Enhanced removal of Cryptosporidium oocysts by filter media coated with hydrous iron aluminum oxides

MISSING IMAGE

Material Information

Title:
Enhanced removal of Cryptosporidium oocysts by filter media coated with hydrous iron aluminum oxides
Physical Description:
vii, 80 leaves : ill. ; 29 cm.
Language:
English
Creator:
Shaw, Kathryn M
Publication Date:

Subjects

Subjects / Keywords:
Water -- Purification -- Filtration   ( lcsh )
Water -- Purification -- Disinfection   ( lcsh )
Cryptosporidium   ( lcsh )
Environmental Engineering Sciences thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Environmental Engineering Sciences -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2001.
Bibliography:
Includes bibliographical references (leaves 74-79).
Statement of Responsibility:
by Kathryn M. Shaw.
General Note:
Printout.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 027753846
oclc - 48540047
System ID:
AA00017643:00001


This item is only available as the following downloads:


Full Text











ENHANCED REMOVAL OF Cryptosporidium
OOCYSTS BY FILTER MEDIA COATED WITH HYDROUS IRON ALUMINUM
OXIDES













By

KATHRYN M. SHAW


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


















To my mother and the three men of my life,
my husband, my father, and Nickolai.














ACKNOWLEDGMENTS

I would like to express my gratitude to the many individuals with which I have had

the pleasure to work during the course of my doctoral studies. I thank Dr. Ben Koopman

for mentoring my research and Dr. Hassan El-Shall for his ability to focus on

fundamental science. I would also like to thank Dr. Joseph Delfino, Dr. Sam Farrah and

Dr. Spyros Svoronos for their valuable recommendations as dissertation committee

members.

In addition, I would like to thank Dr. Jerzy Lukasik of the Department of

Microbiology and Cell Science, University of Florida; and Dr. Joan Rose and Chuck

Gibson of the Department of Marine Science, University of South Florida, St. Petersburg

for guiding an engineer through microbiology.

Finally, I would like to thank Dr. Brij Moudgil and the other members of the NSF

Engineering Research Center (ERC) for Particle Science and Technology for their

support and helpful discussions.

I would like to acknowledge the financial support of the Engineering Research

Center for Particle Science & Technology at the University of Florida, the National

Science Foundation (NSF) grant #EEC-94-02989, and the Industrial Partners of the ERC.















TABLE OF CONTENTS

Pama:

ACKNOWLEDGMENTS............................................................... ....................iii

ABSTRACT....................................... ................. .............. vi

CHAPTERS

1 INTRODUCTION ..................................................................................... 1

2 LITERATURE REVIEW............................................ .........................4

2-1 Cryptosporidium ..................................... ...........................4
2-1-1 G eneral..................................................... ........................... 4
2-1-2 Public Health Perspective........................... ........................7
2-1-3 Filtration of Cryptosporidium.......................... .......................... 7
2-2 Microbial Attachment to Surfaces........................... ...................9
2-2-1 Electrostatic interactions................................................................... 9
2-2-2 van der Waals Forces.............................. ..........................11
2-2-3 DLVO Theory........................................................ 11
2-2-3-1 Zeta potential..................................... ........................ 13
2-2-3-2 Hamaker constant ............................ ....... ........... ............ .. 14
2-2-3-3 Dielectric constant ......................... ........................... 14
2-2-4 Non-DLVO Interactions.......................... .............. .......... 14
2-2-4-1 Hydration force.......................... .......................... 15
2-2-4-2 Hydrophobic forces............................ ........................... 15
2-2-4-3 Macromolecule-induced forces .............................................. 15

3 EFFECT OF HYDROUS IRON ALUMINUM OXIDE COATING ON SAND IN
THE FILTRATION OF Cryptosporidium OOCYSTS.................................... 17

3-1 Introduction............................................................ 17
3-2 Materials and Methods.................................... ........................... 19
3-2-1 Coating M ethod............................................... ......................... 19
3-2-2 Sand Columns............................................................ 19
3-2-3 Counting of Cryptosporidium Oocysts ........................................... 20
3-2-4 Zeta Potential Measurements ...................................................22
3-3 Results and Discussion...........................................................................23
3-3-1 Physical Characterization of Cryptosporidium and Sand ..............23










3-3-2 Attainment of Pseudo-Steady-State in Filtration Experiments ........ 30
3-3-3 Effect of Coating on Filtration Performance..................................33
3-4 Conclusions ...................................................... .......................... 39

4 EFFECT OF IRON ALUMINUM (HYDR)OXIDE COATING ON REMOVAL
OF Cryptosporidium OOCYSTS BY GLASS AND CERAMIC BEADS...............40

4-1 Introduction .......................... ....... ...........................40
4-2 Materials and Methods.................................................41
4-2-1 Filter Media Preparation......................... .................................41
4-2-2 Filter Media Characterization......................................................42
4-2-2-1 SEM ............................. .............. .............................42
4-2-2-2 Surface area ........................ ............. .......................42
4-2-2-3 Zeta potential................................................................... 43
4-2-2-4 Surface metal content.... ........................... ........................44
4-2-3 Performance Testing of Filter Media..............................................44
4-2-3-1 Batch tests ..................................................................44
4-2-3-2 Column tests..................................... ................................45
4-2-3-3 Counting of Cryptosporidium oocysts...................................46
4-3 Results and Discussion..................................... ..........................47
4-3-1 Characterization of Glass Bead Surfaces........................................47
4-3-2 Characterization of Ceramic Bead Surfaces................................... 57
4-3-3 Comparison of Filter Media Performance.......................................62
4-4 Conclusions ................................................. 71

5 CONCLUSIONS ........................................ .........................................72

6 REFERENCES........................................................... ......................... 74

BIOGRAPHICAL SKETCH ............................................ ...............................80















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

ENHANCED REMOVAL OF Cryptosporidium OOCYSTS
BY FILTER MEDIA COATED WITH HYDROUS IRON ALUMINUM OXIDES

By

Kathryn M. Shaw

August 2001

Chairman: Dr. Ben Koopman
Major Department: Department of Environmental Engineering Sciences


In the past 10 years, Cryptosporidium oocysts have been shown to be common

contaminants in water, causing at least 20 outbreaks of cryptosporidiosis affecting almost

500,000 individuals in the United States alone. Disinfection processes generally have

had limited success in removing Cryptosporidium from water. The success of

conventional filtration on the removal of Cryptosporidium oocysts is limited by filter

operating conditions and chemical conditioning.

A surface coating of hydrous aluminum and iron oxide on Ottawa sand was

investigated as a means of improving the removal of Cryptosporidium oocysts from water

by filtration. Coating the sand increased the zeta potential from -40 mV to +45 mV at pH

7.0, enhancing the potential for interaction with the negatively charged (-25 mV at pH

7.0) Cryptosporidium oocysts. Water seeded with Cryptosporidium oocysts was passed

through parallel columns of uncoated and coated sand at superficial velocities of 200 to









800 m/d (3.5 to 14 gal/(ft2 min)) and column lengths of 10 to 40 cm (4 to 16 in.). In

every trial, removals of the oocysts with coated sand were significantly higher than

removals with uncoated sand. Filter coefficients of coated sand were 2.9 times higher

than those of uncoated sand, indicating that performance of granular media filters for

Cryptosporidium removal can be significantly enhanced by the coating technology.

Iron aluminum (hydr)oxide was coated on glass and ceramic beads by in situ

precipitation from metal chloride solutions, followed by contact with ammonium

hydroxide. Zeta potential at pH 6 to 8 of glass beads was increased from negative to

positive values by the coating, whereas zeta potential of ceramic beads was positive both

before and after coating. Removal of Cryptosporidium oocysts in batch tests and

continuous-flow column tests was significantly improved by coatings on both types of

beads. These results show that, while electropositive character is the most important

factor in design of coatings for granular media to enhance filtration of Cryptosporidium,

other factors (e.g., affect of coating on surface area and surface roughness of filter media)

should also be considered.














CHAPTER 1
INTRODUCTION

Waterborne diseases are a danger to public health in the United States and around

the world. The presence of Cryptosporidium oocysts in water supplies is currently a

significant public health concern because of the association of the microbes with surface

runoff, and because of their role in several large outbreaks of gastroenteritis. In the past

several years, Cryptosporidium has been responsible for at least 20 outbreaks of

cryptosporidiosis affecting approximately 500,000 individuals in the United States alone

(Rose, 1998).

Disinfection is a water treatment process commonly used to prevent microbes

from contaminating drinking water. Although the use of disinfection has greatly reduced

the number of outbreaks of some waterborne illnesses, it has had limited success in the

inactivation of some waterbore pathogens, such as Cryptosporidium oocysts. One U.S.

study (Fayer et al., 1996) found that 90% of a sample of Cryptosporidium oocysts were

still viable after spending 2 hours in undiluted household bleach. Since disinfection is

not a reliable barrier to Crypstosporidium contamination, physical removal of the oocysts

is necessary.

Filtration is a process used to prevent microbial contamination of drinking water.

Several studies (Fogel et al., 1993; Ongerth and Pecoraro, 1995; Schuler et al., 1991;

Whitmore and Carrington, 1993) have examined the effectiveness of sand filtration of

Cryptosporidium oocysts. They show that the success of conventional filtration in the

removal of Cryptosporidium oocysts from water is highly dependent upon (and often

1







2

limited by) filter operating conditions and chemical conditioning of the water. The

development of a method of increasing the reliability of sand filtration for the removal of

Cryptosporidium oocysts is needed.

Studies show that coatings of metallic hydroxides, oxides, and peroxides on filter

media enhance the removal of bacteria, viruses, and turbidity from water and wastewater

(Ahammed and Chaudhuri, 1996; Chang et al., 1997; Chen et al., 1998; Farrah and

Preston, 1991; Gerba et al., 1988; Lukasik et al., 1996; Lukasik et al.,1999; Mills et al.,

1994; Shaw et al., 2000; Stenkamp and Benjamin, 1994). The filter media, which are

electronegative in their natural state, become electropositive as a result of the coating

process. Since the particles are electronegative in the natural water pH range of 6 to 9,

the enhanced removal has been attributed to a decrease in electrostatic repulsion between

the particles and the coated filter media.

The size of microbes in previous studies of filter media surface modification ranged

from 0.5 to 2.0 gm for bacteria, down to 40 to 60 nm for viruses. Cryptosporidium

oocysts cover a size range of 4 to 8 pm (Levine, 1984). Filter media coating technology,

to be commercially viable for water purification, must be shown to be effective against

particles in this size range also. Along with size differences, other factors, such as

surface composition, surface characteristics and diffusivity in water can influence the

removal of Cryptosporidium oocysts from water. The first phase of this study examines

the effectiveness of the iron aluminum (hydr)oxide coating on Ottawa sand for the

removal of Cryptosporidium oocysts.

A few investigators have cited other changes in the filter media as a result of the

coating process as contributing factors in enhanced particle removal. Factors include

surface area and surface roughness, surface chemical heterogeneity, and amphoteric







3

surface charge (Stenkamp and Benjamin, 1994; Sansalone, 1999) Long-term studies

show that the electropositive surface of the filter media can revert back to electronegative

in time and still remove significantly more microbes than uncoated media (Chen et al.,

1998). Thus, other changes in the characteristics of the filter media surface due to

coating must also play a role in enhancing filtration performance. The second phase of

this study examined the effect of changing the zeta potential of the filter media on the

removal of Cryptosporidium oocysts from water.

The following specific objectives were chosen for this study because of knowledge

gaps revealed above:

1. Evaluate the aluminum and iron (hydr)oxide coating on filter media in terms

of its ability to enhance the removal of Cryptosporidium oocysts using varied

column lengths and superficial velocities.

2. Quantify the non-electrostatic contribution of iron aluminum (hydr)oxide

coating on granular filter media (glass and ceramic beads) to removal of

Cryptosporidium oocysts.















CHAPTER 2
LITERATURE REVIEW

2-1 Cryptosporidium

2-1-1 General

Cryptosporidium is a genus of 21 known species of enteric coccidian protozoans. It

was first described in 1907 by E.E. Tyzzer (Rose, 1990). The organism was isolated

from gastric mucosa in the stomach of a laboratory mouse, and named C. muris. Four

years later, a second species that inhabited the small intestine of the mouse, was

identified (Tyzzer, 1910; Tyzzer, 1912). This species was named C. parvum.

In 1955, Cryptosporidium was discovered as a source of disease in agricultural

animals. It was described during an outbreak of diarrhea in a domestic turkey flock

(Slavin, 1955), and in 1971, Cryptosporidium was found to be the source of diarrhea in a

herd of dairy cows (Panciera et al., 1971). Many studies over the next 10 years found

that Cryptosporidium was a significant cause of illness, and often mortality, in calves and

lambs (Anderson et al., 1982, Angus et al., 1982, and Current et al., 1983).

Cryptosporidium was not recognized as a source of human illness until 1976. At this

time, an infection was found in an immunocompromised individual (Meisel et al., 1976)

and an apparently otherwise healthy 3-year-old (Nime et al., 1976). Since then, reports of

cryptosporidiosis in humans have increased. Most infections were found in patients with

compromised immune systems (Navin and Juranek, 1984). As the prevalence of

Acquired Immune Deficiency Syndrome (AIDS) increased, Cryptosporidium was shown

to be a causative agent in diarrhea in AIDS patients. Since infection by Cryptosporidium

4







5

was suspected as a major cause of death in these patients, with mortality rates as high as

50% (Rose, 1990), it became critical to develop a rapid diagnostic test and treatment.

Although there are several therapeutic agents currently being investigated (Upton,

1997; Woods and Upton, 1998), none have yet been found to be effective. However, the

diagnosis of infection by oocyst detection in feces in lieu of intestinal biopsies was a

major advancement (Current et al., 1983).

Cryptosporidium undergoes a life cycle similar to that of most coccidia (Current and

Garcia, 1991; Fayer and Unger, 1986; Long, 1982). The infective stage of the disease is

the oocyst and the infective dose, generally reported as 10 to 100 oocysts, may be as low

as a single oocyst (Meinhardt et al., 1996). The oocyst, which is resistant to most

environmental conditions, is released by one host and ingested by its new host. As

shown in Figure 2-1, after ingestion occurs, the oocyst undergoes excystation, releasing

the sporozoites. The sporozoites initiate the infection within the epithelial cells of the

gastrointestinal tract. The sporozoite then differentiates into the trophozoite which

undergoes asexual multiplication to form Type I meronts and then merozoites which may

infect new host cells. Merozoites from Type II meronts produce microgamocytes and

macrogamocytes that undergo sexual reproduction to form the oocyst. The infected

individual excretes oocysts in the feces, and the oocysts are infective immediately.

The oocyst is a spherical particle with a diameter of 4 to 8 microns (Levine, 1984).

The DNA sequencing of Cryptosporidium oocysts' surface proteins suggests a surface-

adherent molecule, rich in cysteine, proline and histidine and capable of forming disulfide

bonds (Tilley and Upton, 1997). However, other studies have shown that the outer wall

of Cryptosporidium parvum oocysts contains galactase/galactosamine and glucose/gluco-




































Microment

Figure 2-1. Life Cycle of Cryptosporidium (Fayer and Unger, 1986)







7

samine residues, possibly with both N- and O-linked glycosylation (Considine et al.,

2000).

2-1-2 Public Health Perspective

Cryptosporidium is one of the primary causes of waterborne illness (Rose, 1990). It

is the most predominant intestinal protozoan, ranking sixth behind the enteric bacteria

(Marshall et al., 1987), and the most predominant parasite infection (Holly and Dover,

1987). In one incident alone, over 403,000 individuals became ill in an outbreak of

cryptosporidiosis in Milwaukee, Wisconsin. This incident was judged to cost the

community some $53 million in lost wages, lost productivity, medical bills and

emergency room visits, as well as $100 million in claims of loss of life (Smith and Rose,

1998). Between 250 and 500 million infections ofC. parvum are reported to occur

annually in Asia, Africa, and Latin America (Current and Garcia, 1991).

2-1-3 Filtration of Crvptosporidium

Several researchers investigated the effectiveness of sand filtration for the removing

Cryptosporidium oocysts from water. These studies are shown in Table 2-1. Ongerth

and Pecoraro (1995) investigated the effect of optimization of chemical conditioning on

the removal of Cryptosporidium oocysts. Near-complete removal was achieved under

optimum conditions. However, when conditions were allowed to drop below optimum

levels, removals decreased. Whitmore et al. (1993) studied the effect of hydraulic

loading on the removal of Cryptosporidum. As hydraulic loading increased, removals

decreased. At a hydraulic loading of 5 m/h, a conservative loading for a rapid sand filter,

the removal rate of Cryptosporidium is quite low.








Table 2-1. Sand Filtration ofCryptosporidium oocysts.


Sand Size, Bed Depth, Hydraulic Loading, Chemical Conditioning Removal Study
mm m m/h
(in.) (ft) (gal/(ft2 min))
0.27 0.90 0.15-0.4 None >99.99% Pilot Scale
(0.01) (2.95) (0.0001-0.0003) (Schuler, 1991)
0.2-0.3 1.05 0.19-0.4 None 48% Full Scale
(0.008-0.012) (3.44) (0.0001-0.0003) (Fogel, 1993)
0.45-0.55 1.2 12.2 Alum Coagulation Pilot Scale
(0.018-0.022) (3.94) (0.009) Optimum Conditions >99.8% (Ongerth and
Suboptimum Conditions 96% Pecoraro, 1995)
0.58-0.63 1.0 0.1-0.4 None 97.3-98.4% Lab Scale
(0.023-0.025) (3.28) (0.00007-0.0003) (Whitmore, 1993)
1.0-19.8 None 2.3-43.2%
(0.0007-0.0.01)











2-2 Microbial Attachment to Surfaces

Removal of Cryptosporidium oocysts and other microbes from water by filtration

depends on the ability of the microbes to attach to the surface of the filter media.

Adhesion of particles to surfaces involves several processes, as shown in Figure 2-2.

Transport of the particle to the surface results from body forces, such as gravitational

forces (sedimentation); hydrodynamic forces (fluid flow) causing movement of the

particle; or, in the case of microbes, the cell's own mobility. As the particle gets closer to

the surface and approaches the diffusive boundary layer, fluid convection in the direction

normal to the surface becomes negligible. At these shorter distances, the particle can be

transported by Brownian motion resulting from thermal fluctuations in the surrounding

fluid. The next step is the attachment of the particle to the surface. This is accomplished

by longer range forces (> 5 nm); namely van der Waals forces, electrostatic interactions,

and the influence of macromolecules on the cell surface. Adhesion of the particle to the

surface requires that the particle resist detachment. Detachment is governed by shorter

range forces (< 5 nm), specifically, shorter range van der Waals forces, electrostatic

interactions and hydrogen bonding.

2-2-1 Electrostatic Interactions

All solid surfaces in an aqueous medium carry a surface charge. The surface

charge is a result of either the ionization of surface groups or the adsorption of ions from

the solution. A charged surface then attracts ions of the opposite charge. The counter

ions can be closely associated with the surface or distributed exponentially into the

solution. When two solid surfaces approach each other, the surfaces with like charges












Convection (Mobility)


Diffusion Boudary Layer


Diffusion


Detachment


Longer Range Forces


Attachment


Figure 2-2. Microbial adhesion as a sequence of dynamic processes. (Dickinson et al., 2000)







11

experience repulsion when their electrical double layers overlap. Solid surfaces with

opposite charges attract each other.

2-2-2 Van der Waals Forces

A molecule, even a nonpolar one, can have a momentary unequal distribution of

electron density. There can be an excess of electron density in one region of the

molecule and a corresponding deficiency in another area. An uneven distribution of

electron density can create an instantaneous dipole, and the instantaneous dipole can

create a dipole in a second non-polar molecule, creating an attraction. These weak

interactions are collectively called van der Waals forces.

2-2-3 DLVO theory

Derjaguin and Landau (1941) and Verwey and Overbeek (1948) proposed that the

total interaction force between surfaces was the additive combination of van der Waals

attraction and electrostatic repulsion. Figure 2-3 shows a plot of the energy of interaction

of two surfaces as a function of their separation distance. The bottom line on this plot

represents the van der Waals attraction. This is a negative term which is an inverse

power law function of separation. The top line on this plot represents the double layer

interactions. This is a positive term which decreases exponentially as separation distance

increases. The net energy of interaction (represented by the middle line) is the sum of

the van der Waals and electrical double layer interactions.

Two minimums are apparent on the line representing the sum of the two

interaction forces. The minimum occurring at the shorter separation distance, the primary

minimum, corresponds to a balance between the short-range repulsive forces and

attractive forces. This minimum is present at low or high electrolyte concentration.

When a particle comes into this region of the graph, it will be pulled to the surface and









30


-20




Borier

0
3 ~ -- 2 4 A 6

-10 Primary Minimum
6 dW

-20
Separation Distance, h (1/K)

Figure 2-3. Hypothetical DLVO energy potential curve.







13

permanently attached to it. A secondary minimum at a greater separation distance

corresponds to a balance between long-range attractive and repulsive forces. The

minimum is present only at high electrolyte concentration, because the double layer term

decays more sharply, allowing the van der Waals term to remain significant at a

separation distance beyond the range of repulsion. The particles can remain in this

secondary minimum, and it results in a much weaker adhesion. The attachment, in this

case, could be reversed by shear or reducing the salt concentration. Because interaction

potentials are directly proportional to particle size, secondary minimum effect is more

significant with larger particles (greater than 1 micron in diameter, Elimelech et al.,

1995).

As previously mentioned, the interaction energy between a charged particle and a

charged surface predicted by DLVO theory is the sum of the electrical double layer

contribution and the van der Waals contribution. Quantitative prediction of these

contributions involves measurements such as zeta potential, dielectric constant and

Hamaker constant, which are briefly explained below.

2-2-3-1 Zeta potential

Double layer interactions are generally quantified by the zeta potential. Surface

potential is an important factor in microbial adhesion. Surface potential can be

approximated by zeta potential or surface charge. In the case of microbial adhesion, it is

often assumed that the surface potential is constant over the entire surface of the microbe

even though the surface of cells is not homogeneous. Surface charge refers to the charge

character of the surface due to its actual molecular composition. Zeta potential is the

potential resulting from that charge at a particular distance from the surface. The

distance is the plane of shear. At this plane, the ion association of the surrounding







14

solution (the electrical double layer) can reduce the magnitude of the charge slightly.

However, when the electrolyte solution is dilute, the number of these associating ions

which shield the surface charge is minimized. In this case, the zeta potential can be

assumed to equal the surface potential.

2-2-3-2 Hamaker constant

Van der Waals attraction is evaluated by two methods. One approach is

microscopic (Hamaker, 1937): the total interaction between two particles is found by a

pairwise summation of all of the relevant intermolecular interactions. The resulting

expressions can be split into a purely geometric element, and the Hamaker constant. The

Hamaker constant is related only to the properties of the particles and the medium. It is a

function of the number of atoms per unit volume of particles and the London dispersion

force constant that is related to the polarizability of the molecule.

2-2-3-3 Dielectric constant

The second method, suggested by Liftshitz (1956), is a macroscopic approach for

evaluating the van der Waals force. If a polar substance is placed in an electric field, its

molecules become oriented so that their positive ends face the negative pole and the

negative ends face the positive pole. This alignment partially neutralizes the field. The

dielectric constant of a substance is a measure of the extent to which the field is

neutralized. Dielectric measurements are taken as a function of frequency. The attraction

between two materials in a particular medium is related to the summation of the dielectric

spectra between the materials and the media.

2-2-4 Non-DLVO Interactions

Since the development of DLVO theory, many researchers have obtained

experimental results that could not be explained by DLVO theory. As a result, several







15

other interactions have been proposed to explain these results. The additional

interactions include hydration forces, hydrophobic forces and macromolecule-induced

forces.

2-2-4-1 Hydration force

One of the non-DLVO interactions that was proposed is the hydration force

(Israelachvili, 1992). Ionic groups or hydrophilic sites on a surface can cause water to be

bound tightly to that surface. If the particle is hydrated in this manner or the collector

surface is hydrated, an extra repulsion to attachment can occur. The extra repulsion

occurs because the surfaces must become dehydrated to allow the particle and collector

surfaces to make contact with each other.

2-2-4-2 Hydrophobic forces

An additional non-DLVO interaction is hydrophobic interactions (Stenstrom,

1989; van Loosdrecht et al., 1987a; van Loosdrecht et al., 1987b; Gerba et al., 1988).

When a surface has no polar or ionic groups or hydrogen-bonding sites, it is hydrophobic.

When two hydrophobic surfaces are brought together, there is an attraction. Hydrogen

bonding between the molecules in bulk water cause it to be more structured than water

that is in contact with a hydrophobic surface. As a result, the free energy of water in the

gap between hydrophobic surfaces increases as the distance between the surfaces

decreases, and the water tends to leave the gap.

2-2-4-3 Macromolecule-induced forces

Macromolecules present in the gap between a particle and a collector surface can

affect the attachment of the particle to the surface (Dickinson, 1997). Polymeric

molecules can adsorb on the surface of a particle or collector. When only a small amount

of polymer is present, it can adsorb on more than one surface, causing a bridging effect.







16

When a larger amount of polymer is present and the surfaces of the particle and collector

are coated, the polymer layers can overlap and cause steric repulsion. The repulsion is a

result of the need for the hydrophilic chains to become dehydrated to allow the polymer

layers to overlap.














CHAPTER 3
EFFECT OF HYDROUS IRON ALUMINUM OXIDE COATING ON SAND IN
THE FILTRATION OF Cryptosporidium OOCYSTS

3-1 Introduction

The presence of Cryptosporidium in water supplies is a widespread problem in the

U.S. and other nations. This protozoan causes severe gastroenteritis which is potentially

fatal to infants and immunocompromised individuals and which does not respond well to

available therapeutics. Rose (1998) found that 77% of rivers and 75% of the lakes

sampled in a U.S. survey had detectable levels of Cryptosporidium. Also, 83% of

pristine waters (with no human activity in the watershed) and 28% of the treated drinking

water samples had detectable levels of Cryptosporidium oocysts. In the U.S., over

431,000 confirmed cases of cryptosporidiosis have been reported since the disease was

identified in 1976 (Rose, 1998). Most (403,000) of these cases are from the 1993

outbreak in Milwaukee, Wisconsin. Between 250 and 500 million infections occur

annually in Asia, Africa, and Latin America (Current and Garcia, 1991). Recently, 1100

cases of cryptosporidiosis were reported in Sydney, Australia.

The infective stage of the disease is the oocyst and the infective dose, generally

reported as 10 to 100 oocysts, may be as low as a single oocyst (Meinhardt et al., 1996).

The oocysts are resistant to disinfectants and most environmental conditions, thus making

physical removal by treatment processes necessary. Chemical conditioning and optimum

operation of filtration systems can achieve greater than 99% removal of oocysts from

water (Ongerth and Pecoraro, 1995; Schuler et al, 1991). Suboptimal flocculation can







18

lead to substantially decreased performance, however, dropping the removal to 96% at a

superficial velocity of 293 m/d (Ongerth and Pecoraro, 1995). Fogel et al.(1993)

obtained removals of only 2 to 43% at hydraulic loads of 24 to 475 m/d when chemical

conditioning was not practiced.

Cryptosporidium oocysts carry a negative surface electrical charge in the pH range

of natural waters (Drozd and Schwartzbrod, 1996; Karaman et al., 1999; Ongerth and

Pecoraro, 1996), whereas natural filter media such as sand and diatomaceous earth also

carry a negative surface electrical charge in this pH range (Chen et al., 1998; Farrah and

Preston, 1991). Thus, removal of the oocysts particles by filtration (without chemical

conditioning) will be difficult due to electrostatic repulsion between the oocysts and the

filter media surface. Coatings of metallic hydroxides, oxides, and peroxides on filter

media have been found to enhance the removal of bacteria, viruses, and turbidity from

water and wastewater (Ahammed and Chaudhuri, 1996; Farrah et al., 1991, Gerba et al.,

1988; Mills et al., 1994, Lukasik et al., 1996). However, no studies have been carried out

to determine if the application of electropositive coatings to granular filtration media can

improve removal of Cryptosporidium. Extrapolation of results from previous studies

with bacteria and viruses is problematical, since the oocysts, which are 5 to 7 pm in

diameter, are significantly larger than the microbes tested previously. The purpose of the

present study was therefore to determine the effect of coating filter media (Ottawa sand)

with hydrous aluminum iron oxide on removal of Cryptosporidium. Performance of

coated and uncoated sand was investigated over ranges of superficial velocity and filter

column length.







19

3-2 Materials and Methods

3-2-1 Coating Method

The fraction of 20 x 30 mesh Ottawa sand (Fisher Scientific) passing a U.S. Standard

#25 sieve was collected to obtain particles between 0.6 and 0.7 mm in diameter. The

sieved sand was filled to a depth of 2.5 cm in flat plastic pans, then covered with a

solution 0.4 M in AICI3-6H20 and 0.2 M in FeC13.6H20 (Fisher Scientific). The solution

was immediately drained from the sand, and the sand was dried at 700 C for 24 h, then

cooled to room temperature. Clumps in the dried sand were broken up, then the sand was

poured into a beaker containing 4 M ammonium hydroxide (Fisher Scientific). The

solution was immediately drained from the sand, then the sand was spread over flat pans,

dried at 700 C for 24 h, and cooled. The dried sand was re-sieved using a #25 screen,

rinsed with deionized water until the water was clear, air-dried, and stored in sealed

plastic containers until use.

To measure iron and aluminum contents, 10.0 g of oven-dried (1050 C) sand was

digested in 25.0 mL of gently boiling aqua regia (1:2:2 HCI:HNO3:H20) for 20 min.

After cooling, the digestate and water from rinsings of the digestion vial and sand were

passed through Whatman GF/C filters, and the final volume was made up to 100 mL.

This solution was analyzed for aluminum and iron by ICP.

3-2-2 Sand Columns

Dry sand was packed into 1.5 cm I.D. acrylic columns of varying lengths, depending

on the experiment. Sand quantities were 39 g in the 10 cm columns, 78 g in the 20 cm

columns, 114 g in the 30 cm columns, and 154 g in the 40 cm columns. Glass wool was







20

used at both ends of the columns to prevent loss of filter media. Columns were used in

the upflow mode.

Columns containing uncoated sand and coated sand were run in parallel in each trial

and were fed artificial groundwater (AGW) containing 35 mg MgSO4-7H20, 12 mg

CaSO4-2H20, 12 mg NaHCO3, 6 mg NaCI, and 6 mg KNO3 per liter deionized water

(McCaulou et al., 1994) with a pH of 7.0. Immediately prior to each experiment, packed

columns were rinsed with 70 pore volumes of AGW at the same flow rate used in the

experiment. Cryptosporidium oocysts (Pleasant Hill Farms, Troy, Idaho) were added to

20 to 70 L of AGW in a plastic container to give approximately 500 to 1000 oocysts/mL.

The AGW was mixed at 60 rev/min with a 5 cm diameter propeller-type impeller

throughout the experiments. Influent to the columns was sampled from the container

while pore volumes 71 to 75 were entering the columns. Composite effluent samples

representing steady state conditions were collected at the 71t through 75* pore volumes.

Samples were then refrigerated overnight prior to enumeration. Three trials were carried

out for each column length or superficial velocity tested.

3-2-3 Counting of Cryptosporidium Oocvsts

The fluorescent antibody method ofEPA (1995) was used in enumerating the

Cryptosporidium oocysts. The waters tested were reconstituted in the laboratory and the

oocysts were added from commercial sources, therefore the positive identification step

involving visualization of internal structures of the oocysts was omitted. For each

sample, one slide was prepared for the influent and one slide was prepared for the

effluent.

Reagents used in enumerating Cryptosporidium oocysts included phosphate buffered

saline (PBS), blocking solution, fluorescent reagent, and mounting medium. The PBS







21

(Sigma) had a pH of 7.6. Blocking solution was composed of 1% bovine serum albumin

(BSA), 10% normal goat serum (NGS), and 0.02% thimerosol in PBS (reagents from

Sigma.) The blocking solution was prepared by combining 10 mL NGS and 90 mL PBS,

then placing 1 g BSA and 0.02 g thimersol (both in powder form) on the liquid surface.

The powders were allowed to dissolve before mixing the solution. Blocking solution was

made up weekly. Mounting medium was prepared by mixing together 90 mL glycerol,

10 mL PBS, and 2 g of 1,4-diazabicyclo(2,2,2)octane (reagents from Sigma). The

fluorescent reagent was prepared with a Crypt-A-Glo kit (Waterborne, New Orleans,

LA). The kit contains 5-carboxy-fluorescein-labeled IgM monoclonal antibody made

against oocysts of Cryptosporidium parvum. The working dilution was obtained by

adding 0.25 mL of the antibody reagent to 5 mL of blocking solution, followed by gentle

stirring. The working dilution was prepared fresh for each assay.

Filters were soaked in PBS for 2 min. A wet support filter (0.45 lpm effective pore

size; Gelman GN6) was placed on each support screen of the vacuum manifold (Hoefer

model FH 225V) and a cellulose acetate filter (0.2 tpm effective pore size cellulose

acetate; Sartorius, Edgewood, N.Y.) was placed atop each support filter. The vacuum

was adjusted to 5 to 10 cm Hg, then a manifold valve was opened briefly to flatten the

filter. (Valves were opened only long enough to drain liquids, in order to avoid drying

the filters.) The filter was rinsed with 2 mL of blocking solution to limit non-specific

background fluorescence. A 5 mL volume of sample was then passed through the filter,

followed by an additional rinse of 2 mL of blocking solution. A volume of 0.5 mL

fluorescent reagent was placed on the filter and left for 45 minutes. Light was excluded

during the contact period by covering the filter funnel with aluminum foil. After draining

the fluorescent reagent, the filter was rinsed 5 times with 2 mL PBS per rinse.







22

The slide was prepared by placing a 75 p.L drop of mounting medium on a slide and

warming the slide to 37 o C. A single filter was placed atop the mounting medium on a

slide, completely wetting the bottom of the filter. A 25 piL drop of mounting medium

was added to the top of the filter. Finally, a glass cover slip was placed over the filter and

weighted with 5g of coins to flatten the filter. The edges of the cover slip were sealed

with quick drying nail enamel.

The 5 to 7 micron diameter Cryptosporidium oocysts appeared apple-green and

spherical with a darker green outline under epifluoresence illumination (Leitz Microlab

with 25x bright field objective and 12x ocular). The entire slide was counted. Since the

filtered suspension was in reconstituted water, very little debris was visible on the

prepared slides. Typical influent counts per slide were approximately 2000 oocysts,

compared to effluent counts of 5 to 800 oocysts.

3-2-4 Zeta Potential Measurements

The zeta potential of the Cryptosporidium oocysts was measured over a pH range

of 5 through 8 using a Brookhaven Zeta Plus. The oocysts were purchased without a

preservative (i.e. in deionized water only). The pH was adjusted using NaOH and HCI

(Fisher Scientific). Each sample was adjusted to 0.001 M NaCl. The Cryptosporidium

oocyst mass concentration used was approximately 0.05 mg/mL The electrode was

cleaned by sonicating for 5 to 10 minutes every 3 to 4 runs.

The zeta potential of the uncoated and coated sand was determined using a

streaming potential apparatus consisting of a flow cell packed with the granular filter

media, a reservoir of electrolyte (1.0 x 10-3 M KC1) for the flow cell, electrodes and a

voltmeter (ExTech 380282) to measure the potential difference across the flow cell, and

manometer to measure the pressure drop across the flow cell. The flow cell was made







23

from clear acrylic tubing with a length of 50 cm and an inner diameter of 3.8 cm. The

electrodes were located at the two ends of the flow cell and were formed from fine silver

mesh was cut to match the cross section of the flow cell. Silver wire (18 gauge) was spot

welded to the center of each electrode. The electrodes were anodized in HCI for 30 to 60

minutes using a copper or platinum cathode and a 10 mA current. The electrodes were

re-plated periodically to ensure proper performance.

Sand samples were uniformly packed into the sample cell via the tap and fill

methodology. The sample cell was flushed with the KCI electrolyte solution to remove

air pockets. Single trials were carried out over a range of pressure drops and

corresponding potential differences in order to find the ratio of streaming potential (Ew)

to pressure drop through the media (p) for a given pH and sand type. This ratio was used

to find zeta potential ( ) according to the equation (Hiemenz and Rajagopalan, 1997):


Estr k(3.1)
I P JE

where Tr is the viscosity of water and k is conductivity. The parameter E in SI units is

found from:

E = Er 0 (3.2)

where Er = 78.54 at 250C and E0 = 8.85 x 1012 C2/(J m).

3-3 Results and Discussion

3-3-1 Physical Characterization of Crptosporidium and Sand

The SEM image in Figure 3-1 shows the spherical shape of a typical

Cryptosporidium parvum oocyst. Due to shrinkage in preparation, the oocyst imaged is






































Figure 3-1. Scanning electron micrograph of Cryptosporidiumparvum oocyst (30,000x)







25

smaller than the typical 5 to 7 um diameter as determined by epifluorescence microscopy

(Rose, 1998). Zeta potential of Cryptosporidium oocysts in 10' KCI was negative over

the pH range 5 to 8, in agreement with other investigators (Drozd and Schwartzbrod,

1996; Karaman et al, 1999; Ongerth and Pecoraro, 1996). Our measured values ranged

from -17 mV at pH 5 to -29 mV at pH 8 (Fig. 3-2). These are in excellent agreement

with the relationship given by Drozd and Schwartzbrod (1996), but are somewhat more

electropositive than values given by Ongerth and Pecoraro (1996) and Karaman et al.

(1999).

At 30x, uncoated and coated grains of Ottawa sand were indistinguishable (Fig.

3-3a). At 2000x, the morphology of coated sand was still essentially the same as the

uncoated sand (Fig. 3-3b). Zeta potential of uncoated sand ranged from 0 mV at pH 2.6

to -55 mV at pH 10.7 (Fig. 3-4). The isoelectric point of uncoated sand (pH 2.5) is

consistent with the value of pH 2 to 3 given by Parks (1965) for the a-quartz form of

SiO2. Coating the sand with aluminum and iron hydoxy(oxides) made the zeta potential

more electropositive (Fig. 3-4). The isoelectric point of the coated sand was 8.0, which is

between the range of 4.8 to 6.8 for iron oxides (Fe203 haematite; FeOOH goethite) and

7.8 to 9.1 for aluminum oxides (A1203 corundum; A1OOH boehmite) as given by

Parks (1965).

The uncoated Ottawa sand had an iron content of 0.11 0.026 mg/g sand and an

aluminum content of 0.014 0.003 mg/g sand (Table 3-1). Coating the sand increased

the iron content to 1.36 0.11 mg/g sand and the aluminum content to 1.22 0.081 mg/g

sand. These iron contents fall within the range of values reported by others (Edwards and











4 5 6 7 8 9
0 1 I
Relationship from Ongerth and Pecoraro
-5 (1996)

-10 Relationship from Drozd and
> \Schwartzbrod (1996)
-15

-20

-25

-30
Relationship from Karaman et al, 1999
-35

-40

Figure 3- 2. Zeta potential of Cryptosporidium oocysts in relation to pH. (Data points from present study. Error bars
represent 1.0 standard deviations. The line for Drozd and Schwartzbrod (1996) was plotted from the
equation in Figure 5 of their paper; the lines for Ongerth and Pecoraro (1996) and Karaman et al. (1999)
are visual fits to their data.











Uncoated















Coated


30x 2000x


30x 2000x


Figure 3-3. Scanning electron micrographs of Ottawa sand before and after coating (a) magnified 30x and (b)
magnified 2000x.









50


25


0

-25


-50


-75


6 8


Figure 3-4. Comparison of zeta potential of uncoated sand to the zeta potential of sand coated with iron /
aluminum (hydroxy)oxide










Table 3-1. Iron and aluminum contents of uncoated and coated sand.


Sand size Fe, Al, Reference
mm (in.) mg/g sand mg/g sand
0.6-0.7 (0.02-0.03) (Uncoated) 0.11 0.01 Present study
0.6-0.7 (0.02-0.03) (Coated) 1.7 1.2
0.6-0.8 (0.02-0.04) 0.7 Edwards and Benjamin (1989)
0.6-0.7 (0.02-0.03) 0.4 Chen et al. (1998)
1.1 (0.04) (effective size) 1.6 0.8 Kang (1998)
0.7-1.2 (0.03-0.05) 1-2 Lo et al. (1997)
"Fines" 0.6-0.7 Stahl and James (1991)







30

Benjamin, 1989; Kang, 1998; Lo et al., 1997; Stahl and James, 1991), whereas the

aluminum contents are somewhat higher than values previously reported (Chen et al.,

1998; Kang, 1998). The Al/Fe molar ratio in the coating was 2:1 in the present study,

compared to 1:1 in the coating solution. The enrichment for aluminum may be because

the aluminum coating binds better to the sand. Truesdail (1999) found that attrition rates

for iron hydroxide coatings were significantly higher than for aluminum hydroxide

coatings. In our work, the rinsing and handling as part of the preparation procedure

could cause loss of iron.

3-3-2 Attainment of Pseudo-Steady-State in Filtration Experiments

Pseudo-steady-state conditions were obtained shortly after starting the filter runs.

This is demonstrated in Figure 3-5, which shows that the performance of filters using

either coated sand or uncoated sand did not vary in relation to the length of the filtration

period up to a filter run length of 420 minutes [782 pore volumes at a superficial velocity

(U) of 407 m/d]. Lines representing least squares linear fits to the data for both coated

and uncoated sand had slopes that were not statistically different from zero (P < 0.05). A

long-term experiment with uncoated sand (Fig. 3-6) showed that pseudo-steady-state

conditions were maintained for at least 30 hours (5000 pore volumes at U = 81 m/d).

Based on an influent concentration of 272 oocysts/mL, an average removal of 97%, 6-

pm diameter oocysts, and 0.65 mm spherical sand particles, the surface coverage of sand

by the Cryptosporidium was estimated to be 3% at the end of the 30-h period. This

degree of surface coverage is evidently too small to affect the rate of attachment of the

oocysts to the sand.

Because the superficial velocities varied depending on the experiment, the

appropriate sampling time was defined in terms of the number of pore volumes







_ I T --I4


SCoated Sand


Uncoated Si


S80-



O 60-


0
a


E 40-
V



20 -



0


100


Time (minutes)

Figure 3-5. Effect of column run time on Cryptosporidium removal efficiency in sand columns (U = 407
m/d, L = 40 cm ; Error bars represent 1.0 S.D.)


1000


I-^


and





420 minutes

782 pore volumes


I


I I














100

... 30 hours -
80-
7 0 5000 pore volumes
o
E 60--
S 0 5 10 15 20 25 30 35

Time, hours






Figure 3-6. Effect of extended column run-time on Cryptosporidium removal efficiency (U= 82 m/d,
L = 5 cm ; Error bars represent 1.0 S.D.)







33

processed. Comparisons of pseudo-steady-state filter performance were made using

samples of the 70th through 75th pore volumes leaving the filter. These were well within

the pseudo-steady-state region of filter performance.

3-3-3 Effect of Coating on Filtration Performance

The change in the concentration of particles with respect to column length in a

granular media filter can be modeled by the first-order relation

ac
= -Ac (3.3)
az

where c is the particle concentration at length z and and X is a coefficient that

characterizes the filter media. Under conditions in which the coverage of filter media by

particles is low enough so that it does not affect the rate of particle attachment, the above

expression may be integrated over the length of the filter column, yielding


Cef = e-AL (3.4)
Cin

where ci = influent concentration of particles and c-m = effluent concentration of

particles. Equation (3-4) can be combined with the definition of fractional removal

efficiency (1 = 1-Ceff /Cin ) to obtain


S= 1-e--L (3.5)

Equation (3-5) indicates that, as the filter coefficient increases, the removal efficiency for

a given column length also increases. Figure 3-7 shows least squares nonlinear fits of

Equation (3-5) to removal fraction vs. column data for uncoated sand and coated sand

loaded at a superficial velocity of 407 m/d. The respective filter coefficients are shown

next to the fitted relationships. The filter coefficient for coated sand (8.8 mn) was









0.8 (2.7 ft-1)


S 0.6

0 Uncoated Sand, = 3.0 m-1
E 0.4 -
0 (0.91 ft-')


0.2 -


0 I I I
0 10 20 30 40 50
Column Length, cm (in.)
Figure 3-7. Effect of column length on removal of Cryptosporidium oocysts in sand filters (Error bars represent
1.0 standard deviations. Some error bars are too small to be seen.)







35
increased by a factor of 2.9 over that of uncoated sand (3.0 min). This difference was

significant at P < 0.05.

The effect of superficial velocity on filter coefficient is shown in Figure 3-8. The

range of superficial velocities investigated started at 204 mg/d (3.5 gal/(ft2 min)), which

is typical of rapid sand filters used in municipal water and wastewater treatment facilities.

The upper end of the range investigated was 814 m/d (14 gal/(ft min)), which is on the

order of loadings used in high-rate filter systems. The coated sand outperformed

uncoated sand over the entire range of superficial velocities tested.

A power law relationship of the form

x oc U-a (3.6)

has been suggested to give the effect of superficial velocity on filter coefficient

(Wennberg and Sharma, 1997), where U = superficial velocity. This coefficient is critical

in determining how sensitive the performance of granular media filters is to superficial

velocity. Previous research has indicated that the alpha coefficient is higher in attractive

systems than in repulsive systems (Ghosh et al., 1975), which would tend to negate the

benefit of coating filter media to achieve a positive zeta potential.

We fit relationship (3-6) to the data in Figure 3-8 by nonlinear least squares

regression and show the resulting alpha values on the figure. The alpha value of 0.66 for

uncoated sand is within the range of 0.43 to 1 found by Ghosh et al. (1975) for systems

with double layer repulsion and high superficial velocities (U > 86 m/d). The value of

0.59 for coated sand is well below the alpha of 2 determined by Ghosh et al. (1975) for

attractive systems under the same conditions, however. As a check on our results, we

computed filter coefficients from data on bacteria removal in columns of coated sand that








14

12

cs 10
S8 I Coated Sand
S(a = 0.59 0.054)
0
U 6 Uncoated Sand

S 4 = 0.66 0.028)

2-

0 I
100 300 500 700 900
Superficial Velocity, m/d

Figure 3-8. Effect of superficial velocity on filter coefficient in sand columns filtering Cryptosporidium (L = 40
cm ; Error bars represent 1.0 S.D.)







37

were reported by Ahammed and Chaudhuri (1996). Fits of relationship (4) to these data

are given in Figure 3-9. As shown, the values of alpha (0.50 and 0.48) are reasonably

close to our results.

This study's average 2.9-fold improvement in filter coefficient for removal ofC.

parvum by coating of sand media with iron/aluminum (hydroxy)oxide is generally better

than reported in previous studies on filtration of bacteria or bacteria-sized particles.

Ahammed and Chaudhuri (1996) reported X/Xu of 1.5 to 1.8 for the filtration of

heterotrophic bacteria and 1.6 to 2.1 when filtering Escherichia coli. At the pH of their

experiments (7.8 to 8.1), the bacteria should have carried a negative charge. Stenkamp

and Benjamin (1994) found XJX, of 1.2 to 2.5 and 1.1 to 1.5, respectively, when filtering

negatively charged (at pH 7.0) latexes and ferrihydrite particles. Kang (1998) studied the

effect of aluminum and iron hydroxide coatings on fabric filters. He found that the filter

coefficients were 2.1-fold greater for coated fabric than uncoated fabric for

Staphylococcus aureus and 3-fold greater for coated fabric for Escherichia coli. No

ratios for filtration of protozoa such as Cryptosporidium or similarly sized particles is

available in the literature. This improvement can be most likely be attributed to the

change in zeta potential (from electronegative to electropositive) resulting from the

coating process since this would decrease the electrostatic repulsion between the sand and

the electronegative Cryptosporidium oocysts. Other factors, such as increased surface

roughness, may also play a role.










100
Heterotrophic
0* Bacteria
h 60

40

-3 20
> 0 10 20 30 40 50
S100 -
S 80Total Coliform
S80 Bacteria

60
0 20 40 60
Superficial Velocity, m/d
Figure 3-9. Effect of superficial velocity on bacteria removal in sand columns (Ahammed & Chaudhuri, 1996)
(L = 15 cm, Sand size = 0.3-0.8 mm)







39

3-4 Conclusions

Based on the present study, the following conclusions can be drawn:

A surface coating of hydrous aluminum and iron oxide on Ottawa sand is

effective in increasing the zeta potential of the sand from negative values to

positive values

Coated electropositivee) sand significantly improves removals of

Cryptosporidium oocysts from water at superficial velocities representative of

rapid sand filters operated at low to high superficial velocities

Based on the almost three-fold improvement in filter coefficient, coated sand

can significantly increases the reliability of rapid sand filtration systems and

prevent breakthrough of Cryptosporidium oocysts during periods of

suboptimal chemical conditioning














CHAPTER 4
EFFECT OF IRON ALUMINUM (HYDR)OXIDE COATING ON REMOVAL OF
Cryptosporidium OOCYSTS BY GLASS AND CERAMIC BEADS

4-1 Introduction

Coatings of the oxides and hydroxides (i.e., (hydr)oxides) of iron and aluminum on

granular filter media such as sand and diatomaceous earth enhance the filtration of

bacteria, viruses, protozoa, and turbidity from water (Ahammed and Chaudhuri, 1996;

Chang et al., 1997; Chen et al., 1998; Farrah and Preston, 1991; Gerba et al., 1988;

Lukasik et al., 1996; Lukasik et al.,1999; Mills et al., 1994; Shaw et al., 2000; Stenkamp

and Benjamin, 1994; Truesdail, 1999). The coatings, which are electropositive, change

the zeta potential of the filter media from highly electronegative values to near-zero or

positive values. In comparison, microbes and other colloidal particles in water normally

carry a negative zeta potential in the natural water pH range of 6 to 8. Improvements in

particle removal by coated filter media have thus been attributed to reduction or

elimination of electrostatic repulsion between the filter media and colloids (Ahammed

and Chaudhuri, 1996; Chang et al., 1997; Chen et al., 1998; Farrah and Preston, 1991;

Lukasik et al., 1996; Lukasik et al.,1999; Mills et al., 1994; Shaw et al., 2000; Stenkamp

and Benjamin, 1994).

A few investigators have cited other changes to filter media surfaces due to coatings

as playing a role in enhanced particle removal. These include surface roughness, surface

chemical heterogeneity, surface area, and amphoteric surface charge (Stenkamp and

Benjamin, 1994; Sansalone, 1999). Chen et al. (1998) observed, in a long-term







41

wastewater filtration study with aluminum (hydr)oxide coated Ottawa sand, that the zeta

potential of the coated sand dropped back to negative values soon after the filters were

put into service. However, the coated sand continued to remove significantly more

Escherichia coli from wastewater than uncoated sand for over 3 months of continuous

service.

The present study was carried out to quantify the non-electrostatic contribution of

iron aluminum (hydr)oxide coating on granular filter media to removal of

Cryplosporidium oocysts. Characteristics and performance of glass beads, which have

negative zeta potential in the natural water pH range of 6 to 8, were compared to those of

coated glass beads, which have positive zeta potential in this pH range. Similarly,

characteristics and performance of ceramic beads, which have positive zeta potential in

the natural water pH range, were compared to those of coated ceramic beads which also

have positive zeta potential in the pH range. The results show that the coating improves

filter media performance, even when it has negligible impact on the zeta potential of the

filter media.

4-2 Materials and Methods

Chemicals used in this study were purchased from Fisher Scientific unless

otherwise indicated.

4-2-1 Filter Media Preparation

Approximately spherical high-silica glass beads (Jaygo Inc., Union, N.J.) and

high-alumina ceramic beads (Ferro Corp, Shreeve, Ohio) were used in the study. The

glass beads were 72% Si02, 9% CaO, 4% MgO, and 1% A1203, whereas the ceramic

beads were 87% A1203, 1.4% MgO, 1.4% CaO, 8.7% SiOz, 0.3% Fe203, 0.13% TiO2,

0.4% Na2O, and 0.4% K20, according to the respective manufacturers.







42

The glass beads had a narrow size distribution, with diameters averaging 0.73 mm.

The ceramic beads had a wider size distribution and were sieved to collect the fraction

passing a U.S. Standard #25 sieve and retained on a #30 sieve (0.6 to 0.7 mm). The

beads were coated with iron aluminum (hydr)oxide according to Shaw et al. (2000).

Beads were filled to a depth of 2.5 cm in glass pans and covered with a solution 0.2 M in

FeCl3.6H20 and 0.4 M in AICI3-6H20. The solution was immediately drained from the

beads and the media were dried at 700C for 24 h, then cooled to room temperature.

Clumps were broken up, beads were added to a beaker containing 4M NH4OH, and the

solution was immediately drained. The beads were spread over flat pans and dried at 700

C for 24 h. After cooling, beads were rinsed with deionized water until the water was

clear, air-dried, and finally passed through a #25 sieve. Coated beads were stored in

sealed plastic containers until use.

4-2-2 Filter Media Characterization

4-2-2-1 SEM

Beads were carbon-coated (Ion Equipment, Santa Clara, CA) for 1 to 2 minutes to

obtain a 100 to 200 Angstrom thick coating and examined by SEM (JSM 6400,

thermoionic emission, Tungsten filament). X-ray mapping and quantitative analysis was

carried out with an Oxford Instruments Link ISIS EDS System (Oxfordshire, UK).

4-2-2-2 Surface area

Surface areas of beads were measured by five-point BET analysis using krypton

adsorbate on a Quantachrome Autosorb-1 (Boynton Beach, FL) physical adsorption

analyzer. Approximately 15 grams of beads were placed in a glass sample holder and

degassed overnight at 2000C. Three complete multi-point analyses were carried out on







43

each sample. The correlation coefficient of each BET analysis exceeded 0.99.

4-2-2-3 Zeta potential

Zeta potentials were measured on coating removed in the final rinse of the coating

procedure and on the beads before and after coating. Rinse water was passed through

Whatman GF/C filters and the filters were dried at room temperature. Dried residue was

scraped from the filters and added to aliquots of MilliQ water (Millipore, Bedford, MA)

that were previously adjusted to pH values in the range of 3 to 10 using NaOH and HCI.

The concentration of coating was approximately 0.03 mg/mL. Each sample was then

adjusted to 0.001 M NaC1. The zeta potential of the filter coating was measured on a

Brookhaven Zeta Plus (Brookhaven Instruments Corp., Holtsville, NY). The final pH of

each aliquot was measured immediately after the zeta potential measurement.

Zeta potential of the beads was determined using an Anton Paar Electro Kinetic

Analyzer (EKA, Anton Paar, Graz, Austria). The cylindrical flow cell (2.0 cm I.D.) was

packed to a depth of 4.0 cm with glass or ceramic beads. An electrolyte solution (1.0 x

10-3 M KCI) was pumped through the flow cell. This caused a charge transfer in the flow

direction, resulting in a potential difference (AU) that was detected by silver electrodes

coated with AgCI, placed at the ends of the cell. A differential pressure sensor was used

to measure pressure drop (AP) across the cell. The pressure was held at 300 mbar for 2

cycles of 2 minutes each and AP and AU were recorded. The ratio of these measurements

was used to find zeta potential (i) according to the Fairbrother and Mastin (1924)

equation:


S(AU ) k (4.1)







44

where 71 is the viscosity and k is the conductivity of water. The parameter E in SI units

is found from:

E = ErEO (4.2)

where 8r = 78.54 at 250C and E0 = 8.85 x 10-' C2 /(J m).

Isolectric points of the coating and beads were determined from quadratic fits of the zeta

potential-pH relationships within approximately 2.5 pH units, (corresponding to five or

more measurements) of the point of zero charge (PZC). The r2 of each fit was 0.99 or

higher.

4-2-2-4 Surface metal content

Acid digestion followed by ICP ofdigestates was applied to triplicate samples of

coated and uncoated beads. A volume of 25.0 mL aqua regia (1:2:2 HCI:HNO3:H20)

was added to 10.0 g of oven-dried (105C) beads and the mixture was boiled for 20

minutes. After cooling, the digestate and water from rinsings of the digestion vial and

beads were passed through Whatman GF/C filters and deionized water was added to

adjust the final volume to 100 mL This solution was analyzed for aluminum and iron by

ICP.

4-2-3 Performance Testing of Filter Media

4-2-3-1 Batch tests

Aliquots (5.0 g) of beads were placed in triplicate 125 mL Erlenmeyer flasks. A

suspension of Cryptosporidium parvum (Pleasant Hill Farms, Troy, ID), containing

approximately 1000 oocysts per mL, was prepared in artificial ground water (35 mg

MgSO4.7H20, 12 mg CaSO4-2H20, 12 mg NaHCO3, 6 mg NaCI, 6 mg KNO3 per litre

deionized water; pH 7.0) (McCaulou et al., 1994). A 20 mL volume of this suspension







45

was added to each flask containing beads, plus three flasks containing no beads (control).

The flasks were shaken at 60 rev/min on an orbital shaker table for 60 min. Supernatants

were sampled after 5 min. settling. The percent removals ofCryptosporidium oocysts

were computed based on the mean control concentration vs. the concentrations in the

respective flasks containing beads.

4-2-3-2 Column tests

Dry filter media was packed into 1.5 cm I.D. acrylic columns of varying lengths.

Glass wool was used at both ends of the columns to prevent loss of filter media.

Columns were used in the upflow mode.

Columns containing uncoated beads and coated beads were run in parallel in each

trial and were fed artificial groundwater at 200C. Immediately prior to each experiment,

packed columns were rinsed with 70 pore volumes of artificial groundwater at the same

flow rate (50 mL/min) used in the experiment. Cryptosporidium oocysts were added to 20

to 70 L of artificial groundwater in a plastic container to give approximately 500 to 1000

oocysts/mL. The water was mixed at 60 rev/min with a 5 cm diameter propeller-type

impeller throughout the experiments. Influent to the columns was sampled from the

container while pore volumes 71 to 75 of the Cryptosporidium suspension were entering

the columns. Composite effluent samples representing steady-state conditions were

collected at the 71" through 75th pore volumes. Shaw et al. (2000) previously determined

that pseudo-steady-state conditions were achieved after 10 pore volumes had passed

through the column and were maintained for at least 780 pore volumes. Samples were

stored at 40C overnight prior to enumeration. Three trials were carried out for each set of

conditions.







46

The pressure drop in 100 cm long columns of filter media was measured at a flow

rate of 50 mL/min (superficial velocity = 407 m/d) using a mercury manometer. The

pressure drop due to the filter media was obtained by subtracting the pressure drop

through a column without beads from the pressure drop through a column containing

beads.

4-2-3-3 Counting of Cryptosporidium Oocvsts

The procedure for counting Cryptosporidium oocysts was based on the fluorescent

antibody method of EPA (1995). For each sample, one slide was prepared for the

influent and one slide was prepared for the effluent. Reagents used in enumerating

Cryptosporidium oocysts included phosphate buffered saline (PBS), blocking solution,

fluorescent reagent, and mounting medium (Shaw et al., 2000).

Filters were soaked in phosphate-buffered saline (PBS) for 2 min. A wet support

filter (0.45 gIm effective pore size; Gelman GN6) was placed on each support screen of

the vacuum manifold (Hoefer model FH 225V) and a cellulose acetate filter (0.2 gm

effective pore size cellulose acetate; Sartorius, Edgewood, N.Y.) was placed atop each

support filter. The vacuum was adjusted to 5 to 10 cm Hg, and a manifold valve was

opened briefly to flatten the filter. Valves were opened only long enough to drain liquids,

in order to avoid drying the filters. The filter was rinsed with 2 mL of blocking solution

to limit non-specific background fluorescence. A 5 mL volume of sample was then

passed through the filter, followed by an additional rinse of 2 mL of blocking solution. A

volume of 0.5 mL fluorescent reagent was placed on the filter and left for 45 minutes.

Light was excluded during the contact period by covering the filter funnel with aluminum







47

foil. After draining the fluorescent reagent, the filter was rinsed 5 times with 2 mL PBS

per rinse.

The slide was prepared by placing a 75 pL drop of mounting medium on a slide and

warming the slide to 37C. A single filter was placed atop the mounting medium on a

slide, completely wetting the bottom of the filter. A 25 p.L drop of mounting medium

was added to the top of the filter. Finally, a glass cover slip was placed over the filter and

weighted with 5g of coins to flatten the filter. The edges of the cover slip were sealed

with quick drying nail enamel.

The 5 to 7 micron diameter Cryptosporidium oocysts appeared apple-green and

spherical with a darker green outline under epifluoresence illumination (Leitz Microlab

with 25x bright field objective and 12x ocular). The entire slide was counted. Since the

filtered suspension was in reconstituted water, very little debris was visible on the

prepared slides. Typical influent counts per slide were approximately 2000 oocysts,

compared to effluent counts of 5 to 800 oocysts.

4-3 Results and Discussion

4-3-1 Characterization of Glass Bead Surfaces

Table 4-1 summarizes the iron and aluminum in coatings relative to surface area

of uncoated filter media. This table includes results from previous studies for which

measured surface areas of filter media are available. Values of surface metal content (mg

metal per g of filter media) and measured surface area (m2 per g of filter media) were

used to compute surface metal ratios (mg metal per m2 of filter media surface). Media

diameters in the table range from 0.6 to 0.85 mm. The surface Fe ratio for glass beads

(33.5 mg/m2) was somewhat higher than the ratio achieved with sand by investigators

using similar coating procedure (Lai et al., 2000; Shaw et al., 2000). Benjamin et al.






Table 4-1. Surface metal contents of filter media coated with iron or aluminum (hydr)oxides


Media Diameter Surface Area of Surface Fe Surface Surface Al Surface Al Reference
Uncoated Media Content Fe Ratio Content Ratio
(mm)
(m2/g media) (mg/g media) (mg/m2) (mg/g media) (mg/m2)

Ottawa sand 0.60-0.85 0.04 28 700 Benjamin et al., 1996

Ottawa sand 0.60-0.85 0.04 25 625 -- Chang et al., 1997

Ottawa sand 0.6-0.7 0.10648 -- -- 0.35 3.3 Chen et al., 1998

Quartz sand 0.85 0.85 5.7 6.7 -- -- Lai et al., 2000

Ottawa sand 0.6-0.7 0.1064 1.25 11.7 1.21C 11.4 Shaw et al., 2000

Glass Beads 0.73 0.009966 0.334' 33.5 0.178C 17.9 Present Study

Ceramic Beads 0.6-0.7 0.01413 1.190 84.2 (b) (b) Present Study

'The surface area of 0.6-0.7 mm Ottawa sand was measured in the present study

lNot able to measure

CS.D. < 11% of the mean (N=3)







49

(1996) achieved a considerably higher surface Fe ratio (700 mg/m2) using a two-step

coating procedure in which iron solution [Fe(NO3)3 or FeCl3] was not drained from the

sand media prior to the drying step. The surface Al ratio for glass beads (17.9 mg/m2)

was somewhat higher than the ratios achieved by Shaw et al. (2000) and Chen et al.

(1998).

Viewed by scanning electron microscopy at 250x, the surfaces of both the coated

(Fig. 4-1c) and uncoated (Fig. 4-1a) glass beads appeared relatively smooth, with patches

of roughness. Patches on the coated beads were scale-like in appearance (see area in and

around the box in the top portion of Figure 4-1c), whereas rough areas on the uncoated

bead (e.g., in and around the box in the top portion of Figure 4-la) had a finer texture. X-

ray maps of the uncoated bead (Fig. 4-1b) showed a homogenous distribution of

aluminum, iron, and silicon on the surface, including the rough patches, indicating that

these patches did not differ from the rest of the surface in their composition. In contrast,

the metals were unevenly distributed on the coated bead (Fig. 4-id). For example, the

relative concentration of aluminum and iron was high (lighter colored in the respective Al

and Fe maps) in and around the boxed region, whereas in the same area, the relative

concentration of silicon was low (darker colored in the Si map). This indicates that the

rough, scaly patches on the coated beads were iron aluminum (hydr)oxide coating, which

masked the silicon signal from the underlying glass bead surface.

The boxed areas in Figure 4-1 were then examined at 1600x (Fig. 4-2). The

differences in texture of these rough patches are clearly evident: on the uncoated bead,

this region was characterized by random texture, whereas on the coated bead, it was

covered by large flakes of material. Iron, aluminum, and silicon were homogeneously

distributed on the uncoated bead (Fig. 4-2b) and unevenly distributed on the coated bead
























Al Fe SI Al Fe Si
b __d





Figure 4-1. SEM images glass beads at 250x; (a) uncoated glass bead, (b) X-ray maps of uncoated glass bead, (c) coated glass
bead, (d) X-ray map of coated glass bead. Examples of rough patches are shown in boxes.




































Figure 4-2. SEM images at 1600x of boxed areas in previous figure; (a) uncoated glass bead, (b) X-ray maps of uncoated glass
bead, (c) coated glass bead, (d) X-ray map of coated glass bead.







52

(Fig. 4-2d). Regions of high aluminum and iron concentration on the coated bead

corresponded to regions of low silicon concentration and reflected the shapes of the

flakes.

Further examinations were carried out at 6000x (Fig. 4-3). The region viewed on

each bead was determined by first finding a rough region on the surface, then rotating the

sample to show the edge of this region, as well as some foreground area. The

foregrounds of coated bead surfaces (Fig. 4-3b, 4-3c) were similar to the uncoated bead

surface (Fig. 4-3a) in appearance. Means of duplicate analyses of the foreground in

Figure 4-3c gave 1.6 % Al, 1.9 % Fe, and 18 % Si, whereas analyses of the raised region

(atop the ledge) in Figure 3c gave higher Al and Fe contents (5.2 % and 9.7 %

respectively) and lower Si content (7.2 %). The measurements indicate that the raised

region is a flake of iron aluminum (hydr)oxide coating overlying the bead surface.

Quantitative X-ray analysis at five different locations on an uncoated bead gave mean

elemental compositions of 0.7 % Al, 0.05 % Fe, and 18.0 % Si. The foreground iron and

aluminum contents were probably influenced by the proximity of the coating, and thus

were somewhat higher than those found on the uncoated bead. The raised region in

Figure 4-3b, by its contrasting appearance to the foreground, can also be concluded to

represent a flake of coating. The thickness of the flakes is 1 to 2 pIm, which is somewhat

lower than the 4 to 7 pm thickness reported by Lai et al. (2000) for iron oxide on Ottawa

sand.

The surface area of sand within media sizes of 0.6 to 0.85 mm was reported as

0.01 to 0.05 m2/g by three investigators (Benjamin et al., 1996; Chang et al., 1997;

Sansalone, 1999), whereas Lai et al. (2000) gave a surface area of 0.85 m2/g (Table 4-2).

In the present study, we measured areas of 0.11 m2/g for Ottawa sand and 0.01 m2/g for












a b c

















Figure 4-3. SEM image at a plane tangent to the surface of(a) uncoated and (b) coated glass magnified 6000x
and (c) coated glass magnified 10,000x.










Table 4-2. Effect of coating on surface area of filter media

Media Diameter Surface Area (m2/g) ASAa Surface Fe Ratio Reference / Notes
(mm) Uncoated Coated (mg/m2)
Ottawa Sand 0.60-0.85 0.04 5.8-9.1 7.4 700 Benjamin et al. (1996)
Ottawa Sand 0.60-0.85 0.04 2.7 2.7 625 Chang et al (1997)
Sand 0.67-0.99 0.85 2.76 1.9 6.7 Lai et al (2000)b
Silica Sand 0.60-0.80 0.01-0.05 5-15 10 Sansalone (1999)'
Ottawa Sand 0.6-0.7 0.106e 0.601e 0.49 11.7 Present study
Glass Beads 0.6-0.7 0.00997r 0.382e 0.37 33 Present study
Ceramic 0.6-0.7 0.0141e 1.403e 1.4 84 Present study
Beads

aChange in surface area based on median values
tNitrogen adsorbate
CEthylene glycol monethyl ether method
dKrypton adsorbate
eS.D. < 10% of the mean (N=3)







55

glass beads. Previous studies indicated that coating sand with iron oxide increased the

surface area of the media by 1.9 to 10 m2/g, whereas in the present study the surface area

of sand was increased by 0.49 m2/g and the surface area of glass beads was increased by

0.37 m2/g. Based on the data shown in Table 4-2, the increase in surface area was

significantly correlated with the surface iron ratio (a = 0.05).

Chang et al. (1997) suggested that iron oxide coating is highly porous. Coating

removed during the rinsing process in the present study had a mean surface area of 285 +

9.4 m2/g (N = 3), which is reasonably close to the range of 75 to 108 m2/g for iron oxide

coating (Sansalone, 1999). The particles of coating were irregularly shaped discs, 8 to 50

pm across and 2 to 3 lpm thick, with a mean pore diameter of 2.6 nm and total pore

volume of 0.19 cm3/g. The calculated external surface area of the coating particles

(approximated as 8 x 2 pm discs to give a high estimate) is 0.46 m2/g, a value far lower

than that measured. Thus, virtually all of the coating surface area must be within its

pores. Normalizing the iron content of the glass beads contributed by the coating to the

iron content of the coating gives a value of 1.6 mg coating per g glass. Multiplying this

value by the surface area of the coating (285 m2/g) gives 0.45 m2 of surface area per g

glass, which is reasonably close to our measured value of 0.37 m2/g.

The zeta potential of glass beads, before and after coating with iron aluminum

(hydr)oxide, is shown in Figure 4-4. The point-of-zero-charge (PZC) of the uncoated

glass beads was 3.7, which is slightly higher than the PZC of 2 to 3 given by Parks (1965)

for the alpha-quartz form of SiO2. Coated glass had a PZC of 8.2, which is consistent

with Parks' (1965) value of 7.8 to 9 for aluminum oxides. Using streaming potential

technique, Shaw et al. (2000) and Stenkamp and Benjamin (1994) found PZC's of 2.5







30
25
20
15
10
5
0
-5
-10
-15


D 2


Figure 4-4. Zeta Potential of glass beads, before and after coating with iron aluminum (hydr)oxide


S
*

SCoated
*


Uncoated o 0

o o o4 (
*







57

and 3 for uncoated Ottawa sand and 8.0 and 7.5, respectively, for coated Ottawa sand.

Using a salt addition technique, Benjamin et al. (1996) and Chang et al. (1997) found a

PZC of 0.7 for uncoated Ottawa sand and 9.3 to 9.8 and 10.3, respectively, for iron oxide

coated sand.

The zeta potential of the material collected in the final rinsing step of the coating

method used in the present study was also measured (Fig. 4-5). The PZC of this material

was 8.8, which is somewhat higher that of the coated glass beads and compares well to

PZCs in the range of 8 to 10.3 for iron oxide particles abraded from coated Ottawa sand

(Stenkamp and Benjamin, 1994; Chang et al., 1997).

4-3-2 Characterization of Ceramic Bead Surfaces

The textured nature of the ceramic bead surfaces is apparent at 250x (Fig. 4-6). The

coated bead (Fig. 4-6c) appears less porous than the uncoated bead (Fig. 4-6a),

suggesting that some of the surface texture was filled by the iron aluminum (hydr)oxide

coating. X-ray mapping shows homogenous iron distribution on the uncoated bead

surface (Fig. 4-6b) and a nonuniform distribution on the coated surface (Fig. 4-6b). At

1600x, a raised mass of material on the bead surface (Fig. 4-7c) corresponds to high-iron

regions in the x-ray map. This mass has a scale-like appearance like that seen on the

glass beads. The iron distribution on the uncoated bead is uniform (Fig. 4-7b). At 6000x,

viewed tangentially, the surface of the uncoated bead appears stringy (Fig. 4-8a), whereas

the surface of the coated bead shows cracks and has a "filled in" appearance (Fig. 4-8b).

The surface iron ratio of the coated ceramic beads was 84.2 mg/m2 (Table 4-1), a

value almost three times higher than the ratio for glass beads. The higher amount of iron







40

S30
E
20

,-
.- 10
a-
g o

N
-10

-20


10 12


Figure 4-5. Zeta Potential of material attrited from the surface of the coated media


4 6


0 2


I / ) / I r
























Al Fe Si d Al Fe Si






Figure 4-6. SEM images ceramic beads at 250x; (a) uncoated ceramic bead, (b) X-ray maps of uncoated ceramic bead, (c) coated
ceramic bead, (d) X-ray map of coated ceramic bead. Examples of rough patches are shown in boxes.



































Figure 4-7. SEM images at 1600x of boxed areas in previous figure; (a) uncoated ceramic bead, (b) X-ray maps of uncoated ceramic
bead, (c) coated ceramic bead, (d) X-ray map of coated ceramic bead.




















Figure 4-8. SEM image at a plane tangent to the surface of(a) uncoated and (b) coated ceramic magnified 6000x.








Figure 4-8. SEM image at a plane tangent to the surface of(a) uncoated and (b) coated ceramic magnified 6000x.







62

on the surface could be a result of the higher surface texture, which can protect coating

from attrition during rinsing and handling. Edwards and Benjamin (1989) found that a

portion of the iron oxide coating on sand would detach when the sand was subjected to

aggressive stirring. The surface area of the ceramic beads was increased from 0.01 m2/g

to 1.4 m2/g by the coating process, a change that we attribute primarily to porosity of the

coating, as discussed previously. Figure 4-9 shows the zeta potential of the ceramic

beads. Uncoated ceramic beads had a PZC of 7.9, which is similar to Parks' (1965) value

of 7.8 to 9 for aluminum oxides. The PZC was raised to 8.4 by the coating, which is

intermediate to the PZC of coated glass (8.2) and the coating material itself (8.8).

4-3-3 Comparison of Filter Media Performance

Batch tests were carried out to avoid column-packing effects and thus provide a

direct indication of the interaction between Cryptosporidium oocysts and beads (Table 4-

3). The tests with glass beads were carried out a pH of 7.0, where the zeta potential of

the uncoated beads was negative and the zeta potential of the coated beads was positive.

(Fig. 4-4). The uncoated glass beads removed only 5% of the Cryptosporidium oocysts

from the suspension. The coated beads were significantly (a = 0.05) more effective,

removing 49% of the oocysts from suspension. Previously, Chen et al. (1998) reported

that coating Ottawa sand with aluminum (hydr)oxide increased zeta potential from

negative to positive, while improving batch removals of Escherichia coli from 10 to 20%

to 45 to 60%.

Batch tests with ceramic beads were carried out at a pH of 6.0. At this pH, the zeta

potentials of both the uncoated and the coated ceramic beads were positive and nearly

identical in magnitude (Fig. 4-9). The uncoated ceramic beads removed 54% of the








30
25
E 20
- 15
4-
10
a)
4-
o 5
0-
N -5
-10
-15


0 2 4 6 8 10 12 14

pH


Figure 4-9. Zeta Potential of uncoated and coated ceramic beads uncoatedd are open points; coated are filled points)


Coated


0*0 oo


a3


Uncoated


I I I I I I







64


Table 4-3. Effect of coating on removal of Cryptosporidium parvum by glass or ceramic beads
in batch tests

% Removala

Uncoated Filter Media Coated Filter Media

Glass 5.3 3.79 49.3 0.58b

Ceramic 54.3 + 1.15 61.3 3.06b

aMean 1.0 S.D.;N = 3

bSignificantly higher than uncoated media at a = 0.05; N = 3







65
Cryptosporidium from suspension. Coating the ceramic beads with iron aluminum

(hydr)oxide had a negligible effect on their zeta potential, but significantly (a = 0.05)

improved their performance, increasing Cryptosporidium removal to 61%. Continuous-

flow column tests were carried out to predict media performance under typical filtration

conditions. Porosity of the packed columns of glass and ceramic beads was 39 to 40%

(Table 4-4) and was not significantly changed by the coating. The measured pressure

drop through a 1.0 m column of uncoated glass beads was 71 mm Hg (superficial velocity

= 407 m/d). This agreed closely with the calculated value of 71.3 mm Hg from the

Carmen-Kozeny equation (Montgomery, 1985) with a shape factor of 1.0 and

temperature of 200C. Using coated glass beads in the column increased the pressure drop

by 7.0%. The measured pressure drop for ceramic beads was 113 mm Hg, which agreed

with the Carmen-Kozeny equation with a shape factor of 0.95. Using coated ceramic

beads increased the pressure drop by 3.5%.

The change in concentration of particles with respect to column length in a granular

media filter may be characterized by the first-order relationship

8ac
= -Ac (4.3)
az

where c is the particle concentration at length z and X is a coefficient that characterizes

the filter media. Integrating this expression gives

Ceffl e-AL (4.4)
Cin

which is valid where the coverage of filter media by particles is low enough so that it

does not affect the rate of particle attachment. In this expression, ci, = influent







66


Table 4-4. Effect of coating on porosity and pressure drop of 1.0 m long packed columns
(Values given are mean 1.0 S.D.; N = 3.)

Porosity, % Pressure dropa, mm Hg

Uncoated Media Coated Media Uncoated Media Coated Media

Glass 39.7 0.7 39.4 0.9b 71 0.58 76 1.2

Ceramic 40.0 0.3 40.1 1.7b 113 2.0 117 +2.0

'Superficial velocity of water through columns = 407 m/d

bDifference between values for coated and uncoated media were not significant at a = 0.05

'Difference between values for coated and uncoated media were significant at a = 0.05







67
concentration of particles and cef = effluent concentration of particles. Combining

Equation (4.4) with the definition of fractional removal efficiency (fl = 1-Cff/Cin )

yields

S= 1-e--L (4.5)

Equation (4-5) indicates that, as the filter coefficient increases, the removal

efficiency for a given column length also increases.

Figure 4-10 shows least squares nonlinear fits of Equation (4-5) to removal fraction

vs. column length for glass beads at pH 7.0. The filter coefficient for coated glass beads

(5.7 nm") was increased by a factor of 3.0 over that of uncoated glass beads (1.9 nfm). The

difference in filter coefficients was significant at a = 0.05. Previously, Shaw et al. (2000)

reported a 2.9-fold increase in filter coefficient using Ottawa sand coated by iron

aluminum (hydr)oxide for removal ofCryptosporidium oocysts at pH 7.0. Since the zeta

potential of these coated media at pH 7 is positive, versus negative zeta potential of the

uncoated media, the dramatic improvement in filtration can, at least in part, be attributed

to elimination of electrostatic repulsion between the negatively charged oocysts (at pH 7;

Shaw et al., 2000) and the media.

Figure 4-11 shows fits to removal fraction vs. column length for ceramic beads.

These data were collected at pH 6.0, where the zeta potentials of both the coated and

uncoated ceramic beads were positive and nearly identical in value, whereas the oocysts

had a negative potential (Shaw et al., 2000). The filter coefficient for uncoated beads was

11.1 m"'. Coating the ceramic beads with iron aluminum (hydr)oxide had a significant (a

< 0.05) impact on filter coefficient, increasing it to 19.3 nf' (1.7 x that of the uncoated






100


*. 80 Coated Glass: A = 5.7 0.2 m-1


ii 60
S4
E

40 Uncoated Glass: A= 1.9 0.1 m-1
o
0 20 -
0-
U-
0
0 10 20 30 40 50 60 70

Column Length (cm)
Figure 4-10. Fraction of Cryptosporidium oocysts removed at pH 7.0 as a function of column length for uncoated
and coated glass beads (Error bars represent mean 1.0 S.D for triplicate experiments at each
column length. The mean filter coefficients from fits to the three sets of data are shown next to each
plot.)







100
>, Coated Ceramic: A = 19.3 + 0.7 m-1
t-/ -
I80


cc 60
S/Uncoated Ceramic: A= 11.1 0.3 m-1
E
2 40

.0
--


5 20

LL
0 I
0 10 20 30 40 50 60 70

Column Length (m)
Figure 4-11. Fraction of Cryptosporidium oocysts removed at pH 6.0 as a function of column length for uncoated
and coated ceramic beads (Error bars represent mean 1.0 S.D for triplicate experiments at each
column length. The mean filter coefficients from fits to the three sets of data are shown next to each
plot.)







70

ceramic beads). This improvement cannot be attributed to change in zeta potential of the

coated vs. uncoated media.

Additional factors contributing to improved Cryptosporidium removal by filter

media coated with iron aluminum (hydr)oxide coating may include higher surface area

available to the oocysts, change in surface roughness, and enhancement of other forces.

Most of the surface area increase attributable to coating is present in nanoscale pores,

which are too small to admit the oocysts. Nevertheless, deposition of coating in patches,

together with its scale-like morphology, should also lead to an increase in the external

surface area, resulting in a proportionate increase in capture of oocysts. An increase in

external surface area is also consistent with the slightly increased pressure drop resulting

from the coatings.

Surface roughness plays a role in interactions between colloids and surfaces, with

roughness on appropriate scale significantly increasing favorableness of interactions.

However, the coatings applied in the present research did not appear to substantially

change the surface roughness of the bead surfaces and, in the case of ceramic beads, may

have decreased surface roughness by filling in micron-scale depressions in the bead

surface.

Local variations in chemical composition of the surface can produce nonuniform

distribution of surface charge and hydrophobicity. Song et al. (1994) have shown,

theoretically, that surface chemical heterogeneities can have a profound effect on the

attachment rate of negatively-charged colloids onto surfaces with overall negative charge,

increasing predicted rates of attachment by orders of magnitude over the predicted rate

for a homogenous surface. In the present study, the zeta potential of the coating was

slightly higher than that of the ceramic beads, thus, the patchy distribution of coating on







71

the bead surfaces could affect the distribution of charges on the bead surfaces. However,

the effect of positive charge heterogeneity on a surface with overall positive zeta

potential has not been investigated. Modification of the hydrophobicity of the media

surfaces could also be a factor. Patchy distribution of the hydrophilic (hydr)oxide coating

on the ceramic surface might lead to more favorable hydrophobic interactions.

4-4 Conclusions

In situ precipitation of iron aluminum (hydr)oxide on glass and ceramic beads

resulted in a patchy distribution of coating on the bead surfaces

Large increases of BET surface area upon coating filter media with iron

aluminum (hydr)oxide coating are largely due to nanoscale pores in the coating

Pressure drop in packed columns was increased by 4-7% by use of coated media,

whereas porosity was unchanged

Coated glass beads with positive zeta potential were much more effective in

removing Cryptosporidium from water than uncoated glass beads with negative

zeta potential

Coated ceramic beads with a positive zeta potential removed significantly more

Cryptosporidium oocysts from water than did uncoated ceramic beads with

positive zeta potential

Future research on filter media coatings for enhancing microbe capture should

identify characteristics, in addition to zeta potential, that can be enhanced by appropriate

coating design. Coating morphology, which affects external surface area and surface

roughness, may offer the best possibility for further improving microbe removal in

granular media filters.















CHAPTER 5
CONCLUSIONS

The following conclusions can be drawn from this study:

A surface coating of hydrous aluminum and iron oxide on Ottawa sand is

effective in increasing the zeta potential of the sand from negative values to

positive values

Coated electropositivee) sand significantly improves removals of

Cryptosporidium oocysts from water at superficial velocities representative of

rapid sand filters operated at low to high superficial velocities

Based on the almost three-fold improvement in filter coefficient, coated sand

can significantly increases the reliability of rapid sand filtration systems and

prevent breakthrough of Cryptosporidium oocysts during periods of

suboptimal chemical conditioning

In situ precipitation of iron aluminum (hydr)oxide on glass and ceramic beads

resulted in a patchy distribution of coating on the bead surfaces

Large increases of BET surface area upon coating filter media with iron

aluminum (hydr)oxide coating are largely due to nanoscale pores in the

coating

Pressure drop in packed columns was increased by 4 to 7% by use of coated

media, whereas porosity was unchanged







73

Coated glass beads with positive zeta potential were much more effective in

removing Cryptosporidium from water than uncoated glass beads with

negative zeta potential

Coated ceramic beads with a positive zeta potential removed significantly

more Cryptosporidium oocysts from water than did uncoated ceramic beads

with positive zeta potential

Future research on filter media coatings for enhancing microbe capture should

identify characteristics, in addition to zeta potential, that can be enhanced by appropriate

coating design. Coating morphology, which affects external surface area and surface

roughness, may offer the best possibility for further improving microbe removal in

granular media filters.














CHAPTER 6
REFERENCES

Ahammed, M.M., and Chaudhuri, M. (1996) Sand-based filtration / adsorption media. J.
Water Supply Res. Technol.-Aqua 45, 67-71.

Anderson, N.J., Kolarik, L.O., Swinton, E.A., and Weiss, D.E. (1982) Color and turbidity
removal with reusable magnetic particles III. Immobilized metal hydroxide gels. Water
Research, 16, 1327-1334.

Angus, K.W., Appleyard, W.T., Menzies, J.D., Campbell, I., and Sherwood, D. (1982)
An outbreak of diarrhea associated with cryptosporidiosis in naturally reared lambs.
Veterinary Record 110:129.

Benjamin, M.M., Sletton, R.S., Bailey, R.P., and Bennett, T. (1996) Sorption and
filtration of metals using iron-oxide coated sand, Wat. Res. 30(11) 2609-2620.

Chang, y., Li, C.-W., and Benjamin, M.M. (1997) Iron oxide-coated media for NOM
sorption and particulate filtration. Journal AWWA. 89, 5, 100-112

Chen, J., Truesdail, S., Lu, F., Zhan, G., Belvin, C., Koopman, B., Farrah, S., and Shah,
D. (1998) Long-term evaluation of aluminum hydroxide-coated sand for removal of
bacteria from wastewater. Wat. Res. 32, 2171-2179.

Considine, R.F., Dixon, D.R., and Drummond, C.J. (2000) Laterally-Resolved Force
Microscopy of Biological Microspheres Oocysts of Cryptosporidium Parvum,
Langmuir, 16, 1323-1330.

Current, W.L. and Garcia, L.S. (1991) Cryptosporidiosis. Clin. Micorbiol. Rev. 4, 325-
358.

Current, W.L., Reese, N.C., Ernst, J.V., Bailey, W.S., Heyman, M.B., and Weinstein,
W.M. (1983) Human cryptosporidiosis in immunocompetent and immunodeficient
persons. Studies of an outbreak and experimental transmission. New England Journal of
Medicine 308, 1252-1257.

Derjaguin, B.V. and Landau, L.D. (1941) Theory of Stability of strongly charged
lyophobic sols and the adhesion of strongly charged particles in solutions of electrolyte.
Acta Physiochim. URSS, 14, 733.









Dickinson, R.B. (1997) A dynamic model for the attachment of a brownian particle
mediated by discrete macromolecular bonds, Journal of Colloid and Interface Science,
190, 142-151.

Dickinson, R.B., Ruta, A.G., and Truesdail, S.E. (2000) Physiochemical basis of
bacterial adhesion to biomaterial surfaces in Anti-microbial/Anti-infective materials:
Principles, Applications, and Devices, Swan, S.P. and Manivannan, G. (eds.),
Technomic, Lancaster, PA.

Drozd and Schwartzbrod (1996) Hydrophobic and electrostatic cell surface properties of
Cryptosporidium oocysts. Applied and Environmental Microbiology 62 (4) 1227.

Edwards, M. and Benjamin, M.M. (1989) Adsorptive filtration using coated sand: a new
approach for treatment of metal-bearing wastes. Journal WPCF, 61, 9, 1523-1533.

Elimelech, M., Gregory, J., Jia, X., and Williams, R. (1995) Particle Deposition and
aggregation, measurement, modelling, and simulation. Butterworth Heinemann, Oxford.

EPA / 814-B-95-003, ICR Protozoan method for detecting Giardia cysts and
Cryptosporidium oocysts in water by fluorescent antibody procedure, Office of
groundwater and drinking water.

Fairbrother, F. and Mastin, H. (1924) Studies in Electro-endosmosis. Part I. J Chem.
Soc., 125, 2319-2330.

Farrah, S. R., and Preston, D. R. (1991). Adsorption of viruses by diatomaceous earth
coated with metallic oxides and metallic peroxides. Wat. Sci. Tech., 24(2), 235-240.

Fayer, R., Graczyk, T.K., Cranfield, M.R., and Trout, J.M. (1996) Gaseous disinfection
of Cryptosporidium parvum oocysts. Appl. Environ. Microbiol., 62, 3908-3909.

Fayer, R. and Unger, B. (1986) Cryptosporidium spp. And cryptosporidiosis. Microbiol.
Review, 50,458-483.

Fogel, D., Isaac-Renton, J., Guasparini, R., Moorehead, W., and Ongerth, J. (1993)
Removing Giardia and Cryptospordium by slow sand filtration. J. AWWA. 77-84.

Gerba, C.P., Hou, K., and Sobsey, M.D. (1988) Microbial recovery and inactivation
from water by filters containing magnesium peroxide. J. Environ. Sci. Health. A23, 41-
58.

Ghosh, M.M., Jordan, T.A. and Porter, R.L. (1975) Physiochemical Approach to Water
and Wastewater Filtration J Env. Eng. Div. 71.

Hamaker, H.C. (1937) The London-van der Waals attraction between spherical particles.
Physica, 4, 1058.









Hiemenz, P.C. and Rajagopalan, R. (1997) Principles of Colloid and Surface Chemistry,
Third Edition, Marcel Dekker, N.Y., N.Y.

Holly, H.P. Jr. and Dover, C. (1987) Cryptosporidium: a common cause of parasitic
diarrhea in otherwise healthy individuals. Journal ofInfectious Disease 153:365-368.

Israelachvili, J.N. (1992) Solvation, Structural, and Hydration forces, Intermolecular and
Surface Forces, Academic Press, N.Y., N.Y., 260-287.

Kang, P. (1998) Surface Modification of Fibous Filter Media to Enhance Filtration
Efficiency, Ph.D. dissertation, Department of Chemical Engineering, University of
Florida, Gainesville, FL

Karaman, M.E., Pashley, R.M., Bustamante, H., and Shanker, S.R. (1999)
Microelectrophoresis of Cryptosporidium oocysts in aqueous solutions of inorganic and
surfactant cations. Colloids and Surfaces A: Physiochemical & Engineering Aspects.
146, 217.

Lai, C. H., Lo, S. L., and Lin, C. F. (1994). Evaluating an iron-coated sand for
removing copper from water. Wat. Sci. Tech. 30(9), 175-182.

Levine, N.D. (1984) Taxonomy and review of the coccidian genus Cryptosporidium
(Protozoa, Apicomplexa). Journal of Protozoology. 31, 82.

Liftshitz, E.M. (1956) Theory of molecular attractive forces. Soviet Physics JETP, 2, 73.

Lo, S.-L., Jeng, H.T., and Lai, C.-H. (1997) Characteristics and adsorption properties of
iron-coated sand. Wat. Sci. Tech. 35, 63-70.

Long, P.L. (ed.) (1982) The Biology of the Coccidia. University Park Press,
Baltimore, MD.

Lukasik, J., Cheng, Y.F., Lu, F., Tamplin, M., and Farrah, S.R. (1999) Removal of
microorganisms from water by columns containing sand coated with ferric and aluminum
hydroxides. Water Research, 33, 3, 769-777.

Lukasik, J., Farrah, S.R., Treusdail, S., and Shah, D.O. (1996) Adsorption of
Microorganisms to sand and diatomaceous earth particles coated with metallic
hydroxides. KONA, 14, 1-6.

Marshall, A.R., Al-Jumailli, I.J., Fenwick, G.A., Blint, A.J., and Record, C.O. (1987)
Cryptosporidiosis in patients at a large teaching hospital. Journal Clinical Microbiology
25,172-173.

McCaulou, D.R., Bales, R.C., and McCarthy, J.F. (1994) Use of short-pulse experiments
to study bacteria transport through porous media. J. Contain. Hydrol. 15,1-14.









Meinhardt, P.L., Casemore, D.P., and Miller, K.B. (1996) Epidemiologic aspect of
human Cryptosporidiosis and the role of waterborne transmission. Epidemilogic
Reviews. 18(2), 118-136.

Meisel, J.L., Perara, D.R., and Rubin, C.E. (1976) Overwhelming watery diarrhea
associated with Cryptosporidium in an immunosuppressed patient. Gastroenterology 70,
1156-1160.

Mills, A.L., Herman, J.S., Horberger, G.M., and DeJesus, T.H. (1994) Effect of
solution ionic strength and iron coatings on mineral grains on the sorption of bacterial
cells to quartz sand. Appl. Environ. Microbiol. 60, 3300-3306.

Montgomery, J.M. (1985) Water Treatment Principles and Design, Wiley, New York,
New York.

Navin, T.R. and Juranek, D.D. (1984) Cryptosporidiosis: clinical epidemiologic and
parasitologic review. Reviews on Infectious Disease 6(B),313-327.

Nime, F.A., Burek, J.D., Page, D.L., Holscher, M.A., and Yardley, J.H. (1976) Acute
enterocolitis in a human being infected with the protozoan Cryptosporidium.
Gastroenterology 70,592-598.

Ongerth, J.E. and Pecoraro, J.P. (1995) Removing Cryptosporidium using Multimedia
Filters. J. AWWA. December, 1995, 83-89.

Ongerth, J.E. and Pecoraro, J.P. (1996) Electrophoretic mobility of Cryptosporidium
oocysts and Giardia cysts. J. Environmental Engineering. 228-231.

Panciera, R.J., Thomassen, R.W. and Garner, F.M. (1971) Cryptosporidial infection in a
calf. Veterinary Pathology 8, 479-484.

Parks, G.A. (1965) The isoelectric point of solid oxides, solid hydroxides, and aqueous
hydroxo complex systems. Chem. Rev. 65,177.

Rose, J.B. (1998) Occurrence and significance of Cryptosporidium in water. J. AWWA.
80(2), 53-58.

Rose, J.B. (1990) Occurrence and control of Cryptosporidium in Drinking water,
Drinking water microbiology : Progress and recent developments, Springer-Verlag, N.Y.,
294-321.

Sansalone, J.J. (1999) Adsorptive infiltration of metals in urban drainage media
characteristics. The Science of the Total Environment, 235, 179-188.

Schuler, P.F., Ghosh, M.M., and Gopalan, P. (1991) Slow sand and diatomaceous earth
filtration of cysts and other particulates. Water Research. 25, 995-1005.









Shaw, K.M., Walker, S., and Koopman, B. (2000) Improving Filtration of
Cryptosporidium, JAWWA, 92,11,103-111.

Slavin, D. (1955) Cryptosporidium meleagridis (sp. nov.) Journal of Comparative
Pathology 65, 262-266.

Smith, H.V. and Rose, J.B. (1998) Waterborne cryptosporidiosis: current status.
Parasitology Today, 14(1) 14-22.

Song, Lianfa, Johnson, P.R., and Elimelech, M. (1994) Kinetics of Colloidal Deposition
onto Heterogeneously Charged Surfaces in Porous Media, Environ. Sci. Technol., 28,
1164-1171.

Stahl, R. S., and James, B. R. (1991). Zinc sorption by iron-oxide-coated sand as a
function of pH. Soil Sci. Soc. Am. J., 55, 1287-1290.

Stenkamp, S.V. and Benjamin, M.M. (1994) Effect of iron oxide coating on sand
filtration. Journal A WWA ,86(8)37-50.

Stenstrom, T.A. (1989) Bacterial hydrophobicity, an overall parameter for the
measurement of adhesion potential to soil particles, Applied and Environmental
Microbiology, 55, 1, 142-147.

Tilley, M. and Upton, S.J. (1997), In Cryptosporidium and cryptosporidiosis; Fayer, R.,
ed.; CRC Press, Boca Raton, FL, 163-180.

Truesdail, S. (1999) Fundamental Forces Important to the Enhancement of Biological
Colloid Removal, Ph.D. Dissertation, Department of Chemical Engineering, University
of Florida, Gainesville, Florida.

Tyzzer, E.E. (1910) An extracellular coccidium, Cryptosporidium muris (gen. Et sp.
nov.), of the gastric glands of the common mouse. Journal of Medical Research 23:487-
516.

Tyzzer, E.E. (1912) Cryptosporidium parvum (sp. nov.), a coccidium found in the small
intestine of the common mouse. Arch. Protistenkd. 26:394-412.

Upton, S.J. (1997) In vitro cultivation of Cryptosporidium. In : Cryptosporidiosis of Man
and Animals, 2nd ed. Chapter 8, Fayer, R., ed.,CRC Press, Boca Raton, 181-207.

vanLoosdrecht, M.C.M., Lyklema, J., Norde, W., Schraa, G., and Zehnder (1987)
Electorphoretic Mobility and Hydrophobicity as a measure to predict the initial steps of
bacterial adhesion, Applied and Environmental Microbiology, 53, 8, 1898-1901.

vanLoosdrecht, M.C.M., Lyklema, J., Norde, W., Schraa, G., and Zehnder (1987) The
role of bacterial cell wall hydrophobicity in adhesion, Applied and Environmental
Microbiology, 53, 8, 1893-1897.







79


Verwey, E.J.W. and Overbeek, J.T.G. (1948) Theory of the stability oflyophobic
colloids, Elsevier, Amsterdam.

Wennberg, K.E. and Sharma, M.M. (1997) Determination of the filtration coefficient
and the transition time for water injection wells, Proceedings of the 1997 European
formation damage conference (Society of Petroleum Engineers), 353-364

Whitmore, T.N., and Carrington, E.G. (1993) Comparison of Methods for Recovery of
Cryptosporidium from water. Water Science and Technology. 27, 69-76.

Woods, K.M. and Upton, S.J. (1998) Efficacy of select antivirals against
Cryptosporidium parvum in vitro. FEMS Microbiology Letters 168,1, 59-63.















BIOGRAPHICAL SKETCH

Kathryn Shaw was born in Norwood, Massachusetts. She received a Bachelor of

Science in Chemistry and Diploma in Engineering from the University of Prince Edward

Island, Canada in 1989 and a Master of Science in (Chemical) Engineering from the

University of New Brunswick, Canada in 1993. She then returned to the United States to

study for a Ph.D. at the University of Florida.









I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philoso hy.


Ben L. Koopmarhairman
Professor of
Environmental Engineering Sciences

I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.


sephJ.' tfino
Professor of
Environmental Engineering Sciences


I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.


Hassan E. El-Shall
Associate Professor of
Materials Science and Engineering


I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

S ,LT,-, r' F. -'-*
Samuel R. Farrah
Professor of
Microbiology and Cell Science


I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.


Spr Svoronos
Professor of Chemical Engineering









This dissertation was submitted to the Graduate Faculty of the College of
Engineering and to the Graduate School and was accepted as partial fulfillment of the
requirements for the Degree of Doctor of Philosophy.

August 2001


Pramod P. Khargonekar
S Dean, College of Engineering




Winfred M. Phillips
Dean, Graduate School



















































I 3 1262 08li 556 6122I I
3 1262 08556 6122




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EEK2YB5NW_V0Y5KY INGEST_TIME 2013-10-24T21:10:28Z PACKAGE AA00017643_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES