2..“
«1. 1. .

a.

(E? .
it!)

I .13;
v3.1.5: .72.:
b. 1
:1:
.
«H .JL:
I

1k. 3

Jr... 1.5..

.1»

{Sir

#Iarafii yisr it

. . 1‘ 42.136.11.11

5 . 2:331 .3 is \o
, 7:: .13 I} . .3

3:;17. 2. )2:

r... a. 1.1. v. a.

r.
16.4.?
.11
.35.»...
~v. ‘ 1.
>9

 

.
3:11? u
. .x.£.‘7.«

3...»:

 

5.3.

 

3.1.3.1.. . $3.10 : , .- . it: .a.l t
.. . , .grfimmgmu
V .f .3 :ka .1}-

i 1.; It}
. sna\ .n..J....

llBRARY
,2 Michigan State a
University

 

 

 

This is to certify that the

thesis entitled

THE REMOVAL OF ARSENIC FROM DRINKING WATER
BY CARBON ADSORPTION

presented by

Masanori Fuj imoto

has been accepted towards fulfillment
of the requirements for

 

M.S. degree in Environmental Engineering

 

@070» we 6 date,

Major professor

 

«3/0:

Date H

1x

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DateDquGS-nts

THE REMOVAL OF ARSENIC FROM DRINKING WATER BY
CARBON ADSORPTION

By

Masanori Fujimoto

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE
Department of Civil and Environmental Engineering

2001

ABSTRACT

THE REMOVAL OF ARSENIC FROM DRINKING WATER BY
CARBON ADSORPT ION

By

Masanori Fujimoto

It is estimated that millions of people who are drinking arsenic
contaminated groundwater are suffering from arsenic related diseases
such as skin, bladder and kidney cancer in Bangladesh. The system to
remove arsenic must be economically feasible for the members of affected
communities. In a recent report, Ansari et al have concluded that carbon
adsorption, which is a comparatively cheap and easy to operate method,
is a potentially feasible method for the removal of arsenic. The present
study focused on activated carbon as an adsorbent and four different
types of activated carbon, which are CPG-LF, F400, OLC and DSR-C,
were tested. Both batch and column experiments were conducted to
observe the adsorptive kinetics and capacity of each activated carbon.
Arsenic is present in natural water as As(V) or As(III), and all four
activated carbons removed As(V) effectively. As(III), which is more toxic
than As(V), was also effectively removed by CPG-LF in the batch tests. By
evaluating the adsorptive kinetics and capacity of each activated carbon,
it was concluded that CPG-LF is the best carbon adsorbent of the four

activated carbons for arsenic removal.

Copyright by
Masanori Fujimoto
200 1

To my parents, fiancée, and people in Bangladesh

ACKNOWLEDGEMENTS

First, I would like to thank my advisor Dr. Voice for his guidance and
support. I also thank my thesis committee members Dr. Long, Dr. Zhao,
and Dr. Hashsham for their assistance. I also thank Dr. Loconto, Mr.
Pan, and Mr. Joseph for their assistance in managing the instrumental

facilities.

I also appreciate my coworkers in Environmental Engineering, Shawn,
Chris, K.C., and Lisveth. I would like to express my special thanks to Ms.
Lisveth Flores for her guidance and assistance in helping with the
laboratory work and providing information about arsenic speciation and
toxicity. I also thank my host family, the Tim Potters who have supported

my stay here at Michigan State University.

Finally, I would like to thank my family in Japan, especially my parents
who always love me and worry me. I must also thank my fiancee Ms.
Atsuko who always cheers me up with a phone call. My work was

supported by all of these people.

TABLE OF CONTENTS

DEDICATION ...................................................................................... iv
ACKNOWLEDGEMENT ..................................................................... v
TABLE OF CONTENTS ...................................................................... vi
INTRODUCTION ..................................................................... 1
1. Background ....................................................................... 3
1.1. Arsenic Regulation ............................................................... 3

1.2. Source ................................................................................. 4

1.3. Speciation ........................................................................... 6

1.4. Toxicity ............................................................................... 7

1.5. Health Effect ....................................................................... 8

1.6. Removal Technology ............................................................ 9

1.6.1 . Coagulation /Precipitation followed by Filtration ........................ 9

1.6.2. Ion exchange ........................................................................... 10

1.6.3. Activated Almina ..................................................................... 1 1

1.6.4. Membrane Process (RO /Nanofi1tration) .................................... 12

1.6.5. Oxidation ................................................................................ 13

1.6.6. Removal Technology in Bangladesh ......................................... 15

1.7. Theory of Carbon Adsorption ............................................. 17

1.7.1. Mechanisms of Adsorption ....................................................... 17

1.7.2. Characteristics of Activated Carbon ......................................... 18

1.7.3. Isotherm .................................................................................. 19

2. Methodology ..................................................................... 20
2.1. Material ............................................................................. 20

2.1.1. Activated Carbon ..................................................................... 20

2.1.2. Treatment by Ferrous Sulfite ................................................... 21

2.2. Chemicals ......................................................................... 23

2.2.1. Arsenate .................................................................................. 23

2.2.2. Arsenite ................................................................................... 23

2.2.3. Reducing Agent ....................................................................... 23

2.3. Instruments ...................................................................... 26

2.3.1. GFAAS .................................................................................... 26

2.3.2. ORP, pH and DO Meter ............................................................ 28

3. BATCH TEST .................................................................... 29
3.1. Arsenate Adsorption .......................................................... 29

3.1.1. As(V) Kinetics Studies .............................................................. 29

3.1.2. Adsorption Kinetics Changes ................................................... 35

vi

3.1.3. As(V) Isotherm ......................................................................... 36

3.2. Arsenite Adsorption ........................................................... 39

3.2.1. As(III) Kinetics Studies ............................................................. 39

3.3. pH Effect on Adsorption ..................................................... 60

3.4. Regeneration ..................................................................... 62

3.5. Summary of Batch Test ..................................................... 64

4. Column Test ..................................................................... 65
4.1. Arsenate Adsorption through Column ................................ 66

4.2. Effect of Bed Volume (Contact Time) .................................. 68

4.3. Comparison of As(V) and As(III) Adsorption ........................ 69

4.4. Summary of Column Test .................................................. 72

5. Conclusion ....................................................................... 72
Appendix .............................................................................. 74
A-l. Chemical Description ........................................................ 75

A-2. Amount of Activated Carbon for Batch Test ....................... 76

A-3. Purging Effect on DI Water ................................................ 77

A-4. Shaking Effect on Sodium Thiosulfate ............................... 78

A-5. Eh (Redox-Potential) .......................................................... 80

A-6. ORP Measurement ............................................................ 80
Reference ............................................................................. 8 1

vii

INTRODUCTION

Arsenic, which is a naturally occurring element, contaminates
groundwater in many countries such as Argentina, Australia,
Bangladesh, Chile, Hungary, India, Mexico, Peru, Taiwan, Thailand, and
the United States (WHO, 2001). Among these countries, the severest
problem occurs in Bangladesh where up to 90% of the rural inhabitants
depend on groundwater for drinking. According to a British Geological
Survey (BGS), 28 to 35 million people are exposed to arsenic
concentrations above 50 ppb and the number of people exposed to more
than 10 ppb is. 46 to 57 million (BGS, 2000). It is estimated that millions
of people in Bangladesh are suffering from arsenic related disease such
as skin, lung, bladder, and kidney cancer, and among them about twenty
thousand people die each year, according to the UN Development

Program.

There are some effective technologies to treat arsenic contamination,
such as chemical precipitation, ion exchange and membrane methods.
However, these treatments are expensive and are not easy to operate, so
the introduction of these techniques to developing nations, such as
Bangladesh, is questionable. In a recent report, Ansari et a1 has
concluded that carbon adsorption, which is comparatively cheap and
easy to operate, is a potentially feasible method for the removal of arsenic

after testing several different kinds of adsorbents, such as zeolite,

1

chabazite, clinoptiolite and activated carbon (Ansari, 2000).

This study focused on only activated carbon as an adsorbent. Four
different types of activated carbon were tested. The primary objective of
this study was to evaluate the best carbon adsorbent for arsenic removal,

observing the differences in the adsorption kinetics and capacity.

The ultimate goal of this study is to propose a treatment for the
groundwater in Bangladesh. Arsenic is found mainly in two forms:
trivalent arsenite [As(III)] and pentavalent arsenate [As(V)]. As(III), which
is more toxic than As(V), becomes predominant species in groundwater.
Though most wells affected by arsenic in Bangladesh are shallow wells
(12-70m depth), arsenic can be in the reduced form As(III) since the
oxidation rate of As(III) is slow (Smedley, 2001). Indeed, arsenic
speciation studies revealed that the proportion of arsenite in well water
in Bangladesh is between 50% to 60% of total arsenic (BGS and DHPE,
2001). Therefore, the effect of arsenic oxidation state in the adsorption
process is significant in addressing this problem. A secondary objective
of this study is to observe how well activated carbon adsorbs arsenic in

the reduced form As(III).

The removal system must be economically feasible for members of

affected communities. Several treatment systems are available in

Bangladesh such as using the Safi filter and Kolshi filter (Jakariya,
2000). Comparison among available techniques with activated carbon
must be done in terms of removal efficiency and cost. Activated carbon is
considered to be cost competitive point of use method. This study will
provide people in Bangladesh with an economical and easy-to-use

treatment process.

1. Background

1. 1. Arsenic Regulation

World:

An international standard of 0.20 mg/L (200 ppb) was established as the
allowable level for arsenic in 1958. In 1963 the standard was reevaluated
and reduced to 0.05 mg/ L (50 ppb). In 1984 this was maintained as
World Health organization’s (WHO) guideline value and many countries
have kept this as the national standard. In 1993, 0.01 mg/ L (10 ppb)
was established as WHO’S provisional guideline value for arsenic in
drinking water based on analytical capability (WHO, 2001). The Japanese
standard for drinking water is 10 pg/ L and the maximum permissible
concentration for As in Canada is 25 ug/L. However, most of the drinking
water standards in developing nations are still in the range of 0.04 to
0.05 mg/ L. This is partly because there is no adequate testing facilities

for lower concentrations in those countries.

United States:

The current arsenic standard of 50 ppb was set by US. Environmental
Protection Agency (USEPA) in 1975 based on a Public Health Service
standard set in 1942. Under the 1996 SDWA (The Safe Drinking Water
Act) amendment, the USEPA was required to propose a new standard for
arsenic in drinking water. In January 2001, the USEPA proposed
lowering the amount of arsenic allowed in drinking water from 50 ppb to
10 ppb based on the evidence of cancer risk from high arsenic doses in
Taiwan and Chile (Smendley, 2001). A few months later, the new
administrator of the EPA announced a delay in the implementation of the
new standard on the basis that the scientific indicators are unclear as to
whether the standard needs to go as low as 10 ppb. Currently it is under
consideration, and a revised standard between 3 ppb and 20 ppb will be

announced by February 2002 (Franz, 2001).

Bangladesh:
The national standard for drinking water in Bangladesh is 0.05 mg/ L (50

ppb) that was established by following the WHO guideline (WHO, 2001).

1.2. Source

Arsenic is widely distributed throughout the earth’s crust and commonly
found in rocks and soils with a high sulfur content. Arsenic is introduced
into water through the dissolution of minerals and ores, and by

agricultural application such as pesticides, herbicides and insecticides.

4

Mining and industrial effluent also contribute arsenic to water in some
areas.

In Bangladesh, dissolution of natural minerals causes arsenic to be
released into groundwater. There appears to be three distinct dissolution
processes that can release arsenic from mineral sources into
groundwater. The first is an increase of pH to levels above pH 8.5. This
increase of pH leads to desorption of adsorbed arsenic from mineral
sources. The second process results from the development of reducing
conditions. The reducing condition converts arsenic species from
arsenate to arsenite and it results in the release of arsenic from minerals.
The third potential process involves the oxidation of sulfide minerals.
Arsenic is released from sulfide minerals as they are oxidized by natural
weathering processes. Released arsenic can leach into the groundwater
and be transported to water supply wells.

Arsenate is well adsorbed onto the surface of minerals. The adsorption of
arsenate by iron oxides plays an important role preventing widespread
arsenic toxicity problems in nature. Arsenite is a neutral molecule at
near neutral pH range and is therefore less adsorbed on most mineral
surfaces than arsenate. Thus, arsenite is generally more mobile than
arsenate and may be more likely contaminate water supplies (Smendley,

2001i

1.3. Speciation

 

1200 '

800 "— l \
HASO;

 

 

 

 

 

 

400 )-
H2ASO42-
’5‘
g H3ASO30 A8043-
.: 0
LL!
HAsSzo 3_
A803
-400 A88; H2A803:
_ HASO32- .
‘800 I 11 1 I l
0 2 4 8 10 12 14

Figurel: Eh-pH diagram for Arsenic Species at 25 °C (Korte, 1991)

(Smedley, 2000)

Figurel is a plot of the redox potential (Eh) versus pH for the arsenic
species. The Eh-pH diagram, which indicates the predominant species in
solution, is the important factor for arsenic speciation. The domain of

each species is defined by the solid lines. At the boundaries of two

domains, two species have equal activities (Cullen and Reimer, 1989).
Arsenic occurs in the environment in several oxidation states (-3, 0, +3,
+5). In natural water arsenic is mostly found in inorganic form as

trivalent arsenite [As(III)] or pentavalent arsenate [As(V)]. Arsenate has

four species: H3AsO4, H2AsO4', HAsO42' and AsO43‘ with pKa values of
2.2, 6.9 and 11.5. Arsenite also has four species: H3AsO3, H2AsO3‘,

HA3032‘ and AsOs3‘ with pKa values of 9.2, 12.1 and 13.4 (Cullen and

Reimer, 1989). It is apparent that in surface water under oxidizing
condition, the As(V) species is predominant. On the other hand, in anoxic

water under reducing condition, the As(III) species becomes stable. At

near neutral pH, the predominant species are H2AsO4‘ and HAsO42' for

arsenate, and uncharged H3AsOs for arsenite.

1.4. Toxicity

The toxicity of arsenic is dependent on its chemical form. Arsenite is
considered to be more toxic than arsenate according to several
researches. Inorganic arsenic is combined with sulfhydryl groups in
proteins and is accumulated in human bodies. Since arsenite has higher
affinity for proteins than arsenate, arsenite is more toxic. Arsenate can
be reduced to arsenite by the activity of glutathione and results in the

same toxicity. However, since not all of the arsenate can be converted to

arsenite, the toxicity of arsenate is less than arsenite.

The toxicity of inorganic arsenic can be reduced by methylation. In the
methylation process, inorganic arsenic is converted to methyl arsenic
which is less toxic and is primarily excreted in urine.

Human sensitivity to the toxic effects of inorganic arsenic exposure is
likely to vary based on genetic differences, diet, health conditions, sex
and other possible factors. These factors are important in the risk
assessment to arsenic exposure. For example, people with less ability to
methylate arsenic might accumulate more arsenic in their tissues and
may be more at risk of developing toxic effects. One study indicates that
children and people in poor nutritional conditions might have a

decreased ability to methylate arsenic (NAS, 1999).

1.5. Health Effect

In 1980, the International Agency for Research on Cancer (IARC)
determined that inorganic arsenic compounds are skin and lung
carcinogens in humans. Over the years arsenic is accumulated in
humans bodies and the physical symptoms emerge after a decade or
more of drinking contaminated water. The evidence from human studies
in Taiwan, Chile and Argentina indicates that long term exposure to
arsenic may cause skin, lung, bladder, and kidney cancer. Levels of
arsenic in water at 50 ppb may result in a combined cancer risk of one

out of one hundred. To lower the cancer risk to 1/10,000, which is

8

considered to be an acceptable risk, the concentration of arsenic in
drinking water needs to be reduced to 2 ug/ L (2 ppb) according to the

National Academy of Sciences (NAS, 1999).

1.6. Removal Technology

1.6.1. Coagulation / Precipitation followed by Filtration

Arsenic coagulation with metal salts has been demonstrated since 1934.
The most commonly used metal salts are aluminum salts such as alum,
and ferric salts such as ferric chloride. Laboratory scale tests report that
both alum and ferric chloride can remove arsenic efficiently under
optimal conditions (Cheng et al., 1994). The optimal conditions vary
among coagulants. Coagulation with ferric chloride works effectively at
pH below 8 (Edwards, 1994). Alum works effectively in a narrower pH
range from 6 to 7 (Hering et al., 1997). In general, high dosage of
coagulant achieves high removal efficiency. Typical doses of ferric
chloride and alum for arsenic removal are 5 to 30 mg/ L and 10 to 50
mg/ L, respectively.

The mechanisms of coagulation are described as follows. Alum and ferric
chloride dissolve into solution and form amorphous hydrous aluminum
oxide (HAO) and ferric oxide (HFO), respectively. These metal hydroxides
form flocs and bind to other flocs. Flocs grow until they get heavy enough

to settle down. Soluble arsenic species are physically involved in the

process of the formation of flocs and are co-precipitated. Soluble arsenic
species are also adsorbed on the external surface of the insoluble metal
hydroxide and are precipitated.

Similar removal efficiencies were achieved with either ferric chloride or
alum for As(V). However, As(III) is less efficiently removed than As(V) by
coagulant especially using alum, so pre-oxidation is necessary for better
removal. The reason ferric chloride removes As(III) more efficiently than
alum is that ferric chloride oxidize the arsenite to arsenate and the
formed arsenate is removed by coagulation (Edwards, 1994).

Lime softening, which is a conventional treatment process, is also
effective to remove the arsenic species. This process works effectively in
the pH range of 10.5 to 12 for the removal of arsenate. The main
mechanism was found to be the sorption of arsenic onto magnesium
hydroxide solids that is formed during softening. Arsenite is not removed

effectively by lime softening process (Johnston et al., 2001).

1.6.2. Ion exchange

Ion exchange is commonly used in water softening, but some ion
exchange resins work effectively for removing arsenic. Strong base anion
exchange resins, which are used to remove anion species, are usually
used to remove arsenic. This resin only works for removing arsenate that
is negatively charged above pH 2.2. Negatively charged arsenate is
strongly attracted to sorption sites on the surface of the resin and

10

effectively removed from the solution. However, arsenite is not removed
by the resin because of its uncharged form below pH 9.2 in solution.
Therefore, if water contains arsenite, a pre-oxidation process is necessary
to achieve better removal. Ion exchange resins are usually used in a
packed bed or column. A bed can treat several hundred to a thousand
bed volumes before arsenic breakthrough occurs. The run length of a
column is dependent on the competing anion species in solution. The
selectivity of sulfate is especially high for anion exchange resins, so the
run length becomes shorter with high concentration of sulfate. Ion
exchange resins work effectively with sulfate concentration of under 120
mg/L according to EPA (USEPA, 2000). Ion exchange capacity is usually
measured in a unit of meq/mL. Typical theoretical capacity of strong
base anion exchange resin ranges from 1.0 to 1.4 meq/mL, which is
converted to 315 mg As/ g (Clifford, 1999). However, an actual study
using laboratory reagents found that the operational capacity of anion
exchange resin to be about 64 mg As/ g (Baes et al., 1997). When the
resin is saturated, ion exchange resin is regenerated with concentrated

salt solution (Johnston et a1, 2001) (Wang et a1, 2000).

1.6.3. Activated Almina
Activated alumina is a granular form of aluminum oxide (A1203) with a

high internal surface area in the range of 200—300 m2/ g. Activated

alumina is usually used in packed beds. The kinetics of arsenic

11

adsorption onto activated alumina surface are slower than those onto ion
exchange resins, but usually it has longer run times than ion exchange.
Typically, several thousands of beds volume can be treated before arsenic
is saturated. One study showed that under optimal condition, activated
alumina treated arsenate with an influent concentration of 100 ug/ L
(100 ppb) for 10,000 to 20,000 bed volumes before gradual breakthrough
was observed (Frank and Clifford, 1986). These investigators also
reported that the capacity of activated alumina for arsenic adsorption is
4 mg of As/ g of activated alumina under optimal condition. Activated
almina works effectively in the pH range of 5.5 to 6.0, and above pH 7.0
removal efficiency drops sharply (Clifford, 1999). Arsenate is more
effectively removed by activated alumina than arsenite. Therefore, for
better results, raw water containing arsenite should be oxidized before
treatment. Saturated activated almina can be regenerated by flushing
with strong base such as sodium hydroxide followed by strong acid,
which re-establishes a positive charge on the surface (Johnston et a1,

2001) (Wang et a1, 2000).

1.6.4. Membrane Process (RO/Nanofiltration)

The membrane process allows water and small molecules to pass
through but rejects larger molecules and particles. The removal efficiency
using the membrane process is independent of pH but depends on the

molecular size of substance being removed. High pressure membranes

l2

such as reverse osmosis (R0) and nanofiltration have appropriate pore
size for removal of dissolved arsenic. Recent studies showed that both
reverse osmosis and nanofiltration are equally effective for both arsenic
in reduced form and oxidized form at pressures ranging from 40 to 400
psi. Therefore, membrane processes seem most suitable for arsenic
removal from groundwater in which most of the arsenic is present in
reduced form. A study has shown that the arsenic removal using a
membrane process is independent of pH, so there is no need to adjust pH
(Waypa et a1, 1997).

The fouling of membranes is a primary disadvantage of this process.
Constituents such as organic matter, iron and manganese can foul
membranes. Therefore, the membrane processes usually require
pretreatment of raw water in order to prevent fouling. Membrane systems
are generally more expensive than other arsenic removal options

(Johnston et a1, 2001).

1.6.5. Oxidation

Oxidation itself does not remove arsenic from drinking water, but
oxidation is often useful to optimize other treatment processes. Most
arsenic removal technologies are not so effective in removing arsenite
because arsenite is uncharged below pH 9.2. Therefore many treatment
systems include the oxidation step in removal processes.

Oxidation of arsenite by dissolved oxygen is a particularly slow reaction.

13

Cherry and others reported that even when distilled water spiked with
arsenite (pH 7) is saturated with oxygen, arsenite stayed uncharged for
days (Cherry et al., 1979). The oxidation rate of arsenic is dependent on
the concentration of dissolved oxygen in solution. Therefore when water
is saturated with air instead of pure oxygen, the oxidation rate becomes
slower. Kim and Nriagu showed that the half-life of arsenite in water
saturated with pure oxygen and air is 2-5 days and 4-9 days, respectively
(Kim and Nriagu, 2000). Another study reports that in natural water half-
life of arsenite is 1-3 years, although the rate might be higher due to the
presence of oxidants in solution (Eary and Schramke, 1990).

Adding oxidants such as chlorine, permanganate, ozone and hydrogen
peroxide helps increase the rate of oxidation of arsenic. Among them,
chlorine, which is widely available in developing nations, appears to be a
rapid and effective oxidant for arsenic. However, it is well known that
chlorine may produce toxic by-products when reacting with organic
matter. If chlorine oxidation is applied to groundwater, the risk of toxic
by-products will be less than that of surface water because groundwater
has a less organic matter due to the natural filtering by soil. Typical
doses of chlorine for oxidation of arsenite are 0.8 to 2.0 mg/L. In Europe,
ozone is widely used as an oxidant in treatment process, though it has
not been used in developing nations. A report showed that in water with
2 mg/L of ozone the half-life of arsenite is approximately 4 minutes (Kim

and Nriagu, 2000). Potassium permanganate which is widely available in

14

developing nations is also effective to oxidize arsenite.

There are several other options to increase the oxidation rate of arsenite
without adding oxidants. A remarkable increase in the oxidation rate of
arsenite was observed on exposure to sunlight (Cullen and Reimer,
1989). This is because ultraviolet radiation can catalyze the oxidation of
arsenic in the presence of oxygen. It is also believed that increasing
temperature or freezing solution helps accelerate the oxidation rate of

arsenite (Johnston et a1, 2001).

1.6.6. Removal Technology in Bangladesh

Safi Filter :

A filter called Safi filter has been developed in Bangladesh by Professor
Safiullah, Jahangirnagar University. The material of the Safi filter
consists of a chemical mixture of laterite soil, ferric oxide, manganese
dioxide, aluminum hydroxide and mezo-porous silica. These materials
remove arsenic effectively by adsorption. The Safi filter can treat
approximately 40 liters per day, which is more than sufficient for the
demand of a family of six. The manufacturers indicate that the materials
need to be replaced after two years of continuous use. The materials cost
about $4 (200 Taka) per unit, and the total cost of this filter is about $
20 (900 Taka) per unit (BRAC, 1999). The main problem of this system is

clogging that results in low flow rate of the system (J akariya, 2000).

15

Kolshi filter (Three-Pitcher Filter) :

The Kolshi filter is made from locally available materials and has been
used for many years in Bangladesh. The Kolshi filter consists of three 18-
liter clay pitchers called “kolshis”. A top pitcher is filled with 2 kg of
coarse sand, and a second pitcher is filled with 1 kg of charcoal and 2 kg
of fine sand, and a third pitcher serves as storage. Small holes are made
at the bottom of the first two pitchers so that water can pass through. To
adsorb more arsenic, the system has been modified by adding 3 kg of
iron filing to the top pitcher, which provides an additional source of iron
oxide (Rasul et a1, 1999). The total cost of this system per unit is $ 5.0
(250 Taka).

Laboratory tests showed that the Kolshi filter removed arsenic effectively
from well water. A wide range of arsenic concentration (from 80 to 1000
ppb) was tested using this system, and the effluent arsenic concentration
of 5 to 30 ppb was measured, which is less than the arsenic standard in
Bangladesh. Field tests of two hundred units confirmed that 90% of the
filters produced water with no arsenic after one week of operation, and
7% produced water in which arsenic was slightly detected. The filters
were still working efficiently after four months of operation. However, the
field tests also showed that some slight contamination with bacteria
occurred before filtration, and bacteria counts increased dramatically
during filtration and storage. Near two thirds of treated water samples

showed counts higher than 100 TC / 100 mL. The media of filter needs to

16

be sterilized before filter construction to avoid the risk of bacterial
contamination. It is also a problem that this system produces waste
material with high concentration of arsenic because no regeneration is

available (Johnston et al, 2001) (Jakariya, 2000).

1.7. Theory of Carbon Adsorption

1.7. 1. Mechanisms of Adsorption

Adsorption is the attachment of a substance onto the surface of
adsorbent, whereas absorption is the penetration of the attached
substance into solid. Since both phenomena often happen at the same
time, the term sorption is sometimes used. Although both adsorption and
absorption onto activated carbon occur in removal process, it is usually
referred to as adsorption.

Adsorption occurs when the attraction between solute and adsorbent is
greater than the attraction between solute and solvent. There are two
main attraction forces between phases: van der Waals forces and
electronic forces. The van der Waals forces are relatively weak
interactions, whereas the electronic forces are relatively strong
interactions which are sometimes irreversible.

The rate of adsorption is related to various mass transfer mechanisms.
First, an adsorbate molecule is transferred from the bulk liquid to the
boundary layer surrounding the adsorbent solid. Second, the adsorbate
diffuses through the boundary layer to the surface of the adsorbent solid.

17

Third, the adsorbate diffuses through the pores into internal adsorption
sites of the adsorbent solid. Usually the rate of adsorption is limited by

boundary diffusion and pore diffusion (Voice, 1997) (Reynolds, 1982).

1.7.2. Characteristics of Activated Carbon

Activated carbon is used to remove wide a variety of contaminants in
both liquid and gaseous phases. It is typically used to remove organic
compounds and some inorganic compounds with the characteristics of
non—polarity and low solubility.

Activated carbon is made from a number of carbonaceous materials such
as coal, wood and coconut shells. The manufacturing process consists of
carbonization of the materials followed by the activation using hot air or
steam, which produces an extensive network of internal pores. The
product that results from this process is highly porous, consisting of
macropores and micropores. The size distribution is largely a function of

the manufacturing process. The total internal surface area is typically in

the range of 500 to 2000 m2 / g. The high adsorptive capacity of activated

carbon is a result of high surface area of the finished product. The total
surface area may provide an approximation of adsorptive capacity.
However, the actual capacity for a specific compound is more closely
related to the available surface area for the specific compound. A number
of standardized tests using adsorbates of known molecular size has been

performed. The common test adsorbate is iodine which correlates with a

18

pore size greater than 10 A. Activated carbon is available at low cost and

can normally be regenerated (Voice, 1989) (Reynolds, 1982).

1.7.3. Isotherm

An isotherm is used to determine the adsorptive capacity of an
adsorbent. The isotherm expresses the relation between the equilibrium
concentration in solution and the amount of adsorbate adsorbed per unit
mass of activated carbon. In practice, an isotherm is measured
experimentally by equilibrating a known mass of adsorbent with a known
initial concentration of adsorbate, and plotting the resultant
concentration. However, an isotherm is quite dependent on the presence
of other adsorbates and temperature. Therefore, it is always necessary to

perform the isotherm using raw contaminated waste.

19

2. Methodology
2. 1. Material

2. 1. 1. Activated Carbon

Four types of activated carbon from Calgon Carbon were tested.

CPG-LF:

CPG-LF is an acid washed granular activated carbon with a low acid
soluble iron content. It is designed for the purification and decolorization
of many aqueous and organic liquids. The particle size of 12x40 mesh
has been selected to give a high rate of adsorption and low resistance
flow. CPG-LF is made from selected grades of bituminous coal combined
with suitable binders to give superior hardness and long life. It is
produced under controlled conditions by high temperature steam
activation, so this carbon provides high surface area. Iodine number: 950

mg of Iodine/g. Cost: $1.69 /1b = $3.72/kg.

F 400:

F 400 is a granular activated carbon designed for the removal of taste
and odor compounds and dissolved organic compounds from liquid
phase. This activated carbon is manufactured from select grades of
bituminous coal. Activation is carefully controlled to produce high
internal surface area for effective adsorption. The particle size of 8x30

mesh has been selected. F 400 is used to treat surface and ground water

20

sources for the production of drinking water. Iodine Number: 1000 mg of

Iodine/g. Cost: $1.21/lb = $2.66/kg.

OLC:
OLC is a coconut base granular activated carbon. The particle size of
12x40 mesh has been selected for this product. Iodine Number: 1100 mg

of Iodine/g. Cost: $1.74/1b = $3.83/kg.

DSR-C:
DSR-C is a granular re-activated carbon designed for the removal of
organic contaminants from industrial wastewater. This carbon is

manufactured by the reactivation of bituminous coal. Iodine Number:

800 mg of iodine/ g. Cost: $0.89 / 1b = $1.96/ kg.

2.1.2. Treatment by Ferrous Sulfite

Huang and Vane reported that treated activated carbon by ferrous salt
enhances the removal efficiency for As(V) because of its high affinity for
arsenate (Huang and Vane, 1989). In this study four activated carbons
were treated with 0.2 M of ferrous sulfate (Sigma Chemicals, St Louis).
FOur glass bottles with the capacity of 1 L were filled with 500 mL of DI
water. The water was purged with pure nitrogen gas for 30 minutes to
remove oxygen, which prevents the oxidation of ferrous sulfate to ferric

sulfate. Then, 27.8 g of ferrous sulfate was weighed and added into each

21

bottle during purging so that the chemical can be mixed well. Five grams

of each of the four types of activated carbon were weighed and added into

each labeled glass bottle. The four glass bottles were put on the shaker

horizontally and shaken at 100 rpm for 20 hours. The pH of solution was

measured before and after shaking. After 20 hours of shaking, the

activated carbons were taken out of each glass bottle and put into four

labeled glass vials, respectively. The activated carbons were put in an

oven and dried at 120°C overnight. The weights of the final products

were measured.

 

 

 

 

 

 

Mass of AC. Mass of Ferrous pH pH Mass of AC.
Before (g) Sulfate (g) Before After After Dry (g)
CPG—LF 5.0072 27.821 2 .94 2.57 5.5805
F400 5.0025 27.806 2.92 2.41 5.7018
OLC 5.0072 27.801 2.89 2.40 5.4560
DSR-C 5.0040 27.802 2.89 2.43 5.2170

 

 

 

 

 

 

Table 1: Weight of Adsorbents and pH Change During Treatment

Table 1 shows the difference in weights and pH before and after

treatment with ferrous sulfate. This gain of weight indicates that ferrous

sulfate was adsorbed onto the activated carbons.

22

 

2.2. Chemicals
2.2.1. Arsenate
Sodium arsenate (NagHAsO4-7H20, FW=312.0) was purchased from

Sigma Chemicals, St. Louis. 1000 mg/ L (1000 ppm) of As(V) standard

solution was prepared by dissolving 0.4164 g of NazHAsO4-7H20 in a

100 ml volumetric flask.

2.2.2. Arsenite

Sodium arsenite (NaAst, FW=129.9) was purchased from Sigma
Chemicals, St. Louis. 1000 mg/ L (1000 ppm) of As(III) standard solution
was prepared by dissolving 0.17343 g of NaAst in a 100 m1 volumetric

flask.

The following reaction occurs in solution. (Korte, 1991)

NaAs02 :> Na+ + A302‘
H+ + AsOg' :> HA802 (below pH 9.2, reducing condition)

H20 + HA302 :> H3AsO3 (below pH 9.2, reducing condition)

2.2.3. Reducing Agent

Usually ascorbic acid is used to prevent oxidation of As(III) to As(V)

according to some researchers. However, adding ascorbic acid results in

23

a low pH. Since the objective of this study is to treat the groundwater in
Bangladesh, which has the pH range around neutral, the low pH is not
preferable for this study. However, it is difficult to keep arsenic in a
reduced form at near neutral pH range because the ORP (Oxidation-
Redox Potential, unit of mV) of the solvent needs to be close to 0 mV (See
Fig.1). Through experimentation, it was found that DI water has the ORP
value of 150 to 190 mV. Purging with nitrogen can remove dissolved
oxygen down to 0.5-0.8 mg/ L. However, nitrogen bobbling does not
change ORP significantly (see appendix). Therefore, to obtain a lower
ORP, a reducing agent must be added into the DI water.

Three types of reducing agents were used in this study, which are
sodium sulfite (NazSOa), sodium meta-bisulfite (Na28205) and sodium
thiosulfate (Na28203). A detailed chemical description is included in the

appendix.

Characteristics of Reducing Agent in Solution:

Na2S03 (Sodium Sulfite):

 

 

 

 

 

 

DO (mgg/L) pH ORP (mV)
Initial Condition of DI Water 6.3 7.3 175
3 minutes after adding 0.25 9.38 5
1 hour after addLnQ 0.11 9.28 5
2 hour after adding 0.07 9.16 10

 

 

 

 

 

Table 2. Characteristics of Sodium Sulfite in Solution

Result by Adding 0.5 g of Na2803 into 70 ml DI Water

24

As observed in Table 2, the pH goes up drastically when sodium sulfite is
added into water. The ORP decreases significantly upon addition of
sodium sulfite. The reaction with dissolved oxygen occurs very quickly (in

a few minutes). The pH goes down slightly with time.

Na2$205 (Sodium Meta-Bisulfite):

 

 

 

 

 

 

DO (mg/L) pH ORP (mV)
Initial Condition of DI Water 6.4 7.2 175
30 minutes after adding 0.54 4.23 105
1 day after adding (without air) 0.23 4.22 100
2days after adding (without air) 0.14 4.19 100

 

 

 

 

 

Table 3. Characteristics of Sodium Meta-Bisulfite in Solution

Result by Adding 0.5 g of Na28205 into 70 ml of DI Water

Table 3 shows that the pH goes down drastically when sodium meta-
bisulfite is added into water. The ORP decreases remarkably after the
addition of sodium meta-bisulfite. The reaction with dissolved oxygen is

as fast as sodium sulfite.

25

Na2S203 (Sodium Thio-Bisulfite):

 

 

 

 

 

 

 

 

 

 

 

DO (mg/L) pH ORP (mV)
Initial Condition of DI Water 6.4 7.3 175
3 minutes after adding 6.45 7.42 85
30 minutes after adding 6.34 7.40 85
1 day after adding (without air) 2.23 7.25 80
2 days after adding (without air) 1.81 7.17 80

 

Table 4. Characteristics of Sodium Thiosulfate in Solution

Result by Adding 0.5 g of Na28203 into 70 ml of DI Water

Table 4 indicates that sodium thiosulfate does not react with dissolved
oxygen as quickly as sodium sulfite and sodium meta-bisulfite. The ORP

is significantly decreased, but the pH does not change significantly.

2.3. Instruments

2.3.1. GFAAS

Graphite Furnace Atomic Adsorption Spectrometer (GFAAS, PerkinElmer
AA Analyst 800 model) equipped with an electrodeless discharge lamp
was used for the analysis of arsenic. The GFAAS can measure most
metallic elements with 20 to 1000 times better sensitivities than using
conventional flame techniques. This increased sensitivity results from an
increase in atom density within a furnace as compared to flame atomic
adsorption. Many elements can be detected at concentration as low as

1.0 pg/ L. The GFAAS is based on the same principle as flame atomic

26

adsorption. First, an element is atomized in a graphite tube at high
temperature. The light beam is directed through atomization area to a
slit, and a detector measures the amount of light absorbed by the
atomized element as a signal. Each metal has its own characteristic
adsorption wavelength. The selected wavelength for arsenic is 193.7 nm
with a spectral slit width of 0.7 nm. The peak area was measured as a
signal.

There are three or more stages in an atomization process. First stage
dries sample at low temperature. The second stage destroys organic
compounds and volatilizes other matrix components at an intermediate
temperature. Finally, high heat is applied to the tube and atomizes the
element being measured. Additional stage is added to clean and cool the

tube between samples.

 

 

 

 

 

 

 

 

 

 

 

 

 

Step # Temp °C Ramsiglme Hol(tsle’l(;11me Inéegzival Reading
1 1 10 1 30 250 /
2 130 15 30 250 /
3 1200 10 20 250 /
4 2000 0 5 0 X
5 2450 1 3 250 /

 

Table 5. Furnace Program for Arsenic Analysis

Matrix modification can be useful in minimizing interference and
background signal. Chemical modifiers generally modify relative volatility

of matrix and metal. Some modifiers enhance matrix removal, which

27

results in the isolation of the target metal. For arsenic measurement,

matrix modifier was prepared by dissolving 0.004 g of PdC12 in 0.1 ml of

concentrated HCl. This solution was mixed with a solution prepared by
dissolving 0.020 g of ammonium paratungstate (NH4)10H2W12024 and

0.400 g of citric acid in 5-10 ml of DI water. The resulting solution was
brought up to 20 ml with DI water.

Zeeman background correction, which provides accurate background
correction, was used for this study. This correction is based on the
principle that magnetic field splits the spectral line into two light beams.
One is called 1: component and the other called 6 component. These two
light beams have exactly the same wavelength and differ only in the
plane of polarization. The 1r line will be adsorbed by both the atoms of the
element of interest and by the background caused by broadband
adsorption and light scattering of the sample matrix. The 0 line will be
adsorbed only by the background. The subtraction of the signal of the 0
line from the it line will give a signal of the element of interest (Standard

Method, 1995).

2.3.2. ORP, pH and DO Meter

The ORP, pH and DO were determined with ORP meter (ORP FMl,
OAKTON), pH meter (D-12 Model, HORIBA) and DO meter (97-08-00

Model, ORION), respectively.

28

3. BATCH TEST
3.1. Arsenate Adsorption

3.1.1. As(V) Kinetics Studies

Sample Preparation :
Since the regulation for arsenic is 50 ppb, the initial concentration of 100

ppb As(V) was prepared for this test. A standard solution of 1000 ppm

was made by adding 0.4161 g of Na2HAsO4‘7H20 into 100 ml of DI

water. 1 ml of 1000 ppm standard solution was added into 100 ml of
volumetric flask to make 10 ppm standard solution. 100 ppb solution
was prepared by taking 1 m1 of 10 ppm solution and diluting it up to 100
m1. A prepared 100 ppb solution was poured into a 160 ml glass bottle
and 0.05 g of either plain or treated activated carbon was added to the
solution.

Glass bottles were used for the batch test in order to avoid the risk of
adsorption onto the container. Moreover, a 160 ml glass bottle was
chosen because it has the 60 ml of head space that can help the mixing
process. To observe the difference of adsorption kinetics, an optimum
amount of 0.05 g of activated carbon was selected. This is because pre-
experiments showed that there was no significant difference in
adsorption kinetics when the amount of 0.05 g, 0.1 g, or 0.2 g of
activated carbon were added into 100 ppb of As(V) solution, respectively

(see the appendix).

29

Sampling Method :

The prepared glass bottles were put on a shaker and mixed at 120 rpm
for 52 hours. Samples of 2 ml were taken after 2 hours, 6 hours, 16
hours, 28 hours and 52 hours of shaking from each respective bottle
during shaking. The parameters of pH, DO, and ORP were measured

before and after shaking.

30

Result & Discussion :

As(v) Adsorption on 0.059 of CPG—LF As(v) Adsorption on 0059 of F400

 

 

 

a 3 7 I I ’ l

O. Q 1

9. i E i E _ , 3

c l c 7 i 7 7 if }

2 .9

ii i r r r m r E r r r

E l ‘c‘ *

8 4 4 4 4 4a 8 4 4 4 4 l

5 ‘ 5 ;

o n V M2 -1 A o - n 2 m .
44 4.. . 44 A :44 . 4 44 .
30 40 50 60 3O 40 50 60

Time (hour) Time (hour)
As(v) Adsorption on 0.059 of OLC As(v) Adsorption on 0.059 of DSR-C

Concentration (ppb)
Concentration (ppb)

 

 

 

0 10 20 30 40 50 60
Time (hour) Time (hour)

Figure 2. As (V) Adsorption on 0.05 g of Plain Activated Carbon,

pH = 7.2, ORP = 175 mV

CPG—LF, F400, and DSR-C showed good adsorption kinetics for arsenate
removal as demonstrated in figure 2. Equilibrium was achieved after 20
hours of shaking with concentration of 1 to 5 ppb. OCL removed less
arsenate than other three activated carbons. The adsorption kinetics of

treated activated carbons is shown in figure 3.

31

As(v) Adsorption on 0059 of treated CPG-LF As(v) Adsorption on 0.059 of treated F400

 

 

   

4 — 4 4 4 4 4 120 4 4 4 - — 4 4— 4 4
3 7 A ! l 3 i — W i i]
Q l o. I
3 , 1n_. 53:
.5 i :5 T _ l l
5 4 - t f - *
z E ,
8 — — 4—l 8 4 4 4 4.;
5 l g , 1
o 4 — 4 41‘ o I — 4 4 4- J.
¢ 6“ J l - ‘ ’7 4) A4. 1
30 40 50 60 0 10 20 30 40 50 60
Time (hour) Time (hour)
As(v) Adsorption on 0.059 of treated OLC As(v) Adsorption on 0.059 09 treated DSR-C
120 --44 44-4 4 44444 444 4 120 -4 — 444- —4 4 ~—
" v A“) 3 ’ 7 F ‘;
‘8 l g
.5 ‘ .5 ;
‘5 < 5 4 4 4 -(
.c. E .
g 2.- -1 g -- _ _ _ .
8 _._J 8

 

   

0 10 20 30 40 50 60 30 40 50 60

Time (hour) ‘fime (hour)

Figure 3. As(V) Adsorption on 0.05 g of Treated Activated Carbon

Treated activated carbon also showed good removal efficiency for
arsenate. The adsorption kinetics of OLC was improved by treating with

ferrous sulfate. Figure 4 shows the comparison of adsorption kinetics on

plain and treated activated carbons.

32

Concentration (ppb)

 

 

 

 

Time (hours)

Figure 4. Comparison of Plain and Treated Activated Carbon,

Represented by 0.05 g of Plain and Treated F400

After 2 hours of shaking, the concentration of arsenate in solution with
treated F400 is slightly lower than that with plain F400. This indicates
that the affinity of arsenate for the adsorbent was slightly increased by
treating with ferrous sulfate. However, the equilibrium concentration of
arsenate in solution with treated adsorbent is slightly higher than that
with plain adsorbent. This means the adsorptive capacity of treated
activated carbon is less than that of plain activated carbon. It is a
reasonable hypothesis that the adsorption sites had been already

occupied by ferrous sulfate, so that the adsorptive capacity available for
arsenic decreased.

33

Eh-pH Change during Batch Test

1200 ::>‘\1

l I
H3 As 040 \

800 _

 

 

 

 

 

 

\ HASO“- \
400 - 2_
A H2ASO4
> 7e
E o 3-
v 0 H3ASO3 A804
.:
LLI
HAsSzo 3_
400 -' A“”~
‘ ASSZ- H2ASO3 ' \‘
HA8032- .
)—
-800 l l J l

 

O 2 4 6 pHS 10 12 14

Figure 5. Eh-pH Change during Batch Test, O—> indicates the

Eh- pH Change of Solution Contains Treated DSR-C.

Figure 5 Shows the Eh-pH change of solution during the batch test. The
solution contains plain activated carbon did not show any significant
change in ORP and pH. However, the ORP and pH were significantly

changed in the solution with treated activated carbon. The ORP

34

increased from 170 mV to about 450 mV and pH decreased from 7.2 to
about 3.9. It may be because adsorbed ferrous sulfate was dissociated

into solution.

3.1.2. Adsorption Kinetics Changes

Comparing the kinetics curve of Aug. 2000 with that of Feb. 2001, there

is a difference between kinetics curves.

WAdsupfiononDSR-Cplain AsMAdsorptiononF400Treated
1204-44 «44—4 4——44 44 4 120, 444- 44 44 44 4

§

 

 

8

 

Concentration (ppb)
8 8 8

N
O

 

O

 

Time (hours)

Figure 6. Change of Adsorptive Capacity on Plain DSR-C and Treated

F400

As you can see in figure 6, adsorptive capacity on plain DSR-C worsened
as time went by (Aug’ 00 to Feb’ 01). This could be because plain DSR-C
was oxidized by air or reacted with moisture in the air while it was kept

in the container. The same thing happened in treated activated carbons.

35

Both treated F400 and treated CPG-LF showed less adsorptive capacity

in Feb. 2001 than in Aug. 2000.

3.1.3. As(V) Isotherm

Sample Preparation

The amount of activated carbon was fixed at 0.05 g. The optimum initial
concentration of arsenate was chosen, so that the equilibrium
concentration can be achieved in the range of 0 to 100 ppb. Judging from
the kinetics curve, solutions with the initial concentrations of 1000 ppb,
800 ppb, 600 ppb, 400 ppb, and 200 ppb were prepared for plain CPG-
LF and plain F400. The solutions with initial concentrations of 600 ppb,
500 ppb, 400 ppb, 300 ppb, and 200 ppb were chosen for plain OLC
since it showed less adsorption kinetics. The isotherm for plain DSR-C
was not performed because its adsorptive kinetics had already changed.
The initial concentration of 500 ppb, 400 ppb, 300 ppb, and 200 ppb
were prepared for treated OLC and treated DSR. The isotherm for treated

F400 and treated CPG was not performed for the same reason.

36

Results & Discussion

   

0.0025 y 4 4 4 4 4 44 4 4 J ‘ 1 1
4 —- I
I I ICPG-LFI
; . , . - E ; -. c.)
0.002 N ‘ F400 l
2,3 0.0015 ( - I; .013
B: i
3 I I
a: 0.001 *I c- _‘
U‘ I 1
0.0005 }- (I
I
o! 4. 4 I4 I_ . -4
0 20 40 60 80 100
Ce (ug/L)

Figure 7. Isotherm of Plain Activated Carbon for As(V) at 23 °C

0.0012 . 4 4 4 --V-, ,
0.001 1 44 4. ._ __1

0.0003 1 4 4

 

0.0006- 4 44

0-0004 I . * ‘ Abr-(AOLC-Treated I“

I | ?
0'0002 W ' W W 7’ l- DSR-C-Treatedl“ I

 

qe (us/us)

 

o I . 4 44—4- " ' " I
0 10 20 30 40 50
Ce (ug/L)

Figure 8. Isotherm of Treated Activated Carbon for As(V) at 23 °C

37

0.0025 4 —4 44-4 44 44 4 4 4 4 4 4 4 4 4-4-

I
0.002 .

l I ;
§ 00015] W -- f e ~I IOLCPIain . - I
B: I W W W W ’ 4
a
g 0.001
0.0005 ‘

 

Ce (ug/L)

Figure 9. Comparison of Plain OLC with Treated OLC

Plain CPG-LF showed the best isotherm' curve as shown in figure 7.
From figure 7, it is found that 1.0 g of plain CPG-LF can adsorb 2.0 mg
of As(V) at equilibrium concentration of 20 ppb. It means that 1.0 g of
plain CPG—LF can treat 20 to 25 L of the contaminated water with
influent concentration of 100 ppb until the effluent concentration is over
20 ppb, if sufficient contact time is given. In the same manner, 1.0 g of
plain F400 can treat 11 to 13 L of wastewater that contains 100 ppb of
influent concentration until the effluent concentration reaches 20 ppb.

From the laboratory test, it was found that 1 g of CPG-LF occupied 2.275

38

cm3. Therefore, if 1 g of CPG-LF is put in the column, it can treat 8,790
to 10,990 bed volumes until the effluent concentration reaches 20 ppb.

From the figure 9, it is found that treated OLC has an improved isotherm
curve when compared with that of plain OLC, but the difference between

these two is not significant.

3.2. Arsenite Adsorption

3.2.1. As(III) Kinetics Studies

Several experiments were performed to study the adsorption kinetics of
arsenite. Each experiment is discussed in the following stage in detail. In
each experiment, DI water was pretreated by purging with nitrogen
and/ or adding reducing agents to keep arsenite in its reduced form at
near pH range. The plastic bottle with the capacity of four liters was used
for pretreatment. The ORP, DO and pH, which determine whether the
arsenite is kept in its reduced form, were carefully monitored and
adjusted in this process.

After the optimum surrounding condition for arsenite was achieved by
pretreatment, the standard solution of 1000 ppm of As(III) was prepared
by dissociating 0.17343 g of NaA802 into 100 ml of pretreated DI water.
One ml of 1000 ppm standard solution was taken and added into 100 ml
of volumetric flask to make 10 ppm standard solution. A solution of 100
ppb was prepared by taking 1 m1 of 10 ppm solution and diluting it up to
100 ml. Prepared 100 ppb solution was poured into a 160 m1 glass bottle

39

and 0.05 g of either plain or treated activated carbon was added into
solution. The prepared glass bottles were mixed well with a shaker for
about 50 hours. Four samples were taken from each bottle during
shaking. Blanks were prepared to see the variation of pH, DO and ORP

during and after shaking.

a). Conditions Under Eh 40 mV and pH 6.5

Sample Preparation:

Four liters of DI water in a plastic bottle was purged with nitrogen gas for
two hours. After purging, 3.5 g of sodium meta—bisulfite (Na2S205) and
1.5 g of sodium sulfite (Na2803) were added to the purged DI water. The

change in pH, DO and ORP during this process are summarized in the

 

 

 

 

 

 

Table 6.

DO (mg/L) pH ORP (mV)
Initial Condition of water 2.88 7.24 185
After purging with N2 0.81 7.44 185
After addition of Reducer 0.1 6.52 40
Before Shaking 0.11 6.48 45

 

 

 

 

 

Table 6. The change of DO, pH and ORP by Pre-Treatment

By this pre-treatment, pH of 6.5 and ORP of 45 mV were achieved, at
which the arsenite can be kept in reducing form (see Figure 14).
The standard solution of 1000 ppm of arsenite was prepared by

dissociating 0.17 343 g of NaAs02 into 100 m1 of the pretreated DI water.

40

One ml of 1000 ppm standard solution was taken and added to 100 m1 of
volumetric flask to make 10 ppm standard solution. A solution of 100
ppb was prepared by taking 1 m1 of 10 ppm solution and diluting it up to
100 m1. Prepared 100ppb of As(III) solution was poured into a 160 m1
glass bottle. This preparation was done under aerobic conditions, so the
exposure of the samples to the air was expected.

Then, 0.05 g of plain CPG-LF, F400 and OLC were added into three glass
bottles filled with pre-treated solutions, respectively. Four blanks, which
have 100 ppb of arsenite and no activated carbons, were prepared. Also
four CPG-LF blanks, which have 100 ppb of arsenite and 0.05 g of CPG-
LF were prepared. Treated activated carbons were not performed because

those adsorbents change the condition of pH and ORP in solution.

Sampling Method:

The prepared samples and blanks, a total of 11 bottles, were shaken at
120 rpm for 50 hours. Samples of 2 ml were taken from each bottle after
2 hours, 6 hours, 18 hours and 50 hours of shaking, respectively. The
pH, DO and ORP of blank and CPG-LF blank solutions were measured

after 2 hours, 6 hours, 18 hours and 50 hours of shaking, respectively.

41

Result & Discussion

 

 

 

 

 

 

120 44 4 4 4 4 4 4 a - -
3 4 4 -
Q.
C
.9 H
g 4 - u_# -_
§ +CPG Plain I
C l } .
o l I +F400 Plain 1
0 20 4'4 4 44 4 4-]

 

 

" I F

30 40 50
Time (Hour)

. +OLC Plain ?

i

60

Figure 10. As(III) Adsorption on 0.05 g of Plain Activated Carbons

 

 

 

 

 

7 4 4 44
.. 44-444- .44,
6. 4—e i +Blankai
51 4 4 444 l+CPGpH ’
4
:5 .
3 ”I
2
1 p ...... it B_V,
0 i _‘ W T 'T W 7’ _ ___W
0 10 20 30 40 50
Time (Hour)

 

Figure 11. The pH Change of Blank and CPG-LF Blank During Shaking

42

Figure 11 shows the change of pH in the blank solution and in the
solution which contains 0.05 g of CPG-LF. The pH of the blank solution
decreased from 6.5 to 3.5 during shaking. This pH reduction was
attributed to the reaction SO31” + 02 :> SO42“. The pH of the solution,
which contains 0.05 g of CPG-LF, showed a lower pH (pH 2.5) than that
of the blank solution. It is hypothesized that the sodium ion (Nat) was
adsorbed onto CPG-LF, or the acid that is used in the manufacturing
process of CPG-LF was leached into solution. It is also one possible
hypothesis that the oxygen which are contained in activated carbons are
dissolved into solutions and lowered pH by accelerating the reaction of

2S032‘ + 202 :> 2SO42'.

From Fig.10, it was found that arsenite, which was present in the
solution as H3As03 in this pH range (6.5 to 2.5), was not adsorbed
effectively onto plain activated carbons. The concentration of arsenite in
solution became stable at about 80 ppb after 20 hours of shaking. It
means that 80% of arsenite was not adsorbed and remained in the

solution.

43

120: 44 44 4 444_ _. .__ __
100,
80

60.

Cone. (ppb)
PH

401

mwhmmw

20 4.

 

Time (hour)

Figure 12. The pH Effect on Arsenite Adsorption Kinetics, 0.05 g of CPG

Figure 12 shows the pH effect on arsenite adsorption kinetics. First 2
hours, arsenite was slightly adsorbed onto the activated carbon. This
small amount of adsorption can be explained by referring to Eh-pH
diagram (see Fig.14). Under the initial condition of the solution (pH 6.5,
ORP 40), a small proportion of arsenic would exist as arsenate since the
condition of the solution is close to the boundary of arsenite and
arsenate. It is considered that the arsenate in solution was adsorbed
onto activated carbon in the first two hours. The pH of the solution
decreased down to 2.5 with time, and the concentration of arsenite in

solution was maintained near 80 ppb though it showed slight desorption.

44

There was no significant difference in adsorption kinetics in the pH range
of 6.5 to 2.5. This is because in this pH range (6.5-2.5) arsenite is stable

as H3ASOs.

 

 

 

 

Time (hour)

Figure 13. The Relation of pH and ORP

Figure 13 shows that the ORP value increased as the pH decreased. As
mentioned before, the conversion of 8032' to SO42- during shaking caused
the pH to decrease. The increase of ORP value was also attributed to this
conversion, especially the increase of sulfate and hydrogen ion

concentration.

45

 

 

 

 

1200 _ I l I I 1 I
H3 As 040 \
800 — . \
\ HASO‘; \
400 "
HzAsof‘
’9
g \ A8043-
_= 0 H3AsO3°
LIJ
HASSzo A8033-
-400 — - (
ASSZ- I H2ASO3 ° “
HAso32"
'800 I I I I I

 

 

 

Figure 14. The Shift of pH and ORP during Shaking

The arrow in Fig.14 shows the change of pH and ORP in the solution
during shaking. The arrow is directed toward the center of the domain.
The more center the condition, the more predominant the species would
be. It means that arsenite was the strong predominant species and stable

in this solution at the end of shaking.

46

 

 

 

0.7 T +Blankoo i
g, 0-6 - ~ + + ~ 1 - +CPG-LF ‘ , i
g 0.5 .2 l BlankDO <4
0 0.4 .1- e - _ -

D 4v

0.3 (L r e 1 I
0.2 + - a .
0.1 . *2 x i .l
0 + $ 12 w j
o 10 20 30 4o 50 60

TIme (hour)

Figure 15. The Change of Dissolved Oxygen in the Solution

From Fig. 15, it was found that DO had increased from 0.1 to 0.4 mg/ L in
the blank solution during shaking. The increase of the dissolved oxygen
was derived from the air in the headspace of the glass bottle. D0 in the
solution that contains 0.05 g CPG-LF increased from 0.1 to 0.8 mg/ L.
This difference indicates that additional oxygen was introduced in the
solution which contains CPG-LF. It is considered that activated carbons
usually contain air and the air in activated carbon dissolved into
solution. Indeed, air bubbles could be seen when the activated carbon
was added to the solution. Dissolved oxygen must be consumed when

the reaction S032' + 02 :> 8042‘ proceeds.

47

b). Conditions Under Eh 90mV and pH 5.0

In the previous experiment, the pH and ORP of solution changed
significantly due to the reaction of the added reducing agent with oxygen.
To keep the condition of the solution constant, another reducing agent

was added into solution.

Sample Preparation:

Four liters of DI water in a plastic bottle was purged with nitrogen gas for
two hours. After purging, 3.5 g of sodium meta-bisulfite (Na28205) and
1.5 g of sodium thiosulfate (Na28203) were added into the purged DI

water. The change of pH, DO and ORP in this process are summarized in

 

 

 

 

 

 

Table 7.

DO (mg/L) pH ORP (mV)
Initial Condition of water 2.51 7.17 170
After purging with N2 0.67 7.42 165
After addition of Reducer 0.34 5.01 90
Before Shaking 0.32 5.03 85

 

 

 

 

 

Table 7. The change of DO, pH and ORP by Pre-Treatment

By this pre-treatment, pH of 5.0 and ORP of 85 mV were achieved, at
which the arsenite can be kept in a reduced form.
The solution of 100 ppb of arsenite was prepared by using the pretreated

DI water in the same manner as in the previous experiment. Pre-treated

48

solution was put into glass bottles. Then, 0.05 g of CPG-LF, F400 and
OLC were added into three glass bottles, respectively. Four blanks, which
have 100 ppb of arsenite and no activated carbons, were prepared. Also
four CPG-LF blanks, which have 100 ppb of arsenite and 0.05 g of CPG-
LF were prepared. Treated activated carbons were not performed because

of the same reason mentioned earlier.

Sampling Method:

The prepared samples and blanks, the 11 bottles, were shaken at 120
rpm for 50 hours. Two ml of each sample was taken from each bottle
after 2 hours, 8 hours, 20 hours and 50 hours of shaking. The pH, DO
and ORP of blank and CPG-LF blank solutions were measured after 2
hours, 8 hours, 20 hours and 50 hours of shaking, respectively. The pH,
DO and ORP of the solutions which contain F400 and OLC were also

measured before and after shaking (not at intermediate times).

49

Result & Discussion :

 

 

 

 

120 r 1 f 4 Ak~~e~ ~r~ if f f e i

100 «w 1.1 ~_ A -__-a, * A a e,

3; \ i

3 80 am A '7 7 E a ,7, _ ,, .v- a 7 ,1

= i

.9 ~—_~.__--__- _- i
g 60 —' 7' T " +CPG Plain 1
. i

g 40 l —--— — — —“-————~ +F400 Plain ! .

0

20 : —4~- AM— --— -5#-f-OLCI:Iain J D

o i .. .a.--~_.____m--._. _..._ __i

o 10 20 30 4o 50 60

Time (Hour)

Figure 16. As(III) Adsorption on 0.05 g of Plain Activated Carbons

 

 

 

 

 

 

 

 

6 i ‘— *‘rr' ‘1

i

T i

a ‘

1 I i5 9.. {aiérfiH l i

5 } +CPG pH

0 5‘ L“ ifil_—m ‘7—_ “V '1‘ _—k7 ififiv—fw _T“_+ " ' _ " J

O 10 20 3O 40 50 60

Time (Hour)

Figure 17. The pH Change in Solution during Shaking

50

Figure 17 shows the change of the pH in the solution during shaking.
The pH of the blank solution was kept around 5.0. This is due to the
characteristics of sodium thiosulfate in solution. It takes time for sodium
thiosulfate to ionize in solution. It means that the reaction Na2S203 :>
2Na+ + S2031" is very slow. Shaking accelerates this reaction and the pH
of the solution increases providing sodium ions into solution. The
reactions 2S2032' + 302 2 48032' and 28032- + 202 :> 28042' are also
slow. However, the shaking also enhances the kinetics of these reactions
and the pH in the solution decreases. Since the reaction Na2S203 2 2Na+
+ 82032' limits the available S2032; shaking sodium thiosulfate itself
increases the pH in solution (see the appendix). The co-existing sodium
meta-bisulfite decreases pH due to the reaction 2S2052' + 302 :> 48042:
Both effects on pH were canceled out and the pH of the solution was kept
constant.

Unlike blank solution, the pH decreased remarkably in the solution
contains 0.05 g of CPG-LF. It is hypothesized that the sodium ion (Nat)
was adsorbed onto CPG-LF, or the acid that is used in the
manufacturing process of CPG-LF was leached into solution. However,
the other solutions containing 0.05 g of F400 and 0LC, which are not
treated by acid in the manufacturing process, showed almost the same
pH values after 50 hour of shaking (pH 2.48 and 2.67, respectively).
Therefore, the latter hypothesis is not convincible. Another hypothesis

suggests that the oxygen, which is contained in activated carbon, is

51

dissolved into solution and accelerated the reaction of 28032' + 202 :>
28042'.

Figure 16 showed almost the same result as the previous experiment.
Arsenite in solution was not removed effectively by the plain activated
carbons. The solution reached the equilibrium concentration of about 80

ppb after 10 hours of shaking.

c). 0.20 g of A.C. Under Eh 90 mV and pH 5.0

To observe the better adsorption kinetics of arsenite, 0.20 g (four times)
of plain activated carbons were added to the solution treated in the same

manner as experiment b). The following is the result.

120 Ti ,7- * * I * "*J’ ’ ” ' ”7

 

 

 

 

 

 

Concentration (ppb)
0)
o
i
i

 

|+ céGfiéifii i

40 ‘ “ ‘ -—- -—- ‘g+F400 Plaini -

20: . . -4. _ i+OLC P'a'll ,
!

05.12% — e— . ~* .____, m .

o 10 20 30 4o 50 60

Time (Hour)

Figure 18. As(III) Adsorption Kinetics on 0.20 g of Plain Activated

Carbons

52

 

 

main... 1

l +ch pH i i
o + . a. - i '. L; 9
0 10 20 30 40 50

Time (Hour)

Figure 19. The pH Change in Solution during Shaking

 

 

 

 

 

 

 

“8040 \
800 — ~ \
\ M04. \
400 - 2
H2A504 '
E
3.
V O Hm A504
.C
|.I.l
HAssz" }
A503
-400 - Ass; “#803: .
HAso32"
'800 I I I I I
0 2 4 6 8 10 12 14

[ii

Figure 20. The Shift of pH and ORP during Shaking

53

Figure 19 shows that the pH of the solution containing 0.20 g of CPG
decreased drastically in the first 5 hours compared to the solution
containing 0.05 g of CPG-LF. It is hypothesized that the mass of
activated carbon was four times larger than 0.05 g and approximately
four times the oxygen contained in activated carbon was dissolved into
solution. It accelerated the reaction of 28032' + 202 :> 28042“ and
lowered the pH. Adsorption of the sodium ion (Nat) onto CPG—LF is still a
possible hypothesis.

Figure 18 shows that the arsenite was not adsorbed effectively onto 0.20
g of plain activated carbons. There was no significant difference in
adsorption kinetics between 0.05 g and 0.20 g of activated carbon in
solution. It is considered that chemical attraction between arsenite and
activated carbon is weak because arsenite exists as a neutral form
H3As03 and stable in this pH range (see Figure 20). There is another
possibility that co—existing ions such as sulfate effect on the adsorption
of arsenite.

Since arsenite is not adsorbed onto both 0.05 g and 0.20 g of activated
carbon, it was impossible to obtain an isotherm curve, which is the
relation of mass of adsorbate per unit mass of adsorbent with the

equilibrium concentration.

54

d). No Reducer

To eliminate the effect of the reducing agent on adsorption kinetics of
arsenite, DI water was treated only by purging with nitrogen. Since the
oxidation rate of arsenite with dissolved oxygen is very slow, arsenite is
supposed to be kept in its reduced form for a certain period. The
oxidation rate is also dependent on the concentration of dissolved
oxygen, so removal of oxygen by the pre-treatment process will be

effective to keep arsenic in its reduced form.

Sample Preparation :

Four liters of DI water in a plastic bottle was purged with nitrogen gas for
two hours. The solution of 100 ppb of arsenite was prepared by using the
pretreated DI water in the same manner as in the previous experiment.
Pre-treated solution was put into glass bottles. Then, 0.05 g of plain
CPG-LF, plain F400, plain 0LC, treated DSR-C and treated OLC were
added to five glass bottles, respectively. This process was done under
aerobic condition, so the exposure of the samples to the air was
expected. The change of pH, D0 and ORP in the preparing process are

summarized in Table 8.

 

 

 

 

 

 

 

 

 

D0 (mg/L) pH ORP (mV)
Initial Condition of water 4.51 7.37 180
After purging with N2 0.78 7.42 175
Before Shaking 2.05 7.54 175

 

Table 8. The change of D0, pH and ORP by Pre-Treatment

55

Sampling Method:

The prepared samples were mixed with shaker at 120 rpm for 52 hours.
Two ml of each sample was taken from each bottle after 6 hours, 18
hours, 30 hours and 52 hours of shaking. The pH, D0 and ORP of each
solution were measured before and after shaking (not at intermediate

times) .

Results & Discussion :

 

 

 

 

 

 

 

+CPG-P , I
i +F400-P i T" ’ 7‘
3 i +£LC-P
0- .kfim -7 71
Q.
I"; a
.9.
E -_ ,_ a
E
cu
0
g 7 , ,_
0 H I
”A A- --i
, ‘ — i
40 50 60
Time (hour)

Figure 21. As(lll) Adsorption on 0.05 g of Plain Activated Carbons without

Reducing Agent, pH 7.5, 0RP=175

56

120 f~ ,2 , ff ,2 i ,2 7 1, ,2 MEI

_.__- __ W i

100 —-—-— --— -— — — — '1—0-0LC-Treated: —

 

 
     

i + DSR-Treated i _ _. l

80

Concentration (ppb)
0)
o

 

0 10 20 30 40 50 60
Time (hour)

Figure 22. As(III) Adsorption on 0.5 g of Treated Activated Carbons

without Reducing Agent

Figure 21 shows the adsorption kinetics of arsenite onto 0.05 g of plain
activated carbons. There are significant differences between adsorption
kinetics of the three plain activated carbons. 0LC plain did not adsorb
arsenite effectively, whereas F400 plain adsorbed arsenite by 70%. CPG-
LF plain appeared to be theibest adsorbent for arsenite. CPG-LF plain
removed 90% of arsenite in this experiment. These differences are
attributed to the oxidation ability of each activated carbon. The ability of
oxidation is dependent on the amount of oxygen contained by the
activated carbons. It is possible that CPG-LF contained a larger amount

of oxygen than F400 and 0LC, and dissolved oxygen released from

57

activated carbon could oxidize arsenite to arsenate, and formed arsenate
was effectively removed. There was no significant difference in the pH or

ORP between before and after shaking.

Figure 22 shows the adsorption kinetics of arsenite onto 0.05 g of treated
activated carbons. These kinetics are definitely different from the kinetics
of plain activated carbons. The arsenite in solution was gradually
removed by treated activated carbon, whereas most of arsenite was
removed by plain activated carbon in the first ten hours. This difference
is derived from the difference in the oxidation process between plain and

treated activated carbons. This is explained by referring to Table 9 and

 

 

 

 

 

 

 

Figure 23.

pH D0(mg/L) 0RP(mV)
CPG-P 7.26 4.28 180
F400-P 7.32 . 3.8 180
OLC-P 7.36 4.83 175
OLC-Treated 4.03 4.8 355
DSR-Treated 4.11 4.98 360

 

 

 

 

 

Table 9. The pH, ORP, and D0 of solutions after 52 Hours of Shaking

58

 

 

 

 

1200 I 1 I I I I
3.43040 \
\ HAso; \
400 "
\ H2A8042-
’9 o .
g H3ASO3 A8043-
; 0
LIJ
HASSZO A50 3_
-400 - - 3“
A882- H2A803 ' ‘1
2- .
— HASO3
'800 I I I I I

 

 

 

O 2 4 6 8 10 12 14

Figure 23. Shift of pH and ORP during Shaking

From Table 9 and Figure 23, it is observed that the pH of the solution
with treated activated carbon decreased to 4.1 and the ORP increased to
360 mV. It is presumed that the adsorbed ferrous sulfate was released
from the treated activated carbons into solution and changed the
condition of the solution. This change of the condition in solution could
make the oxidation rate slow compared to the solution with plain

activated carbons.

59

From the experiments a), b) and c), it is understood that arsenite itself
can not be removed effectively by activated carbons. However, it was
turned out that the removal of arsenite is possible using activated carbon
by oxidizing arsenite to arsenate. In experiments a), b) and c), the
released oxygen from activated carbons was consumed by reducing
agents, and arsenite itself was not oxidized. However, experiment (:1)
showed that arsenite can be oxidized by the oxygen released from

activated carbons when there is no reducing agent in the solution.

3.3. pH Effect on Adsorption

The pH effect on arsenate adsorption was observed using 0.05 g of plain
CPG-LF. The final volume of 100 ml of 100 ppb As(V) solutions were
prepared by adding either hydrochloric acid or sodium hydroxide into
solution to adjust the pH. Prepared solutions were poured into 160ml
glass bottles and 0.05 g of plain CPG-LF was added to each solution.
Prepared bottles were put on a shaker and mixed at 120 rpm for 40
hours until equilibrium condition was obtained. The equilibrium
concentration, pH and ORP were measured by using GFAAS, pH meter

and ORP meter, respectively. The following is the result.

60

120‘ 4 -2- *4

100 .

Concentration (ppb)

   
    

12

 

 

 

 

1200
400 2
H2ASO4 -
9 o
g H3ASO3 A3043-
.: 0 ‘>
LLI
HAsSzo A 0 3_
s 3
—400 — - .
AsSz- H2A503 -
HASO32-
'800 I I I I I

 

 

 

 

O 2 4 6 8

Figure 25. The Eh-pH Variation

61

400

~ 350

f- 300

250

.+Equmbrium U.
I Conc.
9+0RP

200
150

ORP (mV)

From the Figure 24, it is observed that arsenate is removed effectively in
the pH range of 5.0 to 8.5. Arsenate is not removed effectively in the pH
range either below 5.0 or above 8.5. At pH above 8.5, the concentration
of hydroxide ions (OH) is significantly greater than the arsenate
(HASO42') and are the predominant species in the solution. The 0H' ions
are competing with arsenate and prevent arsenate from attaching to the
surface of the adsorbent.

The Eh-pH diagram (Figure 25) shows that the arrow crossed over the
boundary between arsenate and arsenite at pH 4.0. It means that below
pH 4.0 arsenate existed as arsenite. At pH 4.0 (on the boundary), the
proportion of arsenate and arsenite was supposed to be 50 /50,
theoretically. Indeed, Figure 25 shows that 50% of arsenate was removed
at pH 4.0. It means that at pH 4.0 50% of arsenic existed as arsenate
and was removed. The reason why arsenate was not adsorbed at pH
below 5.0 is that the predominant species changed from arsenate to
arsenite below pH 4.0. The pH range of 3 to 5 is a transition range of this

conversion.

3.4. Regeneration

Solutions of 100 ml of 100 ppb As(V) were prepared and poured into 160
ml glass bottles (bottle A and B), and then 0.05 g of plain CPG-LF were
added to each bottle. The bottles were mixed using shaker at 120 rpm for

44 hour until the solution reached the equilibrium condition. Two ml

62

samples were taken from the bottle A and B after 44 hours of shaking
and the equilibrium concentration was measured. Then, 1 ml of 3 N
hydrochloric acid and 1 m1 of l N sodium hydroxide were added to
bottles A and B, respectively. Bottles A and B were mixed again for 24
hours to observe the regeneration kinetics. Samples were taken after 2

hours, 6 hours and 24 hours of shaking. The followings are the results.

 

pH ORP
Bottle A 1.74 400
Bottle B 12.01 40

 

 

 

 

 

 

 

Table 10. The Condition of Solution in the Regeneration Process

120 . 7+ 7 “-__—‘73 22,22,222,

 

 

 

I__ .__ __

+HC| (pH=1.74) i f .

Concentration (ppb)
1

 

|

 

 

_ -__fl T .__ ___._.__,,, I _ ._¥ . ._.,fi,#.

7‘ +Na0H(pH=12.01) ; '

0 5 10 15 20 25 30
Time (hour)

Figure 26. The Difference of Regeneration Kinetics between Acid and

Base

63

More than 90% of arsenate was regenerated when 1 m1 of 3 N HCl was
added (at pH 1.74, 0RP 400), whereas only 60% was regenerated when 1
ml of 1 N NaCl was added (at pH 12.01, 0RP 40). Under the conditions of
pH 1.74 and ORP 400, arsenic was regenerated as arsenite (neutral
H3A803). Under the high pH condition, arsenic was regenerated as
arsenate HASO42’ or As043'. Judging from the pH effect on adsorption, it
was expected that arsenite would be effectively regenerated under the
high pH condition. However, it was not. Unlike the adsorption process,
disorption process is not a competing process with hydroxide ions for
adsorption sites, but the ion-exchange process of arsenate with
hydroxide ions. It may be because the affinity of HASO42' or As043' on
activated carbon is high and irreversible. It is concluded that HCl is a

better regenerant for arsenate regeneration.

3.5. Summary of Batch Test

It was observed from the batch tests that arsenate was removed very
effectively by activated carbons, especially by either plain CPG-LF or
plain F400. Plain DSR-C also removed arsenate effectively, but the
capacity was changed when it was kept in a storage tank. Treated
activated carbons also showed effective adsorption kinetics, but the
adsorptive capacity seemed equal to or less than plain activated carbons.
It was also observed that treated activated carbons changed the

condition of the solution by releasing ferrous sulfate into solution.

64

The isotherm for arsenate showed that plain CPG-LF had the best
adsorptive capacity (twice as much as plain F400) among activated
carbons used in this experiment. There was no significant difference of
adsorptive capacity between plain and treated activated carbons.

The batch tests showed that arsenite was not removed as arsenite, but
removed as arsenate. Plain CPG-LF removed 90% of arsenite by oxidizing
aesenite to arsenate.

Arsenate adsorption onto activate carbon was observed to be pH
dependent. It was found that the optimum pH range for arsenate removal
is 5.0 to 8.5.

When 1 ml of 3 N hydrochloric acid was added to solution, about 95% of
the adsorbed arsenic was regenerated. However, it was observed that

sodium hydroxide was not effective to regenerate arsenic.

4. Column Test

Column tests were conducted to observe the adsorption kinetics of
arsenate and arsenite on activated carbon. In this experiment, a pump
(WATSON-MARLOW 502E) was used and the flow rate was adjusted to
5.43 ml/ min. A column (Flex-Column, KONTES) with the size of 1 cm
diameter and 10 cm length was used. The columns were prepared by

loading a certain volume of activated carbon.

65

4.1. Arsenate Adsorption through Column

10 ml of 10 ppm As(V) stock solution was taken and diluted up to 1 L to
make the influent concentration of 100 ppb As(V) solution. Prepared 100
ppb As(V) solution was poured into a 5 L glass bottle. The columns were
filled with 1 cm (0.785 cm3) of plain CPG-LF, plain F400, plain 0LC and
treated 0LC, respectively. The flow rate was adjusted to 5.43 cm3 / min
and calculated contact time was 8.67 sec. Seven to ten samples were
taken from the effluent during running and were measured. The

following is the result of adsorption kinetics of arsenate through the

 

 

 

 

 

 

 

 

  
 
 

 

 

 

column.
100 WWW _ ~ - , WWW ,
90 “f” ~ ~ l
80 W 1
a 70 / .__. __ _ < _
9’ l
o- 60 I
5 50 l
0 , W e.
8 4° 7 W" i+0LC Plain
3 i
g 30- + *~ ’*“‘“+F400 Plain l’
20 -— I-A—CPG Plain ‘
10 2+ W W W W+E9 “93339 I
0 WW W W. W W— W W . WW
0 5 10 15 2o

Treated Volume (Liter)

Figure 27. Adsorption Kinetics of Arsenate onto Activated Carbons
Through the Column, Flow Rate=5.43 cm3/ min, Bed
Volume=0.7 85 cm3, Contact Time= 8.67 sec, Influent
Conc.=100 ppb, pH = 7.2, ORP = 180 mV

66

Plain OLC did not remove arsenate effectively compared with the other
activated carbons. Plain F400 and plain CPG—LF showed almost the same
adsorption kinetics for arsenate removal with the initial effluent
concentration of 35 ppb and the equilibrium concentration of 40 ppb.
The higher equilibrium concentration compared to the batch test was
due to a higher flow rate and shorter contact time. Treated OLC showed
the lowest initial effluent concentration (24 ppb) among the activated
carbons used in this experiment. It means that the affinity of the treated
OLC to the arsenate is the strongest. However, it saturated sooner than
plain F400 and plain CPG-LF because of the less adsorptive capacity. In
total, plain F400 and plain CPG-LF appeared to be the best adsorbent for
arsenate removal. Further study with the lower flow rate and larger bed

volume is necessary to achieve longer contact time.

67

4.2. Effect of Bed Volume (Contact Time)

Two cm of plain CPG-LF (bed volume 1.57 cm3) was added to the column
to observe the effect of contact time on adsorption. The experiment was
carried out in the same way as the previous experiment. Both the
influent concentration and flow rate were maintained as before. The

following is the result.

30 . V ,, an
70
60 a
50
40 ,
30
20

 

  

_ __'__ ’ l

+ CPG 1cm (8.67 sec)

Effluent Conc. (ppb)

10 . ~ W
l + CPG 2cm (17.35 sec)

0 ? TW ,. - . i i 1
0 5 10 15 20 25 30

 

Treated Volume (Liter)

Figure 28. The Effect of Bed Volume (Contact Time) on Adsorption, Flow
Rate=5.43 cm3 / min, Influent Conc. = 100 ppb, pH = 7.2,
ORP = 180 mV
As the contact time increased, both the initial effluent concentration and
the effluent equilibrium concentration decreased. A larger bed volume

showed a longer running length before the break through occurred. It is

assumed that if enough amount of activated carbon is loaded in the

68

column and sufficient contact time is obtained, the equilibrium
concentration would approach the same concentration as in the batch

test.

4.3. Comparison of As(V) and As(III) Adsorption

It is necessary to maintain the arsenite in reduced form while the sample
runs through the column by adding reducing agent. However, high doses
of a reducing agent can effect on kinetics adsorption. So a small amount
of reducing agent was used for this experiment. The pH of the solution
was set near neutral condition by using reducing agent. The following is
the detail method about sample preparation.

Four liter of DI water in a plastic bottle was purged with nitrogen gas for
two hours. After purging, 1.0 g of sodium meta-bisulflte (Na28205) and
1.0 g of sodium sulfite (Na2803) were added to the purged DI water. The

change of pH, DO and ORP in this process are summarized in Table 11.

 

 

 

 

 

 

 

 

 

DO (mg/L) pH ORP (mV)
Initial Condition of water 2.89 7.22 185
After purging with N2 0.88 7.78 180
After addition of Reducer 0.10 7.02 60

 

Table 11. The change of DO, pH and ORP by Pre-Treatment

By this pre—treatment, the condition of pH 7.0 and ORP 60 mV were
achieved, at which the predominant arsenic species is arsenate not

arsenite. However, since the reaction of arsenite with oxygen is very slow

69

and the concentration of dissolved oxygen was kept at a very low level
(about 0.1 mg/ L), it was assumed that arsenite was maintained in
reduced form during the experiment.

The standard solution of 1000 ppm of As(IlI) was prepared by adding
0.17343 g of NaAsO2 to 100 ml of the pretreated DI water. One ml of
1000 ppm standard solution was taken and added into 100 ml of
volumetric flask to make 10 ppm standard solution. A solution of 100
ppb As(III) was prepared by taking 10 ml of 10 ppm solution and diluting
it up to 1000 ml. This preparation was done under aerobic condition, so
the exposure of the samples to the air was expected.

Prepared 100 ppb As(V) solution was poured into a 5 L glass bottle. The
column was filled with 1 cm (0.785 cm3) of plain CPG—LF. The flow rate
was adjusted to 5.43 cm3 / min and the calculated contact time was 8.67
sec. Nine samples were taken from the effluent during running and were
measured. In the middle of the experiment, the values of pH, DO and
ORP were measured to determine the condition of the influent solution.

The following is the result of adsorption kinetics of arsenate through the

 

 

 

 

 

column.
Run Length (Hour) DO (mg/L) pH ORP (mV)
0 0.10 7.02 60
1 0.10 6.94 65
2 0.13 6.86 70
4 0.14 6.84 70
6 0.17 6.65 75

 

 

 

 

 

 

 

Table 12. The change of DO, pH and ORP during Column Analysis.

70

 

 

   

 

 

 

 

E W W W W W W l

3 .

(.5 _ v .1

C

O —— —W --W

0 l

E __ f w- 2‘22 r ,__m

8 ““ ' ' "“1 .
E 30 ,.__ W WW ,mmc 1+As(lll) on CPGP-W~

20 “w— ' —‘** —_ +As(V) on CPG ( “I

10 I. n “V ,mir ”an” ..-_;;, , :::::‘_, A in

o W - e. W— W W 1

0 5 10 15 20

Treated Volume (Liter)

Figure 29. The Comparison of Adsorption Kinetics of As(III) with As(V)

Flow Rate=5.43 cm3 / min, Bed Volume=0.7 85 cm3,

Contact Time= 8.67 sec, lnfluent Conc.=lOO ppb,

D0= 0.10 mg/L, pH= 7.0 - 6.7, ORP= 60 - 75 mV for As(III)
Table 12 shows that the pH slightly decreases and ORP slightly increases
due to the reaction of 28032‘ + 202 :> 28042“. However, this means that
oxygen was consumed by the reducing agent not by the arsenite, so
arsenite should have been kept in the solution during experiment.
Figure 29 shows the adsorption kinetics of arsenate and arsenite on

plain CPG-LF. Arsenite was slightly removed, but the breakthrough

occurred much faster than that of arsenate.

71

4.4. Summary of Column Test

In the column tests, plain F400 and plain CPG-LF showed almost the
same adsorption kinetics. Treated OLC showed stronger affinity than
plain CPG-LF and plain F400 for arsenate, but it was saturated quicker
because of its small adsorptive capacity. The increase of the bed volume
contributed to lower the equilibrium concentration of effluent. Column
analysis also showed that arsenite was not removed effectively by plain

CPG-LF because of the shorter contact time than batch test.

5. Conclusion

All activated carbon used in this study removed arsenate effectively. The
isotherm curve showed that plain CPG-LF has the best adsorptive
capacity (twice as much as plain F400) for arsenate. In the column tests
plain F400 and plain CPG-LF showed almost the same adsorption
kinetics. Considering both adsorption kinetics and capacity, it was
concluded that plain CPG-LF is the best adsorbent for arsenate. Arsenite
was not effectively adsorbed on any activated carbons as arsenite at pH
below 6.5. However, it was observed that plain CPG-LF oxidized arsenite
to arsenate and removed the formed arsenate effectively.

The column tests showed that arsenite was not removed effectively by
plain CPG-LF. This indicates that the given contact time was not enough
to oxidize arsenite to arsenate. Since the contact time used in this
experiment was not sufficient (about 8.7 seconds), further

72

experimentation with longer contact times is necessary.

It was found that the removal efficiency of arsenic is dependent on the
arsenate proportion of the solution. Since very high removal efficiency
was achieved for arsenate with carbon adsorption under the optimal pH
condition (pH 5.0-8.5), the advantage of a pre-oxidation process are
apparent. Oxidants, such as chlorine and permanganate, which are
available in Bangladesh, will be helpful for high removal efficiency.

The isotherm showed that 1.0 kg of plain CPG-LF can treat 20,000 L of
contaminated water with influent concentration of 100 ppb until the
effluent concentration is over 20 ppb. Suppose a family of six uses 40
L/day, 1kg of CPG—LF can be used for 500 days. Since the cost of CPG-
LF is $3.72/kg, this system will be competitive with available techniques
in Bangladesh such as Safi Filter ($4.0) and Kolshi Filter ($5.0).
Moreover, an advantage of using activated carbon is that activated
carbons are easily regenerated and reused by adding hydrochloric acid.
In practical use, co-occuring ions have an effect on the adsorption of
arsenic on activated carbon. Laboratory tests using raw samples need to
be performed.

This study suggests that activated carbon is a cost competitive point-of-
use method. It is hoped that this study will provide the people in
Bangladesh with the economical and easy treatment process to eliminate

arsenic from their drinking water.

73

Appendix

74

A- 1. Chemical Description

Sodium Arsenate (NazHAsO4-7H20), Sigma Chemical
S—9663, 20k2507

Sodium meta—Arsenite (NaASOg), Sigma Chemical
S-7400, 79H6036

Ferrous Sulfate (FeSO4-7H20), Sigma Chemical
F-7002, 129H0898

Sodium Sulfite (Na2S03), Sigma Chemical
S-8018, 121H0368

Sodium meta-Bisulfite (Na2S2Os), Fisher
S-244, 852987

Sodium Thiosulfate (Na2S203), Mallinckrodt
8096, KLJS

75

A-2. Amount of Activated Carbon for Batch Test

The amount of 0.05 g, 0.1 g and 0.2 g of activated carbon were added

into 100 ml of 100 ppb As(V) solution, respectively.

 

120 E - -. ~ W .

1001+ W ~ W 4+0Eg—6fF4oF— (
E '+0.1OQofF4005 l
3 80 ’ *— F'*1+0.2090f F400! fl
*9: 60 ) - » — +
E
a)
o l
c .
O
0 i

* *1

 

 

 

 

Time (Hour)

Figure 30. Optimum amount of Activated Carbon for Batch Test

There was no significant difference in adsorption kinetics among these as
illustrated in Figure 30. To see the clear difference of adsorption kinetics,

the minimum amount of 0.05 g of activated carbon was selected for the

batch test.

76

A-3. Purging Effect on DI Water

DO, pH and ORP were measured during purging DI water with nitrogen

gas. The following Figure 31 is the result.

Purging Effect on DI Water

 

 
 

 

 

 

 

 

 

180W W W WWW W — r 12
160 ' WW W 9
. 10
140 W a - .
A120 W W W . 8 9,,
> 0)
£100 WW _ -_ _ - is g
g 80 W- __ _'__ ' _ W; 1 g
m_ ‘ _
° 60 ,r - W - ”'SEP *4 E
40 +
L/W; ; i — a I 7 a .
o __ — + — -W+ . — 1.— o
o 20 4o 60 80

Time (Hour)
Figure 31. The Purging Effect on DO, pH and ORP

After 1 hour of purging, the concentration of dissolved oxygen decreased
to 0.8 mg/ L, while there was no significant change on pH and ORP.
However, after 12 hours of purging, pH increased to 9.4 and ORP

decreased to 90 mV.

77

A-4. Shaking Effect on Sodium Thiosulfate

Five grams of sodium thiosulfate was added to 500 m1 of DI water. A
prepared solution of 100 ml was poured into 4 glass bottles of 160 m1
capacity. Two glass bottles were shaken at 120 rpm for 3 days and the
other 2 bottles were just kept for 3 days without shaking. The DO, pH
and ORP were measured during the period. The following Figure 32 and

Figure 33 are the results.

 

 

 

 

 

 

 

 

ORP (mV)
pH, DO (mg/L)

 

 

 

 

 

 

 

 

o _ WW WW W- W----+ o

0 0.5 1 1.5 2 2.5 3 3.5
Time (Hour)

Figure 32. DO, pH and ORP Change without Shaking

78

 

 

 

90 I ~WW W _,_________, W. W , ,, W . 9

~: W W 18

7o -----— — ______ W — W 1 7
9 60 f ' — l +ORP' l 6 g
g 50 W +00 'W ; 5 g
& 40 > W W W W+pH , 4 8
O 30 W W W a l 3 F;

20 4——~~ , , ~ WWW—W WW . 2

10 i W W—A » , { 1

o l -__ - , , . ., . . . 0

o 0.5 1 1.5 2 2.5 3 3.5

Time (Hour)

Figure 33. DO, pH and ORP Change during Shaking

Figure 32 indicates the change of the pH, DO and ORP in the solution
without shaking. The pH and ORP did not change significantly.

Figure 33 shows the change of the pH, DO and ORP in the solution
during shaking. The pH of the solution was raised up to 7.7 and ORP
decreased down to 10 mV. Shaking accelerated the dissociation reaction
Na28203 :> 2Na+ + 82032' so that the pH of the solution increased and
ORP of the solution decreased. The produced 82032' reacted with oxygen
and consumed the dissolved oxygen, so the concentration of D0 in the

solution decreased.

79

A-S. Eh (Redox-Potential)

The Eh value is calculated using the following formula :

Eh = E0 -- (0.059/n) logQ

Example: SO42" + 2H+ + 26 :> 8032- + H2O E0 = -0.04

Eh = E0 — (0.059/n) logQ

Eh = -0.04 — (0.059/2) log [8032'1/ [H+]2[SO42‘]

Substituting each concentration into the equation, Eh can be determined

(theoritically).

A-6. ORP Measurement

The ORP electrode measures the potential of the reaction. The ORP
electrode consists of two half cells: Metallic indicating electrode and
Reference electrode. The Eh value of the solution is determined by adding
each signal:

Eh = E metal + E ref

An (Ag/AgCl) electrode was used for this study.

80

Reference

Ahsan, A. (2000). Arsenic removal from Aqueous Solution Utilizing
Zeolites and Granular Activated Carbon. Michigan State
University

BGS. (2000). Groundwater Studies of Arsenic Contamination in
Bangladesh. UK: British Geological Survey

BRAC (1999) Community-based arsenic mitigation programme. BRAC
report, Dhaka.

Cheng et a1. (1994) Enhanced Coagulation for Arsenic Removal. Journal
AWWA, 86, 9, 79

Clifford, D. (1986). Removing dissolved inorganic contaminants from
water. Environmental Science 85 Technology, 20, 1072-
1080, Easton

Clifford, D and Lin. (1991). Arsenic(III) and arsenic(V) removal from
Drinking water in San Ysidro, New Mexico. EPA/ 600 / SQ-
91/011, USEPA, Cincinnati, OH.

Cullen, W and Reimer, K (1989). Arsenic Speciation in the Environment.
Chemical Review, 89, 713-764

Eary and Schramke. (1990). Rates of Inorganic Oxidation Reactions
Involving Dissolved Oxygen. American Chemical society,
Washington, DC.

Edwards. (1994) Chemistry of arsenic removal during coagulation and
Fe—Mn oxidation. M Edwards. Journal of AWWA
86, 9, 64-7 8

Frank and Clifford. (1986) Arsenic (III) Oxidation and Removal from
Drinking Water. EPA/ 600/ S2-86/ 021

Franz, N. (Apr.25, 2001). EPA commits to lower arsenic standard.
NY: Chemical Week

Hering et a1. (1997) Arsenic Removal from Drinking Water During
Coagulation. Journal of Environmental Engineering, 123,
8, 800

81

Huang and Vane. (1989) Enhancing As(V) removal by a Fe (II) treated

activated carbon. Water Pollution Control Fed, 61,
1596- 1603

Johnston, R. et al (2001). Safe Water Technology. United Nations
Synthesis report on Arsenic in Drinking Water, Chapter 6.

J akariya, M. (2000). The use of alternative safe water options to mitigate
the arsenic problem in Bangladesh. U.K.: University of
Cambridge

Jekel, M (1994). Arsenic in the Environment, Part 1, Cycling and
Characterization, Chapter 6. John wiley 85 Sons, INC.

Kim and Nriagu. (2000) Oxidation of Arsenite in Groundwater using
Ozone and Oxygen. Science of the total environment, 247,
71-7 9

Korte, N. (1991). A review of Arsenic (III) in Groundwater. Critical Review
in Environmental Control, 21, 1-40

NAS (National Academy of Science). (1999). Arsenic in drinking Water.
Washington, D.C.: National Academy Press

Rasul, S.B. et a1 (1999). Scientific appraisal of a simple arsenic removal
method for groundwater of Bangladesh

Reynolds, T. (1982). Adsorption. Unit Operation and Processes in
Environmental Engineering, Chapter 6. Brooks / Cole,
Monterey, California

Smedley, P. (2001). Source and behavior of arsenic in natural water.
United Nations Synthesis report on Arsenic in Drinking
Water, Chapter 1.Huang, C.P. and Vane, L.M. (1989).

Standard Method. (1995). Standard Method for the Examination of Water
and Wastewater. 19th edition. APHA, AWWA, and WEF,
Washington, DC.

Voice, T. (1997 ). Activated Carbon Adsorption. In: Standard Handbook of
Hazardous Waste Treatment and Disposal, 2nd Edition,
Section 6.1, McGraw-Hill

Wang, L et a1. (2000). Arsenic removal from drinking Water by Ion
Exchange and Activated Alumina Plants. EPA / 600 / R-
00/088, USEPA, Cincinnati, OH.

82

Waypa, J et a1 (1997). Arsenic removal by R0 and NF membranes.
Journal AWWA, 89, 102-114

WHO. (2001). Arsenic in Drinking Water. Fact Sheet No. 210.
http: / / www.who.int/ inf-fs / en / fact2 10.html

83

IIIIIIIIIIIIIIIIIIIII

lllllllll'lllzlljljllllllsllllllllllwill)