ROBERT A. TAFT WATER RESEARCH CENTER
REPORT NO. TWRC-4
OZONE TREATMENT OF SECONDARY
EFFLUENTS FROM
WASTEWATER TREATMENT PLANTS
ADVANCED WASTE TREATMENT RESEARCH LABORATORY IV
U.S. DEPARTMENT OF THE INTERIOR
FEDERAL WATER POLLUTION CONTROL ADMINISTRATION
OHIO BASIN REG/ON
Cincinnati, Ohio
-------
OZONE TREATMENT OF SECONDARY
EFFLUENTS FROM WASTE-WATER TREATMENT PLANTS
by
D. Th. A. Huibers, R. McNabney,
and A. Halfon
for
The Advanced Waste Treatment Research Laboratory
Robert A. Taft Water Research Center
This report is submitted in
fulfillment of Contract No.
14-12-114 between the Federal
Water Pollution Control Ad-
ministration and the Air Re-
duction Company, Inc.
U. S. Department of the Interior
Federal Water Pollution Control Administration
Cincinnati, Ohio
April 9, 1969
-------
FOREWORD
In its assigned function as the Nation's principal natural
resource agency, the United States Department of the Interior
bears a special obligation to ensure that our expendable re-
sources are conserved, that renewable resources are managed tc
produce optimum yields, and that all resources contribute their
full measure to the progress, prosperity and security of
America -- now and in the future.
This series of reports has been established to present the
results of intramural and contract research carried out under the
guidance of the technical staff of the FWPCA's Robert A. Taft
Water Research Center for the purpose of developing new or im-
proved wastewater treatment methods. Included is work conducted
under cooperative and contractual agreements with Federal, state,
and local agencies, research institutions, and industrial organi-
zations. The reports are published essentially as submitted by
the investigators. The ideas and conclusions presented are,
therefore, those of the investigators and not necessarily those
of the FWPCA.
Reports in this series will be distributed as supplies per-
mit. Requests should be sent to the Office of Information, Ohio
Basin Region, Federal Water Pollution Control Administration,
4676 Columbia Parkway, Cincinnati, Ohio 45226.
-ii-
-------
TABLE OF CONTENTS
Foreword ii
Table of Contents iii
Abstract iv
Summary and Recommendations v
Introduction 1
Experimental Work 3
Analytical Procedures 4
Secondary Effluent Composition 6
Ozonizer 6
Solubility and Decomposition of Ozone in Water 6
Clarification of Secondary Effluents 7
Ozonation in a Continuous Countercurrent Column 12
Equipment and Procedures 12
Results and Discussion 15
Batch Reactor Tests 22
Equipment and Procedures 22
Results and Discussion 24
Simulation of a Six-Stage Cocurrent Contacting System 37
Equipment and Procedures 37
Results and Discussion 39
Design Concepts and Criteria 44
Ozone Supply 44
Process Design 45
Plant Description 48
Capital Cost Estimate 53
Operating Costs 58
References 61
-111-
-------
ABSTRACT
Ozone effectively lowers the chemical oxygen demar.^l (COD)
and total organic carbon (TOG) content of effluents frcrn waste-
water treatment plants. It removes odors and color from writer
and destroys pathogenic organisms. Residual ozone decomposes
rather rapidly; it has a half life in drinking water of about
20 minutes. Tertiary treatment with ozone has the potential of
an automated, trouble-free operation with low maintenance. The
objectives of this research work were to devise an efficient
contacting process and to make a preliminary evaluation of its
economics.
Ozonation was first studied by low-shear countercurrent
contacting in an 18-ft. column packed with Raschig rings. Later
it was found that the rates of COD and TOC removal could be con-
siderably increased by high-shear contacting using a turbine
agitator. Kinetic data, obtained with batch reactor experiments,
indicated that the optimum treatment of secondary effluent with
ozone required multi-stage cocurrent contacting. A six-stage
system was simulated by mixing effluent with ozone in an injec-
tor and recycling the mixture through the 18-ft. packed tower.
Overall ozone-utilization efficiencies as high as 90% were ob-
tained.
The estimated treatment cost is 7.7^/1000 gal. to reduce
the COD from 35 mg/1 to 15 mg/1 assuming an 80% ozone-utiliza-
tion efficiency at a 10-million GPD plant. This is about equal
to the estimated cost for treatment with activated carbon.
Unless the waste-water treatment plant is particularly
effective, chemical clarification of the secondary effluent
prior to treatment with ozone may be advantageous since that
would reduce ozone consumption and result in a lower final COD
and TOC.
-iv-
-------
SUMMARY AND RECOMMENDATIONS
The use of ozone for tertiary treatment of secondary
effluent has proven to be economically effective. Treatment
of actual effluents from the Totowa, New Jersey, and Berkeley
Heights, New Jersey, trickling-filter, waste-treatment plants
by ozone resulted in a product that met PHS requirements for
potable water. Virtually all color, odor and turbidity were
removed; and oxygen-consuming organic materials, measured as
COD, were reduced to acceptable levels (below 15 mg/1). Bacte-
riological tests showed that no live organisms remained; and
surface-active detergents, which can cause foaming, were removed.
Ozone concentrations from 11 mg/1 to 48 mg/1 in oxygen proved
equally effective.
The experimental work consisted of three main parts:
1. Ozonation in a continuous countercurrent column,
2. Batch reactor tests,
3. Simulation of six-stage cocurrent contacting.
Initially, a continuous, countercurrent, reaction system was
investigated using an 18-foot glass column, 4 inches in diameter.
Batch reactor tests were used to investigate the mass transfer
effects on the rates of COD and TOC removal with ozone. These
tests showed that the rates of COD and TOC removal were very de-
pendent on agitation rates. This dependency results from the
need for high mass-transfer rates for ozone from the gas phase
to the liquid phase. The instability of ozone, coupled with low
chemical-reaction riites, led to the development of a cocurrent
multistage contacting system with the ozone supply, per stage, pro-
portional to the demand as determined by the COD content. Contact
between the incoming streams was effected by a high-shear injector,
A comparison of the performance of this improved system with that
of the countercurrent column is summarized in Table I.
Pretreatment of the secondary effluent by coagulation, sedi-
mentation and filtration (i.e., chemical clarification) resulted
in appreciable reductions of ozone usage for the particular
effluents studied. The coagulant used (either lime, alum, or
alum + acid) had little effect on the efficiency of clarification,
but its control over effluent pH had a marked influence on the
ozonation operation. In general, a low pH for the clarified
effluent resulted in lower reaction rates but in higher ozone-
utilization efficiencies.
-v-
-------
TABLE I
SUMMARY OF CONTINUOUS OZONATION OF SECONDARY EFFLUENT
Effluent
Origin
Totowa, N.J.
Berkeley
Heights,
N.J.
Reactor
Type
Counter-
current
column
Six-stage
cocurrent
Coagu-
lant
None
None
Alum
Alum
+Acid
Alum
+Acid
O-j Concen-
tration in
Gas Feed,
mg/1
43
21-45
23
23
11
COD, mg/1
Raw
Effluent
82
112
89
96
91
After
Filter
60
67
26
41
36
Ozone
Treated
13
24
16
15
17
% COD
Removed
by Ozone
78
64
38
63
53
Ozone Con-
sumption ,
Lbs/1000
gal.
1.6
1-.6
1.4
0.74
0.53
% of
Initial
Ozone
Feed in
Exit Gas
21
21
38
24
40
-------
Economic evaluations of ozone treatment of secondary
effluents indicate that ozone treatment is competitive with
activated carbon systems. Operating costs are highly depen-
dent on plant size and the quality of effluent received.
Effluents having high organic content are most economically
treated by a combination of chemical clarification and ozona-
tion. Effluents with a low organic content require only
ozonation. The need for clarification would roughly double
the plant investment and operating costs. Operating costs
are 7.7jzf/1000 gal. for a 10-million GPD ozone treatment plant
without a clarification step, and 15.8^/1000 gal. if clarifi-
cation is required. The capital investment is $1,080,000
without clarification, and $2,270,000 with clarification.
Activated-carbon treatment cost for a 10-million GPD plant.
without pretreatment has been estimated at 8.3jzf/1000 gal.
Capital investment for a plant of that capacity is $1,670,000.
Pretreatment requirements for the activated carbon and the
ozone treatment plants appear to be similar.
There would be a good potential for cost reduction in the
generation of ozone if ozone treatment were to find application
on a large scale. A 20% reduction in ozone generation cost would
result in a 10% reduction in treatment cost and investment. Since
the present efficiency of commercial ozone generators is only 10%,
there is considerable potential for such reductions in ozone
generation costs.
Because of the promising results in the pre-pilot stage,
this development should be continued with a larger-scale pilot
plant.
The function of the pilot plant would be to:
1. Establish the effect of scale-up on:
a) Material transfer;
b) Ozone-utilization efficiency;
o
c) Final TOG, COD, etc., related to initial values ;
d) Long-term effects of solid surfaces on ozone
decomposition and on operating problems;
e) Effects of pretreatment on efficiency and
on operating problems.
2. Conduct operations over a long enough time to evaluate
the effects of hourly, daily, and seasonal changes in
the composition of the effluent.
3. Establish the economics of the process more firmly by
optimizing the number of stages, residence time, and
pretreatment.
-vii-
-------
4. Obtain data for the design of full-scale plants.
5. Develop specific items of equipment, such as ozone
dispersers, etc.
6. Develop methods of control and evaluation.
7. Develop oxygen-recycle systems.
8. Determine the effects of the treated effluent on
the streams into which it is emptied.
-------
INTRODUCTION
The accelerated population growth in the United States
and the increasing concentration of the population in large
urban centers have created a new order of magnitude in the
problem of waste-water disposal. Existing methods of waste-
water disposal are, in many cases, so inadequate that natural
bodies of water into which waste water is discharged are so
polluted that they constitute health hazards. For example,
many rivers into which wastes are discharged also serve as
sources of drinking water for populations downstream from the
discharge point. If the assimilative capacity of these rivers
is overburdened, conversion of their waters for drinking pur-
poses becomes increasingly difficult.
Conventional, secondary, waste-treatment methods do not
remove all of the complex mixture of dissolved and suspended
organic matter. Flocculation and filtration of secondary
effluents, prior to tertiary treatment, is well advanced and
appears capable of giving a product of relatively low solids
and colloidal organic content. However, many pathogenic
organisms and some organics that cause odor and color will
pass through both of these treatments. Present and future
quality criteria for water reuse and waste-water discharge
will force the adoption of additional or improved treatment
methods.
Oxidative purification with ozone, as a tertiary treat-
ment for sewage, has a number of inherent advantages. Ozone's
high reactivity permits oxidation, in a continuous process,
of many compounds which resist biological oxidation. Ozone re-
moves odor and color. It also destroys pathogenic organisms.
After ozonation, chlorine demand is considerably reduced, thus
eliminating the tastes and odors produced by chlorination of
some residual organics in typical discharged effluents. This is
an important consideration in case of water reuse.
The principal disadvantage to the use of ozone for ter-
tiary treatment has been the expense. The high cost of ozone
generation requires a high ozone-utilization efficiency if
ozone treatment is to be economically competitive. The pur-
pose of this study was to develop experimentally a conceptual
design of a process for the treatment of secondary effluents
from municipal waste-treatment plants and to evaluate process
costs.
-1-
-------
The initial development plan included the following aspects:
1. Design, construction and operation of a
laboratory-scale (about 10 gal/hr) counter-
current column, with auxiliary equipment,
for treating secondary effluents with ozone.
2. Monitoring the performance of the ozonation by
measuring changes of COD (chemical oxygen
demand) and TOC (total organic carbon) of the
water, relative to the ozone consumption.
3. Demonstration of the performance of the ozona-
tion process with actual effluents from local
sewage-treatment plants.
4. Determination of the efficiency of various
tower modifications, packings, and gas-
dispersion means.
5. Investigation of the significant parameters of
the process, such as gas and liquid velocities,
contact time, pressure, pH, ozone concentration,
initial organics content of the effluent, and
the effect of various clarification pretreat-
ments.
6. Obtaining reaction rate information with respect
to changes in impurity level and in ozone con-
sumption, both for use in designing larger-scale
plants and for optimization studies.
7. Development of a full-scale process concept,
followed by a preliminary evaluation of its cost.
-2-
-------
EXPERIMENTAL WORK
The apparatus and tests were chosen to show the effec-
tiveness of ozone treatment in reducing the COD to about
15 mg/1. This is the COD level of a good-quality surface
water.
Ozone has been used in the past for drinking-water steri-
lization. However, Airco had also conducted preliminary tests
on ozone treatment of secondary effluent . The various experi-
mental ozone-treatment systems chosen for the present work
were designed to extend the available knowledge of existing
4
ozonation processes to eventual large-scale treatment of sec-
ondary sewage-plant effluent. Ey studying chemical reaction
rates and the effect of mass transfer of ozone from the gaseous
to the aqueous phase, a design concept was obtained for the
most economical treatment of secondary effluent with ozone.
Initially, a continuous, countercurrent, reaction system
was investigated using an 18-foot glass column, 4 inches in
diameter. The effluent, flowing down, was the continuous
phase. Bubbles of ozone in oxygen rose from a sparger at the
bottom. In order to prevent back mixing, the column was filled
with 3/8-inch Raschig rings. The experiments showed that the
ozone transfer rate of this type of contacting was not very
high.
As a result, we decided to make kinetic studies in a batch
reactor. Batch-reaction tests were made by dispersing a stream
of ozone into effluent contained in a glass round-bottom, flask
or in a cylindrical resin flask. The amount of oxidation, char-
acterized by COD and TOC reduction, was measured as a function
of time. Important variables were :
1. Type of clarification applied to the effluent
prior to ozonation.
2. Type of agitation: High-shear versus low-shear.
3. Ozone concentration in oxygen.
These measurements provided basic reaction-rate data for equip-
ment design. For example, we found that high shear roughly
doubles the rates of COD and TOC reduction.
-3-
-------
Clarification was included because the effluent contained
suspended matter which could easily be removed by coagulation,
followed by sand filtration. The clarification facilities
enabled us to evaluate the combination of clarification and
ozonation, compare the combination with ozonation alone, and
explore the effect of various coagulants on ozonation. Ozone
consumption was considerably reduced by prior clarification.
Clarification removes part of the TOC that is resistant to oxi-
dation by ozone, resulting in a lower final TOC level.
Because the reaction between dissolved ozone and many of
the organic compounds to be removed is slow, and because ozone
has a limited life in aqueous solution, we concluded that the
best treatment of secondary effluent with ozone would be multi-
stage, high-shear, gas-liquid contacting.
In each stage, only the amount of ozone that can be expec-
ted to react was introduced. Because the half-life of ozone
in water is about 20 minutes, we chose a residence time of
10 minutes per stage. About one hour was needed for a COD re-
duction from about 35 or 40 mg/1 to 15 mg/1. Therefore, a six-
stage system was selected for the concept.
The six-stage system was simulated by recycling effluent
through the 18-foot column. Ozone in oxygen, and effluent,
were mixed in an injector which provides a low-cost form of
high-shear contacting. From the injector, the mixture descended
a 20-foot dissolving tube, entered the bottom of the column, and
then rose through the packing. The liquid overflowed at the
top of the column and was pumped back to the injector. Ten min-
utes were required for passage of the liquid around the circuit.
Thus, ten minutes of operation represented the performance of
one stage. In one hour, a six-stage system was simulated.
The rate curves obtained from experiments with the batch
reactor, equipped with the high-shear turbine agitator, could
be used as a design basis for the 6-stage continuous system.
From the experimental curves of COD and TOC reduction versus
time, the amount of ozone to be directed to each stage could be
estimated. In this way, an overall ozone efficiency as high as
90% was obtained.
The experiments outlined above are described in detail in
the sections that follow.
Analytical Procedures
The methods used for sample preparation were adopted from
"Standard Methods for the Examination of Water and Wastewater"5.
Samples of 200 ml were taken for TOC, COD, and ammonia-nitrogen
-4-
-------
analyses. The sample bottles either had glass stoppers or
screw caps with inert plastic liners. Each sample was acidi-
fied with 5 ml of 10% sulfuric acid to stop biological action;
this gave the sample a pl-l of about 2.
COD Determination was made using the dichromate method.
TOG Determination was made with Beckman Carbonaceous Ana-
lyzer, Model 137879 U. The methods given in the instruction
manual for this instrument were followed. Potassium biphthal-
ate, instead of acetic acid, was used for preparing solutions
to standardize the instrument.
Turbidity was measured with a Hach "DR" colorimeter-turbidi-
meter. The scale was calibrated in Jackson Turbidity Units
(J.T.U.). Standards were made up by mixing equal volumes of a
solution consisting of 1 g of hydrazine sulfate in 100 ml of
water and a solution consisting of 10 g of hexamethylene tetra-
mine in 100 ml of water. The mixture was allowed to stand
overnight. Dilution of 1 volume of that mixture with 99 vol-
umes of water gave a turbidity of 40 J.T.U. Further dilution of
the 40 J.T.U. standard with an equal volume of water gave a
20 J.T.U. standard, etc. No correction was needed for the scale
readings of the Hach instrument. The standards and the Hach
turbidimeter readings were compared with a Brice-Phoenix, Uni-
versal, Light Scattering Colorimeter. The readings from the
Hach instrument were more consistent than those from the Brice-
Phoenix Colorimeter because the light scattering from both the
standards and effluents varied with the light angle. This
asymmetric behavior caused relatively large errors in the light
scattering measurements.
Ammonia-nitrogen was determined by direct Nesslerization .
Ozone was determined in effluents and gases by the iodo-
metric method, as follows:
Gas streams were passed through a fritted-glass sparger in
a gas absorber containing 300 ml of water and 100 ml of 10% KI
solution. One-half cubic foot of gas was measured through the
absorber by a wet test meter connected downstream. A 25-ml
sample was pipetted from the 400 ml of solution , 25 ml of 10%
sulfuric acid and 50 ml of distilled water were added, and the
iodine was titrated with standard 0.IN thiosulfate solution.
Ozone in liquids was determined by adding 400 ml of sample
to a mixture of 20 ml of 10% KI solution, 20 ml of 10% sulfuric
acid, and 60 ml of water (all) in a 500-ml glass-stoppered
graduate. The entire 500 ml was titrated with standard 0.IN
thiosulfate solution.
-5-
-------
Gaseous impurities in exit gas and oxygen samples were
analyzed with a gas chromatograph after removal of ozone and
water vapor.
Secondary Effluent Composition
Effluents from the final clarifier of a secondary waste-
treatment plant had the following characteristics:
Color Pale Straw
Odor Variable
Chemical Oxygen Demand (mg/1) 70-150
Total Organic Carbon (mg/1) 20-50
Ammonia Nitrogen (mg/1) 5-30
Turbidity (Jackson Turbidity Units) 30-60
Normally, the effluent contains enough residual detergent to
foam easily.
Initially, we thought that ammonia would react with ozone.
Therefore, a special low-ammonia effluent was obtained from the
waste-treatment plant of the Totowa, New Jersey, Training School
Later we switched to effluent from the municipal waste-treatment
plant of Berkeley Heights, New Jersey.
Ozonizer
Ozone was generated in a stream of pure (>99.5 wt%) oxygen
with a Welsbach, "T-23" , laboratory, ozone generator. This unit
can produce 5 grams of ozone per hour from pure oxygen. It will
give a concentration of about 3.6% ozone, by weight, in an oxy-
gen flow of about 100 liters per hour. Lower concentrations are
obtained by increasing the oxygen flow or reducing the voltage.
Solubility and Decomposition of Ozone in Water
Ozone is about 13 times more soluble in water than oxygen.
The solubility of pure ozone is reported to be between 0.57 and
0.8 g/1 at 20°C. At saturation, an aqueous solution in contact
with oxygen containing 2 wt. % ozone will contain about 11
mg 03/1 and 40 mg O2/1- The instability of ozone makes measure-
ments of its solubility and decomposition rate difficult. The
information given in the literature varies considerably^'^/10,11
The solubility and decay rate of ozone in water were
checked by sparging ozone into water for different periods of
time. The solution was sampled and analyzed for ozone by the
-6-
-------
iodometric method at different times after ozone introduction
had been stopped. The results of these tests gave an ozone
solubility, at equilibrium with 4 wt.% ozone in oxygen, of
10 mg/1 at 25°C starting with distilled water. Decay rates
were 25% in 30 minutes, and 60% in 60 minutes. In tap water,
under similar conditions, the solubility was 6 mg/1. Decay
rates in tap water were 75% in 30 minutes, and 87% in 60 minutes
In a later stage of this project, solubility of ozone in
clarified secondary effluent was determined by sparging 1.56 wt. %
ozone in oxygen into secondary effluent in a cylindrical resin
kettle using a high-shear turbine stirrer. Samples of the
effluent were withdrawn into KI solution after 10 minutes of
ozone treatment, and again after 60 minutes; and each sample
was titrated. The 10-minute sample contained 7.2 mg 03/1,
and the 60-minute sample 8.4 mg/1. Adjusted to 4 wt.% ozone in
oxygen according to Henry's law, this would be equivalent to
21.6 mg/1 of ozone - a much higher value than the 10 mg/1 deter-
mined earlier. The higher value also approaches the solubility
of ozone given by Yost and Russell . Better agitation was
probably responsible for the higher dissolved-ozone values ob-
tained in the later work. Ozone dissolved more rapidly, and
thus more was in solution to react with KI before the ozone
could decompose. Tap water also may contain substances which
decompose ozone.
Clarification of Secondary Effluents
A secondary effluent normally contains suspended organic
matter, part of which can be removed by sand filtration. Addi-
tional colloidal matter can be removed by chemical clarifica-
tion (i.e., coagulation and settling, followed by filtration).
Coagulation and filtration are generally less costly than ozone
treatment on a COD removal basis. Therefore, the burden on_the
ozone treatment can be profitably decreased by a clarification
pre-treatment. A schematic flow diagram of the clarification
equipment is shown in Figure 1.
All secondary effluents used in this investigation were
filtered through a sand bed, 9 inches deep, laid on a 6-inch
layer of graded stone in a 20-gallon stainless steel tank. The
filter was fed by gravity at a uniform rate from a 20-gallon
open feed tank equipped with a variable-speed paddle mixer.
The effluent was spread over the surface of the sand by a per-
forated distributor ring. An outlet in the side of the filter
tank, and located a few inches above the sand surface, served
as an overflow for the backwash. The sand was backwashed with
a large volume of tap water; and its surface was cleaned by
scraping, if necessary, at the end of each day's runs.
-7-
-------
FLOATING SIPHON
LIQUID DISTRIBUTOR
OVERFLOW
TO
RAW EFFLUENT
MIXER
COAGULATION
SETTLING
TANK
DRAIN
SAND
FILTER
BACKWASH
WATER
INLET
CLARIFIED
EFFLUENT
TANK
-oo-
OZONE
TREATMENT DRA|N
Figure I. Schematic flow diagram of the
coagulation- filtration equipment.
-8-
-------
The selected dosage of each coagulant was within the range
normally used for water treatment. The action of the alum,
which was the first one tested, was initially observed on a
one-gallon sample of effluent in a glass jar equipped with a
low-speed laboratory stirrer. The 132-mg/l dosage seemed to
produce a satisfactory floe and did not increase the dissolved
solids content of the filtered effluent excessively. This
dosage was then tried in the 20-gallon apparatus (see Figure 1)
with good results. Lime was compared with alum as a coagulant.
We estimated that a lime dosage of 396 mg/1 would give results
equivalent to 132 mg/1 of alum. However, this was enough lime
to produce a high pH which might increase spontaneous decompo-
sition of the ozone. Conseguently, a lime dosage of 132 mg/1
was also tried. In addition, alum-lime and alum-sulfuric acid
combinations were tested to provide a full range of pH levels
(see Table II) .
Alum was added as a 2.5% solution,and lime as a powdered
hydrate. In both instances, the effluent in the tank was stirred
vigorously. After 5 minutes, the stirring was slowed down, and
very gentle agitation was continued for 30 minutes. The stirring
was stopped, and the floe was allowed to settle for 30 minutes.
The effluent was then drawn out from one inch below the liquid
surface with a floating siphon and was put through the sand
filter.
In earlier runs, the coagulation period was longer, and
the settling was omitted. Settling reduced the possibility of
breaking the floe on the filter and improved TOG and COD re-
movals slightly.
Differences among the various coagulants were usually not
significant with respect to TOG and COD removals (see Table III) .
Most of the coagulant treatments removed about 60% of the COD
from the raw effluent, with the exception of some of the lower
lime dosages which removed only about 40%. The greatest dif-
ference was in the pH of the treated effluents. The high lime
dosage gave a high pH, and the alum-acid coagulants gave the
lowest pH.
Another way of characterizing coagulation is by measuring
turbidity change. Table III also shows how turbidity changed
with coagulation. Turbidities were measured with a Hach "DR"
Turbidimeter and were expressed in Jackson Turbidity Units.
-9-
-------
TABLE II
EFFECT OF COAGULANT ON pH OF CLARIFIED EFFLUENT
AND OF OZONE-TREATED PRODUCT
Coagulant
Ca(OH)2
Ca(OH)2
Ca(OH)2
+A12(S04)3
A12(S04)3
A12(S04)3
+H2S04
Dosage,
rag/1
396
132
132
66
132
110
63-92
Average pH
Raw
Effluent
7.2
7.3
7.5
7.5
7.4
After
Clarification
10.9
8.7
8.8
6.8
5.8
After
Ozone
Treatment
9.6
8.0
7.4
7.8
6.6
-10-
-------
TABLE III
REDUCTIONS OF COD AND TOC BY CLARIFICATION OF BERKELEY HEIGHTS EFFLUENTS
Coagulant
Dosage, mg/1
Ca(OH)2 - 396
Ca(OH)2 - 132
Ca(OH)2 - 132
+A12(S04)3 - 66
A12 (S04) 3 - 132
A12(S04)3 - 110
+H SO4 - 63
None
Coagulated and Filtered
TOC, mg/1
Raw
Effluent
20
25
22
23
21
29
After
Filter
10
11
9
14
12
22
Percent
Removed
50
56
59
39
43
24
COD , mg/1
Raw
Effluent
75
85
81
117
85
111
After
Filter
32
36
34
47
39
68
Percent
Removed
57
58
58
60
54
39
Coagulated, Settled, and Filtered
TOC, mg/1
Raw
Effluent
30
26
30
29
After
Filter
11
16
11
11
Percent
Removed
63
39
63
62
COD , mg/1
Raw
Effluent
87
73
90
91
After
Filter
30
40
33
36
Percent
Removed
66
45
63
60
Turbidity,
J.T.U.
Raw
Effluent
60
39
46
49
40
After
Filter
8
8
6
9
21
-------
OZONATION IN A CONTINUOUS COUNTERCURRENT COLUMN
The most obvious means of ozone transfer from oxygen to
effluent is countercurrent contacting. The experiments dis-
cussed below were designed to evaluate the efficiency of this
process.
Equipment and Procedures
A flow diagram of this system is shown in Figure 2. Oxygen,
containing 1.6 to 3.6 wt.% ozone, was fed continuously through
a stainless steel sparger into the bottom of an 18-foot glass
column, 4 inches in diameter. Effluent was pumped continuously
from the filtrate receiving tank through a flow meter into the
top of the column. It flowed down the column countercurrently
to the rising gas bubbles. Treated effluent flowed up the out-
flow leg to an overflow tee located at the same level as the
liquid surface in the column. Thus, the effluent was the con-
tinuous phase in the column.
In the first tests, the column was not packed. Agitation
by the gas bubbles produced so much turbulence that a high per-
centage of the effluent by-passed to the bottom. As a result,
part of the effluent escaped full treatment; and behavior of the
system was erratic. The column was then filled with 3/8-inch
porcelain Raschig rings to prevent back-mixing. The packed
column was used for all of the runs reported herein.
The duration of most of the tests was about 6 hours. Sam-
ples of liquid were taken from the bottom of the column at
30-minute intervals for TOC, COD, ammonia-nitrogen, and dis-
solved-ozone analyses. The gas leaving the top of tower was
analyzed for ozone by passing it through an absorber containing
KI solution and then through a wet-gas meter. A complete analy-
sis of the gas leaving the top of the tower was carried out during
run 85. A sample of the oxygen, fed to the ozone generator, was
analyzed at the same time. The results are given in Table IV.
The carbon dioxide content corresponds to about one atom of oxy-
gen per ozone molecule reacting to carbon dioxide for 75% of the
entering ozone. About one-third of the nitrogen content in the
exit gas might have been stripped from the water as dissolved
nitrogen. The balance of the nitrogen was probably due to air
leakage and some other forms of nitrogen in the effluent.
In most of the early tests (56-76), an effluent from the
sewage-disposal plant of the Totowa, New Jersey, Training School
was used. The ammonia-nitrogen content of this effluent was low,
about 5-10 ppm. Originally, low-ammonia effluent was chosen in
the belief that ammonia would consume ozone and confuse the test
results. However, the ammonia level decreased very little during
ozone treatment. After independently verifying this with ozone
treatment of an ammonium sulfate solution, permission was obtained
to use Berkeley Heights, New Jersey, effluent which contained
10-30 ppm of ammonia-nitrogen.
-12-
-------
OZONE-OXYGEN
OUTLET AND
EXIT GAS SAMPLE
OVERFLOW-v
TANK A
DRAIN
\
V
i,
(
i
f
tx/e'<<
*££*
'"V
£ £<
t
» l»fcfc
u "
18 FOOT
PACKED
TOWER
i
i
OZONE-OXYGEN
INLET
CONTROL VALVE
7
A
EFFLUENT
FEED PUMP
FLOW METER
FILTERED
EFFLUENT
FEED
SAMPLE TAP
AND DRAIN
Figure 2, Schematic flow diagram of continuous
countercurrent ozonation.
-------
TABLE IV
CONTINUOUS COUNTERCURRENT OZONATION:
COMPLETE ANALYSIS OF THE EXIT GAS FROM THE COLUMN IN
RUN 85, AND OF THE OXYGEN USED IN THAT RUN
Carbon Dioxide
Nitrogen
Methane
Acetylene
Propylene
n-Butane
Isobutane
Nitrous Oxide
Xenon
Krypton
Oxygen, ppm (V/V)
5
10.3
0.07
0.02
0.04
0.02
1.1
0.45
7.2
Exit Gas*, ppm (V/V)
9450
9300
<8.0
<0.01
<0.01
<0.01
<0.01
1.5
0.36
5.0
* Moisture and ozone removed before analysis.
-14-
-------
The effect of pH was tested by adding sujfuric acid or sodium
hydroxide to the filtered feed in runs 70 and 76, respectively.
At the end of run 76, a sample of the ozone-treated effluent
was drawn from the outlet of the column into a sterile container.
This sample was sent to a local water-testing laboratory for
chemical and bacteriological examination. The results are given
in Table V. The sample meets the PHS standards for potable water.
Actually, run 76 was not one of the best runs. Even better re-
sults can be expected with ozone-treated water of later runs.
Res jolts and Discussion
In the early work with the Totowa effluent, different liquid
flow rates, initial COD levels, and pH levels were tested. Table
VI gives a summary of these results.
The COD and TOC efficiency indices given in this and the
following tables are expressed as the ratios of COD (or TOC) re-
moved to the molecular equivalents of ozone fed into the process.
It is assumed that one atom of oxygen in each ozone molecule is
used in oxidation of organics, while the other two are lost as
molecular oxygen. Thus, the definitions are:
COD/ozone efficiency index =
COD removed (mg/1) x 10Q%
1/3 of ozone suppliedfmg/1 effluent)
TOC/ozone efficiency index =
TOC removed (mg/1) x 100%
1/8 of ozone supplied(mg/1 effluent)
If more than one atom of oxygen per ozone molecule reacts, or if
some of the oxygen carrier gas takes part in the oxidation, the
efficiency indices can be over 100%.
In runs 58 and 60, where the COD was reduced to 5 mg/1, the
residence time was 1.8 hours which was considered excessive. The
gas leaving the tower contained 34% of the ozone fed, making the
COD efficiency index low. In runs 66 and 68, a large excess of
COD was fed, resulting in high COD/ozone efficiency indices and
lower ozone contents in the exit gas. However, the treated ef-
fluent had a high COD. In runs 70-76, a flow rate was used which
gave a 0.9-hour effluent residence time. Approximately 30-35%
excess ozone was fed. The results at pH = 7 seemed to be slightly
better than those at higher and lower pH levels. Under these
-15-
-------
TABLE V WATER ANALYSIS
WATER FROM Experimental Work DATE mT.T.Fr.Tim February 5 1Q
Samle No. 1 So
Tab,
Sample No. 2 Source
Tertiary Sewage treated with Uzone Run 76'
Samle No. "? Source
Samle No. L Source
Sample No. 5 Source
Sample No. 6 Source
Sample No. 7 Source
Sample No. 8 Source
No. 2
No.
N
No. 6
No. 7
No. 8
PHYSICAL ANALYSIS - P.P.M.
Color
Turbidity
Hot
Trace
Sep.
Taste
CHEMir.iT. ANALYSTS - PJ>J*.
Alkalinity to Phenolphthalein
T2T
tlVaHnity - Total
Hardness - Total (CaC03>
122
Free CO?
~20lT
Total Solids
Manganese (Mn)
Iron (Fe)
Iron & Aluminum Oxides (R^)
Tract
Calcium (Ca)
Magnesium (Mg)
JteJ
LLL
Chloride fCl
Fluorida ttO
Sulfates fSOtt)
JJJL
7.9
BACTERIOLOGICAL ANALYSIS
0.1 I 10 0-1 I 10 O.I I 10 0.1 I 10 O.I I TlO O.I I 10 O.I
Quantity of Water Tested - ml.
10
O.I
10
Gas in Lactose Broth
Gas in Brilliant Green Bile
Coliforms Present
None
Bacteria per ml.(V7cC)
n
Millipgre Count/ 100 ml
EMARKS; Bacteriological Examination indicates wat.flr- t.n he free of
Bacteria and to meet PHS standards for potable water. Physical F.vam-inat.-ir.n
Color is minimal, clarity good, odors meet annsptahlp st.anHqr^Sl Chemical
tests using very small aliquots(giving ton gT-eat. a margin fn^ orr^r) show
sample as meeting acceptable quality
Examined at 2$9 Woodland Ave, Siin^^t jj.J.
Date February 12. 1968
Analyst
Form W-132
-16-
-------
TABLE VI
CONTINUOUS COUNTERCURRENT OZONATION OF TOTOMA. EFFLUENTS
i
56
58
60
62
64
66
68
70
72
74
76
fj
fy\ fa
22.7
n.4
11.4
34.2
34.2
11.4
11.4
22.7
22.7
22.7
22.7
H
7
7
7
7
7
7
7
5
7
7
».s
O &
5 o
O FH
Q rt
*£
a c
J3
67
62
64
73
70
88
84
69
72
80
75
4"^ 3E
cj Q
ft
o
g
N
O
5
7
5
-
4
28
-
8
7
-
11
Flowt 102 liters/hr at 22"0, ?60 mm of Hg, in all runs.
-------
conditions, the exit gas contained about 20% of the ozone fed,
the COD/ozone efficiency index was about 70%, and the treated
effluent had satisfactory TOC and COD levels.
The ammonia-nitrogen content of the effluent was not af-
fected by ozonation in any of these tests. To investigate this
further, a solution of ammonium sulfate equivalent to a nitrogen
content of 28 mg/1 was treated with ozone in the tower under the
same conditions as had been used in run 76. Two 6-hour tests
were run, and the ammonia-nitrogen content of the treated samples
varied between 24 and 29 mg/1, indicating that ozone did not react
appreciably with the ammonia.
Berkeley Heights effluent was used for the next group of
tests which were intended to test the effect of clarification
pretreatments on ozonation. The results are given in Table VII,
together with the results of runs without clarification pretreat-
ment.
Both the Totowa and the Berkeley Heights waste-treatment
plants used a trickling filter for secondary treatment. The
throughput in the Totowa trickling filter, expressed as a per-
centage of rated capacity, was lower which may be the reason
why the Totowa effluent is somewhat more amenable to ozone treat-
ment. Some substantiation for this may be the lower TOC and COD
levels that were obtained with Totowa effluent after ozonation.
With Berkeley Heights effluent, no satisfactory COD levels were
reached without a clarification pre-treatment. We also noticed
that, with the approach of spring, the COD levels in the Berkeley
Heights secondary effluent decreased from about 115 to 85 mg/1.
Warmer weather increased the efficiency of the secondary treat-
ment.
The ozone efficiency indices decrease with decreasing TOC
and COD levels of the secondary effluent, indicating that the
most-readily bio-oxidizable organic compounds in the effluent
also consume ozone most efficiently. It is much more difficult,
for example, to reduce the COD from 25 to 15 mg/1 than from 65
to 55 mg/1. Figure 3 shows a correlation between the COD/ozone
efficiency index and the COD of the column feed. It appears
relatively independent of the type of coagulant used. The cor-
relation is affected, however, by the liquid residence time in
the column.
A high dosage of lime coagulant (runs 94 and 95) raised the
pH in the column to 10.6. During ozonation, a gradual decline
in the ozone level in the tower was noticed. After 3 hours, no
ozone could be found in the effluent from the bottom of the
tower and very little in the exit gas. This was possibly caused
by the gradual build-up of an alkaline deposit that actively de-
composes ozone. The lower dosage of lime (runs 96 and 97) showed
signs of decomposition catalysis, but operation was much better.
Another effect of the high pH (runs 94 and 95) was a drastic re-
duction of the ammonia-nitrogen content which may have caused
part of the loss of ozone.
-18-
-------
TABLE VII
COHTIHDOnS CX)UHTEHCDHHEHT OZOHATIOH Of BEHXELET HEIGHTS EPFLDEHTS
Run
No.
77'
78-
7Cj.
85'
86-
87'
88-
89'
90"
91"
92"
93"
94.
95'
96-
97'
98-
99'
100-
101*
Day-
of the yeek
3
5
4
4
1
3
2
3
4
5
1
2
4
5
1
2
3
Ozone in Gas
Peed, ng/1
45
47
44
44
44
40
44
44
22
22
23
23
4B
47
48
49
48
47
48
47
Coagulant.
«*/l
None
Alum-
152
None
Aluo-
152
Lime-
396
Llme-
152
Llme-
152
lAlum-
66
Alum-
110
1fV
Alum-
110
'H2S04-
92
Overall Averages:
Without coagulation-
Uith coagulation-
Averages for 22-25 me
With coagulation-
/i or o,
- y
Average pH
In the
Columa
7.4
7.4
7.4
7.*
7.4
6.9
7.0
7.0
7.4
7.4
7.0
7.0
10.6
1O.6
8.4
8,3
8,2
7.9
6.6
6.1
7.4
7.8
7.4
7.0
TOC, «g/l
Ban
Effluent
28
52
27
22
20
55
17
IB
55
18
n
16
21
%
22
22
21
&
26
25
26
25
After
Filter
28
26
\l
20
16
11
21
11
12
11
10
11
12
11
10
a
12
10
19
11
16
11
Ozone
Treated
11
11
10
1»
a
12
4
6
11
7
7
5
6
6
1
3
5
6
4
10
6
9
6
COD, mg/1
Bar
Effluent
110
156
94
120
116
118
119
11»
105
75
as
91
70
80
90
81
68
95
85
89
112
91
90
89
After
Filter
90
88
49
46
46
59
56
24
28
29
56
38
35
35
33
39
'JU
67
35
44
26
Osone
Treated
20
21
26
3*
25
19
13
29
17
17
15
15
10
13
11
16
16
13
17
24
15
25
16
0, Balance , %
U>ed «
Decomp.
in Tower
70
74
73
86
82
81
71
70
77
56
54
60
86
95
80
72
60
54
51
29
74
66
67
57
In Effluent
Leaving
Tower
6
6
6
4
5
6
6
6
4
5
5
4
2
7
a
8
9
8
10
7
5
8
In Exit
Gag
24
19
22
10
1«
14
25
24
19
39
41
55
10
, 3
15
20
32
37
41
61
21
27
29
35
Efficiency
Indices, %
TOC
Ozone
70
57
15
22
44
17
41
19
45
20
26
27
14
18
26
22
23
15
21
20
39
24
32
26
COD
Ozone
86
43
42
51
54
36
14
22
19
56
38
32
27
24
52
13
58
50
45
18
Anunonia-N, mg/1
Raw
Effluent
21
26
28
51
28
30
31
26
29
10
14
19
11
11
17
18
20
19
19
25
19
20
16
After
Filter
15
21
23
25
24
23
19
20
16
5
15
18
8
12
14
15
13
12
15
19
15
10
15
Ozone
Treated
14
20
20
26
23
24
20
21
23
6
14
19
2
12
11
10
10
12
19
13
15
16
Affluent Plow
Gas Flow
- 22.? 1/hx.
- 102 1/hx. (22°c, 760 i
i Hg)
Effluent Flow
Gaa Flow
22.7 1/nr.
170 1/hr. (22°C, 760 nun Hg)
Day of the week 1 - Monday to 5 - Friday
-------
100
LJ
O
U_
u_
UJ
Id
O
N
O
^
Q
O
O
80
o 60
40
20
EFFLUENT SOURCE
D BERKELEY HEIGHTS
o TOTOWA
0 20 40 60 80 100
COD OF COLUMN FEED (mg/l)
Figure 3, Continuous countercurrent ozonation:
COD/ozone efficiency index as a function
of COD of the column feed.
-20-
120
-------
The alum-sulfuric acid combination (run 100), which lowered
the pH to 6.6, gave excellent TOC and COD removals and i relatively
good COD/ozone efficiency considering the high concentration of
ozone in the exit gas. Lowering the pH to 6.1 (run 101) increased
the ozone concentration in the exit gas still further, but seemed
to retard COD removal to some extent.
At the lower ozone concentrations, more gas was supplied,
and a higher fraction of the ozone fed left the column in the
exit gas. This decreased ozone-absorption efficiency was an-
ticipated.
Summarizing the results of the continuous countercurrent
ozonation, we found:
1. Ozonation efficiency of a secondary effluent was
high when its COD and TOC were high. However,
such an ozonated effluent had an unacceptably
high COD and TOC.
2. Berkeley Heights effluent was more difficult to
oxidize to a low COD level than was Totowa efflu-
ent. However, pre-clarification enabled us to
bring the COD of Berkeley Heights effluent from
35 to 15 mg/1 after 0.9 hours of treatment with
ozone.
3. The pH range from 6.6 to 8.3 seemed to be optimum.
-21-
-------
BATCH REACTOR TESTS
Ozonation in the continuous countercurrent column is a
combination of mass transfer and chemical reaction. Since we
knew neither the relative importance of each of these two steps
nor the effectiveness of the column packing for this particular
mode of operation, scale-up predictions were difficult to make.
Therefore, we decided to make the next fundamental ozonation
studies in a batch reactor. The main variables to be investi-
gated were:
1. Agitation rate.
2. Clarification with different coagulants.
3. Ozone concentration in the gas.
Equipment and Procedures
Batch ozonation tests were conducted with both low-shear
and high-shear agitation. The low-shear apparatus consisted of
a 250-ml, three-necked, roundbottom flask fitted with a porous-
glass dispenser for gas introduction and with a low-speed paddle
agitator. The high-shear apparatus, shown in Figure 4, consisted
of a 2000-ml resin flask fitted with a flat-blade turbine agita-
tor driven at 1200 rpm by an electric motor. Gas, introduced
through a 1/4-inch O.D. aluminum tube, was directed to the bottom
of the agitator. Liquid samples were withdrawn through 1/4-inch
O.D. aluminum tubing. A baffle system was suspended above the
agitator to effect proper mixing.
The operating procedures were different for the two types
of batch reactors. For the low-shear reactor, batches of fresh
effluent were treated with excess ozone for various periods of
time, at the end of which the total batch was prepared for TOC
and COD determinations by stabilization with a nitrogen purge
and addition of acid.
For the high-shear reactor, 2000 ml of fresh effluent
(clarified, if desired) were placed in the reactor and treated
with excess ozone. Liquid samples were periodically withdrawn
and prepared for TOC and COD determination by stabilization with
a nitrogen purge and the addition of acid.
Significant differences were noted in the quality of the
secondary effluent each day. Varying levels of TOC and COD of
the effluent as received, foaming tendencies during testing,
and color of the effluent after filtration were indications of
these changes. In order to minimize the effects that effluent
variability would have on our study of process variables, a
series of tests was conducted for each variable, and the arith-
metic mean of the results was used.
-22-
-------
LIQUID
SAMPLE
PORT
I
ro
AGITATOR
DRIVE
1
r^-
N
V.
^
r
G/
1 INL
-^~
u
^s
ET
u
\
I
GAS EXIT
/-RESIN
^ (20(
-AGITATOR
BAFFLE
Figure 4, High-shear batch reactor,
-------
Results and Discussion
The overall reaction rate of ozone with organic contami-
nants in waste water is a combination of the transfer rate of
ozone from the gaseous phase to the liquid phase (mass transfer) ,
and the reaction rate of the transferred ozone with the organics.
Gas-liquid mass transfer is characterized by the mass transfer
coefficient which is unit mass transferred per unit time, unit
cross-sectional area and unit driving force. This mass transfer
coefficient is described as either gas-film, liquid-film, or
combination-film controlled. In mass transfer of relatively in-
soluble gases such as oxygen or ozone, liquid-film transfer con-
trols. This means that the major resistance to mass transfer is
imposed by the presence of a stagnant liquid film, or layer,
12
through which mass transfer must occur . Thus, by increasing
the interfacial area between phases and increasing the agitation
of the liquid phase so as continually to renew the contact sur-
face exposed to the gas phase, resistance to mass transfer would
be reduced. High agitation, coupled with high shearing action,
would achieve these requirements.
A comparison of the overall rates of ozonation in the low
and high-shear batch reactors is shown in Figure 5 as a plot of
TOC versus time. The asymptotic approach of the curve for the
high-shear case indicates that part of the TOC is refractory to
ozone treatment. Thus, the total TOC content can be thought of
as consisting of two general classes of materials: one which is
relatively refractory to ozonation, and one which is readily
destroyed by ozone. A second set of plots in Figure 5 is shown
in which an arbitrarily selected, "refractory fraction" of 30%
of the inlet TOC content was subtracted from the total TOC values,
The resultant plots are straight lines indicating that the reac-
tion is first-order for the non-refractory part of the TOC. The
slopes of the non-refractory TOC levels are 0.027 min~ for the
high-shear case and 0.014 min for the low-shear. This signifi-
cant increase in reaction rate realized with increased agitation
shows that, for effective ozonation of sewage effluent, good
agitation must be considered the prime objective in contactor
designs. All subsequent batch-reactor tests were conducted in
the high-shear batch reactor.
Dissolved ozone levels were determined for the high-shear
batch reactor. The results in Table VIII indicate that the gas-
to-liquid transfer rates were greater than the reaction rates.
The value selected for the refractory TOC in the agitation
studies was based on the asymptotic approach value of the high-
shear runs and was assumed to be equal to 90% of the TOC content
remaining after 120 minutes of ozonation. Similar data handling
-24-
-------
IIII I
o HIGH SHEAR
D LOW SHEAR
TOTAL TOC
^^ Q
NON-REFRACTORY
TOC
L_J I
1 I 1 I I I
40 60 80
TIME (MIN)
Figure 5, Ozonotion in batch reactors;
effect of agitation on TOC
removal with excess ozone.
-25-
-------
TABLE VIII
DISSOLVED OZONE LEVELS FOR HIGH-SHEAR BATCH REACTOR
Time,
Min.
0
10
60
Ozone
In Gas Feed,
mg/1
22.6
22.6
22.6
TOC,
mg/1
15
-
7
COD,
mg/1
28
-
10
Dissolved Ozone
mg/1
0.0
7.2
8.4
% of Saturation
0
72
84
-26-
-------
methods for all batch tests resulted in essentially similar re-
sults; i.e., plots of the logarithm of non-refractory TOG versus
time gave straight lines. In reality, the organic content of
most effluents is expected to be composed of compounds encompas-
sing the full range fro-n totally refractory compounds to com-
pounds readily destroyed by ozonation. If the refractory part
of the TOC were also refractory to acid bichromate in the COD
test, the ratio of COD to TOC could be expected to drop below
the expected ratio of ".67:1 for carbohydrates and most proteins.
Results in all tests show that this ratio (COD:TOC) does, in fact,
decrease with increasing ozonation, thus lending credence to the
concept of an ozone-refractory class of compounds. These tests
also indicate that ozone selectively attacks those compounds
which contribute most ho oxygen demand. This is of particular
importance for maintenance of dissolved oxygen levels in bodies
of water into which tertiary effluents will be discharged.
The effects on ozonation rates of preclarification and of
the type of coagulant used are shown in Figure 6 (a plot of TOC
versus time) and in Figure 7 (a plot of COD versus time). The
results obtained for TOC removals were treated in the manner
previously explained, and the results are shown in Figure 8 (a
plot of non-refractory TOC versus time). The slopes of the lines
indicating relative reaction rates are shown in Table IX.
The type of coagulant used had little effect on the effi-
ciency of coagulation, but had a marked effect on the rates of
COD removal (see Figure 7); in general, low pH resulted in lower
rates. Ozone decomposition in aqueous solution increases rapidly
o
with increased alkalinity ; thus, activity of the dissolved ozone
in solution is enhanced by higher pH. In these batch tests with
excess ozone and effective agitation, the chemical reactions are
rate controlling, and the pH effect is not masked by mass trans-
fer effects.
Figure 6 shows that the coagulants remove part of the re-
fractory TOC, thus making a lower final TOC possible. Alum was
the most effective coagulant in this respect. Figure 8 and
Table IX show that non-refractory TOC removal rates are approxi-
mately the same for alum, lime and no coagulant, but significantly
lower for alum plus acid.
Thus, the decrease in COD is a primary indicator of the
reaction of organics with ozone, and depends directly on the
activity of the ozone which in turn depends on the pH. The TOC
decreases only if the organic molecules decompose by splitting
off carbon dioxide. The rate of this decomposition is reduced
only at a pH below 7. The efficiency of ozone consumption was
not calculated in this series of tests since expected analytical
errors would be great for the excessive quantities of ozone used.
-27-
-------
40
20 -
10
^ 8
o>
1
(
X.
^
fe M
1 1 1 I 1 1 1 1
COAGULANTS
Ol IMP
D ALUM
A ALUM AND ACID
x NONE
v%
^^r^"*^'^-^
*
pH*
9.6
7,1
6,1
7,4
^* * ^^^^^
.
O
- \
AFTER CLARIFICATION
I I I I I I I I I I I
40 60 80
TIME (MIN)
100 120
Figure 6, Ozonation in the high-shear batch reactor!
effect of clarification onTOC removal with
excess ozone.
-26-
-------
100
90
80
70
60
i r
i i r
COAGULAN'
o LIME
a ALUM
A ALUM AND ACID
X NONE
5
4
9,6
7,1
6,1
7,4
^AFTER CLARIFICATION
I I I I I 1 I I I I I I
0 10 20 30 40 50 60 70 80 90 100 110 120 ISC
OZONATION TIME (MIN)
Figure 7, Ozonation in the high-shear batch reactor:
effect of clarification on COD removal with
excess ozone,
-29-
-------
20 r
*I I I
1 I I I
COAGULANTS
I
-o LIME
-n ALUM
T
-&H.'
9.6
7,1
A ALUM AND ACID 6,1 -
x NONE
7,4
o
o
0,5
0,1
AFTER CLARIFICATION
I I I I I I I I
1
I I I
0 10 20 30 40 50 60 70 80 90 100 MO 120 130
OZONATION TIME (MIN)
Figure 8, Ozonation in the high-shear batch reactor:
effect of clarification on non-refractory TOC
removal with batch reactor.
-so-
-------
TABLE IX
FIRST-ORDER RATE CONSTANTS FOR NON-REFRACTORY TOG REMOVALS
WITH EXCESS OZONE IN THE HIGH-SHEAR BATCH REACTOR
Coagulant
Alum + Acid
Alum
None
Lime
pH
After Filter
6.1
7.1
7.4
9.6
Rate Constant,
min~l
0.020
0.034
0.027
0.029
-31-
-------
The majority of the tests in the batch reactor were conducted
with 3.4 wt.% ozone in oxygen (45.2 mg/1). Welsbach data show
that ozone production costs are lowest for 1.7 wt.% ozone in
oxygen (22.6 mg/1). A series of tests was made to investigate
the effect of ozone concentration on the rates of TOC and COD
removal. The results in Figure 9 show no rate difference between
3.4 wt.% and 1.7 wt.% ozone. Filtered uncoagulated effluent
was used in these tests. Each day for three days, one test with
3.4 wt.% ozone and one with 1.7 wt.% ozone were conducted using
a cut of the same effluent. The values plotted are the arith-
metic means of the results.
A second series of tests using alum-and-acid-clarified
effluent was conducted comparing 1.7 wt.% and 0.8 wt.% ozone.
The arithmetic means of the results of this series, shown in
Figure 10 (a plot of TOC and COD versus time) , show no signifi-
cant differences in reaction rates.
q
Kilpatrick etalT, report that light accelerates the
decomposition of ozone in aqueous solution. Since their batch
tests were conducted in glass equipment in the presence of
natural light, there was some question of the applicability of
results obtained for design of commercial units wherein light
would probably be absent. A series of runs was conducted to
study the effect of light. The results in Figure 11 (a plot of
TOC and COD versus time) showed no difference in reaction rates.
Up to this point, our work on ozone treatment had not
proven that reduction in COD to 15 mg/1 would bring a correspond-
ing reduction in BOD. Since BOD is the presently used quality
criterion for regulating effluent treatment, a knowledge of this
relationship is important. Thus, special tests were made for
comparing BOD of ozone-treated effluent with COD.
Filtered effluents from Berkeley Heights and Florham Park,
New Jersey, were treated for one hour in the high-shear batch
reactor with oxygen containing 24 mg ozone/1. Samples were sealed
in sterile bottles and sent to the laboratory of the FWPCA pilot
plant at Blue Plains, Washington, D.C., where BOD determinations
were made. The samples sent to Washington were not stabilized
with acid, while those used for the COD analyses had the usual
acid addition.
Table X summarizes the results for these analyses. With
both effluents, the ozone-treated material had a low BOD of about
3 mg/1. The filtered effluent (11-20-1) from the activated
sludge plant of Florham Park, had dropped to a rather low BOD
(5 mg/1). The oxygen content of this effluent was probably high
enough to continue the BOD removal in transit.
-32-
-------
o>
£
100
80
1
o TOC © 3.4 WT. % OZONE
n TOC (a) 1.7 WT. % OZONE
A COD (S) 3.4 WT % OZONE
x COD (a) 1.7 WT.% OZONE
o
4 -
0
20
40 60 80
TIME (MIN)
100
Figure 9
Ozonation in the high-shear batch reactor;
effect of ozone concentration onTOCand COD
removals with filtered ,unclarified effluent and
with excess ozone.
-33-
-------
100
90
80
70
60
50
3~1 I I I I I I I I I I l_
0.8 WT.% OZONE _
1.7WT.% OZONE _
COD -
10
9
8
7
6
5
4
0.8 AND I,7WT,%OZONE
o TOC AT 0.8 WT. % OZONE
D TOC AT 1,7WT.% OZONE
A COD AT 0.8 WT.% OZONE
X COD AT 1.7 WT.% OZONE
I I I I I I I I I
1
1 1
10 20 30 40 50 60 70 80 90 100 110 120 130
OZONATION TIME (MIN)
Figure 10, Ozonation in the high-shear batch reactor;
effect of ozone concentration on TOC and COD
removals with clarified effluent and with
excess ozone.
-34-
-------
20
10
9
8
7
6
4
3
D
o TOC, EXPOSED TO NATURAL LIGHT
D TOC, COVERED
A COD, EXPOSED TO NATURAL LIGHT
x COD,COVERED
I I I I L
'0 10 20 30 40 50 60 70 80 90 100 110 120 130
OZONATION TIME (MIN)
Figure II. Ozonation in the high-shear batch reactor:
effect of exposure to natural light on TOC and COD
removals with excess ozone In the high-shear
batch 'eactc-.
-35-
-------
TABLE X
BIOCHEMICAL OXYGEN DEMAND OF EFFLUENTS
TREATED 60 MINUTES WITH OZONE IN THE
'HIGH-SHEAR BATCH REACTOR
Berkeley Heights Effluent, Samples 11-14-o through 11-14-4
Florham Park Effluent, Samples 11-20-0 through 11-20-4
Sample
Number
11-14-0
11-14-1
11-14-2
11-14-3
11-14-4
11-20-0
11-20-1
11-20-2
11-20-3
11-20-4
Source
Original Effluent
Filtered Effluent #1
Ozone Treated #1
Filtered Effluent #2
Ozone Treated #2
Original Effluent
Filtered Effluent #1
Ozone Treated #1
Filtered Effluent #2
Ozone Treated #2
Turbidity ,
J.T.U.
30
15
2
15
2
23
9
2
12
2
COD,
mg/1
63
44
16
44
16
61
34
13
36
13
BOD,
mg/1
51
36
4
36
3
-
5
3
-
2
-36-
-------
SIMULATION OF A SIX-STAGE
CQCURRENT CONTACTING SYSTEM
High ozone-utilization efficiency is the key to economic
water treatment. Ozone is most efficiently used when it is
consumed quickly. By cocurrent contacting, ozone and organic
matter are brought together at their highest concentrations,
thus obtaining the maximum, initial, reaction-driving force and
consuming the bulk of the ozone in useful reaction before its
spontaneous decomposition can occur. High-shear contacting^of
entering effluent with entering gas in an injector aids rapid
transfer of ozone from the gaseous to the aqueous phase.
The reaction between dissolved ozone and many of the
organic compounds to be removed is slow. Consequently, all
of the ozone must not be introduced at one time. The use of
several stages is required. The experiments described below
simulated a six-stage cocurrent reaction system. Because ozone
has a half-life in water of about 20 minutes, a 10-minute resi-
dence time per stage was selected. About an hour is needed to
reduce the COD from about 35 to 15 m.g/1. Therefore, six stages
were used.
Equipment and Procedures
The 18-foot column was modified as shown in Figure 12 to
simulate multistage contacting. Ozone in oxygen was mixed with
effluent in an injector. The dispersed mixture descended a
20-foot dissolving tube and entered the bottom of the column.
Both the gas and liquid then flowed upward through the column
packing, and the liquid was returned through an overflow siphon
at the top to the suction of the recycle pump. The liquid flow
rate was adjusted to provide for a recycle time of 10 minutes,
representing the performance of one stage. In one hour, a six-
stage system was simulated. The holdup time for the gas was
about one minute.
The recycle pump was a stainless steel, positive displace-
ment, sanitary pump with a variable-speed drive. A by-pass
valve made fine adjustment of the flow possible.
The vertical 20-foot dissolving tube gave a contact time
for intensive gas-liquid contacting of about 6 seconds. The
increasing hydrostatic pressure exerted, as the mixture flows
down improves ozone transfer from the gaseous to the liquid
phase in the mixture. This principle is also used in the so-
called Otto process for ozonizing drinking water. The Otto
process, however, is limited to one stage since much less ozone
is needed to sterilize drinking water.
-37-
-------
t
GAS INJECTOR
MAKE-UP
WATER
OZONE-
OXYGEN
INLET
*
t
TOP
LIQUID
SAMPLE
01
00
I
FLOW
METER
DISSOLVER
TUBE
*
I
OZONE-OXYGEN OUTLET
EXIT GAS SAMPLE
1
RASCHIG RINGS
-LIQUID
RECYCLE
18 FOOT
PACKED TOWER
DRAIN VALVE
BOTTOM LIQUID
SAMPLE
1
RECYCLE
PUMP
Figure 12, Schematic diagram of the simulated
multistage cocurrent contactor.
-------
Normally, liquid samples were taken for analyses at 10,
20, 40, and 60 minutes after the start of the run. Outgoing
gas samples were taken at 2, 30, and 50 minutes. This timing
was used to avoid taking gas samples while the system was
being disturbed by the liquid sampling. Introduction of un-
treated effluent for make-up, after sampling, caused an increase
in the TOG and COD. To check this effect on the TOG and COD
obtained after 60 minutes, run 117 was made without taking
liquid samples during the run so that no make-up with untreated
effluent was required.
The ozone input could be adjusted at various stages of
the experimental run to match the ozone demand of the effluent.
Thus, the effect of excesses and deficiencies of ozone at
various stages could be analyzed.
The ozone generator could produce mixtures ranging from
21 to 48 mg/1 ozone in oxygen. To obtain lower ozone concen-
trations , oxygen from a separate cylinder was passed through a
flow meter and was then mixed with the stream from the ozonizer.
The settings of the ozonizer and flow controls were governed by
ozone analyses of the mixed stream.
Results and Discussion
Results from representative runs are shown in Tables XI
and XII. All of the effluents used in these simulation tests
were coagulated, settled, and filtered.
The COD/ozone efficiency indices of the first stages of
all the experiments with this system, except the one with the
highest pH, were over 100%. The more-active organic compounds
in the effluent apparently used more than one oxygen atom per
ozone molecule. This phenomenon was also reported recently by
Koppe and Giebler
Runs 109 and 111 were made at a low ozone concentration of
10.6 mg/1 and a constant gas flow rate throughout the run. The
COD/ozone efficiency indices decreased very rapidly from about
200% in the first stage, to about 110% in the second stage, and
to 81% and less in the last three stages. This reflects that a
deficiency of ozone was fed to the first and the second stages,
while an excess of ozone went to the third and subsequent stages.
For runs 113 and 115, the amount of ozone fed to the first
two stages was somewhat more than doubled by increasing the
ozone concentration to 22.6 mg/1. After 25 minutes, the gas flow
rate was halved. In these runs, the COD/ozone efficiency indices
were above 100% only in the first stage.
-39-
-------
TABLE XI
SIHULATIOH 0? SIX-STAGE COOUHHEMT OZOBATIOK OP BERKELEY HEI6BTS EFFLUEHTS:
OZONE A5ALYSE3
Run
No.
109
111
1^7
113
115
117
123
Coagulant ,
mg/1
Lime-132
Alum-110
+H2S04-63
Lime-396
Lime-132
Alum-132
Alun-132
Alum-110
+H2S04-85
pH in
the Column
8.2
6.6
9.0
8.5
7.8
7.8
6.2
O'ZOM t!OH(Jia)TkAl
Gas in,
mg/1
10.6
10.6
22.6
22.6
22.6
22.6
22.6
Gas Out, mg/1
10 rlln
1.1
5.0
0.1
1.3
5.6
*
5-7
bo run
5.0
5.3
0.0
2.1
8.0
10.6
*.9
pfOgg
Liquid* at Bottom, mg/1 at
io Win
3.0
4.7
3.9
12.0
12.0
*
15.0
bt> run
3.5
5.8
0.0
12.0
12.0
10.0
6.0
Liquid* at Top, mg/1 tt
10 HIn
1.0
1.9
0.6
2.0
4.2
* *
6.0
bo run
1.0
2.5
0.0
2.6
4.0
5-7
2.0
O
I
Saturated aqueous solution in contact with 1 atm gas containing 10.6 mg 0,/1 4.6 mg 0,/1
' ' gas containing 22.6 mg Oyl 9-7 "8 0;/1
Saturated aqueous solution in contact with 1 atm
'No samples taken until end of run
-------
TABLE XII
SIMULATION OF SIX-STAGE COCURRENT OZQHATION OF BERKELEY HEIGHTS EFFLUEHTS:
TOG So COD REMOVALS
Run
No.*
109
111
12?
113
115
123
pll
8.2
6.6
9.0
8.5
7-8
6.2
GOD
of
Feed,
mg/1
36
36
30
43
5
41
STAGE NUMBER
;
COD
out,
JHK/1
2?
27
22
31
24
27
0,
ill,
mg
400
400
989
791
791
989
CQD/Ozone
Efficiency
Index,
%
214
190
69
143
124
119
2
GOD
out,
mg/1
23
23
1?
24
20
23
0-,
iri,
rag
400
400
791
791
791
791
GOP/Ozone
Efficiency
Index,
%
108
108
68
96
62
68
3
COD
3Ut ,
mg/1
22
20
16
21
18
20
0,
iri,
mg
400
400
395
593
593
395
COD/Ozone
Efficiency
Index,
%
40
81
27
55
27
82
f _j
COD
out,
rng/1
21
18
18
20
17
18
o,
iri,
mK
400
400
395
395
395
395
COD/Ozone
Efficiency
Index,
%
2?
53
-
27
27
55
5
COD
out,
mK/1
20
17
18
19
16
16
Oz
in,
mg
400
400
395
395
395
395
COD/Ozone
Efficiency
Index,
%
8
25
-
13
14
41
6
COD
out,
mp/1
20
17
18
19
16
15
°3
iri,
mg
400
400
395
395
395
395
COD/Ozone
bf 1 iciency
Index,
-
-
-
-
14
27
Bun '
No.*
109
111
127
113
115
117"
123
Coagulant
fia(oit)
2'
mK/1
132
-
396
132
-
-
-
Alum,
mK/1
-
110
-
-
132
132
110
jj 30
2 4
mK/1
63
- -
-
-
_
85
PH
~ "" Haw
Effluent
7.2
7.4
7.2
7.3
7.5
7.5
7.2
After
Filter
8.2
5.6
10.8
8.7
6.8
6.8
5.3
Ozone
Treated
8.2
6.6
9.0
8.5
7.8
7.8
6.2
Haw Effluent
Tdc,
mK/1
28
29
30
24
21
42
31
COD,
mg/1
75
91
89
70
83
98
96
Turbidity,
JTU
39
49
60
39
39
46
39
After Filter
TOC,
mK/1
17
11
11
15
14
12
22
COD,
mK/1
36
36
30
43
35
32
41
Turbidity,
JTU
8
9
8
8
8
6
10
After 60 min.
0, Treatment
TKr,
mK/1
10
9
9
8
11
9
11
UULI,
mK/1
20
17
18
19
16
13
15
ruroiaity,
JTU
2
3
2
3
2
3
9
" "Overall
Efficiency
Indices, %
TOC/Oj
89
25
18
64
27
25
98
5
76
90
41
78
61
58
88
Effluent feed rate = 9 gal./hr.
"Samples taken only at 60 min.
Initial charge = 9 g9l«» Plus 1 gal-
compensate for samples (except in Hun 117)
-------
In runs 123 and 127, the ozone feed rate was changed twice:
first at 10 minutes, (from 98.9 mg 03/min to 79.1 mg 03/min) and
again at 20 minutes, (from 79.1 mg 03/min to 39.5 mg 03/min) . The
ratio of the ozone feed rates for the first/second/subsequent
stages was 15/12/6. This brought the ozone feed closer to the
expected ozone demand in the first three stages. In run 123,
this resulted in a gain in overall efficiency. In run 127, the
COD of the feed was considerably lower which, in combination
with the high dosage of lime coagulant, reduced the overall
COD/ozone efficiency index to only 41%. The relatively good
performance in run 113, using a lower lime dosage for clarifica-
tion, followed a thorough cleaning of the system with inorganic
alkaline cleaners. After cleaning, the first rinse was neutra-
lized with nitric acid to pH = 7. This was followed by several
water rinses.
Table XI gives the ozone concentrations in the gas and
liquid phases at the beginning and end of each run. The data
show that ozone decomposed more readily in run 127 where the
effluent had a high pH as a result of coagulation with a large
dose of lime. This increased ozone decomposition was accom-
panied by a loss in oxidation efficiency. A time-delay effect
was noticed. Decomposition became more severe with passage of
time during the run and was more sensitive to pH in these experi-
ments than in the countercurrent experiments or in the experi-
ments with the batch reactor. The ozone-water mixture passed
down a long, small-diameter tube from the injector. In this tube,
and on the packing of the column, ozone contacted a large solid
surface in proportion to the volume of the system. At a high
pH, alkaline deposits that catalyze ozone decomposition are
probably formed. A small apparatus of this type should be more
sensitive to surface-catalyzed ozone decomposition than full-
scale equipment with a lower surface-to-volume ratio.
A semi-log plot of some of the COD values against time is
shown in Figure 13. The form of these curves is similar to
those obtained for the batch reactors. They are not corrected
for the addition of untreated effluent to make up for the volume
of the samples taken during the runs. This raised the TOC and
COD above their proper values. Run 117 was made to check the
amount of this increase by taking samples only at the end of
the run so that nothing was added to the original charge. The
average final COD was 13 mg/1, 3 mg/1 lower than in the compara-
ble run 115 with normal sampling. TOC was 2 mg/1 lower.
For scale-up, alum or alum-acid coagulation is preferred,
unless a means can be found to prevent the formation of scale
deposits when using lime.
-42-
-------
0
10
OZONATION TIME (MIN)
20 30 40
50
60
50
f
^
i
30^
^ 20
o»
J
O
o
0 10
9
8
7
6
5
4
3
1 1
k
- Vj^fc^
1
1° " c
^^~
t
A* ==""^* n^ILI *
^"^C
RK^N COAGULANT
- o ||3 Ca(OH)2 132
- ° 123 ALUM 110
_ H2S04 85
- x HI ALUM 110
H2S04 63
_
A 127 Ca(OH)2 396
1 1 1
0123
STAGE
mg/l
mg/l
mg/l
mg/l
mg/l
mg/l
1
_
456
Figure 13. COD removals in the simulated six-stage cocurrent system.
-43-
-------
DESIGN CONCEPTS AND CRITERIA
Ozone Supply
The instability of ozone prevents its sale as a "packaged"
commodity; therefore, it has to be produced at the use point.
Many methods are available for the production of ozone, the
most common being silent electric discharge in oxygen or air,
photochemical conversion of oxygen or air, and electrolysis of
sulfuric acid. The photochemical method of ozone production
is used where small quantities in very low concentrations are
required. For large-scale production, the silent-electric-
discharge method is the only practical and economical method.
The choice of whether to use air or pure oxygen as feed to
a silent-electric-discharge ozonator is essentially a question
of economics. For a given ozonator with a constant power input,
approximately twice as much ozone is generated from oxygen as
from air1-4 . Thus, per Ib. of ozone produced, capital and elec-
tric power requirements would be considerably less with oxygen
usage. Against this saving must be charged the cost for oxygen.
For large ozone installations using oxygen, the reuse of uncon-
sumed oxygen becomes an economic necessity. In general, for
large ozone usage, the most economical system uses oxygen with
recycle. An additional benefit of the recycle system would be
the elimination of air pollution problems caused by the discharge
of spent gas containing unconsumed ozone.
A recycle system in a waste-water ozonation plant requires
an oxygen purification system. The recycled oxygen has a high
humidity and contains carbon dioxide (from the oxidation of organ-
ics) , nitrogen, and possibly hydrocarbons desorbed from the sec-
ondary effluent through which the oxygen has passed.
Studies conducted into the effects of various gaseous con-
taminants on yield of ozone per unit of electrical energy expen-
ded have shown that even minute quantities of water can cause
drastic drops in production efficiency '. The presence of
other contaminants have varying effects; i.e., hydrogen behaves
similarly to water, whereas carbon dioxide, argon, and hydro-
carbons decrease efficiency only gradually. The presence of
nitrogen or carbon monoxide increases efficiency when present
in concentration up to 10%, but causes a decrease in efficiency
at greater concentrations. The recommended practice is to sup-
ply the ozonator with oxygen or air dried to a dew point of
-60*C and with an organic content of less than 15 ppm by volume.
-44-
-------
Process Design
The optimum treatment of secondary effluent with O2one,
as evolved from our experimental program, consists of a pre-
treatment by clarification, if required, followed by deaeration
and a stagewise contacting with ozone for short periods. The
vessel of each contacting stage should provide the needed resi-
dence time for the reaction of ozone with organics.
The requirement for pretreatment will depend primarily on
the quality of effluent from the secondary plant and on the re-
quired quality of the ozonated water. For each case, an economic
comparison between pretreatment costs and the incremental in-
creases in ozone-treatment costs will determine the method to be
used. Our studies have indicated that pretreatment is preferred
for effluents having COD values above 40 mg/1.
Operating costs of ozone treatment reflect two major items
of expense: plant amortization and power costs. The oxygen
requirement is relatively fixed and is determined by COD removal
achieved and oxygen dissolved in the outgoing effluent. Labor
costs and maintenance are quite low. The key to economic water
treatment is high ozone-utilization efficiency. The design and
operational modes of the plant must be directed foremost to this
problem.
The continuous and spontaneous decomposition of ozone in
water requires that the ozone react with organics as quickly as
possible in order to achieve highest utilization efficiency.
Cocurrent contacting of ozone and effluent brings ozone and
organics together at their highest concentrations, resulting in
the maximum driving force for the useful consumption of ozone.
The chemical reaction between dissolved ozone and the organ-
ics in the effluent is a time-dependent process. Experimental
evidence has shown that about one hour is required to reduce the
COD content to an acceptable level of approximately 15 mg/1 from
an intake value of 35-40 mg/1; whereas, the half-life of ozone
in water is about 20 minutes. Consequently, all the ozone must
not be introduced at one time; staging is required. Ten minutes
has been selected as a reasonable residence time per stage. The
approximate removal of COD per stage that occurred in our experi-
ments is shown in Table XIII.
For maximum ozone utilization, the ozone supplied,per stage,
should,be equivalent to the amount that can be usefully consumed.
This latter amount depends on the COD levels and can be evaluated
with the high-shear batch reactor. Table XIII shows how the ozone
supply per stage should be reduced as a function of stage position,
Thus, the first stage should receive more ozone than the second,
etc.
Our experiments showed that overall rates in the two-phase
ozonation systems are dependent on the resistances both to mass
transfer and to chemical reaction with ozone, A significant
-45-
-------
TABLE XIII
l
j*
CTi
SIX-STAGE COCURRENT OZONATION:
COD REMOVAL PER STAGE
Stage
(10 min. per
Stage)
1
2
3
4
5
6
Total
Percent of Total COD Removal
High-Shear Batch Reactor
(Alum- Acid Clarification Runs)
51
19
12
9
6
3
100
Simulated Six-Stage
Contactor (Run 123)
52
16
12
8
8
4
100
-------
decrease in either one of these resistances would increase
overall rates significantly. High-shear contacting of effluent
with ozone in injectors achieved high mass-transfer rates and
appeared more economical than turbine agitation. Thus, the use
of injectors has been specified for the proposed plant design.
The holdup tanks incorporated in the system have a twofold
function: (1) as a secondary absorber for ozone from the gas
that was not transferred in the injector or the dissolver tube?
(2) as a time delay for the effluent to assure that the dissolved
ozone is consumed, thus giving maximum drive for mass transfer of
ozone in the next stage.
The pH of the solution affects the rates of spontaneous
ozone decomposition and of reaction of ozone with organics. The
higher the pH, the greater the rates. For optimum economic op-
eration, the ozonation rates must be balanced against the util-
ization efficiency of ozone. Our experiments indicate that a
slightly acid pH (pH 6.0 to 7.0) is optimum for multistage co-
current ozonation.
-47-
-------
PLANT DESCRIPTION
A schematic flowsheet of the proposed ozone-treatment
plant for waste water is shown in Figure 14. Secondary ef-
fluent, either direct from the secondary treatment plant or
from a clarifier, is pumped via a deaerator to the injector-
type contactor of the first stage. Oxygen containing ozone
is simultaneously fed into the injector to effect complete
mixing of the two phases and to induce high-shear contact
for improved mass transfer. The two-phase mixture is dis-
charged at the bottom of the first-stage holdup tank. The
liquid from the top of the tank is used as a feed for the
next stage. The ozone-depleted oxygen stream from the tower
is combined with gas streams from the other stages and is
recycled. The gas feed to all contactors is parallel and is
proportioned according to requirements. In general, the quan-
tity of ozone fed diminishes between each successive contactor;
thus, the first-stage contactor receives a greater quantity of
ozone than the next stage, etc. As explained before, the liquid
from the last-stage contactor is discharged without any further
treatment.
The recycled, oxygen, gas stream is compressed to about
10-15 psig (the pressure required to overcome losses in lines
and equipment) , is cooled to 35°F by a direct-contact chilled-
water spray tower, and is passed through regenerative gas
dryers (molecular sieve, alumina or silica gel) to dry the
oxygen to the required dew point of -60°F. The dried gas is
then sent to the bank of ozone generators to produce ozone at
1.7 wt.% concentration (22.6 mg/1). Makeup oxygen, as required,
is introduced into the recycle stream between the dryers and the
ozone generators.
The build-up of nitrogen and carbon dioxide in the recycle
gas stream has to be controlled. It was estimated that removal
of dissolved nitrogen in the secondary effluent feed by vacuum
deaeration would eliminate the need of a purge. Vacuum deaer-
ation could be carried out in a relatively simple vessel in
which a pressure of 2 psia is maintained by a small vacuum pump.
The expected oxygen consumption is given in Table XIV. An ex-
tra allowance of 0.12 Ib. of oxygen per 1000 gal. was made for
occasional purge requirements and leakage.
Most of the carbon dioxide produced in ozone treatment
should be discharged from the system as dissolved gas in the
liquid effluent. Further control of carbon dioxide levels in
the recycle stream can be effected by adsorption in the dryers,
if required.
The gas exiting from the contactors is expected to contain
residual ozone in quantities estimated at 10 to 25% of that in
the incoming feed. Commercial operation of ozonation systems
has shown that a portion of the residual ozone would pass through
-48-
-------
OXYGEN
PURGE
1
MUM
OZONATORS
DRYERS
CATALYTIC ft
OZONE-H
DECOMR u
CHILLER
RECYCLE
GAS
FEED
1
rj
DEAERATOR
-*kf^
T~
' f
1
I
r
1
_
i
n*
r
i
» LIQUID
EXIT
I 3E HI EC ^
STAGED CONTACTOR
~3L
FLOW RATES FOR 10 MGD PLANT
) EFFLUENT = 6,940 GPM
OXYGEN MAKEUP = 6,340 Ib/day
) OZONE GAS= 309 SCFM
FEED (26lb/min(a) l.7wt.%
OZONE)
GAS FEED TO STAGES
LINE
NO
4
5
6
7
8
9
FLOW
SCFM
157
59
37
28
19
9
a/
TOTAL
GAS
51
19
12
9
6
3
Figure 14. Ozone treatment plant
-------
TABLE XIV
OXYGEN CONSUMPTION IN LB. PER 1000 GAL. EFFLUENT TREATED
Oxygen Absorbed by Effluent
Oxygen Consumed in Lowering COD
Oxygen Lost by Purge and Leakage
Total
Vacuum
Deaeration
0.35
0.17
0.12
0.64
-50-
-------
recycle systems unaffected and would be available again at the
contactor. The proportion passing through the proposed recycle
system is unknown; and in the estimate, possible recycle ozone
has not been taken into account.
Alternate recovery methods for the residual ozone are
possible. Scrubbing of the exit gases by fresh effluent prior
to recycle is an obvious example. However, because of the low
ozone concentration in the exit gases, effective utilization or
alternate recovery of residual ozone would entail rather elabo-
rate and expensive mass transfer equipment. Savings resulting
from recovery would appear to be rather small compared to total
costs; therefore, no attempt was made to estimate these savings.
We realize that an in-depth study could reveal other potential
savings , and various recovery methods or destruction methods
must be studied for any commercial endeavor.
The holdup tanks were designed for a 10-minute retention
time per stage. As explained in the previous section, one of
the functions of each holdup tank is to act as a secondary ab-
sorber for ozone in the gas stream. In order to achieve effec-
tive gas scrubbing, gas distribution within the tank should be
uniform. The cross-sectional area of the tank should be main-
tained within reasonable limits. A height-to-width ratio per
injector of 2:1 was chosen as a reasonable value. Thus, an
injector placed in a tank with an 18-foot water depth will
cover an area approximately 9 feet wide. The 10-million GPD
contactor system, as shown in Figure 15, consists of six chambers,
each of which is 63 feet long x 18 feet deep. These chambers
represent the six stages. Each stage has seven injectors fed by
one pump. The area covered by each injector is 9 feet wide or
81 square feet.
-51-
-------
OZONE + OXYGEN IN
aGAS RECYCLE
EJECTOR AND /
DISSOLVER ~"K
TUBES \
A
EFFLUENT IN
EJECTOR PUMP ».*
STAGES
i
\
\
1
*o*
T
r
Jb.
X.
»o
0*
I,
i
f
It
i*
>
^ t
*o*
*o*
K>
o*
i.
h
T
'
\B
i*
(
*0*
*o*
K>
^>
T:
<
1
v»
f*
^
o*
ro
,,
i
i
\w
i*
5
^
*o*
*o*
o*
ro
\
L,
2:
e
i
-\ +
i*
^»i
»o*
o*
»0*
KV
0-
0-
»
A
t
PRO(
EFFL
OUT
GAS RECYCLE *
HOLD-UP TANKS
CTOR
LEVEL
? TUBE
XNKS -*
LIQUID
IN
»
~J
*
GAS
IN
jj
,
LIQUID
IN
)
GAS
IN
.J
.
LIQUID
IN
~
i
GAS
IN
iJ
,
LIQUID
IN
GAS
IN
-J
_
\
LIQUID
IN
,
GAS
IN
H
.
|St 2nd 3rd 4th 5th
STAGE STAGE STAGE STAGE STAGE
LIQUID II
-
3
GAS
r*
-
,
6th
STAGE
'
16
,
Figure 15. Schemotic representation of a 10-million gpd plant
for ozone treatment of secondary effluent
-52-
-------
CAPITAL COST
Capital cost requirements fo1' 02one-treatment plants of
three sizes (1-, 10-, and 100-million GPD) were estimated and
are presented in Table XV. The estimates are made for a COD
decrease from 35 to 15 mg/1, and an overall ozone-utilization
efficiency (COD/ozone efficiency index) of 80%. The plant
design is based on existing means of ozone manufacture and on
a direct scale-up of our laboratory findings.
Most existing secondary-treatment plants probably do not
discharge effluent with a COD content as low as 35 mg/1. Higher
COD loadings up to 250 mg/1 are common. Tertiary treatment of
these high-COD effluents by ozone, alone, would be too expensive.
Pretreatment by chemical coagulation and filtration, or in some
cases by filtration alone, can result in lower overall costs for
the tertiary treatment process. The decision for or against this
pretreatment would be based on a comparison of the incremental
cost increase of the ozone treatment plant versus the cost of a
pre treatment plant. The decision must be made on a specific-case
basis for existing, and proposed, secondary-treatment plants.
The average capital requirements for chemical clarification
plants of 1-, 10-, and 100-million GPD are shown in Table XVI.
The estimates of Tables XV and XVI are to be added to the
cost of existing secondary-treatment plants; thus, all further
treatment costs must be borne by the tertiary treatment plant.
Two extreme cases were considered: In the first case, the
effluent requires no pretreatment; in the second case, extensive
chemical clarification is required. In actual practice, other
possibilities could be considered; for instance, an incremental
increase in the secondary treatment facility to eliminate, or
appreciably reduce, the pretreatment requirement. For new
facilities, where it is anticipated that tertiary treatment will
be required, design and operation should be based on a total-
facility concept for minimum costs. For instance, incremental
increases in secondary treatment might eliminate tertiary pre-
treatment requirements; or conversely, incremental increases in
pretreatment (coagulation, filtration) of primary effluent could
totally eliminate the need for secondary treatment, as now prac-
ticed. The study of all these possibilities and their ramifica-
tions is beyond the scope of this program; hence, the simpler
case of tertiary treatment costs, as an added-on feature to an
existing facility, is presented.
It should be noted that the need for pretreatment prior to
a tertiary treatment of waste water is not unique with the ozona-
19
tion method. The study by Bishop et al. and the experience at
o c\
Lake Tahoe have shown that pretreatment is sometimes required
when activated carbon is used for tertiary treatment. The quality
requirements of effluent feed for either ozone or activated carbon
treatment are essentially equal.
-53-
-------
TABLE XV
CAPITAL COST ESTIMATES FOR OZONE-TREATMENT PLANTS
Cost Item
1. Ozone generators (1)
2. Recycle system (2)
3. Contactor (3)
4 . Piping and pumps
5. Facilities & buildings
6 . Vacuum deaeration system
Subtotal
7. Contingency and profit at 20%
8. Total
9 . Total in round figures
Capital Cost Per Item versus Plant Size, $
1-Million GPD
50,100
39,400
31,300
26,500
16,500
4,200
168,000
33,600
201,600
202,000
10-Million GPD
330,000
160,000
207,000
127,000
65,400
7,500
896,900
179,400
1,076,300
1,080,000
100-Million GPD
2,300,000
638,000
1,635,000
1,260,000
362,000
22,000
6,217,000
1,243,000
7,460,000
7,460,000
I
01
it*
I
(1) Cost based on data supplied by the Welsbach Corporation
(2) Recycle system includes compressor, water chiller, and chiller tower,
(3) Contactor includes injectors, dissolver tubes and hold-up tanks.
-------
TABLE XVI
CAPITAL COSTS FOR CHEMICAL CLARIFICATION PLANTS
Ln
Plant Size,
Million GPD
1
10
100
Capital Costs
jd/GPD Capacity
Including Pumping Stations
32.0
15.0
7.0
Treatment Plant Only
23.8
11.9
5.5
Treatment
Plant Only, $
238,000
1,190,000
5,500,000
-------
Appreciable potential capital savings could accrue in
ozone-treatment plants. Two obvious areas are ozone gener-
ators and ozone contactors. In view of the vast new market
that would open up if the ozonation process were to prove
practical, research in new techniques of ozone generation
could be justified by industry, and lowered ozone-treatment
costs might be expected. The depicted contactor design, with
its complexity and multiplicity of units, results in relatively
high costs. Improvements in design and layout can be antici-
pated as ozone-treatment plants are built and operated. A
capital cost estimate of the 10-million GPD plant, incorpo-
rating anticipated savings, is presented in Table XVII. The
savings were estimated to be 20% on ozonation equipment, 10%
on contactors, and 10% on pumps and piping. The savings
amount to 11% of the total capital costs.
-56-
-------
TABLE XVII
CAPITAL COST ESTIMATE FOR IMPROVED OZONE-TREATMENT PLANTS
OF 10-MILLION GPD CAPACITY
I
en
Cost Item
1. Ozone generators
2. Recycle system
3 . Contactor
4 . Piping and pumps
5. Facilities and buildings
6. Vacuum deaeration system
Subtotal
7. Contingency and profit at 20%
8. Total
9. Total in round figures
Capital Cost/Item
Existing
Technology, $
330,000
160,000
207,000
127,000
65,400
7,500
896,900
179,400
1,076,300
1,080,000
Savings
For Improved
Technology/ %
20
10
10
11
Improved
Technology, $
264,000
160,000
186,300
114,300
65,400
7,500
797,500
159,500
957,000
957,000
-------
OPERATING COSTS
Operating cost estimates have been divided into two parts,
corresponding to plants with and without pretreatment (Tables
XVIII and XIX). Included, for comparison, in the estimates of
plants without pretreatment are costs for an activated-carbon
treatment plant and for the ozone plant incorporating the im-
proved technology discussed in the previous chapter. The esti-
mates for the activated carbon treatment plant are those derived
from the operation of the Pomona Pilot Plant as reported by
Parkhurst et al .
Labor and maintenance costs for the ozone treatment plants
are based on the operating experience of the Belmont Water Treat-
ment Plant of the Philadelphia Water Department, as reported by
Bean . The operation of this ozone-treatment plant is very
similar to the ozone-treatment plants proposed herein, and oper-
ating costs should be similar. The maintenance cost at the
Belmont plant averaged $3,300 per year for a plant investment
of $1,000,000. This ratio of costs to investment was maintained
for six years. Operating labor costs at Belmont for the years
reported averaged $7,800. These costs were adjusted to 1968
levels by use of the Maintenance Cost Index reported in the
June 3rd issue of the Chemical Engineering Magazine and were
used for the ozone plants of l~and 10-million GPD. The 100-
million GPD plant was assumed to require twice the labor force.
Labor costs for plants with pretreatment were taken at twice
those required for plants with no pretreatment.
Electrical power required for ozone generation in the im-
proved technology plant was assumed to be 80% of that for exist-
ing methods. This resulted in an overall reduction in electrical
requirements of 11.4%. Oxygen costs were calculated from indus-
trial pricing schedules in current usage. The cost for coagulant
(alum) was taken at $57.00 per ton, and for sulfuric acid at
$32.00 per ton.
-58-
-------
TABLE XVIII
OPERATING COSTS OF TERTIARY TREATMENT PLANTS WITH NO PRETREATMENT REQUIRED
Plant Capacity >
Type Plant >
Capital Cost, $
Operating Cost, $/10 Gal.
Amortization, 15 yrs .
at 4%
Power (1)
Labor
Maintenance
Carbon Regeneration:
Power, Gas, Water
Make-up Carbon (10% Loss)
Oxygen
Total $/106 Gal.
Operating Cost,£/103 Gal.
1-Million GPD
Ozone
Existing
Technology
202,000
49.60
36.00
25.60
1.80
24.00
137.00
13.7
10 -Mi Hi on GPD
Ozone
Existing
Technology
1,080,000
26.50
28.80
2.60
1.00
17.80
76.70
7.7
Improved
Technology
957,000
23.50
25.50
2.60
0.90
17.80
70.30
7.0
Activated
Carbon
Existing
Technology
1,670,000
41.00
8.50
15.00
5.00
2.50
11.00
83.00
8.3
100 -Mil lion GPD
Ozone
Existing
Technology
7,460,000
18.31
21.60
0.50
0.67
8.00
49.08
4.9
I
01
(1) Power Rates = Ijzi/KWH for l-million GPD
0.8£/KWH for 10-million GPD
0.6£/KWH for 100-million GPD
-------
TABLE XIX
OPERATING COSTS OF OZONE TREATMENT PLANTS REQUIRING PRETREATMENT
Capital Cost, $
Operating Cost, $/10 Gal.
Amortization, 15 yrs. at 4%
Power (1)
Labor
Maintenance
Oxygen
Coagulant (Alum) (132 mg/1)
Acid (H2S04) ( 63 mg/1)
Total $/106 Gal.
Operating Cost, ^i/10 Gal.
1- Million GPD
Ozone
440,000
108.20
44.50
51.20
4.00
24.00
31.40
8.40
271.70
27.2
10- Million GPD
Ozone
2,270,000
55.70
37.30
5.20
2.00
17.80
31.40
8.40
157.80
15.8
100- Million GPD
Ozone
12,960,000
31.80
30.10
1.00
1.20
8.00
31.40
8.40
111.90
11.2
(1) Power Rates = 1/zJ/KWH for 1-million GPD,
0 . BjzVKWH for 10-million GPD
0.6/zf/KWH for 100-million GPD
-------
REFERENCES
1. Parkhurst, J.D., Dryden,F.D., McDermott, G.N., English, J.,
"Pomona Activated Carbon Pilot Plant", J. Water Polution
Control Fed. , 39^ (2), R70-81 (1967).
2. Morris, J.C., Weber, W.J.,Jr., Adsorption of Biochemically
Resistant Materials from Solution-2, 999-WP-33, AWTR-16,
pp 93-94, Federal Water Pollution Control Administration,
Cincinnati, O., 1966.
3. Smith, F.E., "Ozone Applications", Report No. I.G.E. 150,
Airco, Industrial Gases Div., Jersey City, N.J., Apr. 18, 1967
4. Toricelli, A., "Drinking Water Purification", Ozone Chemistry
and Technology, p 453, American Chemical Society, Washington,
D.C., March, 1959.
5. American Public Health Association, Standard Methods for the
Examination of Water and Wastewater, Twelfth Edition 1963-66,
p 366, APHA, New York, 1966.
6. Ibid, PP 510-514.
7. Ibid, p 193.
8. Kilpatrick, Mary L., Herrick, C.C., and Kilpatrick, Martin,
"The Decomposition of Ozone in Aqueous Solution",
J. American Chemical Society, 78, 1784 (1956).
9. Bailey, K.C., Retardation of Chemical Reactions, p 295,
Longmans Green, New York, 1937.
10. Kassel, L.S., Kinetics of Homogeneous Gas Reactions, p 264,
Chemical Catalog Co., New York, 1932.
11. Yost, D.M., Russell, H., Systematic Inorganic Chemistry,
p 265, Prentice-Hall, New York, 1944.
12. Leva, M. , Tower Packings and Packed Tower Design, Second
Edition, The United States Stoneware Company, 1953.
13. Koppe, P., Giebler, G., "Ozone Decomposition in Water",
Gass Wasserfach, 107 (8) , 196-200 (1966); Chemical Abstracts,
65_, 489 d (1966).
14. The Welsbach Corporation, Philadelphia, Pa., Bulletin 201,
May 1965.
-61-
-------
15. Cromwell, W.E.,Manley, T.C., "Effect of Gaseous Diluents
on Energy Yield of Ozone Generation from Oxygen", Ozone
Chemistry and Technology, p 304, American Chemical Society,
Washington, D.C., March, 1959.
16. Inoue, E. , Sugino, K., "Inhibiting Action of Hydrocarbons
on Ozone Formation by Silent Electrical Discharge", Ozone
Chemistry and Technology, p 313, American Chemical Society,
22
;h,
Washington, D.C., March, 1959.
17. Guinvarc'h, Pierre, "Three Years of Ozone Sterilization of
Water in Paris", Ozone Chemistry and Technology, p 416,
American Chemical Society, Washington, D.C., March, 1959.
18. Koenig, Louis, "The Cost of Water Treatment by Coagulation,
Sedimentation, and Rapid Sand Filtration", J. Am. Water Works
Assoc., 59_, 290-336 (1967).
19. Bishop, D.F., Marshall, L.S., O'Farrell, T.P., Dean, R.B.,
"Activated Carbon for Waste Water Renovation: II. Removal
of Organic Materials by Clarification and Column Treatment",
Paper presented at 149th National Meeting of the American
Chemical Society, April 6, 1965, Detroit, Michigan.
20. Slechta, A.F., Gulp, G.L., "Water Reclamation Studies at
the South Tahoe Public Utility District", J. Water Pollution
Control Fed., 39^ (5), 787 (May 1967).
21. Bean, E.L., "Ozone Production and Costs", Ozone Chemistry
and Technology, p 430, American Chemical Society, Washington,
D.C., March, 1959
-62-
------- |