Environmental Protection Technology Series
ACTIVATED CARBON ADSORPTION OF
TRACE ORGANIC COMPOUNDS
Municipal Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports -
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-77-223
December 1977
ACTIVATED CARBON ADSORPTION OF
TRACE ORGANIC COMPOUNDS
by
Vernon L. Snoeyink, John J. McCreary and Carol J. Murln
Department of Civil Engineering
University of Illinois
Urbana, Illinois 61801
Grant No. R 803473
Project Officer
Alan A. Stevens
Water Supply Research Division
Municipal Environmental Research Laboratory
Cincinnati, Ohio 45268
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for publica-
tion. Approval does not signify that the contents necessarily reflect the
views and policies of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
-------
FOREWORD
The Environmental Protection Agency was created because of increasing
public and government concern about the dangers of pollution to the health
and welfare of the American people. Noxious air, foul water, and spoiled
land are tragic testimony to the deterioration of our natural environment.
The complexity of that environment and interplay between its components
require a concentrated and integrated attack on the problem.
Research and development is that necessary first step in problem
solution and it involves defining the problem, measuring its impact, and
searching for solutions. The Municipal Environmental Research Laboratory
develops new and improved technology and systems for the prevention, treat-
ment, and management of wastewater and solid and hazardous waste pollutant
discharges from municipal and community sources, for the preservation and
treatment of public drinking water supplies, and to minimize the adverse
economic, social, health, and aesthetic effects of pollution. This
publication is one of the products of that research; a most vital communica-
tions link between the researcher and the user community.
The ability of activated carbon to adsorb certain types of organic
compounds has frequently been demonstrated. In this publication the results
of research on the application of activated carbon to reduce the levels of
organic contaminants in drinking water is examined.
Francis T. Mayo, Director
Municipal Environmental Research
Laboratory
-------
ABSTRACT
This research program was conducted to determine how effectively humic
substances and the trace contaminants 2-methylisoborneol, geosmin, the
chlorophenols and polynuclear aromatic hydrocarbons were adsorbed by acti-
vated carbon under the competitive adsorption conditions encountered in
natural waters. Data were collected using isotherm tests and small-scale
laboratory columns.
Humic substances were obtained from a commercial source, a well water,
leaf extract, and soil extract, and some of the materials were separated
into the humic acid and fulvic acid fractions. The molecular weight distri-
butions of the fractions were determined by gel permeation chromatography
and ultrafiltration. Significant differences in the adsorbability, haloform
formation potential, and fluorescence of the various fractions were observed.
A procedure for easily synthesizing the earthy-musty odor-causing com-
pound 2-methylisoborneol (MIB) was developed, and a gas chromatographic
analytical technique for quantitative analysis of MIB and geosmin down to
0.1 yg/1 was successfully formulated and tested. Humic substances compete
with MIB and geosmin for adsorption sites on activated carbon and signifi-
cantly reduce its capacity for these compounds. These naturally occurring
odorous compounds were found to be much more strongly adsorbed than the
humic substances.
Both the chlorophenols and the polynuclear aromatic hydrocarbons are
very strongly adsorbed. Strong competition was observed between anionic
and neutral species of 2,4-dichlorophenol and 2,4,6-trichlorophenol, even
at the 1 yg/1-concentration level. The Langmuir model for competitive
adsorption, or the Jain and Snoeyink modification of that model, conformed
well to the observed data, with one exception at pH 7.0. The presence of
the various humic substances also caused a significant reduction in chloro-
phenol adsorption capacity. Also, humic acid did not interfere with the
rate of adsorption of a model polynuclear aromatic hydrocarbon, anthracene.
This report was submitted in fulfillment of EPA Grant No. R 803473 by
the University of Illinois under the sponsorship of the U.S. Environmental
Protection Agency. This report covers the period January 6, 1975, to
July 5, 1977, and the work was completed July 5/1977.
-------
CONTENTS
Foreword iii
Abstract iv
Figures vi
Tables ix
Acknowledgment x
1. Introduction 1
Statement of Problem 1
Objectives 1
Adsorbate Characteristics and Background Information . . 2
Approach to the Study 8
2. Conclusions 9
3. Recommendations 12
4. Materials and Methods 13
Adsorbents 13
Humic Substances 13
2-Methylisoborneol, Synthesis and Analysis 23
Geosmin Source and Analysis 31
Chlorophenols 32
Polynuclear Aromatic Hydrocarbons 33
Adsorption Test Procedures 36
5. Results and Discussion 41
Haloform Formation Potential of the Humic Substances . . 41
Adsorption of Humic Substances 41
MIB synthesis and Analysis 53
Adsorption of MIB and Geosmin 65
Adsorption of Chlorophenol 76
Adsorption of Polynuclear Aromatic Hydrocarbons 107
References 110
-------
FIGURES
Number Page
1 Reaction scheme for the chlorination of phenol 6
2 Upflow column for extracting humic material 14
3 Organic carbon separation monitored by TOC analysis 16
4 Typical chloroform chromatogram 22
5 Concentration apparatus 28
6 Standard curve for 2,4-dichlorophenol 34
7 Typical chlorophenol chromatogram 35
8 Typical anthracene chromatogram 37
9 Chloroform formation from humic and fulvic acids 40
10 Chloroform formation from soil fulvic acid 42
11 Chloroform formation from soil humic acid 43
12 Adsorption isotherms for commercial humic acid measured by
fluorescence and UV 44
13 Adsorption of well water organic matter 46
14 Effect of phosphate buffer concentration on adsorption of
soil fulvic acid 47
15 Effect of pH on the adsorption of soil fulvic acid 48
16 Adsorption of various types of humic substances 49
17 Adsorption of molecular weight fractions of soil fulvic acid . . 50
18 Adsorption of molecular weight fractions of soil humic acid ... 51
19 Adsorption of molecular weight fractions of leaf fulvic acid . . 52
20 Stereochemical structures of 2-methylisoborneol and camphor
enantiomers 54
vi
-------
Number Page
21 Chromatogram from the product of the action of
methyl!ithiurn on d-camphor 56
22 Chromatogram from the product of the action of methyl!ithium
on d-camphor after treatment with alkaline hydroxylamine ... 57
23 Chromatogram from the residual aqueous solution from the
hydroxylamine reaction after hexane extraction 58
24 Chromatogram from the hexane extract after the
hydroxylamine reaction 59
25 Chromatogram from the hexane extract in Figure 24 after
exhaustive washing with 2 N sodium hydroxide 60
26 Chromatogram obtained in a study of the recovery of MIB
added to tap water at 0.1 pg/1 66
27 Adsorption of MIB 67
28 Adsorption of geosmin 69
29 Column breakthrough curves for MIB and geosmin 70
30 Column breakthrough curve for geosmin 71
31 Column breakthrough curve for MIB 73
32 Adsorption of MIB--nonequilibrium 75
33 Adsorption isotherms for 2,4-dichlorophenol 77
34 Adsorption isotherms for 2,4,6-trichlorophenol 78
35 Influence of pH on chlorophenol adsorption capacity 79
36 Competitive adsorption capacities predicted by the Langmuir
model for dichlorophenol at pH 5.2 96
37 Competitive adsorption capacities predicted by the Langmuir
model for trichlorophenol at pH 5.2 97
38 Competitive adsorption capacities predicted by the Langmuir
model for dichlorophenol at pH 7.0 99
39 Competitive adsorption capacities predicted by the Langmuir
model for trichlorophenol at pH 7.0 100
40 Competitive adsorption capacities predicted by the Jain
model for dichlorophenol at pH 9.1 101
vii
-------
Number Page
41 Competitive adsorption capacities predicted by the Jain
model for trichlorophenol at pH 9.1 102
42 Kinetics of anthracene adsorption with and without humic acid . . 108
-------
TABLES
Number Page
1 Gel Molecular Weight Exclusion Limits and Ultrafiltration
Membrane Molecular Weight Cut-Off Limits 17
2 Finch Soil Humic Substance Fractionation 18
3 Evaluation of G-50 Coarse Sephadex Fractions of Soil Humic
Acid by Ultrafiltration 19
4 Fluorescence Intensity of 5 mg/1 TOC Fractions of Humic
Substances 20
5 Reagents, Solvents and Adsorbents for MIB Synthesis and Analysis. . 23
6 TLC and MIB and Some Other Camphor Derivatives 25
7 Isolation of MIB by Column Chromatography of a Natural Product . . 26
8 Characteristics of the Chloro- and Bromophenols Studied 32
9 Response Factor of MIB for Various Proportions with the
Internal Standard and Amounts Gas Chromatographed 64
10 Relative Recovery of MIB from Various Waters 65
11 Values of Constants for Equation 3 for Chlorophenol Single
Solute Data 81
12 Chlorophenol Competition at pH 5.2 83
13 Chlorophenol Competition at pH 7.0 87
14 Chlorophenol Competition at pH 9.1 91
15 2,4,6-TCP Competition with Humic Substances - pH 5.2 104
16 2,4,6-TCP Competition with Humic Substances - pH 9.1 105
17 Concentrations of Benzanthracene after Eight Hours 107
-------
ACKNOWLEDGMENT
The assistance of Dennis Beckmann, Paul Boening, David Dunn, Dennis
Herzing, Terry Temperly and Neville Wood in producing material for portions
of this report is acknowledged. The adsorbents used in this study were
supplied by the manufacturers.
-------
SECTION 1
INTRODUCTION
STATEMENT OF PROBLEM
Odor problems plague the majority of water treatment plants. Granular
activated carbon (GAC) beds now are being used in the U.S.A. predominantly
for the removal of odor-causing compounds and are reported to be effective
for certain types of odors of biological origin for up to 3 to 5 years
(Love et al., 1973). Although only limited data have been reported on the
use of GAC for removal of odors of predominantly industrial origin (Postal
et al., 1965), bed life may be much reduced. Unfortunately very few data
are available to indicate the mechanism of removal and to enable the applica-
tion of findings at one water treatment plant to odor problems at other
locations. It also is not possible to say whether selected trace organics
will adsorb similar to the odor compounds or whether they will rapidly
saturate the GAC bed thus making thermal regeneration necessary.
Humic substances in water supplies are also of much concern. These
materials react with chlorine to produce haloforms (Rook, 1976; Stevens
et al., 1976; Symons, 1976), they occupy adsorption sites on the carbon
surface thereby reducing the adsorption capacity for selected trace compounds
(Herzing et al., 1977), and they associate with metal ions, pesticides,
phthalates and possibly other organics (Schnitzer and Khan, 1972). Humic
substances constitute the major fraction of organics in most natural waters.
There exists little information on the magnitude and nature of the competi-
tive effects of humic substances, or information on how humic materials from
various sources can be removed by adsorption. For these reasons humic
substances were an important part of this study.
OBJECTIVES
The objectives of this study were to
1. Characterize humic substances from different sources and to
determine how these materials are adsorbed and the extent to which
they compete with selected trace organics for adsorption sites on
activated carbon.
i
2. Determine how geosmin and 2-methylisoborneol (MIB), causative agents
of earthy-musty odor in water supplies, adsorb on carbon at their
threshold odor level in the presence and absence of humic substances
-------
3. Determine how mixtures of chlorophenols, compounds which are odorous
and which are representative of other undesirable compounds, adsorb
at the yg/1 level in the presence and absence of humic substances.
Especially important is the degree to which the presence of one
chlorophenolic species causes a reduction in the capacity of carbon
for other chlorophenols.
4. Determine whether the adsorption of polynuclear aromatic hydro-
carbons, some of which are proven carcinogens, is affected by the
presence of humic substances.
Achievement of these objectives will permit an assessment of whether
past results on GAC adsorption of odor compounds are generally applicable,
and whether selected, important trace compounds will adsorb similar to the
odor compounds.
ADSORBATE CHARACTERISTICS AND BACKGROUND INFORMATION
Humic Substances
According to Schnitzer and Khan (1972) humic substances are compounds
which are amorphous, brown or black, hydrophilic, acidic, polydispersed
substances of molecular weight of several hundred to tens of thousands. In
contrast, the nonhumic substances are those such as proteins, carbohydrates,
carboxylic acids, etc. which exhibit recognizable chemical characteristics.
The humic substances which are found in water can be classified into two
broad categories. The first is humic acid which is soluble in dilute
alkaline solutions but precipitates in strongly acidic solutions. The second
is fulvic acid which is soluble in both acidic and basic solutions.
These fractions are structurally similar but differ in molecular weight,
ultimate analysis and functional group content. Both fractions generally are
relatively resistant to microbial degradation, form water soluble and insoluble
salts and complexes, and interact with clays and organic matter in solution.
It appears that the carboxyl, hydroxyl, and carbonyl groups are the predom-
inant functional groups in humic materials and the relative proportions of
these are an index of humic substance reactivity (Schnitzer and Khan, 1972).
The carbon content of humic acid ranges from 50 to 60 percent, the oxygen
from 44 to 50 percent, the nitrogen from 1 to 3 percent, and the sulfur from
near 0 to about 2 percent.
Some adsorption studies have been conducted on humic substances, or on
fractions of organic matter which include the humic substances. A pilot
plant with a 30 in. GAC filter operated at 2 gpm/ft2 treating raw Ohio River
water has been evaluated for elimination of haloform precursors. It was
shown that after 3 to 4 weeks of operation, sufficient material was being
passed through the bed to pr-oduce measurable amounts of chloroform when the
effluent was chlorinated, and after 10 weeks the concentration of the GAC
bed effluent was about 50% of the concentration formed when GAC was not used
(U.S. Environmental Protection Agency, 1975).
-------
Sontheimer and Maier (1972) carried out an extensive evaluation of 10
different commercial carbons to determine their ability to remove organic
matter as measured by absorption of UV light at 240 nm. River Rhine water
was used which had been filtered through the river bank, (this removes many
biodegradable organics and some dilution of river water with ground water
takes place), ozonated and filtered. The conclusions which could be drawn
from their results were:
1. The relative positions and slopes of the isotherms were dependent
upon the point on the lower Rhine where the sample was taken and
upon the time of the year and the rate of flow of the river.
2. The phenol number and the BET surface area do not provide a good
indication of carbon effectiveness for organic removal by carbons
prepared from different raw materials or by different activation
processes. However, if the same raw material and the same activa-
tion process are used, better adsorption properties are associated
with the higher values of these parameters. It is also likely that
the more extensively activated carbons will cost more, however.
3. The least effective of the 10 carbons in the past had proven to
be very effective for odor removal.
Their results thus provide strong evidence that the adsorbability of this
organic matter is dependent on the type of material being adsorbed.
Sontheimer and Maier (1972) further evaluated whether the relative
efficiencies of the carbons as indicated by the isotherms could be observed
in pilot scale tests using 1 m deep beds. While the results were not
entirely conclusive, it was observed that the better carbons based on the
isotherm evaluation generally removed more material in the carbon beds. The
total amount of adsorbed material as indicated by extraction with dimethyl-
formamide after a period of operation yielded results consistent with those
obtained by analyzing the column effluent. It was further observed that all
of the carbons tested in beds seemed equally effective for odor removal, but
that saturation of the carbon's capacity for total organics was reached much
before its capacity for odorous compounds was reached. They also found
evidence indicating that some of the odor compounds were removed in the
upper part of the bed by biological activity.
In later work by Schweer and co-workers (1975), the removal of sulfur
containing organics by GAC and other treatment processes was found to
parallel TOC removal.
Sontheimer (1974) reports pilot plant data obtained at the waterworks
in Diisseldorf showing the efficiency of 3 different carbons for removal of
TOC and total chlorinated organic compounds (TOC1) as determined by the
technique of Ku'hn and Sontheimer (1973a, 1973b), as well as specific
chlorinated organics. TOC1 is a measure primarily of the lipophyllic organic
group of compounds. This group of compounds practically never occurs in
nature, are difficult to decompose biologically, and are frequently hazardous
to health. There are currently two methods used to determine these compounds.
-------
The first involves adsorption of them onto carbon or a synthetic resin and
then to extract with dioxane or dimethylformamide. The organic chlorine in
the extract can then be analyzed by combusting the sample in a quartz
combustion tube and then determining the chloride produced by microcoulometry
(Klihn and Sontheimer, 1973a). The second involves direct combustion of the
carbon sample containing adsorbed organochlorine compounds and determination
of the chloride which is produced by either microcoulometry or a chloride ion
electrode. Care must be taken to eliminate interference from inorganic
chloride adsorbed on the carbon (Ku'hn and Sontheimer, 1973a, 1973b). The
phenol number and surface area were again found to have no relationship to
the adsorption efficiency. Also the carbon which adsorbed the most total
organic matter from bank filtered and ozonated Rhine River water over a
6 month period of time, as indicated by TOC measurements on the filter influ-
ent and effluent as well as by extraction of the carbon with dioxane and
dimethylformamide, adsorbed the least TOC1. The efficiencies of adsorption
of specific chlorinated organics, hexachlorocyclohexane, bis (2-chloropropyl)
ether and hexachlorobutadiene, paralleled the removal of TOC1. These results
indicate the importance of knowing the objective for which carbon is to be
used prior to selection of the carbon for an application.
Geosmin and 2-Methylisoborneol
The results of using 6AC for the removal of odors presumably of biolog-
ical origin indicate that odor breakthrough occurs much later than break-
through of organics measured by a more general parameter such as carbon
chloroform extract (CCE) or carbon alcohol extract (CAE) (Robeck, 1975).
Because most GAC adsorbers are now in use with the objective of removing
odors of biological origin, two compounds of biological origin,
2-methylisoborneol (MIB) and geosmin, were chosen for in-depth study. Close
examination of the adsorptive behavior of these species should indicate
why the bed-life for such compounds is so long and whether other trace
compounds are likely to be similarly adsorbed.
Geosmin is produced by some actinomycetes and blue-green algae (Rosen
et al., 1970; Gerber and Lechevalier, 1965; Medsker et al., 1968) and MIB
has been identified as a product of actinomycetes (Rosen et al., 1970;
Medsker et al., 1969). These compounds have been identified as causative
agents of the widespread problem of earthy-musty odors in water supplies
(Rosen et al., 1970; Medsker et al., 1969; Jenkins, 1973). A survey in 1957
showed that two-thirds of the water treatment plants surveyed described
their odor problems as earthy-musty or, similarly, as that of decaying
vegetation.(Sigworth, 1957). Dice (1976) also reports survey results
indicating that odor from actinomycetes is a frequent problem. MIB and
geosmin specifically have been found in surface waters with earthy-musty
odors in Europe by the Dutch (Piet et al., 1972), and Japan (Kikuchi et al
1973a, 1973b).
MIB was identified by Medsker et al. (1969). Geosmin was first isolated
and named by Gerber and Lechevalier (1965), while Safferman et al. (1967)
later isolated an identical compound. Gerber (1967) presented a partial
identification of this compound which was substantiated by Medsker et al.
-------
(1968).
0.1-0.2
below.
The threshold odor concentrations of geosmin and MIB are about
yg/1 (Jenkins, 1973) and their structural representations are shown
CH,
OH -J
2-Methylisoborneol
CH3
Geosmin
Chlorophenols
Chlorophenols were selected as model compounds for our study for a
number of reasons. First, they impart an objectionable taste and odor to
water when present and may have been partly responsible for the reported
chemical tastes and odors in the drinking water that triggered the Lower
Mississippi Study (U.S. Environmental Protection Agency, 1975). In a study
performed to determine the halogenated compounds which result from
chlorinated secondary sewage effluents, trichlorophenol, as well as a number
of other chlorinated aromatic compounds, was found (Glaze and Henderson,
1975). In these two instances, Chlorophenols probably were formed by the
reaction of phenols with aqueous chlorine. The primary sources of phenols
in natural waters include natural decay products, waste effluents of coking
plants, brown coal distillery plants, and the pulp and paper industry.
Phenols are used in the synthesis of a number of organic compounds resulting
in their presence in the effluents from many chemical plants. It has been
estimated that the concentration of free phenols in unpolluted streams is
less than 50 yg/1 while that in rivers receiving industrial and municipal
wastewater is frequently greater than 100 yg/1 (Zogorski and Faust, 1974).
According to Burttschell et al. (1959), the chlorination of phenol
proceeds by stepwise substitution of the 2, 4, and 6 positions of the
aromatic ring, in the manner shown in Figure 1. Below each compound in
Figure 1 is listed its threshold odor concentration. The compounds with the
strongest odor-producing potential are 2-chlorophenol, 2,4-dichlorophenol,
and 2,6-dichlorophenol which are detectable at concentrations from 2 to 3
yg/1. These are the compounds primarily responsible for the taste and odor
in water. The 1962 Public Health Service Drinking Water Standards set the
maximum level for total phenols at 1 yg/1 to prevent odor problems. Toxic
effects are thought to occur only at far higher concentrations partly
because phenols are largely detoxified in the mammalian body (McCaull and
Crossland, 1974).
-------
OH
/\
OH
Cl Cl
/\
Cl
cri
OH
y\
V
[>iooo]
OH
C I
[250]
V
OH
/\
Cl
OH
v
Oxidation
Figure 1. Reaction scheme for the chlorination of phenol.
[ ] indicates odor threshold concentration (yg/1)
(after Burttschell et al., 1959)
-------
The series of reactions leading to the formation and ultimate destruc-
tion of chlorophenolic odors involves complex kinetic interrelationships.
A short time after mixing chlorine and phenol, eight interdependent reactions
are occurring simultaneously. In a study of the kinetics of these reactions
by Lee (1967), the expected concentrations of the individual chlorinated
phenols were determined as a function of time and other variables. By
combining these data with information on the organoleptic properties of the
compounds, Lee determined the threshold odor values of chlorinated waters
which contained phenols as a function of time, pH, and relative concentrations
of chlorine and phenol. Depending on the conditions which existed in the
water, Lee observed that different mixtures of chlorophenols may be present
at a given time.
The second reason for selecting chlorophenols for use in this study was
due to their similarity in structure to a number of pesticides which resist
biodegradation. Roughly 25 percent of the pesticides on the world market
are compounds which possess a substituted phenol moiety which can be cleaved
from the molecule through hydrolysis in natural waters (Friestad et al.,
1969). It has been reported that photodecomposition of the herbicide
2,4,5-T (2,4,5-trichlorophenoxyacetic acid) by sunlight in alkaline natural
waters may result in the formation of 2,4,5-trichlorophenol and 2,5-
dichlorophenol (Crosby and Wong, 1973b). Studies by Crosby and Wong (1973a,
1973b) indicate that this pathway may be general for other commercial
phenoxy herbicides. Therefore, the decomposition of pesticides in natural
waters may lead to trace amounts of chlorophenols.
In spite of their importance little information is available to indicate
how chlorophenols adsorb from waters containing a mixture of chlorophenolic
species and humic substances.
Polynuclear Aromatic Hydrocarbons
Polynuclear aromatic hydrocarbons (PAH) are ubiquitous and are found in
small but detectable concentrations in air, water and soil samples of all
types (McGinnes and Snoeyink, 1974). They are natural products of organic
decomposition and are products of incomplete combustion, petrochemical, coal
and chemical industrial processes. The concentrations found in water typi-
cally range from 0.001 to 10 yg/1 and they are concentrated in the food
chain because of their favorable solubility in fatty material. Some species
of PAH are demonstrated carcinogens at higher concentrations and because
of this they are a potential health hazard in water.
Previous studies have shown naphthalene to be very strongly adsorbed
from Cincinnati tap water (Robeck, 1975), and that river bank filtration
followed by activated carbon treatment removed about 99 percent of PAH. from
River Rhine water (Andelman, 1973).
In this study, our primary goal was to determine whether PAH associated
with humic substances. If this does occur, it is possible that adsorption
onto carbon will be controlled by the adsorption characteristics of the humic
substances and not by those of the PAH.
-------
APPROACH TO THE STUDY
The study was conducted using aqueous solutions prepared in the labora-
tory. Sufficient quantities of the adsorbates were obtained from various
sources. MIB was synthesized by us while the geosmin was obtained from the
U.S. EPA. The chlorophenols and PAH were obtained from commercial sources.
Humic substances were obtained from a commercial source, a local well water,
and from leaf and soil extract. Coal-base activated carbon was used as the
adsorbent. In most cases it was necessary to develop a quantitative anal-
ytical procedure for each of the species.
Small-scale laboratory batch and column tests were used to obtain the
desired information on adsorption characteristics. These tests are very
flexible and enabled us to look at a wide variety of conditions.
The tests were conducted in such a way so that biological activity did
not take place. Short-term tests, and in some cases biocides were used to
insure the absence of significant biological growths. It is recognized that
biological growths are prevalent in GAC beds at water treatment plants,
and that such growths significantly affect the quality of the effluent from
these beds (see McCreary and Snoeyink, 1977 for a review) but the scope of
this study did not permit us to examine it.
-------
SECTION 2
CONCLUSIONS
Activated carbon adsorbed humic substances in all cases that were
studied but the adsorption properties of the substances from different
sources varied widely as did the extent to which they competed with selected
trace organics for adsorption sites on activated carbon. Humic substances
from leaf and soil extract, a well water and a commercial source were
examined in detail. Extent of adsorption depended upon solubility, with the
less soluble humic acid (HA) fraction being more adsorbable than the fulvic
acid (FA) fraction from the same source. The lower molecular weight species
from a given FA or HA fraction are more adsorbable than the high molecular
weight species presumably because more surface area is accessible to them.
The adsorption characteristics of the humic substances are also dependent
on the method of analysis used to quantify them. The species which fluoresce
the most were found to be the lower molecular weight species and these adsorb
best. UV absorbing species did not adsorb as well as those which fluoresce.
Solution pH and phosphate concentration also had a marked effect on adsorba-
bility of the humic materials with adsorption generally improving with
decreasing pH and increasing phosphate concentration. The haloform formation
potential of the humic substances varied widely from source to source, with
only one exception, but no dependence on molecular weight was found for
fractions of FA or HA. Based on the results of this work it is concluded
that when designing an adsorption system to remove haloform precursors or
trace organics, it is important that testing be done using the water to be
treated to determine design and operating parameters. Because of the vari-
ability in adsorption characteristics of the humic substances it would appear
that adsorption results obtained at one location will likely not be those
obtained at another location.
Concentrations of 2-methylisoborneol (MIB) and geosmin can be quanti-
tatively measured down to 0.1 yg/1 by a procedure consisting of extraction,
concentration and gas chromatographic analysis with a flame ionization
detector. The MIB synthesized from d-camphor was found to be identical to
that produced by actinomycetes.
MIB and geosmin are both strongly adsorbed by activated carbon. When
present, the humic substances significantly reduce the capacity of carbon
for these compounds, more so before equilibrium is achieved than at equi-
librium. Commercial HA and the humic substances from well water each had
differing competitive effects on MIB. The capacity of carbon for geosmin
was reduced to a greater extent than was observed for MIB by commercial HA.
The performance of laboratory columns was consistent with the isotherm
-------
results. Application of distilled water to a partially saturated carbon bed
resulted in almost no elution of MIB indicating that it was strongly
adsorbed. Assuming complete saturation of the carbon, no leakage and no
biological activity, predicted bed life for reduction of MIB or geosmin from
10 yg/1 to its threshold odor level of 0.1 yg/1 in a 2-foot deep bed is much
greater (several months to years) than for reduction of humic substances
from 5 to 1 mg/1 (1 to 2 months), for example. When both MIB, or geosmin,
and humic substances must be removed, humic substance removal will control
the life of the bed. In field installations of granular activated carbon,
the possibility that odor compounds are generated, or in some instances
degraded, by biological growths within the bed cannot be overlooked.
Chlorophenols are adsorbed very strongly by activated carbon at the
yg/1 level which is near the threshold odor limit for these compounds. The
extent of adsorption of 2,4-dichlorophenol (DCP) and 2,4,6-trichlorophenol
(TCP) is a function of pH. The neutral species of these compounds predomi-
nate at pH below the pKa values (7.85 and 6.00, respectively, at 25°C) and
are adsorbed more strongly than the anionic species. As the number of
chlorine atoms substituted on the phenol increases, the solubility of the
neutral species decreases and the adsorbability increases; as substitution
increases, the pKa of the species is lowered, however. When water containing
phenol is chlorinated with low levels of chlorine, a mixture of chlorophenols
will form and thus the extent of adsorption of one chlorophenol in the
presence of another chlorophenol is an important consideration. Significant
reductions in adsorption capacity (up to 50 percent) of one chlorophenol was
caused by the presence of a second chlorophenol. The langmuir model for
adsorption was found to be inadequate for fitting single solute adsorption
data over a broad concentration range. To obtain Langmuir parameters for
use in competitive adsorption equations, it was necessary to fit the single
solute data with a polynomial equation and to use the polynomial to calculate
the Langmuir parameters at the desired concentrations. Competitive adsorp-
tion studies between DCP and TCP resulted in verification of the applica-
bility of the Langmuir competitive adsorption equation at pH 5.2, and the
Jain and Snoeyink modification of this model at pH 9.1. At pH 7.0, where
neutral DCP competes with anionic TCP, neither equation was satisfactory.
Evaluation of the competitive effects of commercial HA, soil FA and leaf FA
showed that the presence of these materials decreased the capacity of carbon
for chlorophenol and that each of the materials competed somewhat differently.
However, even in the presence of humic substances and another chlorophenol
species, it appears that the adsorption capacity is even greater for chloro-
phenol th-an it is for MIB and that bed life for chlorophenol will be greater
than for MIB and much greater than for humic substances.
Limited experimentation with the polynuclear aromatic hydrocarbon (PAH)
anthracene led to the conclusion that there was no significant association
between it and humic substances. Thus it appears that the possibility of
PAH passage through carbon beds because of its association with the more
poorly adsorbable humic substances is not a cause for concern.
The general conclusion made on the basis of this study is that adsorp-
tion characteristics of organic substances which one may wish to remove
during water purification are highly variable. Past experience with full-
10
-------
scale carbon beds at water treatment plants, used primarily to remove odor,
provides little indication of how humic substances and selected trace
organics will be removed, for example. Adsorption tests using the water to
be treated should be conducted prior to designing systems to accomplish a
specific objective. On the basis of this study it is expected, however,
that carbon bed life for removal of typical levels of chlorophenols, MIB and
geosmin will be much longer than for humic substances.
11
-------
SECTION 3
RECOMMENDATIONS
Laboratory and field studies should be initiated to develop simplified
procedures for characterizing natural organic matter with respect to adsorb-
ability, competitive effects and tendency to form haloforms when chlorinated.
Fluorescence response, UV absorbance, TOC and density of selected functional
groups should be related to both haloform formation potential and adsorba-
bility and seasonal variability in type of organics should be taken into
account. Research should be conducted on procedures other than adsorption,
such as coagulation, which may be effective in removing certain fractions
of humic substances and thereby increasing carbon bed life.
Field studies should be undertaken at water treatment plants where
granular activated carbon is used and where geosmin or MIB is a problem.
Carbon samples taken from several depths within the bed should be extracted
and the extract should be analyzed for the odor compound to determine the
concentration profile in the bed. Determination of the profile at different
times will permit a determination of the rate of movement of the odor com-
pound through the bed and verification of the prediction that bed life
should be on the order of several months to years. Biological activity in
the bed should also be examined to determine whether it is producing or
degrading the odor compound within the bed.
Research should be undertaken to develop reliable small-scale adsorption
test procedures to be used on-site to determine the best design parameters
and operating conditions to treat a given water. The small-scale test
results would necessarily have to be compared with large-scale results and
this could best be accomplished at locations where pilot- or full-scale
studies are being initiated. If small-scale tests can be used successfully,
significant reductions in the time and funds required to obtain necessary
design and operation information should be possible.
12
-------
SECTION 4
MATERIALS AND METHODS
ADSORBENTS
Two bituminous coal base activated carbons, Filtrasorb 200 (F-200) and
Filtrasorb 400 (F-400) were used for this study. Both carbons were prepared
by grinding, sieving to the desired size range (i.e., 40 x 50, 50 x 60, or
60 x 80 U.S. standard mesh), washing, and then drying to a constant weight
at 120-140°C. See Calgon Activated Carbon Product Bulletin (1969) for
general carbon characteristics.
HUMIC SUBSTANCES
Isolation and Purification
The humic substances used in this study were obtained from a commercial
source (Pfaltz and Bauer), and were extracted from leaves and soil. Humic
substances in well water obtained from a deep well in the Civil Engineering
Building at the University of Illinois were also used. The commercial humic
acid was purified using the procedure outlined by Narkis and Rebhun (1975),
with some modifications. The material was dissolved in 1.5 percent NaOH,
filtered, acidified to pH 1 with HC1 to precipitate the humic acid, washed
until chloride free and dried at 60 to 65°C. The procedure was later changed
to employ centrifugation rather than filtration and freeze-drying concentra-
tion instead of oven drying. The purification step did not have a signifi-
cant effect on the humic acid adsorption isotherm as shown by comparison of
isotherms determined before and after the purification step.
Well water from a deep aquifer at the University of Illinois was also
used as a source of humic material. Prior to use it was aerated and settled
to remove the iron. The well water had a yellowish-brown color and a
chemical oxygen demand (COD) (Standard Methods, 1975) of about 10 mg/1. It
was also analyzed at the USEPA Municipal Environmental Research Laboratory
in Cincinnati and found to have a nonpurgeable total organic carbon (NPTOC)
concentration of 3.3-3.6 mg/1.
A continuous upflow column (see Figure 2) was used to extract humic
material from leaves and a Finch soil. Eight liters of deionized water was
cycled continuously through a bed of leaves acquired from a hardwood forest.
Air was bubbled into the holding tank to prevent septicity. The water
quickly became discolored and after about two weeks of cycling, the organics
13
-------
Soil
Pump
Reservoir
Figure 2. Upflow column for extracting humic material
14
-------
which were extracted were concentrated by freeze-drying. The pH of the
water was 6.8. Very little material precipitated from the solution at pH 1
indicating that almost no leaf humic acid was present.
Eight liters of a 0.1 M sodium pyrophosphate solution at a pH of 10.6
was cycled through the column with Finch soil. The Finch series is a some-
what poorly drained sandy soil with a strongly cemented subsoil. It was
obtained from an area near Traverse City, Michigan. The solution became
immediately discolored and contained fine sediment that would not filter
readily. After about three days of cycling, the solution was centrifuged in
150 ml polyethylene bottles at 8000 rpm for 30 minutes. This served to
deposit the colloidal material. The centrate was then acidified to pH 1
and stirred for approximately one hour. This solution was centrifuged to
deposit the precipitated humic acid while the fulvic acid remained in
solution.
In order to separate the fulvic acid from the remaining solution,
XAD-8 macroreticular resin from Rohm and Haas was used following the pro-
cedure of Leenheer and Huffman (1976). The resin was initially washed with
methanol and then Soxhlet extracted with ether, acetonitrile, and methanol
for 8 hours each. A column 3 cm in diameter and 30 cm long was used for the
resin. The XAD-8 resin was followed by a similar bed of XAD-2 resin which
resulted in little additional removal (see Figure 3 for typical results).
About one liter of fulvic acid solution at pH 1 was applied to the
XAD-8 bed at a flow rate of approximately 2.5 ml/min. This was followed by
one liter of 0.1 M NaOH to regenerate the resin. The desorbed fulvic
material moved as a band through the bed. Backwashing followed by several
bed volumes of deionized water was sufficient to prepare the resin for
adsorption of more fulvic acid. The fulvic acid solution obtained from the
resin was adjusted to pH 7 and freeze-dried.
Dialysis tubing was used initially to prepare salt-free organics.
However, it was observed that the smaller molecular weight organics were
passing through the tubing and into the surrounding solution. A similar
observation has been reported by Stevenson (1965). Freeze-drying was
therefore conducted without prior dialysis.
Molecular Weight (Size) Fractionation
Gel filtration was used for the separation of materials on the basis of
molecular weight or size. Gjessing (1976) has shown that the non-excluded
fraction from a Sephadex gel column is of questionable value as far as a
molecular weight estimate is concerned due to irreversible adsorption inside
the gel particle. For the fractionation of humic material he used a series
of Sephadex gel columns and selected only the excluded fraction from each
column with subsequent concentration and reapplication of the non-excluded
fraction to the column with the next smaller molecular weight exclusion
limit. We used a similar procedure accompanied by ultrafiltration to obtain
our molecular weight fractions.
15
-------
100% TOC initially extracted from Finch soil
I
pH adjusted to 1
centrifuge
\
61% TOC in solution
humic acid precipitate
39% of TOC
adsorption on XAD-8 resin
fulvic acid
41% of TOC
20% of TOC in solution
adsorption on XAD-2 resin
X
17% of TOC in solution
adsorbed organics
3% of TOC
Figure 3. Organic carbon separation from Finch soil monitored by
TOC analysis
16
-------
Table 1 gives the approximate molecular weight exclusion limits for the
gels and the molecular weight cut-offs for the ultrafiltration membranes
used.
TABLE 1. GEL MOLECULAR WEIGHT EXCLUSION LIMITS AND ULTRAFILTRATION
MEMBRANE MOLECULAR WEIGHT CUT-OFF LIMITS
Ami con Diaflo Filters Sephadex Gels
UM 10 10,000 MW G-75 50,000 MW
XM 50 50,000 MW G-50 10,000 MW
G-25 5,000 MW
G-10 700 MW
Our procedure was modified from Gjessing's in that a pH 10, 0.01 M
phosphate buffer was used as the eluant to decrease the adsorption of
charged groups on the Sephadex particles (Pharmacia, 1974). Also, freeze-
drying was used to concentrate the eluted organics rather than roto-
evaporation at 40°C.
Each of the gels was boiled in distilled water and allowed to swell
for at least the time specified by the manufacturer. The columns were
poured through a Buchner funnel fitted on the top of the column with constant
stirring and a slow constant flow rate through the column. Finally a glass
fiber filter was placed on top of the bed to prevent disturbance when adding
sample. The column was thoroughly washed with the phosphate eluant prior to
use. Several bed volumes of buffer were required to halt leaching as
monitored by TOC. The absence of packing irregularities was confirmed by
the application of Blue Dextran dye and analysis of the resulting
chromatogram.
One hundred milligrams of organic material dissolved in 2 ml of water
was applied to the top of the column. When necessary the solution prepared
from the freeze dried organic was filtered to eliminate insoluble residue
before application. Irreversible adsorption on the gels was noted when the
soil and commercial humic acid were applied and excessive head loss developed
with the fine-grade gels. For this reason, only the coarse-grade Sephadex
was used for these materials. The volume containing the excluded organics
was taken as equal to one bed volume after color appeared in the bed effluent.
The first bed volume, consisting of that liquid which passed after application
of the sample and before color breakthrough, was discarded. The flow rate
was approximately 0.5-1.0 ml/min. Sampling was done with an automatic SMI
fraction collector.
17
-------
To further purify the Sephadex fractions, ultrafiltration was used.
As shown in Table 1, the Sephadex G-75 exclusion limit is identical to the
molecular weight cut-off for the Amicon XM-50 membrane. The same is true Tor
Sephadex G-50 and the UM-10 filter. The volume containing the excluded
organics from the G-75 gel was filtered through the XM-50 membrane. The
solution that passed through the membrane was discarded and the organics
retained by the filtration unit constituted the purified fraction of molecular
weight > 50,000. The XM-50 filter was also used to purify the non-excluded
fraction of G-75 by collecting the organics that passed through the filter
and discarding those retained by the unit. The filtrate was freeze-dried to
small volume and applied to the G-50 gel filtration column. An identical
procedure was followed for the G-50 gel and the UM-10 filter. Molecular
weight fractions from the other gels were not further purified by
ultrafiltration.
A mass balance (see Table 2) was conducted on the Finch soil humic
material for the molecular weight fractionation using Sephadex gel filtration.
It was observed, as expected, that the humic acid consisted of generally
larger weight molecules than the more soluble fulvic acid.
TABLE 2. FINCH SOIL HUMIC SUBSTANCE FRACTIONATION
Soil Humic Acid Soil Fulvic Acid
MW > 50,000 56.4% MW > 50,000 29%
5,000 < MW < 50,000 32.3% 5,000 < MW < 50,000 48%
MW < 5,000 11.3% 700 < MW < 5,000 20%
MW < 700 3%
A mass balance was also made on the G-50 gel fractions of Finch soil
humic acid using the Amicon XM-50 and UM-10 ultrafiltration membranes. From
Table 3 we observe that much of the humic substance excluded from this gel
passed through the UM-10 membrane which has the 10,000 MW cut-off.
Similarly, much of the volume of < 10,000 MW organics from this gel was
retained by the filtration unit. Thus we see that the use of ultrafiltration
after Sephadex gel fractionation significantly improves the fractions.
Gjessing (1976), observed the same results and attributed this as a
demonstration of the effectiveness of ultrafiltration over gels as a frac-
tionation tool. However, experience in our laboratory demonstrated problems
with clogging of the pores of the membrane if prior gel filtration was not
used and little difference among the fractions in terms of adsorption
capacity when ultrafiltration was used alone as compared with the non-
fractionated material. It appears from our results that gel filtration
followed by ultrafiltration gives superior results.
-------
TABLE 3. EVALUATION OF 6-50 COARSE SEPHADEX FRACTIONS OF SOIL
HUMIC ACID BY ULTRAFILTRATION
Procedure
Molecular Wt.
of Total
Molecular Wt. % of Total
Gel Filtration > 10,000
100
< 10,000
100
Ultrafiltration
> 50,000
> 10,000
< 10,000
42.6
14.7
42.4
> 10,000
< 10,000
45.4
54.6
Fluorescence, UV Absorbance, Total Organic Carbon --
Three primary analytical procedures, ultraviolet absorbance at 240 nm,
fluorescence, and TOC were used for determination of the concentrations of
the humic substances. The fluorescence measurements were made using a
Turner Model 110 fluorometer with an excitation wavelength of 365 nm, an
emission wavelength of 415 nm and a path length of 1 cm. Fluorescence scans
were obtained with an Aminco-Bowman scanning spectrofluorometer- Excitation
spectra and emission scans were obtained for the various humic materials at
identical NVTOC concentrations. The solutions were buffered at pH 7. We
observed that the maximum excitation wavelength was near 360 nm, comparable
to the 365 nm wavelength used on the Turner fluorometer. The maximum
emission peak was very broad and in the range 450-470 nm as compared to the
415 nm wavelength used on the Turner fluorometer. All of the materials had
essentially the same excitation and emission characteristics.
The fluorescence intensity of the unfractionated and fractionated humic
materials at 5 mg/1 TOC concentration was found to correlate well with
molecular size (see Table 4). The smaller molecular weight materials were
found to have a much higher fluorescence intensity than the larger molecular
weight substances.
A Beckman ACTA III spectrophotometer was used for the UV measurements.
For well water analysis, the average TOC of 3.45 mg/1 of an aerated,
settled sample was used. A standard curve was then prepared by making
dilutions of this sample and determining the UV or fluorescence response of
the dilution. Unknown concentrations were determined by measuring the
fluorescence of a sample and using this standard curve; mass concentrations
are thus based on equivalent TOC and on the assumption that UV absorbance/
unit TOC or fluorescence intensity/unit TOC was the same for the unknown as
for the standard curve samples.
19
-------
TABLE 4. FLUORESCENCE INTENSITY OF 5 MG/L TOC FRACTIONS
OF HUMIC SUBSTANCES
Sample Fluorescence Intensity
Commercial
Ful vie Soil
Fulvic Soil
Ful vie Soil
Fulvic Soil
Soil Humic,
Soil Humic
Soil Humic
Soil Humic
Leaf Fulvic
Leaf Fulvic
Humic Unfractionated
, Unfractionated
>50,000
>5,000
<5,000
Unfractionated
>50,000
>10,000
<10,000
>5,000
<5,000
415
315
94
195
320
185
57
96
200
237
397
Since the soil and leaf humic material was not entirely salt-free,
50 mg/1 TOC solutions were prepared. Standard curves were made by making
dilutions of these solutions and determining the fluorescence response of
the dilutions. Concentrations of unknown solutions could be determined by
comparing their fluorescence to this standard curve; mass concentrations are
thus based on equivalent TOC and the assumption that the fluorescence
intensity/unit TOC was the same for the unknown samples as for the standard
curve samples.
Haloform Formation Potential
The haloform formation potential (Stevens and Symons, 1977) of the
various humic fractions was determined to further characterize the material.
The chloroform analysis procedure reported by Kaiser and Oliver (1976) was
modified to suit our needs. All water used for chloroform analysis was
distilled in glass followed by exhaustive stripping with N2 heated over a
copper catalyst. This water was then buffered at pH 7 with 0.001 M phosphate.
All glassware was thoroughly washed and baked overnight at 450°C before use.
Standard chloroform solutions were prepared by injecting chloroform
(Mallinckrodt Nanograde) into an appropriate amount of water followed by
stirring with minimal headspace overnight. Solutions of the internal
standard, 1,1,1-trichloroethane, were prepared in a similar fashion.
20
-------
Water with chloroform was spiked with a known amount of internal standard
solution for analysis. The sample was then placed in a 60 ml separatory
funnel with a 2 ml headspace and inverted in a 70° water bath. After
equilibration for 45 minutes, about 100 yl of the gas was sampled by gas-
tight syringe and injected directly into a Hewlett-Packard gas chromatograph
with a Ni63 ECD detector. A six foot 0.4% Carbowax 1500 on Carbopack A column
was used with argon-methane carrier gas at a flow rate of 36 ml/min. The
column was preconditioned at 210°C for one week after packing. Temperature
settings were 220°C for the detector, 200°C for the injector port, and 110°C
for the column. At these conditions, the retention times for chloroform and
1,1,1-trichloroethane were 2.2 minutes and 3.6 minutes respectively. A
typical chromatogram is shown in Figure 4.
Chlorine solutions were prepared by bubbling Cl2> (Linde high purity
gas), into an alkaline solution. The pH was then adjusted to 7 and the
solution was stored in a dark bottle at 4°C.
Samples of humic material were prepared at 2.5 mg/1 TOC and chlorinated
with 10 mg/1 chlorine as Cl2 as measured by the DPD titrimetric method,
(Standard Methods, 1975). All water was distilled, stripped, and buffered
at pH 7 with 0.001 M phosphate. The reaction solution was mixed and immedi-
ately transferred to 125 ml Wheaton Scientific glass vials. Solution was
added to fill the bottle without headspace and the caps were crimped over a
a Teflon seal. The reaction bottles were placed in an 18°C constant temper-
ature water bath until sampling. At various times the bottles would be
reopened and transferred to a second vial containing an appropriate amount
of sodium thiosulfate to stop the reaction, crimped again without headspace,
and refrigerated until analyzed.
To avoid problems with evolution of chloroform from the standard
solutions, the solution was transferred to a series of Wheaton bottles and
crimped without headspace. Before each series of separatory funnels was
set up, a new bottle of internal standard was removed from the refrigerator
and used. After opening and using, the remainder of the solution would be
discarded.
It was found that the headspace technique for analysis of chloroform
offers distinct advantages over conventional extraction or purge and trap
techniques. Several injections could be made from one funnel and the peak
height ratios of chloroform to internal standard are reproducible to within
2%. Several funnels could be set up at staggered times to permit considerable
time savings over the purge and trap technique. Like the purge and trap
technique, only peaks due to volatile species are observed and no problem
exists with high retention solvent impurity peaks. Due to the non-linearity
of the ECD detector, concentrations above 50 yg/1 had to be diluted to a
suitable concentration range and the peak ratios had to be compared to a
standard curve of several known chloroform concentrations.
21
-------
Concentration !
Chloroform = 40 ppb
1,1,1 Trichloroethane (Internal
Standard ) = 10.7 ppb
Retention Time !
Chloroform — 2.1 min
1,1,1 Trichloroethane - 3.25 min
Attenuation !
32 xlO
Inject 100 y£ Gas
Figure 4. Typical chloroform chromatogram
22
-------
2-METHYLISOBORNEOL, SYNTHESIS AND ANALYSIS^
The 2-methylisoborneol (MIB) used in this investigation was synthesized
from d-camphor, and extraction, concentration, and GC analysis were used to
quantify concentrations down to 0.1 yg/1.
Reagents. Solvents and Adsorbents
The reagents, solvents and adsorbents used in the synthesis and analysis
are given in Table 5.
TABLE 5. REAGENTS, SOLVENTS AND ADSORBENTS FOR MIB
SYNTHESIS AND ANALYSIS
Substance Source
d-Camphor, mp 178-180° Eastman Kodak Company, Rochester, NY
Methyl!ithium, 1.7 M in ether Ventron, 8 Congress Street, Beverly, MA
Ether, anhydrous analytical reagent Mallinckrodt, St. Louis, MO
Hydroxylamine hydrochloride, analytical
reagent Mallinckrodt, St. Louis, MO
Hexane, distilled in glass Burdick and Jackson Laboratories,
Muskegon, MI
Methylene chloride, distilled in glass Burdick and Jackson Laboratories,
Muskegon, MI
Silica Gel, 0.05/0.2 mm, non-activated Brinkmann Instruments
Florisil, 100-200 mesh, non-activated Fisher Scientific Supply Co.
Hater Samples
The tap water used was that from the Civil Engineering Building, and well
water that from a well in the basement. Surface water was that taken from a
polluted stream running through the center of Urbana and lake water that taken
from a small lake just outside the town.
1. Much of the material in this subsection was taken from "2-Methylisoborneol
Improved Synthesis and a Quantitative Gas Chromatographic Method for Trace
Concentrations Producing Odor in Water," by N.F. Wood and V.L. Snoeyink,
J. Chromatogr., 132, 405 (1977) with the permission of the copyright
holder.
23
-------
Polarimetry
The specific rotation of both natural and synthetic MIB was determined
on a Bendix-NPL Automatic Polarimeter which employs the Faraday electro-optic
effect to measure optical rotation. To measure the very small rotation
observed from the small sample of natural MIB, the instrument had to be
operated at high sensitivity. Full-scale deflection of the meter read-out
needle to the right or the left then corresponded to a rotation of plus or
minus 0.1°, respectively, and rotations could be read to within 0.001°.
With 0.1% sucrose the instrument was found to have a calibration factor of
0.952. The 1-cm cell used in this work would not fit the usual cell holder
for the instrument, so a cardboard holder was constructed that fitted in the
recess in the base of the cell compartment. In this way the cell could be
positioned inside the compartment in the same way every time, and a reading
from a particular solution could be reproduced exactly.
The specific rotation of the synthetic MIB was also determined using a
Carl Zeiss polarimeter.
Thin-Layer Chromatography
Thin-layer chromatography (TLC) was performed on Merck precoated plates
of silica gel GOF-254 (0.25 mm). (Stockists may supply the earlier version
of these plates from old stocks, but results with these are distinctly
inferior in terms of sensitivity and resolution.) Samples were spotted in
amounts up to 200 yg. To develop the plates, suitable mixtures of ethyl
acetate and hexane were used. Spots were visualized by examination under
UV light, treatment with iodine vapor, and by spraying with 1% vanillin in
sulfuric acid. Sensitivity with the spray was 1 yg for 2-methylisoborneol,
5 yg for camphor oxime, and 10 yg for camphor. Visualization of camphor
spots was quite sensitive to the manner of spraying and sometimes required
several hours of standing. Best results with camphor were obtained by a
very light spraying followed by a second spraying after 15 minutes. Spray
reactions and some typical Rf values are shown in Table 6.
Gas Liquid Chromatography
Gas liquid chromatography (GLC) was performed at 140° on a 183 x 0.2-cm
glass column containing Supelcoport (60/80) coated with 3% SP2100 using a
F & M instrument (model 810) equipped with a flame ionization detector. The
glass column was treated initially with 10% dimethyldichlorosilane in
toluene for two hours, washed with methanol, dried, packed with the stationary
phase (Supelco, Inc., Bellefonte, PA), and conditioned at 300° for two hours.
Flow rates were: nitrogen carrier, 35 ml/min; hydrogen, 30 ml/min; and air
350 ml/min. Injections were made on-column at 270° with a Hamilton 10-yl
syringe set at 2-yl with a Chaney adapter. Typical retention times were:
camphor, 1.4 min; 2-methylisoborneol, 1.7 min; and camphor oxime, 3.4 min.
24
-------
TABLE 6. TLC AND MIB AND SOME OTHER CAMPHOR DERIVATIVES
Compound
Spot formed with 1%
vanillin in sulfuric acid
Developing solvent - ethyl acetate:hexane (1:4)
0.30
Camphor Oxime
Borneol
Unknown for HBr on
2-methylisoborneol
Isoborneol
2-Methylisoborneol
Camphor
0.34
0.38
0.41
0.49
0.59
Blue-gray, developing
slowly
Sharp translucent,
developing slowly
Bright crimson
developing immediately
Yellow-brown, turning
blue
Bright crimson,
developing immediately
Translucent, developing
very slowly
GLC proved invaluable for analyzing and for monitoring the formation of
the camphor oxime, the course of liquid-liquid extractions, and the eluants
from the silica gel and florisil columns. Before injection, small aliquots
of both aqueous and organic solutions were diluted 1:250 with hexane.
Specific Rotation of Natural MIB
Natural MIB investigated in this work originated from Streptomyces sp
CWW3 isolated from Lake Michigan. It was contained in a by-product kindly
sent to us by the Environmental Protection Agency from the production of
geosmin by Dr. Nancy Gerber of Rutgers University for the Agency (Gerber,
1974). The by-product, a dark tar, was found by GLC to contain traces of
what were probably olefins (0.6 min and 0.7 min), about 35 mg of geosmin
(4.6 min), and only 14.0 mg of 2-methylisoborneol (1.5 min).
The MIB was isolated by liquid chromatography of the by-product on
silica gel. Though probably not essential to the operation, advantage was
taken of its convenient availability to use modern liquid chromatography
equipment. A stainless steel column (60 cm x 0.71 cm id) was packed with
silica gel and connected with stainless steel fittings to a Waters pump
(model M 6000 A), through a Waters loop injection (model U 6 K). Methylene
chloride was pumped through the column at 2 ml/min and the by-product was
injected in about 1 ml of the same solvent. Fractions of 10 ml were collected
and analyzed without dilution by GLC; results are shown in Table 7. Total
25
-------
TABLE 7. ISOLATION OF MIB BY COLUMN CHROMATOGRAPHY OF A NATURAL PRODUCT
Adsorbent: silica gel
Eluant: methylene chloride
Fractions: 10 ml
Fraction
Peak Height by GLC
Unknown Unknown MIB Geosmin
(r.t.T 0.6 min) (r.t. 0.7 min) (r.t. 1.5min) (r.t. 4.6 min)
2
3
4
5
6
7
8
9
10
7.5 9.6
3.4 4.4
15.0
48.4
5.7
48.4
25.4
4.0
0.1
1. r.t. = retention time
recovery of MIB was estimated as 14.5 mg. Fractions 7-9, that were homogen-
eous in 2-methylisoborneol by GLC, were combined and found to be homogeneous
by TLC also. Removal of methylene chloride was carried out on a rotary
evaporator with the bath temperature at 30° and the evaporation terminated as
soon as it appeared that all solvent had been removed. The residue was a
clear gum weighing about 17 mg.
The residue was transferred with ethanol to a quartz cell (1 cm x 1 cm
id) and made up to a calibration mark of 0.905 ml on the sidearm. The cell
was sealed with a Teflon stopper and an air bubble in the sidearm manipulated
around the entire cell to produce a homogeneous solution. The observed
rotation was -0.023°, and allowing for the ethanol blank of -0.003° and the
calibration factor of 0.952, this corresponded to a corrected rotation of
-0.019°. The concentration of MIB in the cell solution was found to be
0.0136 g/ml after a 9.8 yl portion had been diluted to 1.0 ml with ethanol
and analyzed by GLC.
26
-------
0.1
Mass Spectrometry
The mass spectrum of MIB was recorded on a Finnigan 3300 quadrupole
instrument using the solid probe.
Infrared Spectroscopy
Infrared spectra were recorded on a Beckman IR-20A from samples as
potassium bromide pellets (1 mg in 200 mg) or 10% solutions in chloroform.
Evaporations
Analytical extracts of methylene chloride were concentrated using the
apparatus shown in Figure 5. The concentration vessels were made from 25-ml
round-bottom flasks and are unavailable commercially. Micro Snyder distil-
lation columns (Burke et^ al_. , 1966) like the one shown are supplied by
Kontes, Inc., Vineland, NJ. The heating bath used consisted of a pan
completely filled with water held at 75°C by means of a small hot plate.
The cover was a round aluminum plate (0.5 cm x 25 cm diam.) with a small
hole in the center fitted with a thermometer and several holes (4 cm in diam.)
to accept concentration vessels. The holes held the vessels securely around
the midpoint of the bulb and no additional support was necessary.
Extracts were added to concentration vessels together with a small
boiling chip. Glass joints were sealed with a thin film of water. After
about 5 min. on the water bath, evaporation was complete and the vessel was
cooled in a stream of cold tap water or by allowing it to stand at room
temperature before removing the Snyder column. The condensate was about
200 yl . Where maximum sensitivity by gas chroma tog raphy was desired, this
solution was further concentrated to about 20 yl using a rotary evaporator
without a heating bath. To avoid solutions of this small volume evaporating
to dryness on standing, about 20 yl of ethanol was sometimes added.
Eluates from columns used in the purification of synthetic and natural
MIB were evaporated to dryness using a rotary evaporator with a bath
temperature at 30°.
Stock Solutions of MIB
The following stock solutions of MIB were prepared:
yg/ml
A
B
C
D
27
-------
14/20
CO
14/20
12/18
>©<
14/20
B
Figure 5. Concentration apparatus. (A) and (B) are concentration vessels;
(C) is a micro-Snyder distillation column.
-------
A solution was prepared by making 10.00 mg up to 100 ml with ethanol. The
other solutions were prepared by successive 1 in 10 dilutions with ethanol.
A and B were mixed with similar camphor solutions to form reference mixtures,
C and D were used to make additions of MIB to water at 0.1-10 yg/1 during
recovery experiments.
Stock Solutions of d-Camphor, the Internal Standard
The following stock solutions of d-camphor were prepared:
A1
B1
C1
D1
Preparation was similar to that described for the MIB solutions.
Reference Mixtures of MIB and d-Camphor
Usually solution A was mixed with A1 in the following ratios by volume:
2:1, 1:1, 1:2, 1:5, and 1:10, respectively.
Analytical Method for MIB in Water
An exact volume of internal standard solution was added to 1 liter of
the water sample. For an expected MIB concentration of 1-10 yg/1, 5.0 ml of
C' was usually used; for 0.1-1 vg/1, 5.0 ml of D1 was usually used. The
solution was extracted with methylene chloride (1 x 25 ml and 1 x 10 ml) and
the extracts run directly into a concentration vessel. Total volume of the
combined extracts was 20 ml since about 15 ml of methylene chloride dissolved
in the water. After concentration to about 20-200 yl, the extracts were
analyzed by gas chromatography using a reference mixture with a similar
mixture of MIB to camphor. Peak height ratios were determined by dividing
the peak height of the MIB by that of the camphor.
Now assuming response of MIB relative to camphor is constant:
Wl
hl =kw7
W
where: h, and h2 are peak height ratios for sample and reference mixture,
respectively.
W-i is the weight of MIB in sample.
29
-------
w-i is weight of camphor added to sample.
W2 is weight of MIB per ml of A used to make reference mixture.
w2 is weight of camphor added to reference mixture per ml of A.
k is a constant
Eliminating k and rearranging:
hl w
wi = TT1
I ru
-
h x cr x 2
where: c is the concentration of camphor/ml of A'
d is the dilution factor for the particular stock solution of
camphor used with the sample.
v is the number of ml of camphor stock solution added to sample.
r is the ratio of A1 to A in the reference solution.
°'01 x 5
Thus when 5 ml of C1 (d = 0.01) was used: 2-methylisoborneol = T~ x
hl 5 2 hl
x 100 = fp x 1 ug/1. When 5 ml of D1 (d = 0.001) was used: MIB = ^- x
2 2
Synthesis of MIB
d-Camphor (60 g) in about 80 ml of ether was added dropwise with
magnetic stirring to about 260 ml of 1.7 M methyl lithium in ether at such a
rate as to maintain the solution under gentle reflux. Before use, the
reaction flask was flushed with dry nitrogen fed in through a three-way
stopcock in the top of the reflux condenser in one arm and out through the
dropping funnel in another arm. Since d-camphor dissolves readily in ether,
the solution was prepared in the dropping funnel, so as to avoid exposure of
the ether to the atmosphere. After completion of the camphor addition, the
solution was refluxed for a further 2 hours and allowed to stand at room
30
-------
temperature overnight. The solution, containing a white precipitate, was
then poured onto about 400 g of crushed ice and adjusted to pH 6 with glacial
acetic acid. The ether layer was collected by decantation, and the aqueous
layer was extracted with more ether (2 x 10 ml). The combined ether solutions
were dried over sodium sulfate and evaporated to dryness using a rotary
evaporator.
To the residue in 200 ml of ethanol was added 40 g of hydroxylamine
hydrochloride in 100 ml of water and 64 g of sodium hydroxide in 100 ml of
water to yield two clear layers of approximately equal volume. The solution
was refluxed for 8 hours and allowed to stand at room temperature overnight.
Water (about 230 ml) was then added with shaking,resulting at first in the
formation of a homogeneous solution and finally in the appearance of a
slight permanent precipitate. The solution was extracted with hexane
(2 x 100 ml) and the combined extracts washed first with 2N sodium hydroxide
(15 x 500 ml) and finally with water (2 x 50 ml). The hexane was dried over
sodium sulfate and evaporated on a rotary evaporator to yield 24.87 g (37.5%)
of MIB as a white crystalline solid.
The product was purified by chromatography on a column of silica gel
(6 x 34 cm, 500 g) using methylene chloride as eluant. Eluation of MIB
(23.04 g) commenced after 1 liter of eluate had been collected and was
complete after 3 liters. Only the first and last 3 g of eluted material
showed trace impurities by gas chromatography. The remaining material
was homogeneous by both gas and thin-layer chromatography.
The mass spectrum of the product was closely similar to that already
published (Medsker ejt al_., 1969) with a parent peak at m/e 168 and a very
strong base peak at m/e 95. The infrared spectrum was also closely similar
to that published (Medsker e_t al_., 1969). Specific rotation at 20° for the
D line using the Carl Zeiss polarimeter was -14.7° (c!5.5, ethanol) and
using the Bendix polarimeter was -14.9° (clO.76, ethanol) and -3.2° (clO.O,
hexane).
GEOSMIN SOURCE AND ANALYSIS
The geosmin was supplied by the U.S. EPA Municipal Environmental Research
Laboratory in Cincinnati. Gerber (1974) used Streptomyces sp CWW3 to
produce this compound for the EPA. The microorganism was grown at 28°C
in a broth medium and the resulting mixture was steam distilled and the
distillate extracted with methylene chloride and purified by column and gas
chromatography. This species of actinomycetes yielded both MIB and geosmin.
The analytical procedure for determining geosmin concentrations was
almost identical to that used for MIB. The only differences were the use
of 2-chloronaphthalene as the internal standard and a gas chromatograph
column temperature of 140°C.
31
-------
CHLOROPHENOLS
2,4-Dichlorophenol (2,4-DCP) and 2,4,6-trichlorophenol (2,4,6-TCP)
(Eastman Kodak) were used for the chlorophenol studies.
Chlorophenol Analysis
Analysis of chlorophenols was performed on a 5750B Hewlett Packard gas
chromatograph equipped with a pulsed Nib:i electron capture detector. A
one-foot coiled glass column packed with 10% DEGS on 80/100 mesh Supelcoport
(Supelco, Inc.) was used for the separation of 2,4-DCP and 2,4,6-TCP. The
column was conditioned three hours at 200°C with a flow of 20 ml/min carrier
gas. Operating conditions were: Column temperature = 170°C, Detector
temperature = 230°C, Injection port temperature = 200°C, Carrier gas = dried
5% CH4-95% Argon at 50 ml/min flow, Pulse interval = 150 ysec.
2,4-Dibromophenol (2,4-DBP) obtained from the Aldrich Chemical Company was
selected as the internal standard since it is eluted close to and completely
separate from 2,4-DCP and 2,4,6-TCP at these operating conditions; at a given
concentration, the detector response is intermediate to these two chloro-
phenols; and its physical properties are very similar to the chlorophenols
as shown in Table 8.
TABLE 8. CHARACTERISTICS OF THE CHLORO- AND BROMOPHENOLS STUDIED
(Handbook of Chem. & Physics, 1967)
Compound
M.W.
m.p.
°C
b.p.
°C
Solubility
g/100 ml
H20
PKa
2,4-DCP
2,4,6-TCP
2,4-DBP
163.01
197.46
251.92
45
68
40
210
244.5
238.9
0.46
0.08
0.19
7.85
6.00
7.30
Nanograde toluene (Mallinckrodt) was chosen as the solvent for sample
preparation, for GC analysis, and for extraction of the chlorophenols from
aqueous solution for several reasons: the solvent peak does not overlap
with any of the peaks of interest nor are there any interfering impurities;
it is practically immiscible with water; and the partition coefficients for
extraction of chlorophenols from water are high enough to indicate adequate
recovery (Korenman, 1974).
Stock standard solutions containing 1 yg/yl 2,4-DBP and another
containing 1 ug/yl of both 2,4-DCP and 2,4,6-TCP were prepared as outlined
by the U.S. EPA (1971) for organic pesticides. Working standards were
prepared from the stock solution using a micro-syringe and then stored at
32
-------
5°C. A typical working curve for 2,4-DCP is shown in Figure 6. Ratios of
the peak heights of the chlorophenols to that of 75 yg/1 2,4-DBP were computed
and plotted versus chlorophenol concentration. The chromatogram in Figure 7
showing 80 yg/1 of both 2,4-DCP and 2,4,6-TCP and 75 yg/1 2,4-DBP is typical
of the data used to obtain the curves. Day-to-day variation in detector
response and column performance was significant enough that standard curves
had to be prepared prior to each set of chromatographic analyses.
Extractions were performed using 100 ml of the adsorbate solutions
adjusted to pH 2.0 to ensure that the less soluble neutral species were
present. When the chlorophenol concentration was high, dilutions were
prepared to ensure peak heights fell on the standard curves. Prior to
extraction, 7.5 yl of 50 mg/1 2,4-DBP in distilled water was added to the
solution. Extractions were performed in 250 ml separatory funnels with 5 ml
toluene resulting in a concentrating factor of 20X. The solutions were
shaken for five minutes as recommended by Korenman (1974), and 30 minutes
were allowed for complete separation of the phases. Recoveries of the
chlorophenols relative to 2,4-DBP were found to be 100 percent for 2,4-DCP
and 104 percent for 2,4,6-TCP in the concentration range studies. For this
procedure the estimated sensitivity limits for 2,4-DCP and 2,4,6-TCP were
found to be 1 yg/1 and 0.01 yg/1 in water, respectively.
POLYNUCLEAR AROMATIC HYDROCARBONS
Various methods were evaluated to determine the concentration of the
polynuclear aromatic hydrocarbons (PAH), anthracene (Aldrich) and benz-
anthracene (Eastman). A fluorometer (Turner Model 110) produced a linear
response with concentration. This procedure was satisfactory for pure
solutions of PAH, but in the presence of humic substances which fluoresce,
the contribution to fluorescence by PAH alone would be impossible to deter-
mine. Early attempts were made to extract the PAH from humic acid solution
with toluene and to measure the fluorescence of the extract. Some of the
humic material was also extracted, however, and good results using fluor-
escence were not obtained. When the aqueous solution was made basic to a
pH of 11, less material was extracted into the organic phase. It was also
found that UV absorbance did not offer adequate sensitivity; a 5 cm path
length cell was satisfactory for the determination of 20 yg/1 of benzanthra-
cene as a lower limit, for example.
Gas chromatography was chosen for the analytical determination of PAH
in solution with humic acid. The solution was extracted with cyclohexane
(Mallinckrodt) after being made alkaline with several drops of concentrated
NaOH to minimize the extraction of humic substances. Cyclohexane was used
because it tended to produce less of an emulsion. The SP-2100 column used
for the quantitation of MIB and geosmin was found to be suitable for the
PAH work. A variety of internal standards was checked, among them carbazole,
phenanthrene, fluorene, fluoranthene, 2-methyl naphthalene, 9-methyl anthra-
cene and trans-stilbene. Fluoranthene eluted between anthracene and benz-
anthracene and thus showed promise as an internal standard if competitive
experiments between these two PAH compounds were performed. Fluorene was
masked by the solvent front and carbazole and phenanthrene eluted too close
33
-------
O.7
0.6
CL-
OD
O
* 0.5
CM
O.
O
O
CM
O
0.4
•55 0-2
JC.
O
0)
CL 0. I
Stondord Curve For
2,4— Dichlorophenol
(2,4 — Dibromophenol =
I
I )
20 40 60 80 100 120 140 160
Concentration 2,4—DCP in Toluene, /ig / I
180
Figure 6. Standard curve for 2,4-dichlorophenol.
-------
2,4,6-TCP
2,4-DBP
Retention Time 1
2,4-DCP =0.75min.
2,4,6-TCP - 1.35 min.
2,4-DBP =1.85 min.
Concentrations !
2,4-DCP = 80 fig//
2,4,6-TCP = 80 fig//
2,4-DBP = 75 fig//
t t
Inject Atten. = 10x16
Atten. 10x4
Figure 7. Typical chlorophenol chromatogram
35
-------
to the anthracene peak. Trans-stilbene eluted after the solvent peak and
before anthracene and was a logical choice. A portion of it was lost during
evaporation, however, possibly due to partial isomerization to the cis-
isomer. 9-Methyl anthracene (Aldrich) which eluted after the anthracene
was used instead. Tests were performed with standard solutions and it was
found that extraction of a 75 ml sample successively with 10 ml, 5 ml and
5 ml volumes of cyclohexane removed 99 percent of anthracene from solution.
It was also shown that removal of solvent with the micro Kuderna-Danish
(K-D) evaporator resulted in no absolute losses of PAH.
A 6 foot, 3 percent SP-2100 (Supelco) column was used with a nitrogen
carrier gas at a flow rate of 40 ml/min. The temperature settings were
210° for the column, 260° for the injector, and 280° for the detector. At
these conditions, retention times for the anthracene and methyl-anthracene
were 1.9 minutes and 2.5 minutes, respectively, as shown in Figure 8.
Solutions of anthracene to be analyzed were made alkaline to pH of 11
and extracted with three volumes of cyclohexane. The cyclohexane phases
were combined and spiked with a known amount of internal standard. This
was then concentrated to about 200 yl in a micro K-D evaporator. If further
concentration was necessary it was done on a rotary evaporator without a
heating bath similar to the MIB analytical procedure.
ADSORPTION TEST PROCEDURES
Isotherm Tests
Humic Substances --
The tests were conducted by adding an accurately weighed dose of carbon
to a series of 250 ml bottles containing solution of known concentration.
The solution volume was 150 ml. The bottles were placed on a shaker for 4-7
days at room temperature, after which time the equilibrium solution concen-
tration was determined. Several identical samples were analyzed daily to
verify that equilibrium was achieved.
A buffer was not used for nonfractionated commercial humic acid or well
water, but it was necessary to use it for those humic substances which were
fractionated using a phosphate eluant. In order to prepare a phosphate
buffered solution at a fixed concentration of phosphate when the organics
also contain phosphate, the Vanadomolybdophosphoric Acid test was employed
to determine the proper amount of phosphate to add (Standard Methods, 1975).
The phosphate concentration of the solution containing only the organics was
measured and the additional required phosphate was then added. Equilibrium
concentrations were determined by fluorescence or, in some cases, UV
absorbance.
MIB and Geosmin --
Stock MIB and geosmin solutions were made by dissolving the compound in
ethanol and then diluting with deionized distilled water. The desired
36
-------
Retention Time !
Anthracene ! 1.87 min.
Methyl Anthracene ! 2.5 min.
Concentrations !
Anthracene =640mg/j?
Methyl Anthracene = 965 mg/Jl
t
Inject 2fj.f Atten. 8x10^
Figure 8. Typical anthracene chromatogram,
37
-------
quantity of solution was prepared in this manner immediately before starting
the tests. MIB is very stable in ethanol and the ethanol thus introduced
into the adsorption test systems, on the order of 1 to 5 mg/1, had no notice-
able effect on the isotherms as shown by comparing isotherms prepared using
different ethanol concentrations, consistent with its poor adsorption charac-
teristics.
Most of the batch data for MIB and geosmin were obtained using 1 liter
samples in 2-1/2 liter cylindrical bottles. The tests were generally con-
ducted by setting up a series of bottles each with the same concentration of
adsorbate but with different doses of carbon. Blanks were also used which
contained no carbon. It was determined that 8-9 days were necessary for
equilibration at the yg/1 concentration level on the gyratory shaker while
only 4-5 days were required when there was more vigorous shaking on a
reciprocating shaker or if the bottles were tilted about 30° from the
horizontal on the gyratory shaker. Other tests were conducted by putting a
large dose of carbon into a single bottle with a high concentration of
adsorbate. After equilibration, the solution was decanted and another
solution containing adsorbate was contacted with the carbon and the sample
was then re-equilibrated. This process was repeated several times to
establish an adsorption isotherm.
A considerable amount of data were gathered for MIB using a procedure
which did not result in equilibrium. Some of these data are presented in
this paper because they are important when carbon is used under nonequi-
librium conditions.
Chlorophenols --
Chlorophenol stock solutions were prepared just prior to use by dis-
solving a known amount of compound in distilled water adjusted to about pH 9
with NaOH to ensure that the compound was dissolved. 2,4-DCP and 2,4,6-TCP
are weak acids and were studied in both their dissociated and undissociated
forms. To ensure the predominance of either the neutral or aniom'c form,
the pH of the solution was adjusted to at least one unit below or above the
compound's pKa value (Ward and Getzen, 1970). The solutions were buffered
with 10-2 M phosphate salts. A series of 2-1/2 liter bottles were filled
with one liter of the adsorbate solution. Amounts of Filtrasorb 400 carbon
were added to each bottle to yield the desired equilibrium concentration
with at least two bottles containing no carbon to serve as blanks. The
bottles were sealed with Teflon-lined caps and placed on a model G10 gyra-
tory shaker (New Brunswick Scientific) operated at 200 rpm. The shaker
compartments were angled to ensure adequate mixing. At least 12 days were
allowed for equilibration. All studies were performed at room temperature
which varied ±2°C. Prior to extraction and analysis, the solutions were
filtered through fiberglass filters to eliminate carbon fines. Studies were
performed which indicated that chlorophenols did not adsorb on the filter.
Column Tests
MIB and geosmin breakthrough curves were developed using continuous
flow column systems and deionized water. Two grams of 40 x 50 mesh carbon
(unless stated) were placed in 1.27 cm diameter columns. In some cases, the
38
-------
carbon was sandwiched between several cm of sand, also of 40 x 50 mesh size,
although this had little effect on column performance. Bed depth was 2.7 cm
and empty bed contact time was on the order of 0.1 minute. The flow rate of
2 liter/hr (6.5 gpnvft2) was kept constant by use of a metering pump. The
tests were conducted at room temperature of approximately 23°C. Samples were
collected on a regular basis and then analyzed to obtain a record of effluent
concentration vs. time.
39
-------
SECTION 5
RESULTS AND DISCUSSION
HALOFORM FORMATION POTENTIAL OF THE HUMIC SUBSTANCES
The haloform formation potential of the humic substances used in this
study was determined to further characterize these materials. In Figure 9
chloroform formation with time by the various humic substances is shown. It
is interesting to note that not all humic materials will produce the same
amount of chloroform per unit weight of TOC. The yield of chloroform from
the soil humic acid is lower than from the soil fulvic acid while soil
fulvic acid and leaf fulvic acid give nearly the same result. The large
difference between commercial humic acid and soil humic acid illustrates
that the source of the material is important.
In Figure 10 we observe that various molecular weight fractions of soil
fulvic acid yield about the same amount of chloroform as the unfractionated
soil fulvic acid. This would seem to indicate that molecular weight frac-
tions have the same density of those functional groups which produce chloro-
form as the unfractionated group of compounds. In Figure 11, similar
results are shown for soil humic acid fractions. The differences between
the curves are not considered significant.
Another experiment was performed to evaluate the removal of haloform
precursor by activated carbon. One liter of 5 mg/1 TOC commercial humic
acid was equilibrated with an amount of activated carbon to give about 50
percent removal as indicated by fluorescence. After one week the blank
solution was diluted to give an identical fluorescence reading as the
equilibrated solution and both were chlorinated using the above procedure.
After 8 hours, the chloroform concentration of the diluted blank was
115 yg/1. The chloroform yield of the carbon equilibrated solution was
198 ug/1, a 72 percent increase over the blank. Because it has been shown
that the various molecular weight fractions have the same haloform formation
potential as the unfractionated material, this finding is consistent with
our adsorption results which show that carbon preferentially removes the low
molecular weight compounds which are also the most highly fluorescent. Thus
the decrease in fluorescence is not matched by an equal decrease in chloro-
form formation and is not matched by an equal decrease in TOC.
ADSORPTION OF HUMIC SUBSTANCES
The results of batch studies using commercial humic acid are shown in
Figure 12. The data are based on analysis by fluorescence and UV absorbance,
40
-------
300
- 200
c
o
c
V
o
c
o
O
e
t_
o
o
jO
£L
O
100
10 mq/J( C\2
2.5 mg/Jj TOC
pH =7.0
Commercial Humic Acid
Soil Fulvic Acid
0
1 1 1 1 1 1
) 10 20 30 40 50 60 70 80
1
90 10
Time , hours
Figure 9. Chloroform formation from humic and fulvic acids.
-------
o»
c
o
c
0)
o
c
o
o
E
o
H—
O
_o
JC.
u
10 mg/ CI2
2.5 mg/!l TOO
pH = 7
Unfroctionated
MW > 50,000
MW > 5000
MW < 5000
60
7O
80
Figure 10.
50
Hours
Chloroform formation from soil fulvic acid.
90
100
-------
-pi
GO
240r
220
200
ISO
160
ppb "40
0 Unfroctionoted
• MW > 50,000
• MW < 10,000
10 mg/5? CX2
pH = 7
TOC =2.5 mq/j
IOO
Figure 11. Chloroform formation from soil humic acid.
-------
1000
o>
o»
E
100
o
o
00
10
A F- 400, pH 5.5 (FLUOR.)
D F- 200, pH 8.5 (FLUOR.)
O F-200, pH 5.5 (FLUOR.)
V F-200, pH 5.5 (UV)
0.01
O.I I 10
Equilibrium Concentration, mg/X
100
Figure 12. Adsorption isotherms for commercial humic acid measured by fluorescence and UV.
-------
as indicated. F-400, which has a larger average pore size and about 200
m2/g more surface area than F-200, showed a much higher capacity for humic
acid. The data show no difference in the adsorption isotherm using F-200 as
the pH is changed from 5.5 to 8.5, however. The analytical technique used
to determine humic acid concentrations is shown to have a very significant
effect on the results. The material which absorbs UV light adsorbs less
strongly than the material which fluoresces.
Figure 13 shows data for adsorption of organic matter on F-200 from
well water compared to the isotherm for commercial humic acid reproduced
from Figure 12. There appears to be no significant effect of pH on adsorp-
tion of humic material from the well water. There is also more fluorescent
organic matter removed per gram of carbon from well water than from the
commercial humic acid solution. Based upon our results presented earlier
which showed lower molecular weight material to fluoresce more and be
adsorbed to a greater degree, the presence of a greater number of low molec-
ular weight compounds is indicated. However, this remains to be demonstrated
Figure 14 demonstrates the effect of phosphate buffer concentrations on
the soil fulvic acid adsorption at pH 7. The higher buffer concentrations
yield higher capacities, the specific reason for which is not apparent.
Figure 15 demonstrates the effect of pH on soil fulvic acid adsorption
in the presence of 0.001 M phosphate buffer. Due to the acidic character
of the natural organics, a decrease in pH renders them less soluble and
thus more readily adsorbed on activated carbon. This rather significant
effect is in contrast to the absence of a pH effect in the pH 5 to 8.5 range
reported previously for commercial and well water humic substances. A pH of
7 was chosen for all subsequent isotherms.
Figure 16 compares the adsorption of the various humic materials on
F-400. It is interesting to note the different capacities. The soil humic
acid adsorbs better than the soil fulvic acid due to its lower solubility
in water. This cannot be generalized to all fulvic and humic material,
however, as the leaf fulvic acid was adsorbed best of all. It should be
able to penetrate smaller pores that are inaccessible to the humic acid
molecules.
Figures 17, 18 and 19 show the extent of adsorption of various molecu-
lar weight fractions on adsorption capacity. In all cases we observe a
steady decrease in extent of adsorption as the molecular weight of the
material increases. This probably is due to the inability of the large
molecules to enter the smaller pores of the carbon.
45
-------
O> 10
X.
o»
E
c
o
O
C
o
O
4)
O
O
«4—
W
3
CO
Well Water Organic Matter
pH 5.5 and 10
Humic Acid (From Fig. 12}
pH 5.5 and 8.5
F - 200 Analysis By Fluorescense
l 10
Equilibrium Concentration, mg/fl
Figure 13. Adsorption of well water organic matter.
46
-------
100
o>
E
c
o
-------
IOO
-pi
oo
to
p H = 3.0
F—400
0.001 M PO.
O.I
.01
O.I
10
100
Figure 15. Effect of pH on the adsorption of soil fulvic acid.
-------
100
Leaf Fulvic Acid-
o>
X
o>
E
c
V
u
c
o
o
0)
u
o
10
-Soil Humic Acid
Soil Fulvic Acid
-Commercial Humic Acid
p H =7.0
P04 = O.OOIM
F -400
0.01
O.I I 10
Equilibrium Concentration, mg/J As TOC
100
Figure 16. Adsorption of various types of humic substances,
-------
IOO
en
O
o>
E
o
i ">
4>
O
C
o
O
o
o
p H = 7.0
P04=O.OOI M
F -400
O.Ol
Unfractionaled
MW < 5000
MW >5000
MW > 50,000
O.I I "0
Equilibrium Concentration, mg/* As TOC
too
Figure 17. Adsorption of molecular weight fractions of soil fulvic acid.
-------
100
o>
E
o
I
«
u
c
o
o
Q>
U
O
•*-
w
=J
in
10
pH =7.0
P04 =0.001 M
F -400
MW < 10,000-
Unfractionated
MW 10,000 —50,000
MW > 50,000
O.Ol
O.I I 10
Equilibrium Concentration , mg/> As TOC
100
Figure 18. Adsorption of molecular weight fractions of soil humic acid.
-------
100
O»
^s
o>
E
c
o
I0
u
o
O
o
o
o
Unfroctionated
MW < 5000
MW > 5000
p H =7.0
P04 =0.001
F -400
I
0.01
O.I I 10
Equilibrium Concentration , mg/J As TOC
100
Figure 19. Adsorption of molecular weight fractions of leaf fulvic acid.
-------
MIB SYNTHESIS AND ANALYSIS1
Enantlomeric Form of Natural 2-Methylisoborneol
MIB can exist in two optically-active forms with the current Chemical
Abstracts names of (1-R-exo_)- and (1-S-expJ- 1,2,7,7-tetramethylbicyclo
[2.2.1] heptan-2-ol as shown in Figure 20 (Personal communication from D.
Weisgerber, Chemical Abstracts Service, 1976). These are derived from d-
and 1-camphor, respectively (Figure 20). In the literature natural MIB has
always been depicted as the S^ form, although no evidence for this has ever
been reported. Since we were interested in studying not only the adsorption
of the natural form on carbon beds but also in examining for any concurrent
biological activity, we needed to determine which of the two forms occurred
naturally so as to synthesize the appropriate one for our experiments.
Accordingly, we have measured the specific rotation of a natural sample
of MIB and so determined its enantiomeric form. We isolated the sample by
liquid chromatography on silica gel of a product obtained by Dr. Nancy Gerber
of Rutgers University from the culture of Streptomyces sp. CWW3 (Gerber,
1974), for the U.S. EPA and supplied to us by the U.S. EPA Municipal Environ-
mental Research Laboratory- The product was mostly geosmin and only 14 mg
of pure MIB was obtained. By measuring the optical rotation in ethanol to
within 0.001°, a specific rotation of -14° was obtained, in good agreement
with the -14.8° reported in the literature (Malkonen, 1964) for MIB obtained
from d-camphor. Thus natural MIB exists in the R form.
Synthesis of 2-Methylisoborneol from d-Camphor
Since the first preparation was reported in 1901 (Zelinsky, 1901) MIB
has been prepared many times by the action of methylmagnesium halides or
methyllithium on camphor (Medsker et al., 1969; Rosen et al., 1970; Malkonen,
1964; Capmau et al. , 1968; Fieser and Ourisson, 1953; Zeiss and Pease, 1956).
The difficulty with this seemingly easy preparation is that much of the
camphor reacts to form an enolate and during the work-up this reverts to
camphor, which until now has been difficult to eliminate from the product.
Direct separation of the MIB from the camphor by chromatography has
been reported. Liquid chromatography on alumina (Malkonen, 1964; Capmau
et al., 1968) has been used, though it is apparently difficult to eliminate
the camphor entirely (R. G. Webb and N. N. Gerber, personal communications,
1976). Preparative gas liquid chromatography (Medsker et al., 1969; R. G.
Webb and N. N. Gerber, personal communications, 1976) apparently yields a
pure product but is limited to the preparation of milligram quantities.
Another approach (Malkonen, 1964) to the problem has been to repeatedly
1. Much of the material in this subsection was taken from "2-Methyliso-
borneol, Improved Synthesis and a Quantitative Gas Chromatographic Method
for Trace Concentrations Producing Odor in Water," by N. F. Wood and V. L,
Snoeyink, Jour. Chromatogr. 132, 405 (1977), with the permission of the
copyright owner.
53
-------
d — Camphor
I— Camphor
HO
2— Methylisoborneol
R S
Figure 20. Stereochemical structures of 2-methylisoborneol and
camphor enantiomers.
54
-------
treat the crude product with more reagent until the camphor is reduced to
such a small proportion that it can be eliminated by recrystallization.
Finally, MIB has been separated from the camphor through its chromate
ester (Fieser and Ourisson, 1953; Zeiss and Pease, 1956). Although reduc-
tion of the ester with hypophosphorous acid or saponification yielded impure
MIB (Fieser and Ourisson, 1953), reduction with lithium aluminum hydride on
a small scale appeared to yield pure material (Zeiss and Pease, 1956).
In our synthesis of MIB we chose methyl lithium as reagent rather than
a methylmagnesium halide, partly because this was conveniently available
commercially, and partly because we hoped it would give the higher yield.
Figure 21 shows the gas chromatogram from the crude product obtained from
60 g of camphor and 1.3 equivalents of methyllithium after reaction at
reflux for 2 hours and overnight at room temperature. Since the response of
MIB relative to camphor was 0.90, the observed peak height ratio of 0.50
corresponds to 67 percent of unchanged camphor. When the amount of reagent
was increased to 2.0 equivalents, the observed peak height ratio increased
to 0.56 corresponding to only a slightly decreased amount of 64 percent of
unchanged camphor. Decreasing the reaction time to 2 hours without reflux
had no effect on the amount of unchanged camphor. Malkonen (1964) has
reported only 37 percent of unchanged camphor using 3.0 equivalents of methyl'
magnesium iodide. Thus, although the Grignard reagents may not be as con-
venient to use as methyl!ithiurn, they appear to give better yields of MIB.
To eliminate the large amount of camphor in our crude product we
decided to selectively react the camphor quantitatively to form a derivative
that would be more easily separated than the camphor itself. Many reagents
were tested but most were unsuitable because of the particularly unreactive
nature of camphor. The exception was alkaline hydroxylamine at reflux.
Using the reaction conditions of Lenz (1911) and monitoring the reaction
mixture by gas chromatography, it was found that 19 percent camphor remained
after 2 hours but reaction to the oxime was complete after 7 hours. A chro-
matogram of the reaction product is shown in Figure 22.
With the camphor in the form of the oxime it was rather easily elimi-
nated by taking advantage of its acidic properties. First the alkaline
reaction solution was diluted with water and extracted twice with a small
proportion of hexane. Gas chromatography on the residual solution (Figure
23) and the combined extracts (Figure 24) showed that less than 50 percent
of the oxime and more than 98 percent of MIB was in the extracts. Finally
repeated washing of the extracts with aqueous sodium hydroxide gradually
removed all of the oxime without loss of MIB. The product was essentially
pure by gas chromatography (Figure 25).
In another experiment the reaction solution containing the oxime was
not diluted with water as above but extracted directly with hexane. Par-
titioning into the hexane layer was reduced but selectivity improved. Four
extractions removed a total of 92 percent of the MIB and only 20 percent of
the oxime from the reaction solution, so that complete elimination of the
oxime was achieved with far fewer washings of aqueous sodium hydroxide.
55
-------
Minutes
Figure 21. Chromatogram from the product of the action of
methyl lithium on d-camphor.
56
-------
o
o>
c
k_
o
.0
o
in
I CO
0
Minutes
Figure 22. Chromatogram from the product of the action of methy11ithium
on d-camphor after treatment with alkaline hydroxylamine.
57
-------
4
Minutes
8
Figure 23. Chromatogram from the residual aqueous solution from the
hydroxylamine reaction after hexane extraction.
58
-------
4 6
Minutes
Figure 24. Chromatogram from the hexane extract after the
hydroxylamine reaction.
59
-------
0
8
Minutes
Figure 25. Chromatogram from the hexane extract in Figure 24 after
exhaustive washing with 2 N sodium hydroxide.
60
-------
Thin-layer chromatography in addition to gas chromatography was used to
monitor the above work. Table 6 gives details for MIB and related camphor
derivatives. Note that the bright crimson color reaction of MIB is quite
characteristic and develops immediately with a sensitivity of less than 1 yg.
Unfortunately camphor gives only an uncharacteristic translucent spot that
develops erratically over a long period with a sensitivity of 10 yg at best.
(Other sprays for camphor give even poorer sensitivities [Stahl, 1969].)
Camphor oxime does give a characteristic color reaction but this develops
only slowly and again with a sensitivity of 10 yg. Despite the limitations
with camphor and its oxime, thin-layer chromatography was useful in that it
detected impurities in the MIB not revealed by gas chromatography, and thus
indicated the need for final purification of the product by liquid chroma-
tography.
Liquid chromatography on alumina has usually been used for the purifi-
cation of MIB, but our attempt to use basic aluminum oxide (Woelm) led to
complete decomposition. This also occurred when an old preparative column
of silica gel was used. Even on Florisil there appeared to be slight
decomposition as elution was delayed for 10 hours. This was evidenced by the
appearance of several very slow developing translucent spots on thin-layer
chromatograms of the product. However, MIB homogeneous by thin-layer
chromatography was obtained by chromatographing on fresh silica gel and
eluting with methylene chloride within 2 hours.
Attempted Preparation of Electron-Capturing Derivatives from 2-Methyliso-
borneol
Since we were interested in analyzing aqueous solutions of MIB down to
0.1 yg/1 by gas chromatography, the use of the sensitive electron-capture
detector seemed appropriate. Accordingly, attempts were made to prepare a
suitably electron-capturing ester.
Heptafluorobutyric anhydride alone did not react significantly with
MIB at room temperature, and even at 60°, 12 hours was required before most of
the MIB had reacted. The product was a mixture of two substances whose
retention times by gas chromatography relative to that of MIB (0.56 and
0.60) indicated that they were olefins. In benzene in the presence of tri-
ethylamine at room temperature, reaction was virtually complete in 1 hour but
a similar product resulted. Thin-layer chromatography showed only one major
spot (Rf 0.94), which had a bright crimson appearance. After 24 hours the
initially predominant substance (relative retention time 0.60) appeared to
have partially converted to the other substance (relative retention time
0.56).
Hydriodic acid (49 percent) reacted immediately with a solution of MIB
in acetone to give a product that showed a single crimson spot (Rf 0.96) on
thin-layer chromatography. The product in chloroform showed infrared
absorption at 3080 (medium), 1740 (weak), 1655 (medium), and 880 cnT1
(strong) corresponding to a methylene with standard absorption at 3080
(medium), 1800 ^ 1750 (medium), 1655 (medium), and 890 cm'1 (strong)
(Nakanishi, 1962). Again an olefin was produced rather than the desired
ester.
61
-------
In contrast, hydrobromic acid (22 percent) reacted almost completely
with MIB in ethanol in 30 min at room temperature to produce little olefin.
The major product had a gas chromatographic retention time relative to that
of methylisoborneol of 1:10 and gave a crimson spot on thin-layer chromatog-
raphy with a slightly smaller Rf (0.42) than that of MIB (0.49). The major
change in the infrared absorption in potassium bromide was a shift of the
C-0 stretching band from 1095 cnH in MIB to 1030 crrf' in the product, and
this indicated transformation of the tertiary alcohol to a secondary
(Nakanishi, 1962). Both the chromatographic and the infrared evidence
indicate rearrangement of the MIB to a secondary alcohol, possible 4-methyl-
isoborneol (Toivonen, 1968).
These failures to develop a suitable derivative to use with the
electron-capture detector meant that we were forced to use the flame ioniza-
tion detector. To get the desired sensitivity for trace concentrations of
MIB in water, the techniques explained below had to be developed to make
extreme concentrations of water extracts.
Extraction
MIB resembles its precursor camphor in volatility, in gas chromato-
graphic retention times, and in having extremely high distribution ratios
for partitioning into organic solvents from water. Camphor then quickly
became our choice as internal standard in the analysis of MIB.
The favorable partitioning into organic solvents meant that usually
only small proportions of extractant (1-2 percent) were necessary to extract
MIB from water, and this helped both to reduce the amount of coextractives
and also the degree of concentration necessary before analysis. However,
considerable concentration was still necessary and this severely limited
the choice of solvent for extraction. MIB co-distills with some solvents
and with others the grade readily available has impurities that concentrate
sufficiently to cause interference with the analysis. Chloroform was quite
satisfactory for analysis down to about 1 yg/1 but below this solvent
impurities tended to interfere. Methylene chloride was entirely satisfactory
and in addition was readily removed because of its outstanding volatility.
Concentration of Extracts
To achieve the desired sensitivity of 0.1 yg/1 for MIB in water, it was
evident that the extract from 1 liter of water would have to be concentrated
to as little as 20 yl before gas chromatography. Because both MIB and
camphor readily sublime, special attention had to be given to the method of
concentration.
Removal of solvent through a micro Snyder distillation column (Figure 5)
was first tried. This method has the advantage that several samples can be
concentrated simultaneously with little attention and at the end of evapora-
tion the appartus is self-rinsing (Burke et al., 1966). Concentration
vessels of the type A and B shown in Figure 5 were used. These were fabri-
cated from 25 ml round-bottomed flasks and comfortably held the 20 ml of
62
-------
solvent from the extraction of 1 liter of water for MIB. With type A,
access to the concentrate with a syringe was a little awkward but it was
generally preferred because of its simplicity. Type B had to be assembled
with water as a seal on the lower joint to avoid small losses during the
concentration, but was advantageous in that the lower tube could be removed
later for easy access to the concentrate. The smallest volume attainable
with this apparatus was about 200 yl, which was sufficiently low to allow
analyses down to about 1 yg/1.
Both relative and absolute recoveries for the evaporation were deter-
mined using 1-10 yg of MIB in 20 ml of various solvents. Absolute recover-
ies were determined as near as possible from the volume of concentrate as
calculated from its weight. Disappointingly, pentane and hexane gave only
about 50 percent absolute recoveries and relative recoveries were very
erratic. However, methylene chloride, chloroform, and carbon tetrachloride
gave quantitative absolute and relative recoveries. Of these latter sol-
vents, methylene chloride was the obvious choice for the analysis because
of its greater volatility and availability in pure grades.
To attain the highest sensitivity desired, it was necessary to determine
how to reduce the initial concentrate of 200 yl to about 20 ul without loss
of MIB. This was attempted initially in the lower tube of vessel B by
equipping this tube with a small Snyder column. Although relative recoveries
were excellent absolute losses occurred which largely nullified the effect
of solvent removal.
However, concentration of the initial concentrate in vessel A or the
lower tube of B on a rotary evaporator without a heating bath was successful
without loss of MIB or camphor. The same result was possible by simply
connecting the vessels to a water aspirator, but then there was risk of
losing the sample by bumping.
Evaporation of Column Eluates
Before carrying out the purification of the few milligrams of natural
MIB available by liquid chromatography, losses were anticipated during the
evaporation of the column eluate to dryness. However, preliminary tests
with synthetic MIB showed that, although 2.5 mg in 50 ml of hexane was
evaporated to dryness on a rotary evaporator at 45° with a recovery of only
30 percent, similar evaporation at 35° using methylene chloride gave nearly
quantitative recovery. Accordingly, the liquid chromatography was worked
out on the basis of methylene chloride as eluant.
Gas Chromatography
The initial parts of this work were carried out by gas chromatography
using a commercially prepared glass column of 3 percent OV-1. First tests
on the use of camphor as internal standard gave satisfactory results in
the determination of MIB in water at 100 yg/1. However at 10 yg/1 unfavor-
able adsorption effects were observed with the smaller amount of internal
standard used. Peak height response relative to that of MIB decreased
markedly and peak width and retention time increased to the point that
63
-------
resolution from MIB was mostly destroyed.
The column was therefore replaced with one of DMCS-treated glass packed
with 3 percent SP-2100, a similar stationary phase to that of 3 percent OV-1
but with improved characteristics. No untoward adsorption effects were
noted with the column. Providing it was not subjected to harsh heat treat-
ment, this type of column proved to be very stable in performance over a
period of many months.
In the quantitation of MIB in this work, peak height relative to camphor
for the sample was compared with that of a standard mixture showing a similar
ratio. Standard mixtures in ethanol were prepared by mixing a stock solution
of MIB (10 mg/100 ml) with that of camphor (8 mg/100 ml) in various propor-
tions. When stored in stoppered volumetric flasks at room temperature, they
proved to be very stable over a period of many months. It was assumed that
there was a linear relationship between the amount of MIB in the sample and
the peak height ratio, and that this ratio was unaffected by dilution of the
solution.
To substantiate these assumptions the response factor was determined
for MIB from some standard reference mixtures and their corresponding 1:10
dilutions as shown in Table 9. Evidently the response factor is not appre-
ciably affected by dilution. Also, although there is a slight fall in the
response factor with the undiluted mixtures as the proportion of MIB is
reduced greatly, this is not significant enough to affect accuracy if widely
differing peak height ratios are not compared.
TABLE 9. RESPONSE FACTOR OF MIB FOR VARIOUS PROPORTIONS WITH
THE INTERNAL STANDARD AND AMOUNTS GAS CHROMATOGRAPHED
Ratio of (A/A1)
volumes MIB
stock solution to
that of internal
standard in
reference mixture^
Amount of ?
chromatographed, pg
Internal
MIB Standard
(W)
(Wi)
Peak height
of MIB
(H) relative to
that of internal
standard (Hi)
Response
Factor
H Wi
Hi W
2
1
0.2
133.3
13.33
100.0
10.00
33.33
3.333
53.33
5.333
80.0
8.000
133.3
13.33
2.15
2.14
1.04
1.05
0.196
0.211
0.860
0.856
0.832
0.840
0.784
0.844
Stock solution of MIB (A) = 10.0 mg/100 ml in ethanol.
2Internal standard solution (A1) = 8.0 mg/100 ml
Two microliters of each reference mixture was chromatographed both undi-
luted and as a 1:10 dilution with ethanol.
64
-------
Recoveries from Various Water Types
Recoveries of MIB added to various waters in a small amount of ethanol
are given in Table 10. Blanks equivalent to 0.1 yg/1 were carried out for
these waters and also for a sample of lake water, but no interference was
noted. A typical chromatogram from a recovery at 0.1 yg/1 is shown in
Figure 26.
TABLE 10. RELATIVE RECOVERY OF MIB FROM VARIOUS WATERS
Added,
MIB
yg/1
Camphor
(internal
standard)
Recovery of MIB, %, Relative to Camphor
Type of Water Used
Distilled Tap Well
Polluted
Stream
10
5
1
1
0.5
0.1
4
4
4
0.4
0.4
0.8
100.0 97.8 98.2; 97.5
97.7; 94.9 97.4; 97.3
100.4 100.6 100.7
96.6
100.1
101.8
99.0
103.7
Mean relative recovery,
% ± standard deviation 100.3±0.2 98.5±2.4
98.2+1.4
101.4+3.3
ADSORPTION OF MIB AND GEOSMIN
Results
Isotherm data for MIB are shown in Figure 27. Included are the iso-
therms for MIB on F-200 carbon in the presence of both 10 and 100 mg/1 of
commercial humic acid. Because of the small dosages of carbon used in the
tests, negligible reductions in humic acid concentration were observed and
final concentrations were approximately the same as the initial
1. Much of the material in this subsection has been taken from "Activated
Carbon Adsorption of the Odorous Compounds 2-Methylisoborneol and Geosmin"
by D. R. Herzing, V. L. Snoeyink and N. F- Wood, Jour. Amer. Water Works
Assoc., 69, 223 (1977), with the permission of the copyright owner.
65
-------
Minutes
Figure 26. Chromatogram obtained in a study of the recovery of MIB
added to tap water at 0.1 yg/1.
66
-------
100
CTi
0>
^
o>
E
c
o
o
c
o
o
o
o
10
O D I — Dist. , pH 5.5
D D I - Dist. , pH 8.5
V 100 mg/^ Humic Acid
A 10 mg /^ Humic Acid
• Well Water
VA.
* D I = Oeionized
** Dist = Distilled
0.1
O.I
1.0
10
IOO
Equilibrium Concentration,
Figure 27. Adsorption of MIB.
-------
concentrations. One hundred mg/1 is unrealistic for most natural waters
(typical values are on the order of 1 to 10 mg/1) but the data are useful
for emphasizing the magnitude of the competitive effect. As the equilibrium
concentration of MIB increases, the isotherms for MIB in the presence of
humic acid converge to the isotherm for which there is no competition.
However, the isotherm for MIB in well water does not converge in the con-
centration range studied, probably owing to the nature of the organic matter
in the well water vs. the commercial humic acid used.
Batch adsorption tests using F-200 carbon were conducted with geosmin
in a manner similar to that used for MIB and the results are shown in
Figure 28. The data show that at an equilibrium geosmin concentration of
0.1 yg/1 the capacity of the carbon is 0.54 mg/g in deionized-distilled
water, twice the capacity shown in Figure 27 for MIB at the same equilibrium
concentration. The data for adsorption of geosmin from 10 and 40 mg/1 humic
acid solution are also shown and indicate a greater degree of competition
than there was with MIB. In fact, at 1 yg/1 equilibrium concentration in
the presence of 10 mg/1 humic acid, the surface concentration of geosmin is
essentially the same as for MIB under the same conditions.
Adsorption data were also obtained for MIB and geosmin in deionized-
distilled water using a single test bottle and sequential addition of
adsorbate, with equilibration after each addition, as described in Section
4. The data obtained in both cases varied only slightly from the data which
are shown in Figures 28 and 29.
Figure 29 shows the results of several continuous flow column experi-
ments in which MIB was the adsorbate and F-200 was the carbon. An obvious
characteristic of these curves is the manner in which they level off rather
than converge to C/C0 = 1. This suggests that a portion of the capacity is
being utilized very slowly and that a long operation time would be required
for complete saturation. This effect is probably attributable to a combi-
nation of the high flow rate, short contact time and a slower rate of mass
transport in the small pores of the carbon. As predicted from the isotherms,
the presence of 10 mg/1 of humic acid greatly reduces the capacity of the
carbon for MIB. The data show very little effect due to variation in pH
from 5.5 to 8.5, also consistent with the isotherms. The divergence of the
curve for pH 5.5 with no humic acid from the other curves after 40 hours is
attributed to a decreased flow rate because of column plugging. Copper ion
which was introduced to eliminate possible biological activity in the column
appears to have no significant effect on the breakthrough curve.
Figure 29 also shows data for a column test in which geosmin was the
adsorbate. The greater capacity observed for geosmin compared to MIB is
consistent with the batch equilibrium data shown in Figures 27 and 28 which
show geosmin to be more strongly adsorbed.
The data used to determine the C/Co plot for geosmin in Figure 29 are
shown in Figure 30. As is apparent, geosmin loss from the open reservoir
was significant and amounted to more than 40 percent over the 190 hr of the
test. By comparison, MIB is much less volatile from aqueous solution with
typical losses amounting to less than 10 percent over 100 hours.
68
-------
100
o>
\
o>
E
o
V
o
o
o
o
s-
V-
1.0
0. I
O.I
Dl -Dist-
10 mg/Jt Humic Acid
40 rr\g/!( Humic Acid
l.O
10
100
Equilibrium Concentration ,
Figure 28. Adsorption of geosmin.
69
-------
MIB C0 = 50 ug /
pH 8-8.5, 2mM, HCO,
O pH 5.0, 50mg A Cu
pH 5.5
pH 5-6, IOmg/2 Humic Acid
pH5-6, lOmg/JP Humic Acid
Geosmin C0 = 50 u.g / J(
A pH 5.5
80 100 120
Time , Hours
180
Figure 29. Column breakthrough curves for MIB and geosmin.
-------
u
c
o
o
IOO
Time , Hours
150
200
Figure 30. Column breakthrough curve for geostnin,
-------
Discussion
Even though earthy-musty odors are a common problem at water treatment
plants and activated carbon is commonly used to remove these odors, it is
difficult to say whether adsorption or biological activity, or some other
mechanism, is effecting the removal. Our results show that two causative
agents of earthy-musty odor, geosmin and MIB, are strongly adsorbed even in
the presence of interfering natural organic matter.
Silvey et al. (1976) report that biological degradation of geosmin by
8a.cx££a6 ceAetM readily takes place; because of the biological growth which
develops in beds of GAC it is possible that some removal of earthy-musty
odor can be attributed to it. Nevertheless, in our study biological activity
did not contribute to geosmin or MIB removal. Most of the MIB (70-80 per-
cent) and the geosmin (80-90 percent) was recovered unchanged through simple
dioxane extraction of the carbon. No difference in percent recovery was
noted when low pH systems were used with 50 mg/1 of Cu^"1" serving as a biocide
as compared with systems near neutral pH without a biocide. Also, break-
through curves were essentially the same when Cu2+ was used as when it was
not.
It is not possible to conclude that the capacity of carbon in the
column is the same as the carbon used in the batch tests because the columns
were not run until saturation was achieved, although agreement at saturation
is expected. When the column tests were terminated, approximately 50 per-
cent of the expected saturation capacity for the column had been achieved.
Lower flow rates would have resulted in less severe tailing for the break-
through curve but longer experimental operation times would then be required
to achieve a given degree of saturation and this would severely limit the
number and variety of experiments which could be run.
The commercial humic acid used in this study significantly reduced the
amount of the two odor compounds which would adsorb, with the effect being
somewhat greater for the geosmin. Since humic acids are larger than fulvic
acids and much larger than the odor compounds, it is possible that a greater
reduction in adsorption capacity will take place in waters which have a
larger amount of fulvic acids. The fulvic acids can penetrate smaller pores
and thus may interfere with adsorption to a greater extent. However, as
shown in the following subsections dealing with chlorophenols, the source of
the fulvic acid is also likely to be important. Adsorption of MIB from well
water showed a competitive effect somewhat different than that of commercial
humic acid, and there was a strong indication that the organics in well water
were more adsorbable.
An important question to be answered when GAC beds are used to remove
trace organics concerns "unloading" of the carbon, or appearance of concen-
trations of substances in the effluents of GAC beds greater than are in the
influent. This phenomenon is commonly observed for organic matter parameters
such as COD and CCE while the GAC beds continue to remove odor (Love et al.,
1973). To determine whether MIB would appear in the effluent if the humic
acid concentration in the influent were suddenly increased, a column test
was conducted, the results of which are shown in Figure 31. The bed
72
-------
CO
1.0
e? 0.9
O
c
o
0.8
O 0.7
"c
0>
£ 0.6
o
u
— °-5
c
;=: 0.4
C
0)
^ 0.2
0. I
D I Water With
MIB , pH 7-8
/~'0 mg/^ Humic Acid
Added To MIB Solution
Begin Dl Water Only
100
200 300 400 500
Time , Hours
600
700
Figure 31. Column breakthrough curve for MIB.
-------
contained 5 grams of F-200 carbon in a 1.27 cm diameter column. Bed depth
was 7.6 cm and the flow rate was 2 liters/hr. As C/C0 approached 0.4,
10 mg/1 of humic acid was added to the influent MIB solution. The humic acid
rapidly saturated the bed, with effluent concentration essentially equal to
influent concentration after 70 hours. The effluent concentration of MIB
increased more rapidly, as shown, but did not exceed the influent concen-
tration. Unloading is expected only when no adsorption capacity remains to
re-adsorb the MIB displaced by the humic acid.
After 430 hours all humic acid and MIB were eliminated from the influent
solution and only deionized water was applied to the column. As shown in
Figure 31, the effluent concentration dropped very rapidly. After passage
of more than 200 bed volumes of water only a very small fraction of the
total adsorbed MIB was released. The column run was terminated because of
mechanical difficulties but the results indicate that MIB is not easily
eluted from the carbon after adsorption.
A series of tests using MIB was conducted in a manner similar to the
procedure normally used for determining isotherms except that they were not
carried to equilibrium. All test solutions were agitated on a gyratory
shaker for approximately 4 days. The results are presented in Figure 32,
with the equilibrium curve for MIB adsorption from deionized-distilled water
from Figure 27 reproduced for comparison. The data for these tests had much
more scatter than did the equilibrium data and thus only the line of "best
fit" is shown. A comparison of the curves for pH 5.5 in deionized-distilled
water shows nearly 10 times as much MIB is adsorbed at equilibrium as at
non-equilibrium. It is interesting to note that changing the pH from 5.5
to 8.5 again had little effect on the amount adsorbed, and that humic acid
had a more significant effect than was observed at equilibrium conditions
(see Figure 27). A major difference with the interference effect was that
the curves with and without humic acid did not converge at high MIB concen-
trations as in the equilibrium case. Inclusion of 50 mg/1 Cu*+ in the test
solution as a biocide apparently increased the rate of adsorption, although
there appears to be no good explanation for this phenomenon. As was pre-
viously noted, this effect was not observed in the column tests.
The importance of the non-equilibrium data is that when carbon is used
under conditions where equilibrium is not achieved, such as is likely the
case in many instances when PAC is used, the humic substances still have a
major effect on adsorption capacity for MIB.
The results of our study show MIB and geosmin to be much more strongly
adsorbed than natural organic matter. To illustrate this, if it is assumed
that 10 yg/1 of MIB and 10 mg/1 of humic acid are present in water entering
a 2 foot deep carbon bed at a flow rate of 2 gpm/ft^, it would take on the
order of 200 months to completely saturate the bed with MIB while saturation
with the humic acid would take place in about 1 month. This calculation
assumes that MIB and humic acid are present on a continuous basis and that
all of the substance entering the bed is adsorbed until the bed is saturated.
Of course, in an actual situation some leakage would take place before the
bed was completely saturated; also, some trace compounds may be present
74
-------
Dl - Dist. pH 5.5-
At Equilibrium
o>
- i.o
o
c
0>
o
c
o
o
-------
which also compete with MIB for adsorption sites. Factors which would tend
to make the bed life for MIB even longer are that earthy-musty odors gener-
ally occur periodically through the year and that biological growths may
develop in the bed which would destroy a fraction of the MIB. Larger scale
studies are needed to verify these predictions. All indications are that if
removal of MIB or geosmin is the sole objective for which GAC beds are used,
the bed life will be very long. Currently, data on MIB or geosmin distri-
bution in full scale carbon beds are needed to verify that penetration of
these compounds through the bed is extremely slow.
ADSORPTION OF CHLOROPHENOL
As phenol becomes substituted with chlorine, its solubility decreases
and it is adsorbed more strongly onto carbon as was found by Gauntlett and
Packham (1973). Their results also indicated that the dissociated molecules
found at pH values above a weak acid's pKa value are less strongly adsorbed
than the undissociated form. Ward and Getzen (1970) observed the same
effect of pH in a study performed on the adsorption of aromatic acids in
aqueous solution. They found that maximum adsorption occurred near the
point where pH = pKa for each compound. Their data suggested that greater
quantities of undesirable solutes may be removed from waters by the addition
of carbon at a pH level slightly below this optimum point.
In the research reported herein, adsorption of chlorophenols is examined
down to the threshold odor level in systems containing single solute,
bisolute chlorophenol, and single solute chlorophenol in the presence of
various humic substances. Previous research has not examined these aspects
of chlorophenol adsorption.
Single Solute Systems
Equilibrium data for single solute isotherms of 2,4-DCP and 2,4,6-TCP
in distilled water at pH 5.2, 7.0 and 9.1 are shown in Figures 33 and 34.
At pH 5.2 and 7.0, 2,4-DCP is primarily undissociated (pKa = 7.85) while
at pH 9.1 it is primarily anionic. At pH 5.2, 2,4,6-TCP is primarily
undissociated (pKa = 6.0) while at pH 7.0 and 9.1 it is primarily anionic.
The effect of pH on adsorption capacity for an equilibrium concentration of
2 x 10~8 M chlorophenol species is shown on Figure 35. The results confirm
findings of Zogorski and Faust (1974) for 2,4-DCP in the millimolar concen-
tration range. Adsorption of 2,4-DCP and 2,4,6-TCP rises to a maximum at
pH values near the compound's pKa and falls off rapidly at pH values above
the pKa. The peak in adsorption near pH - pKa where the species is 50
percent ionized results because of synergistic adsorption of the anionic
and neutral species (Ward and Getzen, 1970). Jain and Snoeyink (1973) pro-
pose the alternative explanation that the anionic and neutral species adsorb
at two different kinds of sites thereby increasing the total amount of
adsorption when both species are present above that when one species pre-
dominates .
The results also confirm findings of Gauntlett and Packham (1973) in
that the more highly substituted the molecule is with chlorine, the better
76
-------
in
at
o
E
X
c
o
"o
-------
CO
.d2
o
^ ,6'
X
o
c
0>
o
o
o
o> id4
o
3
CO
165
.
,68
T = 28
0.01 M P04
F-400
pH 5.2
10
,66
Equilibrium Concentration, C (moles
pH 9.1
,65
Figure 34. Adsorption isotherms for 2,4,6-trichlorophenol
-------
'b
X
5
£.
o
J
X
c
o
o
h_
c
«
o
c
o
p
o
o
•*-
J)
\lL
II
10
9
8
7
6
5
4
3
2
1
A
•. 'v '
pj\
~ VTCP
\
\
\
\
t
_ /" ^ \ i N^"DCP
a \ ° N
\ V
— _ A
T =28 °C \
— 0.01 M P04 \
F-400 8 ^
Ceq,DCP = 2 xlO' M \
— Ceq , TCP = 2 x 10 M N
III 1
45678 9 1C
PH
Figure 35. Influence of pH on chlorophenol adsorption capacity.
79
-------
is the adsorption of the undissociated molecule. At pH 5.2, 2,4,6-TCP
adsorbs three times as strongly as 2,4-DCP at an equilibrium concentration
of 2 x 10"8 M. It should be noted that at each pH value studied, 10'2 M
phosphate buffer was employed to ensure that the solution pH would remain
constant throughout the time required to reach equilibrium. Zogorski and
Faust (1976) reported that the adsorption of undissociated 2,4-DCP is
unaffected by the presence of 5 x 10~2 M phosphates while the equilibrium
capacities of dissociated 2,4-DCP were enhanced by about 10 to 20 percent.
Due to the sensitivity of adsorption to pH fluctuations in the region near
the pKa of the chlorophenols, it was felt that use of the buffer was more
important than its potential for causing a shifting adsorption capacity.
An attempt was made to fit the experimental single solute data with the
Langmuir adsorption equation shown below. The model assumes that a mono-
layer of solute molecules is adsorbed on the surface, the energy of adsorp-
tion is constant, and that no interaction occurs between adsorbed molecules
(Langmuir, 1918).
X =
XmbC
eg.
1 + b C
(1)
eq
where:
X = amount of solute adsorbed per unit weight of adsorbent
X = surface coverage corresponding to a monolayer of adsorbate
molecules on the adsorbent surface
b = constant related to energy of adsorption, where 1/b is the
adsorbate concentration at which adsorption attains one-
half of the monolayer coverage
C = equilibrium solute concentration
Linearization of Equation 1 results in the following form:
1
"m b Xm Ceq
1
X
1
X.
(2)
Fit of the data to the Langmuir model is indicated if the plot of 1/X versus
l/Ceq is linear. It was found that the data presented in Figures 33 and 34
did not adhere to the model over the concentration range studied, as was
observed in several other studies (Zogorski and Faust, 1974; Snoeyink et al.,
1969). Instead, a computer program was used to fit the data to a polynomial
by Gaussian elimination and least-squares analysis. The data fit a second
order polynomial of the form:
In (X) = J + K In (Ceq) + L (In Ceq)'
(3)
80
-------
The constants obtained for each isotherm are listed in Table 11
TABLE 11. VALUES OF CONSTANTS FOR EQUATION 3 FOR
CHLOROPHENOL SINGLE SOLUTE DATA
Compound Solution pH
2,4-DCP
2,4,6-TCP
5.2
7.0
9.1
5.2
7.0
9.1
-8.4904
-6.9126
-14.056
-11.646
-13.201
-0.19919
-0.48105
-0.22529
-1.2068
-0.96124
-1.1995
+0.66134
-0.02470
-0.01375
-0.048154
-0.038926
-0.049262
+0.0086726
Bisolute Systems
Many studies of carbon adsorption from single solute systems have been
conducted but natural waters contain a mixture of organic substances. As
the study of the kinetics of phenol chlorination by Lee (1967) showed, there
is commonly a number of chlorophenols present at a given time, for example.
The presence of other compounds results in the occurrence of competition for
adsorption sites which can markedly affect the adsorption of a particular
substance. Mutual inhibition of adsorptive capacity occurs if the adsorp-
tion affinities of the solutes do not differ by more than a few orders of
magnitude, and if there is no specific interaction between solutes which
enhances adsorption (Weber and Morris, 1964). Some information can be
obtained from simplified systems, such as bisolute systems, and other infor-
mation can be obtained from studies using natural waters and simulated
natural waters.
The most common model used to predict equilibrium concentrations in a
multi-solute system is the Langmuir model for competitive adsorption, first
developed by Butler and Ockrent (1930). The model permits calculation of
the amount of species "1" adsorbed per unit weight of adsorbent at equi-
librium concentration Ceqsi in the presence of other species.
Xl =
Xm.l b! Ceq,1
l
(4)
81
-------
The constants Xm and b in Equation 4 are those obtained from single solute
systems. The Langmuir competitive model is not expected to apply to systems
where adsorption of either component of a bisolute system occurs on sites
that are either inaccessible or unavailable to one of the species, e.g.,
where a portion of the adsorption occurs without competition. If the
Langmuir model for competitive adsorption does predict adsorption capacities
from a bisolute system when Xmj f %m,2> i* 1S probably because the species
compete for all available sites. The'difference in X^ values in this
instance would most likely be caused by a difference in surface area covered
by one adsorbate as compared with the competing adsorbate (Jain and Snoeyink,
1973)
Jain and Snoeyink (1973) studied competitive adsorption of bisolute
mixtures of organic anions arid neutral species. Typical results showed that
the presence of 5 x 10"3 M p-bromophenol reduced the capacity for p-nitro-
phenol by nearly 2 orders of magnitude at p-nitrophenol equilibrium concen-
tration of 5 x 10~5 M. It was found that the Langmuir competitive model,
which assumes competition among species for various sites could accurately
describe the adsorption from a solution of these compounds in the neutral
form. The data for other mixtures were varied, however. Some adsorbed
with very little competition and in other cases, electrostatic repulsive
forces were of importance. These latter cases are more difficult to des-
cribe with a model, although a modification of the Langmuir equation to
account for some adsorption without competition was moderately successful.
Their model is based on the assumption that adsorption occurs without com-
petition when Xm,i > Xmj2 and that the number of sites for which there is
no competition is equal to the quantity (Xmj - Xp^)- The equation for
the amount of species "1" adsorbed takes the form b'elow while the equation
for the amount of species "2" adsorbed is the same as that in the Langmuir
model.
. - (Xm,1 ' Xm,2) b1 Ceq,1 . "m.2 b1 Ceq.l ,,>
1 " ! * bl <=e,.l 1 + b, Ceq>1 + b2 Ceq_2 I5'
Tables 12, 13 and 14 show equilibrium data obtained for competitive
adsorption studies using 2,4-DCP and 2,4,6-TCP at pH 5.2, 7.0, and 9.1.
Three different isotherms were set up at each pH to attempt to show the
magnitude of competition between each species by increasing the equilibrium
concentration of one species relative to the other. To indicate the degree
of competition, the amount which would be adsorbed per gram if the species
were present in solution alone, Xss> is tabulated. Also tabulated is the
extent of adsorption as predicted by the Langmuir and the Jain and Snoeyink
competitive models, Equations 4 and 5, respectively. A computer program
was developed to evaluate the constants for the models using the polynomials
which were fitted to the single solute data. The program calculates Xm and
b for each individual concentration and then uses these values in Equations
4 and 5 to calculate the amount of species adsorbed in the presence of the
other chlorophenol.
82
-------
TABLE 12. CHLOROPHENOL COMPETITION AT pH 5.2
Ceq,DCP = °'35 Ceq,TCP
Co,DCP= 2.70 mg/1
Co,TCP = 20'80 mg/1
2,4-Dichlorophenol
Bottle
1
2
3
4
5
6
Bottle
1
2
3
4
5
6
Ceq,DCP
mole/1
5.52xlO"9
Z.OlxlO"8
3.99xlO"8
l.SlxlO"7
6.44xlO"7
3.37xlO"6
Ceq,TCP
mole/1
2.28xlO"8
7.09xlO"8
1.47xlO"7
3.92xlO"7
1.54xlO"6
7.34xlO"6
Ceq,TCP
mole/1
2.28xlO"8
7.09xlO"8
1.47xlO"7
3.92xlO"7
1.54xlO"6
7.34xlO"6
2,4
Ceq,DCP
mole/1
5.52xlO"9
2.01xlO"8
3.99xlO"8
1.61xlO"7
6.44xlO"7
3.37xlO"6
Xobs
mole/g
1.62xlO"4
2.13xlO"4
2.71xlO"4
3.31xlO"4
3.77xlO"4
3.77xlO"4
Xss
mole/g
2.47xlO~4
4.61xlO"4
6. 09x1 O"4
9.20xlO"4
1.25xlO"3
1.66xlO~3
V
Langmuir
mole/g
l.SlxlO"4
2.38xlO"4
2.82xlO"4
3.94xlO~4
3.60xlO"4
3.99xlO"4
XJain
mole/g
l.SlxlO"4
2.38xlO"4
2.82xlO"4
3.94xlO"4
3.60xlO"4
3.99xlO"4
,6-Trichlorophenol
Xobs
mole/5
l.OSxlO"3
1.36xlO"3
1.72xlO"3
2.12xlO"3
2.56xlO"3
2.80xlO"3
Xss
mole/g
1.22xlO"3
1.87xlO"3
2.20xlO"3
2.50xlO"3
2.91xlO"3
3.42xlO"3
V
Langmuir
mole/g
7.64xlO"4
1.15xlO"3
1.45xlO"3
1.82xlO"3
2.48xlO"3
2.63xlO"3
Jain
mole/g
1.04xlO"3
1.55xlO"3
1.91xlO"3
2.32xlO"3
2.84xlO"3
2.89xlO"3
X . = observed X
XLangmuir = X as
from Equation 4
Xcc = single solute X Xla_- = X as calculated from Equation 5
ss jain
(continued)
83
-------
TABLE 12 (continued)
co
Co
q,DCP 1-1 Ceq,TCP
,DCP = 3'68 mg/1
jTCp =8.11 mg/1
2,4-Dichlorophenol
Bottle
1
2
3
4
Bottle
1
2
3
4
Ceq,DCP
mole/1
4. 60x1 O"9
1.47xlO"8
3. 80x1 O"8
1.87xlO"7
Ceq,TCP
mole/1
8.78xlO"9
1.57xlO"8
3.14xlO"8
1.16xlO"7
Ceq,TCP
mole/1
8.78xlO"9
1.57xlO"8
3.14xlO"8
1.16xlO"7
2,4
Ceq,DCP
mole/1
4.60xlO"9
1.47xlO"8
3. 80x1 O"8
1.87xlO"7
Xobs
mole/g
2.14xlO"4
3.39xlO"4
4.46xlO"6
6.31xlO"4
Xss
mole/g
2.28xlO"4
3.95xlO"4
e.ooxio"4
9.61xlO"4
V
Langmuir
mole/g
1.54xlO"4
2.52xlO"4
3.51xlO"4
5.41xlO"4
XJain
mole/g
1.54xlO"4
2.52xlO"4
3.51xlO"4
5.41xlO"4
,6-Trichlorophenol
Xobs
mole/g
3.89xlO"4
6.17xlO"4
8.13xlO"4
1.16xlO"3
Xss
mole/g
7.40xlO"4
l.OlxlO"3
1.41xlO"3
2.10xlO"3
V
Langmuir
mole/g
4.77xlO"4
5.94xlO"4
7.69xlO"4
1.1 6x1 0~3
XJain
mole/g
6. 58x1 O"4
8.24xlO"4
1.07xlO"3
1.60xlO"3
Note: See Table 12A for definition of terms.
(continued)
84
-------
TABLE 12 (continued)
c. c
c
c
f\j
eq,DCP ~
o,DCP = 8<
o,TCP = 6<
'9 Ceq,TCP
,47 mg/1
17 mg/1
2,4-Dichlorophenol
Bottle
1
2
3
4
5
6
Bottle
1
2
3
4
5
6
Ceq,DCP
mole/1
2.61xlO"8
9. 20x1 O"8
1.75xlO"7
4.75xlO"7
1.47xlO"6
5.24xlO"6
p
ueq,TCP
mole/1
3.80xlO"9
1.24xlO"8
2.03xlO"8
5.06xlO"8
1.27xlO"7
5.76xlO"7
Ceq,TCP
mole/1
3.80xlO"9
1.24xlO"8
2.03xlO"8
5.06xlO"8
1.27xlO"7
5.76xlO"7
2,4
p
Ueq,TCP
mole/1
2.61xlO"8
9.20xlO"8
1.75xlO"7
4.75xlO"7
1.47xlO"6
5.34xlO"6
Xobs
mole/g
4.98xlO~4
6.82xlO"4
8.55xlO"4
l.OOxlO"3
1.24xlO"3
1.50xlO"3
Xss
mole/g
5.21xlO"4
8.12xlO"4
9.49xlO"4
1.16xlO"3
1.49xlO"3
1.96xlO"3
Langmuir
mole/g
3. 68x1 O"4
5.45xlO"4
6.54xlO"4
8.23xlO"4
1.02xlO"3
1.12xlO"3
XJain
mole/g
3.68xlO"4
5.45xlO~4
6.54xlO"4
8.23xlO"4
1.02xlO"3
1.12xlO"3
,6-Trichlorophenol
V
Xobs
mole/q
2.99xlO"4
4.11xlO"4
5.51xlO"4
6.07xlO~4
7.65xlO"4
9.89xlO"4
y
Ass
mole/g
4.65xlO"4
8.90xlO"4
1.14xlO"3
1.66xlO"3
2.15xlO"3
2.61xlO"3
Y
Langmuir
mole/g
2.55xlO"4
4.42xlO"4
5.31xlO"4
7.27xlO"4
9.15xlO"4
l.SlxlO"3
Y
Jain
mole/g
3.14xlO"4
5.73xlO"4
6. 98x1 O"4
9.91xlO"4
1.29xlO"3
1.87xlO"3
Note: See Table 12A for definition of terms.
(continued)
85
-------
TABLE 12 (continued)
D. Data Summary
Dichlorophenol
y y Y
C_ obs obs Aobs
Bottle
<=H
mole/1
Isotherm A:
1 5
2 2
3 3
4 1
5 6
6 3
.52x10
.01x10
.99x10
.61x10
.44x10
.37x10
Isotherm B:
1 4
2 1
3 3
4 1
.60x10
.47x10
.80x10
.87x10
Ceq
-9
-8
-8
_7
/
_7
/
-6
\J
Ceq
-9
-/
-8
u
-8
-7
Isotherm C: C
1 2
2 9
3 1
4 4
5 1
6 5
.61x10
.20x10
.75x10
.75x10
.47x10
.34x10
-8
(J
-8
_7
/
-7
-6
-6
Xss
,DCP =
66%
46
44
36
30
23
,DCP =
94
86
74
66
a,
,DCP =
96
83
90
86
83
77
V
Langmui
r XJain
Ceq
mole/1
Trichlorophenol
Xobs Xobs Xobs
Xss
V
Langmuir
XJain
°'35 Ceq,TCP
107%
89
96
84
105
94
1J Ceq,
139
135
127
117
8.9 C
eq,
135
125
131
122
122
134
107%
89
96
84
105
94
TCP
139
135
127
117
TCP
135
125
131
122
122
134
2
7
1
3
1
7
8
1
3
1
3
1
2
5
1
5
.28xlO"8
.09x10
-8
.47xlO"7
.92x10
.54x10
.34x10
.78x10
.57x10
.14x10
.16x10
.80x10
.24x10
.03x10
.06x10
.27x10
.76x10
_7
/
-6
\J
-fi
o
_Q
J
-fi
O
-8
-7
-Q
3
-8
0
o
-8
-7
-7
84%
73
78
85
88
82
53
61
58
55
64
46
45
37
36
38
135%
118
119
116
103
106
82
104
106
100
117
93
97
83
84
75
99%
88
90
91
90
97
59
75
79
73
95
72
74
61
59
53
Note: See Table 12A for definition of terms.
86
-------
TABLE 13. CHLOROPHENOL COMPETITION AT pH 7.0
A" Ceq,DCP = °'14 Ceq,TCP
Co,DCP = 6'58 m9/]
Co,TCP = 10'76 m9/]
Bottle
1
2
3
4
5
6
Bottle
1
2
3
4
5
6
r
ueq,DCP
mole/1
5.37xlO"9
1.17xlO~8
2. 58x1 O"8
6. 87x1 O"8
2. 58x1 O"7
1.27xlO"6
r
Leq,TCP
mole/1
4.25xlO"8
9.62xlO"8
2.15xlO"7
5.19xlO"7
1.65xlO"6
6.58xlO"6
r
eq,TCP
mole/1
4.25xlO"8
9.62xlO"8
2.15xlO"7
5.19xlO"7
1.65xlO"6
6.58xlO"6
2
r
Leq,DCP
mole/1
5.37xlO"9
1.1 7x1 0~8
2. 58x1 O"8
6.87xlO"8
2.58xlO"7
1.27xlO"6
2 ,4-Di chl orophenol
y
xobs
mole/g
4.61xlO"4
5.81xlO"4
7.19xlO"4
8. 50x1 O"4
1.02xlO"3
1.18xlO"3
y
SS
mole/g
5.40xlO"4
6.90x.O"4
8.80xlO"4
l.llxlO"3
1.37xlO"3
1.71xlO"3
Y
Langmuir
mole/g
3.99xlO"4
4.64xlO"4
5.22xlO"4
5.77xlO"4
.5.39xlO"4
2.54xlO"4
y
AJain
mole/g
3.99xlO"4
4.64xlO"4
5.23xlO"4
5.77xlO"4
5.39xlO"4
2.54xlO"4
, 4, 6-Tri chl orophenol
V
xobs
mole/g
6.22xlO"4
7.83xlO"4
9.67xlO"4
1.14xlO"3
1.34xlO"3
1.45xlO"3
y
xss
mole/g
9.60xlO"4
1.25xlO"3
1.37xlO"3
2.15xlO"3
1.39xlO"3
1.70xlO"3
V
Langmuir
mole/q
4.10xlO"4
G.lOxlO"4
8.57xlO"4
l.lSxlO"3
1.73xlO"3
2.39xlO"3
x
AJain
mole/g
6.71xlO"4
9. 74x1 0"4
1.31xlO"3
1.67xlO"3
2.09xlO"3
2.49xlO"3
Note: See Table 12A for definition of terms.
(continued)
87
-------
TABLE 13 (continued)
B" Ceq,DCP = °'45 Ceq,TCP
Co,DCP = 8'16 mg/1
Co,TCP = 6-^mg/l
p
Leq,DCP
Bottle mole/1
1
2
3
4
5
Bottle
1
2
3
4
5
7.05xlO-9
1.23xlO~8
7.36xlO"8
3.25xlO"7
1.44xlO"6
r
ueq,TCP
mole/1
1.06xlO"8
4.35xlO"8
1.49xlO"7
5.98xl0^7
2.53xlO"6
r
Leq,TCP
mole/1
1.06xlO"8
4.35xlO"8
1.49xlO"7
5.98xlO"7
2.53xlO"6
2
r
Leq,DCP
mole/1
7.05xlO"9
1.23xlO"8
7.36xlO"8
3.25xlO"7
1.44xlO"6
2,4-Dichlorophenol
V
Xobs
mole/g
4.75xlO"4
7.15xlO"4
9. 94x1 O"4
1.21xlO"3
1.46xlO"3
,4,6-Trichl
Y
obs
mole/g
2.95xlO"4
4. 44x1 O"4
6.15xlO"4
7.40xlO"4
8.55xlO"4
Y
xss
mole/g
5.84xlO"4
7.00xlO"4
1.12xlO"3
1.41xlO"3
1.75xlO"3
orophenol
X
Ass
mole/g
3.65xlO~4
8.70xlO"4
1.50xlO"3
2.20xlO"3
3.09xlO"3
x
Langmuir
mole/g
4.84xlO~4
5.12xlO"4
7.33xlO"4
8.33xlO"4
6.19xlO"4
x
Langmuir
mole/q
1.64xlO"4
3.91xlO"4
6.61xlO"4
l.llxlO"3
1.87xlO"3
x
AJain
mole/g
4.84xlO"4
5.12xlO"4
7. 33x1 O"4
8.33xlO"4
6.19xlO~4
Y
Jain
mole/g
2.27xlO"4
6.20xlO"4
l.OOxlO"3
1.52xlO"3
2.10xlO"3
Note: See Table 12A for definition of terms.
(continued)
88
-------
Table 13 (continued)
C> Ceq,DCP 1>4 Ceq,TCP
Co,DCP= 12.58mg/l
Co,TCP = 4'98 mg/1
2,4-Dichlorophenol
Bottle
1
2
3
4
5
6
Bottle
1
2
3
4
5
6
Ceq,DCP
mole/1
2.70xlO~8
7.67xlO"8
1.53xlO"7
6.29xlO"7
2.21xlO"6
5.40xlO"6
p
Leq,TCP
mole/1
2.03xlO"8
6.08xlO"8
1.42xlO"7
3.92xlO"7
1.37xlO"6
3.14xlO~6
Ceq,TCP
mole/1
2.03xlO"8
6.08xlO"8
1.42xlO"7
3.92xlO"7
1.37xlO"6
3.14xlO"6
2
r
Sjq.DCP
mole/1
2.70xlO"8
7.67xlO"8
1.53xlO"7
6.29xlO"7
2.21xlO"6
5.40xlO~6
Xobs
mole/g
8.45xlO"4
l.OBxlO"3
1.28xlO~3
1.49xlO"3
1.74xlO"3
1.92xlO"3
Xss
mole/g
8.90xlO"4
1.15xlO~3
1.28xlO"3
1.54xlO"3
1.74xlO"3
2.20xlO"3
Langmuir
mole/g
6.78xlO"4
8.14xlO"4
8.83xlO"4
1.06xlO"3
l.OlxlO"3
6.87xlO"4
XJain
mole/g
6.78xlO"4
8.14xlO"4
8.83xlO"4
1.06xlO"3
l.OlxlO"3
6.87xlO"4
,4,6-Trichlorophenol
X
obs
mole/g
2.76xlO"4
3.43xlO"4
4.16xlO"4
4.82xlO"4
5.55xlO"4
5.90xlO"4
x
ss
mole/g
5.70xlO"4
l.OSxlO"3
1.50xlO"3
2.02xlO"3
2.39xlO"3
2.52xlO"3
x
Langmuir
mole/g
2.28xlO"4
4.05xlO"4
6.02xlO"4
8.55xlO"4
1.33xlO"3
1.91xlO"3
x
Jain
mole/g
2.97xlO"4
5.80xlO"4
8.81xlO"4
1.19xlO"3
1.64xlO"3
2.06xlO"3
Note: See Table 12A for definition of terms.
(continued)
89
-------
TABLE 13 (continued)
D. Data Summary
Dichlorophenol
C. Xobs Xobs Xobs
Bottle mole/1
Isotherm A: C
eq
1 5.37xlO"9
2 1.17xlO"8
-8
3 2.58x10 °
-8
4 6.87x10 °
5 2.58xlO"7
6 1.27xlO~6
Isotherm B: C
eq
-9
1 7.05x10
2 1.23xlO"8
3 7.36xlO"8
4 3.25xlO"7
5 1.44xlO"6
Isotherm C: C
eq
1 2.70xlO"8
2 7.67xlO"8
3 1.53xlO"7
4 6.29xlO"7
5 2.21xlO"6
6 5.40xlO"6
Xss
a/
,DCP ~
85%
84
82
77
74
69
,DCP =
81
102
89
86
83
,DCP =
95
91
100
97
100
87
V
Langmui
0-14 Ceq
116%
125
137
147
189
465
0.45 Ce
98
140
136
145
236
1-4 Ceqj
125
129
145
141
172
279
r XJain
,TCP
116%
125
137
147
189
465
,TCP
98
140
136
145
236
TCP
125
129
145
141
172
279
Ceq
mole/1
4.25xlO"8
9.62xlO"8
_7
2.15x10 '
_7
5.19x10 '
1.65xlO"6
6.58xlO"6
-8
1.06x10 °
4.35xlO"8
1.49xlO"7
5.98xlO"7
2.53xlO"6
2.03xlO"8
6.08xlO"8
1.42xlO"7
3.92xlO"7
1.37xlO"6
3.14xlO"6
Trichlorophenol
Xobs Xobs Xobs
Xss
65%
63
71
53
96
85
81
51
41
34
28
48
33
28
24
23
23
V
Langmuir
152%
128
113
97
78
61
180
114
93
67
46
121
85
69
56
42
31
XJain
93%
80
74
68
64
58
130
72
62
49
41
93
59
47
41
34
29
Note: See Table 12A for definition of terms.
90
-------
TABLE 14. CHLOROPHENOL COMPETITION AT pH 9.
A. C
eq,DCP = °'°44 Ceq,TCP
Co,DCP = 3'23 mg/1
Co,TCP = 4'64 m9/]
2,4-Dichlorophenol
Bottl
1
2
3
4
5
Bottl
1
2
3
4
5
Ceq,DCP
e mole/1
1.38xlO"8
1.84xlO"8
2.94xlO"8
3.68xlO"8
7.51xlO"8
r
Leq,TCP
e mole/1
1.62xlO"7
3.73xlO"7
6.84xlO"7
1.52xlO"6
3.85xlO"6
Ceq,TCP
mole/1
1.62xlO~7
3.73xlO"7
6.84xlO"7
1.52X10"6
3.85xlO"6
2
r
eq,DCP
mole/1
1.38xlO"8
1.84xlO"8
2.94xlO"8
3.68xlO~8
7.51xlO~8
Xobs
mole/g
1.90xlO"4
2.57xlO"4
3.30xlO"4
4.12xlO"4
5.45xlO"4
,4,6-Trichl
y
xobs
mole/g
2.24xlO"4
3-OOxlO"4
3.80xlO"4
4.58xlO~4
5.43xlO"4
Xss
mole/g
3.35xlO"4
3.95xlO"4
S.lOxlO"4
5.70xlO"4
7.53xlO"4
orophenol
y
Ass
mole/g
2.37xlO"4
3.30xlO"4
4.25xlO"4
5.80xlO"4
8.70xlO"4
V
Langmuir
mole/g
1.85xlO"4
2.26xlO"4
3.00xlO"4
3.46xlO"4
S.OlxlO"4
Y
Langmuir
mole/g
1.65xlO"4
2.21xlO"4
2.66xlO"4
3.57xlO"4
4.74xlO"4
XJain
mole/g
2.63xlO"4
2.86xlO"4
3.53xlO"4
3.49xlO"4
S.OlxlO"4
x
Jain
mole/g
1.65xlO"4
2.21xlO"4
2.66xlO"4
3.57xlO"4
5.50xlO"4
Note: See Table 12A for definition of terms.
(continued)
91
-------
TABLE 14 (continued)
R r - n ?i r
dl Leq,DCP u-lil ueq,TCP
Co,DCP = 7'67 mg/1
Co,TCP = 2'89 mg/1
2,4-Dichlorophenol
Bottle
1
2
3
4
5
Bottle
1
2
3
4
5
Ceq,DCP
mole/1
2.61xlO"8
5.67xlO"8
1.29xlO"7
3.60xlO"7
1.23xlO"6
r
eq,TCP
mole/1
7.50xlO"8
2.58xlO"7
l.OOxlO"6
2.63xlO"6
5.19xlO"6
Ceq,TCP
mole/1
7. 50x1 O"8
2.58xlO"7
l.OOxlO"6
2.63xlO"6
5.19xlO"6
2
r
beq,DCP
mole/1
2.61xlO"8
5.67xlO"8
1.29xlO"7
3.60xlO"7
J.23xlO"6
Xobs
mole/g
3.60xlO"4
5.35xlO"4
7.81xlO"4
9.79xlO"4
1.24xlO"3
,4,6-Trichl
Y
obs
mole/g
l.llxlO"4
1.64xlO"4
2.27xlO"4
2.52xlO"4
2.56xlO"4
Xss
mole/g
4.75xlO"4
6.90xlO"4
8.80xlO"4
l.lOxlO"3
1.32xlO"3
orophenol
Y
Ass
mole/g
1.75xlO~4
2.87xlO"4
4.90xlO"4
7.30xlO"4
9.60xlO"4
V
Langmuir
mole/g
2.62xlO"4
4.10xlO"4
6.15xlO"4
9.11xlO"4
1.27xlO"3
V
Langmuir
mole/g
1.16xlO"4
1.66xlO"4
2. 44x1 O"4
2.76xlO"4
1.93xlO"4
XJain
mole/g
4.08xlO"4
5.58xlO"4
7.19xlO"4
9.55xlO"4
1.27xlO"3
V
Jain
mole/g
1.16xlO"4
1.66xlO"4
2.44xlO"4
2.76xlO"4
2.49xlO"4
Note: See Table 12A for definition of terms.
(continued)
92
-------
TABLE 14 (continued)
C' Ceq,DCP=°-45Ceq,TCP
Co,DCP = 10'68 ^
Co,TCP= K 56m9/1
Bottle
1
2
3
4
5
6
r
Leq,DCP
mole/1
3.76xlO"8
7.21xlO"8
1.44xlO"7
3.22xlO"7
6.60xlO"7
1.35xlO"6
Ceq,TCP
Bottle mole/1
1
2
3
4
5
6
8.36xlO"8
1.61xlO"7
3.70xlO"7
8.55xlO"7
1.60xlO"6
2.23xlO"6
r
beq,TCP
mole/1
8.36xlO"8
1.61xlO"7
3.70xlO"7
8.55xlO"7
1.60xlO"6
2.23xlO"6
2
Ceq,DCP
mole/1
3. 76x1 O"8
7.21xlO"8
1.44xlO"7
3.22xlO~7
6.60xlO"7
1.35xlO"6
2,4 Dichlorophenol
y
xobs
mole/g
5.89xlO"4
7.18xlO"4
8.85xlO"4
l.OSxlO"3
1.19xlO"3
1.33xlO"3
,4,6-Trichl
Xobs
mole/g
7.04xlO"5
8.49xlO"5
1.02xlO"4
1.17xlO"4
1.16xlO"4
l.lSxlO"4
x
xss
mole/g
5.90xlO"4
7. 50x1 O"4
9.00xlO"4
l.OSxlO"3
1.22xlO"3
1.33xlO"3
orophenol
Xss
mole/g
1.82xlO"4
2.37xlO~4
3.31xlO"4
4.65xlO"4
G.OOxlO"4
6.87xlO"4
V
Langmuir
mole/g
3.22xlO"4
4.55xlO"4
6.29xlO"4
8.61xlO"4
l.OSxlO"3
1.28xlO"3
V
Langmuir
mole/g
1.15xlO"4
1.34xlO"4
1.61xlO"4
1.82xlO"4
1.76xlO"4
1.26xlO"4
y
Jain
mole/g
4.89xlO"4
6.39xlO"4
8.05xlO"4
l.OOxlO"3
1.17xlO"3
1.33xlO"3
XJain
mole/g
l.lSxlO"4
1.34xlO"4
1.61xlO"4
1.82xlO"4
1.76xlO"4
1.26xlO"4
Note: See Table 12A for definition of terms.
(continued)
93
-------
TABLE 14 (continued)
D.
Data Summary
Ceq
Bottle mole/1
Isotherm A: C
eq
1
2
3
4
5
1.38xlO"8
-8
1.84x10 °
-8
2.94x10 °
3.68xlO"8
7.51xlO"8
Isotherm B: C
eq
1
2
3
4
5
2.61xlO"8
-8
5.67x10 °
_7
1.29x10 '
_7
3.60x10 '
-fi
1.23x10 D
Isotherm C: C
1
2
3
4
5
6
3.76xlO"8
-8
7.21x10 S
1.44xlO"7
3.22xlO"7
6.60xlO"7
1.35xlO"6
Dichlorophenol
Xobs Xobs Xobs
Xss
,DCP =
57%
65
65
72
72
,DCP =
76
78
89
89
94
,DCP =
100
96
98
100
98
100
X X
Langmuir Jain
0.044 Ceq
103%
114
no
119
109
,0.21 C „
eq,
137
131
127
108
98
°'45 Ceq
183
158
141
125
110
104
,TCP
72%
90
93
118
109
TCP
88
96
109
103
98
TCP
120
112
110
108
102
100
Ceq
mole/1
1.62xlO"7
_7
3.73x10 '
_7
6.84x10 '
1.52xlO"6
3.85xlO"6
7.50xlO"8
_7
2.58x10 '
_c
1.00x10 D
C
2.63x10"°
-fi
5.19x10 D
8.36xlO"8
_7
1.61x10 '
3.70xlO"7
8.55xlO"7
1.60xlO"6
2. 23x1 O"6
Tri
Xobs
Xss
95%
91
89
79
62
63
57
46
35
27
39
36
31
25
19
17
chlorophenol
Xobs Xobs
V
Langmuir
136%
136
143
128
115
96
99
93
91
132
61
63
63
64
66
94
XJain
136%
136
143
128
99
96
99
93
91
103
61
63
63
64
66
94
Note: See Table 12A for definition of terms.
94
-------
Competition at pH 5.2 --
At pH 5.2 both 2,4-DCP and 2,4,6-TCP are primarily undissociated.
Table 12 shows data obtained in competitive studies. As the equilibrium
concentration of 2,4,6-TCP rises relative to the equilibrium concentration
of 2,4-DCP the extent of adsorption of 2,4-DCP decreases relative to the
single solute value, as summarized in Table 12D. The same phenomenon was
observed for 2,4,6-TCP in relation to its single solute surface concentra-
tion values which indicates that competition for the same sites on the
carbon is occurring. For both species, as the equilibrium concentrations
increase with the ratio of their concentrations remaining constant, the
extent of adsorption becomes increasingly lower than the single solute value
suggesting that at the higher surface coverage the competition is more
intense.
Table 12D shows the fit of the Langmuir and Jain and Snoeyink competi-
tive models. The Langmuir model does a fairly good job of fitting both data
sets. Since Xmjycp is greater than XmjDCP over the entire concentration
range, the Jain model was applicable to the 2,4,6-TCP data. The Jain and
Snoeyink model provided a better fit of the data where Ceq,TCP > ceq,DCP
but did not improve the fit where Ceg}TCP — ^eq,DCP- As the amount of
2,4,6-TCP increased the amount of anionic species increases since the pH is
fairly close to the pKa. The better fit of the Jain model where there is a
significant portion of anionic TCP may indicate that there is some adsorption
occurring without competition because of differences in adsorption sites for
anionic versus neutral 2,4,6-TCP. Figures 36 and 37 show Langmuir model
predictions for 2,4-DCP and 2,4,6-TCP as simulated using the computer and
the single solute data previously presented in Figures 33 and 34.
Competition at pH 7.0 --
At pH 7.0, 2,4-DCP is primarily neutral while 2,4,6-TCP is primarily
anionic. Table 13 shows data obtained in three studies at this pH value.
As at pH 5.2, competition occurs between the species as evidenced by the
increasing reduction of X0bs relative to XSs as the equilibrium concentra-
tion of the competing species increases. As at pH 5.2, as the individual
equilibrium concentrations increase with their ratio remaining approximately
equal, the extent of adsorption is lowered significantly. The Langmuir model
fits the 2,4-DCP data fairly well at low Ceq but at high Ceq values the fit
is poor. The same is true for 2,4,6-TCP. At all of the concentrations
studied, Xm,TCP is greater than Xm,DCP so the Jain and Snoeyink model
applied. The Jain and Snoeyink model fit the data well for low 2,4,6-TCP
concentrations but did not provide an improved fit at high concentrations.
The Langmuir model predicted much less competition than was observed for
2,4-DCP; in some cases where Ceq,DCP > 10"6 M the observed extent of adsorp-
tion was as much as 200 percent of that predicted by the model as shown
in Table 13D. At high equilibrium concentrations of 2,4,6-TCP the observed
extent of adsorption was less than predicted while at low equilibrium con-
centrations the opposite was true. For isotherm A where Cen}DCP = °-^4
ceq,TCP> at low equilibrium concentrations, the Jain model aid a good job
of predicting extent of adsorption. Apparently the anionic 2,4,6-TCP and
neutral 2,4-DCP which predominated here did not compete entirely for the
95
-------
cn
,62
0>
en
«> 10-
o
E
c
o
O)
O -4
§ I0
o
ft)
o
o
,69
Single Solute
,68
2,4,6-TCP = 10° M-
-7
2,4,6-TCP=IO M
-6
2,4,6- TCP = 10" M
pH 5.2
F-400
0.01 M PO,
,a7
,66
Equilibrium Concentration, Ce (moles
,d5
Figure 36. Competitive adsorption capacities predicted by the Langmuir
model for dichlorophenol at pH 5.2.
-------
,62
0>
O
E
c
o
c
Q>
O
C
o
O
.o4
Single Solute
r2,4- DCP = I0"8 M
-7
2,4- DCP = 10 M
2,4- DCP = I06 M
u
o
pH 5.2
F —400
0.01 M PO,,
,b5
,o9
,68
-7
10
-6
10
Equilibrium Concentration , Ceq (moles/J[)
,o5
Figure 37. Competitive adsorption capacities predicted by the Langmuir
model for trichlorophenol at pH 5.2.
-------
same sites. However at higher equilibrium concentrations the model failed
since 2,4-DCP extent of adsorption was much greater than predicted while
2,4,6-TCP adsorbed lower than predicted. These same trends are evident in
isotherms B and C which represent increasing concentrations of 2,4,6-TCP
relative to 2,4-DCP. Since the pH is near the pKa of both compounds the
system contains four types of species, anionic 2,4-DCP and 2,4,6-TCP and
neutral 2,4-DCP and 2,4,6-TCP which complicates the search for a reason for
this behavior. In studies performed by Jain and Snoeyink (1973) on
neutral p-nitrophenol and anionic benzenesulfonate, at pH 3.8, their model fit
well for data collected in the concentration range 10~5 to 10~2 M which
suggested that some adsorption occurred without competition. Neither species
competed with the other to a great extent. However with neutral 2,4-DCP and
anionic 2,4,6-TCP, 2,4-DCP competes with 2,4,6-TCP while 2,4,6-TCP affects
2,4-DCP only slightly. A possible explanation is that the anionic 2,4-DCP
which is present at pH 7.0 effectively competes with the anionic 2,4,6-TCP
while the anionic 2,4,6-TCP present does not compete with the neutral 2,4-DCP
for the same sites. The small degree of competition of 2,4,6-TCP with
2,4-DCP occurs between the neutral species of which 2,4-DCP is present to a
greater extent. The effective competition of anionic 2,4-DCP with 2,4,6-TCP
occurs, as explained by Ward and Getzen (1969), since the anionic 2,4-DCP
adsorption is favored at pH values just below a compound's pKa- An alterna-
tive explanation may be that the chlorophenol species are changing on the
carbon surface. Computer predictions for pH 7.0 are shown in Figures 38 and
39.
Competition at pH 9.1 --
At pH 9.1 both species are primarily anionic. Table 14 indicates that
as the concentration of 2,4-DCP approaches that of 2,4,6-TCP the extent of
adsorption of 2,4,6-TCP decreases. When Ceq,DCP - 0-45 Ceq5jCP tne observed
value of the extent of adsorption of 2,4-DCP is equal to the values which
would be observed if 2,4,6-TCP were not present while those of 2,4,6-TCP are
approximately 30 percent of those if no 2,4-DCP were present. Since the
adsorption capacities for 2,4-DCP at this pH are nearly four times greater
than those for 2,4,6-TCP, an explanation for which is not apparent, the Xm
value also was greater indicating that some adsorption of 2,4-DCP could occur
without competition. This was indicated by the fit of the Jain and Snoeyink
model to the 2,4-DCP data for all three isotherms as illustrated in Table 14D.
The model fit the 2,4,6-TCP fairly well except when Ceq,DCP became close to
that of Ceq,TCP- ^n this instance the adsorption of Ceq,TCP was somewhat
less than predicted indicating that the stronger adsorbing species predomi-
nated, and that electrostatic repulsion on the carbon surface may have come
into play causing less 2,4,6-TCP to be adsorbed. When Ceq,TCP - 23 Ceq,DCP
the predictions for 2,4,6-TCP were slightly lower than observed which may
also indicate that less repulsion on the surface was occurring since less of
the stronger adsorbing species was present. Figures 40 and 41 show adsorp-
tion capacities for pH 9.1 as predicted by the Jain and Snoeyink model which
was the model which provided the best fit for the system. Apparently for
this system the species adsorbed on the same sites but due to the widely
differing capacities of the carbon for the two species, some adsorption
occurred without competition.
98
-------
0>
^
v>
_0>
1
c
o
.0'
o
o
0)
o
o
io4
10"
Single Solute
.o8
2,4,6- TCP = IO8 M
-7
2,4,6- TCP = 10 M
2,4,6- TCP = 10" M
pH 7.0
F-400
0.01 M PO,
10
,66
,65
Equilibrium
Concentration , Ce (moles/X)
Figure 38. Competitive adsorption capacities predicted by the Langmuir
model for dichlorophenol at pH 7.0.
-------
.62
o
o
o>
x.
in
9>
I
c
0)
o
c
o
o
0>
o
,63
Single Solute
-7
2,4- DCP =10 M
2,4-DCP = 10 M
2,4- DCP =I06 M
pH 7.0
F-400
0.01 M PO.
.65
-9
10
-8
10
Equilibrium
-7
10
-6
IO
Concentration, C (moles/5()
-5
10
Figure 39.
Competitive adsorption capacities predicted by the Langmuir
model for trichlorophenol at pH 7.0.
-------
o
io2
0>
M
1 ,6=
o
5
0)
u
o
-9
10
Single Solute
2,4,6- TCP = 10
-7
2,4,6-TCP = IO M
-2,4,6-TCP = 10 M
pH 9.1
F-400
0.01 M PO,
-8
10
-7
10
-6
10
10
Equilibrium Concentration, Ceq (moles/50
Figure 40. Competitive adsorption capacities predicted by the Jain
model for dichlorophenol at pH 9.1.
-------
.62
pH 9. I
F-400
0.01 M PO,
in
~o
J
X
c
o
"o
c
0)
u
c
o
u
u
o
io3
io4
2,4-DCP = IO8 M
2,4-DCP = IO'6 M
-2,4- DCP = 10 M
l65
10"
.68
.67
Equilibrium Concentration, Ceq (moles/JO
,65
Figure 41. Competitive adsorption capacities predicted by the Jain
model for trichlorophenol at pH 9.1.
-------
According to Weber and Morris (1964), when the concentration of solute
is small the term (E b^ Cn-) in the denominator of the Langmuir equation,
Equation 4, is much less than unity so direct proportionality between solute
equilibrium concentration and the amount adsorbed, X-j, is anticipated. No
competitive effects would then be observed. However, this was not the case
at pH 5.2 or at 7.0 or 9.1. In all instances for both compounds the term
(bi C-| + b2 C2) was approximately equal to unity; as the value of Ci
increased, the corresponding value of bi decreased thus causing the entire
denominator to remain greater than unity. In the Langmuir equation, the
parameter b is proportional to exp(-AH/RT) where AH is the adsorption energy.
Adsorption energy varies with surface coverage; high-energy sites are
occupied first with subsequent adsorption occurring at increasingly lower
energy sites as the surface coverage increases. The net result is signifi-
cant competition between adsorbates even at low concentration.
Chlorophenols in the Presence of Humic Substances
Data showing the adsorption of 2,4,6-TCP from solutions containing
humic substances at pH 5.2 and 9.1 are shown in Tables 15 and 16, respec-
tively. Three types of humic substances were used: commercial humic acid,
leaf fulvic acid, and soil fulvic acid, the characteristics of which are
described in Section 4 and previously in Section 5. Tests at each pH were
performed with two initial concentrations of humic material, 10 and 50 mg/1
as TOC, and also with two F-400 carbon doses to obtain different 2,4,6-TCP
equilibrium concentrations. Since initial tests showed that 2,4-DCP behaved
in a manner quite similar to 2,4,6-TCP in the presence of humic material,
data were collected using 2,4,6-TCP as the sole chlorophenol adsorbate.
At pH 5.2, the presence of humic materials resulted in significant
reductions in the capacity of carbon for 2,4,6-TCP as compared with that
achieved in distilled water systems. Leaf fulvic acid proved to be the
most effective competitor of the three humic substances tested while soil
fulvic acid and commercial humic acids were nearly equally effective as
competitors. As shown in Table 15, 10 mg/1 of leaf fulvic acid reduced
the carbon's capacity at Ceq = 1-99 x 10-6 M 2,4,6-TCP by 52 percent.
Increasing the fulvic acid concentration to 50 mg/1 resulted in a slightly
greater reduction in 2,4,6-TCP capacity. Data for commercial humic and
soil fulvic acids show the same trends. At a 2,4,6-TCP equilibrium con-
centration of nearly an order of magnitude higher resulted in capacity
reductions which were not quite as large.
It was expected that leaf fulvic acid and commercial humic acid would
compete better with the chlorophenols than would soil fulvic acid due to
their greater adsorbability as illustrated in Figure 16. The data in
Figure 24 were collected at pH 7.0, however, so direct comparisons to the
data collected at pH 5.2 cannot be drawn. Humic substance removal as
measured by fluorescence showed the greatest removals for commercial humic
acid and the least for soil fulvic acid. However, leaf fulvic acid was the
best competitor with 2,4,6-TCP for adsorption sites while the other two
materials resulted in roughly equivalent competition. Hence, percent
103
-------
TABLE 15. 2,4,6-TCP COMPETITION WITH HUMIC SUBSTANCES - pH 5.2
CoTCR = 23.80 mg/1
TO"2 M phosphate buffer
Humic Substances
Type
Commercial
Humic
A/- T H
MC 1 U
Leaf
Fulvic
Arirl
Soil
Fulvic
AriH
Initial
Cone. % ,
mg/1 TOC Removed
10 83
70
50 66
48
10 56
37
50 28
16
10 48
23
50 27
0
Ceq
mole/1
4.41xlO"7
5.01xlO"6
l.OSxlO"6
9.18xlO"6
1.99xlO"6
1.04xlO"5
3.80xlO"6
1.73xlO"5
8.80xlO"7
5.72xlO"6
1.47xlO"6
9.94xlO"6
2,4
Xobs
mole/g
1.50xlO"3
2.32xlO"3
1.29xlO"3
1.96xlO"3
1.48xlO"3
2.21xlO"3
1.26xlO"3
l.SOxlO"3
1.50xlO"3
2.30xlO"3
1.27xlO"3
1.94xlO"3
,6-TCP
Xss
mole/g
2.63xlO~3
3.40xlO"3
2.68xlO"3
3.55xlO"3
S.llxlO"3
3.60xlO"3
3.31xlO~3
3.72xlO"3
2.81xlO"3
3.43xlO~3
3.02xlO~3
3.58xlO"3
Y Y
xobs/xss
57%
68
45
55
48
61
38
48
53
67
42
54
Percent removal as measured by fluorescence at pH 5.2.
Note: See Table 12A for definition of terms.
104
-------
TABLE 16. 2,4,6-TCP COMPETITION WITH HUMIC SUBSTANCES - pH 9.1
CoTCp = 23.80 mg/1
10~2 M phosphate buffer
Humic
Type
Commercial
Humic
An' rl
Leaf
Fulvic
An' H
Soil
Fulvic
Acid
Substances
Initial
Cone. % 1
mg/1 TOC Removed
10 75
59
50 57
39
10 61
42
50 44
22
10 30
14
50 21
11
2,4,6-TCP
C
I ec!
1 mole/1
1.05xlO"7
1.32xlO~6
2.35xlO"7
2.55xlO~6
l.OlxlO"7
1.53xlO"6
1.72xlO"7
2.25xlO"6
5.70xlO"8
8.61xlO"7
6.58xlO"8
1.04xlO"6
Xobs
mole/g
1.32xlO"4
3.25xlO"4
l.OSxlO"4
2.45xlO"4
1.30xlO"4
3.21xlO"4
1.09xlO"4
2.50xlO"4
1.30xlO"4
2.35xlO"4
l.lOxlO"4
2.69xlO"4
Xss
mole/g
2.06xlO"4
5.30xlO"4
l.SOxlO"4
6.90xlO"4
2.04xlO"4
5.55xlO"4
2.47xlO"4
6.40xlO"4
1.65xlO"4
4.50xlO"4
1.76xlO"4
4.81xlO"4
obs/Xss
64
61
60
36
64
58
44
39
79
74
62
56
Percent removal as measured by fluorescence at pH 9.1.
Note: See Table 12A for definition of terms.
105
-------
removal as determined by fluorescence apparently is not a good indicator of
the ability of the humic substances tested to compete with chlorophenols.
Further research is required to determine a parameter which better indicates
the ability of the humic materials to compete with the chlorophenols.
Results for 2,4,6-TCP competition with humic substances at pH 9.1 are
shown in Table 16. Commercial humic acid and leaf fulvic acid caused the
greatest reduction in capacity for 2,4,6-TCP while soil fulvic acid was not
as effective. Competition between the humic substances and 2,4,6-TCP was
significant. For example, 10 mg/1 leaf fulvic acid reduced the carbon's
capacity at Ceq = 1-01 x 10~7 M by 36 percent while 50 mg/1 leaf fulvic acid
reduced the capacity at Ceq = 1.72 x 10-7 M by 56 percent. At a 2,4,6-TCP
equilibrium concentration which is higher by an order of magnitude slightly
greater reductions in capacity resulted which were unexpected. At pH 5.2,
the opposite occurred, i.e., less competition was observed as the 2,4,6-TCP
equilibrium capacity was increased. The reasons for this behavior are not
clear at present. Percent removal of the humic substances as measured by
fluorescence at pH 9.1 were lower than those at pH 5.2 which is consistent
with the data shown in Figure 15 for the adsorption of soil fulvic acid as
a function of pH.
The effect of pH on the ability of the humic materials to compete with
2,4,6-TCP is noteworthy. At pH 5.2 leaf fulvic acid was the strongest
competitor while commercial humic and soil fulvic acids resulted in nearly
equal competition. At pH 9.1, commercial humic and leaf fulvic acids com-
peted more strongly with 2,4,6-TCP than did soil fulvic acid. Although more
complete studies are needed, it is expected that the different molecular
size distributions and functional group contents of the humic materials used
in these studies lead to differences in adsorbabilities as a function of pH.
This would explain the differences in competitive ability of the humic
materials observed at pH 5.2 and 9.1.
Gauntlett and Packham (1973) investigated the adsorption of monochloro-
phenol at the 0.1 to 1 mg/1 level in the presence of humic acid, fulvic acid
and in Thames River water. They show a significant reduction in capacity
(approximately 40 percent for the river water as compared with that achieved
in distilled water systems) owing to the presence of these materials which
is consistent with the results presented in this section. Unfortunately,
they did not state the pH of their studies. In our study, for example,
10 mg/1 soil fulvic acid at pH 5.2 reduced the capacity at Ceq = 8.80 x
10-7 M by 47 percent, while at pH 9.1 the capacity at Ceq = 8.61 x 10"7 M
was reduced by only 26 percent. This can be accounted for by considering
differences in adsorbabilities of soil fulvic acid at the two pH values;
at pH 5.2, 48 percent of the soil fulvic acid was adsorbed while at pH 9.1
only 14 percent was removed. Therefore at pH 9.1 less material was com-
peting with 2,4,6-TCP for adsorption sites than at pH 5.2
When the findings on competitive adsorption between chlorophenols and
humic substances together with the information reported previously on com-
petition between individual chlorophenols are considered, it is apparent
that any testing to determine the best carbon and design criteria for a
106
-------
given application should be done using the natural water to be treated.
Such factors as the nature and concentration of competing organic materials,
pH and salt content of the water, among others, play a major role in deter-
mining the removal one can expect of a certain component.
ADSORPTION OF POLYNUCLEAR AROMATIC HYDROCARBONS
An initial test of the adsorbability of benzanthracene from water was
made to first evaluate the effectiveness of the carbon. Solutions of various
concentrations of benzanthracene were shaken with 10 mg of F-400 activated
carbon for eight hours. After this time the concentrations were measured by
fluorescence (see Table 17). Thus these extremely insoluble, non-polar com-
pounds are adsorbed very well on active carbon.
TABLE 17. CONCENTRATIONS OF BENZANTHRACENE AFTER EIGHT HOURS
Volume
(ml)
150
150
150
c
o Concentration Remaining
(yg/D (yq/D
242
50
10
12
<1
<1
Because the PAH was removed to below detection limits at such low
dosages of carbon, the competitive effect of natural organic matter was
examined. A kinetic study was performed to establish whether the presence
of high concentrations of commercial humic acid would interfere with the rate
of adsorption. Bottles with and without humic acid were set up with 1 mg/1
of anthracene and taken off each day. A plot of the residual concentration
vs. time both with and without humic acid is shown in Figure 28. No effect
of the humic acid was observed. A series of bottles with 10 mg/1 of carbon
was set up with half of the bottles containing humic acid and anthracene
and the other half with anthracene alone. Two bottles, one with and one
without humic acid, were taken off the shaker each day and the anthracene
concentrations were compared. Plots of residual concentration vs. time are
shown in Figure 42. No effect of the humic acid was observed.
In our experiments we were using abnormally high concentrations of PAH.
The anthracene was present as a particulate because concentrations exceeded
the aqueous solubility of these extremely hydrophobic compounds. Even with
large aqueous concentrations and extremely low carbon dosages we experienced
removal of the PAH to below the detection limit, approximately 1 yg/1, after
shaking for several days. It was believed that the PAH might associate with
natural organic material in water and experience a decreased rate of adsorp-
tion or a decreased capacity on carbon. This was not found to be the case.
107
-------
IOOO
100
0)
u
c
o
o
10
A Without Humic Acid
O With Humic Acid
Days
Figure Q2. Kinetics of anthracene adsorption with and without humic acid
108
-------
However, it should be stressed that our solutions were not natural and that
the particles of anthracene in solution will not behave as individual mole-
cules with respect to association with humic acid.
On this basis it does not appear likely that PAH will associate with
relatively poorly adsorbable humic substances and result in the "leakage"
of PAH from carbon beds. Based on a review of the literature as presented
in Section 1, it is possible that other organics such as pesticides may
penetrate carbon beds in this fashion.
109
-------
REFERENCES
Andelman, J. B. 1973. World Health Organization, European Standard for
Organic Matter in Drinking Water. In: Proc. 15th Water Quality
Conference, Organic MatteA -in (UateA Sapptiu'- OccuAAence, S+.gnifii.c.anc.e.
and ContAol (V. L. Snoeyink, ed.), University of Illinois, Urbana, IL.
70:55.
Burke, J. A., P. A. Mills, and D. C. Bostwick. 1966. JOUA. A^oc.
. Chum. 49:999.
Burtschell, R. H., A. A. Rosen, F. M. Middleton, and M. B. Ettinger. 1959.
Chlorine Derivatives of Phenol Causing Taste and Odor. JOUA. AmeA.
VJoAkA A64oc. 51:205.
Butler, J. A. V. and C. Ockrent. 1930. Studies in Electrocapillarity. III
JOUA. Phy&. Cham. 34:2841.
Capmau, M. L., W. Chodkiewicz, and P. Cadiot. 1968. Stereochimie de 1 'Addi
tion d'Organometalliques a et 3 Insatures sur le (t) Camphre.
Soc. ChMn. Trfianc-d. p. 3233.
Calgon Corp. 1969. Calgon Activated Carbon Product. Bull. No. 20-2a,
Calgon Corp., Pittsburgh, PA.
Crosby, D. G. and A. S. Wong. 1973a. Photodecomposition of p-Chlorophenoxy-
acetic Acid. JOUA. AgA. Pood Ckzm. 21:1049.
Crosby, D. G. and A. S. Wong. 1973b. Photodecomposition of 2,4,5-Trichloro-
phenoxyacetic Acid (2,4,5-T) in Water. JOUA. Ag/i. food Ck&n. 21:1052.
Dice, J. C. 1976. The Challenge to the AWWA Taste and Odor Control Com-
mittee. Proc. AWWA Conference on Water Quality, Atlanta, Dec. 1975,
AWWA, Denver, CO.
Dostal , K. A., R. C. Pierson, D. G. Hager and G. G. Robeck. 1965. Carbon
Bed Design Criteria Study at Nitro, West Virginia. JOUA. AmeA.
AMCC. 57:663.
Fieser, L. F. and G. Ourisson. 1953. Chromates d'alcools Tertiaires.
Butt. Sec. Ckim. f fiance., p. 1152.
Friestad, H. 0., D. E. Ott and F. A. Gunther. 1969. Automated Colorimetric
Microdetermination of Phenols by Oxidative Coupling with 3-Methyl-2-
Benzothiazoline Hydrazone. AnaJL. Ckzm. 41:1750.
110
-------
Gauntlett, R. B. and R. F. Packham. 1973. The Use of Activated Carbon in
Water Treatment. In: Proc. Conference on Activated Carbon in Water
Treatment, University of Reading, Water Research Association,
Medmenham, England.
Gerber, N. N. 1974. Microbiological Production of Geosmin. EPA-670/2-74
094, U. S. Environmental Protection Agency, Cincinnati, OH.
Gerber, N. N. 1967.
Actinomycetes.
Geosmin, an Earthy-Smelling Substance Isolated from
&Lote.ch. and Rioting. 9:321.
Gerber, N. N. and H. A. Lechevalier. 1965.
Substance Isolated from Actinomycetes.
Geosmin, an Earthy-Smelling
Appt. blLcAob-iot. 13:935.
Gjessing, E. T.
Humus. Ann
1976. Physical and Chemical Characteristics of Aquatic
Arbor Science Publishers, Ann Arbor, MI.
Glaze, W. H. and J. E. Henderson. 1975. Formation of Organochlorine
Compounds from the Chlorination of a Municipal Secondary Effluent.
Joufc. WateJi Polt. Con&iot f&d. 47:2511.
Handbook o£ Cke.mJJ>&Ly and
Cleveland, OH.
1967. 48th Ed., Chemical Rubber Co.,
Herzing, D. R., V. L. Snoeyink and N. F- Wood. 1977. Activated Carbon
Adsorption of the Odorous Compounds 2-Methylisoborneol and Geosmin.
JOU/L. AmcA. WateA. Wonkt, /Woe. 69:223.
Jain, J. S. and V. L. Snoeyink. 1973. Adsorption from Bisolute Systems on
Active Carbon. JOUA. Watax. Pott. Con&tot Fed. 45:2463.
Jenkins, D. 1973. Effects of Organic Compounds - Taste, Odor, Color, and
Chelation. In: Proc. 15th Water Quality Conference, O^igan^a Matt&i -in
Wat&L SupptieA'- QdusJiA.ino.iL, S-ignifi-icanzz and Con&iot (V. L. Snoeyink,
ed.), University of Illinois, Urbana, IL. 70:15.
Kaiser, K. L. E. and B. G. Oliver. 1976.
genated Hydrocarbons in Water by Gas
48:2207.
Determination of Volatile Halo-
Chromatography. Anat. Ck&m.
Kikuchi, T., T. Mimura, Y. Itoh, Y. Moriwaki , K. Negoro, Y. Masada and
T. Inoue. 1973a. Odorous Metabolites of Actinomycetes, Biwake-C and
-D Strain Isolated from the Bottom Deposits of Lake Biwa. Identifica-
tion of Geosmin, 2-Methylisoborneol, and Furfural. Ck&tn. and Pkatm.
KvJUL. 21:2339.
Kikuchi, T., T. Mimura, K. Haranaya, H. Yano, T. Arinoto, Y. flasada and
T. Inoue. 1973b. Odorous Metabolite of Blue-Green Algae-' Sck
-------
Korenman, Y. 1974. Distribution of Chlorophenols between Organic Solvents
and Water. JOUA. Appi. Ckm. - USSR. 47:9 (part 2).
Ku'hn, W. and H. Sontheimer. 1973a. Several Investigations on Activated
Charcoal for the Determination of Organic Chloro Compounds, l/om Wo6-6eA.
41:65.
Ku'hn, W. and H. Sontheimer. 1973b. Analytic Determination of Chlorinated
Organic Compounds with Temperature Programmed Pyrohydro lysis, l/om
41:1.
Langmuir, I. 1918. The Adsorption of Gases on Plane Surfaces of Glass, Mica
and Platinum. JOUA. AmeA. Ckm. Son.. 40:1361.
Lee, C. F- 1967. Kinetics of Reactions Between Chlorine and Phenolic
Compounds. In: Psbinc/iptu and Application* o£ UateA Ckem'iA&ty (S.
Faust and J. Hunter, eds.), Wiley and Sons, Inc., New York.
Leenheer, J. A. and E. D. W. Huffman. 1976. Classification of Organic
Solutes in Water Using Macroreticular Resins. Submitted to JOUA. ojj
ReieoAc^ for publication.
Lenz, V. W. 1911. Zur Prufung des Kampfers. AAC/U.V. dun. P/ioAmaz-te..
249:286.
Love, 0. T., Jr., G. G. Robeck , J. M. Symons and R. W. Buelow. 1973.
Experience with Activated Carbon in the U.S.A. Proc. Conference on
Activated Carbon in Water Treatment, University of Reading, Water
Research Association, Medmenham, England, p. 279.
Malkonen, P. 1964. Ann. Ac.ad. Sex. FemUcae SeA. All. No. 128.
McCaull, J. and J. Crossland. 1974. Wat&i Pollution (B. Commoner, ed.),
Harcourt Brace Jovanovich, Inc., New York.
McCreary, J. J. and V. L. Snoeyink. 1977- Granular Activated Carbon in
Water Treatment. JOUA. AmeA. Wo£e.A WoAfci AMOC., in press.
McGinnes, P. R. and V. L. Snoeyink. 1974. Determination of the Fate of
Polynuclear Aromatic Hydrocarbons in Natural Water Systems. Res.
Rept. No. 80, Water Resources Center, University of Illinois at
Urbana-Champaign, Urbana, IL.
Medsker, L. L., D. Jenkins and J. F. Thomas. 1968. Odorous Compounds in
Natural Waters: An Earthy-Smelling Compound Associated with Blue-
Green Algae and Actinomycetes. Env-ision. Sex., and Tzcknol. 2:461.
Medsker, L. L., D. Jenkins, J. F. Thomas and C. Koch. 1969. Odorous Com-
pounds in Natural Waters: 2-exo-hydroxy-2-methylbornane, the Major
Odorous Compound Produced by Several Actinomycetes. Env-iion. Se-c.
and Tachnol. 3:476.
112
-------
Nakanishi, K. 1962. In^fcuizd AbAostption Spuc&ioAcopy, Holden-Day, Inc.,
San Francisco, p. 24.
Narkis, N. and M. Rebhun. 1975. The Mechanism of Flocculation Processes in
the Presence of Humic Substances. JOUA. Ame/i. Uatoji WoAfe4 Awoc.
67:101.
Pharmacia. 1974. Sephadex: Gel Filtration in Theory and Practice.
Upplands Graf i ska AB, Sweden.
Piet, G. J., C. J. Zoetman and J. A. Kraayeveld. 1972. Earthy-Smelling
Substances in Surface Waters of the Netherlands. PAOC. Soc. Wo^eA
T/izat. and Exam. 21 :281.
Robeck, G. G. 1975. Evaluation of Activated Carbon. Water Supply Research
Laboratory, National Environmental Research Center, Cincinnati, OH.
Rook, J. J. 1976. Haloforms in Drinking Water. JouA. AmeA. (riat&i
68:168.
Rosen, A. A., C. I. Mashni and R. S. Safferman. 1970. Recent Developments
in the Chemistry of Odor in Water: The Cause of Earthy-Musty Odor.
PAOC. Wat&i Tsizat. and Exam. 10:106.
Safferman, R. S., A. A. Rosen, C. I. Mashni and M. E. Morris. 1967. Earthy-
Smelling Substance from a Blue-Green Algae. Envision. Sc.i. and Tzcknot.
1:429.
Schnitzer, M. and S. U. Khan. 1972. Hamic. SubAtanceA -in tkn. Environment,
Marcel Dekker, Inc., New York.
Schweer, K. H., F. Fuchs and H. Sontheimer. 1975. Untersuchungen zur
Summarischen Bestimmung von Organisch Gebundenem Schwefel in Wassern
and auf Aktivkohlen. l/om Wo^eA. 45:29.
Sigworth, E. A. 1957. Control of Taste and Odor in Water Supplies.
49:1507-
Silvey, J. K. G., D. E. Henley, R. Hoehn and W. C. Nuney. 1976. Musty-
Earthy Odors and Their Biological Control. Proc. Conference on Water
Quality, Atlanta, Dec. 1975, AWWA, Denver, CO.
Snoeyink, V. L., W. J. Weber and H. B. Mark. 1969. Sorption of Phenol and
Nitrophenol by Active Carbon. Envision. Sex., and Te.c/mo£. 3:918.
Sontheimer, H. 1974. Use of Activated Carbon in Water Treatment Practice
and its Regeneration. Special Subject No. 3, Lehrstuhl fur Wasser-
chemie der Universitat Karlsruhe, Karlsruhe, West Germany.
113
-------
Sontheimer, H. and D. Maier- 1972. Untersuchungen zur Verbesserung der
Trinkwasseraufbereitungstechnologie am Niederrhein (1. Bericht). Go*
inn A ll\n L f.n H Unnln \\\n A A o h /AhlilnAA Oh . 1 1 "\ ' 1 ff7 .
and WaAAHfi^ack Woi-ieA/AbwiaMe/i. 113:187.
Stahl , E. (ed.). 1969. Tkin-Layeji Ckwmatogsiaphy. Springer-Verlag, New
York. p. 224.
Standard ttnthodb ^on tke. EKcmi.Yiat.J-on ofi Wat&i and Ua&tewateJi, 1975. 14th
Edition, American Public Health Association, Washington, D.C.
Stevens, A. A. and J. M. Symons. 1977. Measurement of Trihalomethane and
Precursor Concentration Changes Occurring during Water Treatment and
Distribution. Accepted by JouA. AmeA. UateA Wonkh A^^oc. for publi-
cation, October 1977.
Stevens, A. A., C. J. Slocum, D. R. Seeger and G. G. Robeck. 1976. Chlori-
nation of Organics in Drinking Water. JOUA. Ame/t. Watasi Woife/i Ai^oc.
68:615.
Stevenson, F. J. 1965. Gross Chemical Fractionation of Organic Matter.
In: M&tkodA o& Soit knalyA-u , Part 2, Agronomy Monograph #9 (C. A.
Black it aJL. , eds . ) , American Society of Agronomy, Ma'dison, WI, p. 1409
Symons, J. M. 1976. Interim Treatment Guide for the Control of Chloroform
and Other Trihalomethanes. Water Supply Research Division, Municipal
Environmental Research Laboratory, Environmental Protection Agency,
Cincinnati , OH.
Toivonen, H. 1968. Uber Die Reaktionen Des 2-Methyl isoborneols und der
2-Methylfenchols in Konzentrierter Salpetersaure. T&&iake.dAon L&ttzu.
p. 3041.
U.S. Environmental Protection Agency. 1971. Methods for Organic Pesticides
in Water and Wastewater. National Environmental Research Center,
Cincinnati , OH.
U.S. Environmental Protection Agency. 1975. Preliminary Assessment of
Suspected Carcinogens in Drinking Water - Interior Report to Congress.
Office of Toxic Substances, Washington, D.C.
Ward, T. M. and F. W. Getzen. 1969. A Model for the Adsorption of Weak
Electrolytes on Solids as a Function of pH--l . Joun.. Cottoi.d and
SCA-. 31:441.
Ward, T. M. and F. W. Getzen. 1970. Influence of pH on the Adsorption of
Aromatic Acids on Activated Carbon. EniMAon. Sc-t. and Jichyiot. 4:64.
Weber, W. J., Jr. and J. C. Morris. 1964. Adsorption in Heterogeneous
Aqueous Systems. Jou/i. Ame/z.. W
-------
Zeiss, M. M. and D. A. Pease. 1956. Chromate Esters. II. Reactions of
Di-2-Methylfenchyl Chromate. JOUA. AmeA. Ckw. Sec. 78:3182.
Zelinsky, N. 1901. Uber eine Synthese der Cyclischentertiaren Alkohole mit
Hulfe von Magnesiumhalogenalkylen. BeA. V&>ck. Cham. Ge/s. 34:2877.
Zogorski, J. S. and S. D. Faust. 1974. Removal of Phenols from Polluted
Waters. New Jersey Water Resources Research Institute, Rutgers Univ.
Zogorski, J. S. and S. D. Faust. 1976. The Effect of Phosphate Buffer on
the Adsorption of 2,4-Dichlorophenol and 2,4-Dinitrophenol . JOUA.
Environ. Set. H&altk. All (8&9):501.
115
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-77-223
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
ACTIVATED CARBON ADSORPTION OF TRACE
ORGANIC COMPOUNDS
5. REPORT DATE
December 1977 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Vernon L. Snoeyink, John J. McCreary and
Carol J. Murin
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Dept. of Civil Engineering
University of Illinois
Urbana, Illinois 61801
10. PROGRAM ELEMENT NO.
1CC614
11. CONTRACT/GRANT NO.
R-S03473
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory- Cin., Of
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final, 1/6/75 to 7/5/77
14. SPONSORING AGENCY CODE
EPA/600/14
15. SUPPLEMENTARY NOTES
Project Officer: Alan A. Stevens, 513/684-7228
16. ABSTRACT
Research was conducted to determine how effectively humic substances
and the trace contaminants 2-methylisoborneol (MIB), geosmin, the
chlorophenols and polynuclear aromatic hydrocarbons were adsorbed by
activated carbon under the competitive adsorption conditions encountered in
natural waters. Humic substances compete with MIB and geosmin for
adsorption sites on activated carbon and significantly reduce its capacity
for these compounds. These naturally occurring odorous compounds were
found to be much more strongly adsorbed than the humic substances.
Both the chlorophenols and the polynuclear aromatic hydrocarbons are
very strongly adsorbed. Strong competition was observed between anionic
and neutral species of 2,4-dichlorophenol and 2,4,6-trichlorophenol. The
presence of the various humic substances caused a significant reduction in
chlorophenol adsorption capacity. Humic acid did not interfere with the
rate of adsorption of a model polynuclear aromatic hydrocarbon, anthracene.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Activated Carbon Treatment,
Potable Water, Odor Control,
Organic Compounds, Actinomyces
Competitive Adsorption,
Humic Acids, Chloro-
phenols, Geosmin,
Methylisoborneo.1,
Taste and Odor
13 B
3 DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
126
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
16
•ft U. S. GOVERNMENT PRINTING OFFICE: 1978-757- 1W6660 Region No. 5-1
------- |