PB84-199579
SYNTHETIC RESIN ADSORBENTS IN TREATMENT OF INDUSTRIAL WASTE STREAMS
U.S. Environmental Protection Agency
ADA, OK
May 84
U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
-------
EPA-600/2-84-105
May 1984
SYNTHETIC RESIN ADSORBENTS IN TREATMENT OF
INDUSTRIAL WASTE STREAMS
by
Linda S. Banner
Rohm and Haas Company
Research Laboratories
Spring House, Pennsylvania 19477
Copyright 1983 - Rohm and Haas Company
Cooperative Agreement
Number CR807315
Project Officer
Thomas E. Short
Robert S. Kerr Environmental Research Laboratory
Environmental Protection Agency
Ada, Oklahoma 74820
ROBERT S. KERR ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ADA, OKLAHOMA 74820
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-84-105
3. RECIPIENT'S ACCESSION-NO.
P38U-199579
4. TITLE AND SUBTITLE
Synthetic Resin Adsorbents in Treatment of
Industrial Waste Streams
5. REPORT DATE
May 1984
J6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Linda S. Benner
|S. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
1BB6K
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Rohm & Haas Company
Research Laboratories
Spring House, PA 19477
11. CONTRACT/GRANT NO.
CR807315
12. SPONSORING AGENCY NAME AND ADDRESS
Robt. S. Kerr Environmental Research Laboratory
U.S. Environmental Protection Agency
P. 0. Box 1198
Ada, OK 74820
13. TYPE OF REPORT AND PERIOD COVERED
Final 6/1/80 - 5/31/82
14. SPONSORING AGENCY CODE
EPA/600/15
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The use of synthetic polymeric adsorbents for removal of organic pollutants
from industrial waste streams is a viable alternative to more common treatment
methods such as carbon adsorption. However, resin technology is not widely
practiced due to the difficulty of selecting the appropriate synthetic
adsorbent/regenerant combination for a particular application. This research
program was undertaken to develop a simple, reliable laboratory test to assess
the feasibility of using synthetic resin adsorbents for treatment of a given .
waste water stream.
The new test method facilitates rapid screening of a large number of
adsorbent/regnerant pairs. The test consists of exposing small bags of
adsorbents in the actual waste stream for predetermined periods of time, followed
by regenerating the exposed adsorbents. The batch portion of the test measures
regenerated saturation capacity of any adsorbent/regenerant pair with a single
analysis. The rate portion measures regenerated capacity as a function of time.
Capacity data obtained by the test method were compared and correlated
with data obtained by conventional isotherm tests and column experiments. In
general, the batch test performs better than the isotherm test in predicting
column saturation capacity of polymeric adsorbents.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS |c. COSATI Field/Group
Adsorption
Waste Water
Waste Treatment
Toluene
Nitrobenzene
2 nitrophenol
2,4-dinitrotoluene
Tetrachloroebhylene
1,2 dichloropropane
acetone methanol
68D
18. DISTRIBUTION STATEMENT
Release to public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
20. SECURITY CLASS (This page/
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
-------
DISCLAIMER
Although the research -described in this article has been funded wholly
or in part by the United States Environmental Protection Agency under
assistance agreement no. CR-807315 to Rohm & Haas Company, it has not been
subjected to the Agency's peer and administrative review and therefore may
not necessarily reflect the views of the Agency and no official endorse-
ment should be inferred.
-ii-
-------
FOREWORD
The Environmental Protection Agency is charged by Congress to
protect the Nation's land, air and water systems. Under a mandate of
national environmental laws focused on air and water quality, solid waste
management and the control of toxic substances, pesticides, noise, and
radiation, the Agency strives to formulate and implement actions which
lead to a compatible balance between.-human activities and the ability of
natural systems to support and nurture life. In partial response to
these mandates, the Robert S. Kerr Environmental Research Laboratory,
Ada, Oklahoma, is charged with the mission to manage research programs to
investigate the nature, transport, fate, and management of pollutants in
ground water and to develop and demonstrate technologies for treating
wastewaters with soils and other natural systems for controlling pollu-
tion, from irrigated crop and animal production agricultural activities;
for developing and demonstrating cost-effective land treatment systems for
the environmentally safe disposal of solid and hazardous waste.
EPA, as provided for in the Clean Water Act, has the responsibility
to promulgate regulations to protect streams. These regulations must be
developed based on consideration of tehcnology for waste treatment. Syn-
thetic resin adsorbents are known to be excellent for the treatment of
specific organic pollutants. However, there has been no reliable, simple
way to assess the practicality and cost of wastewater treatment with these
adsorbents. This project has developed a test procedure which provides the
equilibrium and kinetic data needed for the evaluation of the treatment of
industrial wastewater by synthetic resin adsorbents. Thus, the comparison
of the results of more conventional treatment options such as activated
carbon or biological methods with resins is now possible.
Clinton W. Hall
Director
Robert S. Kerr Enviornmental
Research Laboratory
-------
ABSTRACT
The use of synthetic polymeric adsorbents for removal of organic
pollutants from industrial waste streams is a viable alternative to more
common treatment methods such as carbon adsorption or steam stripping.
However, resin technology is not widely practiced due to the difficulty of
selecting the appropriate synthetic adsorbent/regenerant combination for a
particular application. This research program was undertaken to develop a
simple, reliable laboratory test to assess the feasibility of using
synthetic resin adsorbents for treatment of a given waste water stream.
Since a major advantage of resin technology is ease of resin regeneration
and possible recovery of organic pollutants for recycle, the test was
designed to measure regenerated adsorption capacities directly.
The new test method, the batch/rate test, facilitates rapid screening
of a large number of adsorbent/regenerant pairs. The test consists of
exposing small bags of adsorbents in the actual waste stream for prede-
termined periods of time, followed by regenerating the exposed adsor-
bents. The batch portion of the test measures regenerated saturation
(equilibrium) capacity of any adsorbent/regenerant pair with a single
analysis. The rate portion measures regenerated capacity as a function of
time.
Adsorbents used for the test development included several Rohm
and Haas resins, several Mitsubishi resins, and several Calgon activated
carbons. Adsorbent performance was evaluated in several single component
synthetic streams, a multicomponent synthetic stream, and a multicomponent
real stream. These streams contained one or more of the following or-
ganic pollutants: toluene, nitrobenzene, 2-nitrophenol, 2,4-dinitro-
toluene, tetrachloroethylene, 1,2-dichloropropane. Regenerants evaluated
included acetone, methanol, and/or steam.
Capacity data obtained by the batch/rate test method were compared
and correlated with data obtained by conventional isotherm tests and
column loading/regeneration experiments. In general, the batch test
performs better than the isotherm test in predicting column saturation
capacity of polymeric adsorbents. Because the test method is designed to
measure solvent regenerated capacity, it is not suitable for evaluation of
activated carbons which are usually regenerated in furnaces at very high
temperatures. The batch/rate test provides the equilibrium and kinetic
data needed to predict column performance under a variety of operating
conditions. A simple theoretical model, developed for use with batch/rate
data, satisfactorily predicts column performance to within 10% of experi-
mental results during the first half of column breakthrough. Alternately,
a simple empirical relationship between batch data and initial column
breakthrough can be used to estimate initial leakage to within 15% of
-iv-
-------
experimental results. Thus, with a single measurement, the batch test,
column capacity to 10$ leakage can be roughly estimated to be equal to
70$ of the batch capacity.
This report is submitted in fulfillment of Contract No. 807315-0-10
under the sponsorship of the U.S. Environmental Protection Agency. This
report covers the period June 1, 1980 to May 31, 1982.
-v-
-------
TABLE OF CONTENTS
Page
Abstract iii
List of Tables vii
List of Figures x
Acknowledgement . xiv
1. Introduction 1
. Purpose and Approach 1
. Adsorption Phenomenon 2
2. Conclusions • 4
3- Recommendations 5
4. Materials and Methods 6
. Adsorption Conditioning and Handling Techniques 6
. Analytical Methods 6
. Preparation of Synthetic Waste Streams 8
5. Adsorption Test Methods 9
. Adsorption Isotherms 9
. Batch/Rate Tests 9
. Column Experiments 10
6. Results and Discussion 12
. Adsorption Isotherm Tests ' 12
. Batch/Rate Tests 30
. Column Experiments 56
. Summary of Adsorption Data 86
7. Predictions of Column Performance Using Batch/Rate Data 95
. Predictions Based on the Performance Model 95
. . Applications of the Performance Model 95
. Correlation of Predicted and Observed Column Performance 105
. Estimation of Initial Column Breakthrough 105
from Batch Data
-vi-
-------
Page
References 109
Appendices
A. Abbreviations
B. Adsorbent Conditioning
C. Theoretical Approach to Column Performance Modeling 113
Using Batch/Rate Data
D. Symbols Used in Column Performance Modeling 120
E. Glen Cove Pilot Plant Data 122
-vii-
-------
LIST OF TABLES
Page
1. Physical Properties of Adsorbents ' 7
2. Adsorption Isotherm Testing with 2-Nitrophenol ig
3- Adsorption Capacities for 2-Nitrophenol 17
Measured by Isotherm Tests
4. Adsorption Isotherm Testing with Nitrobenzene 21
5. Adsorption Capacities for Nitrobenzene Measured 22
by Isotherm Tests
6. Adsorption Isotherm Testing with 1,2-Dichloropropane 26
7. Adsorption Capacities for 1,2-Dichloropropane 26
Measured by Isotherm Tests
8. Adsorption Isotherm Testing with 2,4-Dinitrotoluene 29
9. Adsorption Capacities for 2,4-Dinitrotoluene Measured 29
by Isotherm Tests
10. Adsorption Capacities for 2-Nitrophenol Measured 32
by Batch Tests
11. Rate Test Results for Adsorption of 2-Nitrophenol by 33
XAD-4
12. Adsorption Capacities for Tetrachloroethylene Measured 35
by Batch Tests
13. Rate Test Results for Adsorption of Tetrachloroethylene 36
by XAD-4
14. Adsorption Capacities for Nitrobenzene Measured by ••'" 38
Batch Tests
15. Rate Test Results for Adsorption of Nitrobenzene by 38
XAD-4
16. Adsorption Capacities for 1,2-Dichloropropane Measured 41
by Batch Tests
-viii-
-------
Page
17. Rate Test Results-for Adsorption of 1,2-Dichloropropane . 41
by XAD-4
18. Adsorption Capacities for 2,4-Dinitrotoluene 43
Measured by Batch Tests
19. Rate Test Results for Adsorption of 44
2,4-Dinitrotoluene by XAD-4
20. Rate Test Results for Adsorption of 2,4-Dinitrotoluene '44
by XE-340
21. Adsorption Capacities for Components of the 47
Synthetic, Nitroaromatic Stream Measured by
Batch Tests
22. Rate Test Results for Adsorption of the Components 47
of the Synthetic, Nitroaromatic Stream by HP-20
23. Rate Test Results for Adsorption of the Components of 49
the Synthetic, Nitroaromatic Stream by XAD-4
24. Rate Test Results for Adsorption of the Components of 49
the Synthetic, Nitroaromatic Stream by XE-340
25. Comparison of Batch/Rate Data From Studies of 52
Multicomponent and Single Component Nitroaromatic
Streams
26. Rate Test Results for Adsorption of the Components 54
of the Glen Cove Stream by XE-340
27. Summary of Saturation Capacities Measured by Isotherms, 90
Batch Tests, and Column Experiments in Single Component
Waste Streams (mg/g)
28. Summary of Saturation Capacities Measured by Batch Tests 91
and Column Experiments in the Synthetic, Nitroaromatic
Stream (mg/g)
29. Summary of Column Capacity Data for Single Component 93
Waste Streams
30. Summary of Column Capacity Data for the Synthetic, 94
Nitroaromatic Waste Stream
31. Experimental Parameters for Predicting 97
1,2-Dichloropropane Concentration History
on XAD-4 Columns
32. Calculation of D/T From the Rate of Adsorption of 98
1,2-Dichloropropane by XAD-4
-ix-
-------
Page
33- Experimental Parameters for Predicting Nitroaromatics 102
Concentration History on XAD-4
34. Calculation of D/T' From the Rate of Adsorption of 103
Nitroaromatics by XAD-4
35. Summary of Predicted and Observed Column Performance 107
of Amberlite XAD-4
36. Correlation of Batch Data and Initial Leakage in '108
Column Experiments
E-1 Glen Cove Adsorption Study 123
-x-
-------
LIST OF FIGURES
Page
1. Adsorption Isotherms for 2-Nitrophenol (pH 5) 14
on Amberlite Adsorbents
2. Adsorption Isotherms for 2-Nitrophenol (pH 8) 15
on Amberlite Adsorbents
3. Adsorption Isotherms for Nitrobenzene on Amberlite 18
Adsorbents
U. Adsorption Isotherms for Nitrobenzene on Diaion 19
Adsorbents
5. Adsorption Isotherms for Nitrobenzene on Carbonaceous 20
Adsorbents
6. Adsorption Isotherms for 1,2-Dichloropropane on 24
Polymeric Adsorbents
7. Adsorption Isotherms for 1,2-Dichloropropane on 25
Carbonaceous Adsorbents
8. Adsorption Isotherms for 2,4-Dinitrotoluene on 27
Polymeric Adsorbents
9- Adsorption Isotherms for 2,4-Dinitrotoluene on 28
Carbonaceous Adsorbents
10. Rate of Adsorption of 2-Nitrophenol by Amberlite XAD-4 34
11. Rate of Adsorption of Tetrachloroethylene by 37
Amberlite XAD-U
12. Rate of Adsorption of Nitrobenzene by Amberlite XAD-4 39
13. Rate of Adsorption of 1,2-Dichloropropane by 42
Amberlite XAD-4
14. Rate of Adsorption of 2,4-Dinitrotoluene 45
15. Rate of Adsorption of Components of the Nitroaromatic 48
Stream by Diaion HP-20
-xi-
-------
Page
16. Rate of Adsorption of Components of the 50
Nitroaromatic Stream by Amberlite XAD-4
17. Rate of Adsorption of Components of the 51
Nitroaromatic Stream by Ambersorb XE-340
18. Rate of Adsorption of Components of the Glen 55
Cove Stream by Ambersorb XE-340
19. Column Adsorption of 2-Nitrophenol 58
(pH 5) by Amberlite XAD-4
20. Column Regeneration of 2-Nitrophenol (pH 5) 59
Loaded Amberlite XAD-4 with Methanol
21. Column Adsorption of 2-Nitrophenol 60
(pH 8) by Amberlite XAD-4
22. Column Regeneration of 2-Nitrophenol (pH 8) 61
Loaded Amberlite XAD-4
23. Column Adsorption of 2-Nitrophenol by 62
Amberlite XAD-8
24. Column Regeneration of 2-Nitrophenol Loaded 63
Amberlite XAD-8
25. Column Adsorption of Nitrobenzene by 64
Amberlite XAD-4
26. Column Regeneration of Nitrobenzene Loaded 65
Amberlite XAD-4
27. Column Adsorption of Nitrobenzene by 66
Selected Adsorbents
28. Column Regeneration of Nitrobenzene Loaded 67
Adsorbents
29. Column Adsorption of 1,2-Dichloropropane 69
by Amberlite XAD-4
30. Column Regeneration of 1,2-Dichloropropane 70
Loaded Amberlite XAD-4
31. Column Adsorption of 1,2-Dichloropropane 71
by Selected Adsorbents
32. Column Regeneration of 1,2-Dichloropropane Loaded 72
Adsorbents
33- Column Adsorption of 2,4-Dinitrotoluene 73
by Amberlite XAD-4
-xii-
-------
Page
34. Column Regeneration of 2,4-Dinitrotoluene Loaded 74
Amberlite XAD-4
35. Column Adsorption of 2,4-Dinitrotoluene 75
by Ambersorb XE-340
36. Column Regeneration of 2,4-Dinitrotoluene . 75
Loaded Ambersorb XE-340
37. Column Adsorption of Nitroaromatics 78
by Amberlite XAD-4 (8 BV/hr)
38. Column Adsorption of Nitroaromatics 79
by Amberlite XAD-4 (16 BV/hr)
39- Column Regeneration of Nitroaromatics 80
Loaded Amberlite XAD-4 (8 BV/hr)
40. Column Regeneration of Nitroaromatics 81
Loaded Amberlite XAD-4 (16 BV/hr)
41. Column Adsorption of Nitroaromatics 82
by Diaion HP-20 (8 BV/hr)
42. Column Adsorption of Nitroaromatics 83
by Diaion HP-20 (16 BV/hr)
43. Column Regeneration of Nitroaromatics 84
Loaded Diaion HP-20 (8 BV/hr)
44. Column Regeneration of Nitroaromatics 85
Loaded Diaion HP-20 (16 BV/hr)
45. Column Adsorption of Nitroaromatics 88
by Ambersorb XE-340
46. Column Regeneration of Nitroaromatics Loaded 89
Ambersorb XE-340
47. Correlation of Predicted and Observed Column 101
Adsorption of 1,2-Dichloropropane by Amberlite XAD-4
48. Correlation of Predicted and Observed Column 104
Adsorption of Nitroaromatics by Amberlite XAD-4
C-1 Simultaneous Diffusion and Linear Adsorption into
a Porous Sphere
-xiii-
-------
Page
C-2 Generalized Dimensionless Solutions for Heat- ' 119
Transfer and Other Mass-Transfer Processes
E-1 Cis-1,2-Dichloroethylene Removal by Ambersorb XE-3^0 125
at U gpm/ft3-Cycle #1-Glen Cove, NY
E-2 Trichloroethylene Removal by Amber-sorb XE-340 at 126
U gpm/ft3-Cycle #1-Glen Cove, NY
E-3 Tetrachloroethylene Removal by Ambersorb XE-3^0 at 127
4 gpm/ft3-Cycle #1-Glen Cove, NY.
-xiv-
-------
ACKNOWLEDGEMENT
The author gratefully acknowledges the assistance of Dr. James
Neely, Mr. Eric Isacoff and Mr. John Maikner in development of the
batch/rate test protocol. The author also thanks Dr. Sabah Dabby
for the use of his column performance model and acknowledges his
assistance in modifying it for use with the new batch/rate test.
-------
SECTION 1
INTRODUCTION
PURPOSE AND APPROACH
The objective of this project was to develop a simple, reliable
laboratory test to assess the applicability of synthetic resin adsorbents
for removal of organic pollutants from industrial waste streams. The test
has been designated the batch/rate test. Since a major advantage of resin
adsorption technology in waste water treatment is ease of resin regenera-
tion and possible recovery of pollutant materials, the test was designed to
measure regenerated adsorption capacities directly. The test consists of
exposing small bags of adsorbents in a waste stream for predetermined
periods of time followed by regenerating the exposed adsorbents to obtain
regenerated capacities. Typical regenerants include organic solvents
(e.g. methanol, ethanol, acetone, toluene), steam, and dilute caustic.
Usually the choice of regenerant is determined by cost, safety factors, and
process compatibility as well as success in regeneration. The batch
portion of the test measures regenerated saturation (equilibrium) capacity
for any adsorbent/regenerant pair with a single analysis. Batch capacity
corresponds to maximum column saturation capacity'. The rate portion of the
test measures regenerated capacity as a function of exposure time in the
waste stream. From the batch test alone, a large number of adsorbent/
regenerant pairs can be rapidly evaluated to determine if any resins
are worth additional testing in the waste stream. The optimum adsorbent/
regenerant pair(s) can then be used in rate testing to generate the addi-
tional data needed to calculate column performance under a variety of
column operating conditions. Treatment costs can then be estimated.
The adsorbents available for evaluation during batch/rate test
develop ment included several Rohm and Haas Amberlite* and Ambersorb*
adsorbents, several Mitsubishi Diaion** adsorbents, and several Calgon
Filtrasorb*** granular activated carbons. During the initial stages of
test development, adsorbent performance was evaluated in synthetic single
component waste streams by standard isotherm tests, by column loading/
regeneration.experiments, and by batch/rate tests. After batch/rate test
optimization, adsorbent performance was evaluated by batch/rate testing and
column studies. Several more single component waste streams, a synthetic
multi-component waste stream, and a real contaminated groundwater source
were used.
* Amberlite and Ambersorb are registered trademarks of the Rohm
and Haas Company.
** Diaion is a registered trademark of Mitsubishi Chemical
Industries, Ltd.
««* Filtrasorb is a registered trademark of the Calgon Corporation.
-1-
-------
ADSORPTION PHENOMENOM
A brief introduction to the factors which influence adsorption
is presented 'here. More detailed information can be found in the litera-
ture[1-8].
The phenomenon of physical adsoprtion onto solids is due to electro-
static and covalent interactions between the sorbate and the solid sur-
face. Many types of forces are involved including van der Waals forces,
dipole-dipole interactions, and hydrogen and hydrophobia bonding. It'is
not possible to accurately predict which species will be adsorbed well by
a given adsorbent. However, a useful adsorption concept is that hydro-
phobic or non-polar molecules or portions of molecules are attracted to
hydrophobia surfaces, and that hydrophilic or polar molecules are at-
tracted to hydrophilic surfaces. In adsorption of organics from aqueous
streams, the portions of the organic molecules with less affinity for
water, i.e. the hydrophobia portions, are preferentially adsorbed onto the
hydrophobia surfaces of the adsorbent while the hydrophilic portions
of the molecules remain oriented in the aqueous phase. Since synthetic
adsorbents are available over a wide range of surface polarities, they
have the potential for efficiently and selectively adsorbing many types of
organic compounds.
The capacity of an adsorbent for an adsorbed species is highly
influenced by its surface area and pore structure. In general, the
greater the surface area, the greater the amount of adsorbate which
can theoretically be adsorbed. However, the adsorbent's pore size
range and distribution pattern must be sufficient to permit adsorbates
to migrate through the pores to the adsorbing surfaces. Surface area and
pore size are inversely related. Thus pores must be large enough to admit
molecules to the adsorbent interior, yet as small as possible to maximize
available surface areas. It is reasonable to expect that for treatment of
a complex waste stream, several adsorbents of differing surface polarity,
surface area, and pore structure might be needed to remove the organic
contaminants from the aqueous stream.
A principal difference between synthetic resin adsorbents and
granular activated carbons is the lower binding affinities of the resins
for sorbates. Thus, resins can be readily regenerated in situ with
organic solvent or steam whereas carbons are usually regenerated in
furnaces at very high temperature (e.g. 800°C).
The adsorption of organic pollutants from aqueous streams is influ-
enced by many factors other than the physical and chemical nature of the
adsorbent. Some of the more important factors to be considered include
the solubility, polarity, and concentration of the sorbate; the stream
temperature and pH; the presence of competing sorbates; the type and
concentration of dissolved electrolytes; and the quantity and size of
suspended particulates. An increase in solubility, ionization, or tem-
perature generally results in a decrease in adsorption.
-2-
-------
Conversely, the presence of added electrolytes usually improves adsorp-
tion. However, the presence of competing organics to be adsorbed may
enhance, interfere with, or act independently in the adsorption process.
Since so many factors influence adsorption, extrapolation of capacity data
obtained in either synthetic single component streams or in synthetic
multicomponent streams may not accurately reflect adsorption of selected
sorbates from a real waste stream.
-3-
-------
SECTION 2
CONCLUSIONS
The experimental protocol developed for measuring the adsorptive
capacity of synthetic resin adsorbents, i.e. the batch/rate test, provides
a simple reliable procedure for assessing the practicality of using
synthetic adsorbents for toxic waste water treatment. In resin applica-
tions, the test method is superior to conventional isotherm techniques
developed for carbon adsorption because it provides a means of assessing
resin regeneration as well as resin adsorptive capacity. The batch portion
of the test measures regenerated saturation capacity and the rate portion
measures regenerated capacity as a function of time. Using batch/rate data
and a simple, theoretical performance model, resin column performance under
a variety of operating conditions can be satisfactorily predicted during
the first half of column breakthrough. Alternately, using a simple
empirical relationship, initial column leakage can be approximated to
within 15$ of experimental results from batch data alone.
-4-
-------
SECTION 3
RECOMMENDATIONS
The batch/rate test has been shown to be a rapid, reliable method of
assessing the treatability of toxic waste streams with synthetic resin
adsorbents. Although resin technology is a viable alternative to the more
common treatment methods (e.g. activated carbon adsorption, liquid-liquid
extraction, steam stripping or biological treatment), it is not widely used
due to the difficulty of selecting the appropriate adsorbent/regenerant
combination for a particular waste stream. Guidelines need to be
established for choosing which of the numerous adsorbent/regenerant systems
to evaluate in a specific application. Using the batch/rate test and its
associated predictive models, the various commercially available resin
adsorbents should be screened in a broad spectrum of industrial waste
streams. The resulting data, along with other published adsorption data,
should then .be tabulated to provide a guide which correlates the type of
waste stream or class of pollutant with the adsorbent/regenerant system(s)
expected to be most successful in treatment.
-5-
-------
SECTION 4
MATERIALS AND METHODS
ADSORBENT CONDITIONING AND HANDLING TECHNIQUES
Twelve adsorbents, including ten synthetic resins and'two granular
activated carbons, were used in this study. Table 1 is a list of selected
physical properties of these adsorbents. The Mitsubishi and Rohm and Haas
synthetic resins were conditioned with methanol and water prior to use as
described in Appendix B. This preconditioning was done to remove the
preservatives and residual monomers from the pores of these adsorbents.
Moisture content of each conditioned resin was determined by
analytically weighing approximately 5 g of conditioned resin, drying the
resin in a 110°C oven for 2 hours, and reweighing. For all experimental
procedures, resins were-weighed in their wet form and the weight was
multiplied by % solids to determine the dry weight for calculations.
% Solids = (g dry resin/g wet resin) x 100
The granular activated carbons were used in their dry form without any
pretreatment. In the carbon adsorption isotherm study by Dobbs and Cohen,
the carbon was pulverized, screened, and slurried before use[9l.
ANALYTICAL METHODS
Solutions were analyzed on a Hewlett Packard 5710A Gas Chromatograph
equipped with a flame ionization detector for >5 ppm studies and an
electron capture detector for <10 ppm studies. A 6' x 1/8" SS column of
60/80 mesh Tenax GC was used. The data were recorded and analyzed with
either a Spectra Physics Autolab system IVB Integrator or a Hewlett Packard
3390A Reporting Integrator.
Solutions for immediate analysis were placed in 1.5 mL Wheaton
vials filled to about 7/8 capacity, sealed with Teflon-lined serum caps,
loaded onto a Hewlett Packard 7671A Auto Sampler, and analyzed. Solutions
for later analysis were placed in vials filed to capacity (i.e., with no
air space), sealed, and refrigerated. This method of sample handling and
storage was particularly critical for aqueous solutions of the pollutants.
Aqueous solutions in unsealed or improperly sealed vials decreased in
concentration by as much as 25% in 8 hours. Each sample was analyzed in
triplicate and if any value exceeded the average by >5%, the sample was
reanalyzed. In comparing the data presented in this report, this +5% error
limit should be considered.
-6-
-------
TABLE 1
Adsorbent
Amber-lite XAD-2
Amber-lite XAD-4
Amber-lite XAD-7
Amberlite XAD-8
Ambersorb XE-340
Ambersorb XE-347
Ambersorb XE-348
71 Diaion HP-10
Diaion HP-20
Diaion HP-30
Filtrasorb 300
Filtrasorb 400
PHYSICAL PROPERTIES OF ADSORBENTS
Manufacturer
Rohm and Haas
Rohm and Haas
Rohm and Haas
Rohm and Haas
Rohm and Haas
Rohm and Haas
Rohm and Haas
Mitsubishi
Mitsubishi
Mitsubishi
Calgon
Calgon
Chemical
Nature
polystyrene
polystyrene
acrylic ester
acrylic ester
polymer carbon
polymer carbon
polymer carbon
polystyrene
polystyrene
polystyrene
activated carbon
activated carbon
Pore Vol.
(cm3/g)
0.68
0.96
0.97
0.82
0.34
0.41
0.58
0.64
1.16
0.8?
0.85
0.94
Surface
Area(m2/g)
300
725
450
160
400
350
500
500
720
570
1000
1125
Pore DJa.
Ave. (A)
100
50
85
150
200, 15a
200, 5a
200, 15a
_b
70
_b
_b
_b
Surface
Polarity
low
low
intermediate
intermediate
very low
low
intermediate
low
low
low
intermediate
intermediate
a Ave. pore diameter of the macropores and micropores, respectively.
b Ave. pore diameter not available.
-------
Standard solutions for GC calibration were made by analytically
weighing the pollutants and dissolving them volumetrically with purified
water or the organic solvent which was used as the regenerant. Plots of GC
area versus concentration were fitted by linear regression analysis to
obtain the calibration curves. Organic solutions of very high concentra-
tion (e.g., >10,000 ppm) exhibited non-linear behavior. Such solutions
were diluted volumetrically to fall within the linear calibration region.
PREPARATION OF SYNTHETIC WASTE STREAMS
A 25 x 600 mm column was filled to 2/3 height with Amberlite XAD-2
which had been treated with the organic(s) required in the synthetic waste
stream. Resin loading was done approximately on a gram organic per gram
resin basis. For example, 220 g resin in water was treated dropwise with
stirring with 220 g organic. Solid organics were added as a slurry in
acetone/water, methanol/water, or toluene/water as determined by the type
of stream desired. The mixture was then placed in a jar and shaken on a
platform shaker for about a week. The solution was decanted and the loaded
resin was washed with several 200 mL portions of purified water. After the
resin was slurried into the column, it was washed with 10 L of purified
water.
The loaded column was connected via Teflon tubing and Teflon or
stainless steel adapters to a large, stirred reservoir (e.g., 5 L, 4-neck
flask) filled with a nearly saturated aqueous solution of the organic(s).
The synthetic waste stream was continuously pumped through the column and
recycled through the reservoir at a rate of ca. 5 BV/hr (50 mL/min). The
solution contacted only glass, Teflon, or stainless steel. Nearly constant
concentration waste stream was withdrawn as needed for the various tests,
and make-up or dilute solution was added. After initial start-up,
equilibration time for addition of 1 to 4 L of make-up solution was about
16 hours.
-8-
-------
SECTION 5
ADSORPTION TEST METHODS
ADSORPTION ISOTHERMS
Adsorption isotherm tests were performed in Alltech 100 mL Reaction
Vials. Four or five different weights of each adsorbent to be tested were
added to separate vials. The following dosages are provided as a guide
for performing these isotherm tests: 1) for initial pollutant concentra-
tions of about 200 ppm use 0.04 to 0.5 g adsorbent, 2) for initial concen-
trations of about 2000 ppm use 0.2 to 1.7 g adsorbent. At low pollutant
concentrations, carbonaceous adsorbents have higher capacities than
polymeric adsorbents. Thus, lower dosages of carbonaceous adsorbents are
required to achieve similar equilibrium concentrations. After addition of
adsorbent, the vials were filled to capacity (about 120 mL) with the known
concentration waste stream using a repipette dispenser. For the control
experiment, an empty vial was similarly filled with the waste stream. The
vials were sealed with Teflon-lined silicon rubber septa and aluminum
crimp caps, and shaken on a platform shaker for a minimum of seven days
until equilibrium was attained. The reaction vials were opened and
supernatants were transferred to vials for GC analysis (vide supra). From
the weight, volume, and concentration data, the adsorption capacity of
each resin was calculated and plotted according to the Freundlich equation.
BATCH/RATE TESTS
The adsorption procedure of the batch/rate test was performed in
the reservoir portion of the synthetic waste stream system. "Tea bags"
formed from ca. 25 sq in pieces of Teflon macrofiltration material
(mesh size 75 to 150 microns) and stainless steel wire were filled with
0.5 g (dry weight) adsorbent. The bags were suspended in the reservoir
which was stirred rapidly at ~400 rpm. After a predetermined exposure
time, the "tea bag" was dewatered by draining and blotting with filter
paper and then stored in a tightly capped vial with minimum air space
until regeneration was performed.
Regeneration was performed on the entire amount of adsorbent in
an 8 mL mini-column. Twelve bed volumes of regenerant were pumped
through the column at a rate of 6 BV/hr (^15 mL/hr) or, the adsorbent was
steam treated for 2 hours. During adsorption and regeneration, solvents
and solutions contacted only glass, stainless steel or Teflon. The
regenerant solution was collected and its volume was measured. To deter-
mine the amount of eluted pollutant, a sample was placed in a vial and
analyzed by GC (vide supra). The column containing regenerated adsor-
-9-
-------
bent was dried at 110°C for 2 hours and the weight of dry adsorbent was
obtained. The adsorbent's capacity in mg/g was then calculated: regen-
erant concentration (mg/L) x regenerant volume (L>-1 adsorbent weight (g).
Batch Test
1. A "tea bag" containing ca. 0.5 g of each adsorbent to be tested
was placed in the waste stream for a minimum of 48 hours. (Time
sufficient for saturation of the adsorbent with pollutant is
required).
2. After saturation, the adsorbent was dewatered (drained and
blotted), transferred to a mini-column, and regenerated.
3. Adsorbent weight and solution volume and concentration were
determined as described above.
4. The regenerated adsorption capacity at saturation was calculated
for each adsorbent.
Rate Test
1. 5 to 7 "tea bags" containing ca. 0.5 g each of the same adsorbent
, (generally the best adsorbent) were placed in the waste stream.
2. The bags were removed at predetermined times over a 24-hour
period. At least 4 samples were removed during the first hour.
(If waste stream concentrations were much less than 100 ppm, these
times were increased significantly).
3. The adsorbent from each bag was dewatered, placed in a separate
mini-column, and regenerated with a known effective regenerant.
4. Adsorbent weight and solution volume and concentration were
determined as previously described.
5. The regenerated adsorption capacity at each exposure time was
calculated.
6. The "adsorption" rate was obtained from the slope of a graph (or
linear regression analysis) of the adsorption capacity versus time
during approximately the first hour, and the mass transfer
coefficient was calculated.
COLUMN EXPERIMENTS
Column experiments were conducted in Ace 11 x 600 mm jacketed
chromatographic columns fitted with Teflon adapters and removable glass
filter discs. Weighed, conditioned adsorbent was slurried into a column
and backwashed with high purity water to give 100% bed expansion for
30 minutes. The adsorbent was then allowed to settle prior to any down
flow of water. This procedure classified the adsorbent bed, placing the
-10-
-------
sinal lest beads at the top of the column and the largest at the bottom.
Water was then drained to 1 cm above the adsorbent bed and the bed volume
was measured.
The waste stream was pumped through the classified column using a
Fluid Metering, Inc. pump and Teflon tubing. Typical flow rates were 1, 2,
3, or 4 gpm/ft3 (8, 16, 24, or 32 BV/hr). Flows were monitored
periodically during the course of each experiment and adjusted as required
to maintain a constant flow. Samples for GC analysis were collected
periodically in the appropriate vials and analyzed immediately or carefully
stored for later analysis (vide supra).
When column loading was complete (i.e., at 90 +_ 10/6 breakthrough) the
column was drained and regenerant was slowly pumped upflow to about 2 cm
above the bed. This displaced air from the bed. Regenerant was then
pumped downflow at 1 BV/hr until no pollutant was detected in the effluent.
Samples were collected volumetrically in 1/4 BV portions and analyzed by
GC. Generally, regeneration was complete in 3 BV or less.
Excellent, illustrated descriptions of laboratory scale column loading
and regeneration experiments using synthetic resin adsorbents have been
published [2,5].
-11-
-------
SECTION 6
RESULTS AND DISCUSSION
ADSORPTION ISOTHERM TESTS
The isotherm test provides a method of predicting an adsorbent's
equilibrium capacity for adsorption of specific organics from water over a
range of concentrations. This test method was used to generate data for
comparison with capacity data obtained using the new batch test method
developed during this project.
The adsorption data were plotted according to the log form of the
empirical Freundlich equation:
x 1
log = log K + log Cf -
m n
where x = mg compound adsorbed, [(Cjjj-Cf) • volume]
m = g adsorbent
Cf = equilibrium concentration of the solution (mg/L)
C^u = initial concentration of the solution (mg/L)
K,n = constants
The data were fitted by least squares linear regression analysis of the log
form of the Freundlich equation to locate the isotherm line. The fit of
the data to this line is indicated by the correlation coefficient R. When
R?1 there is a perfect fit. When R=0 there is no correlation between the
data and the line. The constant K represents the capacity (mg compound
adsorbed per g adsorbent) at an equilibrium concentration of 1 ppm. These
K values are usually an extrapolation of the data by at least a factor of
10. 1/n, the slope of the line, is an indicator of adsorption intensity.
A low slope indicates low sensitivity to concentration changes. Generally.,
the polymeric absorbents are much more sensitive to concentration changes
than the carbonaceous adsorbents (i.e. the slopes of their isotherms are
much greater).
Filtrasorb 300 carbon adsorption isotherms for a large number of toxic
organics have been reported [93. The work was done under very different
experimental conditions than the isotherm tests reported here: the carbon
was pulverized, screened to 200-400 mesh, and slurried before use;
equilibration time was 2 to 3 hours; initial solution concentration was
less than 10 ppm; and carbon dosage was low (0.001 to 0.5 g/L). Since
-12-
-------
isotherms are not strictly linear over large concentration ranges
(i.e. 1Q2-103), it is not realistic to expect linearity over the large
concentration range represented by these three studies (i.e. .001 to
2000 ppm). Thus, carbon adsorption isotherms for the pollutants reported
in this study cannot be compared to previously reported data.
2-Nitrophenol
Adsorption isotherms for 2-nitrophenol (NPL) in water buffered at pH 5
and pH 8 are shown in Figures 1 and 2, respectively for the Amberlite
adsorbents. At pH 5, NPL (Kase.SxIO"8) is almost entirely in its neutral
form while at pH 8 the ratio of weak acid to conjugate base (anion) is
1:7. In addition to the non-functionalized, high surface area adsorbents
tested, two very low surface area, weak base ion exchange resins were also
tested; (i.e. Amberlites IRA-45 and IRA-93). The Freundlich parameters
obtained from linear regression analyses of the isotherm data are recorded
in Table 2. Assuming isotherm linearity, adsorption capacities at selected
concentrations of NPL from 10*to 1000 ppm were calculated from the Freund-
lich parameters and are recorded in Table 3.
The effect of pH on adsorption capacity is marked. All the resins
have higher capacities at the lower pH where NPL is in its neutral form.
In this NPL concentration range, XAD-4 is the best Amberlite adsorbent
tested. These results indicate that surface area effects (adsorption via
Van der Waals forces) predominate over the ionic interaction occurring in
the adsorption of polarized weak adid NPL by the amine-containing weak base
resins (adsorption via hydrogen bonding) [6,10,11].
Diaion adsorbents were not available for testing at the time these
experiments were performed. Difficulties in pH control and dosage levels
gave rise to very erratic data during testing with carbonaceous ad-
sorbents. Thus, no isotherm data are reported for those adsorbents.
Tetrachloroethylene
Due to problems with the electron capture detector of the GC, the
tetrachloroethylene isotherm data are highly unreliable and are not re-
ported here. However, batch/rate data are acceptable and are reported in a
later section.
Nitrobenzene
Adsorption isotherms for nitrobenzene (NB) in water are shown in
Figures 3, 4, and 5 for the Amberlite, Diaion, and carbonaceous adsorbents
respectively. Freundlich parameters are recorded in Table 4 and calculated
equilibrium adsorption capacities at selected NB concentrations are re-
corded in Table 5. Since the plotted data are not linear (see Figures),
extrapolation of the linear isotherm data below 10 ppm and above 1000 ppm
leads to inaccurate capacity data. For Diaion HP-20 at 1900 ppm NB (in-
fluent concentration), the calculated capacity is 670 mg/g but the capacity
obtained graphically from a "best fit" curve is 910 mg/g. Similarly
for Amberlite XAD-4, at 1900 ppm NB, the calculated capacity is 859 mg/g
but the capacity obtained graphically is 1050 mg/g.
-13-
-------
1000
«0V
i
o
I
o
it
tg
to
a
100
10
• XAD-2
A XAD-4
• XAD-7
* XAD-8
— UNEAR LEAST SQUARES LINE
I
I
10
100 1000
EQUILIBRIUM CONCENTRATION (ppm)
FIGURE 1. ADSORPTION ISOTHERMS FOR 2-NITROPHENOL (pH 5)
ON AMBERLITE ADSORBENTS
-------
1000
E
fc
O
Q.
5
100
a.
•i. g
Ul O
1
10
© XAD-2
£ XAD-4
H XAD-7
V XAD-8
, LINEAR LEAST SQUARES LINE
10
100 1000
EQUILIBRIUM CONCENTRATION (ppm)
FIGURE 2. ADSORPTION ISOTHERMS FOR 2-NITROPHENOL (pH 8)
ON AMBERLITE ADSORBENTS
-------
TABLE 2
ADSORPTION ISOTHERM TESTING WITH 2-NITROPHENOL
Freundlich Parameters
Adsorbent
XAD-2
XAD-2
XAD-4
XAD-4
XAD-7
XAD-7
XAD-8
XAD-8
IRA-45
IRA-45
IRA-93
IRA-93
pH
5
8
5
8
5
8
5
8
5
8
5
8
Corr. Coef.(R)
1.000
0.989
1.000
0.997
0.993
0.962
0.998
0.954
0.960.
0.969
0.995
0.980
1/n
0.440
0.639
0.410
0.470
0.679
1.079
0.605
1.234
0.566
0.563
0.569
1.057
a
K
25.1
1.7
74.3
10.2
9.5
0.1
10.8
0.0
12.7
9.4
9.8
0.3
a K is equal to the equilibrium adsorption capacity at a
residual concentration of 1 ppm.
-16-
-------
TABLE 3
ADSORPTION CAPACITIES FDR 2-NITROPHENOL MEASURED BY ISOTHERM TESTS3
b
Equilibrium Adsorption Capacity (mg/g)
Adsorbent
XAD-2
XAD-2
XAD-4
XAD-4
XAD-7
XAD-7
XAD-8
XAD-8
IRA-45
IRA-45
IRA-93
IRA-93
£H
5
8
5
8
5
8
5
8
5
8
5
8
at 10 ppm
69
7
191
30
45
1
43
1
47
34
36
4
at 100 ppm
190
32
491
89
217
17
175
13
172
126
140
43
at 750 ppm
462
116
1120
228
852
148
592
159
538
390
421
365
at 1000 ppm
524
139
1260
261
1040
202
704
227
633
459
496
495
a The initial concentration of aqueous 2-nitrophenol solution was
975 + 25 ppm and equilibrium concentrations were in the range of
20 to 900 ppm.
b The capacities are calculated from the Freundlich parameters
obtained from a linear regression analysis of the isotherm data.
-17-
-------
1000
o>
a
Q.
<
o
z
o
100
To
W
a
10
• XAD-2
A XAD-4
• XAD-7
LINEAR LEAST SQUARES ONE
"BEST FIT" CURVE
I I I I I I I I
I I I I I I l
10
100 1000
EQUILIBRIUM CONCENTRATION (ppm)
FIGURE 3. ADSORPTION ISOTHERMS FOR NITROBENZENE
ON AMBERLITE ADSORBENTS
-------
1000
O)
E
fc
u
Q.
O
100
«
i O
a
10
• HP-10
• HP-20
A HP 30
—- UNEAR LEAST SQUARES UNE
-— "BEST FIT" CURVE
I I I I I I I
I
i i i i i I
_l I
i i i i i
10
100 1000
EQUILIBRIUM CONCENTRATION (ppm)
FIGURE 4. ADSORPTION ISOTHERMS FOR NITROBENZENE
ON DIAION ADSORBENTS
-------
1000
w«
t
0.
2
100
<=> O
1
10
• XE-340
• XE 347
A XE-348
V FS-300
^ FS 400
— LINEAR LEAST SQUARES LINE
J 1
I I I I
J i
I I I I I
10
100 1000
EQUILIBRIUM CONCENTRATION (ppm)
FIGURE 5. ADSORPTION ISOTHERMS FOR NITROBENZENE
ON CARBONACEOUS ADSORBENTS
-------
TABLE 4
ADSORPTION ISOTHERM TESTING WITH NITROBENZENE
Freundlich Parameters
Adsorbent Corr. Coef.(R) 1/n
HP-10 1.000 0.539 9.90
HP-20 0.997 0.533 11.9
HP-30 0.999 0.491 14.1
XAD-2 0.995 0.469 10.0
XAD-4 0.999 0.48? 21.8
XAD-7 1.000 0.623 4.22
XAD-8 0.998 0.630 3.91
XE-340 0.999 0.163 76.9
XE-347 0.998 0.0665 135
XE-348 0.994 0.0688 217
FS-300 0.990 0.168 155
FS-400 0.988 0.223 142
K is equal to the equilibrium adsorption capacity at a residual
concentration of 1 ppm.
-21-
-------
TABLE 5
ADSORPTION CAPACITIES FOR NITROBENZENE MEASURED BY ISOTHERM TESTS3
b
Equilibrium Adsorption Capacity (mg/g)
Adsorbent
HP-10
HP-20
HP-30
XAD-2
XAD-4
XAD-7
XAD-8
XE-340
XE-347
XE-348
FS-300
FS-400
at 10 ppm
34
41
44
29
70
18
17
119
158
255
228
236
at 100 ppm
118
139
135
90
205
74
71
163
183
298
336
394
at 1000 ppm
409
475
417
255
628
312
304
236
214
350
494
659
at 1900 ppm
578
670
571
345
859
467
456
262
222
365
551
760
The initial concentration of the aqueous nitrobenzene solution
was 1900 ppm and equilibrium concentrations were in the range of
10-1000 ppm.
The capacities are calculated from the Freundlich parameters obtained
from a linear regression analysis of the isotherm data.
-22-
-------
The results of this isotherm testing indicate that at high NB con-
centrations (e.g. >1000 ppm), Amberlite XAD-4 is the best adsorbent. At
low concentrations (e.g. <100 ppm) the activated carbon Filtrasorb-400 is
the best adsorbent.
Based on the results of the NPL and NB isotherm studies, only a
representative group of the 12 available adsorbents were selected for
further testing. The 6 adsorbents chosen include Diaion HP-20, Amberlites
XAD-4 and XAD-7, Ambersorbs XE-340 and XE-348, and Filtrasorb 400. These
were considered to be the best representatives of their particular class
(see Table 1).
1,2°Dichloropropane
Adsorption isotherms for 1,2-dichloropropane (PDC) in water appear in
Figures 6 and 7 for the polymeric and carbonaceous adsorbents respectively.
Freundlich parameters, obtained from linear regression analyses of the
data, are recorded in Table 6. Assuming isotherm linearity, adsorption
capacities at selected equilibrium concentrations of PDC were calculated
from these Freundlich parameters and are recorded in Table 7. Again the
plotted data show nonlinearity so both the linear least squares lines and
the "best fit" curves are displayed. Extrapolation of least squares data
to predict adsorption capacity at initial waste stream concentration
(2500 ppm) produces inaccurate capacity data. For example, for XAD-4 at
2500 ppm PDC, linear extrapolation predicts a capacity of 747 mg/g while
curved extrapolation predicts a capacity of 1000 mg/g.
At. very high PDC concentrations (e.g. >1500 ppm) the polymeric
adsorbents Diaion HP-20 and Amberlite XAD-4 have the highest capacities.
At intermediate PDC concentrations (e.g. 100-1000 ppm), the activated
carbon Filtrasorb 400 is the best adsorbent.
2,4-Dinitrotoluene
Adsorption isotherms for 2,4-dinitrotoluene (DNT) in water are shown
in Figures 8 and 9 for the polymeric and carbonaceous adsorbents
respectively. Freundlich parameters are listed in Table 8 and the cal-
culated adsorption capacities at selected DNT concentrations are listed in
Table 9. The correlation coefficients are excellent and the data appear to
be more linear than in the previous studies. However, the concentration
range in the DNT study is about a factor of 10 less (i.e., 150 ppm versus
1500 ppai).
Filtrasorb 400 is the best adsorbent over the entire concentration
range. At high DNT concentrations, the polymeric adsorbents Diaion HP-20
and Amberlite XAD-4 also have good capacities.
-23-
-------
1000
-
3
Q.
<
o
z
o
i O-
NJ |£
•t o
100
10
A XAD-4
• HP-20
• XAD-7
—— LINEAR LEAST SQUARES LINE
--- "BEST FIT" CURVE
10
100 1000
EQUILIBRIUM CONCENTRATION (ppm)
FIGURE 6. ADSORPTION ISOTHERMS FOR 1,2-DICHLOROPROPANE
ON POLYMERIC ADSORBENTS
-------
1000
£
t
o
Q.
O
z 100
o
o
U)
o
10
A FS-400
• XE-348
• XE-340
— LINEAR LEAST SQUARES LINE
"BEST FIT" CURVE
I I I I I I I
I I I I I I I
I I
1 1 J I
10
100 1000
EQUILIBRIUM CONCENTRATION (ppm)
FIGURE 7. ADSORPTION ISOTHERMS FOR 1,2-DICHLOROPROPANE
ON CARBONACEOUS ADSORBENTS
-------
TABLE 6
ADSORPTION ISOTHERM TESTING WITH 1,2-DICHLOROPROPAME
Adsorbent
HP-20
XAD-4
XAD-7
XE-340
XE-348
FS-400
Freundlich
Corr. Coef.(R)
0.991
0.995
0.999
0.990
0.912
0.975
Parameters
1/n
0.743
0.619
0.801
0.230
0.142
0.315
K*
2.3
5.9
0.8
31.5
97.8
54.6
K is equal to the equilibrium adsorption capacity at a
residual concentration of 1 ppm.
TABLE 7
ADSORPTION CAPACITIES FOR 1,2-DICHLOROPROPANE
MEASURED BY ISOTHERM TESTS3
b
Equilibrium Adsorption Capacity (mg/g)
at 100 ppm
71
102
31
91
188
233
at 1000 ppm
394
424
196
154
262
482
' at 2500 ppm
778
747
409
190
298
644
Adsorbent
HP-20
XAD-4
XAD-7
XE-340
XE-348
FS-400
The initial concentration of aqueous 1,2-dichloropropane
solution was 2330 ppm and equilibrium concentrations were in
the range of 50-2000 ppm.
The capacities are calculated from the Freundlich parameters
obtained from a linear regression analysis of the isotherm data.
-26-
-------
1000
^*r
fc
O
I
o
2
O
100
10
A XAD-4
• HP-20
• XAD-7
LINEAR LEAST SQUARES LINE
I
I I I I I
I
10 100
EQUILIBRIUM CONCENTRATION (ppm)
FIGURE 8. ADSORPTION ISOTHERMS FOR 2,4-DINITROTOLUEIME
ON POLYMERIC ADSORBENTS
-------
1000
**v
E
o
Q.
O
o
t
ti tt
00 O
'8
100
10
• XE-34O
• XE-348
A FS-40O
— UNEAR LEAST SOUARES LINE
I
I
10 100
EQUILIBRIUM CONCENTRATION (ppm)
FIGURE 9. ADSORPTION ISOTHERMS FOR 2,4-DINITROTOLUENE
ON CARBONACEOUS ADSORBENTS
-------
TABLE 8
ADSORPTION ISOTHERM TESTING WITH 2,4-DINITROTOLUENE
Freundlich Parameters
Adsorbent
HP-20
XAD-4
XAD-7
XE-340
XE-348
FS-400
Corr. Coef.(R)
0.998
1.000
0.998
0.959 '
1.000
0.997
1/n
0.504
0.431
0.569
0.154
0.187
0.333
K3
29.5
42.2
13.8
53.2
95.5
104
K is equal to the equilibrium adsorption capacity at a
residual concentration of 1 ppm.
TABLE 9
ADSORPTION CAPACITIES FOR 2.4-DINITROTOLUENE
MEASURED BY ISOTHERM
TESTS3
Equilibrium Adsorption Capacity
Adsorbent
HP-20
XAD-4
XAD-7
XE-340
XE-348
FS-400
at 1 ppm
29.5
42.2
13.8
53.2
95.5
104
at 10 ppm
94.1
114
51.2
75.9
147
224
at 100 ppm
300
308
190
108
226
483
(mg/g)
at 180 ppm
403
397
265
118
252
587
The initial concentration of aqueous 2,4-dinitrotoluene
solution was 180 _+ 10 ppm and equilibrium concentrations were
in the range of 1-140 ppm.
The capacities are calculated from the Freundlich parameters
obtained from a linear regression analysis of the isotherm data.
-29-
-------
BATCH/RATE TESTS
The batch/rate test consists of exposing bags of various adsorbents to
a waste stream for a selected period of time. The exposed adsorbents are
then regenerated, typically using organic solvents, steam, or caustic, and
the regenerated adsorption capacities are obtained. In the batch portion
of the test, the adsorbents are exposed to the waste stream sufficiently
long enough to saturate each adsorbent with pollutants. The regenerated
adsorption capacity at saturation, as obtained from the batch test, is a
measure of maximum column capacity. The batch capacity data are compared
and used as guide for selecting the optimum adsorbent(s) for additional
testing in the waste stream. In the rate portion of the test, bags
of the optimum adsorbent(s) are exposed to the waste stream and removed as
time passes, prior to adsorbent saturation. From the amount adsorbed (i.e.
regenerated) as a function of time, the mass transfer coefficient is
calculated. Using data provided by batch/rate studies (i.e. saturation
capacity and mass transfer coefficient) column performance can be predicted
under a variety of column operating conditions.
During the course of this study, the batch/rate test procedures
were refined. Thus, data obtained during the NPL study (the first study)
are not as reliable as those obtained in the PDC study (the fourth study).
When testing is done in a limited volume waste stream (such as the 5L
reservoir of the synthetic waste stream system), only 2 to 5 bags of
adsorbent should be tested simultaneously. Too much adsorbent can signifi-
cantly depress the reservoir concentration which ideally should remain
nearly'constant. This should not be a problem in testing real waste
streams if bags of adsorbent can be placed into a constantly flowing
stream. When the bags are removed from the waste stream, the adsorbent
should be drained and either regenerated immediately or stored in small,
tightly sealed bottles with a minimum of air space until regeneration is
performed. This minimizes loss of compound due to volatility.
The polymeric adsorbents usually regenerate well with organic solvents
(e.g. methanol, acetone, toluene). Their regenerated saturation capacities
are often higher than predicted by extrapolation of linear isotherm lines
to influent concentrations. As previously discussed, isotherm data cannot
be accurately extrapolated due to isotherm nonlinearity over large concen-
tration ranges. The carbonaceous adsorbents usually have lower regenerated
saturation capacities than predicted by isotherm tests due to incomplete
regeneration by organic solvents under the chosen regeneration conditons.
During the rate test the adsorbent must be well agitated. By using a
fast stirring rate (>400 rpm) and spacious "tea bags" the influences of
diffusion through bulk solution and adsorbent packing on the rate of
adsorption (mass transfer coefficient) are minimized. Due to experimental
design, very short adsorption times (less than 10 minutes) can result in
highly scattered data. Good kinetic data should be fairly linear or follow
a smooth curve during the first 10 to 90 minutes of adsorption.
-30-
-------
2-NitrophenoI
The results of the batch test adsorption of NPL from water buffered at
pH 5 and pH 8 are reported in Table 10. Samples of each adsorbent were
regenerated with methanol and with acetone. The capacities obtained by
regeneration with these two solvents generally agree except for studies
with the weak base resins, IRA-45 and IRA-93. The data are somewhat
scattered, but methanol appears to be a more effective regenerant than
acetone for the weak base resins. As was noted in the isotherm results
section, the polymeric adsorbents have significantly higher NPL capacity at
lower pH. However, for the activated carbon FS-400, pH has much less
effect on capacity, concurring with previous findings for adsorption of-NPL
by carbon [9]. Amberlite XAD-4 and Filtrasorb 400 have the highest re-
generated saturation capacities.
During the NPL adsorption study, the first application of the new
batch/rate test method, a number of experimental difficulties were en-
countered. The test method was refined and resulting procedural changes
were incorporated in later studies with other compounds. The lack of
precision in the NPL data is primarily due to very small sample size (i.e.
10 to 30 mg adsorbent and 1 to 2 mL regenerant).
Comparison of NPL adsorption data for the XAD resins obtained
by batch testing (Table 10) and isotherm testing (Table 3, at 750 ppm)
reveals some discrepancy. The results obtained at pH 8 agree satis-
factorily, allowing for experimental problems encountered in batch testing
(vide supra). However, at pH 5, batch results (saturation capacities
measured by regenerating the adsorbents) are a factor of 2 or 3 lower than
isotherm results (saturation capacities measured by concentration changes
of the waste stream). This indicates that the adsorbents were not re-
generated completely. Since both isotherm and batch results indicate that
the conjugate base of NPL is less well adsorbed than the neutral molecule,
caustic would be good regenerant to study. Related studies of
p-nitrophenol adsorption on synthetic resins confirm this [6,10]. Re-
generation of adsorbents plays a critical role in the economics of waste
water treatment by adsorbents. The batch/rate test method provides a
relatively simple means of assessing both adsorbent capacity and regenera-
tion ease, whereas the isotherm test method measures capacity only.
Rate test results for adsorption of 2-nitrophenol by XAD-4 at pH 5 and
pH 8 are recorded in Table 11 and plotted in Figure 10. The rate of uptake
of NPL is initially much faster at pH 5 than at pH 8. The tests were
performed under conditions of decreasing NPL concentration with increasing
time, and the data do not extrapolate well to equilibrium (batch)
capacities. This portion of the batch/rate test was also revised for later
studies with other compounds.
Tetrachloroethylene
Batch test results for adsorption of tetrachloroethylene (TETRA)
from water are listed in Table 12. XAD-4 has the highest methanol re-
generated capacity of any of the adsorbents tested (1170 mg/g).
-31-
-------
TABLE 10
ADSORPTION CAPACITIES FOR
Adsorbent
XAD-2
XAD-2
XAD-4
XAD-4
XAD-7
XAD-7
XAD-8
XAD-8
IRA-45
IRA-45
IRA-93
IRA-93
XE-340
XE-340
XE-348
XE-348
FS-400
FS-400
Methanol
pH Adsorption
5
8
5
8
5
8
5
8
5
8
5
8
5
8
5
8
5
8
2-NITROPHENOL MEASURED BY BATCH TESTS**
Regenerated
Capacity (mg/g)
199
149
518
311
204
157
242
153
75
134
123
261
189
150
164
68
485
477
Acetone Regenerated
Adsorption Capacity (mg/g)
207
139
519
386
270
148
247
136
109
35
63
119
189
175
242
139
402
498
a Final 2-nitrophenol concentration was 750 ppm at pH 5 and 820 ppm
at pH 8.
-32-
-------
TABLE 11
RATE TEST RESULTS FOR ADSORPTION OF 2-NITROPHENOL BY XAD-4&
Time (min.)
5
20
60
5
20
75
£H
5
5
5
8
8
8
Methanol Regenerated Adsorption Capacity (mg/g)
124
354
395
51
89
183
a Final 2-nitrophenol concentration was 770 ppm at pH 5 and 950 ppm
at pH 8.
-33-
-------
10
20
30 40 50
TIME (MINUTES)
60
70
80
FIGURE 10. RATE OF ADSORPTION OF 2-NITROPHENOL BY AMBERLITE XAD-4
-------
Results of rate testing for adsorption of TETRA by XAD-4 are reported
in Table 13 and plotted in Figure 11. The rate of adsorption increases
linearly with time over the first 60 minutes. Regression analysis yields
an adsorption rate of 0.8 mg/g-min. This corresponds to a mass transfer
coefficient of 0.0053 L/g-min assuming an influent concentration of
150 mg/L.
Nitrobenzene
The results of batch test adsorption of nitrobenzene (MB) from water
are reported in Table 14. Twelve adsorbents were tested and methanol was
selected as the regenerant. Although methanol proved to be a very good
regenerant, methanol may not be the optimum regenerant. Other solvents
which might be tested include acetone and toluene. The polymeric ad-
sorbents regenerated better with methanol than either the carbonaceous
adsorbents or activated carbons. XAD-4 has the highest methanol re-
generated capacity (1310 mg/g), but both HP-10 (1140 mg/g) and HP-20
(1160 mg/g) also perform well.
Comparison of the NB capacity data obtained from isotherm testing
(Table 5, at 1900 ppm) and batch testing (Table 14) reveals that the batch
tests predict much higher adsorption capacities for polymeric adsorbents
than isotherm tests do. This discrepancy is due to the extrapolation of
isotherm data from the equilibrium concentration range of 10 to 1000 ppm to
the influent concentration of 1900 ppm. The inaccuracy of such extrapola-
tion was discussed previously.
Rate test results for adsorption of nitrobenzene by XAD-4 are listed
in Table 15 and plotted in Figure 12. The rate test was repeated. Due to
improper storage of the samples, data from the first kinetic study did not
extrapolate well to the equilibrium adsorption capacity of 1310 mg/g. Data
measured at very short contact times «10 minutes) were omitted from linear
regression analysis due to their non-reproducibility. For 6 points ob-
tained during the first hour, regression analysis yields a rate of adsorp-
tion of 3-19 mg/g-min with a correlation coefficient of 0.98. Using this
rate of adsorption, a mass transfer coefficient of 0.00168 L/g-min is
calculated assuming an influent concentration of 1900 mg/L. Within the
first hour of exposure to the 1900 ppm NB stream, XAD-4 has attained 50$ of
its saturation capacity.
1,2-Dichloropropane
The results of batch test adsorption of 1,2-dichloropropane (PDC) are
reported in Table 16. Samples of each saturated adsorbent were regenerated
with methanol and acetone. The capacities obtained by regeneration with
these solvents are identical within experimental error (i.e., +5$ of the
average). Batch results again do not correlate well with capacities
calculated from Freundlich parameters (Table 7, at 2500 ppm). However,
batch capacities for the polymeric adsorbents do agree fairly well with
capacities predicted by curve fitting of the isotherm data (see Figure 6).
The best adsorbents are Diaion HP-20 and Amberlite XAD-4 with regenerated
saturation capacities of 1075 + 15 mg/g and 1055 ± 25 mg/g respectively.
-35-
-------
TABLE 12
ADSORPTION CAPACITIES FOR TETRACHLOROETHYLENE
MEASURED BY BATCH TESTS3
Adsorbent Methanol Regenerated Adsorption Capacity (mg/g)
XAD-2 966
XAD-4 1170
XAD-7 438
XAD-8 360
XE-340 330
XE-347 466
XE-348 ' 676
a Initial tetrachloroethylene concentration was 150 ppm.
TABLE 13
RATE TEST RESULTS FOR ADSORPTION OF TETRACHLOROETHYLENE BY XAD-43
Time (min.) Methanol Regenerated Adsorption Capacity (mg/g)
3 3.0
5 5.3
7 6.2
10 8.0
15 10.8
23 23.1
40 , 30.5
a Initial tetrachloroethylene concentration was 150 ppm.
-36-
-------
JE?
o>
Q
oc
o
a
<
H
Z
15
J, O
40
30
20
10
0
10 20 30 40 50
TIME (MINUTES)
60
70
80
FIGURE 11. RATE OF ADSORPTION OF TETRACHLOROETHYLENE BY AMBERLITE XAD-4
-------
TABLE 14
ADSORPTION CAPACITIES FOR NITROBENZENE
MEASURED BY BATCH TESTSa
Adsorbent Methanol Regenerated Adsorption Capacity (mg/g)
HP-10 1140
HP-20 1160
HP-30 706
XAD-2 699
XAD-4 1310 ••
XAD-7 881
XAD-8 580
XE-340 283
XE-347 179
XE-348 241
FS-300 406
FS-400 537
a Nitrobenzene concentration was maintained at 1890 + 30 ppm.
TABLE 15
RATE TEST RESULTS FOR ADSORPTION OF NITROBENZENE BY XAD-4a
Time (min.)
1st Kinetic Run
5
15
30
45
97
247
420
1440
2nd Kinetic Run
6
18
35
55
196
Methanol Regenerated Adsorption Capacity (mg/g)
496
546
593
653
747
917
937
1020
186
567
634
673
1083
Nitrobenzene concentration was maintained at 1890 HK 30 ppm.
-38-
-------
800
INFLUENT CONCENTRATION
1890 ±30 ppm
90
TIME (MINUTES)
FIGURE 12. RATE OF ADSORPTION OF NITROBENZENE BY AMBERLITE XAD-4
-------
Although isotherm results predict that Ambersorb XE-348 has a saturation
capacity significantly greater than Ambersorb XE-340, the regenerated
capacities of these adsorbents are essentially identical. This is true for
nitrobenzene adsorption as well as 1,2-dichloropropane adsorption. The
discrepancy is believed to be due to the different chemical and physical
properties of these two carbonaceous adsorbents. XE-340 is less polar and
has a higher percentage of larger pores than XE-348[4]. After exposure
to the particular organics investigated in this study, XE-340 is completely
regenerated by organic solvent but XE-348 is not. Solvent also efficiently
regenerates the polymeric adsorbents.
XAD-4 was chosen as the adsorbent for use in the rate test. Results
of this study are recorded in Table 17 and plotted in Figure 13. A linear
regression analysis of the data from 27 to 99 minutes yields an adsorption
rate of 1.90 mg/g-min. This corresponds to a mass transfer coefficient of
.00076 L/g-min in the 2500 ppm PDC stream. The rate of approach to equili-
brium is rapid. After one hour of exposure to the 2500 ppm PDC stream,
XAD-4 has attained about 70? of its saturation capacity.
2,4-Dinitrotoluene
The results of the batch test adsorption of 2,4-dinitrotoluene
(DNT) from water are reported in Table 18. Both methanol and acetone were
used as regenerants. Since methanol did not regenerate Ambersorb XE-340 as
efficiently as acetone, methanol regeneration of the other carbonaceous
adsorbents was not performed. Both solvents completely regenerated the
polymeric adsorbents. Batch results for the polymerics agree to within
10 ^ 3$ of the capacities predicted by extrapolation of the linear isotherm
lines to 180 ppm. Extrapolation works here because the concentration range
is smaller than in previous studies. Amberlite XAD-4 has the highest
regenerated saturation capacity. As observed in the PDC study, although
isotherm data predict that Ambersorb XE-348 has a significantly higher
capacity than Ambersorb XE-340, again, their regenerated capacities
are essentially identical.
Both Amberlite XAD-4, a polymeric adsorbent, and Ambersorb XE-340, a
carbonaceous adsorbent, were subjected to the rate test method. Rate test
results are recorded in Tables 19 and 20 for XAD-4 and XE-340 respectively.
Data are plotted in Figure 14. A linear regression analysis of each data
set yields adsorption rates of 4.58 mg/g-min for XAD-4 and 0.82 mg/g-min
for XE-340. XAD-4 approaches its equilibrium capacity at a much faster
rate than XE-340. After one hour of exposure to the 180 ppm DNT stream,
XAD-4 has reached 60$ of its capacity while XE-340 has reached only
33$ of its capacity. Repetition of the rate test with XE-340 yielded the
same results. The slower adsorption rate for XE-340 is attributable to
both its smaller surface area and its pore size distribution in macropores
and micropores.[4]
Synthetic, Nitroaromatic Stream
A 4-component, synthetic, nitroaromatic stream was prepared with the
following compounds: toluene (T), nitrobenzene (NB), 2-nitrophenol (NPL),
-40-
-------
TABLE 16
ADSORPTION CAPACITIES FOR 1.2-DICHLOROPROPANE
MEASURED BY BATCH TESTS*
Adsorbent
HP-20
XAD-4
XAD-7
XE-340
XE-348
FS-400
Methanol Regenerated
Adsorption Capacity (mg/g)
1080
1060
543
245
222
452
Acetone Regenerated
Adsorption Capacity (mg/g)
1030
1090
584
227
220
434
a 1,2-Dichloropropane concentration was maintained at 2450 +_ 100 ppm.
TABLE 17
RATE TEST RESULTS FOR ADSORPTION OF 1,2-DICHLOROPROPANE BY XAD-4a
Acetone Regenerated
/ Capacity,Time = T \
le (min.)
13
20
27
35
50
66
99
180
1440
2880
Adsorption Capacity (mg/g)
531
607
693
740
750
756
849
927
1010
1090
\Equilibrium Capacity'
48.7
55.7
63.6
67.9
68.8
69.4
77.9
85.0
92.7
100.0
100
a 1,2-Dichloropropane concentration was maintained at 2500 + 100 ppm.
-41-
-------
1000
. 800
0
ai
£D
QC
§ 600
Q
10 O
' s
400
200
INFLUENT CONCENTRATION
2500 ± 100 ppm
20 40 60 80 100 120
TIME (MINUTES)
140
160 180
FIGURE 13. RATE OF ADSORPTION OF 1,2-DICHLOROPROPANE BY AMBERLITE XAD-4
-------
TABLE 18
ADSORPTION CAPACITIES FOR 2,4-DINITROTOLUENE
MEASURED BY BATCH TESTS3
Adsorbent
HP-20
XAD-4
XAD-7
XE-340
XE-348
FS-400
Methanol Regenerated
Adsorption Capacity (mg/g)
363
460
287
98
-
-
Acetone Regenerated
Adsorption Capacity (mg/g)
362
440
285
138
133
324
a 2,4-Dinitrotoluene concentration was maintained at 180 + 10 ppm.
-43-
-------
TABLE 19
RATE TEST RESULTS FOR ADSORPTION OF 2,4-DINITROTOLUENE BY XAD-4a
Time (min.)
8
14
24
40
55
Acetone Regenerated
( Capacity, Time = T Y_ _
Adsorption Capacity (mg/g) \Equilibrium Capacity/"
58
96
178
239
271
12.9
21.3
39.6
53.1
60.2
a 2,4-Dinitrotoluene concentration was maintained at 175 +_ 10 ppm.
TABLE 20
RATE TEST RESULTS FOR ADSORPTION OF 2,4-DINITROTOLUENE BY XE-340a
Acetone Regenerated / Capacity, Time = T
Time (min.) Adsorption Capacity (mg/g) \Equilibrium Capacity.
12 14 10.1
24 23 16.7
39 37 26.8
52 46 33.3
a 2,4-Dinitrotoluene concentration was maintained at 175 + 10 ppm.
100
-44-
-------
INFLUENT CONCENTRATION
175 ± 10 ppm
AMBERSORB XE-340
30 40 50
TIME (MINUTES)
80
FIGURE 14. RATE OF ADSORPTION OF 2,4-DIIMITROTOLUENE
-------
and 2,4-dinitrotoluene (DNT). These components were chosen for several
reasons: availability of data obtained from testing these compounds
individually (vide supra)t similarity to more complex industrial waste
streams such as those resulting from production of nitroaromatics or
aromatic isbcyanates [12,13], and reasonable ease of analysis using a Tenax
GC column. Because pH changes influence resin adsorption [6,10,14] and
alter chromatographic sensitivity, the stream was buffered with a pH 2
HC1/KC1 buffer.
The nitroaromatic stream was established and maintained using the
column method described in the Experimental Section. Three 70 g portions
of Amberlite XAD-2 were loaded on approximately a gram per gram basis with
one of the nitro compounds. 50 g of each solid (i.e. NPL and DNT) were
slurried in 20 mL of toluene prior to mixing with an aqueous XAD-2 sus-
pension. 70 g of liquid MB were added directly to an XAD-2 suspension.
After a week of agitation on a platform rocker, each sample was batch
washed to remove any excess organic which had not been adsorbed. The three
portions of resin were then mixed and equilibrated prior to use in the
column. Buffer was added to the 5 L reservoir and recirculated through the
column to establish an equilibrated waste stream. This stream contained
approximately 240 ppm T, 710 ppm MB, 510 ppm NPL, and 50 ppm DNT, or a
cumulative organic concentration of 1510 ppm.
The results of batch test adsorption of this nitroaromatic stream are
reported in Table 21. Samples of each saturated adsorbent were regenerated
with acetone. The adsorbent with the highest regenerated saturation
capacities for all components of this stream is Amberlite XAD-4. Diaion
HP-20 also performs well. The cumulative adsorption capacities for XAD-4
and HP-20 are 1191 mg/g and 1105 mg/g respectively. The relative amounts
of each species adsorbed by these polymeric adsorbents parallel their
relative concentrations in the waste stream (i.e. DNT < T < NPL < MB).
The polymeric adsorbents HP-20 and XAD-4 and the carbonaceous ad-
sorbent Ambersorb XE-340 were used in the rate studies. Results for HP-20
are recorded in Table 22 and plotted in Figure 15. Results for XAD-4 and
XE-340 are recorded in Tables 23 and 24 and plotted in Figures 16 and 17
respectively. For all three resins, rate of approach to saturation is
rapid. They attain 75$ of their cumulative saturation capacities within
90 minutes. These adsorption rates exceed those measured in the single
component studies.
As explained in the Introduction, resin adsorption of any given
compound is influenced by a variety of factors including pH, dissolved
electrolytes, and competition between adsorbable species. In the multi-
component stream, the added electrolyte (buffer) and other organic com-
pounds decrease the solubility of each individual component and increase
the component's adsorbability. Previous studies have shown that the
presence of electrolytes increases the adsorptive capacity of XAD-4 for any
given organic[l4]. In Table 25, batch/rate data obtained, for adsorption of
NB, NPL, and DNT from the single component streams are compared to that
obtained in the multicomponent stream. Both the linear equilibrium con-
stants (i.e. batch equilibrium capacities corrected for pollutant
-46-
-------
TABLE 21
ADSORPTION CAPACITIES FOR COMPONENTS OF THE SYNTHETIC.
NITROAROMATIC STREAM MEASURED BY BATCH
TESTSa
Acetone Regenerated Capacity for (mg/g) :
Adsorbent
HP-10
HP-20
XAD-4
XAD-7 '
XE-340
XE-348 •
'Toluene
155
178
183
100
37
18
Nitrobenzene
463
525
575
362
131
112
2-Nitrophenol
241
276
292
210
72
78
2 , 4-Dinitrotoluene1
107
126
141
100
46
32
a Waste stream concentration was maintained at the following levels:
toluene, 239 _+ 12 ppm; nitrobenzene, 708 _+ 35 ppm; 2-nitrophenol,
506 _+ 25 ppm; 2,4-dinitrotoluene, 48 + 3 ppm-
TABLE 22
RATE TEST RESULTS FOR ADSORPTION OF THE COMPONENTS OF
THE SYNTHETIC, NITROAROMATIC STREAM BY HP-2pa
Acetone Regenerated Capacity for (mg/g):
Toluene
79
97
112
126
128
Nitrobenzene
289
331
343
354
363
2-Nitrophenol
153
167
171
173
180
2 , 4-Dinitrotoluene'
32
46
62
69
75
10
18
30
47
80
a See Table 21 for waste stream concentrations.
-47-
-------
o>
CD
I
2-NITROPHENOL
•««i——«—
TOLUENE
2,4-DINITROTOLUENE
10 20 30 40 50 60
TIME (MINUTES)
70
80
FIGURE 15.
RATE OF ADSORPTION OF COMPONENTS
OF THE NITROAROMATIC STREAM BY DIAION HP-20
-------
TABLE 23
RATE TEST RESULTS FOR ADSORPTION OF THE COMPONENTS OF
Time (min.
10
15
25.5
31
55
87
THE SYNTHETIC, NITROAROMATIC STREAM BY XAD-4a
Acetone Regenerated Capacity for (mg/g)
) 'Toluene Nitrobenzene 2-Nitrophenol 2,4-Dinitrotoluene'
59 287 . 152
92 363 186
114 428 208
126 436 217
137 452 218
131 453 222
22
35
54
59
74
82
a See Table 21 for waste stream concentrations.
RATE
Time (min.
15
31
45
85
170
TABLE 24
TEST RESULTS FOR ADSORPTION OF THE COMPONENTS OF
THE SYNTHETIC, NITROAROMATIC STREAM BY XE-340a
Acetone Regenerated Capacity for (mg/g)
) Toluene Nitrobenzene 2-Nitrophenol 2,4-Dinitrotoluene1
11 51 31
17 77 46
20 90 54 ':
26 105 59
28 110 61
6
12
15
23
31
a See Table 21 for waste stream concentrations.
-49-
-------
O>
2,4-DINITROTOLUENE
20
30 40 50 60
TIME (MINUTES)
70
80
FIGURE 16. RATE OF ADSORPTION OF COMPONENTS
OF THE NITROAROMATIC STREAM BY AMBERLITE XAD-4
-------
100 -
o>
Q
UJ
m
oc
O
CO
a
*
2.4 DINITROTOLUENE
20 -
10
20
30
40 50
TIME (MINUTES)
60
80
FIGURE 17. RATE OF ADSORPTION OF COMPONENTS
OF THE NITROAROMATIC STREAM BY AMBERSORB XE-340
-------
TABLE 25
COMPARISON OF BATCH/RATE DATA FROM STUDIES OF
MULTICOMPONENT
AND SINGLE COMPONENT NITROAROMATIC STREAMS
Linear Equilibrium Constants: Batch Capacity/Concentration (L/g).
Adsorbent
XAD-4
XAD-4
HP-20
HP-20
XE-340
XE-340
Stream
Multi
Single
Multi
Single
Multi
Single
Nitrobenzene
0.81
0.69
0.74
0.61
0.19
0.15
2-Nitrophenol
0.58
0.69
0.55
NA*
0.14
0.25
2 , 4-Dinitrotoluene
2.94
2.50
2.63
2.02
0.96
0.77
Mass Transfer Coefficients: Adsorption Rate/Concentration (L/g-min)
Adsorbent
XAD-4
XAD-4
HP-20
HP-20
XE-340
XE-340
Stream
Multi
Single
Multi
Single
Multi
Single
Nitrobenzene
0.0097
0.0017
0.0036
NA
0.0019
NA
2-Nitrophenol
0.0057
0.0054
0.0017
NA
0.0015
NA
2 , 4-Dinitrotoluene
0.037
0.025
0.031
NA
0.0063
0.0047
a NA = not available, required test not performed.
-52-
-------
concentration) and the mass transfer coefficients (i.e. adsorption rates
corrected for pollutant concentration) for NB and DNT are greater when
adsorption occurs in the multicomponent stream. Although NPL data are also
presented, comparison of adsorption data from the multi and single com-
ponent streams is not valid because batch/rate testing of the pure NPL
stream was done under different experimental conditions than all other
batch/rate studies (vide supra).
Glen Cove Groundwater
Batch/rate testing using Ambersorb XE-340 was done on a contaminated
groundwater source in Glen Cove, NY. In an EPA funded pilot plant study
performed by the firm of Nebolsine Kohlman Ruggiero Engineers, Glen Cove
City well water was successfully treated with Ambersorb XE-340 for removal
of chlorinated organics[15]. The removal of cis-1,2-dichloroethylene
(CIS), trichloroethylene (TRI), tetrachloroethylene (TETRA), and
1,1,1-trichloroethane (ETH) by several treatment technologies was studied.
Batch/rate testing of this water assesses the ability of the new test
procedure to predict resin performance in very dilute streams (1 ppm total
organics). The results of batch/rate testing may be compared to and
correlated with column results from the pilot plant study (Appendix E).
During the batch/rate testing period the average pollutant concentra-
tions were 80 ppb CIS, 190 ppb TRI, 40 ppb TETRA, and 5 ppb ETH. A
reservoir (a 55 gal pail) was established at the stream source (a pipe) and
water continually cascaded into the reservoir. Compared to previous
batch/rate tests, larger "tea bags" (100 sq in) and more adsorbent (25 g)
were used for each sample to facilitate examination of both steam and
solvent regeneration. Because pollutant concentrations were so low com-
pared to others used during this project, exposure times had to be
significantly increased.
The results of acetone regeneration of rate test samples exposed to
the groundwater from 1 to 215 hours are recorded in Table 26. Because ETH
contributes less than 2% of the total organic load, it is not included
here. The data from samples obtained during the first 48 hours are plotted
in Figure 18. Linear regression analysis of these data yield the following
adsorption rates: CIS-10.2/(g/g-hr, TRI-25.7^g/g-hr, and TETRA-4.3
g/g-hr. The batch (equilibrium) sample, exposed to the stream for 1224
hours (51 days), was acetone regenerated to yield the following capacities:
CIS-4810 M,s/s, TRI-10,900 j^g/g, and TETRA-2390 M.&/&- Based on Glen Cove
pilot plant column results (Table E-1), resin saturation has not been
achieved. For example, at a flow of .5 BV/hr (4 gpm/ft3), column
capacities to 10$ leakage are: CIS-2330 MS/S, TRI-30,800 j*g/g,and
TETRA-14,600 Ag/g. Cumulative column capacity to 10/& leakage is thus
47,700>oig/g as compared to a cumulative "batch" capacity of l8,100Mg/g.
Since the batch test sample was not exposed long enough to achieve
equilibrium, no prediction of column performance from batch/rate data can
be made.
Steam regeneration of several XE-340 rate samples was performed
according to the published procedure[2]. Approximately 40 mL (19 g) of
-53-
-------
TABLE 26
RATE TEST RESULTS FOR ADSORPTION OF THE COMPONENTS
OF THE aEN COVE STREAM BY XE-340a
Time Acetone Regenerated Capacity for
(hr.)
1
2
2 3/4
11
25
48 1/4
214 3/4
'Cis-1 ,2-Dichloroethylene
9.7
20.2
29.5
129
217
504
2742
Trichloroethylene
27.8
22.4
82.9
339
786
1210
5400
Tetrachloroethylene'
4.1
9.7
10.8
54.0
129
201
. 1170
a Average concentrations: cis-1,2-dichloroethylene, 80 ppb;
trichloroethylene, 190 ppb;
tetrachloroethylene, 40 ppb.
-54-
-------
TRtCHlOROETHYLENE
CIS-1,2-DICHLOROETHYLENE
TETRACHLOROETHYLENE
25
TIME (HOURS)
30
35 40 45
FIGURE 18. RATE OF ADSORPTION OF COMPONENTS OF THE GLEN COVE STREAM
BY AMBERSORB XE-340
-------
resin was steam regenerated at 1 BV/hr for 12 BV. Because resin loading
was so low (Table 26) and total regenerant volume so high (^500 mL),
resulting pollutant concentrations were very low. Per gram resin, about
25 mL condensed steam was used as compared to 10 mL acetone. The low
concentrations and large number of fractions collected resulted in unsatis-
factory steam regeneration data. For example, steam regeneration of the
11 hour rate sample yielded only 65% of the cumulative capacity as obtained
by acetone regeneration. Similarly steam regenerated capacities were only
80$ of the acetone regenerated capacities for the 25 hour rate sample.
Despite the problems encountered here, steam regeneration of Ambersorb
XE-340 used in Glen Cove groundwater treatment has been demonstrated
previously [2,15].
COLUMN EXPERIMENTS
The general procedures for column loading and regeneration were
described previously. Effluent concentration ia measured as a function of
the volume of solution passed through the column. Breakthrough curves are
plotted as effluent concentration as percent of influent versus volume of
treated solution. These curves are usually non-symmetrical and exhibit
tailing as column saturation is approached. Saturation column capacity
(mg/g) is determined by integration of the area above the breakthrough
curve. Capacity can be estimated to 95 ± 5% directly from the graph
by
Vfio«t CT
Saturation Column Capacity = £
where ^(,Q% is volume treated to 60% breakthrough, Cj is influent con-
centration in mg/L, and WR is grams of dry adsorbent.
For column regeneration, regeneration concentration is measured as a
function of volume collected in 1/4 BV fractions. Regeneration curves are
plotted as cumulative amount of pollutant regenerated per weight adsorbent
versus bed volumes of regenerant. Column regeneration was complete in 3 BV
or less. Since regenerated column capacity is the sum of 10 or more
analyses, this measurement is subject to more experimental error than batch
test data. Because of this experimental error, regenerated, column capacity
sometimes exceeds loaded column capacity. Nevertheless, the data for these
three test methods agree very well in almost all cases.
The following operating variables have been found to be suitable for
lab-scale column studies: 1) bed diameter of 11 mm and bed height of
110 mm, 2) flow rates of 8 and 16 BV/hr. Due to the high concentrations of
wastes to be adsorbed, smaller beds and faster flows are likely to produce
premature breakthrough.
2-Nitrophenol
Two columns of Amberlite XAD-4, one 12.0 mL BV and the other 24.0 mL
BV, were loaded with 1080 + 50 ppm aqueous NPL buffered at pH 5, at flow
rates of 16 BV/hr and 8 BV/hr respectively. Breakthrough curves are shown
-56-
-------
in Figure 19. The saturation column capacity is 927 mg/g for the 12.0 mL
BV column and 685 mg/g for the 24.0 mL BV column. Each column was re-
generated with methanol as shown in Figure 20. The shorter column re-
generated 739 mg/g and the longer column regenerated 847 mg/g. This
discrepancy is primarily due to experimental procedures which were modified
after the entire nitrophenol study was completed and the results of
isotherm, batch/rate, and column experiments were compared. Thus NPL
capacity data obtained by these three test methods do not correlate well.
The column performance of XAD-4 was similarly measured for removal of
1065 +, 50 ppm NPL from an aqueous stream buffered at pH 8. Breakthrough
curves for the 12.0 mL and 24.0 mL columns which were loaded at 16 BV/hr
and 8 BV/hr respectively, are displayed in Figure 21. Saturation column
capacities are 282 mg/g and 299 mg/g respectively. Methanol regeneration
of these columns (see Figure 22) resulted in regeneration capacities of
272 mg/g and 231 mg/g. Despite the similar saturation capacities of these
two columns, the almost immediate breakthrough of NPL on the shorter column
indicates that the 16 BV/hr flow rate is too rapid. As demonstrated
previously by both isotherm and batch/rate testing, NPL adsorption by XAD-4
is much greater at lower pH where NPL exists in its neutral form.
Column adsorption of NPL by Amberlite XAD-8 was also examined. A
24.0 mL BV column and an 8 BV/hr flow rate were used. Two columns, one
each at pH 5 and 8, were run to determine if NPL adsorption by an acrylic
resin (XAD-8) was as pH sensitive as adsorption by polystyrene resin
(XAD-4). The results of these column loading experiments are displayed in
Figure 23. Saturation column capacity at pH 5 is 247 mg/g and at pH 8 is
124 mg/g. The methanol regenerated capacities of these columns are
332 mg/g and 82 mg/g respectively (see Figure 24). Results indicate that
NPL adsorption by XAD-8 is significantly better at lower pH.
Nitrobenzene
Two columns of Amberlite XAD-4, one 11.0 mL BV and the other 23.1 mL
BV, were loaded with 1805 +_ 30 ppm aqueous NB at flow rates of 16 BV/hr and
8 BV/hr respectively. The results are plotted in Figure 25. The satura-
tion column capacity is 1260 mg/g for the 11.0 mL BV column and 1330 mg/g
for the 23.1 mL BV column. The results of methanol regeneration are shown
in Figure 26. The 11.0 mL bed regenerated 1080 mg/g or 86$ and the 23.1 mL
bed regenerated 1250 mg/g or 94$. The column saturation capacity agrees
well with the batch capacity of 1310 mg/g.
Similarly, a 11.6 mL BV column of Amberlite XAD-8 and a 24.3 mL BV
column of Diaion HP-20 were loaded with aqueous NB at flows of 16 BV/hr and
8 BV/hr respectively. Breakthrough curves are plotted in Figure 27. The
column capacity is 569 mg/g for XAD-8 and 1150 mg/g for HP-20. Results of
methanol regeneration are presented in Figure 28. The XAD-8 column re-
generated 530 mg/g or 93$ and the HP-20 column regenerated 1090 mg/g or
95$. Again column results agree with the batch results of 580 mg/g for
XAD-8 and 1160 mg/g for HP-20.
-57-
-------
UJ
IL 100
S 90
(O
80
Z 70
O
I- 60
CC
, H 50
co UJ
' O 40
O
O 30
H-
§ 20
t 10
uu
INFLUENT CONCENTRATION
1080 ±50 ppm
BED VOLUME: 12ml
FLOW RATE: 16 BV/HR
BED VOLUME: 24 ml
FLOW RATE: 8 BV/HR
345
VOLUME (LITERS)
8
FIGURE 19. COLUMN ADSORPTION OF 2-NITROPHENOL (pH 5) BY AMBERLITE XAD-4
-------
O)
DOC
Oi"
I
01
'IS
^ n
H-
Z
I
CM
1100
1000
900
800
700
600
500
400
300
200
100
24 mi BV
12 ml BV
0.5 1 1.5 2 2.5 3
BED VOLUMES OF METHANOL
3.5
FIGURE 20. COLUMN REGENERATION OF 2-NITROPHENOL (pH 5)
LOADED AMBERLITE XAD-4
-------
I-
Z
UJ
ul 100
z
u. 90
O
^ 80
(0
< 70
O
P 60
<
fE 50
Z
O 40
Z
8 30
g 20
UL 10
0
INFLUENT CONCENTRATION
1065 ±50 ppm
BED VOLUME: 12ml
FLOW RATE: 16 BV/HR
0.5
BED VOLUME: 24 ml
FLOW RATE: 8 BV/HR
1.5 2 2.5
VOLUME (UTERS)
3.5
FIGURE 21. COLUMN ADSORPTION OF 2-NITROPHENOL (pH 8) BY AMBERLITE XAD-4
-------
O)
"ro 25°
e
II
III
2< 20°
DOC
ujUJ 150
a\
*f
100
cc
t
Z 50
0
0.5
1 1.5 2 2.5 3
BED VOLUMES OF METHANOL
12 ml BV
24 ml BV
FIGURE 22. COLUMN REGENERATION OF 2-NITROPHENOL (pH 8)
LOADED AMBERLITE XAD-4
-------
UJ
i 100
z
u. 90
O
* 80
CO
P 60
H 50
UJ
U 40
Z
8 30
=>
u! 10
u.
UJ
0
INFLUENT CONCENTRATION
675 ± 3O ppm at pH 5
790 ± 4O ppm at pH 8
BED VOLUME: 24 ml
FLOW RATE: 8 BV/HR
0.5
1.5 2 2.5
VOLUME (LITERS)
3.5
FIGURE 23. COLUMN ADSORPTION OF 2-NITROPHENOL BY AMBERLITE XAD-8
-------
0)
I
Q_ """"
h-h-
DOC
500
400
OS
Ul
300
200
o
2 100
CM
0
24 ml BV
pH 5
0.5 1 1.5 2 2.5 3
BED VOLUMES OF METHANOL
3.5
FIGURE 24. COLUMN REGENERATION OF 2-NITROPHENOL
LOADED AMBERLITE XAD-8
-------
INFLUENT CONCENTRATION
1805±30 ppm
BED VOLUME: 11.0 ml
FLOW RATE: 16 BV/HR
BED VOLUME: 23.1 ml
FLOW RATE: 8 BV/HR
345
VOLUME (LITERS)
8
FIGURE 25. COLUMN ADSORPTION OF NITROBENZENE BY AMBERLITE XAD-4
-------
o>
7»
E
UL ~"
OQ
^ oc
O uj
31
111 UJ
>oc
Hill
. <2
-------
UJ
i 100
z
u. 90
O
* 80
CO
z 70
O
H 60
<
H 5O
UJ
4O
8 30
2 20
III *W
u! 10
u.
til
INFLUENT CONCENTRATION
1 SOS ±35 ppm
AMBERUTE XAD-8
BED VOLUME: 11.6ml
FLOW RATE: 16 BV/HR
DIAION HP-2O
BED VOLUME: 24.3 ml
FLOW RATE: 8 BV/HR
0.5
1.5 2 2.5 3
VOLUME (LITERS)
3.5
4.5
FIGURE 27. COLUMN ADSORPTION OF NITROBENZENE BY SELECTED ADSORBENTS
-------
0.5
DIAION HP-20
24.3 ml BV
-•- AMBERUTE XAD-8
11.6 ml BV
1 1.5 2 2.5 3
BED VOLUMES OF METHANOL
3.5
FIGURE 28. COLUMN REGENERATION OF NITROBENZENE LOADED ADSORBENTS
-------
1 ,'2-Dichloropropane
Both a 10.4 mL BV and a 23.9 nL BV column of Amberlite XAD-4 were
loaded with 2350 + 80 ppm aqueous PDC at flow rates of 16 BV/hr and 8 BV/hr
respectively. Results are plotted in Figure 29. Integration of the area
above each breakthrough curve yields saturation capacities of 1240 mg/g for
the shorter column and 1310 mg/g for the longer column. In Figure 30, the
results of acetone regeneration are shown. The 10.3 mL column regenerated
1160 mg/g or 94$ and 23.9 mL column regenerated 1150 mg/g or 88$. Batch
results predict a saturation capacity of 1090 mg/g.
An 11.7 mL BV column of Diaion HP-20 and a 27.1 mL BV column of
Amberlite XAD-7 were loaded with aqueous PDC at flows of 16 BV/hr and
8 BV/hr respectively. The data, plotted in Figure 31, yield capacities of
1200 mg/g for HP-20 and 523 mg/g for XAD-7. Acetone regeneration results
are shown in Figure 32. The HP-20 column regenerated 1030 mg/g or 86$ and
the XAD-7 column regenerated 544 mg/g or 104$. The cumulative total of PDC
regenerated from the XAD-7 column appears to exceed the amount of PDC
loaded. This discrepancy is due to the number of samples analyzed during
regeneration (vide supra). The capacities agree well with those predicted
by batch tests: 1055 mg/g for HP-20 and 564 mg/g for XAD-7.
2,4-Dinitrotoluene
To determine the effects of increased flow rates on column per-
formance, two columns having similar bed volumes of Amberlite XAD-4 were
loaded with 178 ± 10 ppm aqueous DOT. Thus, a 10.4 mL BV column was loaded
at 16 BV/hr and a 11.4 mL BV column was loaded at 24 BV/hr. The break-
through curves are displayed in Figure 33. Although saturation column
capacities are identical for these two columns (i.e., 434 mg/g and
432 mg/g) the slower flow produced significantly better performance in the
early portion of breakthrough. To 10$ breakthrough (18 ppm DNT in the
effluent) the 10.4 mL BV column had treated 6.4 L or 615 BV of solution
while the 11.4 mL BV column had treated 5.6 L or 490 BV of solution. This
is a 20$ difference during initial DNT leakage. The results of acetone
regeneration are plotted in Figure 34. The 10.4 mL BV column regenerated
470 mg/g and the 11.4 mL BV column regenerated 474 mg/g. These capacities
are higher than the loaded capacities due to analytical problems which
developed during the GC analysis of aqueous DOT solutions. The GC data for
acetone solutions of DOT are quite reliable and calibration standards
varied only +5$ during the entire DOT study. These saturation capacities
agree well with the XAD-4 batch capacity of 450 mg/g.
To make it possible to correlate batch/rate data to column performance
for carbonaceous adsorbents, a column study using Ambersorb XE-340 was
undertaken. A 10.8 mL BV column of XE-340 was loaded with aqueous DOT at
16 BV/hr to 95$ breakthrough. Column loading data are plotted in Figure
35; acetone regeneration data are shown in Figure 36. As predicted by
batch testing, acetone fully regenerated the XE-340 column. The loaded
capacity of 148 mg/g and the regenerated capacity of 151 mg/g agree satis-
factorily with the batch capacity of 138 mg/g.
-68-
-------
H
UJ
ui 100
z
u- 90
O
O)
80
2 70
O
H 60
cc
, h- 50
vo uj
1 O 40
O
O 30
g 20
u? 10
u.
UJ
O
INFLUENT CONCENTRATION
2350 ±80 ppm
BED VOLUME: 23.9 ml
FLOW RATE: 8 BV/HR
BED VOLUME: 10.4ml
FLOW RATE: 16 BV/HR
0
45
VOLUME (LITERS)
6
8
FIGURE 29. COLUMN ADSORPTION OF 1,2-DICHLOROPROPANE
BY AMBERLITE XAD-4
-------
10.4 ml BV
0.5
1.5 2 2.5
BED VOLUMES OF ACETONE
3.5
FIGURE 30. COLUMN REGENERATION OF 1,2 DICHLOROPROPANE
LOADED AMBERLITE XAD-4
-------
t-
z
UJ
LL 100
O
3*
(0
<
z
O
5
oc
O
O
u.
a.
ui
90
80
70
60
50
40
30
w 20
10
AMBERLITE XAD-7
BED VOLUME: 27.1 ml
FLOW RATE: 8 BV/HR
INFLUENT CONC: 2330 ±70 ppm
OIAION HP-20
BED VOLUME: 11.7ml
FLOW RATE: 16 BV/HR
INFLUENT CONC: 2420±80 ppm
0.5
1.5 2 0.5 1
VOLUME (LITERS)
1.5
2.5
FIGURE 31. COLUMN ADSORPTION OF 1,2-DICHLOROPROPANE
BY SELECTED ADSORBENTS
-------
1100
Q
LU
1000
900
0<
H £ 800
IS
O CD 700
•Ice
2
PQ[ 500
-^ ^Bfc.
400
r »«•
3g
I
300
200
100
DIAION HP-20
11.7 ml BV
AMBERLITE XAD-7
27.1 ml BV
0.5
1.5 2 2.5
BED VOLUMES OF ACETONE
3.5
FIGURE 32. COLUMN REGENERATION OF 1,2-DICHLOROPROPANE
LOADED ADSORBENTS
-------
n! 100
Z
H 90
O
c£ 80
< 70
O
p 60
EC
4, Z
LJ UJ
1 O
Z
O
O
UJ
D
50
40
30
20
£ 10
0
INFLUENT CONCENTRATION
178±10 ppm
BED VOLUME: 11.4ml
FLOW RATE: 24 BV/HR
BED VOLUME: 10.4 ml
FLOW RATE: 16 BV/HR
4
8
10
14
16
VOLUME (LITERS}
FIGURE 33. COLUMN ADSORPTION OF 2,4-DINITROTOLUENE BY AMBERLITE XAD-4
-------
cn
J. 50°
Q
u_ UJ
!-OC 400
Z UJ
DZ
OUJ
55 3°°
P uj
^§3
t 5 p 200
11.4 ml BV
0.5
1.5 2 2.5 3
BED VOLUMES OF ACETONE
3.5
FIGURE 34. COLUMN REGENERATION OF 2,4-DINITROTOLUENE
LOADED AMBERLITE XAD-4
-------
z
UJ
UL 10O
Z
u- 90
CO
iS
80
7O
6O
SO
O 40
O
O 30
u] 20
t 10
INFLUENT CONCENTRATION
178±10ppm
BED VOLUME: 10.8ml
FLOW RATE: 16 BV/HR
4 5
VOLUME (LITERS)
8
FIGURE 35. COLUMN ADSORPTION OF 2,4-DINITROTOLUENE
BY AMBERSORB XE-340
-------
O)
7»
1. 250
O
200
DZ
O UJ
50
150
5 50
4
0.5 1 1.5 2 2.5 3
BED VOLUMES OF ACETONE
3.5
FIGURE 36. COLUMN REGENERATION OF 2,4-DINITROTOLUENE
LOADED AMBERSORB XE-340
-------
Synthetic, Nitrparomatic Stream
The average concentrations of organic pollutants in the multicomponent
stream used for column studies were: 200 ppm toluene, 679 ppm nitro-
benzene, 506 ppm 2-nitrophenol, and 52 ppm 2,4-dinitrotoluene. Although
the synthetic waste stream system (vide supra) provided a fairly con-
stant total waste stream concentration to within 5% of the average, a 12?
fluctuation was observed in toluene concentration, perhaps due to its
greater volatility as compared to the nitroaromatics . Column loading was
continued until NB and NPL leakage approached (or exceeded) 100$ break-
through. Generally, NPL breakthrough exceeded 100/6 before column loading
was terminated. This was due to displacement of adsorbed NPL by other
waste stream components which were more strongly adsorbed by the resins.
The linear equilibrium constants are a measure of this preferential ad-
sorption (see Table 25).
In all studies the order of leakage was NPL, NB, then T. DNT, which
contributed only 4$ to the total influent concentration, was never detected
in the column effluent. Since the lower limit of GC sensitivity to DNT in
this waste stream was 5 to 10 ppm, DNT breakthrough of 10 to 20$ could go
undetected. This was not a problem in the single component DNT study where
the lower limit of GC sensitivity was 1 to 2 ppm. Acetone regeneration
adequately displaced NB, NPL, and DNT from the loaded columns. However,
large discrepancies between loading and regeneration data for toluene
indicate that the chosen regeneration conditions did not effectively remove
adsorbed toluene (vide infra) .
Two 10.6 mL BV columns of Amberlite XAD-4 were loaded with the nitre-
aromatic stream. The breakthrough curves for the column loaded at 8 BV/hr
are displayed in Figure 37 while those for the column loaded at 16 BV/hr
appear in Figure 38. Cumulative saturation capacities are 1025 mg/g
(8 BV/hr column) and 1079 mg/g (16 BV/hr column). Within the +5$ error
limits, doubling the flow rate did not significantly alter saturation
capacity. Capacity to 10$ breakthrough was also not significantly altered.
For example, for the 8 BV/hr column, NB capacity is 376 mg/g and NPL
capacity is 268 mg/g. At the faster flow rate these capacities are
355 mg/g and and 252 mg/g respectively. Results of acetone regeneration of
these columns are shown in Figures 39 and 40. Regeneration appeared to be
complete in 2 BV. However, toluene was never fully regenerated. Cumula-
tive amounts of organic regenerated are 976 mg/g (8 BV/hr column) and
mg/g (16 BV/hr) for an average regeneration efficiency of ca. 93$.
Similarly, two equal length columns of Diaion HP-20 were loaded at two
different flow rates: 12.5 mL BV loaded at 8 BV/hr and 12.2 mL BV loaded
at 16 BV/hr. The breakthrough characteristics of these columns are shown
in Figures 41 and 42. Pressure drop across these columns was a constant
problem. Pumping rates had to be frequently increased in order to maintain
the desired flow through the beds'. Cumulative column saturation capacities
are 849 mg/g (8 BV/hr column) and 896 mg/g (16 BV/hr column). Again,
these flow rates produced nearly identical capacities to 10$ breakthrough:
328 + 6 mg/g for NB and 224 H- 2 mg/g for NPL. Acetone regeneration results
are displayed in Figures 43 and 44. Excluding toluene which was not fully
-77-
-------
UJ
uf 100
z
u! 90
* 80
CO
z 70
O
P 60
H 50
?O 40
Q 30
H
Z on |_
m ^u r
u! 10 -
u.
tu
BED VOLUME: 10.6 ml
FLOW RATE: 8 BV/HR
0,5
1.5 2 2.5
VOLUME (UTERS)
3.5
FIGURE 37. COLUMN ADSORPTION OF NITROAROMATICS BY AMBERLITE XAD-4
-------
•
ill
100 -
u. 90
O
^ 80
I
O
60
-^
[E 50
, Z
-j Hi
?0 40
Z
8 30 -
f-
g 20k
15
i 10 -
u.
w
OL
0
BED VOLUME: 10.6ml
FLOW RATE: 16 BV/HR
0.5
1.5 2 2.5
VOLUME (LITERS)
FIGURE 38. COLUMN ADSORPTION OF NITROAROMATICS BY AMBERLITE XAD-4
-------
o»
7n
•=• 500
Q
ill
400
UJ
(!)
UJ
cc
300
O
o200
>
§
O
100
LOADED AT 8 BV/HR
NITROBENZENE
2,4-DINITROTOLUENE
0.5 1 1.5 , 2 2.5 3
BED VOLUMES OF ACETONE
3.5
FIGURE 39. COLUMN REGENERATION OF NITROAROMATICS
LOADED AMBERLITE XAD-4
-------
CD
7n
~ 500
Q
ui
400
300
Ul
O
UJ
O
*r< 200
UJ
P
100
o
LOADED AT 16 BV/HR
0.5
"• NITROBENZENE
• 2 NITROPHENOL
• TOLUENE
2,4 DINITROTOLUENE
j i i
1.5 2 2.5 3
BED VOLUMES OF ACETONE
3.5
FIGURE 40. COLUMN REGENERATION OF NITROAROMATICS
LOADED AMBERLITE XAD-4
-------
H
ui
UL 100
u- 90
O
80
70
60
50
40
g
cc
z
o
O 30
I-
§ 20
u. 10
u.
Ul
BED VOLUME: 12.5ml
FLOW RATE: 8 BV/HR
0.5
1.5 2 2.5
VOLUME (LITERS)
2-NlTROPHEIMOL
NITROBENZENE
TOLUENE
3.5
FIGURE 41. COLUMN ADSORPTION OF NITROAROMATICS BY DIAION HP-20
-------
UJ
3
tu 1OO
z
4- 90
CO
80
Z 70
2
i- 60
<
cc
H 50
O 40
O
O 30
uS 20
t 10
ill
0
BED VOLUME: 12.2ml
FLOW RATE: 16 BV/HR
0.5
1.5 2 2.5
VOLUME (LITERS)
2-NITROPHENOL
NITROBENZENE
TOLUENE
3.5
FIGURE 42. COLUMN ADSORPTION OF NITROAROMATICS BY DIAION HP-20
-------
O)
7s»
•=• 500
Q
in
"I 400
IU
CD
HI
cc
H- 300
2
D
O
5 200
ui
>
D 100
D
O
LOADED AT 8 BV/HR
0.5
• NITROBENZENE
2-NITROPHENOL
TOLUENE
2.4-DINITROTOLUENE
I I i
1.5 2 2.5
BED VOLUMES OF ACETONE
3.5
FIGURE 43. COLUMN REGENERATION OF NITROAROMATICS LOADED DIAION HP-20
-------
o>
-- 500
Q
ui
£ 400
z
ui
CD
01
tc
H 300
o
s
,< 200
111
=D 100
£
O
LOADED AT 16 BV/HR
• NITROBENZENE
2.4-DINITROTOLUENE
0.5 1 1.5 2 2.5 3
BED VOLUMES OF ACETONE
3.5
FIGURE 44. COLUMN REGENERATION OF NITROAROMATICS LOADED DIAION HP-20
-------
regenerated, regeneration was complete in 2 BV. Regenerated column
capacities are 844 mg/g (8 BV/hr column) and 801 mg/g (16 BV/hr column)
or an average efficiency of ca. 94/f.
To provide additional data for correlating batch/rate data and column
performance for carbonaceous adsorbents, an 11.0 mL BV column of Ambersorb
XE-340 was loaded with the nitro-aromatic stream at 8 BV/hr as shown in
Figure 45. Initial breakthrough of NPL, MB, and T occurred almost con-
currently unlike breakthrough behavior observed with the polymeric resins.
Saturation capacity is 274 mg/g. Regeneration required 4 BV of acetone
(see Figure 46) and resulted in a regeneration capacity of 249 mg/g or 91$
efficiency.
A report from an EPA funded study of priority pollutant treatability
using three technologies (carbon adsorption, resin adsorption, and steam
stripping) is related to this study. A complex industrial waste stream was
used and treatability was assessed by monitoring five priority pollutants:
chlorobenzene, dichlorobenzene, nitrobenzene, dinitrotoluene, and phenol.
The results indicate that both carbon and resin adsorption are feasible
methods of treating the waste stream in question. The authors observed
that resin (XAD-4) capacity improved with several loading/regeneration
cycles OfcJ. This indicates that the resin was not fully conditioned prior
to use in the isotherm tests and column experiments. Thus, reported
capacities were underestimated and treatment costs were probably over-
estimated. This illustrates the importance of carefully pre-conditioning
synthetic adsorbents prior to use to remove residual salts and organics
from their pores.
SUMMARY OF ADSORPTION DATA
The saturation capacities as measured by isotherm, batch and column
studies of various adsorbents in the single component waste streams are
recorded in Table 27. In general, the batch test performs better than the
isotherm test in predicting column saturation capacities for these com-
pounds. It provides a more accurate evaluation of column saturation than
the isotherm test and requires only a single analysis after adsorbent
exposure to the influent waste stream. 2-Nitrophenol data are fairly
scattered and good comparisons are therefore difficult to make. This is
because test method development occurred primarily during the 2-nitrophenol
study. This study was not repeated after the incorporation of improvements
in experimental procedures. Carbon adsorption data are not included here.
Because the batch/rate test measures regenerated capacity, it is not
generally useful in assessing carbon capacity. Experimental procedures for
measuring the capacity of activated carbon have been reported elsewhere[9].
The saturation capacities as measured by batch and column studies of
various adsorbents in the synthetic, nitroaromatic stream are listed in
Table 28. Based on results from single component stream testing, the
isotherm test was eliminated in the 4-component stream study. Since column
saturation was achieved only with nitrobenzene and 2-nitrophenol, batch and
column data can be compared for these two compounds only. Despite fluctua-.
tions in waste stream concentrations (see Table 28), the single analysis
batch test adequately predicts column saturation capacities with the
multicomponent stream.
-86-
-------
Column loading is typically stopped at some predetermined level of
breakthrough such as 10$ leakage rather than at saturation. To provide a
more realistic picture of usable column capacity, both bed volumes of waste
stream treated and column capacities for each component to 1$ and 10%
leakage have been obtained from column loading experiments. These data are
recorded in Table 29 for three single component streams and in Table 30 for
the synthetic, nitroaromatic stream. To test flow rate effects, column
loading of Amberlite XAD-4 was done at two different flow rates with all
four streams. Diaion HP-20 was also used in flow rate studies with the
multicomponent stream. In the single component stream studies, increasing
the flow from 8 to 16 BV/hr (1 to 2 gpm/ft3) or from 16 to 24 BV/hr (2 to
3 gpm/ft3) resulted in approximately 15% capacity loss to 10$ leakage.
In the multicomponent stream studies, a similar increase in flow only
decreased the capacity about 5%. In real applications, these fairly rapid
flows (16 BV/hr) would provide optimum use of column capacity. Since flow
rates depend on the particular adsorbent, waste stream components and
concentrations, system design, and the tolerable level of breakthrough
components, they must be determined for each application. Past applica-
tions of polymeric adsorbents have typically used flows of 2 to 12 BV/hr
(0.25 to 1.5 gpm/ft3) [5,16,17].
-87-
-------
z
III
=>
EL 100
Z
u. 9O
O
* 80
(0
<
Z
O
60
50
fiC
UJ
O 40
Z
8 30
H
Z on
ui *u
10
u.
UJ
0
0
BED VOLUME: 11.0ml
FLOW RATE: 8 BV/HR
0.2
0.4
0.6 0.8 1.0
VOLUME (LITERS)
2-NITROPHENOL
^
NITROBENZENE
FIGURE 45. COLUMN ADSORPTION OF NITROAROMATICS BY AMBERSORB XE-340
-------
- 220
o>
— 200
Q
HI
b 180
^
cc
^ 160
UJ
g 140
oc
fc 120
O 100
2 8°
H 60
13 40
O 20
LOADED AT 8 BV/HR
2.4-DINITROTOLUENE
0.5 1 1.5 2 2.5 3
BED VOLUMES OF ACETONE
3.5
FIGURE 46. COLUMN REGENERATION OF NITROAROMATICS
LOADED AMBERSORB XE-340
-------
i
ID
O
TABLE 27
SUMMARY OF SATURATION CAPACITIES MEASURED BY ISOTHERMS, BATCH TESTS, AND COLUMN EXPERIMENTS
IN SINGLE COMPONENT WASTE STREAMS (mg/g)
Adsorbate3
2-Nitrophenol(pH 5)
2-Nitrophenol(pH 8)
2-Nitrophendl(pH 5)
2-Nitrophenol(pH 8)
Nitrobenzene
Nitrobenzene
Nitrobenzene
1,2-Dichloropropane
1,2-Dichloropropane
1,2-Dichloropropane
2,4-Dinitrotoluene
2,4-Dinitrotoluene
Flow Column Column
Adsorbent Isotherm13 Batch BV(mL) (BV/hr) Loading Regeneration
XAD-4
XAD-4
XAD-8
XAD-8
XAD-4
XAD-8
HP-20
XAD-4
XAD-7
HP-20
XAD-4
1120, 1260
228, 261
592, 704
159, 227
859
456
670
747
409
778
397
518
348
245
145
1310
580
1160
1070
564
1060
450
XE-340
118
138
24.0
12.0
24.0
12.0
24.0
24.0
23.1
11.0
11.6
24.3
23-9
10.4
27.1
11.7
10.4
11.4
10.8
8
16
8
16
8
8
8
16
16
8
8
16
8
16
16
24
16
684
922
299
282
247
124
1330
1260
569
1150
1310
1240
523
1200
434
432
148
847
739
231
272
332
82
1250
1080
530
1060
1150
1160
544
1030
470
474
151
a Aqueous 2-nitrophenol cone. (mg/L): isotherm = 750, 1000; batch = 750 (pH 5), 820 (pH 8);
column = 1070 (XAD-4), 675 (XAD-8, pH 5), 790 (XAD-8, pH 8)
Aqueous nitrobenzene cone. (mg/L): isotherm = 1900; batch = 1900; column = 1800
Aqueous 1,2-dichloropropane cone. (mg/L): isotherm- 2500; batch = 2450; column = 2350
Aqueous 2,4-dinitrotoluene cone. (mg/L): isotherm = 180; batch = 180; column = 180
Capacities calculated from Freundlich parameters assuming linear isotherms.
-------
TABLE 28
SUMMARY OF SATURATION CAPACITIES
SYNTHETIC,
Adsorbent Adsorbatea
Toluene
XAD-4 Nitrobenzene
2-Nitrophenol
2, 4-Dinitrotoluene
Toluene
, XAD-4 Nitrobenzene
jf 2-Nitrophenol
i 2, 4-Dinitrotoluene
Toluene
XE-340 Nitrobenzene
2-Nitrophenol
2, 4-Dinitrotoluene
Toluene
HP-20 Nitrobenzene
2-Nitrophenol
2, 4-Dinitrotoluene
Toluene
HP-20 Nitrobenzene
2-Nitrophenol
2, 4-Dinitrotoluene
MEASURED BY BATCH TESTS AND COLUMN
NITROAROMATIC STREAM (tng/g)
Flow
Batch BV (mL) (BV/hr)
183
575 10.6 8
292
141
183
575 10.6 16
292
141
37
131 11.0 8
72
46
178
525 12.5 8
276
127
178
525 12.2 16
276
127
EXPERIMENTS
Column
Loadingb
181
479
308
57
206
486
335
52
52
125
83
14
170
392
241
46
180
395
248
43
IN THE
Column
Regeneration^
144
512
273
47
127
490
332
45-
31
138
72
8
126
433
243
42
118
406
241
36
CONTINUED
-------
TABLE 28 (CONTINUED)
SUMMARY OF SATURATION CAPACITIES MEASURED BY BATCH TESTS AND COLUMN EXPERIMENTS IN THE
SYNTHETIC, NITROAROMATIC STREAM (mg/g)
Notes
a Ave. waste stream concentration (mg/L) varied from test to test as indicated, below:
_ Test Toluene Nitrobenzene 2-Nitrophenol 2,M-Dinitrotoluene
Batch (all) 239 708 506 M8
, XAD-U (8) 175 686 ' 507 55
i XAD-4 (16) 205 683 522 51
*? XE-340 (8) 220 678 515 51
HP-20 (8) 190 685 *»95 51
HP-20 (16) 210 665 *»90 50
b Column saturation was attained with nitrobenzene and 2-nitrophenol only. Column capacities
for toluene and 2,1-dinitrotoluene are not saturation capacities. 5,
-------
TABLE 29
vo
u>
I
Adsorbent
XAD-4
XAD-4
XAD-8
HP-20
XAD-4
XAD-4
XAD-7
HP-20
XAD-4
XAD-4
XE-340
SUMMARY OF COLUMN CAPACITY DATA FOR THE
Flow
(BV/hr)
8
16
16
8
8
16
8
16
16
21
16
Adsorbate^Ave .
Conc.(mg/L)
NB-1805
NB-1805
NB-1805
NB-1805
PDC-2350
PDC-2350
PDC-2330
PDC-2420
DNT-178
DNT-178
DNT-178
BV to 1$
Leakage
173
143
34
119
105
77
28
64
438
390
210
SINGLE COMPONENT WASTE STREAMS
Capacity at 1$
Leakage (mg/g)
918
743
239
847
647
490
300
581
211
201
68
BV to 10?
Leakage
199
179
49
133
143
115
37
85
617
490
304
Capacity at 10$
Leakage (mg/g)
1056
930
344
947
881
731
394
771
297
253
99
NB - nitrobenzene
PDC - 1,2-dichloropropane
DNT - 2,4-dinitrotoluene
-------
TABLE 30
SUMMARY OF
Flow
Adsorbent (BV/hr)
XAD-4 8
XAD-4 16
£ XE-340 8
i
HP-20 8
HP-20 16
COLUMN CAPACITY DATA FOR THE SYNTHETIC, NITROAROMATIC WASTE STREAM
Adsorbatea » b-Ave
Conc.(mg/L)
T-175
NB-686
NPL-507
T-205
NB-683
NPL-522
T-220
NB-678
NPL-515
T-190
NB-685
NPL-495
T-210
NB-665
NPL-490
i. BV to 1$
Leakage
236
156
142
212
132
123
38
36
200
120
112
205
117
111
Capacity at 1$
Leakage (mg/g)
114
296
199
121
252
179
17
49
35
141
303
204
156
282
197
BV to 10$
Leakage
»377C
198
172
>368°
186
172
82
64
62
»244c
132
124
»238°
133
125
Capacity at 10$
Leakage (mg/g)
»183°
376
268
355
251
34
81
60
»170<2
333
226
»180°
322
222
a Waste stream components: Toluene(T), nitrobenzene(NB), 2-nitrophenol(NPL), and
2,4-dinitrotoluene(DNT)
D Since DNT (influent cone. 50 mg/L) was never detected in the column effluent, it is not included in
this table.
c Column loading of XAD-4 and HP-20 columns was terminated before toluene leakage reached 10$.
-------
SECTION 7
PREDICTIONS OF COLUMN PERFORMANCE USING BATCH/RATE DATA
PREDICTIONS BASED ON THE PERFORMANCE MODEL
Many models for predicting the performance of ion exchange and
carbon adsorption columns from equilibrium and kinetic data have appeared
in the chemical literature over the past 30 to 40 years [18-26]. More
recently, models for predicting column performance of synthetic adsorbents
also have been published [27,28], These models are generally very complex
and require the determination of many experimental parameters as well as
the execution of complicated (often computer-based) calculations. A
simplified model for predicting column performance of ion exchange resins
has been developed at Rohm and Haas. The theoretical basis for this model
is presented in Appendix C. This model can predict performance under a
wide range of operating conditions from a limited amount of experimental
data, i.e., either a single column loading/regeneration experiment or
simple equilibrium and kinetic batch experiments. This model, used as
developed for ion exchange columns, also works satisfactorily for pre-
dicting performance of synthetic adsorbents from data obtained using the
new batch/rate test method. Although refinement of this model may be
desired to improve predictions, in its present form it is extremely at-
tractive for several reasons: 1) only a very limited amount of readily
obtained data is required, 2) data manipulation is simple, 3) the mass
transfer coefficient is determined graphically from simple capacity mea-
surements (Figure C-1), U) column performance can be obtained graphically
(Figure C-2) or from various tables developed for fixed-bed heat transfer,
the J function [18,29,30].
The equations used in the following sections are developed and
defined in Appendix C. The definitions and units for symbols used in these
equations appear in Appendix D.
APPLICATIONS OF THE PERFORMANCE MODEL
An example calculation using the performance model is presented here
for adsorption of 1,2-dichloropropane on Amberlite XAD-4 columns. In order
to predict column performance and develop a concentration history graph or
breakthrough curve, the linear equilibrium constant, K in cc/g, must be
calculated from batch data and the mass transfer coefficient or rate
constant, k in min~^, must be calculated from rate data. In addition to
these two experimentally determined parameters (K and k) several physical/
geometrical parameters of the resin are required, as well as a set of
chosen column operating conditions. These parameters can be measured by
-95-
-------
conventional methods, obtained from the resin manufacturer, or simply
estimated.
Calculation of the Linear Equilibrium Constant
Using experimental batch data for adsorption of PDC by Amberlite XAD-4
(Table 16), the linear equilibrium constant is calculated as follows:
KB = x 10QO = - x 1000 = 440 cc/g
0 Cs 2450 •
To adjust for non-linearity of the equilibrium behavior at high
solution concentrations, an empirical correction factor has been applied to
Kg as follows:
Kg, = 1.1 KB = 480 cc/g
Calculation of the Mass Transfer Coefficient
The various parameters needed to calculate the mass transfer
coefficient are listed in Table 31. The coefficient is calculated using
the following equation:
k = - x 60
r2 T» min
An average value for D/*^' is obtained using the rate data (Table 17)
and the graph in Figure C-1 . To determine D/^-' , a value for the parameter
A is needed. This is easily calculated:
U-£V)sKB,, (1-.45)(1.08)(480)
A = - - & - = - = 634
€ .45
Table 32 is a listing of experimental data and calculated D/T*' results.
is calculated from:
(x-coord value)2(1+A)r>2
Finally:
\5 (1-.29)
k _ - (.45) (1.19 x 10-5) (60) = 5.5
(.025)2
-96-
-------
TABLE 31 '
EXPERIMENTAL PARAMETERS FOR PREDICTING
PDC CONCENTRATION
Parameter
A'
BV
Ca (feed)
fia
fsb
fec •
KB
k
1
R
R'
r (estimated)
Wads
€d
pore vol.d
s
TjjL
HISTORY ON XAD-4
Column A
.95 cm2
10.4 cc
2370 mg/L
.36 .
.34
.30
480 cc/g
5.5 min-1
10.9 cm
2.79 cm/min
2.65 cc/nrLn
.025 cm
3.84g
.45 mL/mL
1. 08 g/ML
.96 mL/g
21.5
1825 t/s
COLUMNS
Column B
.95 cm2
23.9 cc
2350 mg/L
.37
.35
.28
480 cc/g
5.5 min-1
25.2 cm
3-23 cm/min
3.07 cc/min
.025 cm
9.12g
.45 mL/mL
1.08 g/mL
.96 mL/g
42.9
4346 t/s
an =
3fs =
(pore vol.) (Wads)
BV
Wads
(BV) (pa)
cfe = 1 - (fi + fs)
dReference 14
-97-
-------
TABLE 32
CALCULATION OF D/fr* FROM THE RATE OF ADSORPTION
OF 1,2-DICHLOROPROPANE BY XAD-4a
T (sec)
780
1200
1620
2100
3000
3960
Q
Q sat
.487
.557
.636
.679
.688
.694
r(D/r') T ,i/p
""(1 + A) r2J
.168
.200
.238
.263
.268
.272
D>r'(cm2/sec)b
1.44 x 10-5
1.32 x 10-5
1.39 x 10-5
1.31 x 10-5
.95 x 10-5
.74 x 10-5
a Rate data from Table 17
b Ave. value is 1.19 x 10~5 cm2/sec
-98-
-------
Calculation of Breakthrough Curves
As described in Appendix C the theoretical performance of a fixed bed
of adsorbent is calculated from parameters s and t. The experimental
parameters required for such calculations are listed in Table 31.
1. PDC Adsorption by XAD-4 at 16 BV/hr - Column A
„ _ kl ' (5.5X10.9) y. c
3 ' T" " 27?9-*!•:>
t = -£
Rearranging the above in terms of t/s and solving for T:
T fi'l / t fe
T - -TT- f — •+ • -ft?
\
)
since -- =
T =
For correlation with experimental data already presented, T may be ex-
pressed in terms of volume (mL) rather than time
t.3W.30(1.08)(MO)](,o.9)(2.65) f . ,825 f
&.(j S 3
Using Figure C-2, at a particular value of t/s (corresponding to a
time T in mL), and at a given s (21.5 in this case), the concentration
history \can be determined. Graphs of the theoretical versus experimental
breakthrough curves for PDC on a 10.M mL BV column of Amberlite XAD-4
loaded at a flow of 16 BV/hr are shown in Figure 47. Agreement during the
early stages of breakthrough is good. For example, this column is pre-
dicted to treat 1160 mL (111 BV) to 10% leakage and was found to treat
1240 mL (119 BV}. Somewhat more detailed and accurate breakthrough data
can be obtained using the published mass transfer tables [18,29,30 ].
-99-
-------
2. PDC Adsorption by XAD-4 at 8 BV/hr - Column B
kl (5.5X25.2) _
3 ~ TT - 3^21 "
Again, data needed to plot the concentration history are obtained
using Figure C-2. Graphs of the predicted and observed breakthrough curves
for PDC on a 23.9 mL BV column of Amberlite XAD-4 loaded at a flow of
8 BV/hr are shown in Figure 47. This column is predicted to treat
3200 mL (134 BV) to 10? leakage and was found to treat 3360 mL (141 BV).
3. Nitroaromatic Adsorption on XAD-4 Columns
The performance model was also tested with multicomponent stream
data. Attempts at modelling the column performance of each individual
stream component (i.e. toluene, nitrobenzene, 2-nitrophenol, and
2,4-dinitrotoluene) resulted in poor agreement between predicted and
observed column behavior. However, using cumulative component data for
each test, satisfactory agreement between predicted and observed column
performance was obtained. Because DNT contributed less than 5% to the
total waste stream concentration and was never detected in column effluent
(vide supra) it was excluded from all calculations. Thus, cumulative data
(i.e. the summation of T, NB, and NPL results) for batch and rate studies
were used to predict column performance (i.e. cumulative breakthrough data
for T, NB, and NPL).
The experimental parameters for predicting nitroaromatic concentration
history on two XAD-4 columns are listed in Table 33. Column A refers to
the 8 BV/hr study while Column B refers to the 16 BV/hr study. The data
needed to calculate the mass transfer coefficient are recorded in
Table 34. Using batch data, an equilibrium constant of 720 cc/g is cal-
culated; using rate data, a mass transfer coefficient of 12.2 min~^ is
calculated. Application of these values to equations for calculating the
performance of a fixed bed of adsorbent (Appendix C) results in the follow-
ing values for the column performance parameters s and T: 92.5 and
2736 t/s (8 BV/hr); 48.7 and 2749 t/s (16 BV/hr). The concentration
history is obtained for each column by using the graph in Figure C-2.
Plots of the predicted and observed breakthrough curves for nitroaromatics
loading on XAD-4 columns are shown in Figure 48. To 10J leakage, the 8
BV/hr column is predicted to treat 2200 mL (208 BV) and was found to treat
2100 mL (198 BV). Similarly the 16 BV/hr column is predicted to treat
2060 mL (194 BV) to 10$ leakage and was found to treat 1950 mL (184 BV).
-100-
-------
UJ
-S
u_
O
3
Z
O
§
I
til
100
90
80
70
60
50
O
O 30
h-
ui 20
D
it 10
UJ
INFLUENT CONCENTRATION
2350 ± 80 ppm
COLUMN A
COLUMN B
PREDICTED
PREDICTED
OBSERVED
BED VOLUME: 23.9 ml
FLOW RATE: 8 BV/HR
BED VOLUME: 10.4ml
FLOW RATE: 16 BV/HR
2.0
3.0
4.0 5.O 6.0
VOLUME (LITERS)
7.0
8.0
FIGURE 47. CORRELATION OF PREDICTED AND OBSERVED COLUMN ADSORPTION
OF 1,2-DICHLOROPROPANE BY AMBERLITE XAD-4
-------
TABLE 33
EXPERIMENTAL
PARAMETERS FOR
PREDICTING
NITROAROMATIC CONCENTRATION HISTORY IN XAD-4 COLUMNS
Parameter
A'
BV
Ca (feed)a
fi
fs
fe
KBb
k
1
R
R'
r (estimated)
Wads
ec
PSC
pore vol.0
s
TmL
a Ca(feed) = CaT + C
rQsat (T,NB,NPL)
.
Column A
.95 cm2
10.6 cc
1370 mg/L
.35
.33
.32
720 cc/g
12.2 min-1
11.2 cm
1.47 cm/min
1.40 cc/min
.025 cm
3.83g
.45 mL/mL
1.08 g/mL
.96 mL/g
92.5
2736 t/s
a® + CawpL
Column B
.95 cm2
10.6 cc
1410 mg/L
.35
.33
.32
720 cc/g
12.2 min-1
11.2 cm
2.81 cm/min
2.67 cc/min
.025 cm
3.80g
.45 mL/mL
1.08 g/inL
.96 mL/g
48.7
2749 t/s
£CS (T,NB,NPL)
c Reference 14
-102-
-------
TABLE 34
CALCULATION OF D/j1 FROM THE RATE OF ADSORPTION
OF NITROAROMATICS BY
T (sec)
600
900
1530
1860
3300
Qb
Q sat
.474
.610
.714
.742
.768
.-(D/r1) T ,1/2
L(1 + A) r2J
.163
.225
.283
.300
.318
D/T'(cm2/secf
2.60 x 10-5
3.36 x 10-5
3.12 x 10-5 .
2.89 x 10-5
1.83 x 10-5
a Rate data from Table 23
b £Q (T,NB,NPL)
£0 sat (T,NB,NPL)
c Ave. value is 2.76 x 10-5 cm2/sec
-103-
-------
100
Ul
COLUMN A
BED VOLUME: 10.6ml
FLOW RATE: 8 BV/HR
CUMULATIVE CONCENTRATION: 1370 ppm
COLUMN B
BED VOLUME: 10.6ml
FLOW RATE: 16 BV/HR
CUMULATIVE CONCENTRATION: 1410 ppm
1.0
2.0
4.0 0 1.0
VOLUME (LITERS)
2.0
3.0
4.0
FIGURE 48. CORRELATION OF PREDICTED AND OBSERVED COLUMN ADSORPTION
OF NITROAROMATICS BY AMBERLITE XAD-4
-------
CORRELATION OF PREDICTED AND OBSERVED COLUMN PERFORMANCE
Comparisons of predicted and observed column performance for eight
columns of Amberlite XAD-4 are listed in Table 35. Predicted values are
usually within 10? of experimental values at 1? and 10/6 leakage. Consider-
ing the limited number of experimental parameters required and the
simplicity of the calculations, this is very good agreement. The model
adequately predicts column performance of XAD-4 in treatment of single and
multicomponent waste streams.
Based on experimental results, the following physical parameters
can be assumed for future calculations of XAD-4 column performance:
as = 1.08; €= 0.45; fe = 0.32; fi = 0.34; fs = 0.33. Thus, the only
experimental studies required for purposes of predicting column performance
of XAD-4 are batch/rate tests (i.e., to obtain data to calculate Kg and
k). Reasonable estimates of these physical parameters are also required
for predicting co]iimn performance of other synthetic adsorbents. Pre-
liminary experimental evidence indicates that similar bed packing para-
meters (fe, fi, fs) can be used for Diaion HP-20. However, for the acrylic
adsorbents, Amberlites XAD-7 and XAD-8, the parameters appear to be very
different: fi»fsfc0.22 and fe«0.55.
The performance model does not satisfactorily predict column behavior
of the carbonaceous adsorbents. For example, in studies of adsorption of
DMT by Ambersorb XE-340, batch tests yield a Kg of 770 cc/g and duplicate
rate tests yield a k of 1.5 ± O.U min~1. For a 90-95? correlation
between predicted and observed column performance, a k of 5 j* 0.5 min~^
is required. The failure of the performance model in this application is
believed to be due to the different pore structure (i.e. pore size and
distribution) of carbonaceous adsorbents as compared to polymeric
adsorbents [4].
ESTIMATION OF INITIAL COLUMN BREAKTHROUGH FROM BATCH DATA
Although column studies have demonstrated that batch data satis-
factorily predict column saturation capacity, in normal column operation a
loading cycle is terminated significantly before saturation is reached,
usually at 1 to 10? leakage. For both single and multicomponent waste
streams, a rough estimate of "column performance can be made from batch
data alone. For the multicomponent stream, again the components are summed
and treated as a single component (vide supra). Based on data from fifteen
column studies with four different waste streams at flows of 8 or 16 BV/hr
(Table 36), the column capacity and volume treated to 1$ and 10? leakage
can be approximated as follows:
1. Capacity at 1$ leakage = batch capacity X R-|
or
Volume treated to = batch capacity X grams adsorbent X R-|
1? leakage influent concentration
-105-
-------
2. Capacity at 10$ leakage = batch capacity X RIQ
or
Volume treated to = batch capacity X grams adsorbent X RIQ
10% leakage influent concentration
where, for single component streams:
R! = .65 (at 8 BV/hr)
= .50 (at 16 BV/hr)
R10 = .80 (at 8 BV/hr)
= .70 (at 16 BV/hr)
and, for multicomponent streams:
R-l = .55 (at 8 or 16 BV/hr)
R-10 = .70 (at 8 or 16 BV/hr)
Thus, with a single measurement, the batch test, column capacity to
10$ leakage can be roughly estimated to be 70$ of the batch capacity.
-106-
-------
TABLE 35
SUMMARY OF PREDICTED AND OBSERVED COLUMN PERFORMANCE OF AMBERLITE XAD-4
1
o
-J
1
Adsorbate3
Nitrobenzene
Nitrobenzene
1 ,2-Dichloropropane
1 , 2-Dichloropropane
2 , 4-Dinitrotoluene
2 , 4-Dinitrotoluene
Ni troaromatics
Nitroaromatics
KB
(cc/gm)
760
760
480
480
2500
2500
720
720
k
(rain-1 )
5.0
5.0
5.5
5.5
10.5
10.5
12.2
12.2
Column BV
(mL)
11.0
23.1
10.4
23-9
11.4
10.4
10.6
10.6
Flow
BV/hr
16
8
16
8
24
16
16
8
BV Treated
PRED
113
161
77
100
398
497
142
170
to 1$ BT
OBSV
137
175
73
105
389
435
124
151
BV Treated
PRED
163
183
111
134
548
669
194
208
to 10$ BT
OBSV
181
202
119
141
490
615
184
198
a See Tables 29 and 30 for concentrations.
-------
TABLE 36
CORRELATION OF BATCH DATA AND INITIAL LEAKAGE IN
Adsorbate
Nitrobenzene
Nitrobenzene
1 ,2-Dichloropropane
1 ,2-Dichloropropane
Nitroaromatics
Nitroaromatics
Nitroaromatics
Nitrobenzene
Nitrobenzene
1 ,2-Dichloropropane
1 ,2-Dichloropropane
2 , 4-Dinitrotoluene
2 , 4-Dinitrotoluene
Nitroaromatics
Nitroaromatics
2 , 4-Dinitrotoluene
SH-I = (Capacity at 15
Adsorbent
XAD-4
HP-20
XAD-4
XAD-7
XAD-4
HP-20
XE-340
XAD-4
XAD-8
XAD-4
HP-20
XAD-4
XE-340
XAD-4
HP-20
XAD-4
f Leakage)
Flow
(BV/hr)
8
8
8
8
8
8
8
16
16
16
16
16
16
16
16
24
COLUMN
R-I&
.70
.73
.61
.53
.54
.60
.45
.57
.41
.46
.55
.47
.49
.50
.58
.45
EXPERIMENTS
Rmb
.81
.82
.83
.70
.70
.68
.70
.71
.59
.69
.73
.66
.72
.70
.66
.56
. -Batch Capacity
= Volume treated
to 1$ leakage x
influent cone.
Batch capacity x grams adsorbent
- (Capacity at 10% Leakage)
Batch Capacity
= Volume treated to 10$ leakage x influent cone.
Batch capacity x grams adsorbent
The nitroaromatics stream contains toluene, nitrobenzene,
2-nitrophenol, and 2,4-dinitrotoluene.
-103-
-------
REFERENCES
1. I. H. Suffet and M. J. McGuire, Ed., Activated Carbon Adsorption of.
Organics from Aqueous Phase, Vol. 1, Ann Arbor Science Publishers,
Ann Arbor U980).
2. J. W. Neely and E. G. Isacoff, The Use of Carbonaceous Adsorbents
for Treatment of Ground and Surface Waters, Marcel Dekker, Inc., New
York (1982).
3. R. Kunin, "Porous Polymers as Adsorbents - A Review of Current
Practice", Amber-hi-Lite No. 163, Rohm and Haas Company, Philadelphia
., (1980).
4. Ambersorb Carbonaceous Adsorbents, Publication IE-231, Rohm and Haas
Company, Philadelphia (1980).
5. Summary Bulletin, Amberlite Polymeric Adsorbents, Publication IE-172,
Rohm and Haas Company, Philadelphia (1979).
6. B. R. Kim, V. L. Snoeyink, and F. M. Saunders, "J. Water Poll. Cont.
Fed.", j»8f 120 (1976).
7. J. E. Millar, "J. Poly. Sci., Poly. Symp. 68", 167 (1980).
8. J. R. Perrich, Ed., Activated Carbon Adsorption for Wastewater
Treatment, CRC Press, Inc., Boca Raton (1981).
9. R. A. Dobbs and J. M. Cohen, Carbon Adsorption Isotherms for Toxic
Organics. EPA Report 600/8-80-023 (1980).
10. E. H. Crook, R. P. McDonnell, and J.T. McNulty, "Ind. Eng. Chem.,
Prod. Res. Dev.", Jj», 113 (1975). '
11. F. X. Pollio and R. Kunin, "Env. Sci. and Tech.", J_, 160 (1967).
12. H. E. Wise and P. D. Fahrenthold, "Env. Sci. and Tech.", 15, 1292
(1981). ~~
13. Kirk-Othmer Encyclopedia of Chemical Technology, 3rd Ed., Wiley-
Interscience, New York (1980).
14. Technical Bulletin, Amberlite XAD-4. Publication IE-159-69, Rohm and
Haas Company, Philadelphia (1978).
-109-
-------
15. D. Ruggiero and R. Ausubel, Removal of Organic Contaminants from
Drinking Water Supply at Glen Cove, NY. Phases 1 and 2,USEFA
Cooperative Agreement CR-tiOb355-01, US EPA, Cincinnati (1980, 1981).
16. J. K. Smith, R. J. Planchet, E. J. Westbrook, and F. J. Zak,
Laboratory Studies of Priority Pollutant Treatability, EPA Report
600/2-81-129 (1981).
17. R. M. Carlyle, "Eff. and Water Treat. J.", (Feb. 1982).
18. F. G. Helfferich, Ion Exchange. McGraw-Hill, New York (1962).
19. T. Vermeulen and N. Hiester, "Ind. Eng. Chem.", 44, 636 (1952).
20. N. Hiester and T. Vermeulen, "Chem. Eng. Prog.", 48, 505 (1952).
21. R. Dedrick and R. Beckmann, "Chem. Eng. Prog., Symp. Ser. 74",
63, 68 (1967).
22. K. Hall, L. Eagleton, A. Acrivos, and T. Vermeulen, "Ind. and Eng.
Chem. Fund.", 5, 212 (1966).
23. B. A. Bell and A. H. Molof, "Water Res.", .9, 857 (1975).
24. T. Keinath and W. Weber, J. Water Poll. Cont. Fed.", 40, 741 (1968).
25. W. Weber and M. Pirbazari, Effectiveness of Activated Carbon for
Removal of Toxic and/or Carcinogenic Compounds from Water Supplies,
EPA Report 600/S2-81-057 (1981).
26. W. Weber and M. Pirbazari, "J. Am. Water Works Assoc.", 203 (April,
1982).
27. B. van Vliet, W. Weber, and H. Hozumi, "Water Res.", J4., 1719 (1980).
28. B. van Vliet and W. Weber, "J. Water Poll. Cont. Fed.", 53, 1585
(1981). ~~
29. S. Brinkley, Report 3172, US Dept. Int., US Bureau of Mines (1951).
30. A. Opler and N. Hiester, "Tables for Predicting the Performance of
Fixed Bed Ion Exchange and Similar Mass Transfer Processes", SRI,
Stanford (1954).
-110-
-------
APPENDIX A
ABBREVIATIONS
BV - bed volumes
BV/hr - bed volumes per hour
cc - cubic centimeters (also cm3)
cc/g - cubic centimeters per gram (also cm3/g)
cm - centimeters
cm/min - centimeters per minute
GC - gas chromatograph
. g — grams: . ..
gal •? gallon
gpm/ft3 - gallons per minute per cubic foot
L - liters
L/g-mLn - liters per gram per minute
nrVg - square meters per gram
mg — nri \ H gpama
tng/g - milligrams per gram
mg/g-fflin - milligrams per gram per minute
mg/L - milligrams per liter
mL - millimeters
mL/nrLn - milliliters per minute
mm - milliliters
A^g - micrograms
^g/L - micrograms per liter
ppb - parts per billion (also^g/L)
ppm T parts per million (also mg/L)
rpm -r revolutions per minute
sq in - square inches
CIS f cis-1,2-rdichlorQethylene
DNT. - 2,U-dinitrotoluene
ETH - 1,1,1-trichloroethane
NB.- nitrobenzene
NPL - 2-nitrophenol
.PDC -. 1,2-dichloropropane
TETRA- - tetrachloroethylene
TRI - trichloroethylene
-111-
-------
APPENDIX B
ADSORBENT CONDITIONING
1. Place the resin in a large beaker, add an equal volume of methanol
(acetone), and allow the resin to soak for 30 minutes with occasional
stirring.
2. Slurry the resin into an appropriately sized column and allow it to
settle.
3. Pass deionized water upflow through the column at a rate sufficient
to give 5Q% bed expansion.
4. Backwash with deionized water for 30 minutes or until complete
classification of the bed is achieved (small beads at the top and large
beads at the bottom). Allow the resin to settle.
5. Drain the water to 1 inch above the resin bed and pass 10 bed volumes
of methanol (acetone) at 8 BV/hr downflow through the column.
6. Pass 20 bed volumes of deionized water downflow at a rate of 16 BV/hr.
7. Thoroughly drain the bed and blow 100$ relative humidity air through
the bed to remove as much excess water as possible.
8. Store the conditioned resin in a clean, tightly capped bottle.
-112-
-------
APPENDIX C
THEORETICAL APPROACH TO COLUMN PERFORMANCE MODELING
USING BATCH/RATE DATA
DETERMINATION OF THE LINEAR EQUILIBRIUM CONSTANT
Synthetic adsorbents exhibit a Langmuir-type equilibrium behavior
which can be mathematically described as
Qa c>Qa
a max
where qa is the concentration of a in the solid phase, qa is the
concentration of a in the liquid phase, qa uj^ is the maximum value
of qa, and p is the Langmuir equilibrium constant.
Over a range of low concentrations, where iqa«1, the equilibrium
relation is approximately linear and
Qa =
where K(cc/g) is the linear equilibrium constant. With the high concentra-
tions of typical waste streams (qa MOO ppm), empirical factors can be
applied to correct for non-linearity.
From the standpoint of performance, the equilibrium behavior of a
"good" adsorbent would be described by a large value of K. Using data from
typical isotherm experiments, the linear equilibrium constant can be
determined from the equilibrium uptake, by a known amount of adsorbent, in
a solution of known volume and concentration by:
r Va (Can -
Cs«* - Cao)
All symbols used in column performance modeling are defined in Appendix D.
Using batch data (regenerated batch equilibrium capacity) which is obtained
for uptake from a solution of infinite volume and known concentration, the
linear equilibrium constant Kg is determined by:
KB = x 1000 cc/L
-113-
-------
If batch data must be obtained from a solution of limited volume, the
equilibrium constant is modified and determined by:'
K W: V
B wads
DETERMINATION OF THE MASS TRANSFER COEFFICIENT
The rate limiting step in achieving equilibrium in these systems can
be any one or a combination of the following steps involved in the process:
1. Diffusion of material from the external fluid
stream to the surface of the adsorbent bead
(external mass transfer).
2. Diffusion of material from the surface of the
adsorbent into the inner pore structure
(internal mass transfer).
3. Sorption at the fluid-solid interface.
4. Diffusion of material through the solid
phase (surface diffusion).
Experimental evidence indicates that for synthetic adsorbents, the
rate limiting step is internal mass transfer. Sorption at the fluid-solid
interface is extremely rapid and, in fact, local equilibrium can be as-
sumed. Furthermore, for reasonable flow rates, the external mass transfer
resistance is small compared to the internal resistance.
The mass transfer in a particular adsorbent can be described in terms
of a mass transfer coefficient k (min~1) as defined by:
Mass Transfer = k (Ca - qa)
For the case of internal mass transfer limiting the rate:
k = J§-U=<
The critical parameter that must be determined to estimate k is
i. This quantity is obtained experimentally from the rate of mass
transfer into the adsorbent from a rapidly stirred solution of infinite
volume (the rate test). Rate data are used in the form of fractional
uptake (Q/Qsat) as a function of time, T. In Figure C-1, these data lie
on the y-coordinate. When the rate test is performed in a solution of
infinite volume, the x-coordinate parameter
D T
T"
(1 -f Air2
1/2
-114-
-------
is determined using the 0% curve which denotes no change in solution
concentration. If the test must be done in a solution of limited volume,
one of the alternate curves which best corresponds to the final fractional
uptake of adsorbate is used.
The parameter A is a function of the linear equilibrium constant. It
appears in the calculation of the mass transfer coefficient because both
diffusion and linear equilibrium are occurring simultaneously.
A =
THEORETICAL BEHAVIOR OF A FIXED BED OF ADSORBENT
". The performance of a fixed bed of adsorbent can be described mathe-
matically, subject to the following assumptions:
1.. uniform bed geometry °
2. uniform feed (concentration, flow rate,
• :>,..' temperature)
3. ideal plug flow outside the adsorbent with
negligible axial dispersion (no flow within the
adsorbent).
Under these conditions the governing equation can be obtained from a
mass balance over an infinitesimal volume element A'dl over an infini-
tesimal time interval T.
R RA'CCali-Call-Kil) '=
T
IT
I
idl
Assuming rapid, linear equilibrium such that qa = Kqa, the
equation simplifies to:
where fi' = f i + fs
ns K
-115-
-------
1.0
2
E
GQ
g
2
o
I
o
cc
0.
Q.
O
o
<
cc
u.
w
U
OT
U
SOLUTION OF INFINITE VOLUME
0.0 0.1 0.2 0.3 0.4 0.5 0.6
0.0
FIGURE C-1. SIMULTANEOUS DIFFUSION AND LINEAR
ADSORPTION INTO A POROUS SPHERE
-116-
-------
Letting t' s T-lfe/R and assuming the transfer rate is proportional to
both.a constant mass transfer coefficient (k) and the difference in average
concentrations inside and outside the adsorbent, we obtain
This is the governing partial differential equation with the following
boundary conditions:
ca = ca feed for T X) at 1 = 0
' <0
It can be transformed into conventional dimensionless form to utilize
the published solutions for fixed beds, [18, 20]
where X= 1 at s = 0; CJ= 0 at t = 0
using the transformations:
CJ=
f eed " Cao
r' = 1
kl
s. = TT
kt'
A.is the solution function giving the concentration history, s is the
column-capacity parameter, t is the solution-capacity parameter.
-117-
-------
t/s is the throughput parameter. The generalized dimension!ess
solutions to these equations are [18,20,30]
. fa (-t-z) __
A= 1- / e Io(2Vtz )dz
o
ft (-s-z) _.
CJ= / e Io(2 Vsz )dz
where IQ is a modified Bessel function of the first kind. Helfferich
[18] and others [29,30] present numerical solutions to the integrals in
these equations for a wide range of t and s. The results are presented
more conveniently in Figure C-2 and can be used to predict effluent concen-
tration ( Ca f .) for any value of time (T) given the following
parameters:
1. Operating variables: Ca _-. (influent concentration)
R (linear flow rate)
2. Physical parameters: fe, fi, fs (bed packing parameters)
PS (adsorbent density)
£ (adsorbent porosity)
3. Equilibrium parameter: KB
4. Mass transfer parameter: k
In characterizing the performance of a column, the adsorbent's
physical parameters can be measured or simply estimated. After the com-
pletion of a representative number of column experiments using the various
adsorbents available, a table of appropriate physical parameters can be
included as part of column performance prediction methodology. The
equilibrium and mass transfer parameters are determined from the batch/rate
test as previously described.
-118-
-------
1.0
0.50
0.10 -
; o.os -
0.01
FIGURE C-2. GENERALIZED DIMENSIONLESS SOLUTIONS FOR HEAT TRANSFER AND
OTHER MASS TRANSFER PROCESSES
-------
APPENDIX D
SYMBOLS USED IN COLUMN PERFORMANCE MODELING
A' = bed cross sectional area
BV = bed volume of classified adsorbent (mL)
Ca = concentration of a in external channels - average (mg/L)
Ca - ^ = feed or influent concentration (mg/L)
Cao = initial Ca, prior to adsorbate exposure (mg/L)
Ca = Ca at equilibrium (mg/L)
Cs = solution concentration (mg/L)
Cso = initial solution concentration (mg/L)
CSoo • = solution concentration at equilibrium (mg/L)
D = diffusion coefficient
fe = external void fraction (bed void fraction)
fi = internal void fraction (bead void fraction)
fs = solid fraction
K = linear equilibrium constant (cc/g)
Kg = batch test linear equilibrium constant (cc/g)
KB> = modified batch test linear equilibrium constant (cc/g)
k = mass transfer coefficient (min~1)
IT = total bed length (cm)
1 = distance from top of bed (cm)
qa = concentration of a in adsorbent pores - average (mg/L)
qa = concentration of a in solid phase (mg/L)
Qa max = Qa ^ ^ - >0° (ng/g)
Q = adsorbent capacity (mg/g)
Qsat = adsorbent saturation capacity (mg/g)
R = flow rate into bed per unit cross-sectional area (cm/min)
R1 = flow rate into bed (cc/min)
r = radius of adsorbent bead - average (cm)
s = column-capacity parameter
T = time (sec)
t = solution-capacity parameter
t/s = column throughput parameter
-120-
-------
Vs = solution volume (mL)
^ads = adsorbent volume (mL)
wads = adsorbent weight (g)
fi - Langmuir equilibrium constant
£ = absolute adsorbent porosity (mL pore/mL adsorbent)
pa = adsorbent skeletal density (g/mL)
T1 = tortuosity
-121-
-------
APPENDIX E
GLEN COVE PILOT PLANT DATA
The following Tables and Figures have been reproduced from Ref-
erence 2, Chapter 6, with the authors' permission. The data were obtained
during an EPA funded pilot plant study (U.S.E.P.A. No. CR 806355-01)
investigating treatment methods for Glen Cove City groundwater contaminated
by chlorinated organics.
-122-
-------
TABLE E-1
to
co
I
GLEN COVE ADSORPTION SUMMARY
Cycle
Column
1
1
1
Adsorbent Number
ICI Carbon
ICI Carbon
ICI Carbon
1
1
1
Flow Rate
2
2
2
gpm/ft3
gpm/ft3
gpm/ft3
< Steam
1
1
1
2
2
2
ICI Carbon
ICI Carbon
ICI Carbon
Calgon Carbon
Calgon Carbon
Calgon Carbon
2
2
2
1
1
1
2
2
2
2
2
2
gpm/ft3
gpm/ft3
gpm/ft3
gpm/ft3
gpm/ft3
gpm/ft3
< Steam
2
2
2
3
3
3
Calgon Carbon
Calgon Carbon
Calgon Carbon
Ambersorb XE-340
Ambersorb XE-340
Ambersorb XE-340
2
2
2
1
1
1
2
2
2
4
4
4
gpm/ft3
gpm/ft3
gpm/ft3
gpm/ft3
gpm/ft3
gpm/ft3
< Steam
3
3
3
Ambersorb XE-340
Ambersorb XE-340
Ambersorb XE-340
2
2
2
4
H
4
gpm/ft3
gpm/ft3
gpm/ft3
a
Compound
cis
tri
tetra
Regeneration >
cis
tri
tetra
cis
tri
tetra
Regeneration >
cis
tri
tetra
cis
tri
tetra
Regeneration >
cis
tri <1 ppb
tetra <1 ppb
Bed Volumes
to 10$ Leakage
(Leakage ppb)
15
25
35
5
14
>35
17
38
>49
3
17
>21
27
105
121
38
after 46
after 46
,000
,700
,300
,000
,200
,300
,300
,400
,000
,000
,000
,000
,000
,000
,000
,400
,000
,000
(4)
(17)
(6)
(4)
(15)
(9)
(4)
(17)
(6)
(15)
(12)
(5)
(17)
(7)
(7)
(13)
(17)
CONTINUED
13
37
>92
45
100
130
8
44
>55
35
137
158
50
>60b
-------
TABLE E-1 (CONTINUED)
Column
4
4
4
5
5
5
5
5
5
5
5
5
5
5
GLEN COVE ADSORPTION SUMMARY
Cycle
Adsorbent Number
Ambersorb
Ambersorb
Ambersorb
Ambersorb
Ambersorb
Ambersorb
XE-340
XE-340
XE-340
XE-340
XE-340
XE-340
1
1
1
1
1
1
Flow Rate
4
4
4
2
2
2
gpm/ft3
gpm/ft3
gpm/ft3
gpm/ft3
gpm/ft3
gpm/ft3
< Steam
Ambersorb
Ambersorb
Ambersorb
< —
Ambersorb
Ambersorb
Ambersorb
XE-340
XE-340
XE-340
-This
XE-340
XE-340
XE-340
2
2
2
Column
3
3
3
2
2
2
was
4
4
4
gpm/ft3
gpm/ft3
gpm/ft3
a
Compound
cis
tri
tetra
cis
tri
tetra
Regeneration >
cis
tri 4.1 ppb
tetra 1.4 ppb
Bed Volumes
to 10$ Leakage
(Leakage ppb)
31
69
>77
40
80
83
23
after 23
after 23
,000
,000
,000
,000
,000
,000
,000
,000
,000
(7)
(13)
(7)
(4)
(20)
(6)
(6)
(17)
(9)
Days
40
90
>100
104
208
216
60
>60
>60
stopped prematurely and then Steam Regenerated >
gpm/ft3
gpm/ft3
gpm/ft3
< Steam
Ambersorb
Ambersorb
Ambersorb
XE-340
XE-340
XE-340
n
4
4
4
4
4
gpm/ft3
gpm/ft3
gpm/ft3
cis
tri
tetra
Regeneration >
cis
tri
tetra
15
69
77
28
64
73
,000
,000
,000
,000
,000
,000
(4)
(17)
(6)
(5)
(16)
(11)
20
90
100
36
83
95
cis = cis-1,2-dichloroethylene
tri = trichloroethylene
tetra = tetrachloroethylene
Still in progress
-------
FIGURE E-l
COLUMN 81 - 4gpm/ft3 - CIS 1,2 DICHLOROETHYLENE LEAKAGE
£t
Q.
a.
UJ
O
ui Z
I UJ
X
UJ
O
cc
O
X
O
O
120
110
100
90
80
70
60
50
40
30
20
10
2.5 LITERS OF AMBERSORB XE-340
GLEN COVE WATER AT 4 gpm/fP
O INFLUENT
D EFFLUENT
0 10 20 30 40 50 60 70
BED VOLUMES (THOUSANDS)
80
90
100
-------
FIGURE E-2
a.
a.
•teX
HI
O
UJ
UJ
280
260
240
220
200
180
160
140
;E 120
UJ
§ 100
o
X
o
DC
80
60
40
20
COLUMN #1 - 4gpm/ft' - TRICHLOROETHYLENE LEAKAGE
2.5 LITERS OF AMBERSORB XE-340
GLEN COVE WATER AT 4 gpm/ft*
O INFLUENT
D EFFLUENT
•o—o-r-cH
o
10
20
30
40
50
60
70
80
90
100
BED VOLUMES (THOUSANDS)
-------
FIGURE E-3
COLUMN 81 - 4gpm/ft3 - TETRACHLOROETHYLENE
120
a 110
a
. V**
g 100
UJ
I UJ
i" 2
UJ
O
cc
O
I
O
<
cc
UJ
90
80
70
60
50
40
30
20
0
2.5 LITERS OF AMBERSORB XE-340
GLEN COVE WATER AT 4 gpm/ft?
O INFLUENT
D EFFLUENT
10
20
30 40 50 60 70
BED VOLUMES (THOUSANDS)
90
100
------- |