This item is only available as the following downloads:
THE NATURE OF ORGANIC
COLOR IN WATER
RUSSELL FABRIQUE CHRISTMAN
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
The author wishes to express his earnest appreciation to his
committee chairman, Dr. A. P. Black, for guidance and encouragement
given during this investigation and throughout the author's graduate
training. He would also like to thank the other members of his
supervisory committee, Dr. W. S. Brey, Dr. J. M. Pearce, Dr. H. E.
Schweyer, and particularly Dr. J. D. Winefordner.
Special thanks are due many of the author's associates, princi-
pally Mr. Howard Latz and r.John Wilson for their many helpful sugges-
The author also wishes to express his gratitude to the National
Institutes of Health whose financial support made this project possible.
Finally, the author gratefully acknowledges his indebtedness
to his wife, Sylvia, for her support and understanding.
TABLE CF CONTENTS
LIST OF TABLES .
LIST OF FIGURES .
II. HISTORICAL .
V. SUMMARY .
LIST OF REFERENCES. .
BIOGRAPHICAL SKETCH .
* m 6 0 5 S
* S S 0
f f ft
* 5 0 5 U 5 0 a a S 0 a
S U 6 0 o 0 a
S S S a a 0 O 0 S S 0 0 5 S .a
LIST OF TABLES
1 Classical Fractions of Soil Organic Matter 8
2 Molecular Weights of Some Humic and Hymatomelanic
3 Effect of Storage on Color in a Nitrogen Atmosphere 27
4 Effect of Storage on Color in an Osygen Atmosphere 28
5 Sources and Designations of Colored Waters 29
6 Chemical Analyses of Waters A Through E 31
7 Chemical Analyses of Waters F Through J 32
8 Variation of Iron Content in Colored Water Samples 33
9 Variation of Iron Content in Colored Water Samples 33
10 Variation of Color with pH 34
11 Electrodialysis and Millipore Filtration of Water A 36
12 Scattering Ratios of Latex Suspensions 37
13 Solubility Relationships of Fractions Obtained by
Shapiro Technique 44
14 Weight Distribution in Fractions Obtained by
Shapiro Technique 44
15 Ultimate Analyses of Fractions Ia and Ib 48
16 Weight of Organic Matter in Fractions 49
17 Percentage of Total Organic Matter in Fractions 50
18 Infra-Red Absorption Bands of Fulvic Acid Fractions 53
19 Ultimate Analyses of Fractions of Water A 55
20 Paper Chromatographic Separation of Fulvic Acid
Fraction of Water A 56
LIST OF TABLES (Continued)
21 Paper Chromatographic Separation of Humic Acid
Fraction of Water A 57
22 Paper Chromatographic Separation of Hymatomelanic
Acid Fraction of Water A 57
23 Equivalent Weight and Quantitative Methylation
Data of Fulvic Acid Fractions 60
24 Quantitative Methylation of Model Compounds 61
25 Effects of C12, ClO2 and 03 on Color Removal 64
26 Effects of C12, C10 and 03 on Chemical Oxygen
Demand of Colored Water Concentrates 64
27 Infra-Red Absorption Bands of Ether Soluble Chlorine
Oxidation Products 65
LIST OF FIGURES
1 Effect of pH on Color of Waters A and G 75
2 Effect of pH on Color of Water B 76
3 Effect of pH on Color of Waters D and E 77
4 Dialysis and Millipore Filtration of Waters A Through E 78
5 Dialysis and Millipore Filtration of Waters F Through J 79
6 Light Scattering Properties of Water B at 546 ma 80
7 Light Scattering Properties of Water B at 436 m4 81
8 Effect of pH and Constant Alum Dosage on Coagulation
of Water B 82
9 Effect of pH and Constant Alum Dosage on Coagulation
of Water I 83
10 Variation of pH Width of Optimum Color Removal with
pH Width of Charge Reversal 84
11 Spectrum of Mercury Lamp Source with Distilled Water 85
12 Fluorescence Spectrum of Water B 86
13 Fluorescence Spectrum of Water F 87
14 Effect of pH on Fluorescence Emission of Water B 88
15 Ultraviolet Absorption of Water B as Function of
Color Value 89
16 Ultraviolet Absorption of Water I as Function of
Color Value 90
17 Absorption of Colored Waters at 220 m) 91
18 Absorption of Colored Waters at 300 mn 92
19 Absorption of Colored Waters at 350 mu 93
LIST OF FIGURES (Continued)
20 Effect of pH on Ultraviolet Absorption of Water B 94
21 Effect of pH on Ultraviolet Absorption of Water C 95
22 Infra-Red Spectrum of 'Free Acid" Fraction II 96
23 Fractionation Scheme 97
24 Relation Between Color Value and Total Organic
Content of Colored Waters 98
25 Relation Between Color Value and Concentration for
Humic and Hymatomelanic Acid Solutions 99
26 Infra-Red Absorption Spectrum of Fulvic Acid Fraction
of Water B 100
27 Infra-Red Absorption Spectrum of Fulvic Acid Fraction
of Water C 101
28 Infra-Red Absorption Spectrum of the Methyl Ester of
Fulvic Acid Fraction of Water A 102
29 Potentiometric Titration of Benzoic Acid in
30 Potentiometric Titration of Phenol in Ethylenediamine 104
31 Potentiometric Titration of Resorcinol in Ethylene-
32 Potentiometric Titration of o-Hydroxybenzoic Acid
in Ethylenediamine 106
33 Potentiometric Titration of p-Hydroaqbenzoic Acid
in Ethylenediamine 107
34 Potentiometric Titration of Fulvic Acid Fraction
of Water B in Ethylenediamine 108
35 Potentiometric Titration of Fulvic Acid Fraction
of Water I in Ethylenediamine 109
36 Infra-Red Absorption Spectrum of Chlorine Oxidation
Product from Water A 110
37 Infra-Red Absorption Spectrum of Chlorine Oxidation
Product Extracted from Miami Tap Water 111
LIST OF FIGURES (Continued)
38 Infra-Red Absorption Spectrum of a Fulvic Acid
Oxidation Product 112
39 Infra-Red Absorption Spectrum of a Fulvic Acid
Oxidation Product 113
Waters containing color resulting from their natural environ-
ment are found throughout most parts of the world. While organic color
is most frequently found in surface waters, it is occasionally found in
shallow or deep wells in limestone regions where solution topography
The United States Public Health Service (1) has set a maximum
limit for drinking waters of 15 color units on a platinum-cobalt scale.
This limit has been imposed mainly for aesthetic reasons as the materials
responsible for color in water are not known to be physiologically
harmful. In practice, most municipal treatment plants strive to maintain
finished waters with color values of 10 or less.
Almost all of the research to date involving naturally colored
water has been concerned either with empirical methods of treatment for
color removal or with the mechanism of color coagulation. Little
attention has been directed to the fundamental nature of the organic
materials that are responsible for color in water. Therefore, the pri-
mary purpose of this investigation is to study various physical and
chemical properties of organic color in water, in the hope that such basic
information may eventually lead to a more complete understanding of
Because of the widespread occurrence of this type of water, it
is an additional objective of this research to determine the uniformity
of these properties in colored waters from a variety of sources.
Nature of Organic Matter in Water
Aschan (2) carried out the first extensive research on these
substances with an analysis of six Finnish lake and river waters. The
humus material was precipitated with FeC13 and subjected to ultimate
analysis with the following results: carbon content was found to vary
from 44.99 to 54.10 per cent, hydrogen 3.86 to 5.05 per cent, nitrogen
1.46 to 4.23 per cent and oxygen from 38.76 to 47.93 per cent. He con-
cluded that these materials were acids of strength comparable to phenol.
Birge and Juday (3) observed that the carbon/nitrogen ratio was
larger in more highly colored lakes, a fact which has led many authors
to regard nitrogen as an impurity.
Saville (4) reported from the results of some qualitative
electrophoretic experiments that organic color is present in water as
a negatively charged colloid. Behrman, Kean and Gustafson (5) found
that little color in water would dialize through a parchment membrane,
indicating a colloidal nature. These authors also noted the indicator
action of colored water and that most of the color could be oxidized by
Christman (6) determined the rate of migration in cm/sec/volt
of organic color bodies in 32 natural waters and of two aqueous dye
solutions in an electrophoresis cell. All colloids present in the 32
waters were found to be negatively charged.
In 1910, Dienert (7) detected fluorescent materials present in
natural waters and noted that this property was lost by mild oxidation.
The use of absorption spectra analyses for the investigation
and routine checking of contaminated surface water was first described
in 1936 by Denmering (8). This author found the absorption bands pro-
duced by lignin substance offered a means of checking changes in the
concentration of organic material. Demmering applied this as a
criterion for amount of chemicals to be used for chemical treatment.
Datsko (9) carried out photometric investigations of color re-
actions of sulfuric acid with organic substances in natural waters. He
observed an intense coloration produced in the reaction and noted that
the intensity followed Beer's law. No description was furnished as to
the manner of preparation of his humic acid.
Keiser (10) has reported a differentiation of organic substances
in natural water. According to this author alum removes more of the
oxygen oxidizable material than the chlorine oxidizable material.
More recently, Flaig and Otto (11) have carried out investiga-
tions of humic materials and model substances on plant life. These
authors added polyhydroxyanthraquinone to spring water at a concentration
of 10-5 gms/ml and found that the plants grew normally, using this
material precisely as plants use natural humic acid.
Goryunova (12) has investigated the character of dissolved
organic substances in water of Glubokoe Lake, Russia. The water of this
lake was found to contain 96 gms/m3 of organic matter, the bulk of this
material being made up of high molecular weight fatty acids and colloidal
complexes. He also noted small amounts of substances that would reduce
According to Isachenko and Egorova (13) the earthy smells in
reservoirs are caused by actinomycetes. Odors of this type have been
noted frequently in concentrates of organic colored water.
In 1954 Ponomareva and Ettinger (14) obtained dark, easily
precipitated substances of the humic-ulmic acid type from organic matter
in waters of the Neva River. This material was found to comprise only
5 to 6 per cent of the total organic matter. In addition, apocrenic
acid concentrations were found to reach the same percentage. The
remainder of the organic material was thought to consist of lightly
oxidized fractions of crenic acids.
Goryunova (15) has isolated and identified 50 per cent of the
organic matter in the water of Beloe Lake, Russia. This author concluded
these are humus materials but fatty acids and polysaccharides are also
Skopintsev and Krylova (16) have stated that aqueous humus of
terrigenous origin is responsible for the yellow color of water. These
authors found a linear relation between color intensity determined
colorimetrically and the coefficient of extinction determined spectro-
photometrically. No such relation was found to exist between the total
organic content and either color intensity or extinction coefficient.
Raudnitz (17) isolated a water soluble, surface-active phosphoric
ester of humic acid from rhododendron leaves. Subsequent hydrolysis
yielded a uniform humic acid free of nitrogen and sulfur, insoluble
in common organic solvents but soluble in pyridine and aniline. Infra-
red data showed the presence of hydroxyl and carbonyl groups.
Wilson (18) stated that colored organic materials in water are
fulvic acids, a fraction of natural soil humus. The author believed no
specific chemical test for these materials will be developed because
of the diversity of functional groups present, but a quantitative
estimation of these acids can be obtained from their ultraviolet absorp-
tion spectra. Using a standard solution of fulvic acids from
phragmites peat, Wilson observed that for a given concentration absorption
at 300 m)n was constant in the pH range 1.0 to 5.0 increasing over pH 6.0.
Application of this technique to several natural waters led to the
following conclusions (a) All samples contained appreciable quantities
of fulvic acids. (b) All absorption spectra were similar. (c) Since
fulvic acid is the most water soluble fraction of natural soil humus it
would be expected to be found in natural waters in higher concentrations
than either humic or hymatomelanic acids.
Fouling of anion exchange resins by organic materials in colored
water is a problem of some industrial significance. Numerous authors
have observed organic fouling of strongly basic anion exchange resins
of the quaternary types I and II and have ascribed the differences in
their fouling rates to the structural differences in their quaternary
ammonium substituents. The foulants are held on the resins rather
strongly and are not removed by normal regeneration. Excellent reviews
of this problem are provided by Frisch et al. (19) and Frisch and
Kunin (20). In addition, Ungar (21) has reported fouling of weakly basic
anion exchangers and that resins of this type exhibit a limited exchange
capacity for humic acids.
One of the most recent publications in this area is an extensive
investigation of the yellow organic acids of lake water carried out by
Shapiro (22). This author has outlined a separation scheme for the
isolation of a "free acid" fraction of the organic materials present in
the water. Data obtained by infra-red and ultraviolet spectroscopy and
by various organic spot reactions led this author to report that these
acids are aliphatic unsaturated polyhydroxy dicarboxylic acids with an
approximate molecular weight of 456. No evidence was found in this work
of the existence of aromatic rings and dialysis studies indicated a lack
of colloidal character. Shapiro suggests the name "humolimnic" acids
for these materials to indicate their probable relation but improbable
equivalence with natural humic acids of the type found in soils.
The Nature of Organic Material ig Soils
Because of the relation of these coloring substances to the
organic matter present in soils, it is necessary to review the work in
this area of chemistry. No attempt will be made to critically outline
the early developments in this field as this has been done by Waksman (23).
However, the major findings will be discussed particularly where there is
an appreciable correlation with data obtained on the organic materials in
The term humus dates back to the Roman era when it was generally
applied to the complete soil substance. From this time to around the
17th century little was known about the nature or functions of soil.
At the start of the 18th century Linnaeus classified soils in a manner
similar to his classification of plants, humus daedalea (garden soil),
humus ruralis (field soil), etc. In the latter half of the 18th century,
humus began to be considered in terms of decaying organic matter. How-
ever, the general ideas concerning the chemical nature of humus were
Unfortunately, the term humus came into general use when organic
chemistry was in its infancy and was not always used to designate the
same organic materials. This was responsible for some degree of con-
fusion among workers in the field of soil science, a situation which
has never been entirely remedied.
Perhaps the earliest scientific classifications of humus are
chose of Oden (24) and Ramann (25). Ramann regarded humus as "colloidal
complexes of varying composition, consisting of unchanged colloids of
the original plant substance mixed with carbon-rich decomposition
products." Ramann recognized the importance of micro-organisms in the
synthesis of humus, a fact disregarded by Oden. The latter regarded
humus as "those dark colored bodies of unknown constitution which origi-
nate in nature through the decomposition of organic substances...; they
possess a definite affinity for water and show, if not true solution, a
Oden also supplied the first generally accepted system of naming
the four common fractions or preparations of soil organic matter. Up
until this time there were in common use 43 different names for the four
Oden's classification scheme is shown in Table 1.
In 1930 this system was modified by Page (26), and resulted in
the most widely used classification system to date. He suggested the
name "humin" for preparation 4 and humicc acid" for preparation 2.
The multitude of names and classification systems during this
early period was due to a large extent to the variety of sources from
which the humus materials were obtained. The early theories about the
simplicity and uniformity of the substances in soil were gradually
Classical Fractions of Soil Organic Matter
1. Soluble in alkali, not precipitated Fulvic Acid
2. Soluble in alkali, precipitated by HC1, Huus Acid
insoluble in alcohol
3. Soluble in alkali, precipitated by HC1, Hrmatomelanic Acid
soluble in alcohol
4. Insoluble in alkali Humus Coal
replaced by the 20th century theories that the composition of humus
material and to some extent the properties, were dependent on the source,
manner of extraction, and length of formation time of the organic sub-
One of the earliest and perhaps the most persistent theory of the
origin of humus was that it was produced by the action of mineral acids
on sugars. This trend in research prompted many investigators to produce
artificial humus materials in the belief that these were formed along
the same pathways that nature employs in producing humus. Boullay (27)
calculated the chemical constitution of an artificial preparation as
C30H30015. Malaguti (28) suggested that HN03 changes sugars into oxalic
acid whereas HC1 and H2SO4 give humus matter, Under the influence of
acid, the sugar loses water giving first a compound of the structure
C12H28014 which changes to humic acid. The mechanism was as follows:
C12H22011 > C1281206 + 5 H20
Sucrose Humic acid
Stein (29) suggested the formula C24H1809 for the same material
and Berzellius (30) adopted yet another formula, namely C32H12016 for
Berzellius was one of the first workers to isolate humus
material from a natural water. He believed these compounds were also
present in plants and soils. He suggested the name crenic acid for his
preparation which he believed could be oxidized to apocrenic acid by the
action of air. These preparations had formulas of C24H12016 and
C24H6012, respectively. In addition he noted the salts of these acids
with alkalies. Magnesium and iron were soluble in water and could be
transported from soil to water in this manner.
A most extensive study of dark colored compounds extracted from
soil, peat, coal and prepared artificially in the laboratory was made by
Hermann (31). His system of classification only added to the complexity
of the problem but he did note the precipitative effect of heavy metal
acetates on these substances and the fact that although most humus
materials contain nitrogen, many do not.
Around the turn of the century investigators began to doubt the
validity of explaining natural processes of humus formation by results
obtained from artificial materials produced by treatment of sugars and
other complex organic substances. Attention turned towards investigation
of the physical and chemical properties of the natural humus. Thus,
Sestini (32) demonstrated in the humic complex the presence of ethereal
and anhydride substances, as well as hydrorl, alkyl and ketonic groups.
The recognition of the mixed and indefinite nature of the various
humic acids and the turning of attention towards the elucidation of the
complex chemical composition of the natural materials was the most import-
ant advance in the history of research in this area.
Van Bemmelen (33) was one of the first researchers to investi-
gate the colloidal properties of soil organic matter. He believed that
the humus complexes were amorphous and colloidal in nature and origi-
nated from the colloidal substances in plant life.
Hoppe-Seyler (34) investigated both natural and artificial
products and found that treatment with alkali gave dark solutions in
the presence of air, but acid treatment gave the same result even in a
hydrogen atmosphere. Included in his organic compounds were pyrogallol,
pyrocatechuic acid, pyrocatechin, and quinone, which all gave similar
humic materials on oxidation. He noted xylan, or wood gum, in soil in
an unchanged state, and found that lignin complexes take an active part
in the formation of humic acids.
Miklauz (35) pointed out that the compositions of the alkali
and alcohol soluble fractions are affected by the strength of extract-
ment and the extracting temperature; and that the highest carbon content
is usually found in the alcohol soluble fraction.
At this time a new concept appeared which has been called the
colloid-chemical approach, mostly due to the outstanding works of Van
Bemmelen, Baumann and Gully (36). This new trend was based upon two
main ideas. The first was that the acidity of humus is not due to free
humic acids but is associated with the reactions of humus with mineral
salts, bases being adsorbed on the colloidal surfaces of the organic
matter releasing mineral acids. Second, these workers concluded from all
existing data on natural and artificial preparations that humic acids are
not definite chemical compounds but mixtures of plants and animal
residues, partly decomposed, partly conserved because of resistance to
decomposition, and combined in a colloidal state. The possibility that
pure organic compounds are present in the complex was not excluded but
such a compound as humic acid was believed to be entirely absent from
Baumann and Gully later reported the electrical conductivity
of humic acid to be much less than that of a 0.5 per cent solution of
acetic acid and took this as proof of a lack of free acids.
In 1907-1913, Schreiner and Shorey (37) reported what seemed to
be conclusive proof of the absence of single humic acid structure when
they reported isolating from the humic and the hymatomelanic acid
fractions of soil several pure organic compounds. Among these were
vanillin, salicylic aldehyde, 2-hydroxy- and dihydroxystearic acid,
trithiobenzaldehydes, and various resin acids. The amounts of materials
extracted (50 mg of dihydroxystearic acid from 25 Kg soil) were very
minute and for this reason failed to convince many workers in the field.
Interest in humic acid as a pure substance persisted as the
works of Maillard (38), Samuelly (39) and Gortner (40) postulated humic
acid as a complex of sugars and proteins.
Oden (41) showed by potentiometric titrations that organic acid
ions are present in soils and peats. He attributed the low pH of water
solutions or dispersions of these materials to a release of hydrogen ions
from an acid group rather than as a surface phenomenon. Oden first
reported that these acids were tribasic and later tetrabasic with an
equivalent weight of 339. As had other authors before him, Oden recog-
nized the tendency of these substances to readily yield colloidal dis-
persions in water.
Oden proposed the formula C60H52024(COO014 of molecular weight
1332. He suggested that hymatomelanic acid is a hydrolysis product of
humic acid and forms only during the extraction of humic acid itself.
He reported an equivalent weight of 250 for hymatomelanic acid and noted
that it has a higher carbon content. Because he obtained decreasing
nitrogen content in successive purification procedures he concluded that
the nitrogen in humic and hymatomelanic acids is an impurity. One of
the first authors to quantitatively determine functional groups in these
organic molecules was Eller (42), who concluded from his data on the
relative amounts of carbonyl oxygen in the preparation that the natural
products had little relation to the products obtained by the action of
mineral acids on sugars. Leopold (43) agreed with this theory but found
that, on oxidation of certain phenolic substances, products were obtained
which were similar in physical and chemical properties to the natural
humic acids. Leopold's artificial products prepared by the oxidation of
phenol, quinone, and hydroquinone in alkaline solutions contained 58.05 per
cent carbon of the formula x(C6H4 3)2 with a basic structure of,
lHO H H OOH
Ha'o H H H
The process of oxidation was outlined as follows:
hydroquinone -+ quinone
Phenol oxyquinone -- humic acid
pyrocatechein -+ oxyhydroquinone
His natural humic acid contained 59.6 to 60.2 per cent carbon,
3.2 per cent hydrogen and 1.7 to 2.0 per cent nitrogen, which was
considered as an impurity. The natural products were considered to
arise from the following reaction:
C6H1206 + 0 = C6H403 + 4 H20
The acidity of these materials was attributed to the phenolic hydroxyl
Fuchs (44) and Leopold (43) demonstrated that humic acids pre-
pared from peat contain both carboxyl and phenolic hydroxyl groups. The
carboxyl groups form salts and the hydroxyl groups are capable of com-
bining with bases. These authors prepared salts of humic acid by shaking
the acid with alcoholic potassium acetate. The base content of the salt
was found to correspond quantitatively to the methoxyl content of the
saturated methylated humic acids. The methoxyl content of the natural
humic acid was found to vary considerably with the source of the material.
Fuchs concluded that humic acid contains three to four hydroxyl groups,
three to four carboxyl groups, one methoxyl and one carbonyl group. The
hydroxyl groups were assumed to be subject to methylation. He considered
humic acids then as "oxy-carbonic acids originating in the decomposition
of dead organic material as dark amorphous substances capable of giving
off hydrogen ions and of forming salts and possessing base-exchange
capacity." The following formulae were suggested for the complexes
isolated from peat.
C49H52010 (OCH3)2 (COOH) (CH3CO) (OH)2
C3H46011 (OCH3) (COOH) (OE)4
C59H41017 (COOH)4 (OH) 3(CH2CO)
Fuchs regarded the presence of nitrogen in different humic acids
as due to the bonding of a protein molecule to a nitrogen free compound
or to the replacement of oxygen by ammonia in the cyclic compound.
Leopold (43) determined the carbonyl number of several humic
acid preparations by reacting them with phenylhydrazine hydrochloride.
Fehling's solution was employed to decompose excess phenylhydrazine,
the amount in excess being measured be determining the amount of evolved
N2. His results varied from 0.22 per cent to 2.31 per cent carbonyl
and were at best inconclusive.
Reports of equivalent weights and molecular weights for these
acids vary widely and depend on the source of the material. Molecular
weight data obtained by Samec and Pirkmaier (45) are shown in Table 2.
Molecular Weights of Some Humic and Hymatomelanic Acids
Acid Lignite Peat Brown Coal
Humic 1445 1235 1345 gms/mole
Hymatomelanic 855 761 739 gins/mole
Plunguian and Hibbert (46) were the first to consider the possi-
bility of an enol-keto tautomerism in the structure of natural humic acid,
and ascribed the differences in properties of the natural and artificial
preparations to the relative amounts of aromatic and aliphatic hydroxyl
groups in each. The natural material was found to have mainly aromatic
hydroxyl while the artificial had a predominance of aliphatic hydroxyl
Fuchs (47) studied the relation between humic acids and lignin
and found them to be very similar structurally by testing with various
organic spot reactions. He found the products of nitration to be par-
ticularly similar and concluded they are isonitrosoketohydroxy-carbonic
acids of high molecular weight. According to Fuchs, the difference
between the lignin and humic nitro derivatives is the manner of N-
linkage in each. Molecular weights (determined in (CH3)2CO) ranged from
1410 to 1465 gms/mole for methylated nitro derivatives of lignin and
1535 for the corresponding derivative of humic acid.
Fuchs used bromination as proof that a tetrahydrobenzene ring
is present in natural humic acid and Fuchs and Stengel (48) were success-
ful in identifying various organic acids and nitrophenols by oxidation
of humic acid preparations with HNO3.
Ludmila (49) has presented evidence that humic acids are cyclic
compounds to which are attached carboxyl, hydroxyl, and carboxyl groups
and have an average molecular weight of 1400 gms/mole.
Fromel (50) extracted humic acid from coal and peat with NaOH
and NaF and proved that the Freundlich adsorption isotherm is followed
in the medium concentration range and light absorption in the 2300 to
5000 Ao range follows Beer's law.
Ubaldini (51) determined acid numbers of humic acid by suspending
humus material in ethanol and titration with alcoholic KOH. Total acid
numbers were between 492 and 500 and carboxylic acid numbers between 263
and 267. Methylation with diazomethane showed not all acid groups are
methylated. Carboxyl, phenolic hydroxyl and carbonyl groups were found
in a ratio of 8:7:2.
Scheele and Steinke (52) reported molecular weights of humus
material as high as 9000, and found a slow decomposition takes place
in the presence of alkali giving a water soluble substance that cannot
be precipitated by mineral acid.
Galle and Lodzik (53) investigated the functional groups of
humus material by reacting model substances with MeOl containing HC1o
The acidic MeOH effected an esterification of carboxyl groups and an
esterification of phenolic groups. The authors treated phenylmethanol,
diphenylmethanol and triphenylmethanol in this manner and found a 90.3
per cent methylation. In model compounds containing both carboxyl and
phenolic hydroxyl groups, no distinction was made as to which group was
more completely methylated.
Later Scheele (54) used conductometric titrations to show that
the equivalent weight of natural humic acid is around 200. Molecular
weight determinations by diffusion constants were said to reveal that
the molecular weight of humic acid is twice as large in acid solution
as in basic solution.
Esh and Guha-Sircar (55) prepared humic acids in the following
manner. Soil was washed consecutively with hot alcohol-benzene, water,
2 per cent HCI, and 4 per cent KOH. This filtrate was then precipitated
with HC1. Humic acid prepared in this manner was oxidized with H202
(12 to 15 per cent ) in NaOH at 30-40. Extraction with ethyl acetate
gave acidic and phenolic components. Oxidation with chlorine dioxide
for 20 hours at room temperature yielded butyric and oxalic acids in small
Niklas and Genninger (56) also performed extensive studies on
the oxidation products of humic acids. They found ferric hydroxide
"of a given type and concentration" to act as catalysts in the oxidation
of humic acid with H202. No data were given by the authors but they
claim that a gas, a liquid and a solid were obtained in this manner.
The gas was described as a pungent, yellow acid which could be condensed
to a volatile liquid. A light brown wax was reported as being distilled
under pressure at 2500 from the residue.
Forsyth (57) subjected humic materials to ultimate and group
analyses and arrived at the conclusion that although various humic acids
have similar molecular structures, the number and kind of structural
groups depends entirely on the conditions and length of time of forma-
tion of the humus.
Evstigneev and Nikiforova (58) reported spectral data on hydroxyl-
methylfurfural obtained from a decomposition of glucose. This material
exhibits absorption maxima at 282.5 and 228-230 mu Humus substances
from sugars were found to show an initial absorption in the visible
region increasing rapidly in the ultraviolet.
Kukharenko (59) described a semimicro titrimetric determination
of the functional groups in humic preparations. This process involved
heating the humic acids at 1000C with calcium acetate. The number of
milligram equivalents of NaOH necessary to titrate the AcOH form from
the reaction from one gram of sample with calcium acetate equaled the
number of carboxyl groups in the humic molecule. No data were reported.
Hadzi (60) also reported spectral data on humic acids. These
authors found the infra-red absorption band at 3.25 microns (due to
phenolic hydroxyl) was shifted to 3.88 microns when the material was
deuterated (phenolic OD). In addition they found the carbonyl band at
1700 cm and stated that the absence of marked absorption bands in the
900 cm" indicates a lack of acid dimers and proves that the acidic
groups in humic acid are mainly phenolic.
Botteri (61) describes lengthy extractions of meladoin, humic
acid and lignin. This author concludes on the basis of aromatic sub-
stances obtained from alkaline fusion of these materials that there is
no similarity in their molecular structure.
Higuchi and Shibuya (62) obtained a reaction product with humic
acid in NaOH and the diazonium salt of sulfanilic acid. The purified
product was an orange-yellow dye of the following composition:
R(N:NC6H4SO3Na-p)7 where R equals
C106H85016 (CO2Na) (OH) 7 (CO) 4te.
Recently a good deal of attention has been focused on"the use of
model substances for determination of the structural elements in humic
Ploetz (63) produced 1,4-diquinonylbenzene by adding an aqueous
solution of the tetrazonium salt of p-C64(NH2)2 to a cooled aqueous
or aqueous alcohol solution of quinone. A coupling was noted by the
evolution of nitrogen gas. The author concludes the coupling product
is an intra-molecular quinhydrone complex, and this same structural type
is present in natural humus substance.
Welte et al. (64) used polymers of polyvalent phenols to give X-
Ray amorphous humic materials. Auto-oxidation reactions resulted in the
formation of oxalic acids, acetic acid and CO2 by ring cleavage.
Flaig (65) reported that orthoquinones are intermediate products
in the formation of humic acids. Substituted hydroxyquinones were used
as model substances in investigations on the preliminary stages and
decomposition products of humic acids.
The isolation and identification of organic compound in humus
is described by Wedgwood (66). Polycyclic hydrocarbons, pyrene, fluo-
anthrene, 1,2-benzanthracene, 3,4-benzopyrene, perylene and anthracene
have been detected.
Bremner (67) points out the need for isolative investigation in
soil analysis. Extraction of soil organic matter is usually done with
alkali, however some research has been done using mineral acid as the
extractant. Because, according to this author, acid or base may harm
organic matter extractions)new extraction procedures using neutral salts
were investigated. These results indicate soil organic humus is intimately
tied up with metallic cations. The efficiency of a neutral salt was found
to depend on its ability to remove cations as precipitates or soluble
complexes. The following order in decreasing efficiency was reported:
oxalate> malonate) salicylate> succinate> adipate)
maleate) trichloroacetate> phthalate.
The cations exhibited the following capabilities:
Li> Na) NH4) K) g.
The most effective neutral extractants were found to be those
capable of forming complexes with iron.
According to Bremner there are two distinct theories regarding
the origin of humic materials in soils. The first suggests that hunus
arises from an alteration of plant lignins entering the soil, while the
second suggests they are products synthesized by or formed by autolysis
of soil micro-organisms. Both of these theories have their proponents
among modern soil chemists.
According to Burges (68) all the available evidence suggests that
the humic acid fraction is either a single chemical substance or a group
of very similar substances. He believes humic acid is primarily non-
nitrogenous and nitrogen present in many preparations is due to a
secondary combination with protein or amino acids, the main structural
element being the quinone group.
Steelink et alZ (69) studied alkaline degradation products of
soil humic acid and isolated catechol (1), protocatechuic acid (II) and
resorcinol (III) in a mixture from a KOH fusion.
OH OH OH
I II III
Felbeck (70) detected about 60 components by gas chromatographic
separation of pyrolysis products of soil organic matter. The pyrolysis
was carried out at a reaction temperature of 3000C and in a H2 stream,
to reduce polymerization of phenolic materials. These pyrolysis products
were ether soluble and were further separated by extraction of the ether
with aqueous alkali. Roughly 20 per cent of the total organic matter
charged to the furnace was accounted for in the product mixture.
Steelink at al. (71) have reported the presence of at least two
free radical species in natural humic acid from data obtained in electron
paramagnetic resonance spectra. The authors believed one could be a
semiquinone radical and the other a quinhydrone-type radical.
III. EXPERIMENTAL PROCEDURES
Collection ang Concentration Techniques
The colored water samples secured for this study were collected
in five-gallon polyethylene bags or drums with the exception of the one
local water which was collected in five-gallon polyethylene-lined cans.
Delhez (72) has reported that water stored in polyethylene bottles for
several days may show absorption in the ultraviolet due to a continuous
release of organic matter from the polyethylene. Accordingly, distilled
water samples were stored in each type of container used in the work for
periods up to three months. No spurious absorption was noted in any
sample from 200 to 300 my)
The drums were shipped by truck to the State Boards of Health of
the states of Massachusetts, Connecticut, Virginia, North Carolina and
Wisconsin for sample collection and returned to this laboratory in the
same manner. Numerous samples from the states of Florida and Georgia
were collected by the investigator or members of the laboratory staff.
Because the material responsible for the coloration of water is
present in minute quantities in the stream, concentration of the organic
materials was necessary. Although recovery of the colored materials was
possible by several techniques, only two were used in this study, namely
vacuum distillation and freeze concentration. Gross solid matter was
removed from all water samples prior to concentration by either technique
by filtering through a suitable grade of filter paper.
All concentrated samples were swept with nitrogen and stored
in pyrex containers at 250C + 20 until use.
The distillation apparatus was a laboratory batch-type evapo-
rator manufactured by Precision Scientific Co., Chicago, Ill. The
vacuum source was an ordinary water aspirator which provided an internal
pressure of 79 im hg at 450C. The distillation was carried out between
45 and 500C to lessen any possibility of decomposition or combination
of chemical species by oxidation or other means. In a usual concen-
trating run, 40 liters of water was concentrated to a final volume of
1.0 liter in six hours of continuous operation.
Freeze concentration was carried out in a 16 ft3 Crosbey home
freezer, fitted with a 25-gallon polyethylene drum and an electrical
stirrer. In normal operation, 20 gallons of stream water was concen-
trated to a final volume of 1.0 liter in approximately 110 hours of
All ultraviolet absorption spectra were obtained on a Carey
Model 14 recording spectrophotometer from 200-800 my at a scanning
speed of 25 I/sec, using 1.0 cm quartz cells.
Infra-red absorption spectra were determined on a Perkin-Elmer
Model 137 Infracord. Solid samples were run by the potassium bromide
technique while liquids were determined on sodium chloride plates.
Fluorescence spectra were obtained with a fluorimeter assembled
at the University of Florida. It consisted of a converted Perkin-Elmer
Model 12 B monochromator fitted with a lithium fluoride prism and a
55 watt Oaram mercury lamp. The detector, an IP 21 photomultiplier
tube, was situated at an angle of 900 from the incident beam and was
connected to an El Dorado Model RH 200 photometer and a Varian Model
G-11-A recorder. A color filter that transmitted the 365, 405 and
436 m mercury lines was placed between the source and the sample.
All color values were determined on a Lumetron Model 450 filter
photometer, a product of Photovolt Corporation, New York, New York. The
instrument was fitted with a blue filter having maximum transmission at
420 ma, and provided a 15 cm light path. The reference standard for
color tests, according to the procedure recommended by the APHA for the
sanitary examination of water, is a solution of 1.246 gias potassium
chloroplatinate and 1.0 gm crystallized cobaltous chloride in 100 ml
concentrated hydrochloric acid, made up to 1.0 liter with distilled
water. This solution has an assigned color value of 500 color units.
A standard color curve was obtained by measuring the optical densities
of appropriate dilutions of the color standard on the Lumetron at
420 m It was necessary to correct the optical density values for
turbidity present in the color samples. The general procedure described
in the Lumetron manual (73) was employed.
All pH values were determined on a Beckman Model G pH meter with
glass and saturated calomel electrodes. The potentiometric titrations
in nonaqueous solvents were carried out with the same titrimeter equipped
with platinum and saturated calomel electrodes.
Light scattering data were obtained with a Series 1965 Brice-
Phoenix Light Scattering Photometer. A General Electric D. C. galvaw
nometer with a sensitivity of 0.0015 ""a/an was used with this instrument.
A Welsbach Type T-23 Laboratory Ozonator was used to supply the
gaseous ozone required in the oxidation experiments. Operated at 7 psig
and 100 VAC, this instrument produced 80 zg/1 03 in an oxygen stream at
a flow rate of 0.03 S.C.F.M.
Electrophoretic mobility data were obtained with a Briggs
microelectrophoresis cell using the technique and accessory apparatus
described by Pilipovich et al. (74) and Black and Smith (75).
The methylating agent used throughout this study was diazomethane
and was prepared in the following manner.
A 125 ml distilling flask was fitted with a condenser and a long
stem dropping funnel. To the flask was added 6 gms of potassium hydroxide
dissolved in 10-ml of water, 35 ml of methanol, 10 ml of sodium-dried
ether and a teflon stirring bar. A solution of 21.5 gis (0.1 mole) of
p-tolylsulfonyl-methylnitrosamide (diazalid) dissolved in 125 ml of ether
was added to the dropping funnel. The distilling flask was stirred in
a water bath at 70-750C, while the contents of the funnel was added at
such a rate that the 125 ml volume was delivered in 15 to 20 minutes.
The distillation was continued until the distillate was colorless. The
usual yield was 2.7 to 2.9 gns of diazomethane.
The diazomethane was standardized by reacting an aliquot with a
weighed excess of pure benzoic acid in absolute ether. The excess
benzoic acid was titrated with standardized sodium hydroxide to a phenol-
phthalein endpoint with rapid stirring.
The procedure for the quantitative methylation of the fulvic
acids and model compounds was as follows: A weighed sample was placed
in an excess of standardized diasomethane solution and allowed to stand
in the cold for 48 hours, or for such time as was necessary to achieve
maximum methylation. At this time, a weighed excess of benzoic acid
was added to react with the remaining diazomethane and the excess
titrated with standard aqueous alkali. From these data it was possible
to calculate the moles of CH2N2 consumed per mole of sample, or, in the
case of the fulvic acids, per unit weight of sample. The latter value
is hereafter referred to as n. That is
S=, moles CH2N2 reacted
100 gms fulvic acid
and if it is assumed that both the carboxylic acid and phenolic hydroxyl
groups are subject to methylation, and that these same groups constitute
the titratable acidity,
= ,nga of acid groups (-COOH and -QOH)
100 gis fulvic acid
IV, EXPERIMENTAL RESULTS
Characteristics of Ea Waters
As mentioned in Chapter III, all water samples were passed
through a roughing filter before concentration. Several qualitative
filters were tried and no measurable effect on color or pH of the
samples was noted. Centrifugation at 30,000 rpm also produced no loss
in color or change in pH. Prolonged storage of colored water has only a
slight effect on the color value, as indicated by the following experi-
ment with water A. Approximately 2.5 liters was swept with nitrogen and
stored in a cold room. The data in Table 3 show only a slight loss in
color over a six-month period.
Effect of Storage on Color in a Nitrogen Atmosphere
Storage time O.D.b O.D.r
(weeks) (x 10) (x 10) Color
0 6.4 0.75 204
1 6.0 0.70 188
4 6.0 0.70 188
8 6.1 0.70 192
16 6.0 0.71 188
24 6.1 0.68 188
In the presence of oxygen, however, the color was not quite as
stable. Approximately the same volume of water A was swept with oxygen
and stored in the same manner, with the results as listed in Table 4,
Effect of Storage on Color in an Oxygen Atmosphere
Storage time O,D.i O.D.
(weeks) (x 10) (x 10c Color
0 6.4 0.75 208
1 6.1 0.74 200
2 6.0 0.74 185
4 5.8 0.72 172
8 5.7 0.73 170
All samples used in this study were analyzed within three days
of collection and concentrated within one week.
In general, 25 gallons of each water was obtained. The source
of each water and the designations used to describe them throughout this
report are listed in Table 5.
A partial mineral analysis was performed on each colored water.
These data are presented in Tables 6 and 7.
It is evident that the majority of these colored waters were of
low mineral content with the exception of the two waters from the Florida
Everglades. Several other facts are evident from these analyses. First,
Sources and Designations of Colored Waters
A Creek near Newnan's Lake, Gainesville, Florida
B *Suwannee River at Fargo, Georgia
C Rice Creek near Palatka, Florida
D Lumber Creek near Raleigh, North Carolina
E Juniper Creek near Raleigh, North.Carolina
F Florida Everglades about 20 miles NW of Miami
G Drainage canal near Belle Glade, Florida
H Emerson Brook near Tewksbury, Massachusetts
I Great Dismal Swamp near Norfolk, Virginia
J Unidentified stream near Hartford, Connecticut
*River rises in the Okefenokee Swamp and samples collected in
the swamp near the source.
all of the colored waters contained an appreciable iron content.
Attempts were made at the beginning of this work to obtain a water free
of this metal. Eight Florida waters and one Georgia water, all from
different sources and types of drainage areas, were secured in this en-
deavor and all were found to possess substantial iron contents as seen
in Table 8. In addition, Nordell (76) lists the results shown in Table 9.
Second, there is no relation between iron content and color
value for waters from different sources, although for a given water this
might be expected with seasonal variations.
Third, the chemical oxygen-demand values do not vary linearly
with color values in waters from different sources, although this
relation might hold for a given water. Fourth, biochemical oxygen-
demand values are extremely low in all cases. If the organic substances
responsible for color in water are products of microbiological decom-
position, they are apparently in their final state.
Finally, the amount of dissolved matter in a water sample may
often be estimated by multiplying the specific conductance by an
empirical factor, that varies from 0.55 to 0.90. The data in Tables 6
and 7 show that for soft waters containing organic color this factor is
appreciably higher, varying from 1.47 to 2.21.
Effect of pH on Color
One of the most interesting chemical properties of natural colored
water is the variation of color intensity with pH. This "indicator"
action has been noted by many authors, but again little data on the extent
of this effect has been presented in the literature.
Each colored water sample used in this study was diluted 1:1 with
a Clark and Lub's buffer solution of desired pH value, and the resulting
color determined on a Lumetron. These data for waters A, B, D, E and G
are presented in Figures 1 through 3, and Table 10. In every water used
in this study the variation of color with pH was reversible, but not
linear. In most cases, the slope of the crye in acid solution was con-
siderably less than in basic solution. The data in Table 10 show that
the per cent increase in color value over the pH range employed was quite
variable. However, the greatest increase was obtained in waters with
Chemical Analyses of Waters A Through E
A B C D E
ppm as CaC03 9.6 14 68 40 30
Ca ppm as CaCO3 -- 8.0 40 2.5 2.0
g1 ppm as CaCO3 --- 6.0 28 1.5 1.0
ppm as CaCO3 10 0.0 45 2.0 1.3
Chloride ppm 10 16 116 3.0 3.0
Total iron ppm 0.58 1.0 0.31 1.5 0.40
pH 6.75 4.26 7.28 5.85 5.50
Color 240 352 156 108 68
Turbidity ppm trace 12 trace trace trace
25C 44.0 42.5 256 22.0 17.0
at 180oC ppm 76 94 250 40 25
COD ppm from
dichromate 142 83 46 96 28
5 days at
200C 1.24 -- 0.99 0.60 1.6
Chemical Analyses of Waters F Through J
F G H I J
ppm as CaCO3 295 505 30 16 8.0
Ca ppm as CaC03 250 365 24 10 6.0
* ppm as CaCO3 45 140 6.0 6.0 2.0
ppm as CaCO3 230 370 10 4.0 2,0
Chloride ppm 70 183 5.0 4.0 2.0
Total iron ppm 0.15 0.30 0.12 1.05 0.70
pH 7.85 7.95 6.90 5.10 5.00
Color 76 264 70 424 240
Turbidity trace 4.0 trace 10 4.0
250C 770 1490 88 61 23
at 1800C ppm 469 990 65 113 61
COD ppm from
dichromate 51 142 24 66 64
5 days at
200C 1.7 4.7 0.6 1.1 1.4
Variation of Iron Content in Colored
Water Source Fe (ppm) Color (ppm)
Suwanee River at Santa Fe 1.08 102
Suwanee River at Fargo, Ga. 0.78 316
Fort Pierce 0.50 182
Palm Beach Canal 2.60 88
U. S. 27-South Bay 1.39 65
Pumping Station-U. S. 27 0.36 76
Miami Canal at Bend 1.11 60
Taniami Trail Canal 0.97 45
Newnan Lake tributary 0.58 220
Variation of Iron Content in Colored Water Samples
Water Source Fe (ppm) Color (ppm)
Great Dismal Swamp
Elizabeth City, N. C. 1.8 1200
Burlington, N. C. 1.2 300
Waycross, Ga. 0.11 130
Shannuck, R. I. 0.20 70
Black water River
Burdette, Va. 0.17 80
Birmingham, N. J. 2.00 100
Variation of Color with pH
Color at Color at Per Cent Increase Over
Water pH 2.0 pH 10 Color at pH 2.0
A 160 270 67
B 237 412 74
C 102 182 78
D 71 130 83
E 50 82 64
F 54 82 52
G 190 290 53
H 42 86 105
I 330 440 30
J 150 265 77
lowest original color, and the smallest increase was obtained in waters
with the highest original color.
Because of this pH-color variation, all color values reported in
the course of this work were measured at a constant pH of 8.4.
Particle Size Estimation
Natural color in water has been reported as being in true solution
by some authors (20) and as existing mainly as a colloidal dispersion by
others (77). In all cases, the proposed theories have been backed by
little or no experimental evidence on particle size.
Early in this study it was observed that the smallest com-
mercially available maebrane filter, a 10 my millipore filter, would
retain only 13 per cent of the original color value of water A. Accord-
ingly, two electrodialysis cells were built. They were constructed
from plexiglass, fitted with a copper gauze cathode and a graphite block
anode. These cells were essentially rectangular plastic boxes cut
apart at approximately one-third and two-thirds of their total length
for the insertion of membranes and held together with steel bolts
through plastic cross-braces at either end. The only difference in the
two cells was their capacities, namely one and three liters.
In operation the cell was wired directly to a 125 volt direct
current line in the laboratory. A millimeter was connected in series
with the anode and a voltmeter was tapped across the electrodes. The
electrode compartments were filled with distilled water and changed
The maximum amount of color that would dialyze through a fairly
coarse parchment paper membrane was found to be 82 per cent of the
original color after 16 hours of continuous operation. Unfortunately,
the pore size of this coarse parchment could not be determined. However,
when water A was dialyzed through a cellophane membrane with a pore
size of 4.8 mg, 87.5 per cent of the color was retained after 24 hours.
Collodion membranes held 91 per cent of the color in the same length of
time. These data are summarized in Table 11.
These data indicate that most of the color in this water is
colloidal in nature, and while there is a distribution of particle sizes,
most of the particles are between 3.5 and 10 m, in average diameter.
Electrodialysis and Millipore Filtration of Water A
Maximum Per Cent
Material Pore Size Color Retained
Millipore filter 10 mp. 13.0
Cellophane membrane 4.8 m& 87.5
Collodian membrane 3.5 mul 91.0
All of the waters used in this study were vacuum filtered
through millipore filters of graded pore size down to 10 m,. In
addition, each water was dialyzed for 24 hours using the 4.8 mA cello-
phane membrane. The effects of each on the residual color values of the
waters are presented in Figures 4 and 5.
It is apparent from these figures that a natural distribution
in particle size existed in all of the colored waters used in this study.
Whereas the per cent of original color retained by the 100 m/ filter
varied from 4 to 38 per cent, that retained by the 4.8 mpA membrane
varied only slightly and was approximately 90 per cent. In every
colored water the majority of the particles were between 4.8 and 10 msA
in average diameter. Furthermore, approximately 10 per cent of the color
of each water passed through a membrane with a pore size of 4.8 mny.
The light-scattering properties of turbidity suspensions have
been extensively investigated by Black and annah (78). These authors
characterized various natural and synthetic dispersions by measuring the
intensity of scattered light from 150 to 1370 per unit intensity of
transmitted light at 00. In a given suspension, the ratio of scattered
intensity at 150 per unit scattered intensity at 900 per unit volume
was defined as a scattering ratio. For Latex dispersions, the scatter-
ing ratio increased with increasing particle size as shown in Table 12.
Scattering Ratios of Latex Suspensions
Particle Size Scattering Ratio
365 mp. 136
264 m) 25
188 m)u 10
88 myA 8
The upper limit of this ratio for substances in true solution is
If organic color in water is particulate rather than dissolved,
a scattering pattern should be obtained with a sensitive light scatter-
ing photometer. Accordingly, water B was filtered consecutively through
a 100 m,) and a 10 my millipore filter, as this procedure was found to
produce an optimum turbidity-free water by Black and Hannah. Scattering
intensities of a sample prepared in this manner were measured from 15o
to 137 with light sources of 436 and 546 mA Fluorescence filters were
supplied with this instrument for each wavelength and were color filters
that transmitted no source radiation. The results are shown in Figures 6
and 7, and describe the variation of scattering intensity at any angle per
unit transmitted intensity at 00 with the angle 0. It is evident that
these materials show appreciable scattering properties at 546 mk and
that the scattering ratios are not those of dissolved materials.
Scattering at 436 m,& is more intense and relatively constant
from 400 to 1370 This is caused by fluorescence of the organic color
at this wavelength and is shown in Figure 11 to be very intense and
independent of the angle 9. No fluorescence was observed at 546 mg .
It is interesting to note the effect of pH on the scattering properties
of this water. As shown in Figure 6 raising the pH from 5.0 to 12.0
resulted in a decrease in the scattering ratio, whereas the scattered
intensity at any angle per unit transmitted intensity at 00 was higher.
These data indicate that raising the pH of this colored water effected
both a decrease in individual particle size and an increase in the number
of particles present. The fluorescence intensity was higher in alkaline
solution as had been determined previously.
Black and Willems (77) have shown that coagulation of organic
colored water proceeds by an electrokinetic mechanism. They state,
"When alum or ferric sulfate is added to a colored water whose
alkalinity, naturally present or added, is sufficient to produce a pH
value within the upper portion of the pH range of hydrolysis of the
respective salts, positively charged colloidal picelles .
are formed. These positively charged hydrosols neutralize the negative
charge on the colloidal particles of color resulting in electro-
Although the two colored waters used in their study were of
different total solids and alkalinity content, they were both local
waters and were obtained from areas with similar forest covers.
As a knowledge of the similarity of this behavior in waters
from widely different sources is desirable, jar test and electro-
phoretic mobility data were obtained on seven of the waters used in
this study. Water A was not run as it was one of the waters used by
Black and Willems and had not changed in quality since their deter-
minations. Waters F and G were not coagulated as normal treatment of
these waters would have been lime softening.
The remaining waters were coagulated with alum, adjusting the
pH when necessary with lime or HC1. Figures 8 and 9 show the results
for waters B and I.
The uniformity of coagulation behavior in these colored waters
is apparent from these figures. The zone of optimum color removal was
accompanied by a zone of charge reversal in every case.
In waters B, C, I, and J, both zones occurred within relatively
narrow pH ranges. Waters D, E, and H had considerably wider pH ranges
of optimum color removal. However, the alum dosages employed with
these waters were substantially higher than would be used in normal plant
practice for waters of such low color values.
It is interesting to note that the width of the pH range of
charge reversal for these waters was roughly proportional to the width
of the pH range of optimum color removal as is shown in Figure 10.
These data indicate that in this respect at least, organic
colored waters from various sources are remarkably similar.
The fact that optimum coagulation of colored waters occurs at
relatively low pH values has been ascribed entirely to the low pH
ranges of hydrolysis of the Al (III) and Fe (III) ions (77). It should
be mentioned here that light scattering data obtained in this study
have shown that the particle size of color particles is affected by pH.
In acid solution the color particles were found to be fewer in number
and larger in size than those in the same water in basic solution.
What effect this phenomenon has on coagulation is undetermined but it
should be considered if the exact mechanism of color coagulation is to
Shapiro used the fluorescence of organic color fractions to
locate bands after chromatographic separation on paper. However, the
fluorescense emission spectra of organic color have never been recorded.
Since it was observed in this work that the raw stream waters fluoresced
as brilliantly as some of the fractions isolated from them, the spectra
of both were determined. Only the raw water fluorescence spectra will
be presented here.
Figure 11 is the spectrum of the Osram mercury lamp source with
distilled water, while Figures 12 and 13 are the fluorescence spectra
obtained from waters B and F. It can be seen from these figures that
the emitted fluorescence has a maximum intensity at 490 my although
the bands are relatively broad and of appreciable intensity from 450 to
550 mu The spectra of the remaining waters varied only in the intensi-
ties of emission.
The different mineral content of the waters may account for the
fact that the relative intensities of fluorescence emission were not
proportional to the initial color values.
Adjusting the pH value of any of the colored water samples to
higher values would invariably increase the intensity of fluorescent
emission. This effect is shown for water B in Figure 14.
None of the colored waters investigated in this study showed
any absorption in the visible region and all showed only end absorption
in the ultraviolet region. Each water was diluted, if possible, with
pH 8.0 buffer to color values of 200, 100, 50, 30 and 10 color units as
determined on the Lumetron. These samples were scanned on the Cary
Model 14 from 200 to 600 my against the same concentration of pH 8.0
buffer. Two of these spectra are presented in Figures 15 and 16.
However, all were similar in two respects. First, each water showed
only end absorption and second, dilution of each water had no effect on
the nature of the absorption curve other than to decrease the absorbancy
value. A substantial variation was observed in the absorbancy values
of each water at 200, 300 and 350 mp as shown in Figures 17 through 19.
However, at 350 mA all of the waters with the exceptions of waters F
and G had essentially the same absorbancy value at all concentrations of
organic color. The two exceptional samples were both obtained from the
Florida Everglades, in a region of extremely uniform vegetative cover,
and were noticeably more yellow to the eye than the rest of the samples
used in this study. In addition, waters F and G exhibited very weak
fluorescence spectra, in comparison with the other colored waters.
The effect of pH on the ultraviolet spectra of waters B and C
is shown in Figures 20 and 21. Only two waters are shown here, as
identical results were obtained with every water. The absorbance value
in alkaline solution was always uniformly higher than that in acid
The facts that organic colored waters are invariably colored
yellow-orange to the eye yet show no absorption in the 450-550 mA
region, and that the general shape of the ultraviolet spectra is that
of a scattering rather than an absorption curve suggests that the
color in these waters is due to scattering rather than molecular absorp-
tion of light energy. The difference in absorbance values at the
wavelengths shown in Figures 17 through 19 are probably due to two
factors. First, a natural variation in particle size would account for
a variation in scattering intensity. Second, the color values for each
series were determined on a Lumetron filter photometer which detects
scattered, transmitted and fluorescent radiation. The Cary, however,
eliminates all light of different wavelength from the incident beam
with a second monochromator, and therefore would not detect fluorescent
radiation. Thus, the Cary spectrophotometer was not examining equal
"color" values as determined on the Lumetron.
The variation of absorbancy on the Cary, as well as the Lume-
tron, with pH of the colored water samples may also be explained by
changes in particle size and fluorescence intensity with pH. As men-
tioned in a previous section, both of these effects have been observed.
These data are consistent with the theory that color in water
is of a colloidal nature.
Characteristics of Extracted Organic Material
Nature of the Fractions
Any fractionation scheme that is to be used in a method of
classification of colored waters should include all of the organic matter
present in these waters. The only fractionation method reported in the
literature has been presented by Shapiro (22). However, this author
reported that some organic matter was excluded by this extraction pro-
cedure, which was intended to extract one relatively pure acid fraction.
To determine if this amount was significant, water B was concentrated
and fractionated using Shapiro's techniques.
Shapiro's fractionation scheme was essentially as follows:
The concentrate obtained by vacuum distillation was dried in an oven
at 600C. The solid thus obtained was mixed with 80 per cent ethanol
acidified with HC1. This mixture was taken to dryness, dissolved in
dilute HC1 and extracted with various organic solvents, the most effi-
cient of which was ethyl acetate. Shapiro referred to material obtained
in this manner as the "free acid" fraction. Usual concentrations of
this fraction were reported as 4.16-5.24 mg/1 of raw water.
The solid and liquid materials resulting from the concentration
of 25 gallons of water B were combined and evaporated to dryness in an
oven at 450C over a 48-hour period. This solid residue was treated with
80 per cent ethanol acidified with HC1 to remove mineral materials. The
major portion of the residue dissolved leaving a small but definite
amount of finely dispersed solid (I). The alcohol solution was then
evaporated at 45C leaving a dark colored solid which was mixed with
dilute HC1 and extracted with ethyl acetate thereby separating it into
two more fractions; (II)', acid soluble but insoluble in ethyl acetate
and (II) soluble in ethyl acetate. Fraction (II)' can be further
separated yielding a water-soluble portion (III) and a water-insoluble
Thus four distinct fractions were obtained on the basis of
solubility by Shapiro's extraction scheme.
Tables 13 and 14 suamarize the solubilities and relative dis-
tribution of material in these fractions.
Solubility Relationships of
Fractions Obtained by Shapiro Technique
NaOH NaHCO3 EtOH
5 per 5 per 95 per
Fraction H20 cent cent cent Ethyl Acetate
I + + -
II + + + + +
III + + + +
+ soluble insoluble + partially soluble
Weight Distribution in Fractions Obtained by Shapiro Technique
Fraction Cent Weight (from 25 gal)
I 7 0.36 gma
"free acid" II 26 1.34 gns
III 8 0.41 gmn
IV 59 3.06 gms
Thus, the free acid fraction described by Shapiro is seen to con-
tain less than 30 per cent of the organic material present in this natural
All of these fractions dried to a brown to black solid, I and
III being noticeably lighter in color than II and IV. Fractions I,
III, and IV gave no melting points, i.e., they were stable up to 3600C.
However, fraction II softened at 90-1000C. This is a somewhat
narrower melting point range than that reported by Shapiro, but still
indicates a mixture of several constituents.
All of these materials were insoluble in non-polar organic
solvents such as benzene, petroleum ether and dioxane. I, II, and III
were only slightly acid soluble but were completely soluble in basic
solutions of 5 per cent NaOH and 5 per cent NaHCO3. Fraction IV was
only partly soluble in alkalis but was very soluble in dilute HC1.
The presence of enolic structures in all four fractions was
tested by reacting these materials with bromine water and potassium
iodide. All fractions gave a strongly positive test taking up bromine
immediately and yielding a blue color with starch.
A positive test for aromaticity was obtained on all fractions
with the Chloranil test. However, the Le Rosen test for aromaticity
was negative for each fraction. Comparative tests with pyrogallol and
anthraquinone gave the same results. In the presence of sodium car-
bonate, fraction II gave an Immediate red color with the diazonium salt
of sulfanilic acid, indicating a phenolic nature.
Aqueous titration of fraction II with NaOH yielded an equivalent
weight of 224 gms/eq. Shapiro reported an equivalent weight of 228 gms/eq
for the "free acid" fraction isolated from a Connecticut colored water.
The infra-red absorption spectrum of fraction II, the equivalent
of Shapiro's free acid fraction, is shown in Figure 22. This spectrum
was identical with one obtained in Shapiro's laboratory.
In general, the data obtained in this study on fraction II are
consistent with those reported by Shapiro for the free acid fraction.
As mentioned earlier in this report, many authors have pointed
out the vast variety of humus substances of varying elemental com-
positions that may be obtained from soil organic matter, the particular
compositions depending on the type of soil, manner of extractions and
strength of extractants. Since it is believed that these materials do
originate in the soil and in bottom deposits as well, it was decided to
try a manner of separation more similar to that outlined by Oden,
Waksman and Bremner which has since become the classical method of
fractionation of soil organic matter.
Whereas less than 30 per cent of the total organic matter
present was isolated by Shapiro's scheme, the classical separation of
humic material in water would, if applicable, separate the entire organic
content into generally recognized fractions of humus material as shown
in Figure 23.
The concentrate used in this part of the study was prepared by
both concentrating techniques so that a comparison could be made.
Many soils contain ether or chloroform-soluble fats and waxes,
the actual type and amount depending on the nature of the humus and its
abundance in the soil. Accordingly, one concentrate batch obtained by
freezing was extracted in the following manner. One hundred milliliters
of concentrate was shaken in a separatory funnel with 200 ml of sodium-
dried ether. The ether was drawn off, dried over Na2SO4 and allowed to
evaporate. The residue consisted of a greenish-yellow, odorless wax.
Two extractions of ether were sufficient to remove all the wax from this
volume of concentrate. Only 2.0 mg was obtained from 100 ml of concen-
The wax melted at 55-600C and gave no indication of decomposition
up to 130C. Attempts were made to identify this material in methanol
in a vapor phase chromatography apparatus. However, nothing was recorded
but solvent up to a retention time of 10 minutes.
Solid material was noted in the concentrate obtained by either
the freezing or evaporation methods although there seemed to be less in
freeze concentrates. This solid was at first included in the extraction
process since the entire concentrate was taken to dryness before
Since the first step of the new fractionation scheme required a
precipitation of the liquid concentrate with HC1, it was decided to
remove this solid. A 0.45/A millipore filter was sufficient for this
purpose and 3.5 gms of solid was filtered from one concentrate batch
before the HC1 treatment.
When the filtrate was treated with HC1, a much finer solid was
formed. These solids, namely that produced in the concentration process
and that produced by HC1 treatment, were labeled fractions la and Ib
Samples of each were subjected to an elemental analysis. The
results, on an ash free basis, are shown in Table 15.
Throughout the literature search for this work, all references
to the separation of the combined humic and hymatomelanic acids by the
addition of HC1 referred to the separation as a precipitation process.
The general appearance of both of these solids, and the manner in which
they were formed suggested that they were not true precipitates but
were formed from a colloidal fraction in the concentrate and were
actually flocculated particles. A solid obtained by adding HC1 to an
extract of soil obtained from the stream area could not be distinguished
from one obtained from the water.
Ultimate Analyses of Fractions Ia and Ib
Per Cent C Per Cent H Per Cent N
Ia 35.50 5.40 1.86
Ib 43.90 4.75 1.96
Fraction Ia probably resulted from a compression of the electrical
double layer during concentration of the negative charged colloidal
particles in the water to a point where its stability was overcome.
In order to avoid these physical or chemical changes during
concentration, attempts were made to concentrate smaller original volumes
which would still yield sufficient organic material for subsequent
analyses. The concentration of 20 liters of waters having a color of
200 or more to 1.0 liter produced no settleable solid phase. For
waters having a color of less than 200, a 40 to 1.0 concentration ratio
was used with the same result.
All of the colored waters used in this'study were concentrated
in this manner. The weight of organic material in each fraction obtained
from either 20 or 40 liters of raw water is listed in Table 16, while
the weight percentage of each fraction in the raw waters is shown in
Table 17 .
Organic Matter in Fractions
Volume Weight in Fraction, grams
Water Concentrated (1) I II III
Percentage of Total Organic Matter in Fractions
Fraction Per Cent
Water Fulvic Acid Hymatmcelanic Acid Humic Acid
A 87.0 11.2 1.8
B 87.5 11.6 0.9
C 87.7 10.3 2.0
D 89.6 9.6 0.8
E 88.8 11.1 0.1
F 87.9 11.3 0.8
G 85.0 14.3 0.7
H 89.5 8.4 2.1
I 86.4 12.9 0.7
J 82.8 16.6 0.6
These data show that organic materials in colored waters are
predominately of the fulvic acid type, and that the relative amounts
in each fraction were fairly constant in all the waters examined.
A general relationship was observed between the total organic
matter in each water and the color value of the water. As is shown
in Figure 24, a straight line results if the total organic content,
expressed here as ng/1 of raw water, is plotted against the respective
color values of the waters. Furthermore, samples of humic and hymato-
melanic acids obtained from water A were found to follow Beer's law as
shown in Figure 25. It is apparent from these data that on a unit
weight basis, the fulvic acid fraction is the most color producing, but
only slightly more so than the hymatomelanic acid fraction. Humic acid
produces measurably less color per unit weight than either of the other
Because of the predominance of the fulvic acid fraction in colored
waters, most of the attention during this investigation was directed
Evidence of Functional Groups and Structural Characteristics
The fulvic acid material dried to a brown solid which was slightly
hygroscopic and possessed a faint smell of caramel. The hymatomelanic
and humic acid residues, however, possessed no noticeable odor, were
non-hygroscopic and were of different color; the former drying to a light
brown, flaky solid, and the latter drying to a dense, shiny black solid.
All were stable in the solid form at temperatures up to 3600C, although
the fulvic and hymatomelanic acid materials darkened noticeably at
All fractions gave colored mixtures in 5 per cent NaOH and 5 per
cent NaHC03 leaving no solid material in the bottom of a test tube.
Fractions I and III were partially dissolved or dispersed in EtOAc, and
completely so in ethylenediamine, pyridine and dimethylformamide. All
of these materials were insoluble in non-polar organic solvents such as
benzene, petroleum ether and hexane.
All enols take up bromine instantly with intermediate formations
of dibromo-enols which form labile oc-bromoketones on elimination of HBr,
Such o(-bromoketones oxidize hydroiodic acid and liberate free iodine.
Fulvic acid samples from every water used in this study gave a positive
reaction to this test yielding a blue color with starch. In most cases,
the reaction and subsequent color development was much slower than the
model compounds aceto-acetic and benzoylacetic esters.
The same organic spot tests used with the Shapiro fractions were
employed to determine the presence of aromaticity to the fulvic acid
fractions of each water. The chloranil test was positive for the fulvic
acid fraction of every water except waters C, F and G, while the Le
Rosen test with sulfuric acid and formaldehyde was negative in all cases.
However, the strongest evidence of aromaticity in this fraction was
obtained using the phosphotungstic-phospho-molybdic acid reduction test.
Aromatic hydroxyl compounds reduce these acids with the production of a
characteristic blue color. A strongly positive reaction was observed for
every fulvic acid sample obtained in this study.
Infra-red spectra of the fulvic acid fraction of each colored
water were obtained and were very similar as shown in Figures 26 and
27 and Table 18.
Carboxylic acids typically show broad absorption bands with a
series of minor peaks or satellite bands over the range of 3.33 to 4.0u .
Infra-Red Absorption Bands of Fulvic Acid Fractions
Band location, ,
This is a result of the relative proximity of the associated OH and
normal C-H stretching frequencies. These bands were observed an the
spectra of all fulvic acid samples.
The two strongest absorption bands on every spectrum were
located at 5.8 and 6.15)a The former is the carboxy group stretching
frequency while the latter is the characteristic stretching frequency
of doubly bound carbon atoms in conjugated phenyl groups.
Absorption in the range of 6.9 and 7.3AU in all samples was
probably due to alkane groups. The C-H deformation frequencies of both
the -CH2- and CH3-C- groups are in this region as are the C-H stretching
frequencies of the cis and trans isomers of RlCH=CHR2.
The very broad absorption regions between 8.0 and 10.0/u is
probably due to a combination of the C-C and C-0 stretching frequencies.
The fact that this region is very broad and lacks any more definite
absorption bands indicates that these materials are sructurally complex
and/or contain several carboxylic acids.
Figure 28 is the infra-red spectrum of the methyl ester of the
fulvic acid from water A. Diazomethane was used as the methylating
agent. The presence of the band at 2,9,u indicates that not all of the
aromatic hydroxyl groups were methylated. The increased absorption at
3.4)u must be attributed to the increase in methyl groups. The absorp-
tion at 6.9 is stronger than in the unmethylated fulvic acid while
that at 7.3)A is weaker. In addition several bands were obtained,
namely at 7.95, 9.15, 9.80 and 12.5 u respectively. The band at 12.5)1
may well be the only spectral information obtained to date on the nature
of the substitution of the aromatic nuclei in these molecules. The C-H
out-of-plane deformation frequencies of the three adjacent free hydrogen
atoms in 1,2,3-trisubstituted benzene characteristically occurs at
The infra-red data thus confirm the assumption that these
materials are hydroxy carboxylic acids, and give strong indications of
unsaturation and aromaticity. The negative charge on particles of
organic color may arise from ionization of the carboql and aromatic
Fifty milligram samples of the three fractions of water A were
sent to Galbraith Laboratories, Inc., for ultimate analysis. The results,
on an ash free basis, are shown in Table 19.
Ultimate Analyses of Fractions of Water A
Fraction C H N O
I 41.50 5.72 1.98 50.80
II 29.30 5.94 1.85 62.91
III 49.29 5.11 1.24 44.36
A similar analysis of the fulvic acid fraction of water B showed
40.6 per cent C, 5.45 per cent H, 1.53 per cent N and 49.78 per cent 0.
The fulvic acids of these two waters are indeed very similar in this
respect. In addition, the fulvic acid fraction of water B showed a
methozyl content of 2.64 per cent.
Complexity of Fractions
Some knowledge of the complexity of these fractions was
obtained by the use of paper chromatography. Several solvent systems
were tried but the most useful had the following composition by
volume: pyridine, 50 ml; water, 50 ml; 2-methyl-2-butanol, 50 ml;
(C2H5)2NH, 2 ml.
An excellent elution solvent was composed of methyl ethyl
ketone, formic acid and water of the following volume composition:
methyl ethyl ketone, 37.5 ml; water, 15 ml; formic acid, 7.5 ml.
The following tables summarize the results obtained by chro-
matographing the three fractions of water A on Whatman paper number 3
by the descending technique. The bands were located by irradiation
with ultraviolet light.
Paper Chromatographic Separation of Fulvic Acid Fraction of Water A
Band No. Rf Range Description
1 0.0 dark brown (origin)
2 0.10-0.16 dark yellow (major)
3 0.36-0.38 yellow (major)
4 0.39-0.63 yellow smear
5 0.67-0.80 yellow mear
Paper Chromatographic Separation of Humic Acid Fraction of Water A
Band No. Rf Range Description
1 0.0 brown (origin)
2 0.40-0.50 yellow (major)
3 0.44-0.61 yellow
4 0.93-0.98 yellow
Paper Chromatographic Separation of bymatomelanic Acid
Fraction of Water A
Band No. RE.Rage Description
1 0.0 brown (origin)
2 0.08-0.11 yellow
Several of the bands obtained by chromatographing the fulvic
acid fraction of water A were eluted for infra-red analysis. No signi-
ficant difference in the absorption spectra of any band from that of the
unseparated fulvic acid was noted.
Titration Non-Aqueous Solvents
The acid strength of a phenolic hydroayl group is influenced by
the nature and position of any other substituents which may be present.
Phenol itself, and the naphthols and their homologues are only weakly
acidic in aqueous solution. Similarly, the fulvic acids isolated from
colored waters in this study, which have been shown to contain both
carbowylic acid groups and phenolic hydroxyl groups, exhibited only a
slightly acidic nature in aqueous solution.
In order to increase the acidity of these groups and obtain
more well-defined equivalence points, it was necessary to employ basic
organic solvents. Dimethylformamide, butylamine and ethylenediamine
were tried but the most useful was ethylenediamine. As the solutions
of fulvic acid in this solvent were intensely colored, it was necessary
to titrate them potentiometrically. A platinum-saturated calomel
electrode combination was found to be very suitable when sodium amino-
ethoxide was used as the titrant. Standardization of the sodium amino-
ethoxide was accomplished with benzoic acid. A typical potentiometric
titration curve of this acid is shown in Figure 29.
All the solvents were redistilled before use, as even traces of
water were found to significantly decrease the magnitude of the potential
change at the equivalence point.
Several model compounds were titrated to evaluate the magnitude
of this break for the neutralization of carbonylic and phenolic acid
groups. Figure 30 shows the curve obtained by titrating 53 ag of phenol
with 0.148 N sodium aminoethoxide. A sharp break of 125 millivolts was
obtained and calculation revealed an experimental equivalent weight of
94.0 compared with the theoretical value of 94.11. Two gradual breaks
were observed for resorcinol as shown in Figure 31.
This procedure will distinguish between carboxyl and phenolic
acidity in the same molecule as shown by the ortho and parahydro~ybenzoic
acid curves in Figures 32 and 33. These data show that the o-hydroxy
acid is the stronger acid of the two, but that the phenolic hydrogen
on the p-hydroxy acid is more acidic than that on the o-hydroxy acid.
This difference in acidity may be attributed to hydrogen bonding in
and subsequent stabilization of the salicylate anion.
The fulvic acid fractions of waters A through J were titrated
in this manner and the calculated equivalent weights are shown in
Table 23. The shapes of the titration curves were very similar and each
contained one well-defined equivalence point, as shown in Figures 34 and
35, for the fulvic acid fractions of waters B and I. These data in-
dicate that if different types of acid groups exist in these materials,
they are essentially of the same acidity. The other titration curves
were similar in appearance although the equivalent weights calculated
from them varied significantly. This apparent natural difference in the
number of acid groups per mole was also shown from quantitative methyla-
Methylation of several model compounds using the procedure
described in the chapter on Experimental Procedures gave the results
listed in Table 24. The values listed are the maxima that could be
obtained with this methylating agent. It is apparent from these data
that diazomethane will not completely methylate all of the phenolic
hydroxyl groups in these compounds. If the structural relationships
in the fulvic acid molecules are at all similar to the structures of
these model compounds, the validity of the assumptions listed above is
doubtful. Nevertheless, the former definition of the n value is valid
and should be an approximation of the number of acid groups in these ful-
vic acid fractions per unit weight. Table 23 lists the R values
Equivalent Weight and Quantitative Mathylation Data of
Fulvic Acid Fractions
Quantitative Methylation of Model Compounds
Moles CH2N2 Reacted
Compound Theoretical Experimental
hydrocinnamic acid 1.00 1.01
acid 2.00 2.05
acid 2.00 1.80
hydroquinone 2.00 1.55
resorcinol 2.00 1.62
obtained for the fulvic acid fraction of waters A through J,
the equivalent weights of each as determined by titration in
It is interesting to note that the lowest equivalent weight was
obtained for the fulvic acid fraction of water C, and that this same
fraction showed the largest A value. In general, there was an inverse
relationship between the equivalent weight and n values, although this
relationship was not linear. The fact that the n values are not exactly
1.00 per unit equivalent weight indicates, for the 5 values less than
1.0, that some titratable groups are not methylatable, and for n values
greater than one, some methylatable groups are not titratable.
Effects of Gaseous Oxidants
The most common method of color removal consists of coagulation
with alum or ferric salts. However, several gaseous oxidants have been
employed in plant practice to reduce organic color in water. The most
extensively used of these is chlorine.
For over fifteen years, the city of Miami, Florida, has been
treating a hard, colored water by lime softening followed by heavy
chlorination. The softening process itself reduces the color value of
the water from approximately 80 to a color value of 25 to 30. A chlorine
dosage of 12 ppm reduces the remaining color to less than 10. Oxidation
with ozone was tried on a pilot plant scale for a period of six months
with excellent results. However, the power cost for the production of
the ozone was found to be prohibitive. Although ozone is rarely used
for water treatment in this country, over 200 municipal plants throughout
Europe employ ozone for water disinfection, and for the control of tastes
Chlorine dioxide is used extensively in water treatment for
taste and odor control, and in the paper manufacturing industry for the
bleaching of wood pulp. This chemical, although more effective than
chlorine as a disinfectant at high pH values, is generally too expensive
for routine water treatment problems.
The effects of chlorine, chlorine dioxide and ozone on the color
values of the waters used in this study were evaluated.
The gases were bubbled through duplicate 200 ml samples of each
colored water concentrate until a constant value was obtained. The
chemical oxygen demand of one sample was then determined according to the
procedures listed in Standard Methods, and the residual color value was
measured on the other sample after it has been appropriately diluted.
It was found that color values determined in this manner agreed with
those resulting from oxidation of the raw water directly. In the cases
of oxidation with chlorine and chlorine dioxide, it was necessary to
destroy excesses of these chemicals with crystals of sodium thiosulfate
before the color values could be determined on the Lumetron. The
results are listed in Table 25.
It is apparent from these data that ozone was more effective
than chlorine dioxide or chlorine for removing organic color, and
chlorine dioxide was more effective than chlorine. Both chlorine and
chlorine dioxide produced their maximmn effects within a contact time of
5 to 10 minutes, whereas ozone generally required anywhere from 20
minutes to 1.0 hour. This is probably due to the relatively low concen-
trations of ozone employed.
In actual practice where these oxidants have been used for color
removal, it has always been assumed that color removal corresponded to
oxidation of the organic matter to carbon dioxide and water. That this
is not true is shown by the results listed in Table 26. These data show
that even in the absence of a color value the waters still exerted a
substantial dichromate chemical oxygen demand. In addition, ether
extraction of each concentrate sample that had been oxidized with chlorine
yielded approximately 10 to 30 mg of a yellow solid which gave the infra-
red spectra shown in Figure 36 and Table 27. These spectra are very
similar and much more definitive than those of the fulvic acid samples.
Sodium fusions of each solid showed that no C-C1 bonds were present.
Effects of C12, C102 and 03
on Color Removal
Original Color after Oxidation
Water Color Cl2 C102 03
A 240 -- 8 0
B 352 98 14 0
C 156 72 10 0
D 108 30 3 0
E 68 25 0 0
H 70 15 15 3
I 424 22 18 3
J 240 5 0 0
Effects of C12, C102 and 0 on Chemical Oxygen Demand of
Colored Water Concentrates
COD of COD after Oxidation, ppm
Water Concentrate C12 C102 03
A 2800 -- 750 615
B 1394 1060 80 220
C 996 118 40 152
D 1316 408 120 129
E 554 0 0 82
H 960 780 751 794
I 1320 830 652 742
J 1280 616 514 810
I I I i I I
N N 4 rN N
I in .-l -4 M-
I I I I I I
0 8% 84
41 a a a
VU CM r4 4 L,-
|3 r*i l ri 4
o o a
C4n 0 0% 0 1 a
CS Q a.
0! Ci !
0 tn 00 o I i
U U 1 1 1 1 1
U2 U 8r5 0 0 CD c 0
0 0 L V!
Sao 0G0 0 o
Sa S S a1 a S
Q 0 % o0s
S* N 0 5
S0 10 10 ID 10 10 '0
I C00 0 t0CO 00 00
cc c ci c c cc c
u 4 rm u r4 Ht, i
To determine if similar materials were present in a municipal
water that had been treated with chlorine for color removal, 25 gallons
of Miami tap water was concentrated to a final volume of 1.0 liter and
extracted with ether. The extraction yeilded 65 mg of a solid which was
similar in appearance to those previously obtained, and gave the Infra-
red spectrum shown in Figure 37.
The infra-red region from 11.1 to 16.7 .& is a region of strong,
aryl, out-of-plane C-H deformation vibration absorption which are
valuable for identifying substitution patterns in aromatic ring systems.
Generally, monosubstituted ring systems show absorption at 13.0 to 13.7u.
and 14,1 to 14.5A o-disubstituted at 13.0 to 13.6,u, m-disubstituted
at 12.4 to 13.3,A and 11.1 to 11.6A and p-disubstituted at 11.6 to
12.5.u The data in Table 27 show that with the exception of water A,
all spectra gave bands that are characteristic of p-disubstitution.
Since these materials have given every indication of being structurally
complex, it must be concluded that this p-disubstituted ring is only
a fraction of the whole structure and is perhaps an end group. All of
these solids darkened at 285-3100C without melting and gave positive
spot tests for aromatic hydroxyl groups.
Degradative Oxidation Studies
It is noteworthy that despite the many extensive investigations
concerned with the structural nature of soil organic matter, no pure
organic compound has ever been isolated that has been proven to exist in
the natural organic complex. In addition, the small number of pure com-
pounds that have been isolated have always represented yields of 10 per
cent or less.
Similarly, the organic materials extracted from the colored
waters used in this study were of such complexity as to defy identifi-
cation by normal chemical procedures. It was observed that oxidation
of fulvic acid with KMh04, Ce(SO4)2, C102 and H202 produced ether-
soluble oxidation products which gave infra-red spectra similar to the
unoxidized fulvic acids. In all cases, mild oxidation resulted in a
loss of fluorescence. However, some useful information was obtained
by degrading these molecules by peroxy-acid oxidation. These studies
were performed on the fulvic acid concentrates of waters A, C, I and J
as these were the most highly colored.
Periodic acid has a selective oxidizing action on l:2-glycols
and on ot-hydroxy aldehydes and ketones. Compounds of this type are
cleaved at the C-C bond between the adjacent hydroxyl or carbonyl groups
with the formation of aldehydes or acids of shorter chain length.
The procedure used for the oxidation of the fulvic acid con-
centrates with periodic acid was as follows. Two hundred milliliters
of colored water concentrate was adjusted to pH 1.0 with dilute sul-
furic acid and filtered to remove fractions II and III. This volume
was then mixed with a solution of 1.0 gm of sodium paraperiodate in
50 ml of approximately 7 N H2S04 in a 500 ml round bottom flask equipped
with a reflux condenser. The solution was then refluxed at atmospheric
pressure for 1.0 hour, cooled and extracted with ether.
In general, the extraction with ether left an aqueous phase
that was only slightly colored whereas the ether solution was invariably
a deep, orange-red color. Evaporation of the aqueous phases to dryness
at 400C under nitrogen and extraction with EtOH yielded solids which
gave infra-red spectra typical of the unoxidized fulvic acids. These
solids represented approximately 10 to 20 per cent of the total weight
of organic matter present after oxidation.
Chromatographic separation of the materials in the ether extracts
was possible on a 10-inch column packed with adsorptive alumina. In
general, the extracts gave one rather diffuse yellow band near the top
of the column and one relatively narrow yellow band approximately 4
to 6 inches down the column. Neither of these bands fluoresced under
Following elution with 95 per cent EtOH, the lower band dried
to a sweet smelling yellow oil while the upper diffuse bend dried to an
orange solid that decomposed in a Bunsen flame.
For water A, less than 15 ng of the oil was obtained, whereas
135 mg of material was obtained from the diffuse layer. The infra-red
spectra of these materials resulting from the oxidation of water A are
shown in Figures 38 and 39. The absorption data obtained with the yellow
oil indicate that it is not a carboxylic acid and does not contain
hydrosyl groups. The only absorption below 4,0 is due to the presence
of C-H. The shoulder at 3.25 may be due to aromatic hydrogen while
the strong peaks at 3.4 and 3.48 are the assymetric and symmetric C-H
stretching frequencies of the CHl-C group. The presence of the strong
carbonyl band at 5.80/M indicates that fulvic acids may contain these
groups outside of carboxylic acid configurations. The doublet at 6.30
and 6.40)1 is probably phenyl ring absorption. This doublet frequently
occurs in conjugated ring systems. The medium intensity bands at 6.90
to 7.30,A are probably due to C-H deformations of the -CH2- and CH -C
groups respectively. Absorption at 7.90/A is a result of C-C
deformations while that at 8.95pc is due to C-O stretches. The origins
of the bands at 9.40. 9.65 and 10.5 are uncertain while those at 13.5
and 14.25 are indicative of monosubstitution. Absorption in the 3.0-
4.0 A region in Figure 39 indicates that the solid materials obtained
from this oxidation are a mixture of carboxylic acids. The band assign-
ments for all bands up to 8.0 are the sane as for Figure 38. Absorption
at 9.3 and 9.60/p is probably due to -OH deformations. The large
band at 12.5)A may be a result of skeletal vibrations. Attempts at
further separation of these acids met with little success.
The data in Figures 38 and 39 suggest that fulvic acids contain
either oA, -dihydroxy, or an c -hydroxy ketone groupings of the type,
R C C -R
R C C R'
where both R and R' contain aromatic nuclei. The facts that the oily
substance showed the presence of a carbonyl group but was not an acid,
and that such small quantities of the oil were obtained, suggest that
the following groupings may also be present in fulvic acids, namely
OH OH OH OH
I I I I
R C C R"' R C C R"
I I I I
R' R" R' H
R C C- R
where R and R' contain monosubstituted phenyl rings.
The physical and chemical properties of ten naturally colored
waters have been investigated and have been found to be relatively in-
dependent of the source of the water.
Angular light scattering data and ultraviolet spectroscopy
have shown that the materials responsible for color in water exist pri-
marily in colloidal suspension in the water. It is suggested that the
color in these waters is due to light scattering rather than molecular
absorption of light energy. The pH value of the waters was found to
affect both the particle size and number of particles in a given sus-
pension. This effect was proposed as an explanation for the variation
of color value with pH.
The particle size of the organic matter in each water was
estimated at constant pH by dialysis and millipore filtration and was
found to be generally less than 10 ma.
The coagulation behaviors of these waters with aluminum sulfate
were compared and it was observed that the pH zone of optimum color
removal was accompanied by a zone of minimum zeta-potential and typically,
of charge reversal, of the coagulated color particles.
The fluorescence spectra of these colored waters were recorded
and revealed in all cases a maximan energy of emission at 490 ma with
the combined excitation wavelengths of 365, 404 and 436 myn.
Organic materials were isolated from these waters and were shown
to be similar to natural soil organic matter. The predominate fraction
was found to be fulvic acid whereas the combined humic and hymatomelanic
acid fractions accounted for less than 10 per cent of the total organic
matter present in each water. For all the waters examined a linear
relationship was observed between the amount of organic matter and the
color value of the water.
Data obtained from solubility relationships, chemical spot tests
and infra-red absorption spectra indicated that fulvic acids are aromatic
polyhydroxy methoxy carboxylic acids. Ionization of the carboxyl and
aromatic hydroxyl groups was offered as an explanation of the negative
charge of color particles.
Equivalent weights of the fulvic acids were determined by
potentiometric titration in ethylenediamine and were found to vary from
89 to 133 gms per equivalent among the waters studied. Quantitative
methylations of the fulvic acids with diazomethane revealed that there
were generally 0.40 methylatable groups present per 100 gas for fulvic
acids of high equivalent weight and 1.35 methylatable groups per 100 gas
for fulvic acids of low equivalent weight.
Ozone was found to be more effective than either chlorine or
chlorine dioxide for the removal of color from these waters. It was
shown that color removal did not correspond to oxidation of the organic
matter to carbon dioxide and water. Chlorine oxidation was found to
produce ether-soluble oxidation products. Infra-red spectra of the
substances suggested that they are polyhydroxy carboxylic acids con-
taining disubstituted phenyl rings. Similar materials were extracted
from a municipal water system that routinely uses chlorine for color
Degradative oxidation of the fulvic acids produced ether-
soluble mixtures of complex carboylic acids from which pure compounds
could not be isolated. However, infra-red spectra of ether-soluble
oxidation products of the fulvic acids with paraperiodic acid indicated
that these acids contain cc, @, dihydroxy, or oc-hydroxy ketone con-
A- water G
-- water A
0 2.0 4.0 6.0 8.0 10.0
Figure L Effect of pH on Color of Waters A and G
Figure 2. Effect of pH on Color of Water B
A- water D
-A- water E
2.0 4.0 6.0 8.0
Figure 3. Effect of pH on Color of Waters D and E
A water A
--- water B
-0-- water C
80 -0- water D
-&- water E
S 40 -
o I I I I I-l-I--I--I
10 20 30 40 50 60 70 80 90 100
Pore Size my
Figure 4. Dialysis and Millipore Filtration of Waters A
--- water F
-- water G
80 water H
--- water I
-A- water J
0 II I I I I
10 20 30 40 50 60 70 80 90 100
Pore Size my
Figure 5. Dialysis and Millipore Filtration of Waters F
A Scattering at pH 12.0
--- Scattering at pH 5.0
1.0 0 -
20 40 60 80 100 120
Figure 6. Light Scattering Properties of Water B at 546 m)m
Scattering at pH 12.0
-- Scattering at pH 5.0
1.0 Fluorescence at pH 12.0
-0- Fluorescence at pH 5.0
I I I I I I
0 20 40 60 80 100 120
Figure 7 Light Scatterng Properties of Water B at 36
Figure 7. Light Scattering Properties of Water B at 436 m)
Alum Dose 120 ppm
Effect of pH and Constant Alum Dosage on
Coagulation of Water B
ALUM DOSE 120 ppm
Effect of pH and Constant Alum Dosage on
Coagulation of Water I
S 1.0 O
0 I I I I I
0 0.5 1.0 1.5 2.0 2.5
pH Width of Charge Reversal
Figure 10. Variation of pH Width of Optimum Color Removal
with pH Width of CMarge Rversal
350 400 450
/, X ^
Figure 11. Spectrum of Mercury Lamp Source with
400 450 500 550
Figure 12. Fluorescence Spectrum of Water B
Figure 13. Fluorescence Spectrum of Water F
Figure 14. Effect of pH
500 550 600
on Fluorescence Emission
,f Water B
As 0.8 200
200 250 300 350 400 450
Figure 15. Ultraviolet Absorption of Water B as Function of
200 250 300 350 400 450
Figure 16. Ultraviolet Absorption of Water I as Function of
0 I I I 1
0 50 100 150 200
Figure 17. Absorption of Colored Waters at 220 mu
As 0.5 E, B, D, J
0 I I ----
0 50 100 150 200
Figure 18. Absorption of Colored Waters at 300 mny
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EPK0TV9US_KQBOR0 INGEST_TIME 2012-03-09T00:17:42Z PACKAGE AA00003977_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC