removal of the carbon, followed by sand filtration for the secondary
treatment of the domestic waste water. Shell and Burns (14) designed a
scheme in which the chemical treatment was followed by powdered carbon
adsorption and granular media filtration. The effluent characteristics
were reported as; suspended solids-3 mg/1, total phosphorus as P-0.3 mg/1,
total organic carbon-12 mg/1, COD-2 mg/1, BOD-10 mg/1, and pH-7.0 to 8.0.
The cost was estimated as 18 to 22£ per 1000 gallons for a 10 MGD plant.
A key to the economic use of the adsorption process is the regenera-
tive capacity of the spent carbon. One theory predicts that in the pro-
cess of regeneration the kinetic energy of the adsorped impurity is in-
creased to overcome the surface bond energy by subjecting it to a driving
force, usually thermal energy. Thermal regeneration is thought to occur
in three steps: first, drying at approximately 220°F, followed by pyroly-
sis of adsorbed molecules at 500-1550°F, and activation at 1600 to 1700°F
(15). At such high temperatures the chemical reactivity of the activated
carbon is also enhanced and if the atmosphere is not carefully controlled,
the carbon itself may be lost due to chemical reaction. The proper con-
trol of the atmosphere is critical for effective regeneration. From many
experiments the conductive atmosphere to minimize carbon losses are found
to be vacuum, nitrogen, and steam. In normal industrial practice, careful
operation could minimize carbon losses to 5 to 10 percent of the original
quantity. As far as the economics of regeneration are concerned, the
19
-------
capital and operating cost of regeneration vary substantially with the
amount of carbon to be regenerated, type of industry, the spent carbon
handling characteristices, method of collection and original carbon ac-
tivity.
In a laboratory scale experiment, -Juhola(i5) regenerated spent
granular carbon, from a tertiary treatment of wastewater, with flue gas
and steam activation. The data indicated that the accumulation of alkaline
and iron oxide ash catalyze the oxidation of carbon. Leaching with hydro-
choloric acid was found to remove the metallic elements and allow regenera-
tion to proceed with less destruction of carbon. Uraono (16) used a 0 to 10
percent nitrogen atmosphere for the regeneration of spent carbon from a sew-
age treatment plant. Westvaco Corporation (17) developed and commercialized
a method using steam activation for the regeneration of powdered activated
carbon. A 5 to 10 ton per day system is estimated to cost $400,000 and
$650,000, respectively, and the operating costs are estimated to vary from
2.3 to 5t per regenerated pound of carbon, depending on unit size. Kampt
(18) used a multiple hearth furnace for thermal regeneration of granular
carbon. The first two hearths of the furnace were used for drying, the
third and fourth for the pyrolysis of volatile matter, and the last two for
the removal of remaining impurities and reactivation, with a total detention
time of 30 minutes. The loss of carbon during reactivation averaged 5 to 7
percent. Knopp and Gitchell (13) reported that the composition of carbon
gradually changed as it was reused. This was probably due to certain ele-
ments of original structure which are preferentially removed by partial oxi-
dation in the regeneration, or to materials which may be deposited from
wastewater by precipitation, and subsequent to oxidation, remain as ash. In
general, the inorganic content was found to increase. The amount of phosphorus
20
-------
Increased 36 times due mainly to the precipitation of calcium, iron, or
aluminum phosphates. The amount of carbon decreased to about 76 percent
and the composition of carbon changed from C,, g Hg .0 to C^ g H~ gO.
A second method of regeneration is chemical regeneration. Shribaand
Purcupile (19) developed a chemical regeneration method called immisible
fluid displacement. A solvent, such as carbon disulfide, is contacted with
the spent carbon subjecting the adsorbed molecules to a driving force due
to their solubility in the solvent. However, the driving force provided by
the solubility is far weaker than the temperature gradient in thermal regen-
eration. Moreover, since the driving force is due to the solubility of the
solute in the solvent, desorption can be very selective requiring more than
one solvent to treat a waste normally encountered in industry. This method
of regeneration can normally be used to recover a particular constituent of
the waste. Yet another method of regeneration, primarily of academic interest,
is electro-chemical regeneration (20).
Adsorption processes finds wide application in wastewater treatment for
many industrial wastes. Azuma (21) treated wastewater from the plating in-
dustry to remove chromates and cyanides. Kalinske (22) enhanced the biologi-
cal oxidation of organic waste in microbial suspension by the addition of ac-
tivated carbon, for microbial activity. DuPont (23) developed a method in
which the addition of powdered carbon to the aerator of a biological treat-
ment resulted in effluent as good as tertiary treatment. The activated car-
bon adsorption process has been used to treat the wastewater from coke plants
(24) and textile plants (25,26). The process has also been applied to oily
wastes (27) and radioactive wastes (28).
Utilization of activated carbon for removal of organic waste materials
from domestic and industrial wastes is usually provided by sorption columns,
21
-------
generally the waste stream contacts the carbon column at a predetermined
rate in either a downflow or upflow configuration. The contact sorption
process is effective in removing waste materials from the waste stream and
produces waste saturated carbon or spent activated carbon. In order to re-
store the sorption capacity of the spent carbon, it is necessary to regen-
erate the carbon. Methods available for regenerating activated carbon are
varied and range from simple displacement reactions to complex high temper-
ature reactivation processes.
Most activated carbon regeneration methods consist of processes directed
towards removing the sorbate from the spent carbon. Consequently, practical-
ly any process which effectively removes the sorbate, either by removal, per
se, or by initiation of a physical or chemical reaction which culminates in
removal of sorbate material without disturbing the activated carbon stucture
might be an effective regeneration process. Recently, it has been demon-
strated that sewage saturated activated carbon can be effectively regener-
ated by biological oxidation (29, 30).
Inherent in most on-going carbon regeneration methods Is the theory
that the sorption nature and capacity of activated carbon is an Intrinsic
property developed by activation which Is diminished by the sorbate loading.
If this is the case, the specific sorption capacity of activated carbon sor-
bent is not depleted by utilization, and the original sorption capacity can
be reestablished by regeneration. In this context, regeneration and sorp-
tion are considered reversible processes.
It is difficult to generalize about the regeneration properties of
activated carbon since it is a proprietary material, however, it Is gener-
ally recognized that high temperature and high pressure treatment, together
with selected chemical reagents, are used to achieve activation during
22
-------
manufacture. Perhaps for this reason, the most popular method of activated
carbon regeneration is high temperature treatment. Currently, high temper-
ature regeneration of activated carbon is employed and temperatures ranging
from about 200 to 1800°F are used to achieve 95 to 100 percent regeneration
(31). The high temperature regeneration process requires either a multiple
hearth furnace or rotary kiln equipment. In practice a carrier gas is
forced through the carbon waste material during thermal regeneration. The
purpose of the carrier gas is two-fold: (1) to provide efficient convec-
tive heat transfer to the granular carbon and (2) to protect the carbon
and furnace or kiln from burn out. The carrier gas is generally formulated
to contain Hp, Np, H^O and small amounts of oxygen, and serves as a mild
oxidant.
Since the process of thermal regeneration of spent activated carbon is
very similar to the process of activation used in preparation of this mater-
ial, consideration of thermal activation response is appropriate. A compar-
ative study of nine activated carbons, representative of nine different acti-
vation processes, showed that the final product exhibited dissimilar specific
sorption capacities for methylene blue, phenol, iodine, molasses, and caramel
sorbates (31). This finding suggests that the thermal activation process may
determine the sorbate-sorbent affinity and the sorption capacity of the acti-
vated carbon. This was verified in an additional study wherein a common pine
wood char was activated using high temperature air and steam carbonization
treatment (31). Results of this study showed that the sorption affinity and
capacity of pine wood chars was determined by the activation process.
The role of the activating gas has been identified as a specific mild
oxidizing agent. Hassler (31) states "When proper activation conditions are
23
-------
provided, the oxidizing action of the activating gases does not consist of
an indiscriminate removal of successive layers of atoms from the surface".
Consequently, oxidation of a sorbent surface probably proceeds as selective
errosion of the surface which increases surface area and porsity and gener-
ally improves the specific affinity of the sorbent surface.
The most interesting features of thermal activation, relative to this
investigation, is the pattern of sorption capacity increase with increase
in activation time exposure. Results of two studies (32) show that there
is a dissimilar response in sorption capacity increase using malachite green,
methylene blue, molasses, and phenol. Malachite green and phenol produced
a smooth convex curve, i.e., the sorption capacity increased rapidly at small
activation times and then leveled out. Methylene blue exhibited a linear re-
sponse and sorption capacity increased about 12 percent for each 10 minutes
of activation exposure. All of the other sorbate materials, aniline blue
and molasses, showed a smooth concave curve type of response wherein the
sorption capacity increased rapidly after about 40 minutes of activation
exposure. The most interesting result of this study was the convergence of
all curves to a common maximum point after about 50 minutes activation ex-
posure.
Unfortunately, the specific experimental conditions used in the above
studies were not described, yet it seems probable that the activated time
basis was probably developed by limiting the exposure time of a carbon sam-
ple. If this were the situation, then the activation time basis is very
similar to regeneration time, and/or temperature, used in this study and
the opportunity for comparison is provided.
24
-------
SECTION 6
ACTIVATED CARBON-PETROCHEMICAL WASTE SORPTION
The purpose of this portion of the study was to determine the speci-
fic sorption characteristics of four petrochemical liquid wastes and to
determine the vacuum thermal regeneration response of the waste carbon
combination.
Four petrochemical wastes, provided by Gulf South Research Institute,
were used to determine the sorptive capacity of four representative acti-
vated carbons. The physical properties of the activated carbons, labeled
A, B, C, and D are presented in Table 2.
The petrochemical wastewater analyses, labeled I, II, III, and IV are
representative of typical petrochemical industries. The chemical compo-
nents for each waste were different depending on the products and processes
of the industry. Efforts were not made to identify each component present
nor to analyze quantitatively for specific products. The information pro-
vided about the waste characteristics is listed below.
Waste I
The composition of Waste I is as follows:
Methyl chloride 2 ppm
Chloroform 683 ppm
Ethylene-di-chloride 8331 ppm
COD 16100 ppm
Chloride as C1 7960 ppm
Alkalinity 0.0
pH <0.1
Waste II
The components of this waste are naphthyl and similar ring structured
compounds together with unsaturated hydrocarbons. Other components were:
COD 106680 ppm
TOC 1810 ppm
-------
TABLE 2
PHYSICAL PROPERTIES OF ACTIVATED CARBON USED IN STUDY
Carbon
Manufactorer
Trade Mark
Mesh Size
2
Surface Area m /g
App. Density g/cc
Real Density g/cc
Part Density
Effective Size
Uniform Coeff.
Pore Volume
Iodine No.
Ash %
Moisture Max °l°
A
Westvaco Chem
Corp.
Nuchar
12 x 40
850
1.4
0.65
1.60
850
7
2
B CD
Witco Chem Am. Norit Calgon
Corp. Corp. Corp.
Witco Norit Filtra-
sorb
12 x 30 8 x 20
1050 1100
0.48 0.44
2.1
0.92 1.35
0.60
I1-9
0.6 0.94
1000
0.5 8.5
2
26
-------
Alkalinity 1475 ppm as CaCOg
Chloride as Cl 14 ppm
pH 9.7
Waste III
This waste contained chlorinated methyl group hydrocarbons. Other
known details were:
pH approximately 2.0
Waste IV
The essential components were:
Methyl Chloride 0-40 ppm
Ethyl Chloride 0-30 ppm
Ethylene/dichloride 9122 ppm
COD 1764 ppm
Alkalinity 1369 ppm as CaC03
Chloride as Cl 2662 ppm
pH 11.0
Total organic carbon was used as the waste concentration parameter.
Since some of the waste materials were highly volatile, the actual pro-
cedure of acidifying the sample and purging with nitrogen gas, for mea-
suring organic carbon, would have resulted in substantial organic loss
from solution. In the carbon analyzer, the instrument used for the
quantitative analysis of total carbon, the sample was carried by oxygen
gas through a furnace at 960°F. At this temperature the organic and
inorganic carbon compounds present in the sample would be decomposed to
carbon dioxide and water in the furnace. The water formed in the process
is trapped in the manometer and carbon dioxide is carried through an in-
frared analyzer. The response proportional to the carbon dioxide present
is transmitted to the recorder. The concentration corresponding to the
response was obtained from a standard curve. It should be indicated that
by this method of analysis of the waste, before and after the carbon adsorp-
tion treatment, the extent of removal of total organic carbon was obtained.
27
-------
It was not possible to Identify the particular component that was removed.
The carbon composition of the wastes is presented in Table 3.
The removal of organic carbon from the wastes is more important for
analysis of carbon adsorption efficiency than the total carbon. However,
it is possible to estimate the inorganic carbon in each waste from the
general characteristics of the wastes. Wastes I and III have a pH of <0.1
and 2.0, respectively, under this condition no inorganic carbon remains in
solution. Therefore, for Wastes I and III the total carbon is equal to or-
ganic carbon. For Wastes II and IV the estimation was made from the results
of column study by making the assumption that the initial leakage was purely
inorganic carbon, which was 55 mg/L and 80 mg/L for Waste II and IV respec-
tively. The inorganic portion of total carbon was also estimated from the
pH and alkalinity data provided by Gulf South Research Institute as 300 Mg/L
and 182 mg/L, respectively. As the results of the isotherm study support the
values obtained from the column study, the values obtained from pH and alka-
linity were Ignored.
Isotherm Study
The Isotherm study is a laboratory simulation of a batch process in
which the activated carbon is contacted with the waste under continuous
stirring and constant temperature until the adsorption reaches equilibrium.
The batch isotherm results are a true measure of sorptlon affinity, as there
is no other physical process contributing to the removal of solutes.
The Warburg apparatus was used for this study because it offered tem-
perature control and continuous mixing. The extent of adsorption of solute
onto activated carbon 1s a measure of the equilibrium in the adsorption pro-
cess. At a constant temperature it is a function of the concentration of
solute in the bulk solution. The functional expression of the distribution
28
-------
TABLE 3
RESULTS OF ANALYSIS OF RAW WASTE
FOR ORGANIC AND TOTAL CARBON
Waste
Total
Carbon
Inorganic
Carbon
Organic
Carbon
I
II
III
IV
4368 + 288
2885 + 365
43 + 7
350 + 6
0
55
0
80
4368 + 288
2830 + 365
43 + 7
350 + 6
29
-------
ratio and the solute concentration in the bulk solution is called the ad-
sorption isotherm. The distribution, expressed as the amount of solute
adsorbed per unit mass of activated carbon, is a function of the concen-
tration of solute remaining in solution at equilibrium. This is not only
a convenient form for representing experimental data, but also a useful
starting point for the development of theoretical treatment of adsorption
equilibria. Theoretical and empirical models have been developed to re-
present the experimental data. These models are discussed below.
Langmuir Model
This model can be deduced from either kinetic considerations or from
the thermodynamics of adsorption with three principal assumptions: 1) max-
imum adsorption corresponds to a saturated monolayer of solute molecules on
the adsorbent surface, 2) the energy of adsorption is constant and, 3) there
is no transmigratory activity of adsorbate molecule in the adsorption surface.
If an adsorption process satisfies these conditions, it will follow
m
0)
where C is solute concentration
X/M is quantity of adsorbent per unit mass adsorbent
X is the adsorbate per unit mass adsorbent
b is the energy of adsorption
BET Model
This model could also be deduced from either kinetic considerations or
from the thermodynamics of adsorption; but unlike the Langmuir model, adsorp
tion is not restricted to a monolayer. This model assumes that a number of
30
-------
layers of adsorbate molecules form at the surface and the Langmuir equation
applies to each layer. A further assumption of BET model is that a given
layer need not complete formation prior to the initiation of subsequent
layers. The equilibrium conditions will, therefore, involve several types
of surfaces in the sense of number of layers of molecules on each surface
site. This model could also be applied to those cases where adsorption
takes place at preferential sites. In a mathematical form, this model
could be represented as
C _ 1 . E-l C
EXm
where E is the constant of energy of interaction with the surface
C is the saturation concentration of the solute.
Freundlich Model
This model is a special case for heterogeneous surface energies in which
the energy term varies as a function of surface coverage due to variations in
the heat of adsorption. The equation has the general form
1
X/M = E Cn (3)
and the data fitted to the logarithric form as
log X/M = log E + 1 log C (4)
The intercept is roughly an indicator of sorption capacity and the slope,
the adsorption intensity.
Column Study
The column study was designed to analyze the dynamics of the adsorption
31
-------
process under a constant flow of the waste through the void space of a fixed
bed of activated carbon. Column operations have a distinct advantage over
batch operations because rates of adsorption depends on the concentration of
solute in solution and, for column operation, the column is in continuous
contact with fresh solution. Whereas, in the isotherm study, no attempt was
made to study the effect of transport mechanism on adsorption. The column
study is particularly designed to evaluate the transient characteristics of
adsorption.
As the waste moves down by gravity through the voids of the carbon bed,
the organic carbon is adsorbed rapidly by the upper layer of fresh carbon,
the primary adsorption zone, because this layer is in contact with the solu-
tion at the highest concentration. The small amount of organic carbon es-
caping adsorption in this layer is removed in the lower strata of the bed and,
initially, no impurity escapes from the adsorber. As the process continues,
the top layers of the carbon become partially saturated and become less effec-
tive for further adsorption; then the primary adsorption zone moves through
the column to regions of fresher adsorbent. As the primary zone moves down,
more and more solute tends to escape in the effluent. The breakthrough occurs
when there is the appearance of an adsorbate of interest 1n the carbon bed ef-
fluent at a predetermined concentration.
The breakthrough curve, for a column operation, is obtained by plotting
concentration against volume of flow. For most operations in wastewater
treatment, the breakthrough curve exhibits a characteristic "/" shape, but
with varying degree of steepness and position of break point. The factor in-
fluencing the break through curve are the flow rate, bed depth, detention time,
and all other factors influencing the process of adsorption. The break point
is chosen at some low value and exhaustion point is chosen close to the initial
concentration of the waste. When the effluent concentration reaches or passes
32
-------
the break point value and reaches an arbitrary allowable value, the operation
should be discontinued and the bed regenerated for further use.
In this study, a 50 ml burette packed with 0.25 to 4.0 grains of activated
carbon was used for the packed column and the flow was adjusted to a few drops
per minute to maintain a contact time almost equal to that of a conventional
pilot plant study. Such a simple experimental set up was resorted to due to
paucity of resources. Although the quantitative nature of the results are
not identical to those of pilot or scale up processes, it is believed that
the results at least describe the nature of the transport mechanism qualita-
tively.
Regeneration Study
For the regeneration study, Carbon B was chosen as it exhibited slightly
superior adsorption characteristics to the other carbons. The choice of
waste depended not only on the adsorption characteristics but also on availa-
bility. Waste I was not available in the required quantity. Waste II, al-
through not available in its original strength, was available in diluted
form after being used for the isotherm study. As the chemical nature of the
waste may change when it is subjected to adsorption, Waste II may have a dif-
ferent composition from the original waste. Therefore, it is labeled as
waste II-A, since it is believed that the original properties were not changed
drastically. Waste III exhibited very poor adsorption characteristics, so it
was not considered for regeneration study. Waste IV was available in suffi-
cient quantity and was employed in conjunction with Carbon B and Wastes II-A
and IV.
About three grams of carbon was contacted with a large volume of waste
until the steady state concentration was reached. The saturated carbon was
divided into five parts and each part was regenerated at temperatures in the
-------
range of 0 to 1000°F. Then, an adsorption isotherm was obtained using 10 ml
of the waste and varying amounts of carbon - using the regenerated carbon to
study the effect of temperature on regeneration.
The spent carbon was regenerated in a high temperature furnace using a
vacuum of 50 mm of mercury. The spent carbon was transferred to a boat and
inserted into the furnace. After checking for air tightness, the vacuum pump
was switched "on" and the vacuum was allowed to build up before switching the
furnace "on". As the furnace was heated, the emf generated in the thermo-
couple was transmitted to the strip chart recorder. The heating unit was
turned off when the chart recorded an emf equivalent to the temperature of
regeneration since the control system was not designed to maintain a constant
temperature. Then the furnace was allowed to cool to ambient temperature un-
der vacuum conditions.
Results and Discussion
The results of the isotherm study are presented in Table 4 for four car-
bons and four wastes. The first column indicates the quantity of activated
carbon M, in milligrams per liter, in a 25 ml sample. The waste may contain
volatile matter that escapes from solution when subjected to agitation so, in
order to estimate only that portion available for adsorption, the first sam-
ple, or reference sample, was run without activated carbon 1n the flask. The
initial concentration of the waste was assumed to be the equilibrium concen-
tration of the reference sample, and the removal of organic carbon, for any
other sample was obtained by subtracting the equilibrium concentration of the
sample and it is expressed as milligrams per liter. The ratio of X/M was ob-
tained as a dimensionless number and was the mass of organic carbon adsorbed
per unit mass of activated carbon.
The characteristics of adsoption isotherms, proposed by Weber (29) are
34
-------
TABLE 4
RESULTS OF ISOTHERM STUDY
Waste
I
II
Carbon A
Act. Car.
M
mg/L
0
60
200
500
1000
2000
0
644
1092
5035
10030
15025
25100
Eqm.
Cone.
mg/L
4080
3888
3792
3528
3720
3720
3185
2885
2645
1445
197
0
0
Removal
x
mg/L
0
192
288
552
360
360
0
300
540
1740
2988
3185
3185
X » x/M
_
3.2
1.44
1.10
0.36
0.15
-
0.466
0.498
0.346
0.296
0.212
0.127
M
0
60
200
500
1000
2000
5024
10016
25048
49888
0
700
1150
5000
10080
15040
25100
Carbon
Eqm.
Cone.
4656
4512
4320
4164
3792
3276
2760
2360
2090
1120
3205
3355
2845
1725
725
145
145
B
X X
0
144 2.4
336 1 .63
492 0.98
864 0.85
1380 0.69
1896 0.38
2296 0.23
2566 0.10
3536 0.07
0
.
360 0.313
1480 0.296
2480 0.246
3060 0.203
3060 0.122
Carbon C
M Eqm. x X
Cone.
0 2465
1896 1985 480 0.253
7000 1047- 1418 0.202
9088 749 1716 0.187
12116 421 2044 0.168
17048 115 2350 0.138
20136 53 2412 0.120
23120 53 2412 0.104
M
0
35
175
515
900
2015
0
2292
4000
6000
8005
11068
12808
14080
Carbon
Eqm.
Cone.
4656
4200
3620
3488
3060
3060
2465
1373
905
578
353
105
35
35
D
x X
0
456 13.04
1036 5.92
1168 2.27
1596 1.63
1596 0.79
.
1092 0.477
1560 0.390
1887 0.315
2112 0.264
2360 0.215
2430 0.191
2430 0.173
CO
en
-------
TABLE 4 (CONT'D)
RESULTS OF ISOTHERM STUDY
Waste
III
IV
Carbon A
Eqm.
M Cone.
0 50.0
95 29.2
204 27.6
348 21.2
420 27.0
584 18.4
840 15.6
1028 15.6
0 264
1992 189
4004 149
5188 134
6040 117
7116 108
8000 104
X
_
20.8
22.4
28.8
28.8
31.6
34.4
34.4
.
75
115
130
147
155
160
X
.
0.217
o.no
0.083
0.066
0.054
0.041
0.035
_
0.038
0.029
0.025
0.024
0.022
0.020
Carbon B
M
0
68
192
320
448
608
840
1024
0
2024
4152
5228
6000
7112
7944
Eqm.
Cone.
50.0
28.0
21.2
18.0
18.4
17.2
15.2
15.2
264
189
159
134
132
130
117
X
„
22.0
28.8
32.0
31.6
32.8
34.8
34.8
„
75
105
129
132
134
147
X
_
0.324
0.150
0.100
0.071
0.054
0.042
0.034
.
0.037
0.025
0.024
0.022
0.019
0.018
Carbon C
Eqm.
M Cone. x X
0 36.0
68 32.4 3.6 0.053
176 30.0 6.0 0.034
360 27.6 8.4 0.023
436 27.2 8.8 0.020
620 28.8 7.2 0.011
856 26.8 9.2 0.011
1032 27.6 8.4 0.008
0 277 -
2084 234 43 0.021
3944 202 75 0.019
5152 196 81 0.016
5992 185 92 0.015
7132 182 95 0.013
8044 176 101 0.012
Carbon D
M
0
100
216
308
404
636
844
980
0
2092
3948
5012
6008
6928
8298
Eqm.
Cone, x
36.0 -
28.8 7.2
27.2 8.8
27.0 9.0
27.6 8.4
26.3 9.7
23.6 12.4
24.4 11.6
277
213 64
188 89
179 98
166 111
164 113
160 117
X
_
0.072
0.041
0.029
0.021
0.015
0.015
0.012
-
0.031
0.022
0.020
0.018
0.016
0.014
CO
cr>
-------
presented in Figure 5. Curves 1 and 3 indicate the curvilinear dependence
of favorable and unfavorable adsorption pattern, respectively, and curve 2
represents a combination of adsorption and absorption which occur in direct
proportion to concentration. Using this general criteria, the experimental
data was analyzed and the results presented in Table 5.
The pattern of favorable and unfavorable adsorption appears to depend
on the components of each waste and their respective adsorbability. The
recent study of Giusti e_t a]_., (33) has established that the amenability
of typical organic compounds to activated carbon adsorption from pure com-
ponent systems increases with increasing molecular weight and decreasing
polarity, solubility, and branching. Although multicomponent systems, as
encountered in petrochemical waste, have not been studied to the extent nec-
essary, it is probably reasonable to assume that, as the average molecular
weight of the components of a multicomponent system increases, the process
of adsorption becomes more favorable provided polarity, solubility, and
branching do not affect adsorption significantly. As Waste II is composed
of components of higher molecular weight, such as anthryl and naphthyl ring
structured compounds (34), a favorable adsorption pattern is obtained. It
appears that short chain organic chlorides, principal components of Wastes
I and IV are particularly resistant to carbon adsorption. This may be be-
cause chloride compounds are lyophylic and the intermolecular forces are
better balanced when adsorbed. Waste III, on the other hand, composed of
methyl group hydrocarbons, exhibited poor adsorption characteristics pos-
sibly due to the low average molecular weight of the components. The rela-
tive adsorption characteristics of Wastes I, II, III, and IV, with Carbon B,
are presented in Figure 6.
The process of adsorption also depends on the nature of the carbon. In
37
-------
Figure 5. Types of sorption separations after
Weber (29): 1. Favorable adsorption;
2. Linear adsorption and adsorption;
3. Unfavorable adsorption.
38
-------
TABLE 5
SORPTION CHARACTERISTICS OF WASTES STUDIED
Carbon A B C D
Waste
I x x - x
II + + + +
III x x x x
IV x x x x
+ favorable adsorption
x unfavorable adsorption
39
-------
00
o
.o
o
O
ro
CD
CM
- 9
o
E
o
0>
(X
TJ
0)
.0
•o
o
o
o
o
o
o>
o
K>
O
CM
O
q
o"
CM
o o
1000
2000
3000
—i az
10
15
20
25
30 — 3
100
125
150
175
200— 4
Concentration of Organic Carbon in mg/l
Figure 6. The influence of the source of petrochemical waste on
adsorption characteristics: Adsorption Isotherm for
Wastes I, II, III, and IV using Carbon B.
40
-------
Figure 7, the Isotherm for Waste II, with all four carbons Is presented.
Carbon C exhibited by far the poorest adsorption characteristics followed
by Carbon B. At concentrations above 400 mg/L of the waste, Carbon D ex-
hibited the best adsorption pattern and below 400 mg/L Carbon A was best.
It Is Important to note that the suitability of a particular carbon depends
also upon the range of waste concentration. This phenomenon was noted even
for wastes with an unfavorable adsorption pattern. In the case of Waste I,
Figure 8, Carbon D was best suited 1n the concentration range higher than
3000 mg/L, but below this concentration it loses its superiority to Carbon
B. However, for Wastes III and IV, Figures 9 and 10 show that Carbons B
and A, respectively, were found to give the best pattern over the entire
range of waste concentration. Although it was impossible to pick out a
carbon that was capable of providing the best pattern for all wastes, the
one that exhibited the poorest pattern in all cases was Carbon C.
The results of the experiments do not necessarily have to conform to
any of the isotherm models discussed earlier. However, analysis based on
these models hopefully will show that one of the models describes the nature
of adsorption both quantitatively and qualitatively. The experimental points
were fitted to the linearized form of the three models, equations (1), (2),
and (4), and it was found that the Freundlich model more accurately described
the process than the others. In Figures 8 through 11, the linearized forms
of the Freundlich model are presented for all wastes and carbons. The Lang-
mulr and BET models were not as applicable since the linearized form usually
resulted In a curve, whereas the Freundlich model always produced a linear
response. The lack of fit of the Langmuir and BET models may be attributed
to the limitations of the theoretical assumptions used in deriving Langmuir
and BET isotherms, or that factors not considered In these two models in-
fluence and alter the adsorption pattern. In Figures 12, 13 and 14, the
41
-------
O.6 -
c
o
-O
L_
o
o
•o
0)
o
o
E
o
0)
a
•o
41
O
10
TJ
O
O
O
O
O
0>
E
o
0.5 -
IOOO 2000
Concentration of Organic Carbon in mg/l
3000
Figure 7. The influence of the nature of activated carbon on
adsorption characteristics; Adsorption Isotherm for
for Waste II with Carbons A, B, C and D.
42
-------
o
.0
L.
O
o
E
o
o>
Q.
O
o
c
o
t-
o
o
_o
c
o
o>
6
o
w
o
5.0
4.0
3.0
2.0
1.0
0.8
0.6
0.4
0.2
O.I
- • — Carbon A
I i I
1000 2000 3000 4OOO 6OOO
Concentration of Organic Carbon in mg/l
8000
Figure 8. Results of Isotherm Study for Waste I and
Carbons A, B and D in the linearized form
of Freundlich Isotherm Model.
43
-------
•o •£
c O
o *
O O
O •£
'c o
o <
O» -_
6 =
Is
I »
O w
1.0 -
0.5
0.4
0.3
0.2
O.I
20
100
1000
6000
Concentrotion in mg/l
Figure 9. Results of Isotherm Study for Waste II and Carbons A, B, C and D in the
linearized form of Freundlich Isotherm Model.
-------
o
.o
o
I
o
•o
0
.a
o
•o
o
I
o
o
o
'£
o
o>
E
o
0.40
0.30
0.20
0.10
0.08
0.06
0.04
O.O3
0.02
O.OI
—Carbon A
—Carbon B
— Carbon C
—Carbon D
I
I
10 15 20 25 30 40 50
Organic Carbon mg/1
Figure 10. Results of Isotherm Study for Waste III
and Carbons A, B, C and D in the linearized
form of Freundlich Isotherm Model.
45
-------
o
.0
£
o
o
E
o
W
a.
o>
.o
s
•o
o
c
o
.O
O
o
0.04
0.03
0.02
W
E
o
I—Carbon A
f— Carbon B
> — Carbon C
> — Carbon D
0.01
I
100
200
Organic Carbon, mg/l
300
400
Figure 11. Results of Isotherm Study for Waste IV and
Carbons A, B, C and D in the linearized form
Freundlich Isotherm Model.
46
-------
-0.5
-0.6
-0.7
-0.8
-0.9
-1.0
2.0
Waste H
I L
2.5 3.0
Log (C)
3.5 4.0
0.0
Waste
-0.5
-1.0
-1.5
I.I
1.3
Log (C)
1.5
Figure 12. Linearized form of Freundlich Isotherm
Model for wastes exhibiting favorable
adsorption, Waste II and unfavorable
adsorption, Waste III with Carbon B.
47
-------
8000
6000
4000
Woste XL
2000
1000 2000 3000
C
500
Waste HL
35
Figure 13. Linearized form of Langmuir Isotherm
Model for wastes exhibiting favorable
adsorption, Waste II and unfavorable
adsorption, Waste III with Carbon B.
48
-------
20
o
I
Waste H
0.2 0.4
0.6
C
C.
0.8 1.0
T 8
"- 6
Waste HT
0.2 0.3
0.4 0.5
C
C,
0.6 a?
Figure 14. Linearized form of BET Isotherm Model
for wastes exhibiting favorable adsorption,
Waste II and unfavorable adsorption, Waste
III with Carbon B.
49
-------
three models are presented for wastes which exhibit favorable and unfavorable
adsorption, Wastes II and III, respectively, with Carbon B.
In summary, not all petrochemical wastes are amenable to carbon adsorp-
tion. Wastes with short chain compounds do not always exhibit favorable ad-
sorption pattern. As the average molecular weight of the waste increases the
better the adsorption pattern. No one carbon exhibited the best adsorption
for all wastes. The choice of carbon depends on the waste as well as the
waste concentration ranges. The empirical Freundlich model more accurately
represents petrochemical waste adorption than the Langmuir or BET models.
Packed Column Study
The results of packed column studies, for Waste II and Carbon A, are
presented in Figure 15. It was expected that the effluent would approach
the initial concentration when the bed reached saturation. However, the
initial concentration was not reached by the effluent even though over two
liters of the waste was treated. It is believed that the difference between
the initial concentration and breakthrough concentrations was caused by some
kind of filtration effect. The overall adsorption resulted in the removal
of 0.286 grams of organic carbon per gram of activated carbon and adsorp-
tion alone contributed to the removal of 0.235 gm/gm concentration and break-
through concentrations were caused by some kind of filtration effect. The
overall adsorption resulted in the removal of 0.286 grams of organic carbon
per gram of activated carbon and adsorption alone contributed to the removal
of 0.235 gm/gm.
The results of the packed column study, for Waste III and Carbon A indi-
cated a wide scattering of experimental data points. This was attributed to
poor adsorbability of the waste - further complicated by inaccuracy in the
measurement of total carbon in such low concentration.
50
-------
o
.o
o
o
3600 -
3200 =
24OO —
1600 —
800 —
100 200 300 400 500
Volume of Waste treated, ml
600
700
800
Figure 15. Results of packed column study for Waste II and Carbon A.
-------
In Figure 16, the results of one of the three-column studies for Waste
IV and Carbon A is presented. All three have similar breakthrough curves
but are vastly different from the characteristic shape observed for Waste
II. This may be attributed to the nature of adsorption as presented in
Figure 6. The waste that exhibited a favorable adsorption isotherm, Waste
II, also yielded the characteristic breakthrough curve in the column study,
whereas, the waste that had an unfavorable isotherm deviated from this pat-
tern.
In Table 6, the results of three column studies, with Waste IV and Car-
bon A, are summarized. The comparison of X/M values of isotherm and column
studies, Tables 4 and 6 respectively, indicates that two of the three values
of X/M for column study are greater than obtained in isotherm study indicating
better removal in the column study. This is in contrast with the results for
Waste II in which the column study X/M value of 0.235 is between the maximum
0.494 and the minimum 0.127, obtained for isotherm study. It is interesting
to note that in the column study, for Waste IV, as the residence time In-
creases the removal per unit mass decreases, a phenomenon also observed by
Cohen and Kugelman (6) in their studies. This indicates that the determina-
tion of an optimum residence time for column operation 1s essential.
In summary, the characteristic breakthrough curve was exhibited only by
the waste that exhibited favorable adsorption pattern. In the case of Waste
II, the effluent concentration after the saturation of the bed was lower than
the raw waste concentration, which probably Increased the removal by some
type of filtration mechanism. The residence time Is a critical factor for
efficient operation of a column.
Regeneration Study
In the regeneration study for Waste II-A and Carbon H, one liter of the
52
-------
in
to
E
•»
g
o
o
c
o
250
200
ISO
100
50
100 200 300 4OO
Volume of Woste Treated, ml
500
600
700
800
Figure 16. Results of packed column study with Waste IV and Carbon A
-------
TABLE 6
SUMMARY OF THE RESULTS OF THREE PACKED COLUMN
STUDIES FOR WASTE IV, CARBON A
1.
2.
3.
* 4.
5.
** 6.
++ 7.
8.
+ 9.
Columns
Activated Carbon (gm)
Time of Operation (hrs)
Volume of Waste Treated (ml)
Carbon Adsorbed (mg)
x/M
Volume of Activated
Carbon (ml)
Void Volume (ml)
Flow Rate (ml/hr)
Residence Time (hrs)
1
2.0385
511:45
731
93
0.048
4.72
3.59
1.428
2.52
2
2.0204
511:45
511
37.7
0.0184
4.67
3.55
1.00
3.55
3
0.3364
90
526
277
0.082
0.779
0.592
5.87
0.1
* Obtained from the respective figures by counting
squares.
** Volume of activated carbon: weight/0.433, 0.433
being the density.
+ Void volume/flow rate.
++ Void volume is 76.1 of total volume.
54
-------
waste was contacted with 3.019 grams of carbon for a period of about 48 hours.
During this period the initial concentration of organic carbon, 2585 mg/L, was
reduced to 1889 mg/L resulting in overall adsorption of 0.23 grams of organic
carbon per gram of activated carbon. The saturated carbon was divided into
four parts; one part was dried at 70-80°C and the other three regenerated at
400, 700 and 1000 degrees fahrenheit. After the regeneration of spent carbon,
an isotherm study was conducted with 10 ml of waste. The results are presented
in Table 7. The Freundlich isotherm model was used to compare the performance
of regenerated carbon, Figure 17. It is apparent that an increasing regenera-
tion temperature improves the adsorption pattern. As indicated in Table 8,
the parameters E and n increased with temperature, a positive sign that regen-
eration improved with increasing temperature.
In the study with Waste IV and Carbon B, 4.033 grams of carbon were con-
tacted with two liters of the waste to reach a steady state concentration of
214 mg/L from the initial concentration of 292 mg/L to give an adsorption of
0.0387 grams of organic carbon per gram of activated carbon. The saturated
carbon was divided into five parts for regeneration. The regenerated car-
bon was contacted with 10 ml of waste, the results are presented in Table 9.
The linearized form of Freundlich model is also presented in Figure 18.
Comparison of the results of the two studies shows that the system ex-
hibiting favorable adsorption, Waste II-A, requires a high temperature of
regeneration and more energy expended for regeneration. In the case of Waste
II-A, the optimum temperature of regeneration was higher than 1000°F, but for
Waste IV, it is around 600°F. Therefore, it Is necessary that the optimum
temperature of regeneration be established for every system. A higher than
optimum temperature would probably reduce the efficiency of the process and
also damage the carbon.
55
-------
TABLE 7
RESULTS OF REGENERATION STUDY WASTE 11-A: CARBON B
M
0
550
1040
1930
3100
4020
5410
6170
7270
Virgi
Eqm. Cone.
2345
2225
2125
-
1765
1625
1385
1245
n
X
-
120
220
-
580
720
960
1100
1085 1260
X
-
0.218
0.216
-
0.187
0.179
0.177
0.178
0.173
70CTF
0
450
1050
1970
3210
3990
4980
6120
7330
2245
2185
.2115
2005
1875
1805
1755
1725
1645
_
60
130
240
370
440
490
520
600
-
0.133
0.124
0.122
0.115
0.110
0.098
0.085
0.082
0UF
M
0
480
1050
2060
3020
4140
5000
6150
7150
0
560
1080
2020
3080
4010
4930
6250
7680
Eqm. Cone
2295
2295
2295
2295
2295
2295
2295
2295
2295
1000°F
2295
2215
2145
-
1925
1845
1705
1665
1385
. x
0
0
0
0
0
0
0
0
0
-
80
150
-
370
450
490
630
910
X
0
0
0
0
0
0
0
0
0
_
0.143
0.139
-
0.120
0.112
0.100
0.101
0.118
400WF
M Eqm. Cone, x
0 2245
590 2185 60
970 2155 90
1970 2065 180
3170 1975 270
4110 1915 330
5210 1855 390
6150 1815 430
7260 1765 480
X
-
0.1018
0.0929
0.0914
0.0852
0.0803
0.0748
0.0699
0.0662
01
en
-------
o
o
0.20
E
o
Q.
c
o
0.10
0.08
6
E
o
o
0.06
— Virgin Carbon
' — 400° F Regn.
I—7OO°F Regn.
>—IOOO° F Regn.
i i t i i i I i i i i i i i i i
1000
2000 3000
Organic Carbon, mg/l
Figure 17. Effect of temperature on regeneration for
Waste II-A and Carbon B.
57
-------
TABLE 8
Co-efficients and Exponents of Freundlich Isotherm Model, Equation (3) Evaluated
From the Results of Regeneration Study for Wastes II-A and IV With Carbon B
en
00
Waste II-A - Carbon B
Freundlich
Model
Parameter
E
n
Reqeneration Temperature
Virqin 0°F 400°F 700°F
1.7108xlO"2 8.734 x 10"8 1.77xlO"7
3.0684 0.5512 0.5663
1000°F
7.637 x 10"6
0.7830
Waste IV - Carbon B
Freundlich
Model
Parameter
E
n
Regeneration Temperature
Virgin 0°F 400°F 600°F 800°F
1.048xlO"5 0.8989 2.03 x 10"5 35.79 x 10"5 0.097 x 10"5
0.6434 -1.1551 0.7641 1.2977 0.535
1000°F
0.379 x 10"5
0.6298
-------
TABLE 9
RESULTS OF REGENERATION STUDY WASTE IV: CARBON B
Virgin
Eqm.
M Cone.
0 264
2040 189
4152 159
5228 135
6000 132
7112 130
7944 1 27
Carbon
X
_
75
105
129
132
134
147
600°F
0 290
610 275
1100 262
2090 234
3110 216
4270 196
5170 186
6140 176
7160 166
0
15
28
56
74
94
104
114
124
0° Regeneration 400°F
Eqm. Eqm.
X M Cone. x x M Cone. x x
0 290 - 0 330 -
0.0368 520 290 0 0 570 310 20 0.0351
0.0253 1000 290 0 0 1160 292 38 0.0328
0.0247 2200 276 14 0.0064 2020 264 66 0.0327
0.0220 3040 268 22 0.0072 3150 242 88 0.0279
0.0188 4090 262 28 0.0068 4050 224 106 0.0262
0.0185 5130 256 44 0.0086 5060 212 118 0.0233
6030 256 44 0.0073 6140 204 126 0.0205
7130 246 54 0.0076 6950 200 130 0.0187
800°F 1000°F
0 0 330 - 0 345 -
0.0246 530 304 26 0.0491 490 340 5 0.0102
0.0255 1040 290 40 0.0385 1080 334 11 0.0102
0.0268 1980 274 56 0.0283 2080 312 33 0.0159
0.0238 3020 248 82 0.0272 3130 264 81 0.0259
0.0220 4020 224 106 0.0264 4070 248 97 0.0238
0.0201 5070 212 118 0.0233 5000 232 113 0.0226
0.0186 6150 202 128 0.0208 6060 220 125 0.0206
0.0173 7760 196 134 0.0173 7140 216 129 0.0184
01
10
-------
c
o
.0
o
o
E
o
•o
w
43
C
o
43
O
o
o
a*
6
•^
o
w
E
o
O
0.05
0.04
0.03
0.02
0.01
0.008 -
0.006
Q — Virgin Carbon
• —0° F Regeneration
< — 400°F Regeneration
D — 600° F Regeneration
O — 800° F Regeneration
O—IOOO°F Regeneration
OO
100
200
300
400
500
Organic Carbon, mg/l
Figure 18. Effect of temperature on regeneration of Waste IV
and Carbon B.
60
-------
It should be Indicated that even at an optimum regeneration temperature,
the regenerated carbon did not match the performance of virgin carbon. Al-
though 1t 1s hard to pin point the reason within the limitations of this study;
a host of reasons can be cited In which the damage to pore structure of carbon
due to thermal stress Is the most significant.
In summary, the optimum temperature of regeneration for each system should
be established. Even regeneration at optimum temperature would not return the
carbon to the performance level of virgin carbon, which 1s probably due to the
damage caused by thermal stress to the pore structure.
61
-------
SECTION 7
VACUUM-THERMAL REGENERATION OF ACTIVATED CARBON
The purpose of this study was to determine the specific regeneration
response of activated carbon, saturated with selected sorbates, to high tem-
perature vacuum regeneration.
Conventionally spent activated carbon 1s regenerated by exposure to high
temperature conditions In multiple hearth furnaces or rotary kiln devices.
In this process, carrier gases such as carbon dioxide and stain are used to
prevent oxidation of the carbon sorbent. The role of the carrier gas, in
terms of protecting the carbon sorbent, 1s well defined; however, the role
of the carrier gas in terms of regeneration and/or reactivation is not de-
fined. In order to minimize the effect of the carrier gas on thermal regen-
eration, this study was designed on the basis of using vacuum thermal regen-
eration techniques.
Prior to development of special vacuum thermal regeneration equipment,
samples of various commercial grades of activated carbon saturated with petro-
chemical waste and methylene blue sorbates were prepared. These samples were
subjected to vacuum differential thermal analysis (DTA). Following differen-
tial thermal analysis, the samples were examined for sorptlon capacity. Re-
sults of this study established that vacuum thermal regeneration was possible
and potentially promising.
Consideration of the factors Involved in testing carbon sorbates, together
with the relatively large amounts of activated carbon required, showed that a
special vacuum furnace device would be required. The DTA apparatus was capa-
ble of containing a 50 mg sample; however, 50 gram samples of material were
required for sorption, regeneration, and analysis.
The sorbates selected for this study were Iodine, methylene blue, para-
62
-------
chlorophenol, and a petrochemical waste material. Iodine is a standard sor-
bate material routinely used to evaluate activated carbon sorption capacity
and micro pore capacity. In practice, activated carbon materials which ex-
hibit an iodine affinity of 600 mg per gram of carbon, or greater, as deter-
mined by a standard solution exposure and contact procedures, are considered
to be high quality sorbent materials. It is generally agreed that the iodine
sorption capacity of activated carbon is a measure of the micropore spaces (35,
36). Usually the numerical value of the iodine number for an activated car-
bon sorbent is very similar to the surface area per unit mass value. Thus,
for very porous activated carbon materials, the iodine number and the sur-
2
face area per unit mass (m /gm) values are nearly the same. The established
nature of iodine sorption by activated carbon was the first basis for selecting
this material, however, iodine has an additional characteristic which favored
inclusion in this study. It is an inorganic material which shows well de-
fined sharp melting and boiling points at 113°C and 184°C, respectively. More-
over, it is only slightly soluble in water (<162 mg/liter at 20°C) and sublimes,
without melting, below a pressure of 100 mm Hg. In addition, this element has
a relatively high heat of vaporization value - 10,388 cal/mole.
Methylene blue was included in this study because it is also a standard
material used in sorption studies. The major reason for using methylene blue,
in activated carbon sorption studies, is related to the optical properties of
this substance. Dilute solutions of methylene blue exhibit a pure blue color
which conforms to the Beer-Lambert relationship at 665 millimicrons. Thus,
it is convenient and easy to measure the sorption uptake of methylene blue by
colorimetric measurement. Traditionally methylene blue has been used in san-
itary engineering to evaluate sorption and the relative stability of biooxi-
dation processes. Methylene blue is a relatively large molecule having a
63
-------
molecular weight value of 320. Probably the most Interesting feature of
methylene blue is that 1t decomposes before melting.
The third sorbate material selected for this study was parachlorophenol.
Generally the para-chlorophenol molecule 1s considered to be similar to many
components found in organic chemical and petrochemical trade wastes. Para-
chlorophenol 1s a dichloro derivative of benzene which is slightly soluble
1n water and is a solid at 23°C. This material exhibits a sharp melting
point at 42°C and boils at 217°C. In contrast to methylene blue and iodine,
para-chlorophenol is a moderately soluble low molecular weight material pos-
sessing a relatively high boiling point.
The fourth sorbate material selected for this study was a petrochemical
waste material. Specifically, the petrochemical waste material was collected
from a cold temperature quenching operation waste stream by representatives
of Gulf South Research Institute. The waste sample was highly colored, had
a COD value of 10,180 mg/liter, and contained 1,810 mg/liter dissolved or-
ganic carbon. Chloride concentration of the waste was 14 mg/liter and the
pH value was 9.7. This waste material was representative of a specific petro-
chemical waste stream containing relatively large quantities of benzene deri-
vatives.
Activated carbon materials used in this study were Flltrasorb 400 and
2
Nuchar WVL . Preliminary laboratory study showed that both carbon materials
were effective sorbents for the selected sorbates and exhibited stable response
to high temperature exposure. Prior to application, stock quantities of the
selected carbon materials were ground up and sieved to pass 40 mesh size,
soaked in distilled water, and dried at 103°C for 24 hours.
Product of Calgon Chemical Co., Pittsburgh, Pennsylvania
Product of Westvaco Chemical Division, Covlngton, Virginia
64
-------
Analysis of the organic sorbate concentration was accomplished on the
basis of total organic carbon using a Sectarian Carbon Analyzer instrument .
Procedure developed for determining the total organic carbon concentration
of para-chlorophenol, methylene blue, and the petrochemical waste consisted
of removing a 1.0 ml portion of the carbon-wast solution, diluting to 100
ml with distilled water, acidification with HN03 to pH value of 2, and the
injection of a small sample (10-50 microliter) into the carbon analyzer.
The carbon analyzer instrument was standardized just prior to use with
standard glucose solutions.
Residual iodine concentration was measured by starch-iodine titration
using sodium thiosulfate as a titrant (35). The complete detailed analysis
of iodine and iodine number determination are presented in Appendix B.
The laboratory procedure used to establish sorption consisted of con-
tacting the sorbate solutions with varying amount of activated carbon and
determining the change in sorbate concentration with time. Specifically
this procedure involved contacting 1.0, 3.0, 5.0, 7.0, 10.0, 20.0, 50.0,
100 and 250 milligram portions of activated carbon with 100 ml of sorbate
solutions. The sorbate solutions were prepared and applied in the fol-
lowing concentrations: iodine 1000 mg/liter, para-chl orophenol 1000 mg/Hter,
methylene blue 100 mg/liter and petrochemical waste 1,810 mg/liter organic
carbon. In practice, weighed amounts of activated carbon were transferred
to a series of 125 ml Erlenmeyer flasks containing 100 ml of sorbate solu-
tion. The activated carbon-sorbate solution mixture was then placed in a
22°C temperature controlled modified Warburg apparatus. Aliquot samples
of the incubated solutions were removed by pipetting at five hour intervals
Product of Beckman Instruments, Inc., Fullerton, California
65
-------
diluted with distilled water and reserved for analysis. Following deter-
mination of sorption equilibrium, samples of the respective mixture (acti-
vated carbon-sorbate) were prepared for thermal regeneration.
Preliminary study showed that all sorbate materials reached sorption
equilibrium in 24 hours or less. On the basis of this determination, all
of the experimental solution mixtures were subjected to a 24 hour contact
period. A wide range of carbon to sorbate ratios and concentrations were
contacted in the preliminary study and representative loading values of
0.25, 0.75, 0.90 gms of sorbate/gm of carbon were used for the balance of
the investigation.
Thermal regeneration was provided by exposing 1 gram portions of sor-
bate- carbon mixtures (spent carbon) to temperatures ranging from 200°F to
1400°F in a vacuum tube furnace. The procedure of thermal regeneration
involved transferring weighed samples of activated carbon-sorbate mixtures
to a small ceramic sample boat. The ceramic sample boat and contents were
then located in the center of a 1.5 inch diameter 32 inch long ceramic fur-
nace combustion tube. A contact thermocouple was placed in the sample boat
just prior to evacuating the combustion tube. After location of the sample,
the combustion tube was sealed with rubber plugs and evacuated to 50 mm Hg
pressure. Temperature of the sample was monitored continuously with a cromel
alumel thermocouple attached to a Minneapolis-Honeywell, Model 2H-11, strip
chart recorder . Pressure In the combustion tube was determined using an
electronic vacuum gauge. A sketch of the experimental equipment is shown
In Figure 19.
Following completion of the heating regime, the sample boat and contents
Product of Minneapolis-Honeywell Corporation, Pittsburgh, Pennsylvania
66
-------
Sample Holder
Thermocouple
*
^^T— — u— ^X"
Combustion Tube
Combustion
Tube
Temp. Recorder
Furnace
Vacuum Pump
Figure 19. Experimental carbon regeneration equipment.
67
-------
were permitted to cool to ambient temperature In the 50 mm Hg pressure com-
bustion tube. Samples of the thermally regenerated mixture were placed in
a dessicator and prepared for sorption study.
Determination of the effect of thermal regeneration was accomplished
by determining the sorption loading achieved by the regenerated carbon sor-
bent. This operation involved exposure of a weighed portion of the regen-
erated sorbent to various concentrations of sorbate. The same procedure,
sorbate, equipment, and analysis were used in this test as in the Initial
sorption test.
It should be noted that all samples of activated carbon sorbate mix-
tures were heated in a 50 mm Hg vacuum. A control manometer was not used
to control the vacuum; the 50 mm Hg pressure was the equilibrium pressure
attained with the porous ceramic combustion tube apparatus. Preliminary
experimentation showed that this pressure was sufficient to prevent oxi-
dation of the activated carbon. All samples of the activated carbon sor-
bate mixture were heated in ceramic sample boats and were spread out over
the total surface of the boat 1n order to promote good heat transfer. The
location of the sample boat in the furnace combustion tube Is shown in
Figure 19.
Results and Discussion
Results of the Initial sorption regeneration study, using Filtrasorb
#400 activated carbon and methylene blue dye, showed that increased temper-
ature favored regeneration response. The specific response of FUtrasorb
#400 activated carbon samples, exposed to temperatures ranging from 220°F
to 1200°F, is shown in Figure 20. The effect of sorption capacity loading,
in terms of grams of sorbate sorbed per gram of carbon, is displayed 1n
Figure 20 as four separate curves. The sorption regeneration response, for
68
-------
each sample, was determined and tabulated by comparison with virgin Filtra-
sorb #400 activated carbon sorptlon values and plotted as a regeneration
ratio (Rr) value. The regeneration ratio value was calculated using the
following relationship.
Rr _ (X/M) reg
Kr (X/M) virg
where
Rx = regeneration ratio
X = mass of sorbate (methylene blue), grams
M = mass of sorbent (activated carbon), grams
This procedure permitted direct evaluation of regeneration response In terms
of experimentally determined sorptlon capacity values for Flltrasorb #400
virgin activated carbon.
The sorptlon loading values and their associated thermal regeneration
response, In terms of Rr values In the temperature range5 200°F, describe a
family of smooth curves. Each sorptlon capacity curve showed a linear In-
crease 1n sorptlon capacity from 200°F to about 800°F each curve exhibited
a rapid Increase in sorption capacity which continued to 1000°F. The curves
show that, 1n the 1000°F to 1200°F region, the Increase 1n sorptlon capacity
was slight. Each of the regeneration ratio versus temperature curves 1s spe-
cific with respect to the Initial regeneration ration (Y axis intercept) and
rate of change In the 200°F to 800°F region. The Initial sorptlon values
(plotted on the ambient temperature scale value) represent the used sorptlon
capacity of Flltrasorb #400 virgin carbon and the slope of each curve describes
how rapidly the used capacity Is regained as a function of temperature.
The most interesting features of the curve, shown In Figure 20, 1s the
threshold response to temperatures exhibited by all the curves at about 800°F
and the convergence of the curves at 1200°F. These conditions predicate that
69
-------
1.00-
o»
«
oc
o
oa
o
9
9
V
0)
CK
200 400 600 800
Temperature (°F)
IOOO 1200
Figure 20. Effect of specific sorption capacity on
regeneration.
70
-------
800°F is a realistic minimum value for thermal vacuum regeneration and show
that maximum gain in regeneration capacity per degree temperature is achieved
in the 800°F to 1000°F range. Further consideration of Figure 20 suggests
that rate of response, i.e., the increase in regenerated capacity as a func-
tion of temperature, increased with increasing loading capacity. This fea-
ture was selected for additional study and is discussed in the following
section.
Plotting the data displayed in Figure 20 as regeneration ratio versus
specific sorption capacity, using temperature as an independent variable,
yielded the family of curves shown in Figure 21. The term, specific sorp-
tion, developed for use in this analysis was calculated by determining the
difference in sorption capacity occasioned by saturation with methylene
blue. Thus, the specific sorption values plotted in Figure 21 represent
the spent sorption capacity of the Filtrasorb #400 sorbent.
An interesting finding, illustrated in Figure 21 is the linear rela-
*-•
tionship of regeneration gain versus temperature exhibited by each curve.
The effect of specific sorption capacity is shown to be most pronounced
in the low temperature region (300°F - 500°F) decreasing in the 800° region
and decreasing further in the 1000°F to 1200°F region. The most interesting
feature displayed by the data plotted in Figure 21 is the linear relation-
ship of temperature, regeneration capacity, and specific sorption capacity.
The relationship between sorption capacity and regeneration ratio Is de-
scribed by the following
Rr = 1 - (£) sp Mt
where
Rr = regeneration ratio
y
(—)sp = specific sorption capacity, mg/mg
Mt = temperature rate coefficient.
71
-------
I200°F (M = 0.067)
? .80-
o»
«
oc
o
o:
c
o
0
-------
The temperature coefficient term, Mt, 1s an empirical constant and spe-
cific for each waste sorbate-sorbent combination, temperature region, and
sorbate loading value. Seemingly the Mt term 1s somewhat similar to a rate
constant. A test of this idea was developed by plotting Mt values versus
temperature, Figure 22. Although the curve shown 1n Figure 22 exhibits sig-
nificant curvature, it indicates a proportional relationship between Mt and
temperature in the 70°F to 800°F region. Moreover, the shape of the curve
presented 1n Figure 22 indicates that the Mt term depicts the sorbate loss.
Consequently direct comparison of the Mt and Rr values for a specific sor-
bate loading value, should show similar changes. The curves shown in Figure
23 illustrate that this relationship is valid. In the 70°F to 1200°F region,
the Mt versus temperature curve is an inverse reflection of the Rr versus
temperature curve. Thus, the Mt term is a measure of sorbate removal and
the Rr term is a measure of regeneration.
Collectively the results of the Filtrasorb #400 - methylene blue study
showed that thermal regeneration proceeded in a regular progression depen-
dent upon temperature and sorptlon loading. Moreover, the gain In regener-
ated sorption capacity appeared to be directly related to the specific sorp-
tion capacity. As an aid to development of data analysis, a composite term,
regeneration efficiency (Re), was devised. The purpose of introducing the
Re term is to minimize the number of factors and to develop and expression
relating regeneration response and temperature. The Re expression combines
the former Rr and specific sorption loading term into a single expression
defined by the following relationship
(X/M) reg - (X/ML
Re =
(X/M) vlrg
73
-------
4.00 —
0
"o
0>
o
o
a>
ci
(T
3
«••
O
k.
«
o.
E
0 200 400 6OO 8OO 1000 1200
Regeneration Temperature (°F)
Figure 22. Temperature rate coefficient versus
temperature.
74
-------
o
ct
&
1.00
.90
.80
70
.60-
.50
.40
.30-
.20
.10
4.0O
3.OO
2.OO
I.OO
0 200 400 600 800 IOOO I20O
Temperature (°F)
Figure 23. Comparison of M_ and regeneration
response.
75
-------
where
X/M reg = regenerated sorption capacity, gm/gm
X/M virg = virgin carbon sorption capacity, gm/gm
(X/M)Q = residual sorption capacity, gm/gm
The relationship between the initial measurement of sorption, regeneration,
regeneration ratio (Rr), and regeneration efficiency is
Re = Rr - RrQ
where Rr is the regeneration ratio normalized to residual sorption capacity.
The above relationship requires that Re equal Rr when the residual sorption
capacity is zero.
A graphical presentation of regeneration efficiency (Re) versus tempera-
ture for methylene blue sorbed on Filtrasorb #400 and Nuchar WVL activated
carbons is shown in Figure 24. The most interesting and potentially useful
finding illustrated 1n Figure 24 is the similar linear increase in regenera-
tion efficiency as a function of temperature exhibited by Filtrasorb #400 and
Nuchar WVL activated carbon. Both sorbent materials show a similar response
to the thermal regeneration response curve. Specifically, Nuchar WVL material
showed an increase of 53 percent in the 100°F to 500°F temperature ranges;
whereas, Filtrasorb #400 material evidenced a 54 percent increase. The one
percent difference in response is well within the range of experimental error
and is probably not significant. It 1s apparent that both sorbent materials
increased 1n sorption efficiency at about the same rate.
The general shape of the thermal regeneration response curves, shown
In Figure 24, suggests that there 1s a dissimilar response characteristic of
each type of carbon. Generally, the inflection point common to each sorbent,
500°F for Nuchar WVL and 800°F for Filtrasorb #400, is the threshold tempera-
ture value for thermal regeneration, the specific response of each carbon at
76
-------
lOOh
oc
«»
>»
o
UJ
c
o>
«
o:
V
O.
'100
1000
2OOO
Temperature (°F)
Figure 24. Methylene-blue regeneration efficiency.
77
-------
temperatures greater than the threshold value are different. The Filtrasorb
#400 material showed a slightly increased rate of regeneration response in
the high temperature region, whereas, the Nuchar WVL material showed a de-
creased rate of response. In both cases the change from the initial linear
course is slight and both curves coincide at about 1000°F.
In order to test the fidelity of the response pattern and the data anal-
ysis process, the thermal response data for iodine, petrochemical waste, and
para-chlorophenol were examined. Results of this operation, Figures 25, 26,
and 27, show that the thermal regeneration response was similar to the re-
sponse pattern of methylene blue.
All of the thermal regeneration response curves shown in Figures 25, 26,
and 27, describe a family of curves very similar to the methylene blue thermal
regeneration prototype (Figure 24). There is an obvious upward displacement
of the iodine, parachlorophenol and petrochemical waste curves directly pro-
portional to the specific sorption capacity of each material. The most atypi-
cal regeneration response pattern is illustrated by iodine. Iodine is a unique
material in that it is not decomposed during thermal regeneration, rather it
vaporizes. Consequently, the response pattern of iodine is indicative of the
behavior of a volatile non-reactive sorbate. On this basis, it seems proba-
ble that the initial portion of each thermal regeneration response curve il-
lustrates the volative loss of sorbate during thermal regeneration.
Comparison of the thermal regeneration response patterns shown in Figures
24, 25, 26, and 27, reveals consistent response. In terms of temperature re-
quired for regeneration, the carbon-iodine samples achieved full regeneration
at about 300°F, whereas, carbon-para-chlorophenol, carbon-methylene blue, and
carbon-petrochemical wastes required 800°F, 1000°F and 1000°F resepctively.
The rate of regeneration response exhibited by each of the sorbate materials
was not related to the threshold temperature. Comparison of the rate values
78
-------
100
o
«c
«.
UJ
c
_o
o
c
0>
oc
s.
10
= .30T2-°°)
Nuchor FS 400
100
1000
2000
Temperature (°F)
Figure 25. Iodine regeneration efficiency.
79
-------
100-
9
|5
•fc.
-------
100-
UJ
o
o
s.
100
Temperature (°F)
1000
20OO
Figure 27. Chlorophenol Regeneration Efficiency
81
-------
listed 1n Figures 24, 25, 26, and 27 shows that the carbon-petrochemical
waste material regenerated at the fastest rate, followed in order by para-
chlorophenol, methylene blue, and iodine. The specific rate of thermal
regeneration, as a function of temperature, is apparently unique for each
waste sorbate; however, the rate values for each waste-carbon combination
range from 1.21 to 2.00. The differences exhibited by the rate values are
not particularly distinctive. Rather, the similarity in these values sug-
gests that the initial portion of the thermal regeneration process proceeds
at about the same rate irrespective of the sorbate materials.
On the basis of the above considerations the vacuum thermal regenera-
tion response can be described by the following relationship
Re = KTn
where
Re = regeneration efficiency, percent
K and n = empirical constants
T = temperature, °F.
The above relationship predicates that vacuum thermal regeneration is
an exponential function of temperature. The empirical constants associated
with temperature, K and n, are directly proportional to specific sorption
values and rate of regeneration response, respectively. The relationship
proposed in this equation is approximate since it does not account for the
inflection point which occurs at the threshold temperature. However, the
inflection point, or threshold temperature, is specific and unique for each
waste sorbate-carbon combination. Moreover, the equation is significant in
that it shows: (1) overall response, in terms of thermal regeneration re-
sponse, as an exponential function of temperature and (2) the relationship
between specific sorption capacity, and/or loading rate of regeneration re-
sponse.
82
-------
SECTION 8
DIFFERENTIAL THERMAL ANALYSIS OF ACTIVATED
CARBON REGENERATION
The purpose of this portion of the study was to determine 1f differen-
tial thermal analysis could be used to evaluate activated carbon sorption in
terms of temperature and diagnostic exothermic-endothermic reactions.
Activated carbon Is a proprietary material and the specific physico-
chemical reactions requisite to activation are considered to be well guarded
trade secrets. Regeneration of spent activated carbon has been achieved using
a variety of processes and techniques ranging from solvent extraction to high
temperature pyrolysis. The mechanism of thermal regeneration of activated car-
bon is probably related to activation since both processes are accomplished by
heat treatment. Ideally the thermal regeneration of spent activated carbon
provides a material for reapplicatlon with the same sorption affinity and ca-
pacity as the parent material. Field and laboratory investigations have shown
that thermally regenerated activated carbon achieves about 95 percent original
5--
sorption capacity per regeneration cycle together with a 5 percent loss of ma-
terial. In general, the behavior and characteristics of the regenerated car-
bon are very similar to virgin activated carbon except for increased density
of the regenerated carbon.
The most significant problem associated with the large scale application
of activated carbon is determination of sorption capacity depletion and regen-
eration response. Traditionally sorption capacity has been determined by analy-
sis of sorbate solution rather than by direct analysis of the sorbent.
The major Interest in this study was directed toward development of analy-
tical procedures based on differential thermal analysis (DTA) and thermal char-
acteristics of the activated carbon sorbate which could be used to detail sorp-
tion reactions and/or thermal regeneration. In order to minimize the complexity
83
-------
of thermal regeneration a basic low pressure thermal regeneration procedure
was employed in this study.
Differential thermal analysis (DTA) is a unique form of analysis. In
concept, DTA produces a series of patterns which represent the enthalpic re-
actions specific to selected materials or substances during heating or cooling.
In this context, DTA is basically a thermal spectrometer, since the output data
usually consists of a series of peaks representing intensity vs thermal energy
level. Previous study of DTA has demonstrated that the enthalpic changes or
reaction exhibited by samples of pure compounds, materials and mixtures vs.
a stable inert thermal standard such as SiOp constitute unique qualitative
characteristics of the sample. Moreover the quantitative characteristics,
such as sample mass or activity are frequently proportional to peak area.
Thus DTA is potentially useful for detailing the qualitative-quantitative
features of sample materials in a single determination.
The basic components of DTA are; thermal detector, signal amplifier,
variable rate of heating furnace and data display system. Most of the com-
mercially available DTA machines provide a variable rate of heating furnace
and direct data display. Usually the data output is protrayed as a thermo-
gram, a series of endothermic-exothernic peaks plotted against temperature of
the furnace. The basic differences 1n DTA machines are; the quality of con-
struction, design of the furnace temperature control (heating rate) and data
display system design. Basic DTA machines are designed to directly record
temperature differences between sample and standard as a function of fixed
heating rates; whereas, research DTA machines are equipped with variable
heating rate controls and provision for heating samples under low pressure
(vacuum) and inert gas atmospheres. In addition, the research machines pro-
vide data output via an XY recorder. The advantage of the XY recorder sys-
tem is that the temperature differences (delta temperature - AT) is plotted
84
-------
directly as a function of the reference standard.
Differential thermal analysis was selected for application to this study
because 1t offered the following advantages:
1) Provided a qualitative - quantitative method for monitoring
thermal reactions related to sorptfon and thermal regeneration.
2) All of the requisite parameters relative to thermal regenera-
tion were available 1n one Instrument.
3) Promised a mechanism for developing direct measurement of
thermal reactions unique to thermal regeneration of activated
carbon.
85
-------
SECTION 9
METHODS, MATERIALS, AND PROCEDURES
In accordance with the objectives of this study, DTA was employed to
detail selected sorbate uptake and experimental determination of thermal re-
generation. The general plan of study developed for this investigation con-
sisted of a series of broad range preliminary exploratory determinations fol-
lowed by narrow range definitive determination. The first portion of the study
was directed towards developing qualitative data relevant to DTA response and
the thermal regeneration response of activated carbon sorbed materials.
The activated carbon used in this study was Filtersorb 400*. This mater-
ial was initially applied in granular form, however analysis of this material
required that the grain size be reduced to about 40 mesh. This factor was
established by comparing the DTA thermograms developed by various sized grains
of Filtrasorb 400. Results of this study showed that granular activated carbon
= 1/8 - 3/10" size produced a variable geometry in the DTA sample holders which
resulted in spurious response. The sample cups employed in this study were
capable of containing 5 and 50 mg of activated carbon. The relatively small
size required that homogenous small grained sample materials be used. On
the basis of this finding a modified method of sorption contact was developed
using a thin film of 40 mesh activated carbon.
In order to develop quantitative data relative to DTA response, a series
of organic materials were employed as standards and the appropriate peak area
for selected endothermic-exothermic reactions were measured. The selected
organic materials were benzoic acid, phenol and methylene blue. These ma-
terials were selected on the basis of established thermal standard charac-
teristics, sorption affinity for activated carbon and analytical chemistry
86
-------
response. Benzole acid is a primary standard employed in calorimetry; whereas,
methylene blue and phenol are readily sorbed by activated carbon and their
concentration can be determined by colorimeteric or spectro-photometric analysis.
DTA was initiated by exploring the thermal characteristics of activated
carbon. The apparatus selected for this work was a Model 202 DTA system man-
ufactured by the R. L. Stone Company*. This system was equipped with 3/8" dia-
meter 1/2" deep cylindrical nickel sample holders equipped with internal con-
tact thermocouples and a 1/4" diameter 0.010" thick platinium disk sample
holder supported on position indexing ring thermocouples. Each of the sample
holders were placed in a pressure furnace module equipped with water cooling
coils. A diagram of the thermocouples and furnace units are shown in Figure
28. This instrument was also equipped with a variable heating rate programmer,
temperature limit switch and XY data plotter. The furnace unit for this in-
strument was constructed as a pressure vessel and used for low and high pres-
sure applications.
Preliminary testing of the DTA system showed that activated carbon sam-
*-•
pies exhibited exothermic peaks in the 500-700 C region when an inert gas,
argon at 20 psi gauge, was used. Further study showed that the exothermic
peak (evidence of combustion) was also produced under evacuated pressure con-
ditions (10 mm Hg). Experimental evaluation of this situation showed that
the exothermic peak formation could be eliminated by increasing the low pres-
sure (10 mm Hg) degassing period to 30 minutes or by employing three five-
minute serial pruges of the furnace with argon. During the course of this
preliminary study a variety of activated carbon samples representative of
several trade names were subjected to DTA to determine if the persistent
exothermic reaction was related to the product or manufacturing process.
Results of the study, Figure 29, showed that the exothermic reaction was
common to all samples. Since low pressure degassing and pruging eliminated
87
-------
DTA FURNACE
Sample Holder
Got
Platinum Dish
Reference Thermocouple
Bead
Differential Thermocouple
Bead
DTA Sample Holder
Ring Thermocouples
Figure 28. Diagram of DTA apparatus
-------
Activated Carbon (standard) vsActivated Carbon (sample) 5.0mg samples;
Temperature increase IO°C/min IOmm Hg pressure
lOmin purge
20min purge
30 min purge
40min purge
I
I
ZOO
400 600
Temperature, °C
800
1000
Figure 29. Effect of low pressure (10 mm Hg) degassing period on high
temperature exothermic reaction of activated carbon.
89
-------
the problem it seems likely that sorbed oxygen was released from the activated
carbon in the 500-700 C range. The balance of DTA work conducted in this
study used either a preliminary 30 minute 10 mm Hg pressure degassing period
for low pressure determinations or three 5 minute duration purges with Argon
gas.
Qualitative thermal characteristics of activated carbon were determined
under differential thermal conditions using a variety of sample materials,
thermal standards and heating rates. Initially, basic standard materials
such as aluminum oxide, silicon dioxide and quartz flour were employed as
thermal reference standard materials in conjunction with virgin activated
carbon samples. The objective of this exploratory study was to determine
the baseline characteristics of the activated carbon under low pressure and
inert atmosphere heating regimes. Results of this study showed that the
standard reference materials provided conditions requisite to achievement
of a well defined flat baseline; however, the standard thermal materials
such as aluminum oxide, silicon dioxide and quartz flour diminished the
sensitivity of the system to exothermic-endothernric responses. Further ex-
perimentation showed that the system of using virgin activated carbon as a
thermal standard in conjunction with a sample of virgin activated carbon
produced a relatively smooth baseline. In this arrangement the small dif-
ferences in sample composition, grain size etc. were diminished and differ-
ences in thermal characteristics are recorded directly. Additional experi-
mental study showed that the activated carbon (standard) vs. activated car-
bon (sample) system was the most sensitive on the basis of peak height re-
sponse per unit mass of material. Moreover this system was most realistic
for this study since the thermal energy difference between virgin activated
carbon, organic sorbed activated carbon and regenerated activated carbon could
be determined directly.
90
-------
Following determination of baseline development and sample response, a
systematic study of sample geometry and heating rate effects on repsonse was
completed. Results of this study showed that the small sample (= 3 to 8 mg)
located on the 1/4" diameter platimium foil crucibles - ring thermocouple
sample holders provide optiminium qualitative response; whereas, the 3/8" dia-
meter 1/2" deep nickel crucibles - contact thermocouples sample holders pro-
vided the most consistent quantitative response. Direct comparison of low
pressure thermograms for various samples of activated carbon and organic ma-
terials showed that heating rates ranging from 10 to 12°C per minute were nec-
essary for this work.
Evaluation of the quantitative nature of the DTA method and apparatus
was accomplished by determining the relative peak area developed by various
sized samples of benzoic acid. Results of this determination are shown in
Figure 30.
91
-------
3.0
2.0
1.0
00
Benzole Acid
Endofherm
I
5 10
Relative OTA Endothermic Peak Area
Figure 30. Quantitative response of DTA - bassed on
benzoic acid standard.
92
-------
SECTION 10
RESULTS
The broad range exploratory studies directed toward determining the
factors affecting differential thermal analysis of activated carbon showed
that sample grain size, sample preparation and heating rate were the domi-
nant factors. Specifically the results of this study showed that:
1) activated carbon samples must be reduced to 40 mesh grain size
2) activated carbon samples must be degassed for 30 minutes at 10 mm
Hg pressure prior to differential thermal analysis and
3) a minimum heating rate of 10°C/min and maximum of 15°C/min was
necessary.
Experimental study of the thermal characteristics of phenol sorbed
activated carbon produced a typical response pattern shown in Figure 31.
Specifically the phenol-activated carbon shown in Figure 31 indicated that
the phenol sorbate produced a sharp well defined endothermic peak in the
300-310 C range. The characteristic endothermic pattern produced by phenol
remained very nearly constant and demonstrated that the phenol was removed
from the activated carbon in the 300-310°C range.
Further study of the phenol-activated carbon thermal disassociation
characteristics by DTA showed that the 300-310°C region was the threshold
temperature for thermal regeneration. This feature was established by
measuring the sorption uptake of phenol following thermal regeneration of
spent (phenol sorbed) samples of activated carbon. The response of this
system in terms of sorption uptake vs. regeneration temperature values plotted
in Figure 32 show that sorption regeneration remains at a constant low value
in the 23°C to 110 C temperature range and increased linearly with tempera-
ture to the 310°C region, thereafter it remained constant to about 500°C.
93
-------
AT
Activated Carbon (Std) v$ Phenol sorbed Activated Carbon (sample)
Temperature increase, IO°C/min;IOmmHg pressure; 5mg samples
ir
•V
0.25 mq phenol
0.68mg phenol
0.95mg phenol
50 100 150 200 250 300
Temperature, °C
350
4OO
Figure 31. Phenol sorbed activated carbon DTA thermogram.
94
-------
Temperatures greater than 500°C reduced the regeneration sorption capacity.
Also shown in Figure 32 are the results of a temperature equilibrium study
of phenol sorbed-activated carbon thermal regeneration. In this study phenol-
sorbed activated carbon was heated to a selected temperature value and held at
this value for 10 minutes. Results of this study plotted in Figure 32 showed
a similar response to the continuous DTA results.
The most interesting feature of the DTA thermal regeneration study is the
relatively high temperature required for thermal regeneration. Pure phenol is
a solid at room temperature and exhibits a broad range melting point at about
42-43°C and a sharp boiling point at 181.5°C. Moreover phenol exhibits a
relatively high vapor pressure and is a deliquiesent matertial. On the basis
of these characteristics, if phenol were weakly sorbed to activated carbon
it would be expected to be displaced from the activated carbon sorbent at
some temperature value close to phenol boiling point value. At a low pres-
sure value of 10 mm Hg, the boiling point value for phenol is 73.5 C; signifi-
cantly less than the 310°C value determined by the DTA study.
3
Experimental study of the high temperature characteristics of activated
carbon sorbed phenol by DTA established that the temperature ranges determined
in the thermal regeneration were realistic. Experimentation with various
sized samples of phenol and phenol sorbed activated carbon verified the boiling
and melting point values for pure phenol and showed that activated carbon
sorption displaced the melting point endotherm pattern from 182 to 310°C.
This finding suggests that the amount of thermal energy requisite to
thermal regeneration of phenol sorbed activated carbon might be propor-
tional to the temperature difference between the boiling point for pure
phenol and the threshold thermal regeneration point. Verification of this
relationship was considered to be beyond the scope of this study and fur-
ther study of this relationship is required.
95
-------
100
o
JO
o
o
•I 75
Conventional DTA
O
a
(A
0
or
o
50
S 25
0
or
in Thermol
equil. period
100 200 3OO 400 500
Regeneration Temp. ° C
600
700
800
Figure 32. Regenerated sorption capacity developed by DTA study
of phenol-activated carbon system.
96
-------
Experimental study of the methylene blue-activated carbon system
showed a thermal response pattern somewhat similar to the phenol sorbed
activated carbon thermal regeneration study. The significant exceptions
were; 1) methylene blue did not exhibit a sharp melting point rather it
decomposed on heating, 2) Methylene blue thermograms developed by low pres-
sure (10 mm Hg) heating exhibited broad exothermic patterns similar to the
curve shown in Figure 33. Samples of methylene blue sorbed activated car-
bon subjected to low pressure (10 mm Hg) and evaluated for sorption uptake,
Figure 34, yielded thermograms similar to the curves shown in Figure 33. As
in the phenol sorbed-activated carbon regeneration study, the threshold tem-
perature of methylene blue thermal decomposition determined by DTA established
the lower limit temperature for thermal regeneration. This factor was reeval-
uated independently during the course of the thermal regeneration study of
methylene blue-activated carbon system. Results of the DTA thermal regenera-
tion study, shown in Figure 34, predict that thermal regeneration in the 100°C
to 800°C region would progress exponentially. Plotted values in Figure 34
also showed that thermal regeneration above 800°C was not effective in re-
storing sorption capacity. Generally the efficiency of sorption regenera-
tion as a function of temperature yields a response directly proportional to
the thermal reaction curve derived from activated carbon-methylene-blue sorbed
carbon was significantly different from phenol sorbed carbon.
97
-------
46.Omg Carbon A+4mg Methylene Blue
S.Omg Carbon A+2 mg Methylene Blue
4mgCarbon B+9.6mg Methylene Blue
Carbon B + l2.5mg Methylene Blue
IO°C/min Temp. Rate
10 mm Hg Pressure
Act Carbon (Std.)v* Act.Corbon + Methylene Blue (Sample)
100 200 3OO
4OO 500 600
Temperature, °C
700 800 900
Figure 33. Methylene blue sorbed activated carbon DTA thermogram.
98
-------
100-
75
o
v»
o
o
(T
50
25
1472 °F
100 200 300 400 5OO 600
Regeneration Temp.
700 800 °C
Figure 34. Regenerated sorption capacity developed by DTA study
of methylene-blue activated carbon system.
99
-------
SECTION 11
SUMMARY
Results of the DTA study of thermal regeneration of activated carbon
under low pressure conditions showed that DTA is a realistic and promising
tool for detailing qualitative and quantitative regeneration response.
Comparison of the results derived from the DTA study of thermal regen-
eration with the results derived from the bench scale vacuum regeneration
study showed that the sorption regeneration response was similar and that
the temperature limits were similar. In addition the results of both studies
showed that various sorbates exhibit unique thermal properties and require
systematic study.
Potentially DTA provides a method for studying thermal regeneration on
a direct determination of sorbent uptake capacity basis. Tentatively this
method could be used to determine specific temperature response of thermal
regeneration as a function of sorption loading or pre-treatment conditions.
The direct application of DTA to the analytical problems involved in activated
carbon sorption and thermal regeneration require careful consideration of the
sorbate and probable conditions. Optimum utilization of DTA for this purpose
may require development of special techniques and DTA instrumentation.
100
-------
REFERENCES
1. Joyce, R. S., J. B. Allen and V. A. Sukenlk. "Treatment of Municipal
Wastewater by Packed Activated Carbon Beds", J.W.P.C.F. Vol. 38, No. 5,
May 1966. pp. 813-823.
2. Parkhurst, John D., Franklin D. Dryden, Gerald N. McDermott and John
English. "Pomona Activated Carbon Pilot Plant", J.W.P.C.F. Vol. 39,
No. 10, Part 2. October 1967, pp. R70-R81.
3. Weber, Walter J., Jr., Charles B. Hopkins and Ralph Bloom, Jr. "Physico-
Chemical Treatment of Wastewater", J.W.P.C.F. Vol. 42, No. 1, January
1970, pp. 83-99.
4. Evans, David R. and Terry C. Wilson. "Capital and Operating Cost - AWT",
J.W.P.C.F. Vol. 44, No. 1, January 1972, pp. 1-13.
5. Sawyer, George, "new Trend in Wastewater Treatment and Recycle", Chemical
Engineering. Vol. 79, No. 16, July 24, 1972, pp. 120-128.
6. Cohen, J. M. and I. J. Kugelman. "Physical-Chemical Treatment for Waste-
water", Water Research. Vol. 6, No. 415, April/May 1972, pp. 487-492.
7. Shuckrow, A. J., W. F. Bonner, N. L. Presecan and E. J. Kazmierczak.
"A Pilot Study of Physical-Chemical Treatment of the Raw Wastewater at
the Westerly Plant in Cleveland, Ohio", Water Research. Vol. 6, No. 415,
April/May 1972, pp. 619-626.
8. Friedman, L. D., R. Bloom, Jr., and W. J. Weber, Jr. "Think Physico-
Chemical", Water and Waste Engineering. Vol. 9, No. 9, September 1972,
pp. 67-68 and 79-80.
9. Villiers, Noland V. "Single Stage Process Checks Out", Water and Waste
Engineering. Vol. 9, No. 1, January 1972, pp. 32-35.
10. Bishop, Dolloff F., Thomas P. O'Farrell and John B. Stamberg. "Physical-
Chemical Treatment of Municipal Wastewater", J.W.P.C.F. Vol. 44, No. 3,
March 1972, pp. 361-371.
11. Peoples, R. F., P. Krishna, and R. N. Simonsen. "New Biological Treat-
ment of Refinery Wastewater", J.W.P.C.F. Vol. 44, No. 11, November 1972,
pp. 2120-2128.
12. Huang, J. C. and M. G. Hardie. "Wastewater Reclamation by Physico-Chemical
Treatment of Wastewater", Pollution Abstracts. Vol. 44, No. 7, July 1971,
pp. 327.
13. Knopp, Paul V. and Wayne B. Gitchell. "Wastewater Treatment with Powdered
Activated Carbon Regenerated by Wet Air Oxidation", Proc. 25th Ind. Waste
Conf., Purdue University Extn. Service, pp. 687.
101
-------
14. Shell, G. L. and D. E. Burns. "PAC-PCT Process for Wastewater Treat-
ment", Public Works. Vol. 103, No. 2, February 1972, pp. 71-73.
15. Juhola, A. J. "Optimization of the Regeneration Procedure for Granular
Activated Carbon", Chemical Abstracts. Vol. 77, No. 12, September 18,
1972, pp. 284, 791230":
16. Urano, Roher. "Absorbents and Ion-Exchange Materials from Coal, II.
Regeneration of Activated Carbon Used for Sewage Treatment", Chemical
Abstracts. Vol. 77, No. 8, August 21, 1972, pp. 250, 52086S.
17. Hoches, Charles F. and Stanton B. Smith. "Reactivate Powdered Carbon",
Chemical Engineering. Vol. 79, No. 9, May 1, 1972, pp. 46-47.
18 Kampt, H. J. "Repeated Use of Granular Activated Carbon in Wastewater
Treatment", Water Research. Vol. 6, No. 4/5, April/May 1972, pp. 493-494.
19. Shriba, M. C. and J. C. Purcupile. "Powdered Carbon Regeneration by
Immiscible Fluid Displacement", Unpublished Study.
20. Wills, G. B., P. B. Rept. 202628, (1971). U.S. Nat'l Tech. Inf. Ser-
vice, Springfield, Va.
21. Azuma, Kawazoe and Matushima Hitomi. "Treatment of Waste from Plating
by Use of Activated Coal", Chemical Abstracts. Vol. 77, No. 6, August 7,
1972, pp. 306, 38862K.
22. Kalinske, A. A. "Enhancement of Biological Oxidation of Organic Wastes
using Activated Carbon in Microbial Suspensions", Vol. 119, No. 6, June
1972, Water and Sewage Works, pp. 62-64.
23. Browning, John E. "New Water-Cleanup Roles for Powdered Activated Carbon",
Chemical Engineering. Vol. 79, No. 4, February 21, 1972, pp. 36-40.
24. Van Stone, G. R. "Treatment of Coke Plant Waste Effluent", Iron and Steel
Engineering. Vol. 49, No. 4, April 1972, pp. 63-66.
25. Porter, John J. "Concepts for Carbon Adsorption in Waste Treatment",
Journal of the American Association of Textile Chemists and Colorists.
Vol. 4, No. 2, February 1972, pp. 29-35.
26. Porter, John J. "Treatment of Textile Waste with Activated Carbon",
American Dyestuff Reporter. Vol. 61, August 1972, p. 24.
27. Thompson, C. S., James Stock and P. L. Mehta. "Cost and Operating
Factors for Treatment of Oily Wastewater", Oil and Gas Journal. Vol. 70,
No. 47, November 20, 1972, pp. 53-56.
28. Matsumura, T. and T. Ishiyama. "Decontamination of Radioactive Nitrosyl
Ruthenium Complexes by Activated Carbon", Nuclear Science Abstracts. Vol.
26, No. 4, February 29, 1972, p. 630. '
29. Weber, W. J., "Physical-Chemical Process for Water Quality Control",
Wiley-Interscience, New York, 1972.
102
-------
30. Rodman, C. A. and E. L. Shunney, "Bio-Regenerated Activated Carbon
Treatment of Textile Dye Wastewater," U.S. Environmental Protection
Agency Publication No. 12090DWM01/71, Washington, D.C., January 1971.
31. Hassler, J. W., "Activated Carbon," Chemical Publishing Company, Inc.,
171-193, New York, New York, 1963.
32. Hassler, J. W., "Proceeding of the Anthracite Conference," Pennsylvania
State University, October 18, 1956.
33. Giusti, D. M., R. A. Conway, and C. T. Lawson, "Activated Carbon Ad-
sorption of Petrochemicals," J.W.P.C.F. Vol. 46, No. 5, May 1974, pp.
947-965.
34. Valentine, M., "Vacuum-Thermal Regeneration of Activated Carbon,"
Unpublished M;S. Thesis, University of Missouri-Columbia, May 1974.
35. Pittsburgh Activated Carbon Company, "Basic Concepts of Adsorption on
Activated Carbon," Calgon Center, Pittsburgh, Pennsylvania.
36. Johnson, R. L. et al., "Evaluation of the Use of Activated Carbons
and Chemical Regenerants in Treatment of Wastewater," AWTR-11. U.S.
Public Health Service Publication No. 99-WP-13, May 1964.
103
-------
TECHNICAL REPORT DATA
ir'tnase read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-78-103
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
THERMAL REGENERATION OF ACTIVATED CARBON
5. REPORT DATE
May 1978 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO
Louis Hemphill
9. PERFORMING ORGANIZATION NAME AND ADDRESS
University of Missouri - Columbia
Department of Civil Engineering
Columbia, Missouri 65201
10. PROGRAM ELEMENT NO.
1BB610
11. CONTRACT/GRANT NO.
Grant No. 800554
12. SPONSORING AGENCY NAME AND ADDRESS
Robert S. Kerr Environmental Research Lab.
Office of Research and Development
U. S. Environmental Protection Agency
Ada, Oklahoma 74820
- Ada, OK
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/600/15
15. SUPPLEMENTARY NOTES
16. ABSTRACT
A three-part experimental study of activated carbon adsorption and thermal
regeneration has been completed. The project included an experimental pilot plant
thermal regeneration study to determine specific thermal regeneration character-
istics of selected petrochemical waste materials and a low-pressure thermal in-
vestigation. Results of these studies showed: (1) Small polar molecular species,
or species highly soluble in water, are resistant to carbon sorption. (2) Extreme
acidic or basic waste streams may require pH adjustment to promote carbon sorption.
(3) Batch isotherm values provide basic information relative to activated carbon-
petrochemical waste column design. (4) Vacuum regeneration of petrochemical
saturated activated carbon is effective and efficient. With most sorbent-sorbate
combinations tested, the carbon sorbent could be regenerated to 95% of original
sorption capacity. (5) Temperature required for carbon regeneration was a function
of waste type (composition) and sorption capacity.
The report describes experimental methods of differential thermal analysis,
vacuum thermal regeneration, isotherm and column derived sorption values, and
quantitative relationships of temperature and thermal regeneration response.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COS AT I Field/Group
Industrial wastes
Activated carbon
Petrochemistry
Adsorption
Pilot plants
Isotherms
Selected petrochemical
waste materials
Regeneration
68D
13. DISTRIBUTION STATEMENT
Release to public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
116
20. SECURITY CLASS (Thispage)
Ilnrl afigi f-fg»H
22. PRICE
EPA Form 2220-1 (9-73)
104
HWmKOfnCt 1971-757-140/6838
------- |