ENVIRONMENTAL HEALTH SERIES
Water Supply
and Pollution Control
ADSORPTION
OF BIOCHEMICALLY
RESISTANT MATERIALS
FROM SOLUTION. 2.
AWTR-16
Federal Water Pollution Control
Administration
-------
ADSORPTION OF BIOCHEMICALLY
RESISTANT MATERIALS
FROM SOLUTION. 2.
by
J. Carrell Morris and Walter J. Weber, Jr.
for
The Advanced Waste Treatment Research Activities
Cincinnati Water Research Laboratory
Robert A. Taft Sanitary Engineering Center
This publication, a combination of reports submitted by the authors in
June 1963 and August 1964, is a report on the continuation of studies
under Contract No. SAph 76295, between the Public Health Service and
the Division of Engineering and Applied Physics, Harvard University.
Federal Water Pollution Control Administration
Basic and Applied Sciences Program
Cincinnati, Ohio
March 1966
-------
The ENVIROMENTAL HEALTH SERIES of reports was
established to report the results of scientific and engineering
studies of man's environment: The community, whether urban,
suburban, or rural, where he lives, works, and plays; the air,
water, and earth he uses and re-uses; and the wastes he pro-
duces and must dispose of in a way that preserves these natural
resources. This SERIES of reports provides for professional
users a central source of information on the intramural research
activities of Divisions and Centers within the Public Health Ser-
vice, and on their cooperative activities with State and local
agencies, research institutions, and industrial organizations.
The general subject area of each report is indicated by the two
letters that appear in the publication number; the indicators are
WP - Water Supply
and Pollution Control
AP - Air Pollution
AH - Arctic Health
EE - Environmental Engineering
FP - Food Protection
OH - Occupational Health
RH - Radiological Health
Triplicate tear-out abstract cards are provided with reports
in the SERIES to facilitate information retrieval. Space is pro-
vided on the cards for the user's accession number and additional
key words.
Reports in the SERIES will be distributed to requesters, as
supplies permit. Requests should be directed to the organization
identified on the title page or to the Publications Office, Robert
A. Taft Sanitary Engineering Center, Cincinnati, Ohio 45226.
Public Health Service Publication No. 999-WP-33
-------
ADVANCED WASTE TREATMENT
RESEARCH PROGRAM
The Advanced Waste Treatment Research Activities of the Federal
Water Pollution Control Administration has two ultimate goals: One is to
help abate our Nation's growing water pollution problems, and the other,
more startling in concept, is to renovate waste water for direct and
deliberate re-use.
Conventional water and waste treatment practices have little or
no effect on many simple inorganic salts and permit an ever-increasing
number and amount of highly complex, synthetic organic wastes to con-
taminate drinking water supplies. If we are to protect these water
supplies for the future, it is mandatory that new technologies for water
and waste treatment be developed.
Advanced waste treatment may be looked upon as a two-step
process: (1) Separating concentrated contaminants from the purified
water "product/' and (2) disposing of the concentrated contaminants in
a way that will render them forever innocuous. This "permanent dis-
posal" of separated wastes is essential so that useable surface or
ground waters will not be contaminated and so that the same contamin-
ants need not be removed again and again from water supplies.
In one step, then, advanced waste treatment, i.e., water renovation,
could alleviate both water pollution and water supply problems — prob-
lems of increasing concern both in this country and in the rest of the
world. The importance of advanced waste treatment was recognized by
Congress in its 1961 amendments to the Federal Water Pollution
Control Act (Public Law 87-88). This act directs the Secretary of
Health, Education, and Welfare ". . . to develop and demonstrate
practicable means of treating municipal sewage and other water-borne
waste to remove the maximum possible amounts of physical, chemical,
and biological pollutants in order to restore and maintain the maximum
amount of the Nation's water at a quality suitable for re-use."
It is too early to predict accurately the cost of advanced waste
treatment technology, but the need is inevitable. Because of population
increase and industrial expansion, we are approaching the time when
our rivers and streams will no longer be able to assimilate our wastes
and when there will be no more developable supplies of fresh water.
Whatever the cost of this technology, one thing is certain — it must be
the very lowest science and engineering can achieve. Anything more
iii
-------
will impose an inexcusable, perhaps intolerable burden on our country's
economic growth and on its ability to compete in tomorrow's world.
The authors of this report collaborated on two previous reports in
this series (AWTR-2 and AWTR-9). This report, together with AWTR-9
(999-WP-ll), presents the results of 3 years of work on the contract by the
authors at Harvard University.
iv
-------
CONTENTS
Page
Abstract vii
Introduction 1
Adsorption in Columns of Fluidized Media 3
Experimental 4
Apparatus 4
Adsorbents and Adsorbates 4
Analytical Methods 6
Experimental Method 6
Results 7
Activated Aluminum Oxide 7
Activated Carbon 8
Solute Effects 13
Temperature Effects 18
pH Effects 19
Multicomponent Solutions . 21
Two Components 22
Three Components 22
Four Components 31
Eight Components S3
Self Regeneration of Activated Carbon 35
Limited Model for Qualitative Interpretation 35
Implications for Advanced Waste Treatment 43
Adsorption on Carbon in Batch-Type Systems 49
Rates of Adsorption 49
Adsorption on Pittsburgh Carbon 49
Adsorption from Mixtures of Organic Compounds . 51
Equilibria and Capacities for Adsorption 52
Effect of Particle Size 52
Isotherms for Different Carbons 53
Isotherms for Adsorption of Nitrochlorobenzenes
on Carbon 54
Isotherm for a Mixture of Organic Compounds ... 54
Adsorption of Organic Pesticides on Carbon 57
Experimental Details 58
Analytical Methods 58
Experimental Methods 68
Rates of Adsorption 70
Adsorption Equilibria 75
Examination of Effluents from Activated Carbon Columns .... 91
Total Organic Carbon 93
Chemical Oxygen Demand 93
-------
Page
Acid-Base Titrations 94
ABS Determinations 97
Miscellaneous Specific Tests 97
Discussion 98
References 101
Appendix A: Notations 105
Appendix B: Advanced Waste Treatment Research Publications 107
VI
-------
ABSTRACT
Earlier studies (reported in PHS Publication No, 999-WP-ll -
AWTR-9) showed that activated carbon for -waste water renovation could
best be used in continuous-flow columns. Such techniques should result
in an adsorptive capacity of greater than 10 percent. Results on studies
of adsorption of organics from single- and multi-component systems in
fluidized carbon are reported herein. The adsorbability of organic
pesticides on activated carbon was investigated in some detail. Studies
were undertaken to characterize those types of organic pollutants that
are not adsorbed on activated carbon.
vn
-------
ADSORPTION OF BIOCHEMICALLY
RESISTANT MATERIALS
FROM SOLUTION. 2.
Introduction
This is the second and final report, covering the last 2 years of
work, on a contract with Harvard University initiated in May of 1961.
The first year's work was covered in an earlier report (No. 999-WP-ll)
in this series. In our second contract year with the Advanced Waste
Treatment Research Activities, studies were directed toward the
definition and evaluation of the characteristics of adsorption from
aqueous solution in columns of fluidized media. During the third year,
the types of studies begun in the previous years were extended.
The first three sections of this report deal with the use of activated
carbon and activated aluminum oxide to remove organic contaminants
that might be found in a municipal secondary effluent. To minimize
analytical complications and allow closer interpretation of results, a
synthetic solution of one or more organic compounds in water was used
as the feed material.
To take advantage of the previously determined fact that the adsorp-
tion rate is inversely proportional to the size of the adsorbent granule,
the investigators utilized a fluidized column technique. This allowed the
use of small granules without the attendant problem of plugging that
would be encountered in downflow operation.
To supplement and strengthen the findings reported earlier on the
effect of adsorbent particle size on adsorption rate and capacity, the
authors conducted additional experiments on batch-type systems. These
studies also involved single- and multi-component systems.
Because of the increasing occurrence of organic pesticides in
water and waste water, the batch adsorption studies were extended to
include pesticides. Analytical procedures were developed for deter-
mining concentrations of various compounds typically found in commerci-
ally available herbicides, acaricides, and insecticides.
In the early work on adsorption of organics from municipal waste
water, a significant amount of the organic material, as measured by
chemical oxygen demand, remained unadsorbable in activated carbon
regardless of the amount of contact time. Identification of these unad-
sorbed materials was of interest, and so samples of carbon treated
effluent were concentrated up to 60-fold and sent to Harvard for exam-
-------
ination. A preliminary study of these concentrates indicated that the
organic content was composed primarily of partially oxidized substrates
and metabolites of microorganisms.
Because of the wide scope of the research and the vast number of
organic materials that might be involved, none of the research projects
that have been pursued, except possibly the determination of organic
carbon (not reported herein), have been concluded with definitive thor-
oughness. A great deal has been learned about the characteristics of
adsorption from dilute aqueous solution, but the information is not
complete enough for conclusive generalizations. Many gaps remain in
the data, and the implications of some studies have not been followed up
sufficiently. Numerous instances are pointed out in this report.
Nonetheless, the fundamental ground work has been laid for more
detailed and extended studies of adsorption by activated carbon on a
larger scale. Advantageous possibilities for the use of activated carbon
in adsorption have been shown; more practical studies are now needed
to define efficient and economical conditions and procedures of operation.
This does not mean that fundamental studies in this area should be
abandoned. New basic information and theory are always needed for
stimulation and understanding. These studies should be undertaken now
in smaller areas of interest, however, and with greater thoroughness
and refinement.
-------
ADSORPTION IN COLUMNS
OF FLUIDIZED MEDIA
Previous studies of the kinetics of adsorption from solution on
granular adsorbents in agitated nonflow systems have shown that particle
size is an important determinant of the rate of uptake of organic sub-
stances by an adsorbent.1 Within the range of experimentation in these
previous investigations, rate of uptake was found to vary as the inverse
square of the diameter of the adsorbent particle. Hence, for application
of adsorption to treatment of waters and waste waters, keeping particle
sizes as small as conditions of efficient operation allow appears desir-
able so that high rates of adsorption may be attained. Columnar opera-
tion in which the solution to be treated flows upward through an expanded
bed of fineparticulate adsorbent is one method of taking advantage of small
particle size while avoiding the problems of excessive headloss common
to standard packed-bed operation with fine media. Furthermore, other
problems common to packed beds, such as air binding and fouling with
particulate matter, are not usually encountered with fluid-bed operation.
Design of a continuously operating, countercurrent, fluidized bed should
be relatively simple. Fresh solid adsorbent may be added at the top of
the column and spent adsorbent withdrawn from the bottom on a continu-
ous basis, possibly by a lift-screw arrangement, while the solution from
which adsorption is taking place flows upward from partially utilized
adsorbent to freshly added material. Segregation of the adsorbent so
that nearly its full capacity can be realized with this form of operation
is aided by the fact that porous adsorbent particles — of activated car-
bon, at least ~ become denser with increasing adsorption, and tend,
therefore, to concentrate in the lower strata.
Fluidization as a method of columnar operation for achieving im-
proved fluid - solid contact has found most extensive investigation and
application in the areas of heat transfer,2-6 mass transfer in gas - solid
systems,2,5,7,8 and catalytic reactors.2,9,10 Because of its rather ex-
tensive use in the aforementioned applications, fluidization has also been
treated in great detail from the point of view of mechanics by a number
of investigators. 11-15 Reports of fluidized-column operations in which
the primary function of the solid has been adsorption have, however,
been few; even these few have dealt principally with gas-phase adsorption
and have thus had no particular significance for adsorption of organic
materials from aqueous solution. Perhaps the closest example has been
the use of a fluidized column of charcoal for removal of hydrocarbons
from a stream of gas.I6 Much of the work that has been carried out
with liquid - solid systems has dealt with the dissolution of a soluble
solid rather than with the uptake of a solute by the solid. 17,18
The kinetics of adsorption was investigated in terms of uptake
profiles relative to the velocity dimensions, time and length, for selected
values of these variables and for variously sized particles of adsorbent.
-------
Adsorption of Biochemically
Two different types of adsorbent, activated carbon and activated alumi-
num oxide, were investigated. In the initial column studies, only single-
solute solutions were used in order to facilitate interpretation of the
relative significance of more fundamental variables. Later, column
studies were undertaken with multicomponent solutions to observe how
solutions more nearly resembling waste waters behave in continuous-
flow operation.
Experimental
APPARATUS
A schematic representation of the experimental equipment is
shown in Figure 1. The column, a 1-inch-diameter Luc it e tube, 3 feet
in length, is fitted at the bottom with a velocity-head dissipation chamber
filled with 2-millimeter glass beads. In runs in which a greater quantity
of carbon was used, longer columns were needed to allow ample head
space for expansion. A 140-mesh stainless-steel screen separates the
adsorbent in the column from the glass beads in the dissipation chamber.
Sampling ports are positioned at 3-inch intervals along the length of
the column, flow to the column is regulated by means of two valves,
one of which is a needle valve, in the 0.5-inch-diameter feedline from
the constant-head tank. A float valve in the constant-head tank controls
the flow from two 200-liter plastic storage reservoirs. Flow from the
top of the column passes through an air gap, to prevent siphoning and
pressure fluctuations in the column, through a 1-liter settling tank, for
removal of any fine material that may have been carried over into the
effluent, and into one of several volume-calibrated 200-liter collection
reservoirs.
ADSORBENTS AND ADSORBATES
hi most of the column systems studied, Columbia* LC carbon was
used as the adsorbent because this material had been thoroughly tested
in earlier batch-type studies* and thus provided a logical starting point
for extension of the present investigations to column-type operation.
Activated carbon obtained from the Pittsburgh Chemical Co. was also
used in several of the studies to provide some basis for comparison of
different types of carbon.
A number of column studies were also carried out with an activated
aluminum oxide ("regular-iron" Porocel from Minerals and Chemicals
Philipp Corp.). This material was found, however, to be less effective
than either of the two types of activated carbon.
Before use, each adsorbent was separated into uniform particle
sizes by thorough sieving; portions of suitable size range were then
washed In distilled water to remove leachable impurities and adherent
powder and were next dried at 105°C. The size chosen for most ex-
tensive study consisted of particles passing a United States Standard
* Mention of commercial products does not imply endorsement by tfte Federal Water Pollution
Control Administration.
-------
Resistant Materials from Solution
Sieve No. 50 and retained on a No. 60; the mean particle diameter for
this size range is 0.273 millimeters.
200-liter STORAGE
RESERVOIRS
20-liter CONSTANT-HEAD TANK
- FLOAT VALVE
l-in.-DIAM, 3-ft LUCITE COLUMN
SAMPLE PORTS
AIR GAP
1-liter SETTLING
TANK
VELOCITY- HEAD
DISSIPATION
CHAMBER
NEEDLE VALVE
200-liter COLLECTION RESERVOIR
Figure 1. Schematic diagram of experimental apparatus.
The adsorbates used in these studies included a 92.9 percent
active material, dodecylbenzenesulfonate, obtained from the Petro-
chemical Department of the Continental Oil Company; a similar
material, Ultrawet D.S., obtained from the Atlantic Refining Company;
and Ultrawet K, a branched-chain sulfonated alkylbenzene obtained as
a 93 percent active material sodium salt from the Atlantic Refining Co.
The first two materials were found to have a molecular weight of 372,
and the Ultrawet K, 396.
-------
Adsorption of Biochemically
Other solutes studied included nonylphenoxypolyethoxyethanol (a
nonionic surfactant); 2-sec-butyl-4> 6-dinitrophenol (an herbicide-
insecticide); phenyl-N, N-dimethylphosphorodiamidate (an insecticide);
and 2,4-dichlorophenoxyacetic acid (an herbicide). These substances
were obtained from the Dow Chemical Company. Other adsorbates in-
cluded sodium dodecylsulfate, a Matheson, Coleman, and Bell product,
which was better than 95 percent dodecylsulfate and 99.5 percent total
alkyl sulfate by analysis; triethanolamine; quinine sulfate; phenol; 2, 4-
dichlorophenol; and sodium g-toluenesulfonate, which were obtained as
reagent-grade products from Eastman Organic Chemicals,
The adsorbents and the adsorbates were prepared for the studies
according to procedures detailed in a previous report in this series.*
ANALYTICAL METHODS
Ultraviolet absorption^ Was used for determination of the concen-
tration of the sulfonated alkylbenzene, Ultrawet K, in experiments in
which this solute was the only adsorbate present in solution. When this
substance was a component of solutions in which other solutes caused
interference with the ultraviolet method, the Longwell and Maniece
modification of the Jones method using methylene blue was employed.20
Ultraviolet absorption as previously described*^ was used also for
determination of j>-toluenesulfonate.
Many of the other substances could also be determined readily by
light absorption in the visible or ultraviolet region. Quinine exhibits a
maximum in light absorption at 330 millimicrons with a molar absorpti-
vity of 4.70 x 106 square centimeters per mol; 2,4-dichlorophenol has a
maximum in absorption at 284 millimicrons with a molar absorptivity of
2.05 x 10^ square centimeters per mol; and 2-sec-butyl-4, 6-dinitro-
phenol (DNOSBP) has a maximum at 375 millimicrons with a molar
absorptivity of 13.80 x 106 square centimeters per mol. The absorption
of each of these substances was found to follow the Beer-Lambert law
for the range of concentrations studied. Thus measurements of absorb-
ance at the indicated wavelengths permitted rapid and accurate deter-
mination of the corresponding substances.
The remaining solutes were used solely as components of complex
mixtures of organic compounds, and determinations of their individual
concentrations were not made. Effectiveness of adsorption in experi-
ments with complex mixtures was evaluated in terms of the overall
removal of organic pollutants, which was measured in terms of organic
carbon. Determinations of organic carbon were made according to the
wet oxidation technique.21
EXPERIMENTAL METHOD
An accurately weighed quantity of carbon, usually 45 grams, was
mixed well with approximately 1 liter of distilled water, and the mixture
was introduced into the adsorption tube. The column was set in place in
-------
Resistant Materials from Solution
the experimental apparatus as shown in Figure 1. Distilled water was
run through the apparatus and the column to permit adjustment of flow
to the desired rate with the needle valve. The gate valve immediately
following the needle valve in the flow train was then closed, the carbon
was allowed to settle, and the water in the column was drawn down through
a sample port to a level approximately one-fourth of an inch above that of
the settled carbon. The distilled water was drained from the remainder
of the system. The feed solution was then made up in the storage reser-
voirs and introduced to the constant-head tank. The gate valve was
opened, the carbon expanded, and samples taken at appropriate intervals
after displacement of the initial volume of distilled water remaining in
the column at the beginning of the run. In addition to "instantaneous"
sampling, samples of the total effluent were taken at appropriate intervals
of time after initiation of the run. Instantaneous samples were taken at
selected ports along the length of the expended column of adsorbent as
well as from its top.
Temperature control internal to the system was deemed unneces-
sary in those studies in which the effect of temperature was not being
specifically investigated because the room temperature, and consequently
the temperature of the feed solution, varied but little from approximately
28° to 3QOC. On the basis of temperature effects noted in batch systems,1
this slight variation was thought to have no significant effect on the results
of the column studies.
Results
Eighteen different column runs were carried out, including three
with activated aluminum oxide, two with packed beds of carbon, and one
with a 2-inch column.
ACTIVATED ALUMINUM OXIDE
The activated aluminum oxide used in the column studies was a
"regular-iron" Porocel obtained from Minerals and Chemicals Philipp
Corporation. This material differed considerably in its adsorptive
properties and physical behavior from a "low-iron" Porocel used in
earlier batch studies.22 The "regular-iron" Porocel exhibited poor
adsorptive characteristics relative to the carbons tested, and in addition,
created considerable operational difficulty because it tended to disinte-
grate and form fine suspensions in water when subjected to the slightest
agitation. This tendency was so great that it was impossible to pre-wash
the material with any success and to stop continual carryover of colloidal
aluminum oxide in the effluent from the columns during experiments.
The bauxite tended, furthermore, to agglomerate and be lifted in slugs
rather than as free, discreet particles during upflow operation.
The results of the three runs with bauxite columns may be sum-
marized very briefly. All studies were carried out in 1-inch columns,
-------
Adsorption of Biochemically
•with 150 grams of activated aluminum oxide in each; the adsorbate in
each case was technical alkylbenzenesulfonate in an initial concentration
of approximately 60 micromoles per liter (22.4 nog/liter). Hie first
experiment was conducted with 0.273-millimeter bauxite and a flow rate
of approximately 2.5 gallons per minute per square foot; the initial
portion of effluent from this column had a residual concentration equal
to about 80 percent of that of the influent, and after 3 hours the concen-
tration in the effluent had reached over 90 percent of that in the influent.
After 25 hours a considerable quantity (approximately 10 to 20%) of the
adsorbent had been carried out of the column into the effluent, and the
remaining material had begun to agglomerate into relatively large
cylindrical masses within the column, rendering maintenance of the rate
of flow extremely difficult; at this point the run was terminated.
To provide for more rapid removal of adsorbate from the solution
passing through the expanded adsorbent, very fine bauxite, passing a
U.S. No. 140 sieve, was used in a second experiment. This material
became agglutinated very soon after initiation of the run, forming
masses that were forced up the column by the ascending stream of in-
fluent. Even as these slugs of material were broken up by vigorous
tapping along the walls of the LAIC it e tube, new slugs formed below them
in the column. The experiment had, therefore, to be stopped shortly
after its initiation.
A third study in which an intermediate-size bauxite, 0.200 milli-
meter, was employed failed for the same reasons as did that in which
very fine material was used. Collection of valid data on the effluent
from either of the latter columns was impossible.
ACTIVATED CARBON
Thirteen fluidized-bed studies were conducted with activated
carbon. Columbia LC carbon was used in 10, and 3 were carried out
with Pittsburgh carbon. The salient properties of all 13 experimental
columns are listed in Table 1. The three columns in which Pittsburgh
carbon was used are designated K, L, and M. With the exception of M,
which was a 2-inch-diameter column, the columns were 1 inch in
diameter. Four experiments were conducted to study the effects of
flow rate (columns A to D), and five, to determine effects of particle
size (columns E to I).
Typical data for the fluidized-carbon experiments are shown in
Figures 2 and 3. The data for column study C, in which the highest flow
rate was used, are shown in Figure 2, and those for the experiment in
which the lowest flow rate was used, column study O, are shown in
Figure 3. In Figures 4 to 6, summaries of effluent concentrations as a
function of time are shown for the flow rate studies, for the particle
size studies, and for the remainder of the fluidized-column experiments,
respectively.
A few experiments were continued until the concentration of solute
in the effluent was equal to that in the influent, although the extended
-------
Resistant Materials from Solution
Table 1. PROPERTIES OF EXPERIMENTAL
FLUIDIZED-CARBON COLUMNS
Column
A
B
C
D
E
F
G
H
I
J
K
L
M
Particle
size,
mm
0.273
0.273
0.273
0.273
0.178
0.503
0.359
0.230
0.273
0.126
0.359
0.200
1.545
Total
weight of
carbon, g
45
45
45
45
45
45
45
45
45
45
50
61
500
Concentration of
adsorbate in
influent, A«nol/liter
67.0
65.0
68.4
69.4
76.1
78.5
78.5
74.9
71.7
59.9
72.4
79.7
80.4
Flow rate,
gpm/ft2
2.75
1.43
6.14
0.68
2.56
2.52
2.58
2.52
2.58
0.68
2.68
2.55
5.12
Column
cm
Unex-
panded
21
20
22
21
22
18
19
22
21
15
26
29
57
height
5
Ex-
panded
42
30
67
26
60
23
28
41
40
31
37
55
57
curves are not shown in the summary plots. Complete breakthrough in
Column I, for example, occurred after about 325 hours.
The columns used in experiments A and I were made nearly
identical to check reproducibility of the experimental data; the results
for these runs, shown respectively in Figures 4 and 5, accord rather
well with each other.
Two experiments were conducted with unexpanded downflow col-
umns, identical in all other respects to the columns used in experiments
G and I, to compare this method of operation with fluidization. Despite
the relatively large particle sizes {0.359 mm and 0.273 mm respective-
ly), short columns, and an available head of nearly 10 feet, maintenance
of the initial flow rate of approximately 2.5 gallons per minute per
square foot was not possible. In both experiments the flow rate had
dropped to less than one-half of the initial rate within 24 hours, and to
about one-tenth of the initial rate within 30 hours.
The relatively sparse data collected during the packed-column
experiments indicate that values of C/CO for each of these columns were
less than 0.1 throughout the 30 hours of operation; these data are for
rapidly decreasing flow rates, however, and may not be compared with
those for the constant-rate experiments with columns of fluidized carbon.
The packed-column experiments serve to underscore the advantage of
much smaller headlosses afforded by fluidization.
-------
10
Adsorption of Biochemically
MT
0.5
h = HEIGHT OF SAMPLE POINT
ABOVE BOTTOM OF EXPANDED
COLUMN
h =TOTAL HEIGHT OF EXPANDED COLUMN
TIME.hr
Figure 2. Ratio of concentration at points in fluidized column to influent concentration as a
function of time. Column C, 45-g,0.273-mm Columbia carbon; flow rate = 121.5 ml/min; face
velocity = 24.4 cm/min; unexpended height = 22 cm; expanded height = 67 cm.
-------
Resistant Materials from Solution
11
h =HEIGHT OF SAMPLE POINT
ABOVE BOTTOM OF
EXPANDED COLUMN
h ^TOTAL HEIGHT OF EXPANDED
COLUMN
240
280
320
TIME.hr
Figure 3. Ratio of concentration at points in fluidized column to influent concentration as a
function of time. Column D, 45-g, 0.273-mm Columbia carbon; flow rote = 13.5 ,-nl/min; face
velocity - 2.7 cm/min; unexpanded height = 21 cm; expanded height = 26 cm.
-------
12
Adsorption of Biochemically
340
Figure 4. Ratio of effluent concentration to influent concentration as a function of
time for different flow rates, 45-g 0.273-mm Columbia carbon.
I I I I I I I I I
0 10 20 30 40 50 60 70 80 90 100 110 120
0.2 -
Figure 5. Ratio of effluent concentration to influent concentration as a function of
time for different particle sizes of adsorbent, 45-g Columbia
carbon, flow rate = 2.5
-------
Resistant Materials from Solution
13
0.8
0.6
UU 0.4
0.2
0.359-mm PITTSBURGH
(COLUMN K)
0.200-mm PITTSBURGH
(COLUMN L)
0 126-mm COLUMBIA
(COLUMN J)
1.545-mm PITTSBURGH
(2-in. COLUMN M)
150
175
200
Figure 6. Ratio of effluent concentration to influent concentration as a function of
time for additional column studies.
SOLUTE EFFECTS
Previous experiments on adsorption by carbon in fluidized columns
had been conducted with highly branched sulfonated alkylbenzenes. It
•was of interest to explore possible differences in adsorption pattern for
different solutes upon passage through fluidized-carbon columns and
ascertain whether observed differences can be correlated with character-
istics found through more readily accomplished measurements on batch
systems. Hence, in one experiment, parallel-columns were run on two
solutes with considerably different molecular weights but of the same
chemical class of compound. The solutes used were Ultrawet K and
sodium j>-toluenesulfonate. The latter is the simplest alkylbenzenesul-
fonate, with a molecular weight of 171 for the anion; the former is a
branched-chain alkylbenzenesulfonate with about 15 carbon atoms in the
alkyl group.
One reason for the choice of these two solutes was that rates of
adsorption and capacities for adsorption on activated carbon from solute
solution have been studied previously for p-toluenesulfonate and for
materials very similar to the Ultrawet K in batch-type systems, and
marked differences in these adsorption characteristics have been
observed.!
Figure 7 presents data for the adsorption of Uitrawet K from a
solution with an initial concentration of 73.3 micromoles per liter in a
column containing 45 grams of 0.273-millimeter Columbia carbon; this
quantity of the 0.273-millimeter carbon had an unexpanded column height
of approximately 22 centimeters. The now rate for the study was 2.5
gpm per square foot. The form of the curve representing the relative
effluent concentration C/CO as a function of time is similar to that
obtained previously for other highly branched sulfonated alkylbenzenes
under similar experimental conditions.*
-------
14
Adsorption of Biochemically
In Figure 8, the curve shown in Figure 7 for Ultrawet K is com-
pared with a similar curve for sodium £-toluenesulfonate. The initial
concentration of j>-toluenesulfonate in the solution used for this experi-
ment was 79.6 micromoles per liter; carbon size and quantity as well as
flow rate were identical to the experiment with Ultrawet K.
0-8
0.7
0.6
0.5
0.3
0.2
0.1
20
40
60
80
100
120
140
160
TIME,
Figure 7. Breakthrough curve for branched-chain sulfonated alkylbenzene (ABS).
Several points of comparison should be noted. First, there was an
initial leakage of |>-toluenesulfonate considerably greater than that for
the ABS, the leakage of the £-toluenesulfonate being 6 percent of its
influent concentration to about 2 percent for the ABS. Second, the effluent
leakage for the p_-toluenesulfonate increased very little for the first 20
hours of operation whereas that for the ABS increased almost from the
start and after 16 hours of operation became greater than the leakage
from the ja-toluenesulfonate column. Third, as a corollary to the second
point, the total breakthrough of p-toluenesulfonate was much sharper than
that of the ABS. After about 60 hours, the uptake from the ABS solution
was once again more effective than that from the p_-toluenesulfonate
solution. Fourth, after about 100 hours the total integrated uptake from
the two solutions had been about the same, but about one-third of the in-
fluent ABS was still being removed whereas the capacity of adsorption
of £-toluenesulfonate had apparently been exhausted.
-------
Resistant Materials from Solution
mu 0.5 -
p - TOLUENESULFONATE-
80
100
TIME, ht
Figure 8. Breakthrough curves for ABS and p-foluenesulfonate.
The fact that the Ultrawet K is taken up more completely during
the very early stages of the run than the £-toluenesulfonate is in accord
with the findings of previous equilibrium studies in batch systems. Dur-
ing the first few hours of the experiments, the concentration of each
solute in the water passing up through the expanded bed of fresh carbon
is reduced to a low figure beyond the first few inches of the column.
Previous investigations of the equilibria of adsorption on carbon have
indicated that the energy of adsorption for a high-molecular-weight sul-
fonated alkylbenzene is greater than that for p_-toluenesulfonate.l In
other words, effective adsorption is observed for much smaller concen-
trations with the ABS than with £-toluenesulfonate. Thus, removal by
-------
Adsorption of Biochemically
adsorption of the last traces of the high-molecular-weight Ultrawet K
should be considerably easier than complete cleanup of the jj-toluene-
sulfonate.
The increasing leakage of the ABS in contrast to that of the p_-
toluenesulfonate over the first 20 hours of column operation can be
associated with the relative rates of adsorption observed in batch studies
on these solutes. Because of the slow rate of penetration of the ABS
into the interior surfaces of the carbon granules, the active external
points soon reach a significant fraction of saturation, and subsequent
adsorption is less rapidly effective. With the £-toluenesulfonate a
relatively rapid migration of adsorbed material to the interior permits
renewal of active external surface and substantial maintenance of effi-
ciency of adsorption for a longer period. When nearly full capacity has
been attained, a relatively sharp breakthrough occurs.
0.2 -
0.1 -
100 120 140 160
TIME, hr
180
200 220
Figure 9. Breakthrough curves for ABS in solutions of different initial concentration.
Maintenance of adsorption capacity in the column adsorbing ABS
beyond the time required to exhaust the column adsorbing n-toluenesul-
fonate is expected on the basis of the greater adsorption capacity for the
former material found in the batch equilibrium studies.
The separate-column studies with Ultrawet K and p-toluenesulfonate
thus indicate that at least qualitative prediction of column behavior may
be possible from knowledge of batch-system behavior, at least for cases
in which a single-solute solution is passed through the column.
-------
Resistant Materials from Solution
17
Figure 9 presents breakthrough curves for three experiments in
which the effects of different concentrations of Ultrawet K were studied.
The flow rate for each experiment was 2.5 gpm per square foot, and each
column contained 45 grams of 0.273-millimeter Columbia carbon.
The effect of concentration on breakthrough of solute is as would be
expected, that is, the greater the concentration the more rapid the break-
through. Note, however, that at the points corresponding to C/CO = 0.1
the total weights of ABS adsorbed were about the same in all three runs.
Although leakage appears greater at the smaller influent concentrations,
note that the actual effluent concentrations at zero time were 0.85, 1.5,
and 0.8 ppm for increasing influent concentrations. Apparently then,
initial effluent concentrations at a fixed face velocity will be determined
primarily by the depth of the adsorption column rather than by the influent
concentration.
One other noteworthy point is that with the lowest influent concen-
tration, 13.2 micromoles per liter or 5.2 milligrams per liter, better
than 90 percent adsorption was maintained with just a 9-inch settled
depth of carbon for more than 5 days. This would mean about 40 days'
effective operation for a 5-foot settled depth with this concentration of
material to be removed.
320
280 -
I
120 -
80 _
40 -
0 20 40 60 80 100 120 140 160 180 200 220
TIME, hr
Figure 10. Cumulative adsorption of ABS from solutions of
different initial concentrations.
Figure 10 shows total quantity of solute removed from solution by
the 45 grams of carbon in each experiment. As would be expected on the
-------
18 Adsorption of Biochemically
bases of the increased driving force for transfer with increased concen-
tration of solute in solution and of greater capacity of the carbon at the
greater influent concentration, both the rate of uptake and total quantity
of solute removed are greater at the higher influent concentrations of
Ultra-wet K for the same period of time. After 100 hours of operation,
the carbon in the experiment in which the influent concentration of solute
was 73.3 micromoles per liter had removed a total quantity of solute
equal to approximately 11 percent of the weight of the adsorbent, and for
the experiment with Co = 13.2 micromoles per liter, the solute removed
was only about 3 percent of the weight of the carbon. Since continuing
the experiments at the lower initial concentrations was impractical until
the carbon was exhausted, full capacities as a function of concentration
could not be measured.
TEMPERATURE EFFECTS
Three experiments were carried out with Ultrawet K as solute at
solution temperatures of 36OC, 28oc, and 16QC. Flow rates of 2.5 gpm
per square foot were maintained in columns containing 45 grams of
0.273-millimeter Columbia carbon. The initial concentrations of solute
were closely the same, in the range of 77 to 80 micromoles per liter.
The experiment at 28°C was at room temperature; for the experiment at
16°C the influent was chilled by means of a conventional Freon compres-
sor with the cooling coil immersed in the constant-head tank; a blade-
type heater was used in the constant-head tank to heat the influent for the
experiment at 36°C. Each temperature given is an average of the tem-
peratures of the influent and effluent for a column. Temperature differ-
ences between the influent and effluent for each experiment were less
than 3°C.
As shown in Figure 11 little difference was observed in the total
uptake of solute as a function of time for any of the temperatures studied.
Such a result is to be expected, for the total uptake is primarily a function
of the adsorptive capacity of the carbon with beds of sufficient depth and
depends very little on the rate of adsorption. Adsorptive capacity does
not change greatly over a small temperature range such as that used in
these experiments.
The effect of temperature should be noted most significantly for
column work in the steepness of the breakthrough portion of the curve of
relative effluent concentration as a function of time. At lower tempera-
tures, where the rate of adsorption is less, increasing leakage should
begin sooner and a more gradual approach to full saturation should be
observed. A close look at the experimental points in Figure 11 reveals
an indication of this effect, for the data at 36°C beyond 25 hours lie con-
sistently above the mean curve and more nearly maintain the linear
relationship characteristic of complete adsorption, while the data at
18°C lie consistently below the mean curve, indicating increasing leak-
age through the column.
-------
Resistant Materials from Solution
19
20C
4C _
20 _
TIME, hr
Figure 11. Cumulative adsorption of ABS from solutions at different temperatures.
pH EFFECTS
A preliminary investigation of the effect of solution pH on uptake
of solute by carbon in a fluidized bed was made by passing a solution of
Ultrawet K, with an unadjusted pH slightly below neutral, at a flow rate
of 2.5 gpm per square foot through a column containing 45 grams of
0.359-millimeter Columbia carbon. Hie larger particle size was used
so that more rapid breakthrough would be obtained. When the value of
C/Co for the experiment reached 0.55, the pH of the influent was de-
creased to 2.5 by addition of concentrated phosphoric acid. As illus-
trated by the data for this run plotted in Figure 12, the decrease in
leakage after this acidification of the influent was dramatic.
-------
20
Adsorption of Biochemically
To test the effect of solution pH more extensively, four experi-
ments were conducted at different pH values, each with a column con-
taining 45 grams of 0.273-millimeter Columbia carbon and a flow rate
of 2.5 gpm per square foot. The initial concentration of solute, Ultrawet
K, in each experiment was within the range from 79 to 81 micromoles
per liter. Adjustment of pH in the solutions to values of 3.9, 6.6, 7.9,
and 10.0, respectively, was accomplished by appropriate addition of
phosphoric acid or phosphate. The pR values given are for the influents
to the columns.
0.7
0,6
0.5
0.4
0.3
0.2
0.1
I
I
I
10
20
30
50
60
70
90
TIME, hr
Figure 12. Effect of acidification on the breakthrough curve for ABS.
Data for the pH studies are presented in Figure 13, which shows the
total quantities of solute removed from solution by the carbon as a func-
tion of time. A notable difference exists between the total amounts of
solute removed from solution in the experiments at the highest and low-
est pH values tested, the total amount at any time being considerably
greater at low pH. The experiments at intermediate pH levels are
observed to fit into the pattern of increasing removal with decreasing pH.
The observed dependence of solute removal on pH for the column studies
accords with previous work in which both rates and final positions of
equilibrium for adsorption of sulfonated alkylbenzenes on the experimental
carbon in batch-type systems were found to increase with decreasing pH.l
-------
Resistant Materials from Solution
21
500
u
100 —
50 —
20
40
60 80
TIME, tir
Figure 13. Cumulative adsorption of ABS from solutions at different pH values.
Even so, in order to show marked changes in adsorption with pH,
use of very high or low pH values was necessary. Over the pH range of
particular interest, 6 to 8, differences in adsorption efficiency were
minor. More pronounced effects may be obtained with other types of
pollutants, but in general it can be expected that normal pH variations
will not affect significantly the operation of activated-carbon columns.
MULTICOMPONENT SOLUTIONS
That solutes in multiple-component solutions compete with one
another for adsorption on carbon has been demonstrated for agitated
nonflow systems.*3 Studies have been carried out during the present
investigation to explore the nature and extent of competitive interactions
-------
22 Adsorption of Biochemically
in continuous-flow systems. These studies are significant for evaluation
of the suitability of continuous-flow adsorption systems for waste-water
treatment because the waste water to be treated rarely contains but one
organic pollutant. Rather, waste waters are likely to be heterogeneous
mixtures of complex organic substances of different chemical types or
classes.
In these investigations, solute mixtures of two, three, four, and
eight components have been studied.
Two Components
Some pronounced effects of sodium dodecylsulfate on the adsorp-
tion of a sulfonated alkylbenzene in batch systems have been noted pre-
viously.23 Hence, mixtures of sodium dodecylsulfate and the branched
sulfonated alkylbenzene, Ultrawet K, were employed as solute in initial
studies of adsorptive competition on columns.
A solution containing 71.6 micro moles of Ultrawet K per liter and
75.0 micromoles of dodecylsulfate per liter was passed at 2.5 gpm per
square foot through a column containing 45 grams of 0.273-millimeter
Columbia carbon. Because a simple analytical technique for dodecyl-
sulfate in dilute solution was not available, only the concentration of the
ABS was determined in the effluent. Figure 14 presents the break-
through curve found for the Ultrawet K (ABS) in this study and compares
it with a breakthrough curve for a pure solution of the ABS. The experi-
ment with the ABS alone was carried out under conditions identical to
those with a solution containing 73.3 micromoles of ABS per liter.
In Figure 14, the presence of the dodecylsulfate apparently has a
marked effect on the over-all adsorption of the sulfonated alkylbenzene.
The competitive effect appears, however, to be not significant in the
early stages of the run; a major difference between the breakthrough
curves for ABS in mixed and pure solution is noted only after about 55
hours of operation of the column. During the first 55 hours the carbon
apparently had sufficient capacity for both solutes so that little effective
competition for adsorption occurred. Since in this experiment the
dodecylsulfate was added to the ABS concentration rather than in place
of part of it, whatever dodecylsulfate was adsorbed would be expected to
reduce the adsorptive capacity for ABS.
Three Components
To test further the effects of solute interactions on adsorption
in fluidized columns from solutions containing mixtures of organic
compounds, a column study was run with a solution containing three
different classes of organic compounds. The influent solution consisted
of Ultrawet K at a concentration of 43.6 micromoles per liter, 2-sec-
butyl-4, 6-dinitrophenol (DNOSBP) at a concentration of 31.8 micromoles
per liter, and quinine sulfate at a concentration of 26.9 micromoles per
-------
Resistant Materials from Solution
23
liter (as the quinine base). The flow rate for this experiment was 2.5
gpm per square foot and the column consisted of 45 grams of 0.273-
millimeter Columbia carbon.
i.o
0.9
0.8
0.7
0.6
uiu"
0.4
0.3
0.2
0.1
ABS
MIXTURE
ABS, PURE SOLUTION
20 40 60 80
TIME, hr
100
120
140
160
Figure 14. Breakthrough curves for ABS for a single-solute solution and for a
bisolute solution with dodecylsulfate
Analyses for concentration of the ABS in the mixed solution were
carried out -with the methylene blue procedure mentioned previously; no
interference from the quinine or the DNOSBP was experienced with this
procedure. Determinations of quinine and DNOSBP were made by
spectrophotometric measurements at 330 and 375 millimicrons, re-
spectively. The molar absorptivities for these substances at their
respective wavelengths for peak adsorption have been listed earlier in
this report. At 375 millimicrons the molar absorptivity of the quinine
-------
24
Adsorption of Biochemically
was 0.07 x 10^ square centimeters per mole, and that for the DNOSBP at
330 millimicrons was 4.71 x 10^ square centimeters per mole. Deter-
minations of individual concentrations of these two solutes thus required
measurement of each sample at each of the two wavelengths, and simul-
taneous solution of two absorbance-concentration equations.
0.3 —
0.2
60
80
100
TIME, hr
Figure 15- Breakthrough of organic solute with a mixture of
ABS, DNOSBP, and quinine.
The overall breakthrough curve for the mixture is shown in Figure
15. The concentration ratio plotted in Figure 15 is for the total concen-
trations of all the organic solutes present in the solution, computed by
summing the results of the individual determinations. The overall
breakthrough curve is similar in pattern to those observed previously for
single-solute solutions.
-------
Resistant Materials from Solution
25
Figure 16 shows breakthrough curves for each of the components
of the mixture. The breakthrough for the ABS in the mixture is much
more rapid than that for a pure solution of similar initial concentration
of ABS is. This fact is illustrated by the curves plotted in Figure 17. The
breakthrough curve shown in Figure 17 for the single-solute solution was
obtained under identical flow and adsorbent conditions with a 39.4-micro-
moles-per-liter solution of the Ultrawet K.
u|u
0.2 —
0.1 —
40
60
80
100
TIME, hr.
Figure 16. Breakthrough curves for the components with s solute mixture of
ABS, DNOSBP, and quinine.
On the other hand the breakthrough is not as rapid as would have
been expected for ABS alone at a concentration equal to 102 micromoles
per liter, the sum of the component concentrations. In fact the curve is
rather similar to that obtained with 73 micromoles of ABS per liter.
This tends to indicate some enhancement of total adsorptive capacity for
-------
26
Adsorption of Biochemically
i.o
U|U 0.5 -
0.4 -
0.3 -
0.2 -
0.1 —
Figure 17.
100
TIME, hr
Comparison of breakthrough curves for ABS with a single-solute solution and
with a solute mixture with DNOSBP and quinine.
the mixed components, like that found earlier in the batch studies of
competitive adsorption.
Here were no single-solute column studies for quinine and
DNOSBP with which the other curves in Figure 16 might be compared.
Hie curves shown in Figure 16 do not bring out completely the
adsorption picture for this experiment with a multicomponent solution.
Figure 16 alone gives the impression that the amount of ABS adsorbed
was less than the amount of DNOSBP adsorbed, and that the amount of
quinine adsorbed was greatest of all. Actually, because of the initial
differences in concentration, the amounts of individual solutes removed
from the mixed solution flowing through the column were the reverse of
this order. This is shown by the cumulative uptake plots in Figure 18.
-------
Resistant Materials from Solution
27
Comparison of Figure 18 -with Figure 10 shows that the extent of
adsorption of the ABS from the mixture was less than that from a pure
solution of approximately the same concentration for long periods of
operation. For example, the quantity of ABS adsorbed from the mixture
after 100 hours was about 130 micromoles per gram of carbon, while for
a pure solution of similar initial concentration of ABS, the quantity
adsorbed after 100 hours was about 210 micromoles per gram of carbon
(Figure 10). The total organic matter adsorbed after 100 hours—the sum
of ABS, DNOSBP, and quinine—was, however, 350 micromoles per gram.
In an effort to determine whether the observed adsorption patterns
for the mixture could be predicted from batch-system studies, kinetic
and equilibrium experiments with nonflow systems were conducted with
the same mixture of the three solutes and also with solutions of the three
individual components in each of which the initial concentration of solute
was the same as its initial concentration in the mixture.
140
—O- ABS
-•—DNOSBP
—•-QUININE
10
Figure 18. Cumulative adsorption of trie components oi a mixture of
ABS, DNOSBP, and quinine.
Figure 19 shows a plot of data for the rate of adsorption of quinine,
obtained from the batch-type experiments. Both the pure solution and the
mixture studies were conducted at 29°C. Fifty milligrams of 0.273-milli-
meter Columbia carbon per liter was used in the experiment with the pure
quinine solution, and 250 milligrams of the same carbon per liter was
used for the experiment with the multicomponent solution to take account of
the difference in total concentration of organic solutes present. The data in
-------
28
Adsorption of Biochemically
Figure 19 have been linearized for the initial portions of the rate studies
by plotting the amount removed from solution as a function of the square
root of the time, according to an intraparticle-transport rate step.l
200
QUININE,
PURE SOLUTION
,O.S(hr0.5
Figure 19. Rates of adsorption of quinine from a single-solute solution and from a
mixture of AB5, DNOSBP, and quinine under agitated non-flow conditions.
Hiere is a clear reduction in the rate of uptake of the quinine from
the mixture as compared with the rate for the pure solution. Note also,
however, that while the amount adsorbed per unit of carbon from the
multicomponent solution in an hour was about half that in the pure solu-
tion, the mole fraction of quinine in the multicomponent solution was only
about 0.23.
-------
Resistant Materials from Solution
29
Similar data -were obtained for the other solutes; rate constants
derived from the data are listed in Table 2. The values for the relative
rate constant, k, are squares of slopes of plots of the type shown in
Figure 19. These data indicate that the rate of adsorption of DNOSBP
is by far the most adversely affected in the mixture. There seems to
be some tendency for all the rates to reduce to a more nearly common
value in the multicomponent solution. This is, also indicated by the
parallel slopes of the breakthrough curves in Figure 16.
Table 2. RELATIVE RATES OF ADSORPTION FROM PURE
SOLUTIONS AND MULTICOMPONENT SOLUTION 0.273-mm
COLUMBIA CARBON. AGITATED NONFLOW SYSTEM, 28 °C
Relative rate constant, k, f(^mol/g)2ArJ x 10"3
Pure solution
Multicomponent solution
sulfonated alkylbenzene
quinine
2-se_c-butyl~4, 6-dinitrophenol
3.60
7.23
57.60
1.85
1.60
0.26
Mixture: 43.6 i*mol sulfonated alkylbenezene per liter,
26.9 pmol quinine per liter, 31.8 Mmol
2-sec-butyl-4 , 6-dinitrophenol per liter
Table 3.
LANGMUffi CONSTANTS FOR ADSORPTION ISOTHERMS;
0.273-mm COLUMBIA CARBON, 30°C;
XmbC
LANGMUIR EQUATION:
X =
1+bC
m
b,
sulfonated alkylbenzene
quinine
2-sec-butyl-4,6-dinitrophenol
455
605
1,850
0.478
4,13
0.174
Isothermal equilibrium studies at 30°C for the individual com-
ponent solutions were also conducted with 0.273-millimeter Columbia
carbon. A plot of the equilibrium adsorption data for quinine is given
in Figure 20. The curve fitted to the points in Figure 20 is a calculated
Langmuir-type curve; the linearized form is given in Figure 21.
Langmuir constants for all three solutes are listed in Table 3; X
is the amount of adsorbed material per gram of carbon, Xm is the limit-
-------
30
Adsorption of Biochemically
ing value of X for monolayer adsorption according to the Langmuir model,
C is the concentration of solute in solution at equilibrium, and b is a con-
stant expressive of the energy of adsorption and equal to the reciprocal
of the concentration at which adsorption is equal to Xra/2.
280
200
40
Ce<, fimol liter
Figure 20. Adsorption isotherm for quinine sulfote.
- Jit«r//imo1
Figure 21. Langmuir plot of adsorption isotherm for quinine sulfote.
-------
Resistant Materials from Solution 31
Attempted correlation of these batch studies with the previous
column studies is difficult and probably premature, for neither the batch
nor the column studies have been sufficiently complete. The only direct
comparison to be made between the batch studies on the rates of adsorp-
tion and the continuous-flow studies is with the slopes of the breakthrough
curves for the columns. Unfortunately the slopes of the breakthrough
curves with the shallow beds used in these studies have not been well
enough defined to make such a comparison meaningful. Conversely, the
relative amounts of uptake for the individual solutes in the mixture might
be compared with data on individual adsorption isotherms in the multi-
component solution, but these latter data have not yet been obtained,
E is tempting to relate the great reduction of rate of uptake of
DNOSBP in the multicomponent solution as compared with pure solution
with the comparatively low "b" value obtained in the isotherm studies
and with the reduced capacity for adsorption shown in the mixed-column
studies. Yet, fitting of the Langmuir equation to the equilibrium adsorp-
tion data for a substance like DNOSBP that exhibits an apparent Xm value
equal to better than 40 percent of the weight of carbon must be admitted
to be theoretically meaningless, since the monolayer assumption is
certainly invalid.
For the present, then, these results must simply be added to the
fundamental data on adsorptive behavior being accumulated for later use
when more complete delineation of the systems has been accomplished.
Four Components
Another column study was conducted with a solution containing four
different classes of organic compounds, both to determine the adsorptive
patterns of some other types of solutes in multicomponent solutions and
to investigate breakthrough effects in deeper columns of a carbon more
suitable for practical use in waste-water treatment.
A multiple-component solution containing (in mg/liter) Ultrawet
K, 24.5; triethanolamine, 15.5; 2,4-dichlorophenol, 16.9; and nonylphen-
oxypolyethoxyethanol, 53.7, was passed upward through a 1-inch column
containing 200 grams of 0.273-millimeter Pittsburgh carbon, at a rate
of 2.5 gpm per square foot. The expanded height of the carbon column
was 209 centimeters. In addition to routine determinations for concen-
trations of total organic carbon, ABS, and 2,4-dichlorophenol in the
effluent from the column, several 20-liter samples of the effluent were
concentrated by vacuum distillation and analyzed by infrared spectro-
photometry.
Figure 22 is a plot of data for the uptake of organic material by
the Pittsburgh carbon during this 180-hour experiment. The data have
been plotted in terms of the organic carbon removed from solution as a
function of time. It was possible, as illustrated in Figure 22, to account
for the fractions of organic carbon existing as ABS and 2,4-dichloro-
-------
32
Adsorption of Biochemically
phenol because these materials were determined individually. Although
one fraction is represented as being composed of triethanolamine plus
nonylphenoxypolyethoxyethanol (9N10), it should be kept in mind that in-
frared analyses showed complete uptake of 9N10 and considerable leak-
age of triethanolamine.
200
180
160
140
120
100
80
60
40
20
I i I \ J I I I S\ I
TOTAL ORGANIC CARBON
ORGANIC CARBON
AS TEA t 9N1O
ORGANIC CARBON
AS ABS
ORGANIC CARBON
AS 24 OCR
20
40
80
100
120
140
160
180 200
TIME, hr
Figure 22. Cumulative adsorption ot organic carbon from a mixture of ABS,
triethanolatnine (teo), nonylphenoxypolyethoxyethanol
(9N10), and 2, 4-dichloroph«nol (2,4-DCP),
After about 150 hours of operation, maintaining the flow-rate of
2.5 gpm per square foot with the available head of 10 feet became diffi-
cult because of increased density of the carbon and clogging of the
material in the column. At 180 hours, holding the rate with the experi-
mental apparatus became impossible, and the run was terminated.
-------
Resistant Materials from Solution 33
After 180 hours the C/CO value for total organic carbon in the
effluent had reached a value of only about 0.2. No significant concentra-
tions of ABS or 2,4-dichlorophenol were found in the effluent up to this
point by standard determinations. The periodic infrared analyses of con-
centrated effluent indicated that the only material being carried over in
the effluent in any significant amount was the triethanolamine; approxi-
mately 50 percent of this material was passing completely through the
column.
If the weight of the organic carbon in the mixture of organic solutes
is taken as one-half the total weight of organic material, then as shown by
Figure 22, the total amount of organic material removed from solution by
the Pittsburgh carbon during the 180-hour run was about 400 milligrams
per gram or a tremendous capacity of 40 percent-by-weight for the car-
bon to this point. To check this, a mass balance was run by removing
the carbon from the column and drying it in a very thin layer at 105°C
for 24 hours. The carbon was weighed after drying and found to weigh
279.5 grams! Thus the mass balance confirmed the very high capacity
of 40 percent-fay-weight. The good capacity realized in this experiment
indicates that very high capacities with resulting economy in cost of
carbon may be realized for some pollutants.
On the other hand, the early breakthrough of triethanolamine shows
that some selectivity in removal of pollutants will be observed in practice
and give some indication of the types of compounds that may not be
effectively removed by carbon. Batch studies of rate of adsorption and
adsorption equilibria should be carried out with triethanolamine.
Eight Components
One column study was run with a very complex mixture of assorted
classes of organic solutes. The solutes comprising the mixture, along
with their respective concentrations in the influent, are listed in Table 4.
Table 4. COMPOSITION OF EIGHT-COMPONENT SOLUTE MIXTURE
Concentration in
Solute influent, mgAiter
phenol 2.4
quinine 9.3
dodecylsulfate 7.2
sulfonated alkylbenezene 10.6
2-sec-butyl 1-4,6-dinitrophenol 6.0
2,4 -dichlorophenoxyacetic acid 5,5
nony Iphenoxypolyethoxyethanol 16.5
phenyl N,N'-dimethylphosphorodiamidate 5.0
-------
34
Adsorption of Biochemically
The column used in this experiment contained 100 grams of 0.273-
millimeter Pittsburgh carbon, and the flow rate was 2.5 gpm per square
foot. Total organic carbon was the only analytical measure used to
determine the effectiveness of the column.
Continuous flow was maintained for approximately 100 hours; then
experimental difficulties forced termination of the experiment: for some
unexplainable reason prolific microbiological growths developed in the
influent solution, presumably made up largely of resistant substrates.
At the termination of the run the value of C/CO for total organic carbon
had reached a level of about 0.1, and the total quantity of organic carbon
removed from solution up to the time of termination was 8.5 grams.
"Fius the capacity attained by the adsorbent, up to the point when the
experiment was stopped, was on the order of 16 percent on the assump-
tion that the weight of organic matter was twice that of the organic carbon.
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0,2
0.1
o
I
I
I
0 10 20 30 40 50
TIME, hr
Figure 23. Breakthrough curve for ABS in a column of reused carbon.
60
-------
Resistant Materials from Solution 35
SELF REGENERATION OF ACTIVATED CARBON
In a previous study it had been noted that activated carbon tended
to regain some of its "initial capacity" upon standing, out of solution.
This phenomenon was assumed to occur as a result of continued trans-
port of solute within the pores of the carbon.24 In further exploration of
this phenomenon, carbon from one of the previous column experiments
with Ultra-wet K (the carbon used in the high-concentration run during
the concentration-effect studies) was dried at 105°C for 48 hours and
stored for 2 weeks. After drying, the quantity of carbon weighed 52.1
grams, whereas its initial weight before use in the concentration-effect
study had been 45 grams. The carbon was then replaced in a column,
and a solution containing 77.1 micromoles of Ultrawet K per liter was
passed through it at a rate of 2.2 gpm per square foot.
The results of this experiment are shown in Figure 23. While the
carbon apparently did regain some of its "initial capacity" in that a
typical breakthrough curve starting near C/CO = 0 was found, the gain in
capacity was not great, as indicated by the rapid and complete break-
through after a relatively short period.
Limited Model for Qualitative Interpretation
Previous research on the kinetics of adsorption on granular carbon
from solution in rapidly stirred batch systems has indicated that rates
of uptake of organic adsorbates by granular carbon are controlled by
intraparticle transport phenomena. 1,25 The curves shown in Figures 4,5,
and 6 for the fluidized-column experiments were found, however, not to
accord well with curves predicted by mathematical solution of theoretical
equations, such as those given by Rosen26 or Kasten etjal.,2?for pro-
cesses in columnar or continuous systems controlled by intraparticle
diffusion.
Values of the Reynolds numbers for the fluidized systems studied,
calculated on the basis of superficial velocities in the columns, were in
the range of 0.1 to about 1.0. In this region of apparent laminar flow it
is not unreasonable to expect that the over-all rate of transfer of adsorb-
ate from solution phase to the solid adsorbent is controlled by film dif-
fusion processes rather than by intraparticle transport.
Several hypothetical models may be advanced for description of
rates for transfer processes, each with its degree of accuracy in repre-
senting an experimental system dependent upon the accuracy and the
number of limiting assumptions imposed by the model. The most ele-
mentary, and therefore most limited, of the several models that may be
used to represent the present experimental system will be considered
herein; more sophisticated models are presently being developed and
will be presented in a subsequent discussion.
-------
36 Adsorption of Biochemically
The elementary model, which may be more properly termed the
"completely mixed-column model," involves assumption of a uniform
concentration of adsorbate in the bulk solution over the length of the
adsorption column and the assumption that the quantity of adsorbate re-
moved from solution by each particle of carbon in the column is uniform
throughout the expanded media. A further limiting assumption imposed
by the elementary model is that the equilibrium relationship between
adsorbed solute and solute in solution is linear. Obviously, these assump-
tions are restrictive for application of the elementary model to quantita-
tive description of transfer rates in the experimental system. For ex-
ample, the experimental isotherms have been determined to follow a
Langmuir adsorption pattern, and thus the assumption of linearity in-
volved in the elementary model can provide only an order-of-magnitude
approximation for the distribution of adsorbed solute and solute in solu-
tion. Further, Figures 2 and 3 show that the distribution of solute in
bulk solution is not uniform over the entire length of the adsorption
column, and thus the assumption of uniform distribution of solute will
result in only an approximation of the actual condition. Observation of
considerable particle mixing within the fluidized columns during opera-
tion has indicated that the assumption relating to the uniformity of up-
take of solute by each particle of adsorbent should not be as restrictive
as the two other assumptions.
The authors are fully aware of the limitations of the elementary
model and emphasize that this model is not quantitatively representative
of the experimental system. The elementary model has, however, a
marked advantage of simplicity over more sophisticated models, while
still incorporating the major variables of the experimental system. The
greatest value of the elementary model, and the reason for its presenta-
tion herein, is that it is very useful for delineating qualitatively the
effects of the major systemic variables.
Based on the foregoing hypotheses, a relatively straightforward
rate equation for transfer of adsorbate from solution to the exterior
surface of the adsorbent in a fluidized column may be written. The rate
for steady-state transfer of adsorbate across a film to the carbon may
be expressed as the ratio of driving force to transfer resistance.
driving force
Rate of transfer = resistance = r (1)
The driving force in Equation 1 is the product of the effective con-
centration difference between the solution and adsorbent and the total
area across which transfer occurs. The resistance to transfer is a
function of the system and is assumed uniform throughout the system.
The transfer resistance is then a coefficient, the reciprocal of which
may be termed the transfer coefficient, k.
Equation 1 then takes the form
r = kamMaAC (2)
-------
Resistant Materials from Solution 37
where:
k = transfer coefficient,
am = effective transfer area per unit mass of adsorbent,
Ma = mass of adsorbent, and
Ac = concentration difference across film.
The experimental column operations were carried out under non-
steady-state conditions, for which the instantaneous rate of transfer of
adsorbate from solution can he expressed in terms of the concentration
remaining in solution at any time during the transfer and the concentra-
tion gradient across the film at the same instant. Because the adsorbate
in the experimental systems had a strong preference for the adsorbent,
application of a partition factor, ir , to the concentration of adsorbate in
the adsorbent is necessary in defining the concentration gradient of the
transferable agent across the film. The factor, IT, which is equal to the
value of C/GS at equilibrium, serves effectively to increase the concen-
tration-gradient term. Then,
Qm(C0 - C) = kT am Ma (C -TTCS) (3)
In Equation 3, Qm is the mass flow rate of solution (MT~1); Co is the
mass concentration of adsorbate in the influent (MM-1); C is the mass
concentration of adsorbate in the effluent at time, t, during the transfer
process (MM-1); kT is the mass transfer coefficient per unit driving force
(MT-lL-2); and Cg is the mass concentration of adsorbate in the adsorb-
ent at time, t, (MM-1). The value for the instantaneous concentration-
gradient term, which incorporates the unknown quantity Cs, is difficult
to evaluate experimentally. The value of Cs, however, can be determined
by graphical integration of an appropriate expression for the differential
quantity of adsorbate transferred to the adsorbent, since the variation of
the quantity (C0-C) with time, t, may be evaluated readily from experi-
mental observations of the variation of C with time.
The differential quantity of adsorbate transferred from solution to
the adsorbent may be expressed by the equation
dCs = Qm(Co-C) (4)
dt W
in which W is the total weight of adsorbent (M), and all other terms are
as previously defined.
As stated previously, Equation 4 may be integrated graphically to
determine a value for Cs, which may in turn be used in Equation 3 to
solve for values for the transfer coefficient, kT. The problem posed
to the evaluation of kT by the appearance of the unknown term, Cs, in
-------
38 Adsorption of Biochemically
Equation 3 may be circumvented more conveniently in yet another
manner. Equation 3 may be differentiated with respect to C to give an
expression for dCg in terms of k*p» dC, and the known quantities Qm,
am> T, and Ma- Rearrangement of Equation 3 gives the following ex-
pression fbnrCg:
Qmc Qmc0
cs = C+
Equation 5 is then differentiated with respect to C to give
I Qm
dCs = IT (1 + kTamMa ) dC
The right side of Equation 6 is now substituted for dCg in Equation 4
to give
(1 + Qrn/kTamMa) dC =7rQm(Co - C)
dt W
hence:
C
(7)
J
-------
Resistant Materials from Solution
39
example, the value of kj is not constant over an entire run, it would not
be valid to compare values of this parameter computed for a fixed time
interval for different systems. Such plots have been constructed for the
experimental data from the 1-inch column studies, as shown in Figures
24 and 25. Tin Figure 24 values of ln(l-C/Co) versus t are shown for the
flow rate studies with 0.273-millimeter Columbia carbon, and in Figure
25 they are shown for the particle size studies with Columbia carbon,
and the experiments with Pittsburgh carbon. The value of Cj/Co = 0.1
has been taken, arbitrarily, as the initiation of significant carryover of
adsorbate into the effluent, and C2/CO = 0.3, as the point at which the
concentration of adsorbate in the effluent has reached an undesirably
high level. These plots indicate that the experimental data accord well
with the transfer hypothesis over the ranges of C/CO used.
u|u°
0.75
0.70
J_
20
40
60
TIME,
80
100
120
HO
Figure 24. Logarithm oi (1 - •=• ) as a function of time.
Values for 2 have been calculated from the slopes of the straight
lines shown in Figures 24 and 25 and have been tabulated in Table 5
along with those of the parameters needed to compute the corresponding
values for k-p, by using the expression
(ID
Q
m
-------
40
Adsorption of Biochemically
Table 5. TRANSFER COEFFICIENTS AND CORRESPONDING
REYNOLDS NUMBERS
SxlO2,
Column
olu
0.90
0.85
0.80
0.75
g/hr
AxlO
cm
~2
kTx!02,
A
B
C
D
E
F
G
H
I
K
L
2.38
1.56
8.33
0.47
2.00
3.84
3.84
2.08
2.50
1.11
0.89
3.27
1.71
7.29
0.81
3.04
3.00
3.06
3.00
3.07
3.18
3.03
203
203
203
203
299
110
154
240
203
154
375
19.0
17.7
200.4
3.1
9.8
502.4
272.8
13.3
23.2
35.8
15.0
0.55
0.29
1.22
0.14
0.33
0.92
0.68
0.42
0.52
0.70
0.37
0.70
20
40
60 80
TIME, tir
100
T20
Figure 25. Logarithm of (1- -p- ) as a function of time.
.o
140
The term A in Table 5 represents the total external surface area
across which transfer occurs, and thus is equivalent to the product
Values of A have been derived from experimental determinations
-------
Resistant Materials from Solution
41
of the number of particles per gram for each type of carbon used in the
columns, from microscopic observations of shapes and sizes of indivi-
dual particles,^ and with knowledge of the mesh sizes of standard sieves
through which the particles for each size range have passed and by which
they have been retained.
A value of """ = 5.6 x 10-4 for the Columbia carbon over the con-
centration range of interest has been taken from the best linear trace of
the data for the adsorption isotherm for this carbon; the isotherm for
adsorption of the technical ABS on 0.273-millimeter Columbia carbon is
shown hi Figure 26. The value for irhas been taken from a trace from
the origin and intersecting a point corresponding to a solution concen-
10
20 30 40 50 60
EQUILIBRIUM CONCENTRATION, /mol/liter
Figure 26. Adsorption isotherm for technical ABS on 0.273-mm
Columbia carbon, 30 C.
tration of 70 micromoles per liter and an adsorption level of 310 micro-
moles per gram. For calculation of tr , a value for the density of the
carbon of 0.406 gram per cubic centimeter was used. The density of
the carbon was determined by weight count and microscopic examina-
tion of several hundred particles of the carbon. The value of 0.406 gram
per cubic centimeter corresponds to published values for activated
carbons.28
-------
42
Adsorption of Biochemically
No steady-state experiments have been carried out during the pre-
sent investigations for adsorption of the technical ABS on the Pittsburgh
carbon. Isothermal equilibrium experiments have, however, been con-
ducted with straight-chain alkylbenzenesulfonates on both Pittsburgh and
Columbia carbons. These have demonstrated that the particular grade
of Pittsburgh carbon used in the present experiments has a consistently
greater capacity for adsorption of the straight-chain compounds than does
the grade of Columbia carbon, of comparable size, used in the experi-
ment.
Figure 27 is a. plot of typical isotherms for the two carbons; the
isotherms shown are for adsorption of 2-decylbenzenesulfonate on
0.273-millimeter carbon. From comparison of these isotherms for the
two carbons for comparable systemic conditions and extrapolation of
Z
LU
CD
«
s
o
z
x
Q
Ul
tfl
o;
o
700
600
500
400
300
200
I I I
0.273-mm Pittsburgh Corbon
Q,
0,273—mm Columbia Carbon
J I I I I \
04 8 12 16 2° 24 28 32
EQUILIBRIUM CONCENTRATION, /zmol/lite,
Figure 27. Adsorption isotherms for 2 decylbenzenesulfoncte
on carbon, 30 C.
the difference to the isotherm for adsorption of the technical ABS on
the Columbia carbon, and by use of an experimentally determined value
of 0.560 gram per cubic centimeter for the density of the Pittsburgh
carbon, a value of » = 2.7 x 10-4 was computed for the latter adsorbent
in the experimental system.
-------
Resistant Materials from Solution 43
Particle size and flow rate are both characterized in the dimen-
sionless Reynolds number, N'. Thus this group is useful for correla-
R
tion of the transfer coefficients determined from the experiments with
the fluidized columns under different conditions of flow rate and particle
size. For calculations of values for a modified Reynolds Number,
N' = Y. j the velocity term, v, has been taken as the superficial face
velocity of flow through the column. The terms D, P , and n in the
Reynolds group are, respectively, the particle diameter, and the density
and viscosity of the water.
Values for NR are listed in the last column of Table 5; in Figure
28 values of In k,p are shown plotted against values of In N™. This type
of plot is typically used for correlation of heat and mass transfer co-
efficients with dimensionless groups similar to the modified Reynolds
number used in the present correlation.^9 The data are rather well
represented by the trace drawn through the points in Figure 28. Thus the
postulated model seems adequate over the range taken for description
of the rates of adsorption in the experimental fluidized columns.
Implications for Advanced Waste Treatment
Results of laboratory studies indicate that fluidized operation of an
activated carbon bed gives qualitatively the same sigmoid relationship
for the residual concentration of pollutant that characterizes ordinary
packed-bed operation in adsorption or ion exchange. Although many of
the experiments do not exhibit the same sharpness of breakthrough
associated with columnar operation, it must be kept in mind that these
experiments were conducted with very shallow beds of adsorbent and for
the most part with relatively high flow rates. The developing sharpness
with decreasing flow rate and with decreasing particle size implies that
with deeper columns (several feet of adsorbent) and with flow rates of a
few gallons per minute per square foot, sigmoid curves in which a large
fraction of the ultimate capacity will be used before appreciable break-
through occurs should be realized.
That such breakthrough patterns are observed in spite of the
apparent complete mixing of adsorbent particles in the column seems to
be related to the fact that adsorption of material by the carbon results in
an increased density for the carbon. As adsorption proceeds, the rate
of flow of liquid up through the carbon bed must be continuously increased
to maintain the same degree of expansion.
As adsorption proceeds, the nearly saturated particles are lifted
less than fresh ones. Consequently, there is stratification in the bed,
with nearly exhausted adsorbent at the bottom at the point of inflow and
-------
44
Adsorption of Biochemically
6.0
4.0
2.0
1.0
0.8
0.6
0.4
0.2
0.1
0.08
0.06
0.04
0.02
o
_L
O
J I
O.I 0.2 0.4 0.6 0.8 1.0 2.0
NlR
Figure 28- Correlation of transfer coefficient with Reynolds number.
-------
Resistant Materials from Solution
45
with fresh particles at the top to provide a polishing effect. Thus true
counter-cur rent operation is approximately maintained despite the con-
siderable continuous agitation of the bed.
Such segregation is advantageous also because continuous with-
drawal of material from the bottom of the bed will remove only substan-
tially saturated carbon, thus essentially full utilization of capacity can
be obtained before regeneration.
Summarization of data for an appropriate set of comparable
column experiments in a form that perhaps more clearly justifies these
conclusions is shown in Table 6. For compilation of this table from data
for the initial column experiments, A to I, the point at which the concen-
tration of organic matter in the effluent was equal to 0.3 of that in the
influent was taken as representing practical exhaustion of the column.
This arbitrary figure was convenient to use, and the conclusions would
not have been changed qualitatively had another fraction been selected.
Table 6. CAPACITIES OF FLUTOIZED CARBON COLUMNS
TO 0,3 BREAKTHROUGH
Volume of % of total input
effluent at Adsorption to Adsorption to of adsorbate
Flow rate, C/C0 = 0.3, C/C0=0.3, C/C0Z0.3, adsorbed to
Column liters/hr liters (imol, g/liter ^mol/g C/C0=0.3
A
B
C
D
E
F
G
H
I
3.27
1.71
7.29
0.81
3.04
2.99
3.06
3.00
3.07
68.7
77.0
43.7
102.1
149.0
20.9
39.8
138.0
67.5
1.4
1.4
1.3
1.5
1.6
1.3
1.6
1.5
1.5
96.2
107.8
56.8
153.2
238.4
27.2
63.7
207.0
101.3
94
97
86
97
95
75
92
90
94
Column three of Table ( shows the volume of liquid passed through
each bed up to the selected br ^akthrough point. The. e volumes vary with
the rate of flow and particle size of the carbon. The fourth column is
simply the influent concentration less the mean effluent concentration
divided by the weight of carbon in each column. When the figures in
columns 3 and 4 are multiplied, the amount of adsorption that has
occurred up to the selected breakthrough period is obtained. The ulti-
mate capacity of the carbon is about 300 micromoles per gram. Most
of the experiments, therefore, showed about a third to a half of ultimate
capacity at the time of breakthrough with just 8 to 9 inches of settl d
bed. The fine-particle carbon (run E), however, showed almost 80 per-
-------
46
Adsorption of Biochemically
cent realization of capacity at this point, while the coarse-particle
material (run F) had reached only 10 percent of capacity. Variations
•with flow rate are less dramatic, but there is a marked increase with
decreasing flow rate as shown by runs A, B, C, and D, in which the other
variables all have the same value.
Nonetheless, all of the beds except runs C and F show better than
90 percent total removal of organic matter over the period up to the
selected breakthrough point. If multiple columns in parallel are used -
some fresh and some nearly at the breakthrough point - or if an effluent
holding reservoir is provided, use of 0.3 as the breakthrough point is not
unreasonable from a practical standpoint.
200
O
D
O
ex.
X
o
O
CL
100
80
60
40
20
100
80
60
40
20
10
0.1
0.2 0.3 0.4
PARTICLE SIZE, mm
0.5
0.6
Figure 29- Total solute adsorbed per unit weight of adsorbent to 0.3 breakthrough as a
function of particle size, 45 g Columbia carbon, flow rote = 2.5
Figure 29 shows more quantitatively the variation of capacity to
this 0.3 breakthrough point. The plot would indicate that full capacity
should be realized at a rate of 2,5 gallons per minute per square foot
in a 9-inch bed with particle diameters less than about 0.13 millimeter.
-------
Resistant Materials from Solution
47
Undoubtedly the relationship actually tapers off as adsorption near ca-
pacity is reached; however, additional relationships of this sort obtained
with deeper beds should be significant and interesting.
160
140
o
a.
X 120
O
tt
UJ
K
CO
u
0-
u
100
80
60
40
50
40
30
20
0.5
1.0 2.0
FLOW RATE, sp
4.0
6.0 8.0
Figure 30. Total solute adsorbed per unit weight of adsorbent to 0.3 breakthrough gs a
function of flow rate, 45 g 0.273-mm Columbia carbon.
Similarly Figure 30 shows the quantitative relationship between
flow rate and realized capacity. Here the need for data with deeper beds
is even more pronounceo to reach a higher percentage of ultimate capa-
city at the time of breakthrough.
Figure 31 illustrates the difference between the breakthrough
curve and the total uptake indicated by the figures in the last column of
Table 7. The illustration is for run D, the one with the low flow rate.
The ratio of the area under the locus of the experimental points to that
under the straight line is the percent uptake of the organic material.
Finally, Figure 32 shows plots of runs A, B, C, D, with scales of
reduced variables that seem useful in correlating data from these and
subsequent runs. It should also serve as a good representation for
initial determination of the practicability of a given method of operation.
-------
48
1~
uj 160
CO
<£
g 140
o
1 120
= .100
to
O£
O
-------
ADSORPTION ON CARBON IN
BATCH-TYPE SYSTEMS
Extensive investigations of kinetics and equilibria of adsorption
in batch-type systems have been described in an earlier report.1 To
supplement some of the information provided by the earlier investiga-
tions, additional experiments have been conducted with batch systems.
These experiments have included studies of adsorption kinetics and
adsorption equilibria with simple systems composed of different types
of carbon and different compounds than those previously reported, and
some studies with complex mixtures of organic compounds.
The inferences that may be drawn from the present investigations
accord with those set out in the earlier report: thus the data are pre-
sented mainly for comparison of additional systems with those pre-
viously described.
The Columbia LC carbon used in the earlier studies of adsorption
in batch systems was chosen for study because it satisfied operational
requirements, and not because it is the most desirable material for
practical application to waste-water treatment. Indeed, the Columbia
carbon is relatively expensive in comparison with other commercially
available carbons and does not provide as high values for rates and
capacities of adsorption as some others do. Testing the performance of
other commercially available carbons was, therefore, desirable. The
materials chosen for study were Pittsburgh carbon (Pittsburgh Chemical
Co.), Darco carbon (Atlas Powder Co.), and Norit carbon (American
Norit Co.). Some rates of adsorption with the Darco and Norit carbons
relative to that of the Columbia carbon have been reported previously; 1
the additional rate studies have been limited to the Pittsburgh carbon.
All four carbons have been comparatively tested in studies of adsorption
equilibria.
Rates of Adsorption
ADSORPTION ON PITTSBURGH CARBON
Rates of uptake of technical grade ABS -were studied for various
particle sizes of Pittsburgh carbon. Results of these experiments are
shown in Figure 33.
The relative rate constants for these systems taken from the
e 33 range from 20 micromoles per
49
-------
50
Adsorption of Biochemically
carbon. Previous studies have indicated that the value of the relative
rate constant for adsorption of the technical ABS by 0.126-millimeter
Columbia carbon is 52 micromoles per gram/(hour)V2 [3.9.4 mg/g
(hr) V?]. 1 The relative rate constant predicted for 0.200-millimeter
Columbia carbon, from the dependence of this parameter on the inverse
square of particle diameter, 1 is about four-tenths of 52, or approxi-
mately 21 micromoles per gram/(hour)l/2. Thus the Pittsburgh carbon
has been found to exhibit a rate constant for adsorption about five times
as large as that for Columbia carbon in systems with technical ABS.
The notably greater rate for the Pittsburgh carbon in the rapidly stirred
batch system, in which intraparticle transport is quite probably rate
controlling, can presumably be attributed to larger pore diameters,
which permit more rapid transport of the highly branched technical
ABS, This effect should decrease with decreasing molecular size and
complexity of the adsorbate.
240
I I I I i JO
— 74.4
— 59.5
_ 44.6
— 29.8
— 14.9
TIME, hr
Figure 33. Rates of adsorption of technical ABS on Pittsburgh carbons of
different particle size, 100 mg/liter carbon, 30 C.
The relative rate constants for the different sizes of Pittsburgh
carbon correlate well with the square of the inverse of particle di-
ameter, as illustrated in Figure 34. This correlation accords with that
noted for different sizes of Columbia carbon, 1 and identical conclu-
sions may be drawn.
-------
Resistant Materials From Solution
120
^ 100
i-
z
z
o
o
BO
40
S 20
I
37.2
29.8
22.3 ^
v,
01
E
14.9
7.4
10 15 20
DIAMETER-2,mm-2
25
30
Figure 34. Relative rate constant as a function of the inverse of the square of particle
diameter; lOOmg liter Pittsburgh carbon technical ABS 30 C.
ADSORPTION FROM MIXTURES OF ORGANIC COMPOUNDS
An investigation of the rate of adsorption on carbon from a com-
plex mixture of organic compounds typical of the resistant materials
that might be expected to occur in some combination in waste effluents
was carried out to gain some insight into the characteristics of adsorp-
tion from complex mixtures. The results of the rate experiment with
a mixture of technical ABS, nonylphenoxylnonylethoxyethanol,
dichlorodiphenyltrichloroethane, nicotinic acid, and triethanolamine are
illustrated in Figure 35.
The relative rate constant for the data shown in Figure 35 is
15 milligrams per gram/(hour)l/2 or 1.5 percent by-weight adsorption
per (hour)l/2 for the 0.273-millimeter Columbia carbon. For adsorp-
tion of the technical ABS alone on 0.126-millimeter Columbia carbon,
the relative rate constant has a value of about 52 micromoles per
gram/(hour) 1/2 Or 19.4 milligrams per gram/(hour)1/2.1 Extrapola-
tion of this relative rate constant to 0.273-millimeter carbon gives a
value of approximately 4 milligrams of total material per gram/(hour)1/2
as against nearly four times this much for adsorption of organic carbon
alone from the mixture. The increased total rate thus noted is in
accord with previous observations for studies of competitive adsorption
in bisolute systems.28 The much larger initial concentration in the
-------
52
Adsorption of Biochemically
mixed solutions (Co = 29.5 mg organic carbon/liter) must contribute
in part to the increased rate of adsorption.
40
30
at
o
t: 20
I
z
o
u
y
2
O
SOLUTION; C0- 293 mg ORGANIC CARBON/lit«r
".O ing/liter TECHNICAL ALKYLBENZENESULFONATF
10.2 mg/liter NONYLPHENOXYNONYLETHOXYETHANOL
9.5 mg/liter DICHLORODIPHENYLTRICHLOROETHANE
|5.2 nig/liter NICOTINIC ACID
5.7 ma/liter TRIETHANOLAMINE
I 2
TIME, hr
Figure 35. Rate of adsorption organic carbon from a mixture of
organic compounds by carbon;
500 mg/liter 0.273-mm Columbia carbon, 30 C.
Equilibria and Capacities for Adsorption
EFFECT OF PARTICLE SIZE
The increase of capacity of carbon with decreasing particle size
noted tentatively in the first report1 has been investigated in more
detail. Data from isotherms for the adsorption of 3-dodecylbenzene-
sulfonate on three different sizes of Columbia carbon at 30°C are
shown in Figure 36. An increase in capacity with decreasing particle
size is apparent and appears significant. The quantity of solute adsorbed
per unit weight of carbon for equilibrium concentrations in solution
corresponding to apparent monolayer saturation of the carbon has been
found to correlate with the inverse of the diameter squared for the data
-------
Resistant Materials from Solution
53
plotted in Figure 36. This correlation is illustrated in Figure 37. Thus
an Increase in capacity of approximately 50 percent results in reduction
of diameter by a factor of two.
600 —
z
2
ct
u
I-
*
Of
8
Q
UJ
H
400
200
0.126-mm
1—o—(p—o—o—o •
0.503-mm
21
Q
LU
CD
14 2
UJ
*
t-
UJ
0 10 20 30
CONCENTRATION OF SOLUTE IN SOLUTION AT EQUILIBRIUM, ftmol/liter
Figure 36. Isotherms for adsorption of 3-dodecylbenzenesulfonate by carbons
of different particle size; Columbia carbon, 30 C.
ISOTHERMS FOR DIFFERENT CARBONS
In Figure 38, data obtained in studies on equilibria of adsorption
of 2-decylbenzenesulfonate on different commercially available 0.273-
millimeter carbons have been plotted. The Pittsburgh carbon apparently
has a much greater capacity for this surfactant than does any of the
other carbons tested, more than 20 percent by weight for solution con-
centrations of 20 micromoles per liter (6.5 mg/liter) or greater. More-
over, the capacity of the Pittsburgh carbon equilibrium increases more
with increase in solute concentration than does that for any of the other
adsorbents studied.
Because of its relatively large capacity and high rate of uptake,
the Pittsburgh carbon appears more suitable than the other studied for
application to •waste-water treatment.
-------
54
Adsorption of Biochemically
600
o
E
a.
o
en
ar
<
u
t-
Z
\
a
UJ
ta
cc
o
UJ
I-
3
500
400
21.0
m
tc.
O
17.5
UJ
u
QL
UJ
14.0 o-
10
20
30
4C
50
60
DIAMETER'2,"™-2
Figure 37. Plateau value for adsorption isotherm as a function of the inverse of the square
of particle diameter; 3-dodecylbenzenesulfonate, Columbia carbon, 30* C.
ISOTHERMS FOR ADSORPTION OF NITROCHLOROBENZENES
ON CARBON
The occurrence of nitrochlorobenzenes as discharged waste in
river waters has been noted by Middleton.30 Previous studies on rates
of adsorption of this compound by carbon have been reported.* To
supplement the previous information on the characteristics of adsorp-
tion of this pollutant, investigations of adsorption equilibria have been
carried out.
The Columbia carbon used in the present experiments appears to
have a remarkably large capacity (nearly 40% by weight) for adsorption
of nitrochlorobenzene, as may be observed from the data shown in Fig-
ure 39. This large capacity may be due in part to the fact that the
nitrochlorobenzenes are only slightly soluble in water (the solute was
originally dispersed in the water by dilution of a 10~2 M solution of
nitrochlorobenzene in methanol). Thus the equilibrium is strongly in
favor of deposition of material in the pore spaces of the carbon.
ISOTHERM FOR A MIXTURE OF ORGANIC COMPOUNDS
The equilibria of adsorption from a complex mixture of organic
compounds, identical to the mixture used in the rate studies described
previously, were investigated by using 0.273-millimeter Columbia
-------
Resistant Materials from Solution
55
700
600
500 —
400 —
300
200
100 —
0 4 8 12 16 20 24 28 32
CONCENTRATION OF SOLUTE IN SOLUTION AT EQUILIBRIUM, ^utiol/liter
Figure 38. Isotherms for adsorption of 2-decylbenzenesulfonate on different carbons;
0.273-mm particle size, 30° C.
0
E
a.
z"
o
D;
U
i-
*
D
UJ
m
o:
8
o
I I t I ^
E
E
•z.
o
m
a.
*
H
Z
2 0.8
400
— 300
- 200
- TOO
0 10 20 30
CONCENTRATION OF SOLUTE IN SOLUTION AT EQUILIBRIUM, /jmol/liter
Figure 39. Isotherms for adsorption of nitrochlorobenzenes on carbon;
0.503-mm Columbia carbon, 30 C.
-------
56
carbon. "Die isotherm obtained from these studies is shown in Figure
40. Organic carbon was used as a measure of the extent of adsorption.
Obviously, from Figure 40, the capacity of the carbon is enhanced
considerably in the presence of several compounds as compared with
that obtained in single-solute solutions.*
t-
z
ILI
CO
DC
o
I-
*
z
o
n
ae.
U
u
o
BL
O
150
100
SO
0
1
-CT
-?'
SOLUTION; C
| 20.0 ing/liter
10J mg/liter
*- 9.5 ing/liter
5.2 mg/liter
5.7 mg/liter
1
1
— —
1 0
_
o
o
= 29.5 mg ORGANIC CARBON/titer
TECHNICAL ALKYLBEN2ENESULFONATE
NONYLPHENOXYNONYLETHOXYETHANOL
DICHLORODIPHENYLTRICHOROETHANE —
NICOTINIC ACID
TRIETHANOLAMINE
1
t 1
04 8 12 16
CONCENTRATION OF ORGANIC CARBON IN SOLUTION AT EQUILIBRIUM, mg/liter
Figure 40- Isotherm for adsorption of organic carbon from a mixture of organic
compounds by carbon; 0.273-mm Columbia carbon, 30* C.
GPO 8*5-048-3
-------
ADSORPTION OF ORGANIC
PESTICIDES ON CARBON
Environmental dispersal of organic pesticides has been a matter
of much recent and widespread concern. The increased use of organic
pesticides in agricultural, domestic, and industrial applications and the
consequent increase in the occurrence of these materials in water and
wastes have brought about a potential hazard for the health and well-
being of the public.31
Adsorption by carbon appears to have promise as a technique for
removing potentially harmful pesticides from waters and wastes.
Accordingly, studies of rates and equilibria of adsorption have been ex-
tended to a number of representative pesticides or substances mole-
cularly related to them.
Pesticide materials chosen for study include compounds of dif-
ferent chemical classes, typical of various types of herbicides, acari-
cides, and insecticides. The pesticides studied, along with some of
their significant properties, are listed in Table 7.
As shown in Table 7, the chemical classes represented by the
pesticides studied include thiophosphates (0,0-diethyl-O-g-nitrophenyl
phosphorothioate), carbamates (1-naphtyl N-methylcarbamate), dinitro-
phenols (2-sec-butyl-4, 6-dinitrophenol and 2-cyclohexyl-4,6-dinitro-
phenol), and chlorophenoxy acids [2,4-dichlorophenoxyacetic acid, 2,4,
5-trichlorophenoxyacetic acid, and 2-(2,4,5-trichlorophenoxy) propionic
acid].
The 0,0-diethyl-0-2-nitrophenyl phosphorothioate (Parathion) was
obtained from the -American Cyanamid Company; the 1-naphthyl-N-
methylcarbamate (Sevin), from the R.A. Taft Sanitary Engineering
Center (U.S. Public Health Service); and the remainder of the compounds,
from the Dow Chemical Company. All compounds as obtained were
described by the suppliers as being chemically "pure" relative to the
active chemical agent.
Columbia LC coconut carbon was used as the adsorbent in all
experiments with the pesticides. Prior to use in the experiments, the
adsorbent was separated by thorough sieving into uniform particle
sizes, after which portions of suitable size range were washed in dis-
tilled water to remove any leachable impurities and adherent powder,
and dried at 105°C. The size range chosen for the present studies in-
cluded those particles passing a U.S. Standard Sieve No. 50 and being
retained on a No. 60 sieve; the mean particle diameter for this size
range is 0.273 millimeter.
57
-------
58
Adsorption of Biochemically
Table 7. SELECTED ORGANIC PESTICIDES
Compound
Primary
application
Molecular
weight
Approximate
solubility in
water at25"C,
mg/liter
2,4-dichlorophenoxyacetic
acid herbicide
2,4,5-trichlorophenoxyacetic
acid herbicide
221.0
255.5
900
280
2 - (2 , 4 , 5 -trichlorophenoxy) -
propionic acid
2-sec-butyl-4 , G^dinitro-
phenol
2-cyclohexyl-4 , 6-dinitro-
phenol
1-naphthyl N-methylcarbamate
0 , 0-diethyl-0-p-nitrophenyl
phosphorothioate
herbicide
herbicide ,
insecticide
acaricide,
insecticide
insecticide
acaricide,
insecticide
269.5
240.2
266.2
201.2
291.3
180
52
10
99
24
To eliminate from the experimental systems all extraneous
material that might have introduced sources of error and interference,
all solutions of the pesticides used in the experimental systems were
prepared with twice-distilled water. Cambridge tap water was first
distilled in a conventional tin-lined still, and the condensate was trans-
ferred into an all-glass still for redistillation. Preliminary boiling with
escape of steam was used to purge the water of dissolved carbon dioxide
and chlorine before collection of the condensate from the redistillation.
Experimental Details
ANALYTICAL METHODS
Of the substances investigated, only Parathion does not exhibit a
characteristic absorption spectrum suitable for direct measurement of
concentration within the ranges used in the present experiments. Spectra
for the six other pesticides are shown in Figures 41 through 46.
The analytical method employed for Parathion takes advantage of
the ease with which the phosphorus-oxygen bond in that material is
cleaved by hydroxide. The hydrolysis produces a mole of £-nitrophenol
-------
Resistant Materials from Solution
59
for each mole of Parathion originally present in solution. Because
o-nitrophenol has a well-defined absorption spectrum with a high molar
absorptivity at the wavelength of maximum adsorption (400 m/x), and be-
cause the reaction to form p-nitrophenol from Parathion is rapid and
quantitative, the method provides a reliable and sensitive means for
analysis of Parathion.
o
m
tt
8
o
0.2 —
0.1 —
o L
250
260
270
280
290
300
310
WAVE LENGTH (X),nyi
Figure 41- Absorption spectrum for 2,4-dichlorophenoxy-ocetic acid;
cell fength 5 cm, C = 70 /unoMiter.
For fairly concentrated samples (1-100 MM), about 10 milliliters
of 1 N KOH were added to 50 milliliters of sample. The mixture was
heated to boiling, allowed to cool, diluted to 100 milliliters, and then
the absorbance at 400 millimicrons was measured in a cell of appropri-
ate length. For more dilute samples, 3 KOH pellets were added to a
100-milliliter sample, and the procedure outlined above was repeated.
The lower limit of the method is about 50 nanomoles per liter with a
10-centimeter cell. The absorption spectrum for p-nitrophenol is shown
in Figure 47.
-------
60
Adsorption of Biochemically
UJ
-------
Resistant Materials from Solution
61
I I 1 1
260
270
280
290
300
310
320 330 340
WAVELENGTH
Figure 43. Absorption spectrum for 2-(2,4,5-trichlorophenoxy)-propicnic acid;
cell length 5 cm, C = 60 /imol/liter.
-------
62
Adsorption of Biochemically
to
Of
o
0.7
0.6
0.5
0.4
0.3
0.?
I I
I
I
_L
J_
I
340 350 360 370 380 390 400
WAVELENGTH (X),mfi
425
Figure 44. Absorption spectrum for 2-sec-butyl-4.6-dinitrophenol;
cell length 1 cm, C - 50 ^mol/liter.
U4
O
z
m
K
s
tn
0.7
0.6
0.5
0.4
0.3
0.2
f
I
I
I
I
340
350
425
360 370 380 390 400
WAVELENGTH (X),mp
Figure 45. Absorption spectrum for 2-cylonexyl-4,6-dinitrophenol;
cell length 1 cm, C = 50 ^mot/liter.
-------
Resistant Materials from Solution
63
After measurement of the absorption spectrum for each solute had
been made with a Beckman Model DK-2 spectrophotometer, the absorp-
tion maximum for solutions of each of the substances was located pre-
cisely with a Beckman Model DU spectrophotometer by measuring in-
tensities of absorption for a localized region on both sides of the wave-
length of maximum absorption indicated by the spectra obtained with the
Model DK-2. Subsequent spectrophotometric measurements were made
with the Model DU.
0,7
0.6
0.5
0.4
o
z
o
i/J
2 0.3
0.2
0.1
250
260
270
280
290
300
3)0
WAVELENGTH
Figure 46. Absorption spectrum for 2-nophthyl-n-methylcarbamate;
cell length 5 cm, C = 23
Calibration curves for all the solutes are given in Figures 48
through 54. Wavelengths of maximum absorption, and observed values
for the molar absorptivity for each material are listed in Table 8.
-------
64
Adsorption of Biochemically
I I I I 1
340 350 360 370 380 390 400
WAVELENGTH (A), mfi
425
Figure 47- Adsorption spectrum for £-nitrophenol; cell length 1 cm,
C = 40 /imol/liter.
-------
Resistant Materials from Solution
65
Table 8. ANALYTICAL DATA FOR ORGANIC PESTICIDES
Compound
Wavelength of maximum
absorption,
Molar absorptivity,
{cm2/mol) x 10-6
2,4-dichlorophenoxy-
acetic acid
2,4,5-trichlorophenoxy-
acetic acid
284
289
1.9
2.5
2-(2 , 4 , 5-trichloro-
phenoxy) propionic acid
2-sec-butyl-4 , 6-dinitrophenol
2-cyclohyxyl-4 , 6-dinitrophenol
1-naphthyl-N-methylcarbamate
£ -nitrophenol (for Parathion)
289
375
376
279
400
2.5
14.5
14.0
5.9
18.0
10
20 30 40 50
CONCENTRATION, fimol/liter
Figure 48. Calibration curve for 2,4-dichlorophenoxyacetic acid;
cell length 5 cm, X= 284 m^.
-------
66
Adsorption of Biochemically
1.0
0.8
Z1 0.6
-------
Resistant Materials from Solution
67
4 6 8 10
CONCENTRATION, litnol /liter
12
Figure 51. Calibration curve for 2-sec-butyl-4, 6-dinitrophenol;
cell length 5 cm, \- 375 ffl(i.
4 6 8 10
CONCENTRATION, /uno|/|iter
12
Figure 52. Calibration curve for 2-cyclohexyl-4,6-dinitrophenol;
cell length ] cm, X= 376 mfi-
-------
68
Adsorption of Biochemically
UJ
O
<
m
a.
s
0.8
0.6
0.4
0.2
I
J_
f
I
_L
J_
' 1 2345 67
CONCENTRATION, /unol/liter
Figure 53. Calibration curve for p-nitrophenol; cell length 5 cm.
0.8
2
a
0.2 -
20
too
40 60 80
CONCENTRATION, /jmol/liter
Figure 54. Calibration curve for 1-naphthyl-N-methylcorbomate;
cell length 1 cm, X= 279 m^t-
EXPERIMENTAL METHODS
Agitated nonflow experiments were used in investigating both the
kinetics and equilibria of adsorption of the pesticides on carbon. The
batch technique was selected because of its relative simplicity. Advan-
tages of this type of system that figured significantly in its selection
were its freedom from complex hydraulic parameters indigenous to
nowthrough systems, its adaptability to small-volume work, the ease of
investigation of variation of conditions, and the general facility of
operation.
Although column operation seems likely to be used for large-scale
technical applications of carbon adsorption to waste treatment, evalua-
tion of the fundamental characteristics of adsorption is simpler with a
batch technique. Furthermore, once suitable functional relationships
for the variables have been established, extrapolation of data from batch
-------
Resistant Materials from Solution 69
systems to the prediction of behavior in continuous systems should be
feasible.
The reaction vessels for the kinetic studies were 4,000-milliliter
resistant-glass bottles, in which solutions were agitated with Teflon-
coated stirring rods extending directly into the adsorbate solution and
connected to synchronous motors operating at 1,550 rpm. Previous
kinetic studies in similar systems had shown that the rate of adsorption
was independent of stirring rate at rotations greater than a few hundred
revolutions per minute.^2 Temperature control for the rate studies was
maintained by immersing the reaction vessels in water baths, thermo-
regulated to 0.5°C. All kinetic studies have been conducted at 28°C.
For each experiment on rate of adsorption, an 8-liter volume of
adsorbate solution was prepared at a concentration slightly larger than
that desired for the kinetic experiment. Spectrophotometric measure-
ment of the exact concentration of solute in the solution was then made,
and the original 8-liter volume diluted appropriately to give more
closely a solution of the desired concentration. Exactly 3,000 milliliters
of the adjusted solution was then placed in each of two 4-liter reaction
vessels. The reaction vessels were then placed in a water bath for
approximately 24 hours to allow the solution to attain the desired tem-
perature and to reach adsorptive equilibrium with the surfaces of the
vessels. Before introduction of a suitable, accurately weighed quantity
of carbon to each vessel, an initial sample was removed and analyzed
for concentration 01 adsorbate. This gave an accurate check on the
preparation of the solution and served as the reference concentration
for the rate study. The carbon was then introduced into the vessel and
rapidly dispersed by the motor-driven stirrer operating at 1,550 rpm.
At appropriate intervals the stirring was briefly interrupted while
samples of the supernatant solution were pipetted from each of the
reaction vessels; the samples for analysis resulted in some small de-
crease in total adsorbate available to the adsorbent, but the cumulative
error of 2 to 3 percent was not significant for this type of experiment.
The running of two parallel experiments, identical in all details, pro-
vided a check on the reproducibility of the data.
For investigation of adsorptive capacities and equilibria, large
volumes of adsorbate solution were prepared and 250 milliliter mea-
sures of this solution were dispensed into 300-milliliter resistant-glass
reaction flasks. A suitable, accurately weighed amount of adsorbent
was then added to each flask, the weight being varied so as to cover the
range of equilibrium solute concentrations of interest. For each study,
several flasks of solution were left without carbon for use as blanks.
The ground-glass-stoppered reaction flasks were then sealed and placed
in an oscillating shaker to be shaken for approximately 2 weeks, the
times required for equilibrium to be obtained having been determined
previously for each type of system. Samples from each of the solutions
were analyzed at the end of the appropriate period; the amount of
adsorption was computed from the difference between the concentration
of adsorbate measured for the blanks and for each reacted solution.
-------
70 Adsorption of Biochemically
Most studies were conducted in a room maintained at a reasonably uni-
form temperature, so that fluctuations in the temperatures of the solutions
•were no more than 2 to 3°C over the duration of the study. The mean
temperature for the equilibrium studies was 25°C.
Rates of Adsorption
Previous investigations have indicated that the rate-limiting step
for removal of organic solutes from dilute aqueous solution by porous
activated carbon in agitated nonflow systems is one of intraparticle
transport of the solute in the pores and capillaries of the adsorbent.32
For systems in which intraparticle transport is the rate-limiting step,
data for uptake of solute from solution should give a linear plot as a
function of the square-root of time from introduction of the adsorbent to
the system.^ In accord with this method of representing data for sys-
tems in which intraparticle transport is the rate-limiting mechanism,
the data for the present experiments have been plotted as a function of
the square-root of time.
The rate data for the experiments on adsorption of the pesticides on
activated carbon are plotted in Figures 55 through 61. The (Cq-C)/m
values in these plots represent the amount of solute, both in micromoles
and milligrams, removed from solution per gram of carbon. Good linear-
ization of the data is observed for the experiments, in accord with ex-
pected behavior for intraparticle-transport rate control. The linear
traces facilitate comparison of relative rates of adsorption of the organic
pesticides, and this comparison is made in Table 9 by using the square
of the slope of each plot as the relative rate constant for the experiment.
Table 9. RELATIVE RATE CONSTANTS FOR ADSORPTION OF
ORGANIC PESTICIDES ON CARBON;
28°C, 25 mg/liter 0.273-mm COLUMBIA CARBON, CQ = 10 /«nol/liter
Relative rate constant, k,
Compound [
-------
Resistant Materials from Solution
71
300
200
100
- (66.3)
- (44.2)
Q/O
°
y
o/o
t0.5,h,0.5
Figure 55. Rate of adsorption of 2, 4-dichlorophenoxyacefic acid.
200 i— (5i.i)
o
II
u
100
— (25.6)
8
,0.5,hr0.5
Figure 56. Rate of adsorption of 2,4,5-trichlorophenoxyacetic acid.
-------
72
Adsorption of Biochemically
Note that the rate constants for adsorption of the various organic
pesticides on the 0.273-millimeter carbon are remarkably similar. The
data suggest that, even for a rather broad spectrum of different types
of organic pesticides, similar rates of removal from solution should
obtain. Thus, the effectiveness of activated carbon as related to rate
of removal of organic pesticides from solution should be relatively
independent of the type of organic pesticide, at least within the classes
of compounds and conditions used for the present experiments.
300
200
IT-
100
(53.9)
frt
— (27.0)
8*°'
r
,0.5. hrO-5
Figure 57. Rate of adsorption of 2-(2,4,5-trichlorophenoxy)-
propionic acid.
Concentration of solute in solution was noted to have a significant
effect upon the rate of removal by activated carbon. The data shown in
Figures 55 through 61 are for experiments conducted with a solution con-
taining initially about 10 micromoles of solute per liter. Figure 62
shows some data for experiments conducted with some of the pesticides
at concentrations greater than and slightly smaller than 10 micromoles
per liter. The uppermost curve in Figure 62 is for adsorption of 2,4-
dichlorophenoxyacetic acid from a solution having an initial concentra-
tion of 52.2 micromoles per liter. The middle curve is for 2-(2,4,5-
trichlorophenoxy) propionic acid in a solution having an initial concen-
tration of 45.2 micromoles per liter, and the lower curve is for adsorp-
-------
Resistant Materials from Solution
73
tion of 2-cyclohexyl-4,6-dinitrophenol from a solution with an initial
concentration of 7.7 micromoles per liter.
300
0.5
Figure 58. Rate of adsorption of 2-sec-butyl-4,6-dinitrophenol.
For the 2,4-dichlorophenoxyacetic acid, the increase by a factor
of 5.1 in the initial molar concentration, over that used in the experi-
ment represented by the data in Figure 55, resulted in a new rate con-
stant of 16.0 x 104 (micromoles/gram)2/hour as compared with a
relative rate constant of 1.44 x 104(micromoles/gram)Vhour for the
experiment in which the initial concentration was 10.3 micromoles/liter.
The experiment at high concentration with 2-(2,4,5-trichiorophenoxyJ
propionic acid, which represented an increase in initial molar concen-
tration by a factor of 4.6 times that used for the experiment represented
by the data in Figure 57, yielded a new relative rate constant of 7.56 x
I04(micromoles/gram)2/hour as compared with a value of 0.71 x 10*
(micromoles/gram)2/hour, for the experiment with an initial molar
concentration of 9.8 micromoles/liter.
Note that the ratio of the relative rate constants for the two solutes,
2,4-dichlorophenoxyacetic acid and 2-(2,4,5-trichlorophenoxy) propionic
-------
74
Adsorption of Biochemically
acid, obtained from the experiments at small concentration is 2.03,
while the ratio of the relative rate constants obtained from the experi-
ments at the large concentrations is 2.11. Apparently, therefore, the
effect of initial concentration upon the rate at which organic pesticides
are removed from water by activated carbon is relatively uniform, at
least for similar classes of organic pesticides.
300
200
100
- (53.2)
,0.5,
Figure 59. Rate of adsorption of 2-cyclohexyl-4,6-dinitrophenol.
The relative rate constant for the 2-cyclohexyl-4,6-dinitrophenol
obtained from the experiment with an initial molar concentration of 7.7
micromoles per liter was 0.90 x 104(micromoles/gram)2/hour as com-
pared with a value of 1.12 x 104(micromoles/gram)2/hour obtained for
the experiment with an initial molar concentration of 10.2 micromoles/
liter. Thus the effect of initial concentration appears qualitatively uni-
form for different classes of organic pesticides.
-------
Resistant Materials from Solution
75
°
V
u
loo —
°-5
Figure 60. Rate of adsorption of 1-naphthyl-N-methylcarbamate.
Adsorption Equilibria
The extent to which the full surface area of an activated carbon
can be used for adsorption depends on the concentration of solute in the
solution with which the carbon is mixed. The specific relationship be-
tween concentration of solute and the degree of its removal from solu-
tion by an adsorbent at constant temperature and for conditions of equili-
brium defines the adsorption isotherm. The preferred form for repre-
senting the adsorption isotherm is to express the quantity of solute ad-
sorbed per unit weight of adsorbent as a function of C, the equilibrium
concentration of solute remaining in solution.
A number of different types of adsorption relationships may obtain
under different circumstances. The most common relationship between
the amount of solute adsorbed per unit of adsorbent and the equilibrium
concentration in solution is obtained for systems in which it appears that
adsorption from solution leads to the deposition of only a single layer
-------
76
Adsorption of Biochemically
of solute molecules on the surface of the solid. This type of adsorption
equilibrium is best represented by the Langmuir model for adsorption,
which assumes that maximum adsorption corresponds to a saturated
monolayer of solute molecules on the adsorbent surface, that the energy
of adsorption is constant, and that there is no movement of adsorbate
molecules in the plane of the surface after initial adsorption.34
300
200
-
100
— (58.2)
Figure 61. Rote of adsorption of 0, O-dietfiyl-0-p-nitrophenol.
Hie form of the Langmuir adsorption isotherm is
X = XmbC
+bC
(12)
in which C indicates the measured concentration in solution at equili-
brium, Xm is equal to the number of moles of solute per gram of carbon
adsorbed in forming a complete monolayer on the carbon surface, X
represents the number of moles of solute adsorbed per gram of carbon
at concentration C, and b denotes a constant related to the energy of
adsorption. The reciprocal, 1/b, is the concentration at which adsorp-
tion attains half its limiting Xm value.
-------
Resistant Materials from Solution
77
Two convenient linear forms for the Langmuir equation are:34
bxmc
and
c =
X
(13)
(14)
Either Equation 13 or 14 may be used for linearization of data that
conform to the Langmuir equation. The form chosen usually depends on
the range and spread of the data and on the particular range of data to
be emphasized.
600 —
U
400 ~
Figure 62- Rates of adsorption of 2,4-dichlorophenoxyacetic acid, 2-(2,4>trichlorophenoxy)
pfopionic acid, and 2 -cyclohexyl-4,6-dinitrophenol from solutions
of different initial concentration.
All the organic pesticides tested exhibit equilibrium adsorption on
the experimental carbon in reasonable accord with the Langmuir model
for adsorption. The data for isothermal equilibrium adsorption experi-
ments are plotted in Figures 63 through 68. The points on these plots
are the experimental data; the solid curves through the points are cal-
culated Langmuir adsorption isotherms. Figures 69 through 74 are
linear plots of the data according to the linear form of the Langmm
-------
78
Adsorption of Biochemically
adsorption isotherm given by Equation 13. The intercept, with the
ordinate, of each line drawn through the data in Figures 69 thru 74
represents the reciprocal of the ultimate capacity for adsorption, _!,
X
and the slope of each plot represents the reciprocal of bXm. From the
intercepts and slopes of the linear Langmuir plots, values for the
Langmuir constants for each solute have been calculated; these values
for the Langmuir parameters are listed in Table 10.
1,800
(397.8) -,
Figure 63- Adsorption isotherm for 2,4-dicKlorophenoxyocetic acid.
As was the case for the relative rate constants for the different
organic pesticides, values of Xm, listed in Table 10, are similar in
magnitude for the different adsorbates, on the order of 1,800 micro-
moles of solute per gram of carbon. Thus, the ultimate capacity
appears to be relatively independent of the type of organic pesticide
tested, at least within the experimental ranges. With the exception of
the Parathion, values for the parameter b, which is a measure of the
energy of adsorption, are similar for the organic pesticides. The fact
that the value for b for the Parathion is an order of magnitude higher
-------
Resistant Materials from Solution
79
Table 10. LANGMUIR CONSTANTS FOR ADSORPTION ISOTHERMS OF
PESTICIDES ON ACTIVATED CARBON;
25CC, 0.273-mm COLUMBIA LC CARBON
Compound
2,4-dichlorophenoxyacetic acid
2,4,5-trichlorophenoxyacetic acid
2-(2,4,5-trichlorophenoxy)
propionic acid
2-sec-butyl-4 , 6-dinitrophenol
2-cyclohexyl-4 , 6-dinitrophenol
0 , 0-diethyl-0-£-nitrophenyl
phosphorothiate
Xm,
-------
80
Adsorption of Biochemically
than values of b for the other organic pesticides indicates that this
material has a high energy of adsorption and attains adsorptive capa-
cities near the ultimate capacity at relatively small equilibrium con-
centrations of solute in solution.
1.800
1,600 -
1.400 -
1,200 -
1,000 -
o
O
-J (485.1) -
(269.5) -
(161.7) -
(mg/liter)
(5.39) (10.78)
I I I
(53.9) -
(16.17)
I
10
20
30
40
50
60 70
C*q '
Figure 65. Adsorption isotherm for 2-(2,4,5-trichlorophenoxy)-propionic acid.
Comparison may also be made between the values of Xm and of b
obtained -with these pesticides and those obtained earlier for ABS and
other compounds. Results of some of these earlier studies along with
those for the pesticides are shown in Table 11 with somewhat different
units than those given in Table 10.
The capacity values for the pesticides are very large, even greater
on a weight basis than that previously reported for p_-nitrochloroben-
zene. On the other hand, the 1/b values, except for that of Parathion,
indicate approach to the saturation adsorption only at relatively large
residual concentrations of the adsorbate.
The capacity figures are probably the most significant for column
operation, for countercurrent operation makes realizable capacities
-------
Resistant Materials from Solution
81
Table 11. COMPARISON OF LANGMUIR CONSTANTS FOR
VARIOUS ADSORBATES ON ACTIVATED CARBON
Compound
Xm> 1/b,
mmol/g mg/g AtmolAiter mgAiter
2,4-D
2,4,5-T
Silvex
DNOSBP
DNOCHP
Parathion
Phenol
2-dodecylbenzenesulfonate
l-chloro-4-nitrobenzene
1.75
1.75
1.72
1.85
1.88
1.82
1.09
0.40
2.52
387
448
464
444
500
530
103
139
400
10.5
6.7
6.9
5.8
6.8
0.82
9.3
0.22
1.4
2.32
1.71
1.86
1.39
1.81
0.24
0.87
0.078
0.22
1.800
(432.4)-,
(336.3) -
200
(mg/g) (240.2) -
(144.1) -
r
1
0
(rug/liter)
(4.80) (9.61)
1 | 1 1
10 20 30 40
C ol/fiter
eq ( r^
(48.0) _
(14.41)
' "
50 60 7C
Figure 66- Adsorption isotherm for 2-5ec-butyi-4,6-dinitrophenol.
-------
82 Adsorption of Biochemically
those corresponding to initial concentrations rather than to allowable
effluent concentrations. The 1/b values give an indication of the antici-
pated sharpness of the breakthrough front and to some extent provide
guidance for the optimum depth of column for satisfactory removal of
the contaminant, although rate of adsorption is an equally important
factor in this respect. Larger 1/b values in general would call for
greater depths of adsorption column.
Analysis of this type can give only very qualitative conclusions
with regard to the pesticides, for obviously, the basic assumption of
the Langmuir treatment, that only a mono molecular layer of adsorbed
material is formed, does not hold when one is considering weights of
adsorbed material in the neighborhood of 40 percent of the weight of
adsorbent. While the Langmuir formulation does provide an adequate
representation of the data for adsorptive equilibrium over a large and
relatively high concentration range, it seems likely that more refined
experimentation in a smaller concentration range (say 1 to 100 ppb)
would show considerably greater adsorption of the pesticides than that
predicted by the Langmuir constants presented here. Indeed, the
experimental results for the lower end of the concentration range shown
in Figures 66 and 67 provide some evidence of this. Two different
Langmuir formulations may be needed for more exact description such
as those found for phenol in smaller and larger ranges of concentration.
There the value of Xm at larger concentrations was found to be 3,32
millimoles per gram or 310 milligrams per gram rather than the fig-
ures recorded in Table 11. Note that the phenol concentrations giving
the higher capacity were from 6 to 14 percent of the saturation values
in water. Many of the pesticides were studied in about this range of
fractional concentrations of their saturation values.
Consequently, cleanup of small concentrations of pesticides by
activated carbon seems likely to be considerably more efficient than
indicated by the Langmuir constants obtained in these studies. Prob-
lems of initial leakage of these materials are not expected to occur
when beds of adequate depth (5 to 10 ft) of activated carbon are used for
advanced waste treatment.
-------
Resistant Materials from Solution
83
1,800
Figure 67. Adsorption isotherm for 2-cyclohexyl-4, 6-dinitrophenol.
-------
84
Adsorption of Biochemically
III
10 12 '4 16
400
200
Figure 68. Adsorption isotherm for 0,0-diethyl-O-p-nitrophenol-
phosphorothioote.
-------
Resistant Materials from Solution
85
10
(J
o
o
o
O
O
0.02
0.04 0.06 0.08 0.10
1/Ceq
Figure 69. Langmuif plot for 2,4-dichlorophenoxyacetic acid.
-------
86
Adsorption of Biochemically
u
o
~
22
20
18
16
14
10
I
0.1
O
_L
0.2
0.3
0.4
1 7Ceq, liter/^mol
Figure 70. Langmuir plot for 2,4.5-trichtorophenoxyacetic acid.
-------
Resistant Materials from Solution
87
E
U
X
E
0.4
/Ceq, liter/fi
Figure 71. Longmuir plol for 2-(2,4,5-trichlotophenoxy)-propionic acid.
-------
88
Adsorption of Biochemically
10
i I i
1 1 1 1
0 0.02 0.04 0.06 0.08 0.10 0.12 0.14
l/Ceqj liter//unol
Figure 72. Longmuir plot for 2-sec-butyl-4.6-dinitrophenol.
0.16
GPO 825—048-^4
-------
Resistant Materials from Solution
89
0.02
0.04 0.06
l/Ceq, liter/(lmol
0.08
0.10
Figure 73- Longmuir plot for 2-cyclohexyt-4,6-dinitrophenol.
-------
90
u n -
u
~C
5 —
l/Ceq. ll,.r,>
Figure 74. Langrauit plot for 0,0-di*lhyl~Q-p-mtrophenof.
-------
EXAMINATION OF EFFLUENTS FROM
ACTIVATED CARBON COLUMNS
Studies on the removal of residual organic pollutants from con-
ventionally treated sewage by means of activated carbon have indicated
that there is initial leakage of organic material regardless of the extent
of carbon treatment.35 These studies at the Pittsburgh Chemical
Company involved conventional filtration of secondary sewage effluents
through beds of granular activated carbon 5 feet in depth, one to four of
these units being used in series. Leakage of organic material was
measured as chemical oxygen demand in the effluents and was found to
be substantially the same after passage through four units as after
filtration through a single unit.
The nature of the unadsorbed organic material was not known and
is difficult to determine because of the small total concentration and the
variety of possible components.
Nonetheless, obtaining some idea of the tyjes of substances not
effectively removed by carbon adsorption is of great importance. If
these are compounds that have great hygienic significance, then the
value of adsorption on carbon as a method for advanced waste treatment
is seriously diminished. On the other hand, if they are normal bacterial
metabolites that have escaped oxidation during conventional sewage treat-
ment, or similar compounds, then their presence in small quantity may
not represent a significant degradation in effluent quality.
It seemed possible that useful information on the nature of the
organic materials might be obtained after concentration of effluent by
vacuum distillation. Such treatment would change the effluent by loss of
volatile constituents, but these are not likely to be of importance. Other
changes that might occur or their effects are more difficult to assess,
but any thermal changes should be minimized by concentrating at as low
a temperature as possible.
Accordingly, samples of effluent from the carbon columns of the
Pittsburgh studies were concentrated by vacuum evaporation at the R.A.
Taft Sanitary Engineering Center, and two of the concentrates were then
shipped to Harvard for examination. Sample number one, dated February
7, 1964, had been concentrated by a factor of 36:1, and sample number
two, dated May 5, 1964, had been concentrated by a factor 60:1.
Because of limitations of time, not nearly as much was done with
these samples as would have been desirable. Besides very general
tests, only a few determinations of specific types of substances were
attempted, and the results of these had more significance in a negative
91
-------
92 Adsorption of Biochemically
sense than in a positive one. Nonetheless, it is believed that the
methods of approach have value and that more extensive and detailed
examination of similar concentrated effluents from carbon columns
should be undertaken by others.
Hie concentrated effluents as received contained considerable
sediment. Much of this was inorganic material such as CaCOs and
CaSO4 that had precipitated during the evaporation. Some organic
matter, however, had also coagulated or precipitated to such an extent
that it was removable by filtration through a membrane filter. This
may indicate that a part of the organic matter leaking through the col-
umns is colloidal and perhaps, therefore is either carbohydrate or
protein. It would be desirable to pass some unconcentrated effluent
through a Seitz or other molecular filter to see how much of the effluent
COD would be removed by such treatment. Presumably material of the
types indicated would not be particularly deleterious.
Note also that sample number one exhibited rapid growth of
microorganisms both during shipment and after receipt. Such growth
implies that much of the unadsorbed organic matter in the effluent was
readily oxidizable and that failure to be oxidized earlier was the result
of one or more of four factors:
1. The oxidizable materials were originally at too small a concen-
tration for ready biological oxidation, and evaporation increased the
concentration enough to support good growth of microorganisms.
2. Organic materials adsorbed by the carbon had been acting as
inhibitors for biological oxidation; once they were removed, biological
metabolism could proceed.
3. Time of contact in the sewage treatment plant was not great
enough for complete oxidation of metabolites.
4. Short-circuiting in the sewage treatment plant allowed carry-
over of some readily oxidized substances.
In any event, there is preliminary evidence in this simple observa-
tion that a substantial fraction of the organic material leaking through
the carbon column is not composed of so-called "persistent" compounds,
but may instead be carbohydrates, amino acids, or low-molecular-weight
carboxylic acid salts that would contribute little nuisance to receiving
waters and would tend to be removed by biological oxidation processes
in lakes or streams. Moreover, it means that further search for specific
unadsorbed materials should not be confined to "exotic" chemicals but
should include common metabolites and their intermediates.
Because of this experience, a portion of the sample number two
was acidified with HC1 prior to shipment from Cincinnati to suppress
bacterial action.
-------
Resistant Materials from Solution 93
General measurements made on the samples included organic
carbon and COD of both total and filtered material, potentiometric acid-
base titrations, and infra-red examinations of chloroform extractables.
Individual measurements of ABS, carbohydrates, and dicarboxylic acids
were attempted.
Total Organic Carbon
Organic carbon in the effluent samples was determined by the wet
oxidation method described by Weber and Morris.21 For each sample,
total organic carbon determinations were performed on the unfiltered
material so as to include insoluble organic matter, and on samples of
the effluent that had been subjected to filtration. The samples were first
acidified with sulfuric acid and then oxygen was bubbled through them for
10 minutes to remove CO2-
Results of the determinations, the average of two or three tests in
each instance, are shown in Table 12 for the samples as received in con-
centrated form and as computed for the original effluents.
Table 12. DETERMINATIONS OF ORGANIC CARBON IN EFFLUENTS
FROM PITTSBURGH CARBON FILTERS
Organic carbon Organic carbon
Sample mg/liter in concentrate mg/liter in effluent
No. 1, unfiltered
No. 1, filtered
No. 2, unfiltered
No. 2, filtered
92
66
143
50
2.6
1.8
2.4
0.83
Chemical Oxygen Demand
Determinations of COD were also carried out by the standard
technique on unfiltered and filtered samples except that, because of the
bacterial decomposition in sample No. 1 and other limitations on the
sample, a measurement on filtered concentrate was not possible for this
material. The results of these determinations are shown in Table 13.
Ordinarily one may expect the COD value for a waste to be about
2.5 to 3 times the organic carbon, the theoretical ratio for carbon in the
zero oxidation state as in carbohydrate being 32/12 or 2.67. Further-
more, the total weight of organic matter will range around 2.5 times the
organic carbon, 30/12 or 2.5 being the value for a simple carbohydrate.
-------
94 Adsorption of Biochemically
Table 13. DETERMINATIONS OF COD ON EFFLUENTS
FROM PITTSBURGH CARBON COLUMNS
COD COD
Sample mg/liter in concentrate mgAiter in effluent
No. 1, unfiltered
No. 2, unfiltered
No. 2, filtered
398
236
113
11.0
3.9
1.9
Thus, for organic material with average oxidation number near zero,
COD and total weight of organic matter will be approximately the same
numerically.
The two factors tend, however, to change in opposite directions
as the average oxidation number departs from zero. With butyric acid,
^4^10^2' w*tt* a mean oxidation number of -1.5 for the carbon, the
weight is 1.875 times the organic carbon, while the COD should be 3.67
times the organic carbon. Similarly, for glyoxylic acid, C2H4O3, with
net carbon oxidation number of +1, the weight is 3.17 times the organic
carbon, while the COD is only 2.0 times it.
Filtered sample No. 2 shows an expected ratio between COD and
organic carbon, 113/50 or 2.26. If this ratio is accurate, the mean
oxidation state of the carbon in the effluent is relatively high, about
+0.6. Such a result seems in accord with the fact that this effluent had
already been subjected to considerable oxidative treatment and is also
in accord with the intuitive feeling that oxygenated organic compounds
such as dicarboxylic acids, which will exist in wastes as ionized mater-
ials, will be less readily adsorbed by carbon than other types of organic
substances will.
The ratios of COD to organic carbon for the unfiltered effluents,
4.33 for sample No. 1, and 1.65 for sample No. 2, are completely out-
side the range of legitimate values for the type of material being
studied. Variations in sampling of the suspended matter or failure to
achieve complete COD for suspended organic matter may have caused
these anomalies.
Acid-Base Titrations
Because it seemed likely that carboxylic acid salts might be
largely responsible for the observed leakage of organic material
through carbon beds, it was felt that potentiometric acid-base titration
curves on the effluent concentrates would be useful in characterizing
the organic materials.
-------
Resistant Materials from Solution
95
Accordingly, these titrations were performed on portions of the
concentrates that had been filtered through membrane filters. Titrations
•were carried out on 25-milliliter portions of sample No. 1 and 50-milli-
liter portions of sample No. 2, pH readings with a glass electrode being
taken frequently throughout the titrations with tenth-normal HC1 or KOH.
The results of these titrations, shown in Figures 75 and 76, -were un-
satisfactory largely because the dominating effect of the large concentra-
tions of carbon dioxide, and bicarbonate alkalinity had been overlooked.
0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2
miNimoles H«0 ^ ~i —^. mHlimoles OH
Figure 75. Titration of effluent No. 1.25-ml sample.
I I I I I— I 1 1
1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8
tnillimoles H., 0
OH
1.0 1.2
Figure 76. Titration of effluent No. 2,50-ml sample.
-------
96
Adsorption of Biochemically
Additional titrations were performed, therefore, on portions of
filtered sample No. 2, which had been acidified and aerated to remove
CO2 before titration with tenth-normal hydroxide. Results of this work
are shown in Figure 77.
X
CL
1.2 1.6 2.0 2.4 2.R
2 —
0,0 0.4 0.8
77. Titrotioo of acidified effluent No- 1.50 mi-sample.
Again the results were unsatisfactory, because the excess of acid
used had been so great that neutralization of this strong acid tended to
obscure the finer points of the titration curve. There was some indica-
tion of weakly basic constituents in the form of the titration curve near
pH 10, possibly an indication of amino acids, but because precautions
to remove ammonia and phosphate had not been taken, the data must be
considered inconclusive.
-------
Resistant Materials from Solution 97
fci spite of the failure of these titrations, it is still believed that
valuable information can be obtained from similar titrations carried out
after more careful preparation of the initial solutions. Based on the ex-
perience obtained, preliminary treatment should include the following
steps: (1) Acidification of total unflitered sample to pH 4 + 0.5 followed
by bubbling ol nitrogen through the sample to remove CO2 fully; neutral-
ization with CC^-free KOH, possibly accompanied by addition of Ca++
or some other precipitating ion to remove PC«4- completely, followed by
filtration (not necessarily through a membrane filter; it may be desirable
to include colloidal material in the titration); (2) addition of CC^-free
KOH to pH 9.5 followed by bubbling of nitrogen to remove NHs fully. At
this point the sample should be separated into two portions, one to be
titrated carefully with dilute acid, the other with dilute base. As nearly
continuous records of pH as a function of acid or base added should be
obtained and compared with similar curves for the titration of CC>2-free
distilled water.
ABS Determinations
Each of the concentrates was analyzed for ABS concentration by
the methylene blue method as given in Standard Methods.36
Concentrations were determined from a calibration curve obtained
for a technical ABS, Ultrawet DS, obtained from the Atlantic Refining
Company, having a molecular weight of 372.19 The Ultrawet DS solutions
used for the calibration of the methylene-blue method were standardized
by ultraviolet spectrophotometry.1^
Sample No. 1 was determined to have an ABS concentration of 6.0
milligrams per liter, and sample No. 2 an ABS concentration of 7.0
milligrams per liter. Considerable difficulty was experienced with both
effluents during the ABS tests owing to emulsion formation.
These values for ABS concentration in the concentrates corres-
pond to 0.17 milligram per liter in the original effluent No. 1 and to
0.12 milligram per liter in effluent No. 2, values that accord well with
expectations for effluents from carbon columns and with similar analyses
performed at Cincinnati.
Miscellaneous Specific Tests
All the additional work carried out was on sample No. 2 because
of the bacterial degradation that occurred rapidly in sample No. 1.
Attempts were made to find chloroform-extractable materials in
the concentrate by treatment with this reagent followed by infra-red
examination of the dried chloroform extracts. Only negative results
-------
98 Adsorption of Biochemically
were obtained with the concentrates as received, and attempts to work
with more highly evaporated solutions were frustrated by the formation
of highly stable emulsions. Time did not permit further explorations
along this line.
Glucose and total carbohydrate determinations were carried out,
the former by the glucostat enzymatic method, the latter by the anthrone
method as described by Gaudy.37 Results for glucose were negative, as
might well have been expected, but carbohydrate was found at a concen-
tration of 7 milligrams per liter in the concentrate, corresponding to
0.12 milligram per liter in the initial effluent. One might guess that
this material would be dextrin-like, degraded glycogen or cellulose
perhaps.
Some experimentation was done with thin-layer chromatography on
the effluent samples. Qualitative indication of the presence of small
quantities of dicarboxylic acids was obtained, but methods for quantita-
tive estimation of the amount were not at hand. The technique does show
promise for examination of these effluents and probably should be pur-
sued more intensively in the future.
Unfortunately, circumstances prevented extension of these in-
vestigations; additional general information could have been obtained
readily. For example, determinations of Kjeldahl nitrogen and organic
phosphorus would have helped greatly in characterizing the residues,
and these can be performed easily. More sophisticated techniques such
as gas chromatography were proposed, but could not be used because of
delays in obtaining appropriate equipment.
Discussion
Perhaps the most significant result of this preliminary and partial
investigation into the nature of the organic material unadsorbed by
activated carbon columns is a negative one. No indications were found
of important amounts of the persistent, noxious organic chemicals for
which methods of removal are most important. What positive data
could be obtained make it seem likely that partially oxidized substrates
and metabolites of microorganisms make up a major portion of the
effluent organic matter. Although this conclusion is not firm, nothing
that has been found gives any reason to suppose that further develop-
ment of full-scale equipment for advanced waste treatment by activated
carbon should be drastically modified or abandoned.
Besides the obvious step of continued experimentation along the
lines followed in these tests to find individual compounds or classes of
compounds contributing to the organic leakage, another method of
approach is to test likely individual substances or classes of materials
to see how they behave on carbon columns. Dilute solutions of indivi-
dual carbohydrates, carboxylic acids, amino acids, and so forth might
-------
Resistant Materials from Solution 99
be passed through laboratory-scale columns to see how effectively
they are removed.
Note that in the small-column studies described earlier the
material that gave greatest difficulty with leakage was triethanolamine,
a small, hydroxylated molecule existing largely in salt-like cationic
form in neutral aqueous solutions. In contrast, all materials containing
phenolic groups seem to be effectively adsorbed.
Apparently, if additional removal of organic matter beyond that
now obtainable with reasonable depth of activated carbon is desired, one
should follow the activated carbon with a contrasting type of adsorbent.
One possibility seems to be to make use of porous or so-called "macro-
reticular" ion-exchange resins in this way, taking advantage of their
adsorptive abilities rather than specifically of their ion exchange prop-
erties. Organic "fouling" of ion-exchange resins is well known and
there is reason to suppose that the types of materials causing fouling
may be like those leaking through carbon columns.
It is recommended that experimentation be initiated to investigate
these possibilities.
-------
REFERENCES
1. Morris, J.C., and Weber, W.J., Jr., Adsorption of Biochemically Re-
sistant Materials from Solution. 1., Public Health Service Publica-
tion No. 999-WP-ll. Number 9 in the series of Advanced Waste
Treatment Research Publications.
2. Frantz, J.F., Design for Fluidization, I., Chemical Engineering, 69
161 (1962).
3. Wamsley, W.W., and Johanson, L.N., Fluidized Bed Heat Transfer,
Chemical Engineering Progress, 50, 7, 347 (1954).
4. Mickley, H.S., and Trilling, C.A., Heat Transfer Characteristics of
Fluidized Beds, Industrial and Engineering Chemistry, 41, 6, 1135
(1949).
5. Kettenring, K.N., Manderfield, E.L., and Smith, J.M., Heat and
Mass Transfer in Fluidized Systems, Chemical Engineering Pro-
gress, 46, 3, 139 (1950).
6. Leva, M., Weintraub, M., and Grummer, M., Heat Transmission
Through Fluidized Beds of Fine Particles, Chemical Engineering
Progress, 45, 9, 563 (1949).
7. Chu, J.C., Kalil, J., and Wetteroth, W.A., Mass Transfer in a
Fluidized Bed, Chemical Engineering Progress, 4^, 3, 141 (1953).
8. Resnick, W. and White, R.R., Mass Transfer in Systems of Gas and
Fluidized Solids, Chemical Engineering Progress, 45, 6, 377 (149).
9. Murphree, E.V., Gohr, E.J., and Kaulakis, A.F., The Fluid-Solid
Technique: Application in the Petroleum Industry, Jour. Institute
Petroleum, 33, 286 (1947).
10. Harang, R.A., Fluid Catalytic Hydroforming Units, Petroleum
Refiner, 32, 10, 135 (1953).
11. Matheson, G.L., Herbst, W.A., and Holt, P.H., Characteristics of
Fluid-Solid Systems, Industrial and Engineering Chemistry, 41, 6,
1099 (1949).
12. Lewis, W.K., Gilliland, E.R., and Bauer, W.C., Characteristics of
Fluidized Particles, Industrial and Engineering Chemistry, 41., 6,
1104 (1949).
13. Morse, R.D., Fluidization of Granular Solids, Industrial and Engi-
neering Chemistry, 41, 6, 1117 (1949).
14. Leva, M., and Grummer, M., Correlation of Solids Turnover in
Fluidized Systems, Chemical Engineering Progress, 48, 6, 307
(1952).
101
-------
102 Adsorption of Biochemically
15. Lewis, E.W., and Bowerman, E.W., Fluidization of Solid Particles
in Liquids, Chemical Engineering Progress, 48, 12, 603 (1952).
16. Etherington, L.D., Fritz, R.J., Nicholson, E.W., and Scheeline,
H.W., Fluid Char Adsorption Process, Chemical Engineering Pro-
gress, 52, 7, 274 (1956).
17. McCune, L.K., and Wilhelm, R.H., Mass and Momentum Transfer in
a Solid-Liquid System, Industrial and Engineering Chemistry, 41 6,
1124(1949).
18. Hobson, M., and Thodos, G., Mass Transfer in the Flow of Liquids
Through Granular Solids, Chemical Engineering Progress, 45, 8,
517 (1949).
19. Weber, W.J., Jr., Morris, J.C., and Stumm, W., Determination of
Alkylbenzenesulfonates by Ultraviolet Spectrophotometry, Analytical
Chemistry, 34, 13, 1844 (1962).
20. Rosen, M.J., and Goldsmith, H.A., Systematic Analysis of Surface-
Active Agents, Interscience Publishers, Inc., New York, N.Y. (1960).
21. Weber, W.J., Jr.,and Morris, J.C., Determination of Carbon
in Waste Waters by High-Temperature Wet Oxidation, Journal of
the Water Pollution Control Federation, 36, 573-86 (1964).
22. Morris, J.C., and Weber, W.J., Jr., Removal of Biologically-Resist-
ant Pollutants from Waste Waters by Adsorption, Proc. First Inter-
national Conf. on Water Pollution Research, Pergamon Press Ltd.,
Oxford, England, (1962) (in press).
23. Weber, W.J., Jr., and Morris, J.C., Adsorption in Heterogeneous
Aqueous Systems, Journal of the American Water Works Associa-
tion, 56, 447-56 (1964).
24. Weber, W.J., Jr., and Morris, J.C., Equilibria and Capacities for
Adsorption on Carbon, Journal of the Sanitary Engineering Division,
Proceedings of the American Society of Civil Engineers, 90, SA3,
79-107 (1964).
25. Edeskuty, F.J., and Amundson, N.R., Effect of Intraparticle Diffusion:
L - Agitated Non-Flow Adsorption Systems, Industrial and Engineer-
ing Chemistry, 44, 1698 (1952).
26. Rosen, J.B., Kinetics of a Fixed Bed System for Solid Diffusion into
Spherical Particles, Jour, of Chemical Physics, ^0, 3, 387 (1952).
27. Kasten, P.R., Lapidus, L., and Amundson, N.R., Mathematics of
Adsorption in Beds. V. Effect of Intraparticle Diffusion in Flow
Systems in Fixed Beds, Jour, of Physical Chemistry, 56, 6, 683
(1952). ~~
-------
Resistant Materials from Solution 103
28. Eshbach, O.W., Handbook of Engineering Fundamentals, Second
Edition, p. 1-146, John Wiley & Sons, Inc., New York (1952).
29. Klinkenberg, A., and Mooy, H.H., Dimensionless Groups in Fluid
Friction, Heat, and Material Transfer, Chemical Engineering Pro-
gress, 44, 1, 17 (1948).
30. Middleton, F.M., Report on the Recovery of Orthonitrochlorobenzene
from the Mississippi River, Report from Robert A. Taft Sanitary
Engineering Center, Cincinnati, Ohio, June 1959.
31. Faust, S.D., and Aly, O.M., Water Pollution by Organic Pesticides,
Journal of the American Water Works Association, 56, 3, March
1964, pp. 267-79.
32. Weber, W.J., Jr., and Morris, J.C., Kinetics of Adsorption on Carbon
from Solution, Proceedings of the American Society of Civil Engineers,
Journal of the Sanitary Engineering Division, 89, SA 2, Proc. Paper
3483, April 1963, pp. 31-59.
33. Crank, J., The Mathematics of Diffusion, Oxford at the Clarendon
Press, London, England, 1956.
34. Langmuir, I., The Adsorption of Gases on Plane Surfaces of Glass,
Mica, and Platinum, Journal of the American Chemical Society,
40, 1918, p. 1361.
35. Joyce, R.S., and Valentine, A.S., Activated Carbon in Waste Water
Treatment - Adsorption and Thermal Reactivation. Paper presented
at Seminar on Advanced Waste Treatment Research, Cincinnati, Ohio,
May 1962; see also Middleton, F.M., Advances in Water Pollution
Research, Proceedings of the First International Conference, London,
September 1962, Vol. 2, pp. 258-265; Pergamon Press (1964).
36. Standard Methods for the Examination of Water and Wastewater,
Eleventh Edition, American Public Health Association, Inc.,
New York (1960).
37. Gaudy, A.F., Colorimetric Determination of Protein and Carbohy-
drate, Industrial Water and Wastes, 7, 17-22 (1962).
GPO 825—048—5
-------
APPENDIX A
NOTATIONS
A = total external surface area of adsorbent, [L] ^
b = constant of Langmuir adsorption isotherm related to the
energy of adsorption
- effective transfer area per unit mass adsorbent, [L]
C = mass concentration of solute in effluent [M] [M]~l
Co = mass concentration of solute in influent [M]
Cs = mass concentration of solute in adsorbent [M]
AC = concentration potential across liquid film [M] [M]~l
D = diameter of adsorbent particle, [L]
£ =kTamMaQmT/W OcramMa+Qm)
h = height measured from base of expanded medium, [L]
he = total height of expanded medium, [L]
k & k-p = mass transfer coefficients, [M] [Lj"2 [l]~VA c
m & Ma = total mass of adsorbent, [M]
M = dynamic viscosity of solution, [M] [L]~l [T]~*
= modified Reynolds number, dimensionless
= partition factor = {-/£— | > dimensionless
V cs /equil.
= mass flow rate, [M] [T]"1
r = rate of transfer, [M] [T]'1
p = mass density of solution, [M] [L]"3
t = time, [T]
v = superficial face velocity, [L] [T]'1
W = total weight of carbon, [M]
105
-------
106
= quantity of solute adsorbed per unit weight adsorbent,
[M]
Xoo = capacity of adsorbent at equilibrium with a solution of
concentration CQ, [M] [M]~l
Xm = moles of solute in forming a complete monolayer on the
carbon surface, [M] [M]"1
-------
APPENDIX B
ADVANCED WASTE
TREATMENT RESEARCH PUBLICATIONS
Although over-all guidance of AWTR research projects is provided
by the AWTR Program, the results obtained, opinions expressed, and con-
clusions reached by the contractors are theirs and are not necessarily
endorsed by the Federal Water Pollution Control Administration. Except
for editorial review and abridgments as necessary, these reports appear
as received from the contractors.
AWTR-1 Summary Report, June 1960 - December 1961
AWTR-2 Preliminary Appraisal of Advanced Waste
Treatment Processes
AWTR-3 Ultimate Disposal of Advanced-Treatment
Waste
Part 1. Wet Oxidation
Part 2. Incineration
AWTR-4 Waste-Water Renovation
Part 1. A Design Study of Freezing
and Gas Hydrate Formation
Part 2. Feasibility Tests of
Freezing
AWTR-5 Contaminant Removal from Sewage Plant
Effluents by Foaming
AWTR-6 Cost of Purifying Municipal Waste Waters
by Distillation
AWTR-7 Advanced Waste Treatment by Distillation
AWTR-8 Ultimate Disposal of Advanced-Treatment
Waste
Part 1. Injection
Report
Number
W62-9
W62-24
999-WP-3
999-WP-4
999-WP-5
999-WP-6
999-WP-9
999-WP-10
107
-------
108
Report
Number
Part 2. Placement in Underground
Cavities
Part 3. Spreading
AWTR-9 Adsorption of Biochemically Resistant
Materials from Solution. 1. 999-WP-ll
AWTR-10 Feasibility of Granular, Activated -
Carbon Adsorption for Waste-
Water Renovation 999-WP-12
AWTR-11 Evaluation of the Use of Activated
Carbon and Chemical Regenerants in
Treatment of Waste Water 999-WP-13
AWTR-12 Evaluation of Various Adsorbents and
Coagulants for Waste-Water Renovation 999-WP-14
AWTR-13 Electrochemical Treatment of
Munic ipal Waste Water 999- WP-19
AWTR-14 Summary Report: The Advanced Waste
Treatment Research Program.
January 1962 through June 1964 999-WP-24
AWTR-15 Feasibility of Granular, Activated Carbon
Adsorption for Waste-Water
Renovation. 2. 999-WP-28
AWTR-16 Adsorption of Biochemically Resistant
Materials from Solution. 2. 999-WP-33
GPO
-------
BIBLIOGRAPHIC: Morris, J. Carrell, and Walter J.
Weber, Jr. Adsorption of biochemically resistant
materials from solution. 2. PHS Publ, No
999-WP-33. 1966. 108pp.
ABSTRACT: Earlier studies (reported in PHS Publi-
cation No. 999-WP-ll - AWTR-9) showed that
activated carbon for waste water renovation could
best be used in continuous-flow columns. Such
techniques should result in an adsorptive capacity
of greater than 10 percent. Results on studies of
adsorption of organics from single- and multi-
component systems in fluidized carbon are reported
herein. The absorbability of organic pesticides on
activated carbon was investigated in some detail.
Studies were undertaken to characterize those
types of organic pollutants tbat are not adsorbed
on activated carbon.
BIBLIOGRAPHIC: Morris, J. Carrell, and Walter J.
Weber, Jr. Adsorption of biochemically resistant
materials from solution. 2. PHS Publ. No.
999-WP-33. 1966. 108pp.
ABSTRACT: Earlier studies (reported in PHS Publi-
cation No. 999-WP-U - AWTR-9) showed that
activated carbon for waste water renovation could
best be used in continuous-flow columns. Such
techniques should result in an adsorptive capacity
of greater than 10 percent. Results on studies of
adsorption of organics from single- and multi-
component systems in fluidized carbon are reported
herein. The absorbability of organic pesticides on
activated carbon was investigated in some detail.
Studies were undertaken to characterize those
types of organic pollutants that are not adsorbed
on activated carbon.
BIBLIOGRAPHIC: Morris, J. Carrell, and Walter J.
Weber, Jr. Adsorption of biochemically resistant
materials from solution, 2. PHS Publ. No.
999-WP-33. 1966. 108pp.
ABSTRACT: Earlier studies (reported in PHS Publi-
cation No. 999-WP-ll - AWTR-9) showed that
activated carbon for waste water renovation could
best be used in continuous-flow columns. Such
techniques should result in an adsorptive capacity
of greater than 10 percent. Results on studies of
adsorption of organics from single- and multi-
component systems in fluidized carbon are reported
herein. The absorbability of organic pesticides on
activated carbon was investigated in some detail.
Studies were undertaken to characterize those
types of organic pollutants that are not adsorbed
on activated carbon.
ACCESSION NO.
KEY WORDS:
Harvard Rept 2
Advanced Waste
Treatment
Adsorption
Waste Water
Renovation
Activated Carbon
Refractory Materials
Equilibrium
Kinetics
Regeneration
Pesticides
ACCESSION NO.
KEY WORDS:
Harvard Rept 2
Advanced Waste
Treatment
Adsorption
Waste Water
Renovation
Activated Carbon
Refractory Materials
Equilibrium
Kinetics
Regeneration
Pesticides
ACCESSION NO.
KEY WORDS:
Harvard Rept 2
Advanced Waste
Treatment
Adsorption
Waste Water
Renovation
Activated Carbon
Refractory Materials
Equilibrium
Kinetics
Regeneration
Pesticides
------- |