Internal Report
Assessment of High-Rate
Sedimentation Processes:
Microcarrier Weighted
Coagulation Jar Tests
by
Yuan Ding, Robert Dresnack, and Paul C. Chan
Department of Civil and Environmental Engineering
New Jersey Institute of Technology
Newark, New Jersey 07102
Contract No. 7C-R364-NAFX
August 1999
Project Officer
Chi-Yuan Fan
Water Supply and Water Resources Division
National Risk Management Research Laboratory
U.S. Environmental Protection Agency
Edison, New Jersey 08837
-------
Disclaimer
The study described in this document had been funded wholly or in
part by the United States Environmental Protection Agency
(Contract No. 7C-R364-NAFX) and the New Jersey Institute of
Technology.
It has been subjected to the Agency's technical peer review.
Currently, the microcarrier weighted coagulation jar test
procedure is being further evaluated. It does not necessarily
reflect the views of the Agency, and no official endorsement
should be inferred. Also, the mention of trade names or
commercial products does not imply endorsement by the United
States government.
11
-------
Abstract
Past studies have identified that a significant amount of wet
weather flow pollutants is associated with colloidal and larger
particulate solids. These particles can play an important role
in water treatment and pollutant transport due to their large
specific surface area and high energies that facilitate the
sorption of significant quantities of substances. Since the
colloidal particles adsorb heavy metal and organic ions and water
borne microorganisms, removal of these particles is of paramount
importance in the water treatment process. In this process,
colloidal particles, coagulated with microcarriers (MC), can be
removed by a high-rate sedimentation process. The MC plays a
crucial role in enhancing settling properties, and in particular,
the removal of colloidal particles and associated contaminants.
A detailed testing procedure and a method of experimental
analysis using a modified jar test for the MC process have been
developed. A series of MC weighted jar tests were undertaken on
parking lot storm runoff, synthetic samples, and combined sewer
overflow mixed with a MC, coagulant (electrolyte) and coagulant
aid (polyelectrolyte). Two particle analyzers with a range of
0.002 to 5 micrometers (|im) and 0.1 to 2,000 |im, respectively,
were used to determine the full range of particle size
distribution. Different materials were used as the MC in this
study. The operational parameters being evaluated include
coagulant dosage, coagulant aid type and dosage, mixing- and
flocculation-induced hydraulic shear or velocity gradients and
duration, and characteristics of the MC. The pH, turbidity,
particle size distribution, total solids, total volatile solids,
suspended solids, and zeta potential were determined. The
experimental results reveal that the MC weighted coagulation
dramatically reduced coagulation (< 3min.) and settling time (< 8
min) producing high settling velocity floes and high quality
supernatant (turbidity from > 80 to < 2 NTU).
111
-------
Table of Contents
Number Pag
Disclaimer ii
Abstract iii
Table of Contents iv
Figures vii
Tables xi
Chapter 1 - Introduction 1
1.1 Background 1
1. 2 Purpose of Study 3
Project Objectives 3
1.3 Colloids Coagulation Analysis 4
Mass Transport Coefficient (/3) 4
Colloidal Stability (a) 5
Particle Interactions 5
Fractal Approach 6
Relevant Aspects 6
1.4 Organization of Report 6
Chapter 2 - Experimental Instruments and Testing Procedures 8
2 .1 Test Apparatus and Instruments 8
2.2 Testing Procedures 14
MC Weighted Jar Test 15
Particle Size Determination 16
Zeta Potential Measurement 20
iv
-------
Table of Contents (continued)
Number Page
Chapter 3 - Experimental Program 23
3.1 Sample Preparation 23
Surface Runoff 23
Combined Sewer Overflow
27
3.2 Measurement Parameters 27
Control Variables 27
Response Variables 28
3.3 Experimental Design 28
MC 29
Coagulant 29
Coagulant Aid 31
Experiments 35
Chapter 4 - Experimental Results: Surface Runoff 43
4.1 Prescreening Tests 43
4 .2 Effectiveness of MC Process 45
4.3 Screening Tests 49
Level One 49
Level Two 50
Level Three 51
Summary 52
4.4 Confirmative Tests 54
Chapter 5 - Experimental Results: Combined Sewer Overflow....87
5.1 Prescreening Tests 87
5 .2 Effectiveness of MC Process 87
5.3 Control Variable Determination 91
Coagulant Concentration 91
Coagulant Aid Concentration 94
MC Concentration 96
MC Size 96
Settling Time 99
5.4 Response Variable Evaluation 101
pH 101
Suspended Solids 102
Total Solids 103
Total Volatile Solids 103
Total Organic and Inorganic Carbon 104
Fecal Coliform 107
Relationship of Particle Count Rate and Turbidity ...107
Zeta Potential 112
v
-------
Table of Contents (continued)
Number Page
Chapter 6 - Summary and Recommendations 110
6 .1 Summary 110
6.2 Recommendations Ill
References 113
Appendix A - Quality Assurance Statement A-l
Appendix B - Zeta Potential Measurement Procedures B-l
Appendix C - Test Data C-l
VI
-------
Figures
Number Page
2-1. Particle Size Analyzer I 10
2-2. Particle Size Analyzer II with Zeta Potential Meter....11
2-3. Six Stirrer Jar Test Apparatus 12
2-4. Square Jars 13
2-5. Principles of Particle Size Analyzer I 17
2-6. Principles of Particle Size Analyzer II 18
2-7. Principles of Zeta Potential Measurement 22
3-1. Zeta Potential Distribution of Aluminum Sulfate Solution. . .30
3-2. Zeta Potential Distribution of Ferric Chloride Solution .... 30
3-3. pH Distribution of Aluminum Sulfate Solution 31
3-4. pH Distribution of Ferric Chloride Solution 31
3-5. Zeta Potential Distribution of POL-EZ-2466 33
3-6. Zeta Potential Distribution of POL-EZ-3466 33
3-7. Zeta Potential Distribution of POL-EZ-2696 34
3-8. Zeta Potential Distribution of POL-EZ-7736 34
3-9. A Comparison of Zeta Potential of Coagulant Aid 35
4-1. Particle Size of Raw and Treated Samples
(Cumulative) 45
4-2. Particle Size of Raw and Treated Samples
(Distributions) 46
4-3. Effect of the MC 43
4-4. Typical pH Distributions (Level-1) 56
VII
-------
Figures (continued)
Number Page
4-5. Turbidity Versus Coagulant Concentration
(Level-1) 57
4-6. Zeta Potential Distributions (Level-1) 58
4-7. Correlation of Zeta Potential and Turbidity
(Level-1) 59
4-8. Particle Count Rate for the Best Coagulant
Concentration (Level-1) 60
4-9. Typical pH Distributions (Level-2) 61
4-10. Summary of pH Distributions (Level-2) 62
4-11. Typical Turbidity Distributions (Level-2) 63
4-12. Summary of Turbidity Distributions (Level-2) 64
4-13. Total Solid Distributions (Level-2) 65
4-14. Total volatile Solid Distributions (Level-2) 66
4-15. A Comparison of Turbidity and Count Rate
(Level-2) 67
4-16. Zeta Potential Distributions (Level-2) 68
4-17. Correlation of Zeta Potential and Turbidity
(Level-2) 69
4-18. Summary of pH Distributions (Level-3) 70
4-19. Turbidity Distributions (Level-3; POL-EZ-2696) 71
4-20. Zeta Potential Distributions
(Level-3; POL-EZ-2696) 72
4-21. Correlation of Zeta Potential and Turbidity
(Level-3) 73
4-22. Turbidity Distributions (Level-3; POL-EZ-2466) 74
4-23. Zeta Potential Distributions
(Level-3; POL-EZ-2466) 75
4-24. Correlation of Zeta Potential and Turbidity
(Level-3) 76
Vlll
-------
Figures (continued)
Number Page
4-25. Count Rate Distribution (Level-3; POL-EZ-2466) 77
4-26. Total Solids Distributions
(Level-3; POL-EZ-2466) 78
4-27. Total volatile Solids Distributions
(Level-3; POL-EZ-2466) 79
4-28. Turbidity for Different MC Concenttrations
(small MC) 80
4-29. Turbidity for Different MC Concenttrations
(large MC) 81
4-30. Turbidity for Different MC Sizes (Low MC Dosage) 82
4-31. Turbidity for Different MC Sizes (High MC Dosage) 83
4-32. Turbidity Summary for Confirmative Tests
(by MC Group) 84
4-33. Turbidity Summary for Confirmative Tests
(by Coagulant Aid group) 85
4-34. Correlation of Particle Count Rate and Turbidity 86
5-1. Effectiveness of the MC Process 89
5-2. Particle Sizes of Raw and Treated Samples
(Cumulative) 89
5-3. Particle Sizes of Raw and Treated Samples
(Distributions) 90
5-4. Coagulant Concentration Selection (Test-1) 92
5-5. Coagulant Concentration Selection (Test-2) 93
5-6 . Particle Count Rate Versus Turbidity 94
5-7. Coagulant Aid Concentration Selection (Low Dose) 95
5-8. Coagulant Aid Concentration Selection (High Dose) 95
5-9 . MC Concentration Selection 97
5-10 . MC Size Selection 98
5-11. Turbidity Versus Settling Time with Optimal Condition ... 99
5-12. Turbidity Versus Settling Time 100
IX
-------
Figures (continued)
Number Page
5-13. pH Versus Settling Time with Optimal Condition 101
5-14. Suspended Solids Versus Settling Time 102
5-15. Total Solids Versus Settling Time 103
5-16. Total Volatile Solids Versus Settling Time 104
5-17. Total Organic and Inorganic Carbon Distributions
(with Different Raw Samples) 105
5-18. Total Organic and Inorganic Carbon Distributions
(with Different Mixing Duration) 106
5-19. Fecal Coliform Distributions 107
5-20. Particle Count Rate Versus Turbidity 108
5-21. Particle Count Rate Versus Turbidity (log-log) 108
5-22. Turbidity Distribution Versus Zeta Potential 109
x
-------
Tables
Number Page
1-1. Metal Distribution Versus Particle Size 2
2-1. Specifications of Major Instruments 9
2 -2 . Parameter Measurement Procedures 14
3-1. Preservation Condition and Holding Time For
Sample Analysis 24
3-2. Dry Sample Size Characteristics 26
3-3. Composition of Synthetic Samples 26
3-4. List of Coagulant Aids 32
3-5. Zeta Potential of Coagulant Aids 35
3-6. Parameter Setup for Prescreening Jar Tests 36
3-7. Determination of Rapid Mixing Rate with Duration
and Flocculation Rate with Duration 37
3-8. MC Identification 38
3-9. Screening and Confirmative Tests Parameter
Evaluation 38
3-10. Screening Tests -- Level 1 39
3-11. Screening Tests -- Level 2 40
3-12. Screening Tests -- Level 3 41
3-13. Confirmative Tests 42
4-1. Summary of Experimental Settings 44
XI
-------
Tables (continued)
4-2. Particle Size Distributions of Raw and
Treated Samples 47
4-3. Relationship of Turbidity and Zeta Potential 53
5-1. Summary of Experimental Settings for CSO Treatment 88
A-l. QA Objective for Precision, Accuracy, MDL, and
Completeness A-2
XII
-------
Chapter 1
Introduction
1.1 Background
Urban wet-weather flow (WWF), which includes sanitary sewer
overflow (SSO), combined sewer overflow (CSO), and stormwater
discharge (SWD), contains significant quantities of toxic
substances. WWF related toxic pollutants are major contributors
to the degradation of receiving waters. Urban WWF contains a
greater variety of toxic pollutants than sanitary wastewater.
Pollutants carried off urban catchments by drainage systems
during wet weather originate from many sources, e.g., commercial,
industrial, and residential parking areas; roadways; automobile-
service stations; sewer infiltration from leaking underground
storage tanks; accidents and spills; park and residential lawns;
construction sites; and active and inactive industrial sites.
Past studies indicate that SSO, CSO, and urban SWD contain
significant quantities of toxic substances; a number of the
hazardous-waste priority pollutants have been identified.
Without consideration of urban and industrial stormwater-runoff
toxic-substance control, the various hazardous-substances-cleanup
programs will not be effective in controlling total area wide
emissions of these substances.
Toxic-organic chemicals (e.g., benzene, polynuclear aromatic
hydrocarbons [PAH], polychlorinated biphenyls [PCB] , etc.) and
heavy metals (arsenic [As], cadmium [Cd], chromium [Cr], copper
[Cu], lead [Pb], mercury [Hg], and zinc [Zn]) in storm-induced
discharges contribute to receiving-water degradation (Pitt, et
al., 1995). The US Geological Survey reported that urban storm
runoff collected from residential, commercial, and industrial
areas around Phoenix, AZ was found to be toxic to fathead minnows
and water fleas (Lopes and Fossum, 1995). Industrial and
commercial parking lots, material storage areas, and vehicular
service stations are the most significant contributors of such
pollutants to WWF as reported by Pitt, et al. (1995) .
Sansalone et al. (1995) observed that lead, predominantly
associated with the particle fraction, was more mobile than
copper for highway runoff. Water quality in a creek below a
highway construction site indicated that the sediment
concentrations of total hydrocarbons, aromatic hydrocarbons, and
heavy metals and water concentrations of heavy metals and
selected anions increased downstream of roadway runoff (Maltby et
al. 1995). Hydrocarbon contamination of sediments was positively
correlated with potential contaminant loading functions (that is,
length of road drained/stream size). Boudries, et al. (1996) and
Estebe, et al. (1996) reported that heavy metals and aliphatic
and aromatic hydrocarbons bound to particles in the River Seine
sediments near Paris are due to urban WWF discharges. Such toxic
-------
substances in sediments create a long-term impact on ecological
systems.
Most of the solids finer than 50 micrometers (|im) are the principal
vector of pollution in urban stormwater (Chebbo et al., 1990) .
Table 1-1 that summarizes results from urban storm runoff
characterizations by Ellis and Revitt (1982) and Vignoles and
Herremans (1995) indicates the majority of heavy metals are
associated with particles less than 10 |im which are well into the
colloidal range. Thus, in order to effectively control toxic
heavy metals in urban stormwater, treatment processes that are
capable to remove fine particles (<10 |im) and to be able to handle
unsteady-nonuniform stormwater flow must be used. However, very
few of the commonly applied physical-chemical processes are cable
to remove such fine particles effectively.
Table 1-1. Metal Distribution Versus Particle Size
Suspended
solids Size
(micrometers)
> 100
10 -- 100
< 10
Metal Distribution (%)
Cd
18
36
46
Co
9
31
60
Cr
5
24
71
Cu
7
30
63
Mn
8
21
71
Ni
8
29
63
Pb
4
23
73
Zn
5
35
60
One approach of removing small particles in urban WWF is to apply
coagulant to promote colloid sorption during floe formation prior
to sedimentation than followed by filtration. These processes
achieve better solids removal than the plain sedimentation
(retention tank); but unsteady stormwater flow can detrimentally
affect the process efficiency. In recent years, micro-sand has
been used as weighted microcarrier (MC) in ballast-coagulation of
colloidal particles accelerating settling velocity. This is a new
high-rate physical-chemical clarification process. It was
originally designed for drinking water treatment and recently
being tested for treating wastewater and WWF. This process
consists of the addition of a coagulant in the influent pipe, MC
and coagulant aid (i.e., flocculant, polymer, polyelectrolyte) in
a mixing chamber, and than followed by maturation (flocculation)
and sedimentation tanks. The initiation of the reaction of
coagulation-flocculation processes is improved by the presence of
the MC and polymer, which increases the bonding of the floe to
-------
the MC, resulting in higher settling velocities. The MicrosepR
(U.S. Filter Corporation) and the ActifloR (Omnium de Traitement
et de Valorisation [OTV]) are two commercially available systems
(R)
that use recycled MC (e.g., microsand), while DensaDeg (Infilco
Degremont, Inc.) recycles its sludge. Recently, these processes
are being evaluated at an increasing number of pilot units for
treating CSO.
1.2 Purpose of Study
In view of the difference between MC and conventional
coagulation-flocculation processes, the existing bench-scale
testing procedures or standard Jar Test may not be adequate.
Thus, there is a need for developing a new set of bench-scale
testing procedures. The main purpose of this study is to develop
and evaluate new MC bench-scale (Jar Test) procedures for
engineers or plant operators to screen and select the effective
combination of type-dosage of coagulant, MC, and coagulant aid
for removing WWF colloidal particles. Determination of the
particle size distribution and zeta-potential of colloids enable
a better selection of coagulation-flocculation agents for the MC
weighted coagulation process. Results of the experimental study
herein may offer useful information to provide a framework for
further evaluation on MC weighted coagulation for subsequent
researchers.
Objectives
The objectives of this investigation are to:
Evaluate the applicability of conventional jar test
procedure to the MC weighted coagulation, and if needed,
modify the jar test procedure for screening MCs, coagulants,
and coagulant aids;
Test the effects of different types and dosages of coagulant
and coagulant aid in conjunction with MC for selection of
the most effective combination;
Investigate the effect of MC on particle size distributions
and zeta potential of colloids in urban WWF by using the
modified jar test procedure.
Presented below is a brief review related to the phenomenon of
coagulation.
1.3 Colloids Coagulation Analysis
Colloid coagulation is a complex process which depends on a
number of factors such as colloid type, particle concentration,
pH, coagulant concentration, particle size distribution, surface
area, surface charge, interfacial reactions and collisions
between suspended particles (Gregory, 1993). Mathematically, the
-------
coagulation process is known to be a binary process that has been
modeled as a second-order rate process (for example, O'Melia,
1993) : dnjdt=krf , where n is the concentration of particles in
suspension at time t, and k is a second order rate constant.
From a mechanistic point of view, coagulation depends on two
distinct influences: (a) particles must move in such a way that
collision occurs; and (b) interaction between colliding particles
must be such that permanent contacts can be formed. These two
factors are related to the rate constant mathematically as
k=a$, where a is a dimensionless sticking coefficient
associated with the colloidal stability while (3 is a mass
transport coefficient depending on the transport mechanics and
their particle interactions. Various approaches have been
proposed for determining these coefficients. In general, (3 is
evaluated from a micro-hydrodynamics standpoint and a is
determined indirectly from experimental results. It should be
pointed out that although micro-mechanisms of colloidal particle
coagulation have been studied extensively for chemical and
biological applications, interest in water and wastewater
treatment applications only began in the late 1960's. In any
case, the interest in environmental studies was provided by the
advances in environmental engineering, especially in the theory
of interparticle forces, coupled with the development of new
experimental techniques.
Mass Transport Coefficient (ft)
Suspended particle collisions occur in water due to velocity
variations caused by three different processes, namely, Brownian
diffusion, fluid shear, and differential settling. Generally,
particles less than 0.1 micrometer in diameter may be dominated
by Brownian diffusion while particles larger than about 5
micrometers may be transported by settling. Particles in the
range of 0.1 to 5 micrometers are too large for Brownian
diffusion and too small for settling.
By assuming that all particles move in straight lines until
contacts occur between them, a rectilinear model was established
by Smoluchowski. The expressions of (3 for each of the three
different transport mechanisms have been determined analytically
(e.g., Clark, 1996). This model is considered to be the most
fundamental approach in colloidal coagulation analysis. However,
the rectilinear model neglects the hydrodynamic influence and the
short-range interactions between approaching particles.
Moreover, as aggregates grow in size, transport mechanisms
change, aggregate morphology is altered, and breaking up by
hydrodynamic forces can occur. Recently, questions were raised
with regard to the validity of the rectilinear modeling approach.
Han and Lawler (1991) modified this model using a curvilinear
track for particle motions by adding hydrodynamic considerations.
-------
However, interparticle reactions have not been included in the
analysis.
Colloidal Stability (a)
Given the possible collisions between particles, stable particles
have to be transformed into unstable ones in order to complete
the coagulation process. Experiments to determine the colloidal
stability in aquatic systems are scarce. Theories about the
origins of colloidal stability in aquatic systems are few and
qualitative and none has been tested experimentally. The present
status is described well by Professor O'Melia: "Theories of
colloidal stability are helpful in understanding why particles
are stable and in identifying important chemical properties of
solutions and solids that affect stability, but they are not able
to provide accurate quantitative predictions of coagulation rates
when chemical repulsive interactions produce low attachment
probabilities (low alphas)." (O'Melia, 1993).
Particle Interactions
The earliest and still the only quantitative analysis of colloid
stability is the DLVO theory, developed independently by
Derjaguin and Landau (1941) and Verwey and Overbeek (1948).
Essentially, they combine the Van Der Waals attraction (caused by
dipole moments in the constituent molecules of two approaching
colloidal particles) and the electrical double layer repulsion to
give the total energy of interaction between particles as a
function of the separation distance. All other types of
interactions are ignored. In addition, the theory is dependent
on factors such as colloid type, particle concentration, pH and
coagulant concentration.
Analytical formulation of the double layer was developed
independently by Gouy and Chapman who derived the fundamental
equation relating electrical potential to charge in the diffuse
layer (Russel, et al, 1989). The potential at the plane of
shear within the double layer is known as the Zeta potential. It
depends on the thickness of the double layer and its value
determines the extent of the electrostatic forces of repulsion
between charged particles. The seta potential is a useful and
important tool in providing information on the optimum coagulant
dosage for the microcarrier process.
Fractal Approach
The problem of characterizing particle size becomes somewhat
complicated for natural particles and their random aggregates.
Fractal theory has been explored to study this process (e.g.,
Lin, et al., 1989; Meakin, 1988). Since random aggregates
contain greater surface area than any equivalent sphere or other
-------
shape, the aggregate is more of a 'tenuous' surface (two-
dimensional) rather than a three-dimensional object. For a large
number of aggregates, if the mass is plotted against aggregate
size on a log-log scale, the plot may be linear with a non-
integer slope. The relationship between aggregate mass (M) and
size (L) can be expressed as M = Ld , where the 'slope' (d) of
the line is called the fractal dimension. For a regular two or
three-dimensional object, the slope is either two or three,
respectively. The lower the fractal dimension, the more open the
aggregate structure. As the coagulation of solid particle
proceeds, fluid is incorporated into pores in the aggregates that
are formed. Aggregate density decreases and total aggregate
volume increases as the process continues. The result is that
the target cross sections or collision diameters of the
aggregates increase thereby increasing the rates of interparticle
contact brought about by Brownian diffusion, fluid shear and
differential sedimentation. Observations of natural and
technological systems indicate that the aggregates in these
systems are fractals and have fractal dimensions less than three.
Relevant Aspects
In the past, considerable attention has been given to describing
airborne-particle capture in flow past simple collector
geometries, especially cylinders (Pruppacher and Klett, 1978;
Wen, 1996). There are similarities between the capture of gas-
borne particles and liquid-borne particles based on computed
forces from fluid mechanics theory. However, important
dissimilarities exist for kinds and magnitudes of other forces
not based upon the field of fluid mechanics such as electrostatic
force. Other factors such as different molecular mean free paths
for gases and liquids also play an important role in the
coagulation process.
1.4 Organization of Report
This report consists of six chapters and two appendixes with
associated references. First, Chapter I presents a background
review on the MC process followed by statement of the primary
objectives of the study and a brief review of colloid coagulation
phenomena of colloids related to water treatment. Chapter 2
describes the major apparatus, instruments and testing procedures
used in this study. The experimental program is presented in
Chapter 3, which includes sample preparation and three stages of
experimental design, namely, prescreening, screening, and
confirmative tests. Chapters 4 and 5 present experimental
results for surface runoff and CSO jar-tests, respectively.
Summary and recommendations are discussed in Chapter 6. Appendix
A presents a quality assurance statement. Appendix B describes
the zeta potential measurement procedures recommended by the
manufacturer. In addition, experimental results are presented in
Appendix C.
-------
Chapter 2
Experimental Instruments and Testing Procedures
2.1 Test Apparatus and Instruments
The descriptions for the major test apparatus and analytical
instruments used for the investigation, including the measured
parameters, range, model number, and manufacturer, are listed in
Table 2-1 and briefly described as follows:
Particle size analyzer I, designated as PSA-I, is a large
scale particle size analyzer with a measurement range from
0.1 to 2000 |im. Figure 2-1 is a photograph of the PSA-I
instrument.
Particle size analyzer II, designated as PSA-II, is a
small scale particle size analyzer with a measurement
range from 0.002 to 5 |im. PSA-II is equipped with
disposable sample cells that eliminates the cross sample
residue influence. Figure 2-2 is a photograph of PSA-II
instrument.
Zeta potential meter. The zeta potential meter and PSA-II
are integrated in one unit. The measurement range of the
zeta potential meter is from 0.1 to 200 mV with the particle
size range from 0.002 to 30 |im. The resolution is sample
dependent and in the range of 0.1% to 5%.
Jar test apparatus: A Phipps and Bird (Model PB-700 as
shown in Figures 2-3 and 2-4) was used. Dimensions of each
jar are 11.5 X 11.5 X 21 cm depth which is capable for
testing a volume of 2,000 ml water sample. Each jar is
equipted with a flat stirring paddle (7.6 X 2.5 cm or 19.3
cm2) . For MC jar test, the area of the paddle was increased
to 38.7 cm2 for MC jar test in order to generate more
rigorous turbulence for keeping micro-sand in suspension.
Turbidity meter.
pH meter.
Balance.
-------
Table 2-1. Specifications of Major Instruments
Apparatus/
Instrument
PSA- 1
PSA- I I
Zeta
potential
meter
pH meter
TOC
Analyzer
Stirrer
Turbidity
meter
Balance
Measurement
Parameter
Particle
size
Particle
size
Zeta
potential
pH
TOC
Jar Test
Turbidity
Weight
Measurement
Range
0 . 1 2 000 |im
0.002 5 |im
0.1 200 mV
1 14
4 10000 ppb
0 300 rpm
0 1000 NTU
0 210 grams
Model No .
Master-
Sizer X
90 Plus
combined
with
ZetaPlus
700 TOC
PB-700
Jar
tester
DRT-15CE
XS-210
Manufacturer
Malvern
Instruments Inc.
Southbo rough, MA
Brookhaven
Instruments
Corporation
Holtsville, NY
O.I. Analytical
College Station,
TX
Phipps & Bird
Richmond, VA
HF Scientific
inc. Fort Myers
FL
Denver
Instrument Co.
Arvada, CO
-------
Figure 2-1. Particle size analyzer
10
-------
Figure 2-2. Particle size analyzer II with zeta potential meter
11
-------
Figure 2-3. Six stirrer jar apparatus
12
-------
2.2 Testing Procedures
With the exception of zeta potential measurement, all parameter
measurement procedures were based on USEPA procedures or Standard
Methods. The measurement procedures are summarized in Table 2-2.
Table 2-2. Parameter Measurement Procedures
Parameter
Particle
Size
Zeta
Potential
MC
Weighted
Jar Test
pH
Volatile
Solids
Turbidity
Suspended
Solids
Total
Solids
Sample Type
Stormwater
Microcarrier
Stormwater
Stormwater
Microcarrier
Stormwater
Stormwater
Stormwater
Stormwater
Stormwater
Method
No.
2560
Appendix
B
1-1
150.1
160.4
180.1
160.2
160.3
Method Title
Particle count
and size
distribution
Zeta potential
measurement
Coagulation and
flocculation
pH
(electrometric)
Residue,
Volatile
Turbidity
(Nephelomet ric )
Residue, Non-
f ilterable
Residue, Total
Reference
Standard
Methods (1)
Manufacturer
AEEP(2)
(modified)
EPA(3)
EPA(3)
EPA(3)
EPA(3)
EPA(3)
Standard Methods, 18th edition supplement (1995) . Environmental
Engineering Unit Operations and Unit Processes Laboratory Manual, Association
of Environmental Engineering Professors (1971). Methods for Chemical
Analysis of Water and Wastes, EPA-600/4-79-020 (1983) .
13
-------
Figure 2-4. Square jars
14
-------
MC Weighted Jar Test
The jar test has been used as a method for evaluation of the
effectiveness of coagulants and coagulant aids for removal of
solids in water treatment for many decades. Detailed jar test
procedure can be found in numerous publications (Cohen, 1957;
Black et al., 1957, 1969; Camp, 1968; AEEP, 1971; ASTM, 1996).
However, the MC weighted jar test is a new application due to the
different physical characteristics of the mixture.
In the MC weighted jar tests, water samples of equal volume
(1,000 ml) were poured into a series of six 2-litter square
beakers on a multiple stirring machine equipped with a variable
speed drive. After precalculated dosages of the microcarrier,
coagulant, and coagulant aid (i.e., flocculant, polymer, or
polyelectrolyte) had been added to the beakers, the contents were
rapidly stirred to simulate flash mixing and then reducing
stirred to simulate flocculation. After a given period of time,
the stirring was stopped and the floe formed was allowed to
settle.
During the process, illumination aids were used in watching floe
formation; however, heating effects from the light were avoided.
The controlling parameters are enumerated as follows:
1. The volume of the sample.
2. The size and shape of the container.
3. Peripheral speed and time of rapid mixing.
4. Peripheral speed and time of slow mixing.
5. Type and dosage of microcarrier, coagulant and coagulant
aid (i.e., flocculent, polymer, or polyelectrolyte).
The principal procedures include the following steps:
1. Collect storm surface runoff sample, prepare synthetic
sample (see Section 3.1), or CSO sample. Measure the
sample for pH value and turbidity reading.
2. Pour 1,000 ml of the water sample into each two-liter jar
on the jar-test apparatus and check stirrer operation. A
light table facilitates viewing of the contents of the
beakers.
3. Add controlled amounts of MC, coagulant, and flocculant
dosage to the designated jars.
4. Flash mixing for 2060 seconds at 100200 rpm.
5. Slow mixing for 10120 seconds at 3060 rpm. Record the
elapsed time before a visible floe is formed. If large
floes are formed, it may be desirable to reduce the paddle
speed. Record the appearance of the floe formed.
6. After flocculation, remove the paddles and settle for 230
minutes.
15
-------
7. Collect the supernatant from the sampling port on each jar
and measure the turbidity; the settled solids should not
be disturbed during sampling. Select and recored the
dosage of coagulant and flocculant based on the
supernatant clarity and settleability of floe.
Particle Size Determination
Principle
Both particle size analyzers (PSA-I and PSA- II) used in this
study are based on light-scattering techniques using a Helium-
Neon laser as the light source. However, the signal collection
and conversion for the two instruments are different.
In the PSA-I system, the direct path of the light beam through
the flow cell is scattered by a particle as it flows through the
measurement zone with the fluid (see Figure 2-5) . Scattered light
over a fixed range of angles is collected by a photo-voltaic
cell. Based on the principles of Fraunhofer diffraction,
particle size can be determined from the angle (6) and intensity
(I) of scattering as follows:
) 2na
sin 9 A
where a is the particle radius; Jl is the first order Bessel
function; and A is the wavelength. For multi-particles, the
resulting responses from all particles are collected and
mathematically deconvoluted to generate the size distribution.
In the PSA-II system, the scattered light is collected at a 90
degree angle to the light source (see Figure 2-6) . The photon
correlation spectroscopy of quasi-elastically scattered light
technique, based on correlating the fluctuations about the
average scattered light intensity, is the measurement mechanism.
The total measurement time is divided into small intervals called
delay times. These intervals are selected to be small compared
16
-------
(S
8
03
M
M
«
N
N
H
rH
O
-H
*>
H
«
01
«H
o
Oi
H
O
fl
H
14
Ol
in
ot
3
0)
H
17
-------
a
H
fr*
(D
n
43
wO
d
-------
with the time it takes for a typical fluctuation to relax back to
the average. The scattered light intensity in each of these
intervals, as represented by the number of electrical pulses
registered during each delay time, fluctuates about a mean value.
The intensity auto-correlation function is formed by averaging
the products of the intensities in these small time intervals as
a function of the time between the intervals (delay times) . As
the delay time increases (t) , the correlation (c) decreases and
the function approaches the constant background term B. In
between these two limits the function decays exponentially for a
monodisperse suspension of rigid, globular particles and is given
by
where A is an optical constant determined by the instrument
design, and F is related to the relaxation of the fluctuations
by
where D is the transitional diffusion coefficient. The value of
q is calculated from the scattering angle (9 = 90 degrees), the
wavelength of the laser light (A = 0.635 |im) , and the index of
refraction (n) of the suspending liquid. The equation relating
these parameters can be expressed as
Inn
« =
For a sphere, there is
kT
d =
37T77D
where d is the particle diameter; k is Boltzmann's constant; T is
the temperature; and 77 is the viscosity of the liquid in which
the particle is moving. The above equation is based on the
assumption that the particles are moving independently of one
another. In case a particle is not spherical, the d calculated
from the above equation is considered as a particle size
indicator.
Measurement Procedure
The principal steps for particle size distribution measurement,
in accordance with the Standard Methods For Examination of Water
and Wastewater (Standard Methods, 1995) , are enumerated as
follows:
1. Preparation. The instrument and any sample handling unit
should be switched on and any connections between the
optical unit, sample handling unit and computer should be
in place. The correct range lens should be fitted to the
19
-------
instrument and the lens caps removed. Any sample cell
should be correctly fitted and the windows should be
clean. In particular, the correct instrument range should
be selected.
2. Background measurement. A background measurement is
necessary before any sample measurement.
3. Blank sample measurement. Measure at least one blank
sample of particle-free water.
4. Calibration. Calibrate by determining the channel number
into which particles of known size are sorted by the
instrument. Use spherical particles manufactured for this
purpose. Use three sizes of calibration particles in
similar concentrations to calibrate a sensor. Calibrate
under conditions identical with those of the sample
measurement, e.g., settings on the instrument, flow rate,
and type of sample cell.
5. Measurement of samples. The light scattered by the
particles must be measured for a suitable period to ensure
that all particles are represented in the measurement and
to average out fluctuations caused by the dispersing
medium. A suitable measurement period is 10 to 30 seconds
depending on the size range of the distribution.
6. Data reporting. Particle concentrations should be shown
in both tabular and graphical formats.
In the course of experiments, it was found that the large
particles (> 5 |im) produced interferences during the measurement
of small particles (< 5 |im) . Furthermore, the measurement of
small particles was found to be inconsistent in the presence of
large particles, even if a low concentration of large particles
existed. In order to eliminate the interferences from large
particles, a special filtration process was necessary before
measuring the small particle size. An attempt was made with
different types as well as different pore size filters. It was
concluded that filter paper (regardless of type) was not suitable
for this experiment. Having experimented with other filtering
processes, a disposable nylon syringe filter with 5 |im pore size
was found to be suitable for the study.
Zeta Potential Measurements
Principle
The basic principles of zeta potential measurements include three
different aspects. First, the velocity (V) of charged colloidal
particles in liquids between the electrodes is measured by a
laser Doppler shift. Second, the electrophoretic mobility (/2)
is determined based on the measured velocity and the electric
field strength (E) by the equation V = jjE . The zeta potential (£)
can be calculated from the solution conditions and the mobility
20
-------
by the equation ^ = ^ri/e where 77 is the viscosity of the liquid
and £ is the dielectric permittivity. However, the equation is
only correct for certain combination of particle size and ionic
strength. Either mobility or zeta potential may be used as
measures of dispersion stability although zeta potential is used
more widely. The principles of zeta potential measurement are
illustrated in Figure 2-7.
Measurement Procedure
A detailed testing procedure that is recommended by the
manufacturer is given in Appendix B. Major steps are outlined as
follows:
1. Instrument preparation.
2 . Background measurement.
3 . Blank sample measurement.
4. Calibration.
5. Sample preparation.
6. Clean the electrodes and insert the electrode assembly
into the cell.
7. Insert the cell into the cell compartment.
8 . Make a measurement.
21
-------
s
"3
m
d
u
OJ
w
4J
1
I
O
fc
SB
N
ft
to
22
-------
Chapter 3
Experimental Program
3.1 Sample Preparation
Past studies have identified urban stormwater runoff as a major
contributor to the degradation of many urban lakes, streams, and
rivers. Industrial and commercial parking lots, material storage
areas, and automobile service stations are the most significant
contributors of a variety of pollutants to wet weather flow.
Therefore, an attempt was made to analyze both stormwater runoff
and synthetic samples. The former was intended to provide a
preliminary guideline in parking lot runoff characteristics,
while the latter allowed a consistent background for
experimentation. Preparation and evaluation procedures for each
type of sample are described in this section. In addition, CSO
samples were also used for this study.
Surface Runoff
During the initial stage of the investigation, surface runoff
samples were collected from a parking lot. In addition to
natural storm surface runoff samples, simulated runoff was
generated by spraying the above-noted parking lot with city water
during dry weather periods.
The parking lot used in this study was the Otto H. York Center
Parking Lot #3 at the New Jersey Institute of Technology (NJIT)
campus in Newark, New Jersey. The parking lot has a capacity of
60 cars with a dimension of approximately 100 meters by 30 meters
with a 4% grade. It operates seven days per week with five days
at full capacity.
Sampling of actual stormwater runoff events is weather dependent,
and hydrologic factors such as the interval between storm events,
rainfall intensity, total rainfall, etc. are random in nature,
and, as such, greatly affect runoff characteristics. Because of
this, the time to conduct the experiments and the ability to
replicate results with varying samples would have required a
substantially longer period than is programmed in the
investigation. Therefore, actual runoff samples were used only
in the preliminary stage of this study to test the feasibility of
the MC process.
Sampling.
Bulk runoff samples generated by natural rain or by spraying a
predetermined area of the parking lot with city water were
sampled at the beginning of the precipitation event directly from
the drainage inlet chamber by a hand-bucket and industrial vacuum
apparatus. The pavement was washed from the highest elevation to
the lowest elevation. The total area washed for each of the
23
-------
tests was approximately 3,000 square feet. Sampling setup and
operations were prepared at least two days prior to rainfall
events anticipated in accordance with the weather forecast.
Sample Preservation, Transportation, and Storage.
Samples were transported to the laboratory within five minutes
after finishing the sample collection process. The preservation
and holding time of samples are listed in Table 3-1.
Table 3-1. Preservation Condition and Holding Time
For Sample Analysis
Parameter
Particle size
distribution
Zeta potential
Suspended solid
Total solid
Volatile solids
pH
Fecal coliform
Total organic
carbon
Turbidity
Container
P, G*
P, G
P, G
P, G
P, G
P, G
P, G
P, G
P, G
Preservation
Cool, 4°C
Cool, 4°C
Cool, 4°C
Cool, 4°C
Cool, 4°C
None
Cool, 4°C
Cool, 4°C
Cool, 4°C
Holding Time
48 hours**
48 hours**
7 days
7 days
7 days
Immediately
6 hours
28 days
48 hours
*
* *
P -- plastic container; G -- glass bottle
No standard holding time is given in the currently available
literature. This number is selected based on turbidity
holding time.
Synthetic Samples.
As indicated above, evaluation of the MC process by both natural
and simulated runoff requires a substantial amount of sampling
due to the randomness of runoff samples. Since the dry residual
materials from a parking facility also possess the basic
constituents of parking lot runoff, such residues were collected
24
-------
and used to prepare synthetic samples. It was found that two
types of dry materials existed:
Type I -- Low organic content materials. The type I material is
a low organic content sandy-like residue that was found in the
area along the walls of the parking deck. The total volatile
solid content of this type of material was found to be less than
1% by weight. This material was sieved and divided into the
seven particle size ranges as shown in Table 3-2.
A series of sample preparation evaluation tests were conducted to
determine the quantity of each range for the aqueous sample to be
analyzed. It was found that the material with grain size larger
than 106 |im would settle too fast to make a uniform sample. It
was concluded that particles smaller than 106 |im were more
suitable for suspension than the larger particles. Subsequently,
particles smaller than 106 |im were used to prepare synthetic
samples. The constituents of the synthetic samples are
illustrated in Table 3-3.
25
-------
Table 3-2. Dry Sample Size Characteristics
Particle Diameter
(|im)
less than 53
53 75
75 106
106 150
150 250
250 425
larger than 425
Control Sieve No.
(ASTM)
270
270 200
200 140
140 100
100 60
60 40
40
Table 3-3. Composition of Synthetic Samples
Material
Type
Type I
Low organic
content
material
Type II
with
Clay
Total
Specification
size < 53 (|im)
53 < size < 75
(|im)
75 < size < 106
(|im)
High organic
content
0.1 < size < 3
(|im)
Relative
Weight*
38.9%
19.4%
8.3%
1.9%
31.4%
100%
Remarks
By sieve analysis
High organic content
materials were ground
with clay
* based on the total weight of dry sample
26
-------
Type II -- High organic content material. The type II material
is a high organic content residue that was attached to the ground
surface under each parked car. The total volatile solids for
this type of material were found to be approximately 18% by
weight. In performing sample preparation tests, it was observed
that this material either settles at the bottom of the jar or
floats on the surface. In order to obtain suspended volatile
solids, the high organic content material was first ground with
clay and then used to prepare an aqueous sample. After grinding
with clay, it was observed that the type II solids remained in a
suspended state, which was confirmed by total volatile solid
testing. A series of tests were conducted to determine the
proportion between clay and the original sample. Both low and
high organic content materials were combined and used to generate
a synthetic sample. The composition of the type II material
ground with clay is shown in Table 3-3.
Combined Sewer Overflow
The City of Perth Amboy operates a combined sewer system and
wastewater transfer pumping station that collects combined
sanitary sewage, industrial wastewater, and storm runoff from an
approximately 7 square kilometers drainage area to a regional
wastewater treatment plant owned and operated by the Middlesex
County Utilities Authority. The wastewater transfer pumping
facility is located at the junction of Water Street and Sadowski
Parkway. A CSO regulator is directly located about 20 feet below
the Sadowski Parkway with an overflow weir and 84" diameter CSO
tide-gated outfall to the Arthur Kill.
The inflow from the interceptor discharges into one of two
screening channels each equipped with a mechanics coarse bar
screen for removing screenings and protecting the sewage pumps.
The screen chamber inflow was collected for the MC coagulation
study.
The channel is housed in an enclosed building with 24 hours a day
access and a person always on duty. Grit, that accumulates in the
channel, is removed from the facility monthly.
3.2 Measurement Parameters
In this experimental program, the parameters were divided into
control and response variables.
Control Variables
Control variables are independent variables of a system whereas
response variables are dependent variables (results). In this
study, the control variables included the following parameters:
27
-------
- MC type, size range, and concentration.
- Coagulant type and concentration.
- Coagulant aid (coagulant aid) type and concentration.
- Rapid mixing time and rotation rate.
- Slow mixing (flocculation) time and rotation rate.
Response Variables
Response variables are indicators used for determining the
effectiveness of the control variables. Response variables may
provide useful information in determining the optimal setup for
control variables. In this study, the following parameters were
used as response variables:
- Supernatant turbidity.
- Supernatant pH.
- Supernatant particle-size-distributions.
- Supernatant zeta potential.
- Supernatant suspended solids.
- Supernatant total solids.
- Supernatant volatile solids.
- Supernatant fecal coliform.
- Supernatant total organic carbon.
Raw Stormwater Characterization. The raw samples including
surface runoff, synthetic samples, and CSO (see Section 3.1) were
characterized by the following parameters:
pH Total volatile solids
Turbidity Zeta potential
Total solids Total organic carbon
Particle size distribution Fecal coliform
Suspended solids
3.3 Experimental Design
The experimental design of this study consisted of a three-layer
experimental design: prescreening tests, screening tests, and
confirmative tests. The purpose of the prescreening tests during
which different experimental conditions were evaluated by visual
observation was to determine the range of operational parameters.
Upon the completion of the prescreening tests, screening tests
were performed followed by confirmative tests. The purpose of
screening and confirmative tests was to provide a quantitative
analysis of the MC process. Detailed descriptions of MC,
coagulant, coagulant aid, prescreening tests, screening tests,
and confirmative tests are given in the following sections.
28
-------
Microcarrier (MC)
Two types of material were selected to use as MC, namely, Ottawa
sand and a beach sand obtained from Sandy Hook, NJ. Both sands
were used in prescreening tests. Ottawa sand was selected in
screening and confirmative testing, due to its durability and
size uniformity over the Sandy Hook beach sand and commercial
available in large quantity. The size range of Ottawa sand tested
was between 100 to 500 |im. The MC size ranges and concentrations
were determined by prescreening tests.
All containers and MC were first washed thoroughly with a
detergent and hot water, then rinsed with hot water to remove all
traces of residual washing compound, and finally rinsed with
particle-free water.
Coagulant
Alum (aluminum sulfate, A12 (S04) 3»18H20) and ferric chloride
(FeCl3»6H20) were used as the coagulants for this study. Aluminum
sulfate has been employed extensively in water and wastewater
treatment because it is usually less expensive than other
coagulants and it operates effectively close to neutral pH's
while ferric chloride is effective over a wider pH range.
Concentrations of coagulant reported in this report are the
concentration of aluminum sulfate as mg of A12 (S04) 3»18H20/L and
the concentration of ferric chloride as mg of Fe+++/L-
The zeta potential and pH of aluminum sulfate and ferric chloride
were measured at different concentrations. Stock coagulant test
solutions were prepared daily by mixing chemicals with deionized
water to a concentration of 10 g/L (1 mL of stock solution when
added to 1 L of sample is equivalent to 10 mg/L). The zeta
potential and pH values of the deionized water were close to 0 mV
and 7, respectively.
The zeta potential distributions of aluminum sulfate and ferric
chloride solutions for different concentrations are presented in
Figures 3-1 and 3-2. The zeta potential increases with the
increase of concentration within the range of 10100 mg/L for
both coagulants. For concentrations higher than 100 mg/L, no
significant changes for zeta potential was noticed.
29
-------
10
100
A Ijm Jnum sulfate cxDncsatatbn tn g/L)
-H
1000
Figure 3-1. Zeta Potential Distribution of Aluminum Sulfate Solution
10 100
Ferric chloride cnnaenlzatiQn (as Fe+ + + m g/L)
1000
Figure 3-2. Zeta Potential Distribution of Ferric Chloride Solution
Figures 3-3 and 3-4 illustrate variation of pH values with
different concentrations of coagulant. The pH value decreases
from 5.2 to 3.5 for aluminum sulfate concentrations increases
from 10 to 1000 mg/L as A12 (S04) 3»18H20, and the pH values vary
from 3.5 to 2.3 for ferric chloride solution when the
concentration of ferric chloride increases from 10 to 1000 mg/L
as Fe+++ (or 48 to 4,800 mg/L as FeCl3»6H20) .
30
-------
a
5.5 -
- - X
4.5 --
4 --
3.5 --
3
10
100
A Ijm dnum Sulate cxDnaentatbn tn g/L)
1000
Figure 3-3. pH Distribution of Aluminum Sulfate Solution
3.5
a
3 --
2.5 --
10
100
Ferric chloride cxDncsatation (as Fe+ m g/L)
1000
Figure 3-4. pH Distribution of Ferric Chloride Solution
Coagulant Aid
There are many commercially available coagulant aids or
polyelectrolytes. Five polyelectrolytes from two different
manufacturers (see Table 3-4) were used in the experiments.
Among them, four (POL-EZ-2466, POL-EZ-3466, POL-EZ-2696, and POL-
EZ-7736) were used in surface runoff tests and one (309C) for CSO
tests. Zeta potential distributions versus polyelectrolytes
concentrations are illustrated in Figures 3-5, 3-6, 3-7, and 3-8,
respectively. For cationic and anionic polyeletrolytes, the
charge strengths are stronger for higher concentrations, while
for non-ionic polymer, there is no significant change in zeta
potential when concentration increases.
31
-------
The relationship between coagulant aid concentrations and zeta
potential values based on log-linear regression is summarized in
Table 3-5 and illustrated in Figure 3-9. One can see that the
cationic (POL-EZ-2466 and POL-EZ-3466) and anionic (POL-EZ-7736)
coagulant aids have similar strengths of charge while the zeta
potential of the non-ionic coagulant aid (POL-EZ-2696) is
comparatively insignificant.
Table 3-4. List of Coagulant Aids
Coagulant Aid
(Polyelectrolyte)
POL-EZ-2466
POL-EZ-3466
POL-EZ-2696
POL-EZ-7736
309C
ID
PE-1
PE-2
PE-3
PE-4
PE-5
Charge
Cationic
Cationic
Non- ionic
Anionic
Cationic
Test Stage
Surface
Runoff
CSO
Manufacturer
Calgon
Corporation
Pittsburgh
Pennsylvania
Polydyne, Inc
Riceboro
Georgia
32
-------
POL-EZ-2466 (cationi:)
70 T
60 -
50 -
4J
S 40
O
n3
4J
(U
IS]
30 -
10 -
0.1
Z = 26 + 13 log C
CorrE]atrmCoefEic:ent= 0.9
M easuianent
Regression
1 10
Polyelectolyte Concentration
100
1000
Figure 3-5. Zeta Potential Distribution of POL-EZ-2466
POL^IZ-3466 (cationi:)
n5
-H
4J
fl
(U
4J
O
ft
(C
4J
(U
IS]
100
80
60
40
20
0.1
Figure 3-6.
M easuianent
RegiBSSxn
Z = 19 + 23 !bg C
ConE]ationCoefEic:ent= 0.9
H
100
1 10
Po]ye]ecto]yte C oncentration
Zeta Potential Distribution of POL-EZ-3466
1000
33
-------
POL-EZ-2696 (non-:bn:ic)
Zeta Potential
u -
-2 -
-4 -
-6 -
-8 -
10-
Standard D ev:atbn = 1 .6 m V M easuiem ent
i\ vuiuyu
*
T* T 1*1
II 1 1
-12041
1 10 100
PolyelectiDlytB C oncentcatbn fa g/L)
1000
Figure 3-7. Zeta Potential Distribution of POL-EZ-2696
rtf
-H
QJ
4->
o
-p
QJ
N
POL^Z-7736 (aniDnJc)
Z= -33-20]ogC ConelatbnCc
it= 0.9
-100
0.1
1 10 100
PolyelectiolytE C oncentiatbn fa g/L)
1000
Figure 3-8. Zeta Potential Distribution of POL-EZ-7736
34
-------
Table 3-5. Zeta Potential of Coagulant Aids
Coagulant aid
POL-EZ-2466
POL-EZ-3466
POL-EZ-2696
POL-EZ-7736
Zeta Potential versus Concentration
Z = 26 + 13 log C
Z = 19 + 23 log C
Z = -7.4
average
Standard deviation = 1.6 mV
Z = - 33 - 20 log C
cti
-H
4->
fl
QJ
-p
O
a
cti
-p
QJ
N
100 -,
r~ n
bU -
0 -
-50 -
i nn
0
--*"*"
_ "'
_.-- --^l
^ *
^ .-^ """
~ ~~ -_.__
i i i i
I I I I
.1 1 10 100 1000
POL-EZ-
°1fifi
3466
2696
7736
i
1
10000
C onceatcatiDn |ng/L)
Figure 3-9. A Comparison of Zeta Potential of Coagulant Aid
Experiments
The experimental program consisted of three phases, namely,
prescreening, screening, and confirmative tests. The
prescreening phase was a qualitative determination that provided
operational setups for the screening and confirmative tests. The
screening tests evaluated various coagulant and coagulant aid
concentrations for the MC jar tests. Finally, the confirmative
tests assessed the effects of MC size and concentration in the
jar test. The function and program of each phase is described in
detail in the following sections.
35
-------
Prescreening Jar Tests
The prescreening tests consisted of a qualitative
characterization to provide testing ranges for the control
variables (see Section 3.2). In the prescreening tests, a series
of jar tests was performed on natural and artificial storm runoff
samples mixed with MC, coagulant and coagulant aid. The
operational parameters included: coagulant aid type and dosage,
coagulant dosage, rapid mixing rate and duration, flocculation
mixing rate and duration, and MC type, concentration, and size.
The degree of agitation, time required for good floe formation,
and the time required for settling were noted. Since the results
are largely comparative, visual observation was employed.
Quantitative analysis was considered unnecessary. The parameters
and their ranges are listed in Table 3-6.
Table 3-6. Parameter Setup for Prescreening Jar Tests
Parameter
MC Size
MC Concentrations
Coagulant
Coagulant aid
Coagulant aid
Coagulant aid
Coagulant aid
Rapid Mixing Rate
Flocculation Rate
Rapid Mixing
Duration
Flocculation
Duration
Settling Time
Material
Ottawa sand
Beach sand
Ottawa sand
Beach sand
Aluminum sulfate
POL-E-Z 3466
POL-E-Z 2466
POL-E-Z 7736
POL-E-Z 2696
Value Ranges
100 500 |im
3 10 g/L
10 120 mg/L
0.3 1.5 mg/L
0.3 1.5 mg/L
0.3 1.5 mg/L
0.3 1.5 mg/L
60 200 rpm
10 60 rpm
10 120 sec
10 sec 30 min
1 30 min
36
-------
The minimum energy and time that can initiate all the MC into
suspension were selected as the rapid mixing rate and duration
for the jar tests. In addition, the minimum rate to keep all MC
in suspension was selected as the optimal flocculation rate. The
optimal flocculation time was determined based on floe formation.
The rapid mixing rate and duration as well as the flocculation
rate and duration were determined by visual observation. A
summary of parameters with respect to this criteria selection is
shown in Table 3-7.
Five MC configurations with different sizes and concentrations
were selected based on prescreening tests and they are identified
in Table 3-8.
Table 3-7. Determination of Rapid Mixing Rate with Duration and
Flocculation Rate with Duration
Parameter
Rapid mixing rate
Rapid mixing
duration
Flocculation rate
Flocculation
duration
Selection Criterion
Minimum energy to initiate MC into the
suspension
Minimum time to achieve fully mixed
condition
Minimum energy to keep all MC and floe
the suspension
Best floe formation (size and density)
observation intervals of 1, 3, 5, 10,
20, and 30 min
in
at
15,
Screening Tests
The screening tests focused on determining the optimal doses of
coagulant and coagulant aid for the MC process. There were three
test levels with different combinations of MC, coagulant and
coagulant aid. A summary of parameter evaluation for screening
as well as confirmative tests is presented in Table 3-9.
37
-------
Table 3-8. MC Identification
MC type
MC-1
MC-2
MC-3
MC-4
MC-5
Material
Ottawa sand
Ottawa sand
Ottawa sand
Ottawa sand
Ottawa sand
Size
53 150 |im
(Sieve #270 #100)
150 250 |im
(Sieve #100 #60)
53 150 |im
(Sieve #270 #100)
150 250 |im
(Sieve #100 #60)
53 75 |im
(Sieve #270 #200)
Dosage (g/L)
3
3
10
10
3
Table 3-9. Screening and Confirmative Tests Parameter Evaluation
Test
Screen level 1
Screen level 2
Screen level 3
Confirmative
Parameter
MC-1 and
Coagulant
MC-1,
Coagulant, and
Coagulant aid
MC-1,
Coagulant, and
Coagulant aid
MC-1, MC-2,
MC-3, MC-4,
Coagulant, and
Coagulant aid
Control Variable
Coagulant
concentration
Coagulant
concentration
Coagulant aid
concentration
MC size
MC concentration
Rate and duration for rapid mixing and flocculation determined in
prescreening tests were used for the screening tests. Turbidity
was used as a water quality indicator. Supernatant samples were
taken at settling times of 3 and 8 minutes, respectively.
38
-------
Determination of optimal dose was based on the following
criteria:
Select the jar with the lowest turbidity, unless
The difference between the lowest turbidity jars is less
than 20%, in which case, select the jar with the lower
chemical dosage.
Level I. The purpose of Level 1 testing was to determine the
best coagulant dosage in the absence of coagulant aid. MC-I was
used for each jar. Coagulant concentration setup, ranging from 0
to 80 mg/L, is illustrated in Table 3-10.
Table 3-10. Screening Tests -- Level 1
Test
Set
1-1
Additive
MC
Coagulant
(mg/L)
Jar No .
1
2
3
4
5
6
M C - 1
0
10
20
40
60
80
Level 2. The purpose of level 2 testing was to determine the
best coagulant dosage in the presence of coagulant aid. The
influence of coagulant aid on the optimal dosage of coagulant was
thus determined. Coagulant, MC-I, and 1 mg/L coagulant aid were
used in the tests. The coagulant concentration setup, ranging
from 0 to 80 mg/L, is illustrated in Table 3-11.
Level 3. The purpose of level 3 testing was to determine the
influence of coagulant aid concentrations. The optimal coagulant
concentration based on the test results of level 2, and MC-I was
used for each jar. The coagulant aid concentration setup,
ranging from 0.3 to 1.5 mg/L, is presented in Table 3-12.
39
-------
Table 3-11. Screening Tests -- Level 2
Test
Set
2-1
2-2
2-3
2-4
Additive
MC
Coagulant (mg/L)
Coagulant aid
PE-1 (mg/L)
MC
Coagulant (mg/L)
Coagulant aid
PE-2 (mg/L)
MC
Coagulant (mg/L)
Coagulant aid
PE-3 (mg/L)
MC
Coagulant (mg/L)
Coagulant aid
PE-4 (mg/L)
Jar No .
1
2
3
4
5
6
M C - 1
0
1
10
1
20
1
40
1
60
1
80
1
M C - 1
0
1
10
1
20
1
40
1
60
1
80
1
M C - 1
0
1
10
1
20
1
40
1
60
1
80
1
M C - 1
0
1
10
1
20
1
40
1
60
1
80
1
40
-------
Table 3-12. Screening Tests -- Level 3
Test
Set
3-1
3-2
3-3
3-4
Additive
MC
Coagulant
(mg/L)
Coagulant aid
PE-1 (mg/L)
MC
Coagulant
(mg/L)
Coagulant aid
PE-2 (mg/L)
MC
Coagulant
(mg/L)
Coagulant aid
PE-3 (mg/L)
MC
Coagulant
(mg/L)
Coagulant aid
PE-4 (mg/L)
Jar No .
1
2
3
4
5
6
M C - 1
based on results from test set 2-1
1
0.3
0.5
0.7
1.2
1.5
M C - 1
based on results from test set 2-2
1
0.3
0.5
0.7
1.2
1.5
M C - 1
based on results from test set 2-3
1
0.3
0.5
0.7
1.2
1.5
M C - 1
based on results from test set 2-4
1
0.3
0.5
0.7
1.2
1.5
41
-------
Confirmative Tests
In confirmative tests, it was intended to determine the impact of
the MCs with respect to their size and concentration. In these
tests, one coagulant, four MCs, and four coagulant aids were
used. The experimental setup is outlined in Table 3-13.
Table 3-13. Confirmative Tests
Test
Set
C-l
C-2
C-3
C-4
Additive
MC
Coagulant
(mg/L)
and
Coagulant
aid
(mg/L)
MC
Coagulant
(mg/L)
and
Coagulant
aid
(mg/L)
MC
Coagulant
(mg/L)
and
Coagulant
aid
(mg/L)
MC
Coagulant
(mg/L)
and
Coagulant
aid
(mg/L)
Jar No .
1
2
3
4
5
M C - 1
Based
on
Test
Set
1-1
Based
on
Test
Set
3-1
Based
on
Test
Set
3-2
Based
on
Test
Set
3-3
Based
on
Test
Set
3-4
M C - 2
Based
on
Test
Set
1-1
Based
on
Test
Set
3-1
Based
on
Test
Set
3-2
Based
on
Test
Set
3-3
Based
on
Test
Set
3-4
M C - 3
Based
on
Test
Set
1-1
Based
on
Test
Set
3-1
Based
on
Test
Set
3-2
Based
on
Test
Set
3-3
Based
on
Test
Set
3-4
M C - 4
Based
on
Test
Set
1-1
Based
on
Test
Set
3-1
Based
on
Test
Set
3-2
Based
on
Test
Set
3-3
Based
on
Test
Set
3-4
42
-------
Chapter 4
Experimental Results: Surface Runoff
4.1 Prescreening Tests
The function of the prescreening tests was to determine the
ranges of operational parameters to be utilized in the study for
the control variables employed. As stated in Chapter 3, E-l and
E-2 represent aluminum sulfate and ferric chloride, respectively.
Names and identifications of coagulant aids and MCs are indicated
in Tables 3-4 and 3-8, respectively. A series of jar tests was
performed with conditions stated in Table 3-7.
Rapid mixing. Based on the prescreening trials, two stages of
rapid mixing were used in the MC weighted jar tests. In the
first stage, a mixing rate of 150 rpm for 10 seconds was required
to lift MCs from the bottom of the jar into suspension. In the
second stage, a mixing rate of 100 rpm was required to keep the
MC in suspension. It was observed that floe growth began 10
seconds from the beginning of the second rapid mixing stage.
Floe shear would occur due to the rapid speed if the rapid mixing
rate of 100 rpm were continued. Based on this observation, a 10-
second duration was selected in the second stage of rapid mixing.
Slow mixing (flocculation). It was observed that both MC and
floes would settle without further flocculation if the slow
mixing rate were lower than 60 rpm. 60 rpm was found to be an
appropriate mixing rate to avoid floe shear but still keep MCs
and floes in suspension. A 10-second duration was found to be
sufficient for the flocculation process.
In the MC weighted jar test, the total mixing time (approximately
30 seconds) is much shorter than those used in conventional jar
tests. This indicates that MC might be an effective approach in
reducing the treatment time and in turn, the treatment cost,
compared to conventional coagulation processes. Due to the
presence of MCs, however, both rapid and flocculation mixing
rates of MC weighted jar tests were higher than those of
conventional jar tests.
In addition, it was observed that the supernatant was rather
clear after 3 minutes of settling. As a result, 3 and 8 minute
sampling times were selected to study the settling kinetics.
43
-------
Table 4-1 is a summary of experimental settings based on
prescreening tests. These settings were used as operating
conditions for both screening and confirmative tests.
Table 4-1. Summary of Experimental Settings
Parameter
Rapid mixing rate -- stage- 1
Rapid mixing duration -- stage- 1
Rapid mixing rate -- stage-2
Rapid mixing duration -- stage-2
Slow mixing (Flocculation) rate
Slow mixing duration
MC concentration -- 1
MC concentration -- 2
MC size -- 1
MC size -- 2
Settling time -- 1
Settling time --2
Coagulant concentration
Coagulant aid concentration
Value
150 rpm
10 sec
100 rpm
10 sec
60 rpm
10 sec
3 g/L
10 g/L
53 150 |im
150 250 |im
3 min
8 min
10 80 mg/L
0.3 1.5 mg/L
44
-------
4.2 Effects of MC Coagulation
>y
Figure 4-1 illustrates cumulative volume distributions versus
particle size before and after the MC weighted coagulation. For
the raw sample (before treatment), the measurable range of
particle size is from 1 to 170 |im which consists of 81% from 10
to 100 |im and 14% smaller than 10 |im. After the MC coagulation,
the particles in the supernatant of the jar were found to be < 2
|im, which indicated that all particles > 2 \im were removed.
Figures 4-2 (A) and (B) illustrate the particle size distribution
characteristics of the raw sample and supernatant sample after
treatment, respectively. For the raw sample, the distribution
peak is at approximate 50 |im while the small particles (from 0.7
to 2 |im) were either undetectable or only a low percentage
compared to the peak.
100 -r
(V
-x- T reefed sam p]e
Raw sample
0.1
1 10 100
Particle Size fn licrom eters)
1000
MC-l: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l: Aluminum Sulfate Concentration = 40 mg/L
PE-1: Polyelectrolyte POL-EZ-2466 Concentration = 1 mg/L
Figure 4-1. Particle Sizes of Raw and Treated Samples (Cumulative)
45
-------
Q
QJ
tn
5 -r
4 --
Raw Sample
QJ 3 ..
o o f
QJ
1 --
O 0
13 2 - - 2 m licrom el
0.1
10
Particle Size tn licrom eters)
100
1000
Q
QJ
QJ
o
40 -r
30 --
QJ
8 20 +
QJ
10 --
0
0.1
(B) Treated Sample
2 m licrom eters
1 10 100
Particle Size |n licrom eters)
1000
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l: Aluminum Sulfate Concentration = 40 mg/L
PE-1: Polyelectrolyte POL-EZ-2466 Concentration = 1 mg/L
Figure 4-2. Particle Sizes of Raw and Treated Samples
(Distributions)
46
-------
After the treatment, only particles smaller than 2 |im were found
to remain in the supernatant of the sample. Since the larger
particles were all removed from the sample, the percentage of
small particles increases. A summary of particle size
distributions in different ranges is shown in Table 4-2. One can
see that more than 96% (> 2 |im) of particles were removed from
the raw sample. The removal efficiency for the 4% smaller
particles (< 2 |im) was indicated by the particle count rate
discussed below.
Figures 4-3 (A) and (B) present comparisons of turbidity and
particle count rate (PCR) results with and without using an MC
for different coagulant aids. Turbidity has been used
extensively as a water quality indicator in water and wastewater
treatment process analysis. The count rate, utilizing a unit of
kilo-count per second, is directly related to the particle
concentration in the solution; therefore, it is a good indicator
for colloidal particles (containing particles smaller than 0.45
micrometer) for which TSS measurement is usually not applicable.
Both turbidity and particle count rate of supernatant samples
with MC are much lower than those without MC. It is apparent
that the addition of MC improves the treatment process
effectively.
Table 4-2. Particle Size Distribution of Raw and Treated Samples
Size Range (|om)
Volume
(%)
Raw
Treated
0 . 6 9 1
< 0.1
84
1 2
4
16
2 10
10
< 0.1
10 100
81
< 0.1
100 170
5
< 0.1
47
-------
-H
TJ
-H
8
6
4 --
2
0
Tutbddiy of Raw Sample = 74 ntu
W ihM C
W ihoutM C
(A) PE-4 (Anionic)
o
QJ
rH
u
-H
rd
20 -r
15 --
10 --
5 --
0
ParticGe CountRatE of Raw Sample = 35 kops
W ihM C
W ihoutM C
(B) PE-3 (Non-ionic)
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l: Aluminum Sulfate Concentration = 40 mg/L
PE-3: Polyelectrolyte POL-EZ-2696 Concentration = 1 mg/L
PE-4: Polyelectrolyte POL-EZ-7736 Concentration = 1 mg/L
Figure 4-3. Effectiveness of the MC Process
48
-------
4.3 Screening Tests
For each test set, supernatant samples were taken at 3 minute and
8 minute settling times, respectively. The time difference for
sampling six jars was less than 15 seconds, which was negligible.
Screening tests include three levels of setup conditions that are
summarized in Tables 3-10 through 3-12 (Section 3.3). Results of
these three levels of tests are described in the following
sections.
Level One
In this level, only an MC and a coagulant were used for the
treatment process. The concentration of coagulant range from 0
to 80 mg/L (see Table 3-10).
pH
Figure 4-4 illustrates the relationship of pH and coagulant
concentration. It is seen that pH decreases as coagulant
concentration increases. All pH values are in the range of 6.5
to 7.5 except for jar number six with 80 mg/L coagulant. For
coagulant concentration between 20 and 80 mg/L, pH values with 8
minute settling time are closer to neutral than those with 3
minute settling time. The changes of pH values may be due to
that some of the immediate pH drop (as aluminum hydroxide is
formed from alum) is counteracted by a slow loss of carbon
dioxide to the atmosphere, raising the pH slowly.
Turbidity
Figure 4-5 shows the turbidity distribution versus coagulant
concentration. Here the lowest turbidity for both 3 and 8 minute
settling times was found at 60 mg/L. In addition, the 8-minute
settling time yields better results (lower turbidity) than the 3
minute settling time.
Zeta Potential
Figure 4-6 illustrates zeta potential values versus coagulant
concentrations. It can be seen that the zeta potential basically
increases as coagulant concentration increases. At coagulant
concentrations between 60 and 80 mg/L, the differential of zeta
potential becomes greater between settling times of 3 minutes and
8 minutes.
Turbidity versus Zeta Potential
It is of interest to see the correlation between zeta potential
and turbidity presented in Figure 4-7. The turbidity decreases
for less charge in the zeta potential range of -10 to -25 mV. A
log-linear regression was performed based on measurements within
the negative zeta potential range. The straight line in the
positive zeta potential side is the imaging line of the negative
49
-------
side. Since there is only one point on the positive side,
regression at this side is impossible. Generally, organic
colloids require zeta potentials near zero for optimal
coagulation, while clay-related turbidity is best removed at
slightly negative zeta potentials. In this case, the fact that
the measured point is very close to the image line indicates that
the zero potential is the image symmetrical point, and so, is the
best for turbidity removal.
Particle Count Rate
The particle count rate for 60 mg/L coagulant was found to be 1.2
and 1.0 at settling times of 3 minutes and 8 minutes,
respectively (see Figure 4-8). Results showed that approximately
50% of the small particles (< 5 |im) were removed from the raw
sample in 3 minutes.
Level Two
Level two parameters include MC, coagulant, and four types of
coagulant aids with a fixed concentration of 1 mg/L. For each
coagulant aid, coagulant concentration was varied from 0 to 80
mg/L (see Table 3-11) .
pH
Figure 4-9 shows a typical pH distribution versus coagulant
concentration. It appears that the relationship between pH and
coagulant for both level one and level two tests is similar. A
summary of pH behavior for the four coagulant aids at two
different settling times (3 and 8 minutes) is shown in Figure 4-
10. pH values were found to be close to 7 when coagulant
concentrations were in the range of 20 to 40 mg/L.
Turbidity
The lowest value of turbidity was obtained at 40 mg/L coagulant
as shown in Figure 4-11. A plot of turbidity distributions (with
3 and 8 minute settling times) versus coagulant concentration
with the four coagulant aids is illustrated in Figure 4-12.
Based on the turbidity results, the optimal coagulant
concentration for level two was 40 mg/L for all four coagulant
aids. It was indicated that turbidity with 8 minute settling
time is lower than that with 3 minute settling time. This is
true because the longer the settling time, the more the particles
will be removed.
Total Solids and Total Volatile Solids
Figure 4-13 and Figure 4-14 illustrate the total solids and total
volatile solids distributions, respectively. It can be seen that
8 minutes settling time yields better results for both total
solids and total volatile solids. The total solids results
demonstrate that 40 mg/L was the optimal coagulant concentration.
50
-------
For total volatile solids results, the optimal coagulant
concentration was between 20 and 40 mg/L. Similar results were
obtained for the total solids and total volatile solids removals.
Particle Count Rate
Figure 4-15 illustrates a comparison between particle count rate
and turbidity. It can be seen that a similar trend indicates 40
mg/L of coagulant as the optimal concentration for turbidity and
particle count rate removal.
Turbidity versus Zeta Potential
Figure 4-16 illustrates zeta potential distributions versus
coagulant concentrations. Since anionic coagulant aid was added
into the mixture, all supernatant samples, with the exception of
80 mg/L of coagulant samples, have larger negative charge than
the raw sample. Figure 4-17 shows the correlation between
turbidity and zeta potential by using a log-linear regression
based on the negative zeta potential data. In addition, an image
line was drawn for the positive zeta potential side with respect
to the image centerline of - 10 mV zeta potential (Z). Similar
to level one, it can be seen that turbidity decreases with lower
zeta potential.
Level Three
In level three, four coagulant aids with concentration ranging
from 0.3 to 1.5 mg/L were tested with MC-I and 40 mg/L coagulant
which was determined in level two tests to be the optimal dosage
for all four coagulant aids.
pH
Figure 4-18 shows the pH value versus coagulant aid concentration
for the four coagulant aids at 3 and 8 minute settling times.
Results indicate that coagulant aid concentration had no major
influences on pH values.
Turbidity and/versus Zeta Potential
Based on the test results, the trends of turbidity removal are
different for various types of coagulant aids. Figure 4-19
reveals that turbidity decreases as coagulant aid (POL-EZ-2696)
concentration increases up to 1 mg/L, but varies between 1 mg/L
to 1.5 mg/L. In this test set, 1 mg/L is considered as the best
coagulant aid dosage. Similar observations can be found from
zeta potential distributions (see Figure 4-20). By comparing the
above two figures, one can see that there exists a mirror image
phenomenon. As a result, a correlation between turbidity and
zeta potential in the presence of coagulant aid can be
established. Figure 4-21 shows the correlation of zeta potential
with turbidity. The results show that the turbidity decreases
for less negative charge in the range between -25 and -8 mV.
51
-------
Figure 4-22 shows that turbidity decreases as coagulant aid (POL-
EZ-2466) concentration increases up to 1.2 mg/L and reaches its
lowest point between 0.7 and 1.2 mg/L. Figure 4-23 shows zeta
potential distributions. Again, the mirror image phenomenon can
be observed from these two figures. The correlation between zeta
potential and turbidity is presented in Figure 4-24. It can be
seen that in the zeta potential range of -3 to -15 mV, turbidity
has no significant change while in the zeta potential range of -
15 to -22 mV, turbidity decreases as zeta potential approach less
negative charge. For other coagulant aids, the results are
summarized in Table 4-3.
Particle Count Rate, Total Solids and Total Volatile Solids
Particle count rate distribution versus coagulant aid
concentration is illustrated in Figure 4-25. At this point, no
definite conclusion can be drawn as a result of the twin peaks.
Similar trends were observed in total solids and total volatile
solids versus coagulant aid concentration as shown in Figures 4-
26 and 4-27, respectively. Therefore, there appears to be a
functional relationship among particle count rate, total solids,
and total volatile solids.
Summary
Based on the screening test results, the coagulant concentrations
of 60 mg/L without coagulant aid and 40 mg/L with 1 mg/L of
coagulant aid yielded better turbidity reduction efficiency.
Relationships between turbidity and zeta potential under
different conditions are listed in Table 4-3. For level one with
coagulant in the absence of coagulant aid, the lowest turbidity
point (symmetrical point) is at zero potential. For level two
with additions of both coagulant (various concentrations) and
coagulant aid (constant concentration), the lowest turbidity
point shifted from zero potential to the range of -10 to +10 mV.
For level three, for all four coagulant aids with various
concentrations, the relationships between turbidity and zeta
potential are either monotonously increasing (for anionic POL-EZ-
7736) or decreasing (for cationic POL-EZ-2466, POL-EZ-3466 and
non-ionic POL-EZ-2696).
52
-------
Table 4-3. Relationship of Turbidity and Zeta Potential
Condition
Level 1
Correlation Equation
Coagulant concentrations: 10 80 mg/L
Without Coagulant aid
log T = - 1 - 0.1 Z (- 25 < Z < 0)
log T = - 1 + 0.1 Z (0 < Z < 15)
Level 2
PE-1
PE-2
PE-3
PE-4
Coagulant concentration: 10 80
Four coagulant aids (concentration
log T = - 0.87 - 0.08 (- 25 < Z < -
log T = - 0.38 (- 6 < Z < 15)
log T = 0.87 - 0.02 Z (- 25 < Z <
log T = 0.19 + 0.038 Z (10 < Z <
log T = - 0.9 - O.KZ + 7) (- 25 < Z <
log T = - 0.9 + O.KZ + 7) (- 7 < Z <
mg/L
= 1 mg/L)
6)
10)
15)
- 7)
10)
log T = - 0.5 - 0.08(Z+10) (- 30 < Z < - 10)
log T = - 0.5 + 0.08(Z+10) (- 10 < Z < 10)
53
-------
Level 3
PE-1
PE-2
PE-3
PE-4
Coagulant concentration = 40 mg/L
Four coagulant aids (concentrations: 0.3 1.5 mg/L)
log
log
log
log
T = 0.89 -
(-22 < Z <
T = 0.79 -
(-20 < Z <
T = 0.33 -
(-25 < Z <
T = 1.5 +
(-25 < Z <
0.02 Z
- 3)
0.007 Z
30)
0.02 Z
- 8)
0.05 Z
-10)
Note:
T = Turbidity (ntu); Z = Zeta potential (mV)
4.4 Confirmative Tests
Based on the results from screening tests, confirmative tests
were conducted to examine the influence of MC size and
concentration on the MC process. Two MC size ranges, (53150 |im
defined as small MCs and 150250 |im as large MCs) and two MC
concentrations (3 g/L as low concentration and 10 g/L as high
concentration) were tested. Coagulant and coagulant aid
concentrations of 40 mg/L and 1 mg/L were used, respectively, for
each jar test set.
MC-1 versus MC-3 and MC-2 versus MC-4
Figure 4-28 illustrates the turbidity distributions for small MCs
with different MC concentrations. It is apparent that the high
concentration yields lower turbidity than the low concentration
for the four coagulant aids. Figure 4-29 shows the turbidity
distributions for large MCs with different MC concentrations.
The trend is similar with the small MC except for coagulant aid
POL-EZ7736 (anionic). Comparing the above two figures, one can
54
-------
see that the influence of MC concentration is more pronounced in
the smaller diameter range.
MC-1 versus MC-2 and MC-3 versus MC-4
Figure 4-30 compares the results of different MC sizes for low MC
concentrations. Except for coagulant aid POL-EZ-2696 (non-
ionic) , large MCs yield lower turbidity. Figure 4-31 compares
the results of different MC sizes for high MC concentrations.
Unlike the situation for low MC concentration, the lower
turbidity results were obtained from small MCs for all four
coagulant aids. Figure 4-32 is a turbidity summary based on
coagulant aid grouping.
Based on the turbidity analyses illustrated in Figure 4-33, MC-3,
namely, 53150 |im size range and 10 g/L dosage, yields the best
results.
Turbidity versus Particle Count Rate
Figure 4-34 presents the correlation between particle count rate
and turbidity. The correlation coefficient of turbidity and
particle count rate was 0.5.
55
-------
Figure 4-1. Particle size Distributions of raw and treated samples
Figure 4-2. Effectiveness of the MC process
Figure 4-3. Effectiveness of the MC process
55
-------
pH of law sam p]e = 7.7
7.5 T
7.0 --
ft 6.5 t
6.0 -
5.5
10
20 30 40 50
E ]ectK)]yte C onoentiatbn frn g/L)
60
70
80
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 1080 mg/L
Settling Time: 3 and 8 minutes; Without Polyelectrolyte
Figure 4-4. Typical pH Distributions (Level-1)
56
-------
100 T
-U
fi
3 io
-H
Turbiiiiy of law sam p]e = 79 ntu
4-
4-
10
20 30 40 50
E]ec±to]yteConaentatbn frng/L)
60
70
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 1080 mg/L
Settling Time: 3 and 8 minutes; Without Polyelectrolyte
Figure 4-5. Turbidity Versus Coagulant Concentration (Level-1)
80
57
-------
Z eta potential of raw sample = -20 mV
ns
-H
20 T
10 -
g °
O
nJ -10
-U
cu
IS]
-20 -'
-30
4-
+
+
10
20
30 40 50
E ]ecttiD]yte C oncentratbn frn g/L)
60
70
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 1080 mg/L
Without Polyelectrolyte
Figure 4-6. Zeta Potential Distributions (Level-1)
80
58
-------
Raw sample: Z eta potential = -20 mV; Tuttoiiiiy = 79 ntu
100.0 T
10.0 "
-H
d
-H
1.0 "
0.1
]ogT= -1-0.1Z
]ogT= -1+ 0.1Z
C ortelatbn coefficient = 0.6
1 1 1
-30 -25 -20 -15 -10 -5 0
Zeta Potential faV)
10
15
20
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 1080 mg/L
Without Polyelectrolyte
Figure 4-7. Correlation of Zeta Potential and Turbidity (Level-1)
59
-------
100 T
10 --
"* Tuttoiiiiy
"" Countiate
x
x
X
Raw sample
3 m ii
8 m ii
T 2.5
-- 2
1.5 0)
4J
-- 1
-- 0.5
4-)
fi
^3
O
U
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 60 mg/L
Settling Time: 3 and 8 minutes; Without Polyelectrolyte
Figure 4-8. Particle Count Rate for the Best Coagulant Concentration (Level-1)
60
-------
ffi
a
7.5 T
7.0 "
6.5 "
6.0 "
5.5 -'
pH of raw sam pie = 7.6
~*~ 3 m dn
" 8 m dn
5.0
10
20
30 40 50
E lectrolyte C oncentration (fa g/L)
60
70
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 1080 mg/L
PE-3: Polyelectrolyte POL-EZ-2696 Concentration = 1 mg/L
Settling Time: 3 and 8 minutes
Figure 4-9. Typical pH Distributions (Level-2)
80
61
-------
7.5 T
7.0 "
6.5 "
6.0 f
5.5 "
5.0 "
4.5
x
X
X
X
X
X
X
pH of Raw Sample: 7.5 8.0
x
X
10
20 30 40 50
Electrolyte Concentration frng/L)
60
70
80
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 1080 mg/L
PE-1: POL-EZ-2466; PE-2: POL-EZ-3466; PE-3: POL-EZ-2696; PE-4: POL-EZ-7736
Polyelectrolyte Concentration = 1 mg/L; Settling Time: 3 and 8 minutes
Figure 4-10. Summary of pH Distributions (Level-2)
62
-------
10.0 T
4J
-H
d
-H
1.0
10
Tuifoiiily of raw sam pie = 74 ntu
20
30 40 50
E lectrolyte C oncentration (fa g/D
60
70
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 1080 mg/L
PE-4: Polyelectrolyte POL-EZ-7736 Concentration = 1 mg/L
Settling Time: 3 and 8 minutes
Figure 4-11. Typical Turbidity Distributions (Level-2)
80
63
-------
100.0 T
-H
d
-H
10.0 -:
1.0 -:
0.1
Tuibiiiiy of raw samples: 74 137 ntu
. Tuibiiiiy ]
3 m in. average
8 m in. average
10
20 30 40 50
Electrolyte Concentration frng/L)
60
70
1
1 1 1 1 1 1 1
80
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 1080 mg/L
PE-1: POL-EZ-2466; PE-2: POL-EZ-3466; PE-3: POL-EZ-2696; PE-4: POL-EZ-7736
Polyelectrolyte Concentration = 1 mg/L; Settling Time: 3 and 8
Figure 4-12. Summary of Turbidity Distributions (Level-2)
64
-------
Total solids of law sample = 580mg/L
10
20 30 40 50
E lecutxlyte C oncBitiatbn frn g/L)
60
70
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 1080 mg/L
PE-3: Polyelectrolyte POL-EZ-2696 Concentration = 1
Figure 4-13. Total Solids Distributions (Level-2)
80
65
-------
250 T
200 -'
-H
i-H
O
CO
0)
rH 150 t
-H
4J
O
>
4J
O
EH
100 --
50 "
Total volatile solid of raw sam pie = 280m g/L
~*~ 3 m in
~* 8 m dn
x
X
X
X
X
X
X
X
X
X
X
X
X
X
10
20 30 40 50
E lectrolytE C oncentxation (m g/L)
60
70
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 1080 mg/L
PE-3: Polyelectrolyte POL-EZ-2696 Concentration = 1 mg/L
Settling Time: 3 and 8 minutes
Figure 4-14. Total Volatile Solids Distributions (Level-2)
80
66
-------
7 T
6 "
5 "
O
en
a
u
cu 3
4-)
PJ
O
U
2 "
1 --
Raw sample: CountRate= 33.8kcps; TutbJdiy = 85 ntu
"" Count Rate
"* TutbJdiy
10
20
30 40 50
E ]ecteD]yte C oncsentcatiDn frn g/L)
60
70
80
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 1080 mg/L
PE-3: Polyelectrolyte POL-EZ-2696 Concentration = 1
Figure 4-15. A Comparison of Turbidity and Particle Count Rate (Level-2)
67
-------
Z eta potential of raw sample = -16 mV
10
20
30 40 50
E lecutxlyte CDnaentiatbn frn g/1)
60
70
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 1080 mg/L
PE-4: Polyelectrolyte POL-EZ-7736 Concentration = 1 mg/L
Settling Time: 3 and 8 minutes
Figure 4-16. Zeta Potential Distributions (Level-2)
80
68
-------
Raw sample: Z eta potential = -16 mV, Turbiiity = 74 ntu
10.0 x
2 i.o ±
]ogT = -0.5 -0.08 (Z+ 10)
C onelatbn coefficient = 0.8
]ogT = -0.5 + 0.08 (Z+ 10)
0.1
-30
-25
-20
-15 -10 -5
ZetaPotential (mV)
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 1080 mg/L
PE-4: Polyelectrolyte POL-EZ-7736 Concentration = 1
Figure 4-17. Correlation of Zeta Potential and Turbidity (Level-2)
10
69
-------
7.5 T
pH of Raw Samples: 7.5 8.0
7.0 "
ffi f- ,- I
a 6.5 --
6.0 "
5.5
0.2
0.4 0.6 0.8 1
Po]ye]ectED]yte C oncentzation (m g/L;
1.2
1.4
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 40 mg/L
Polyelectrolyte: PE-1: POL-EZ-2466; PE-2: POL-EZ-3466;
PE-3: POL-EZ-2696; PE-4: POL-EZ-7736
Polyelectrolyte Concentration = 0.31.5 mg/L
Settling Time: 3 and 8 minutes
Figure 4-18. Summary of pH Distributions (Level-3)
1.6
70
-------
Turbiiiiy of law sam p]es = 152 ntu
-U
-H
7 -'
6 -'
5 -'
\
3 -
2 -
1 -
*~ 3 m in
Srnxi
V
i i i i i i i i
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Po]ye]ecttxD]yte concenttatbn frn g/L)
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 40 mg/L
Polyelectrolyte: PE-3: POL-EZ-2696; Polyelectrolyte Concentration = 0.31.5 mg/L
Settling Time: 3 and 8 minutes
Figure 4-19. Turbidity Distributions (Level-3; POL-EZ-2696)
71
-------
Z eta potential of raw sam p]es = -21 m V
0.2
0.4 0.6 0.8 1
Polyetectrolyte concentration (m g/L)
1.2
1.4
1.6
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 40 mg/L
Polyelectrolyte: PE-3: POL-EZ-2696; Polyelectrolyte Concentration = 0.31.5 mg/L
Settling Time: 3 and 8 minutes
Figure 4-20. Zeta Potential Distributions (Level-3; POL-EZ-2696)
72
-------
Raw sample: Turbiiily of = 152 ntu; Z eta potential = -21 mV
10 T
-H
d
-H
M easaiEm ent
Regression
logT = 0.33 -0.02 Z
C onelation coefficient = 0.7
-25
-20
-15 -10
Z eta Potential (m v )
-5
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 40 mg/L
Polyelectrolyte: PE-3: POL-EZ-2696; Polyelectrolyte Concentration = 0.31.5 mg/L
Settling Time: 3 and 8 minutes
Figure 4-21. Correlation of Zeta Potential and Turbidity (Level-3)
73
-------
30 T
25 --
20 +
-H
10
-H
10 t
5 -
TutbJdiy of raw sam p]es = 158 ntu
4-
4-
0.2
0.4 0.6 0.8 1
PcGyetecttrxlyte CDnaentiatbn frn g/1)
1.2
1.4
1.6
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 40 mg/L
Polyelectrolyte: PE-3: POL-EZ-2696; Polyelectrolyte Concentration = 0.31.5 mg/L
Settling Time: 3 and 8 minutes
Figure 4-22. Turbidity Distributions (Level-3; POL-EZ-2466)
74
-------
nJ
-H
-------
100 T
10 :
-25
Raw sample: Turbiiiiy = 158 ntu; Z eta potential = -21mV
M easuiament
Regiessbn
tog T = 0.89 - 0.02 Z
4-
C onelatbn coefficient = 0.7
4-
-20
-15 -10
Z eta Potential frn V)
-5
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 40 mg/L
Polyelectrolyte: PE-1: POL-EZ-2466; Polyelectrolyte Concentration = 0.31.5 mg/L
Figure 4-24. Correlation of Zeta Potential and Turbidity (Level-3)
76
-------
0>
O
3.5 --
3.0 --
2.5 --
2.0"
1.5"
1.0"
0.5 --
o.o --
0
C ount rate of raw sam pies = 11.2 kcps
A
\
\
\
\
\
\
0.2
0.4 0.6 0.8 1
Polyelectroilyte concentrat±5n (fa g/D
1.2
1.4
1.6
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 40 mg/L
Polyelectrolyte: PE-1: POL-EZ-2466; Polyelectrolyte Concentration = 0.31.5 mg/L
Settling Time: 3 minutes
Figure 4-25. Particle Count Rate Distribution (Level-3; POL-EZ-2466)
77
-------
Total solids of raw samp]e= 960 mg/L
300 -
250 -
d 200 -'
-H
i-H
O
co 150
100 -
50 -
0.2
0.4 0.6 0.8 1
Polyelectrolyte C oncentration fa g/1)
1.2
1.4
1.6
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 40 mg/L
Polyelectrolyte: PE-1: POL-EZ-2466; Polyelectrolyte Concentration = 0.31.5 mg/L
Settling Time: 3 and 8 minutes
Figure 4-26. Total Solids Distributions (Level-3; POL-EZ-2466)
78
-------
200 T
-H
iH
w 150 t
cu
-H
nl 100 +
50 -
o
Total volatile solids of law sam p]e = 480 m g/L
0.2
0.4 0.6 0.8 1
Polyelecutxlyte conaentiation ^n g/1)
1.2
1.4
1.6
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 40 mg/L
Polyelectrolyte: PE-1: POL-EZ-2466; Polyelectrolyte Concentration = 0.31.5 mg/L
Settling Time: 3 and 8 minutes
Figure 4-27. Total Volatile Solids Distributions (Level-3; POL-EZ-2466)
79
-------
4-1
-H
d
-H
ft
12 T
M C Size: 53-150 m Jciom eters
M C Cone. = 3g/L
M C Cone. = lOg/L
Sett
±ig tin e =
8 m ]
i
n
i
i
i
2696
7736 3466
PotyetectzoilytE Type (PO L -E Z ) N urn ber
2466
MC-1: MC Size Range = 53150 |im; MC Concentration = 3 and 10 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 40 mg/L
Coagulant Aid (Polyelectrolyte) Identification in Table 3-4
Polyelectrolyte Concentration = 1 mg/L; Settling Time = 8 minutes
Figure 4-28. Turbidity for Different MC Concentrations (Small MC)
80
-------
12 T
10 "
8 "
4J
-H
-H 6 "
X3
4 --
2 "
M C Sdze: 150250 m licrom etErs
Settling tim e = 8 m dn
+
M C Cone. = 3g/L
M C Cone. = lOg/L
2696
7736 3466
Po]ye]ecttD]ytE Type Q?0 L -E Z) N urn ber
2466
MC-1: MC Size Range = 150250 |im; MC Concentration = 3 and 10 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 40 mg/L
Coagulant Aid (Polyelectrolyte) Identification in Table 3-4
Polyelectrolyte Concentration = 1 mg/L; Settling Time = 8 minutes
Figure 4-29. Turbidity for Different MC Concentrations (Large MC)
81
-------
T3
3
EH
12 T
10 "
M C C concentration = 3 g/L
Settling tim e = 8 m in
53 150 m icrom eters
150 250 m icrom eters
-I-
2696
7736 3466
Poly electrolyte Type (PO L -E Z ) N urn ber
2466
MC-1: MC Size Range = 53150 |im and 150250 |im; MC Concentration = 3 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 40 mg/L
Coagulant Aid (Polyelectrolyte) Identification in Table 3-4
Polyelectrolyte Concentration = 1 mg/L; Settling Time = 8 minutes
Figure 4-30. Turbidity for Different MC Sizes (Low MC Dosage)
82
-------
4J
-H
^
-H
£3
^
EH
12 T
10 "
8 "
4 --
2 --
M C C oncentratbn = 10 g/L
Settling tim e = 8 m in
53 150 m doom etEis
150 250 m dcrom etEis
2696
7736 3466
Po]ye]ecttD]ytB Type (PO L -EZ ) N urn ber
2466
MC-1: MC Size Range = 53150 |im and 150250 |im; MC Concentration = 10 g/L
E-l (Electrolyte): Aluminum Sulfate Concentrations = 40 mg/L
Coagulant Aid (Polyelectrolyte) Identification in Table 3-4
Polyelectrolyte Concentration = 1 mg/L; Settling Time = 8 minutes
Figure 4-31. Turbidity for Different MC Sizes (High MC Dosage)
83
-------
4-1
-H
d
-H
ft
12 T
10 "
4 ~-
2 ~-
POL-EZ-2466
POL-EZ-3466
POL-EZ-2696
POL-EZ-7736
M C-I
M C-H
M C-m
M C-IV
MC Group Identification, Size Range and Concentration in Table 3-8
E-l (Electrolyte): Aluminum Sulfate Concentrations = 40 mg/L
Coagulant Aid (Polyelectrolyte) Identification in Table 3-4
Polyelectrolyte Concentration = 1 mg/L; Settling Time = 8 minutes
Figure 4-32. Turbidity Summary for Confirmative Tests (by MC Group)
84
-------
-H
T!
-H
12
10
4 ~-
M C -I
M C -II
M C -III
M C -IV
POL-EZ-2696
POL-EZ-7736
POL-EZ-3466
POL-EZ-2466
MC Group Identification, Size Range and Concentration in Table 3-8
E-l (Electrolyte): Aluminum Sulfate Concentrations = 40 mg/L
Coagulant Aid (Polyelectrolyte) Identification in Table 3-4
Polyelectrolyte Concentration = 1 mg/L; Settling Time = 8 minutes
Figure 4-33. Turbidity Summary for Confirmative Tests (by Coagulant Aid Group)
85
-------
C ortHJation ooefficdait = 0.5
CR = 2.3 + 0.03T
16 T
V\A AA
. v .A . . A A .
10
TutbJdiy (atu)
100
1000
Figure 4-34. Correlation of Particle Count Rate and Turbidity
86
-------
Chapter 5
Experimental Results: Combined Sewer Overflow
The purpose of this chapter is to examine the effectiveness of
the MC process as a treatment technology for combined sewer
overflows (CSO) (as was previously analyzed in the treatment for
surface runoff). In developing this process, an experimental
program was carried out. The analyses were based on past
experience and information noted in Chapter 4. The experimental
results are summarized hereby in four phases, namely,
prescreening tests, effectiveness of MC process, control variable
optimization, and response variable evaluation.
5.1 Prescreening Tests
Mixing Parameters. The mixing parameter setup for the CSO
treatment is similar to those of the surface runoff tests as
described in Chapter 4. For CSO samples, it was observed that
floes grow gradually within one to two minutes during slow mixing
(flocculation). Along with the growth of floes, the mixing rate
should vary from 60 rpm at the beginning to 20 rpm at the latter
part of mixing process in order to reduce shear stress and avoid
floe break down. The total mixing time is less than two minutes.
Table 5-1 presents a summary of experimental settings to be used
in subsequent tests in control variable optimization and response
variable evaluation. These settings were based on prescreening
tests.
5.2 Effect of the MC
The effect of the MC weighted coagulation was evaluated via
turbidity indicator as well as particle size distribution at the
pre- and post-jar test of raw and supernatant samples,
respectively. Figure 5-1 presents results of turbidity versus
settling time with and without the use of an MC. With the MC
process, the turbidity was reduced from 85 ntu (for the raw
sample) to 5.0 and 3.1 NTU at the 3- and 10-minute settling
times, respectively. At the 3- and 10-minute settling times, the
turbidity without MC is 10 and 5.6 times that of the turbidity
levels with MC, respectively. Thus, the MC treatment process is
effective for removal of turbidity in CSO samples.
Figures 5-2 and 5-3 illustrate cumulative and non-cumulative
volume distributions versus particle size before and after the MC
treatment process, respectively. For the raw sample (before
treatment), the measurable range of particle size was from 0.5 to
60 |im with 93% particles larger than 2 |im in size. After the MC
treatment, the particles in the supernatant of the sample were
found to be smaller than 2 |im, thus indicating that particles
larger than 2 |im in the raw sample were totally removed.
87
-------
Table 5-1. Summary of Experimental Settings for CSO Treatment
Parameter
Rapid mixing rate -- stage- 1
Rapid mixing duration -- stage- 1
Rapid mixing rate -- stage-2
Rapid mixing duration -- stage-2
Slow mixing (Flocculation) rate
Flocculation mixing duration
MC concentration
MC size range - 1
MC size range - 2
Settling time
Ferric chloride concentration
Coagulant aid (polyelectrolyte)
concentration
Value
150 rpm
10 sec
100 rpm
10 sec
20 60 rpm
1 1 . 5 min
17 g/L
53 75 |im
150 250 |im
1 20 min
10 100 mg/L(as Fe+++)
0.5 15 mg/L
88
-------
Turbiiiy for Raw Sam pie = 85 ntu
50 -,
Turbidity
h-1 DO U> d^
3 O O O O
1 1 1 1 1
1
DwihMC
D w JthoutM C
1
i
10
Setting Tine
MC-5: MC Size Range = 5375 |im; MC Concentration = 3g/L
E-2: Ferric Chloride Concentration = 40 mg/L(as Fe+++)
PE-5: Polyelectolyte 309C Concentration = 2 mg/L
Figure 5-1. Effect of MC
e
o
cu
-H
4->
05
U
U
Tieated sample
-Raw sample
1 10 100
Particle Size fa raom eters)
1000
MC-5: MC Size Range = 5375 |im; MC Concentration = 3 g/L
E-2: Ferric Chloride Concentration = 40 mg/L (as Fe+++)
PE-5: Polyelectolyte 309C Concentration = 1.5 mg/L
Figure 5-2. Particle Sizes of Raw and Treated Samples (Cumulative)
89
-------
0.1
Raw Sample
10
Particle Size tn licrcm eters)
100
1000
Q 40-r
0.1
(B) Treated Sample
1 10 100
Particle Size tn licrcm eters)
1000
MC-5: MC Size Range = 5375 |im; MC Concentration = 3g/L
E-2: Ferric Chloride Concentration = 40 mg/L(as Fe+++)
PE-5: Polyelectolyte 309C Concentration = 1.5 mg/L
Figure 5-3. Particle Sizes of Raw and Treated Samples
(Distributions)
90
-------
5.3 Control Variable Determination
>y
In the determination of the effectiveness of the MC process,
turbidity and particle count rate were employed as primary and
secondary indicators, respectively. Six parameters, including
coagulant concentration, coagulant aid concentration, MC size, MC
concentration and settling time were identified as control
variables. Results from each of the control variables with
respect to the turbidity and particle count rate are presented in
this section.
Coagulant Concentration
Turbidity and particle count rate of post treatment supernatant
samples with respect to coagulant concentration are illustrated
in Figure 5-4 (A) and (B), respectively. It can be seen that the
optimal coagulant concentration is 40 mg/L for both turbidity and
particle count rate indicators. The removal rates are 98% for
turbidity and 95% for particle count rate at the optimal
coagulant concentration. These results are confirmed in Figure
5-5 that also shows that the optimal dosage for coagulant is 40
mg/L. Although the raw samples for these two tests are from
different batches, the results are similar.
It is of interest to note that the distribution trends for
turbidity and particle count rate are very similar. Figure 5-6
indicates a linear relationship between these two parameters.
91
-------
-H
d
-H
120 T
100 --
80 --
60 --
40 --
20 --
0
0
Tuzbiiiy for Raw Sam p]e = 176 ntu
20 40 60 80
E -2 C oncHQttation in g/L)
(A) By Turbidity Indicator
PCR forRaw Samp]e= 353 kcps
20
40 60 80
E -2 C oncmttation in g/L)
100
120
100
120
(B) By Particle Count Rate Indicator
MC-5: MC Size Range = 5375 |im; MC Concentration = 3g/L
E-2: Ferric Chloride Concentration Range: 10100 mg/L as Fe+++
PE-5: Polyelectolyte 309C Concentration = 2 mg/L
Settling Time = 3 min
Figure 5-4. Coagulant Concentration Selection (Test-1)
92
-------
-H
TJ
-H
150 -T-
120
90 --
60 --
30 --
500 -T-
Turbidiy for Raw Samp]e= 161 ntu
20 40 60
E -2 C onceatzatiDn fn g/L)
(A) By Turbidity Indicator
PC R for Raw Sample = 303 kcps
80
100
20 40 60 80
E -2 C onoentiatiQn fn g/L)
100
(B) By Particle Count Rate Indicator
MC-5: MC Size Range = 5375 |im; MC Concentration = 3 g/L
E-2: Ferric Chloride Concentration Range: 080 mg/L as Fe+++
PE-5: Polyelectolyte 309C Concentration = 1 mg/L
Settling Time = 3 min
Figure 5-5. Coagulant Concentration Selection (Test-2)
93
-------
400 T
20
40
60
80
100
120
MC-5: MC Size Range = 5375 |im; MC Concentration = 3g/L
E-2: Ferric Chloride Concentration Range: 2110 mg/L as Fe+++
PE-5: Polyelectolyte 309C Concentration = 2 mg/L
Settling Time = 3 min
Figure 5-6. Particle Count Rate Versus Turbidity
Coagulant Aid Concentration
Figure 5-7 presents a comparison of the turbidity distribution
versus coagulant aid concentration for two coagulant
concentrations (20 and 40 mg/L). The coagulant aid concentration
appears to have less influence on final turbidity in low dosages
(ranged 02.5 mg/L) for both coagulant concentrations.
For higher coagulant aid concentration ranged 315 mg/L, 6 mg/L
dosage resulted lowest turbidity as illustrated in Figure 5-8.
94
-------
Tu±>iliy forRaw Sampfe= 176 ntu
100 -T
-H
T5
-H
10 -:
20mg,tJofE-2
. 40mg^ofE-2
0
05
1 15 2
PE -5 C onceatedbn frng/L)
25
MC-5: MC Size Range = 5375 |im; MC Concentration = 3 g/L
E-2: Ferric Chloride Concentration = 40 mg/L as Fe+++
PE-5: Polyelectrolyte 309C; Settling Time = 3 min
Figure 5-7. Polyelectrolyte Concentration Selection (Low Dose)
95
-------
-H
-H
4 -T-
3-1-
2 --
1 --
TuibJdJiy for Raw SampJe = 87 ntu
6 9 12
PE -5 C oncmtzatbn tn g/L)
15
18
MC-5: MC Size Range = 5375 |im; MC Concentration = 3 g/L
E-2: Ferric Chloride Concentration = 40 mg/L as Fe+++
PE-5: Polyelectrolyte 309C; Settling Time = 3 min
Figure 5-8. Polyelectrolyte Concentration Selection (High Dose)
MC Concentration
Four different MC concentrations, ranging between 1 and 7 g/L
were utilized to determine an optimal value. Figures 5-9(A) and
(B) illustrate turbidity and particle count rate distributions
versus four different MC concentrations, respectively. In Figure
(A), a sharp decrease in turbidity was observed for MC
concentrations between 1 and 3 g/L, with slight fluctuations at 5
and 7 g/L of MC. Although 7 g/L MC yields the lowest turbidity,
the improvement is only 23% greater compared with 3 g/L MC while
the MC concentration level is increased by 133%. Using particle
count rate as an indicator in Figure (B), 3 g/L MC appears to
yield the lowest result. Based on the results of both turbidity
and particle count rate, a 3-g/L of MC was selected as the
optimal concentration.
MC Size
Five different size ranges for MCs were employed in the tests.
Turbidity and particle count rate versus the five different MC
sizes, along with one sample without MC, are shown in Figure 5-10
(A) and (B), respectively. It can be seen that the size range
from 53 to 75 |im yields the best results for both turbidity and
particle count rate. Thus, this size range was selected as the
optimal MC size.
96
-------
-LU
8 -
- H 6 -
d
-H
a 4-
EH
2 -
n
i
i
i
40 -r
3 5
M JciDCBrrpr C oncmtetbn
(A) By Turbidity Indicator
PCR forRaw Samp]e= 353 kcps
4->
05
4J 30 -
m
i-(
Particle Coi
h-1 DO
3 O O
1 1 1
1
1357
M izrocariderConcaitiatiDn (g/L)
(B) By Particle Count Rate Indicator
MC Size Range = 5375 |im;
E-2: Ferric Chloride Concentration = 40 mg/L as Fe+++;
PE-5: Polyelectolyte 309C Concentration = 2 mg/L
Figure 5-9. MC Concentration Selection
97
-------
12 T
10 -
* 8
4J
-H
0 ,
-H 6
ft
iM
3
H A
2 -'
Turbrfty for Raw Sample = 82 ntu
+
+
4-1
03
O
U
U
-H
4J
5 T
3 -
None
< 53 53-75 75-106
M irEocander Size frn
(A) By Turbidity Indicator
106-150
150-250
Partdcle CountRatE for Raw Sample = 269 kcps
+
None
< 53 53-75 75-106 106-150
M icrocarrierSdze tn dcrom eters)
150-250
(B) By Particle Count Rate Indicator
MC Concentration = 3g/L
E-2: Ferric Chloride Concentration = 40 mg/L as Fe+++;
PE-5: Polyelectolyte 309C Concentration = 2 mg/L
Figure 5-10. MC Size Selection
98
-------
Settling Time
Settling time required is one of the key factors in the
coagulation-flocculation treatment process since settling
kinetics governs the treatment efficiency, duration, and thus the
size of the treatment facility. The MC process appears to provide
very efficient settling characteristics for CSO treatment.
Figure 5-11 shows results of turbidity versus settling time from
1 to 10 minutes. In both tests, the post-treatment turbidity was
reduced to 1.3 ntu with a minor fluctuation within a two-minute
settling period. The average turbidity removal rate was 98% over
the settling duration of 2 to 10 minutes.
Figure 5-12 illustrates a comparison of settling kinetics with
and without the use of MC, respectively. In this test, it can be
seen that the MC process enhances both short-term (3 minutes) and
long term (20 minutes) settling characteristics.
4J
-H
d
-H
ft
12 T
9 --
6 -
3 -
\
Raw SampteCcodidcn:
*~ Tudoiiily = 52 ntu
Tudoiiily = 72 ntu
4 6
Setting tin e frn ±L)
10
MC-5: MC Size Range = 5375 |im; MC Concentration = 3 g/L
E-2: Ferric Chloride Concentration = 40 mg/L as Fe+++
PE-5: Polyelectolyte 309C Concentration = 6 mg/L
Figure 5-11. Turbidity Versus Settling Time with Optimal Condition
99
-------
4J
-H
20 T
15 -'
io -
5 -
\
\
Tuitodiy for Raw Sample = 165 ntu
\
\
\
\
W ihoutM C
W ihM C
10
Setting Tine fo ±L)
15
20
MC-5: MC Size Range = 5375 |im; MC Concentration = 3 g/L
E-2: Ferric Chloride Concentration = 40 mg/L as Fe+++
PE-5: Polyelectolyte 309C Concentration = 1 mg/L
Figure 5-12. Turbidity Versus Settling Time
100
-------
5.4 Response Variable Evaluation
During the CSO treatment stage, the response variables included
the following parameters: turbidity, particle count rate, pH,
suspended solids, total solids, total volatile solids, particle
size distribution, total organic carbon, total inorganic carbon,
fecal coliform, and zeta potential. Among these parameters,
turbidity and particle count rate have been presented in Section
5.3. In this section, the results of the other response
variables are evaluated.
pH
Figure 5-13 illustrates pH distributions for two different raw
samples along with settling time from I to 10 minutes. It can be
seen that under the optimal control variable setup, all pH values
are within the range of 6.0 to 7.5.
Raw SampteConcLitian
4 6
Setting Tine fail)
10
MC-5: MC Size Range = 5375 |im; MC Concentration = 3 g/L
E-2: Ferric Chloride Concentration = 40 mg/L as Fe+++
PE-5: Polyelectolyte 309C Concentration = 6 mg/L
Figure 5-13. pH Versus Settling Time with Optimal Condition
101
-------
Suspended Solids
Figure 5-14 illustrates suspended solids versus settling time
with two different mixing periods, i.e., 2 and 4 minutes,
respectively. Excellent results for both samples were found to
be less than 10 mg/L with the exception of one sample with a one-
minute settling time. Suspended solids were found to be less
than 2 mg/L with a 4-minute mixing time. The removal efficiency
is 97% after 2 minutes of mixing (average from 3 to 10 minute
settling) and 99.5% after 4 minute of mixing (average from 1 to 7
minute settling), respectively. For both tests, the supernatant
suspended solids reach the lower measurement limit (0.1 mg/L)
after 7 minutes of settling.
Suspended Solids for Raw Sam p]e = 87 m g/L
ra
d
-H
rH
O
CO
d
-------
for surface runoff treatment, then the total solids removal could
be effective. However, if dissolved solids were the major
component in total solids, the total solids removal efficiency
would be significantly limited by the content of organic portion
of suspended solids in the raw sample. For the CSO treatment,
total solids removal rates are approximately 50% as illustrated
in Figure 5-15.
500 T
_ 400 -
ra
d
-H
rH
O
CO
nJ
4J
O
300
200 -
100
TotalSoUdsforRaw Sample = 780 mg^
10
Settling tin e frn ±L)
15
20
MC-5: MC Size Range = 5375 |im; MC Concentration = 3 g/L
E-2: Ferric Chloride Concentration = 40 mg/L as Fe+++
PE-5: Polyelectolyte 309C Concentration = 1 mg/L
Figure 5-15. Total Solids Versus Settling Time
Total Volatile Solids
As with total solids removal, the total volatile solids removal
efficiency was also limited by the content of organic portion of
suspended solids in the raw samples. It was found that the total
volatile solids removal rate was in the range of 60% to 80%. The
results of the two sets of jar tests with similar conditions but
different sample batches are illustrated in Figure 5-16.
103
-------
-H
600 T
500 +
CO
cu 400 t
rH
-H
nj 300 +
4J
200 -
100 -
Raw Sample
TVS = 440 mg^
TVS = 460 mg^
10
Settling Tine fail)
15
20
MC-5: MC Size Range = 5375 |im; MC Concentration = 3 g/L
E-2: Ferric Chloride Concentration = 40 mg/L as Fe+++
PE-5: Polyelectolyte 309C Concentration = 1 mg/L
Figure 5-16. Total Volatile Solids Versus Settling Time
Total Organic and Inorganic Carbon
Figures 5-17 (A) and (B) show the results of total organic and
inorganic carbon with respect to settling time for two different
raw samples. It appears that a 3-minute settling time would be
sufficient to achieve a stable condition. It was observed that
the average total organic and inorganic carbon removal rates over
the settling times from 3 to 20 minutes are approximately 50 and
70 percent, respectively. It was found that these removal rates
remained the same although the post-treatment total organic and
inorganic carbon values varied for different raw sample
conditions. These removal rates were limited by the fact that
only total organic carbon associated with the suspended solids
were removed while total organic carbon in the dissolved format
still remained in the solution after the MC process treatment.
Figure 5-18 illustrates comparison results from two different
mixing durations, 2- and 4-minute, respectively. It appears that
the influence of mixing duration (2- to 4-minute) is
insignificant with respect to total organic and inorganic carbon
removals.
104
-------
100 T
80 -
i-q
6
O 40
EH
20 -
50 -T-
R aw Sam p]e C ondition:
TOG = 137 mg^
TOG = 76 mg/L
5 10
Setting T in e {n in)
(A) Total Organic Carbon
R aw Sam pile C ondidon:
15
20
Oi
u
H
EH
40 -
30 -
20 -
10 -
n _
-^A TIC = 34 mg/L
TIC = 27 mg/L
V ^_ __ __
^ *~ ""*
« ^_ O
1 1 \ 1 \ 1 1 1 1 1
0 5 10 15 20 25
Setting T in e fn in)
(B) Total Inorganic Carbon
MC-5: MC Size Range = 5375 |im; MC Concentration = 3 g/L
E-2: Ferric Chloride Concentration = 40 mg/L as Fe+++
PE-5: Polyelectolyte 309C Concentration = 1 mg/L
Figure 5-17. Total Organic and Inorganic Carbon Distributions
(with different raw samples)
105
-------
Cn
30 -T-
20 --
TOC for Raw Samp]e= 44mg/L
^
u
o
EH
10 -
n _
1 U
2 m in m idng
-^A 4 m in m idng
1
l 1 1
0
30 -T-
20 --
468
Setting T in e fn in)
(A) Total Organic Carbon
TE for Raw Samp]e= 35mg/L
10
12
U
0
- 2mdnm dxdng
....
-^A 4 m urn King
1
D 2
i i i
468
Setting T in e fn in)
i i
10 12
(B) Total Inorganic Carbon
MC-5: MC Size Range = 5315 |im; MC Concentration = 3 g/L
E-2: Ferric Chloride Concentration = 40 mg/L as Fe+++
PE-5: Polyelectolyte 309C Concentration = 6 mg/L
Figure 5-18. Total Organic and Inorganic Carbon Distributions
(with different mixing duration)
106
-------
Fecal Coliform
Figure 5-19 shows fecal coliform distribution versus settling
time. It appears that for a 7-minute settling time, the fecal
coliform is reduced to approximately 300 per 100 mL. The removal
rate was observed as greater than 75% (considering the initial
concentration was 1200 per 100 mL).
FecalCoUfcim forRaw Sample > 1200 per 100 mL
31
e
4-1
-H
1000 T
800 -
600 -
Q 400
oj 200 f
> 1200
> 1200
> 1200
> 1200
3 5
Setting Tine fr ±L)
10
MC-5: MC Size Range = 5375 |im; MC Concentration = 3 g/L
E-2: Ferric Chloride Concentration = 40 mg/L as Fe+++
PE-5: Polyelectolyte 309C Concentration = 6 mg/L
Figure 5-19. Fecal Coliform Distributions
Relationship of Particle Count Rate and Turbidity
Figure 5-20 illustrates the particle count rate distribution
versus turbidity. A linear regression between these two
parameters was performed and the correlation coefficient is 0.94,
which indicates a strong relationship. Since the data ranges
over more than two magnitudes, a log-log distribution of particle
count rate and turbidity is shown in Figure 5-21. The log-log
correlation of these two parameters was found to be 0.8. A log-
log linear regression was also performed. The regression
equations are expressed as follows:
PCR = 3(Turbidity) +4.4
Log(PCR) = 0.91Log(Turbidity) + 0.54
For linear regression
For log-log regression
107
-------
500 T
C onslation C oeffiraent = 0.94
50
100
150
200
Figure 5-20. Particle Count Rate Versus Turbidity
1000 T Conslation Coeffir:iait = 0.8
S 100
O
U
(U
rH
U
-H
4J
10 "
1 "
0.1
Log(PCR)= 0.91Log(T)+ 0.54
0.1
10
ity (phi)
100
1000
Figure 5-21. Particle Count Rate Versus Turbidity (log-log)
108
-------
Zeta Potential
The relationship between zeta potential and turbidity is
illustrated in Figure 5-22. It appears that the treatment is
more effective when particles have less charge. A semi-log
regression was performed based on test results. However, it was
noted that in most CSO samples, zeta potential data fluctuated,
which may be due to the varying charge characteristics of raw
samples. The relationship between zeta potential and turbidity
in CSO treatment is not as clear as in surface runoff treatment.
100 T
4J
-H
d
-H
ft
10 -:
Raw Samp]e Ccodidcn
Tutbrliy = 58 ntu; Zeta potential = 6.9 mV
M easuianent
Regression
T = 20exp (0.05Z)
CortEbtion Coeffir:iait = 0.7
1
-15
-10 -5
Zeta Potential foV)
MC-5: MC Size Range = 5375 |im; MC Concentration = 3 g/L
E-2: Ferric Chloride Concentration: 2040 mg/L as Fe+++
PE-5: Polyelectolyte 309C Concentration: 618 mg/L
Figure 5-22. Turbidity Distribution Versus Zeta Potential
109
-------
Chapter 6
Summary and Recommendations
6.1 Summary
Based on the results of the MC weighted jar tests, the principal
conclusions can be summarized as follows:
1. Conventional jar test procedure was modified to fit the MC
weighted coagulation assessment.
2. In performing the MC weighted jar tests, three stages of
mixing were used. First, a rate of 150 rpm with 10 seconds
duration was used to mix coagulant with liquid and lift the
MC from the bottom of the jar into suspension. Second, a
mixing rate of 100 rpm with 10 seconds duration was used to
mix flocculent and maintain the MC in suspension. Third, a
rate of 60 rpm with 10 seconds duration was found to be
sufficient for the flocculation process in treating storm
runoff samples. In the CSO treatment process, one to two
minutes was needed for the third stage (flocculation
mixing).
3. For a better result of surface runoff treatment, the dosage
of aluminum sulfate concentration was found to be 40 mg/L
with the addition of 1 mg/L coagulant aid (flocculent).
4. From the CSO tests, the coagulant concentration was found to
be 190 mg/L as FeCl3»6H20 or 40 mg/L as Fe+++ with the
addition of 6 mg/L coagulant aid (flocculent).
5. Among the four MCs tested for surface runoff treatment, the
MC in a size range of 53 to 150 |im at a dosage of 10 g/L
yielded the best results in conjunction with the coagulant
and polyelectrolyte. For the CSO treatment, MC in a size
range of 53 to 75 |im at a dosage of 3 mg/L yielded better
results.
6. After three minutes of settling, the following removal rates
were achieved:
(A) Surface runoff testing:
98% for turbidity
94% for particles smaller than 5 |im
99% for particles larger than 2 |im
83% for total solids
95% for total volatile solids
(B) CSO testing
98% for turbidity
98% for particles smaller than 5 |im
99% for particles larger than 2 |im
110
-------
99% for suspended solids
50% for total solids
7. Relationships between the zeta potential and turbidity were
quantified by log-linear expressions for different
conditions (as presented in Table 4-3).
8. The relationship between particle count rate and turbidity
was quantified (as indicated in Figures 4-35, 5-20 and 5-
21) .
9. In view of high percentage of fine particles (< 5 |im)
reduction in both storm runoff and CSO, one may conclude
that the associated toxic heavy metals (as indicated in
Table 1-1) might be substantially removed by the MC weighted
coagulation.
6.2 Recommendations
The MC weighted coagulation is a new process, and many factors
that affect the aspects of coagulation and flocculation bench-
test (jar test) are still not well understood. For future
studies, the following investigations are recommended:
1. Other types of coagulants and coagulant aids (flocculent or
polyelectrolytes) should be evaluated.
2 . In order to verify the effect of MC weighted coagulation on
toxic heavy mattes removal as evidenced in the high
percentage (> 90%) of fine particles (< 5 |im ) reduction in
MC jar test, measurement of additional parameters as listed
in Table 1-1 should be included in experimental design as
respond water quality parameters.
3 . MC weighted coagulation kinetics should be studied in
detail. The floe growth and shear rates are related to
particle and floe concentrations and sizes, particle
collision and stability coefficient, floe breakup constant,
and velocity gradient. To investigate the flocculation
kinetics, the velocity gradient, particle size and
concentration should be controlled to evaluate the particle
collision and stability coefficients as well as to develop a
floe breakup constant. A model of MC kinetics should be
investigated in order to develop greater MC process
efficiency.
3 . Prototype studies are necessary to expand the present bench
scale level (i.e., jar testing) in order to provide
engineering design parameters and operational procedures for
utilization in a treatment plant. In order to perform on-
site evaluation, a mobile unit should be developed for this
purpose.
Ill
-------
References
American Society for Testing and Materials. "Annual Book of ASTM Standards." Vol 11.01, Philadelphia, PA.
1996.
Association of Environmental Engineering Professors. "Environmental Engineering Unit Operations and Unit
Processes laboratory Manual." University of Texas at Austin, Texas. 1971.
Black, A., A. Buswell, andF. Eidsness. "Review of the Jar Test." J. AWWA, Vol. 39, No. 11, pp 1414, 1957.
Black, A. and R. Harris. "New Dimensions for the Old Jar Test." Water and Wastes Engineering, Dec:49 (1969).
Boudries, H., C. Broguet, C., J-M. Mouchel, andD.R. Thevenot. "Urban Runoff Impact on Composition and
Concentration of Hydrocarbons in River Seine Suspended Solids. " Proc. 7thlnt. Conf. on Urban Storm Drainage.
Hannover, Germany, 569-574. 1996.
Camp, T. "Floe Volume Concentration." J. AWWA, Vol. 60, No. 6, pp 656. 1968.
Chebbo, G., P. Musquere, V. Milisie and A. Bahoc (1990). "Characterization of Solids Transferred into Sewer
Trunks During Wet Weather. " Wat. Sci. Tech., 22(10/11) 231-240.
Clark, M. M. 1996. "Transport Modeling for Environmental Engineers and Scientists." John Wiley & Sons, Inc.,
NY.
Cohen, J. "Improved Jar Test Procedure." J. A WWA, Vol. 49, No. 11, pp 1425, 1957.
Derjaguin, B. and L. Landau. Theory of the Stability of Strongly Charged lyophobic Sols and of the Adhesion of
Strongly charged Particles in Solutions of Electrolytes. Acta Physicochim. USSR, Vol. 14, pp 733-762, 1941.
Ellis, J.B. and Revitt, D.M. Incidence of Heavy Metals in Street Surface Sediments: Solubility and Grain Size
Studies. Water, Air, and Soil Pollution, 17(1)87-95. (1982).
Estebe, A., Belhomme, G., Lecomte, S., Videau, V., Mouchel, J-M., and Thevenot, D.R. "Urban Runoff Impacts on
Paniculate Metal Concentrations in River Seine: Suspended Solid and Sediment Transport." Proc. 7th Int. Conf. on
Urban Storm Drainage, Hannover, Germany, 575-580. (1996).
Gregory, J. The role of Colloid Interactions in Solid-Liquid Separation. Water Science and Technology, Vol. 27, No.
10, pp 1-17,1993.
Han, M., and Lawler, D. F. Interactions of Two Settling Spheres: Settling Rates and Collision Efficiency. Journal of
Hydraulic Engineering, Vol. 117, No. 10, pp 1269-1289, 1991.
Lopes, T. J. and Possum, K. D. Selected Chemical Characteristics and Acute Toxicity of Urban Stormwater,
Streamflow, and Bed Material, Maricopa County, Arizona. Water-Resources Investigations Report 95-4074, U.S.
Geological Survey, Denver, Colorado (1995).
110
-------
High Rate Microcarrier-Weighted Coagulation for Treating
Wet Weather Flow
By
Chi-Yuan Fan (1), Yuan Ding(2), Shih-Long Liao (3),
Richard Field (4), Paul C. Chan (5), and Robert Dresnack(5),
ABSTRACT
In wet-weather flow (WWF) a significant amount of toxic pollutants are associated with
colloidal solids. Sedimentation with coagulation is considered to be an effective method
of removing large quantities of suspended solids from water and wastewater. A new
high-rate settling process, using microsand or a microcarrier (MC) as a settling carrier of
colloids, has been developed and applied for treating both wet- and dry-weather flows
(WWF/DWF). The addition and recirculation of the MC results in higher settling
velocities and allowable tank overflow rates. The MC plays a crucial role in enhancing
settling properties, and in particular, the removal of colloidal particles and associated
contaminants.
A detailed testing procedure and a method of experimental analysis using a modified jar
test for the MC process have been developed. A series of MC weighted jar tests were
undertaken on parking lot storm runoff, synthetic samples, and combined sewer overflow
(CSO) mixed with a MC, coagulant and coagulant aid. Two particle analyzers with a
range of 0.002 to 5 micrometers (|im) and 0.1 to 2,000 |im, respectively, were used to
determine the full range of particle size distribution. Different materials were used as the
MC in this study. The operational parameters evaluated include coagulant dosage,
coagulant aid type and dosage, mixing- and flocculation-induced hydraulic shear or
velocity gradients and duration, and characteristics of the MC. The pH, turbidity, particle
size distribution, total solids, total volatile solids, suspended solids, and zeta potential
were determined. The experimental results revealed that MC weighted coagulation
dramatically reduced coagulation-flocculation duration (< 3min) and settling time (< 8
min) producing floes with high settling velocity and high quality supernatant. Reductions
of turbidity from > 80 NTU (raw samples) to < 2 NTU (supernatant samples) were
achieved.
U) Environmental Engineer,(3) ORISE Research Engineer at the time of this study, andt4) Wet-Weather
Flow Research Program Leader, Urban Watershed Management Branch, Water Supply and Water
Resources Division, National Risk Management Research Laboratory, U.S. Environmental Protection
Agency, Edison, NJ 08837.
(2)Assistant Professor and (5)Professor, Department of Civil and Environmental Engineering, New Jersey
Institute of Technology, Newark, NJ 07102.
-------
INTRODUCTION
A significant amount of toxic pollutants in wet weather flow (WWF) are associated with
smaller particles (< 10 |im), including toxic-organic chemicals (e.g., benzene,
polynuclear aromatic hydrocarbons [PAH], and polychlorinated biphenyls [PCB]) and
heavy metals (e.g., arsenic, cadmium, chromium, copper, lead, mercury, and zinc).
Sedimentation with coagulation is generally considered to be an economical method of
removing large quantities of suspended solids (SS) from water and wastewater. A new
high-rate settling process, using microsand or a microcarrier (MC) as a settling carrier of
colloids, has been developed and applied for treating both wet- and dry-weather flows
(WWF/DWF). The addition and recirculation of MC will result higher settling velocities
and allowable tank overflow rates. This process is also known as "ballasted-
coagulation." It has been reported that an MC unit would increase solids removal by
more than 80% at a range of overflow rates of 50 to 100 m3/h/m2 (20 to 40 gpm/ft2). This
high-rate process consists of the addition of MC, coagulant, and coagulant aid with the
influent in a mixing chamber followed by coagulation-flocculation and sedimentation.
The MC plays a crucial role in enhancing settling properties, and in particular, the
removal of colloidal particles and associated contaminants.
Objectives. The objectives of this investigation were:
Evaluate the applicability of conventional j ar test procedure to the MC weighted
coagulation, and if needed, modify the jar test procedure for screening MCs,
coagulants, and coagulant aids;
Test the effects of different types and dosages of coagulant and coagulant aid in
conjunction with MC for selection of the most effective combination;
Investigate the effect of MC on particle size distributions and zeta potential of
colloids in urban WWF by using the modified jar test procedure.
METHODOLOGY
Materials
MC. Ottawa sand was selected, due to its abrasion resistance, and small and uniform
size. The size range of Ottawa sand tested was between 100 and 500 |im. The MC size
ranges and concentrations were determined by prescreening tests.
Coagulants. Alum (aluminum sulfate, Al2(SO4)3*18H2O)and ferric chloride
(FeCl3*6H2O) were used as the coagulants for this study.
Coagulant Aids. Five polyelectrolytes from two different manufacturers were used in the
experiments. Among them, four (POL-EZ-2466, POL-EZ-3466, POL-EZ-2696, and
POL-EZ-7736, by Calgon Corporation, Pittsburgh, Pennsylvania) were used in surface
runoff tests and one (309C, by Polydyne, Inc., Riceboro, Georgia) for combined sewer
overflow (CSO) tests.
-------
Apparatus and Instruments
The descriptions for the major test apparatus and analytical instruments used for the
investigation, including the measured parameters, range, model number, and
manufacturer, are briefly described as follows:
Particle size analyzer I, designated as PSA-I (Master-Sizer X, Malvern
Instruments Inc.), is a large particle size analyzer with a measurement range from
0.1 to2000|im.
Particle size analyzer II, designated as PSA-II (90Plus with ZetaPlus, Brookhaven
Instruments Corporation), is a small particle size analyzer with a measurement
range from 0.002 to 5 |im. PSA-II is equipped with disposable sample cells that
eliminates the cross sample residue influence.
The zeta potential meter and PSA-II are integrated in one unit. The measurement
range of the zeta potential meter is from 0.1 to 200 mV with the particle size range
from 0.002 to 30 |im. The resolution is sample dependent and in the range of 0.1% to
5%.
Jar test apparatus: A Phipps and Bird (Model PB-700). Dimensions of each jar
are 11.5 X 11.5 X 21 cm depth which is capable of testing a water sample volume
of 2,000 ml. Each mixer is equipped with a flat stirring paddle (7.6 X 2.5 cm or
19.3 cm ). The area of the paddle was increased to 38.7 cm for the MC jar test in
order to generate more rigorous turbulence for keeping the microsand in
suspension.
Turbidity meter.
pH meter.
Balance.
MC Weighted Jar Test
The procedures include the following steps:
1. Collect storm surface runoff sample, prepare synthetic sample, or CSO sample.
Measure the sample for pH and turbidity.
2. Pour 1,000 ml of the water sample into each two-liter j ar of the j ar-test apparatus
and check stirrer operation. A light table facilitated viewing the contents of the
jars.
3. Add controlled amounts (dosages) of MC, coagulant, and coagulant aid to the
jars.
4. Flash mix for 20 to 60 s at 100 to 200 rpm.
5. Slow mix for 10 to 120 s at 30 to 60 rpm. Record the elapsed time before a
visible floe is formed. If large floes are formed, reduce the paddle speed. Record
the appearance of the floe formed.
6. After flocculation, remove paddles and settle for 2 to 30 min.
-------
7. Collect the supernatant from the sampling port on each jar and measure turbidity;
the settled solids should not be disturbed during sampling. Select and record the
dosage of coagulant and coagulant aid based on the supernatant clarity and
settleability of floe.
RESULTS
The followings summarize the results of MC jar tests with surface runoff and CSO:
Surface Runoff
Figure 1 illustrates cumulative volume distributions versus particle size before and after
the MC weighted coagulation. For the raw sample (before treatment), the measurable
range of particle size is from 1 to 170 |im which consists of 81% from 10 to 100 |im and
14% smaller than 10 |im. After the MC coagulation, the particles in the supernatant of
the jar were found to be < 2 |im, which indicated that all particles > 2 |im were
removed.
ctf
o
100 -,
80 -
60 -
40 -
20 -
0 -
?
-
-
-
1
->*!-
0.1 1
x Treated sample
Raw sample
10 100
Particle Size (am)
1000
MC Size Range: 53 to 150 |j,m; MC Concentration: 3 g/L
Aluminum Sulfate Concentration: 40 mg/L
Cationic Polyelectrolyte (POL-EZ-2466) Concentration: 1 mg/L
Figure 1. Particle Sizes of Raw and Treated Surface Runoff Samples (Cumulative)
Figure 2 (A) and (B) illustrate the particle size distribution characteristics of the raw
sample and supernatant sample after MC treatment, respectively. For the raw sample, the
distribution peak is at approximate 50 |im while the small particles (from 0.7 to 2 |im)
were either undetectable or only at a low percentage compared to the peak. After
treatment, only particles < 2 |im remained in the supernatant. Since the larger particles
were all removed from the raw sample, the percentage of particles < 2 |im increased
significantly. The removal efficiency for particles < 2 |im was measured by the particle
count rate (PCR) that is directly related to the number of particles counted per unit time
in a solution expressed as kilo-count per second (kcps). Results of PCR are indicated in
Figure 3.
-------
QJ
O
1000
cu
40 -T-
30 --
20 --
> 10 +
(B) Treated Sample
0.1
1 10
Particle Size
100
1000
MC Size Range: 53 to 150 |im; MC Concentration: 3 g/L
Aluminum Sulfate Concentration: 40 mg/L
Cationic Polyelectrolyte (POL-EZ-2466) Concentration: 1 mg/L
Figure 2. Particle Size of Raw and Treated Surface Runoff Samples (Distributions)
MC jar tests were conducted with both coagulants and coagulant aids. Figure 3 (A) and
(B) present supernatant turbidity and PCR results with and without MC. Both turbidity
and PCR of supernatant samples with MC are much lower than those without MC. It is
apparent that the addition of MC improves the treatment process effectively.
-------
Tutbiiiy of Raw Sample = 74NTU
to ~~'
m >i
S -H
CU 13
a H
CO &-I
H-i
8 -
6 -
4 -
2 -
n _
1 1 1
1
W d
C
W JihoutM C
(A) Coagulant Aid (Anionic Polyelectrolyte [POL-EZ-7736])
1
20 -y PCR of Raw Sample = 35 kops
~ 10 -
u 5 -
W JfllM C
W JihoutM C
(B) Coagulant Aid (Non-ionic Polyelectrolyte [POL-EZ-2696])
MC Size Range: 53 to 150 |j,m; MC Concentration: 3 g/L
Aluminum Sulfate Concentration: 40 mg/L
(A) and (B) Coagulant Aid Concentration = 1 mg/L
Figure 3. Effectiveness of the MC Process (Surface Runoff Samples)
CSO
Figures 4 and 5 illustrate cumulative and discrete volume distributions versus particle
size of raw sample and supernatant after MC weighted coagulation, respectively. For the
raw sample (before treatment), the measurable range of particle size is from 0.5 to 60 |im
with 93% (by volume) >2|im. After the MC coagulation, the particles in the
supernatant of the jar were found to be < 2 |im, indicating all particles > 2|imwere
removed.
-------
-H
-U
m
iH
3
e
3
u
U
^_^
cu
g
3
iH
£
100 -,
80 -
60 -
40 -
20 -
0 -
0.1
1
10
Treated sample
Raw sample
I
100
1000
Particle Size
MC Size Range: 53 to 75 \im; MC Concentration: 3 g/L
Ferric Chloride Concentration: 40 mg/L(as Fe^)
Cationic Polyelectrolyte (309C) Concentration: 1.5 mg/L
Figure 4. Particle Sizes of Raw and Treated CSO Samples (Cumulative)
15 -
0.1
(A) Raw Sample
10 100
Particle Size (//m)
1000
0.1
(B) Treated Sample
1 10
Particle Size (//m)
100
1000
MC Size Range: 53 to 75 |im; MC Concentration: 3g/L
Ferric Chloride Concentration: 40 mg/L (as Fe^+)
Cationic Polyelectrolyte (309C) Concentration: 1.5 mg/L
Figure 5. Particle Sizes of Raw and Treated CSO Samples (Distributions)
-------
The effect of the MC was evaluated using turbidity and particle size distribution as
indicating parameters. Figure 6 presents results of supernatant turbidity versus settling
time with and without MC.
Turbidity of Raw Sample = 85 NTU
^ 50 -p
S H 40--
13
" s
* *
i i
0 5 10
-^ Without MC
With MC
1
1
15
A
I
1
20
Setting Tin e tn in)
MC Size Range: 53 to 75 |im; MC Concentration: 3 g/L
Ferric Chloride Concentration: 40 mg/L (as Fe^+)
Cationic Polyelectrolyte (309C) Concentration: 1 mg/L
Figure 7. Turbidity versus Setting Time
-------
CONCLUSIONS
Based on the results described above, conclusions can be summarized:
1. A testing procedure for experimental analysis of the MC process has been
developed.
2. The MC process was found to be effective in reducing coagulation-sedimentation
time, which, in turn, will reduce the sizing requirements of the process needed
compared to a conventional physical-chemical treatment facility. The total
treatment time could reduce to < 10 min utilizing the MC-weighted coagulation-
sedimentation process. Although stormwater and CSO samples with varying
influent characteristics can be anticipated to react differently to the MC process, it
is expected that the tremendous reduction in settling time found in this study will
generally extend (in varying degrees) to all storm runoff and CSO samples.
3. In performing the MC-weighted jar tests (i.e., the modified jar test developed
under this study), three stages of mixing, with 10s duration in each stage, were
used: (1) 150 rpm to lift the MC from the bottom of the jar into suspension, (2)
100 rpm to keep the MC in suspension, and (3) 60 rpm for the coagulation-
flocculation process to treat storm runoff samples. For CSO treatment, the third
stage duration for coagulation-flocculation was found to be optimal at 1 to 2 min.
4. The best combination coagulant/coagulant aid in the surface runoff MC study,
was found to be 40 mg/L aluminum sulfate (coagulant) and 1 mg/L cationic
polyelectrolyte, POL-EZ-2466 (coagulant aid). For CSO 40 mg/L coagulant
(ferric chloride, as Fe+++) with 1 mg/L coagulant aid (cationic polyelectrolyte,
309C) resulted in the best turbidity removal.
5. In the MC size and concentration study, the range of 53 to 150 |im at a dosage of
10 g/L yielded the best results for surface runoff. For CSO treatment, MC in a
size range of 53 to 75 |im at a dosage of 3 g/L resulted in the lowest supernatant
turbidity.
RECOMMENDATIONS
The following is recommended for the future:
1. Additional jar tests to refine the MC coagulation process for removal of toxic
pollutants and microorganisms in urban WWF with different types of coagulant
and coagulant aid.
2. Pilot studies to verify and expand upon the present bench-scale level (i.e., jar
testing) in order to provide engineering design and operational procedures for
full-scale MC systems.
-------
BIBLIOGRAPHIES
1. American Society for Testing and Materials. 1996. Annual book ofASTM standards,
Vol. 11.01, Philadelphia, PA.
2. Association of Environmental Engineering Professors. 1984. Environmental
engineering unit operations and unit processes laboratory manual., University of Texas at
Austin, Texas.
3. ASCE and AWWA, 1971. Water treatment plant design, AWWA Inc., Denver, CO
4. Black, A., Buswell, A., and Eidsness, F. 1957. "Review of the Jar Test," J. AWWA,
Vol. 39, No. 11, p. 1414.
5. Black, A. and Harris, R. 1969. "New dimensions for the old jar test," Water and
Wastes Engineering, Dec:49.
6. Camp, T. 1968. "Floe volume concentration," J. AWWA, Vol. 60, No. 6, p. 656.
7. Cohen, J. 1957. "Improved jar test procedure." J. AWWA, Vol. 49, No. 11, p. 1425.
8. Guibelin, E. 1994. "Actiflo process: a highly compact and efficient process to prevent
water pollution by stormwater flows," Water Science and Technology, Vol. 30, No. 1 .
9. "Methods for chemical analysis of water and wastes," EPA-600/4-79-020, Revised
March 1983.
10. Standard Methods for the Examination of Water and Wastewater, 18th edition
supplement, 1995. American Public Health Association, Washington, DC.
10
-------
Maltby, L., etal. The Effects of Motorway Runoff on Freshwater Ecosystems: 1. Field Study. Environ. Toxicol.
AndChem. 14(6): 1079 (1995).
McGregor, L, Ashley, R. M., and Oduyemi, K. O. K., Pollutant Release from Sediments in Sewer Systems and Their
Potential for Release into Receiving Waters. Wat. Sci. Tech. 28(8/9)161-168 (1993).
O'Melia, C. R. and Tiller, C. L. "Physicochemical Aggregation and Deposition in Aquatic Environmental." in
Environmental Particles, Vol. 2, Edited by Buffle, J. and van Leeuwen, H., Lewis Publishers, N.Y. 1993.
Pitt, R., R. Field, M. Lalor, and M. Brown Urban Stormwater Toxic Pollutant Assessment, Sources, and Treatability.
Water Environ. Research. 67(3) 260-275 (1995).
Pruppacher, H. R., and J.D. Klett. Microphysics of Clouds and Precipitation. D. Reidel Publishing Company,
Dordrecht, Holland, 1978.
Russel, W. B., D. A. S. Saville, and W.R. Schowalter. Colloidal Dispersions. Cambridge University Press. NY.
1989.
Sansalone, J.J., etal. Correlations Between Heavy Metals and Suspended Solids in Highway Runoff: Implications
for Control Strategies. Transportion Research Record, 1483(7): 12, 1995.
Standard Methods for the Examination of Water and Wastewater, 18th edition supplement. American Public Health
Association, Washington, DC. 1995.
Stumm, W. and J. Morgan. Aquatic Chemistry. Third edition. Wiley, N.Y. 1995.
U.S. Environmental Protection Agency Methods for Chemical Analysis of Water and Wastes, EPA-600/4-79-020.
Cincinnati, OH 1983.
Verwey, E. J. W. and J. Overbeek. Theory of the Stability ofLyophobic Colloids, Elsvier Publisher, N.Y. 1948
Vignoles, M. And L. Herremans. Metal Pollution of Sediments Contained in Runoff Water in the Toulouse City.
NOVATECH 95, Second International Conference on Innovative Technologies in Urban Storm Drainage, May 30-
June 1, 1995. Lyon, France, Organized by Eurydice 92 and GRAIE. 611-618.
Wen, C. S. The Fundamentals of Aerosol Dynamics. World Scientific Publishing Co, River Edge, NJ, 1996.
Ill
------- |