20.01.2015 Views

Arsenic Coagulation with Iron, Aluminum, Titanium, and Zirconium ...

Arsenic Coagulation with Iron, Aluminum, Titanium, and Zirconium ...

Arsenic Coagulation with Iron, Aluminum, Titanium, and Zirconium ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

<strong>Arsenic</strong> Water Technology Partnership<br />

<strong>Arsenic</strong> <strong>Coagulation</strong> <strong>with</strong> <strong>Iron</strong>, <strong>Aluminum</strong>, <strong>Titanium</strong>,<br />

<strong>and</strong> <strong>Zirconium</strong> Salts


<strong>Arsenic</strong> <strong>Coagulation</strong> <strong>with</strong> <strong>Iron</strong>, <strong>Aluminum</strong>, <strong>Titanium</strong>,<br />

<strong>and</strong> <strong>Zirconium</strong> Salts<br />

Prepared by:<br />

Divagar Lakshmanan, Dennis A. Clifford, <strong>and</strong> Gautam Samanta<br />

Department of Civil <strong>and</strong> Environmental Engineering<br />

University of Houston<br />

Houston, TX 77204-4791<br />

Jointly Sponsored by:<br />

Awwa Research Foundation<br />

6666 West Quincy Avenue, Denver CO 80235-3098<br />

<strong>and</strong><br />

U.S. Department of Energy<br />

Washington, D.C. 20585-1290<br />

Published by:<br />

WERC, a Consortium for<br />

Environmental Education <strong>and</strong><br />

Technology Development at<br />

New Mexico State University<br />

Awwa Research Foundation


DISCLAIMER<br />

This study was jointly funded by the Awwa Research Foundation (AwwaRF) <strong>and</strong> the U.S.<br />

Department of Energy (DOE) under Grant No. DE-FG02-03ER63619 through the <strong>Arsenic</strong> Water<br />

Technology Partnership. The comments <strong>and</strong> views detailed herein may not necessarily reflect<br />

the views of the Awwa Research Foundation, its officers, directors, affiliates or agents, or the<br />

views of the U.S. Federal Government <strong>and</strong> the <strong>Arsenic</strong> Water Technology Partnership. The<br />

mention of trade names for commercial products does not represent or imply the approval or<br />

endorsement of AwwaRF or DOE. This report is presented solely for informational purposes.<br />

Copyright 2008<br />

By Awwa Research Foundation<br />

<strong>and</strong> <strong>Arsenic</strong> Water Technology Partnership<br />

All Rights Reserved<br />

Printed in the U.S.A.


CONTENTS<br />

LIST OF TABLES ......................................................................................................................... ix<br />

LIST OF FIGURES ....................................................................................................................... xi<br />

FOREWORD .............................................................................................................................. xvii<br />

ACKNOWLEDGMENTS ........................................................................................................... xix<br />

EXECUTIVE SUMMARY ......................................................................................................... xxi<br />

CHAPTER 1 INTRODUCTION .................................................................................................... 1<br />

<strong>Arsenic</strong> Occurrence, Health Effects <strong>and</strong> Regulations .......................................................1<br />

<strong>Iron</strong> <strong>Coagulation</strong>-Filtration <strong>and</strong> Adsorption Processes .....................................................1<br />

Background of the Study ..................................................................................................2<br />

Significance of In-situ Formed Hydroxides in <strong>Arsenic</strong> Removal ........................... 2<br />

Alum as Coagulant for <strong>Arsenic</strong> Removal ............................................................... 2<br />

<strong>Zirconium</strong> Oxide/Hydroxide Adsorbent Media for <strong>Arsenic</strong> Removal ................... 3<br />

Basis of the Project ................................................................................................. 3<br />

Objectives .........................................................................................................................4<br />

Anticipated Practical Benefits of the Project ....................................................................4<br />

CHAPTER 2 MATERIALS AND METHODS ............................................................................. 7<br />

Reagents <strong>and</strong> Stocks .........................................................................................................7<br />

Preparation of NSFI-53 Challenge Water .........................................................................7<br />

Preparation of Coagulant Stock ........................................................................................7<br />

<strong>Coagulation</strong>-Filtration Procedure ......................................................................................8<br />

Preservation <strong>and</strong> Speciation ..............................................................................................9<br />

Instrumentation ...............................................................................................................10<br />

Total <strong>Arsenic</strong> <strong>and</strong> <strong>Arsenic</strong> <strong>and</strong> <strong>Arsenic</strong>(III) Analysis Using FI- HG-AAS ....................10<br />

Freundlich Isotherm ........................................................................................................11<br />

CHAPTER 3 As(III) VS. As(V) REMOVAL .............................................................................. 13<br />

Introduction .....................................................................................................................13<br />

<strong>Arsenic</strong> Removal Study Procedure .................................................................................14<br />

<strong>Arsenic</strong> Removal Using Ferric (III) Chloride ....................................................... 14<br />

<strong>Arsenic</strong> Removal Using Alum .............................................................................. 15<br />

<strong>Arsenic</strong> Removal Using <strong>Zirconium</strong> (IV) Chloride ............................................... 16<br />

<strong>Arsenic</strong> Removal Using <strong>Titanium</strong> (IV) Chloride.................................................. 17<br />

<strong>Arsenic</strong> Removal Using <strong>Titanium</strong> (III) Chloride .................................................. 18<br />

<strong>Arsenic</strong> Removal Using <strong>Titanium</strong> (IV) Oxychloride ............................................ 19<br />

<strong>Arsenic</strong> Removal Using <strong>Zirconium</strong> (IV) Oxychloride ......................................... 19<br />

<strong>Arsenic</strong> Adsorption Isotherms ........................................................................................20<br />

Comparison of Adsorption Capacities ............................................................................24<br />

v


<strong>Arsenic</strong> Adsorption Comparison on a Mass Basis ................................................ 24<br />

<strong>Arsenic</strong> Adsorption Comparison on a Molar Basis .............................................. 25<br />

Oxidation Study of <strong>Arsenic</strong>(III) in the Presence of <strong>Titanium</strong>(III) chloride ....................28<br />

TiCl 3 Oxidation Study Results .............................................................................. 28<br />

Proposed Mechanism for the Oxidation of <strong>Arsenic</strong>(III) in Presence of<br />

<strong>Titanium</strong>(III) ................................................................................................... 29<br />

Summary <strong>and</strong> Conclusions .............................................................................................30<br />

CHAPTER 4 EFFECT OF COMPETING IONS ON ARSENIC (III)/(V) ADSORPTION........ 31<br />

Introduction .....................................................................................................................31<br />

Effect of Silica in NSFI Challenge Water Containing Phosphate ..................................32<br />

Effect of Silica in NSFI Water <strong>with</strong> Phosphate Using Ferric (III) Chloride as<br />

Coagulant ........................................................................................................ 32<br />

Effect of Silica in NSFI Water <strong>with</strong> Phosphate Using Alum as Coagulant .......... 34<br />

Effect of Silica in NSFI Water <strong>with</strong> Phosphate Using <strong>Zirconium</strong> (IV)<br />

Chloride as Coagulant ..................................................................................... 34<br />

Effect of Silica in NSFI Water <strong>with</strong> Phosphate Using <strong>Titanium</strong> (IV)<br />

Chloride as Coagulant ..................................................................................... 36<br />

Effect of Silica in NSFI Water <strong>with</strong> Phosphate Using <strong>Titanium</strong> (III)<br />

Chloride as Coagulant ..................................................................................... 37<br />

Effect of Phosphate in NSFI Challenge Water Containing Silica ..................................38<br />

Effect of Phosphate in NSFI Water <strong>with</strong> Silica Using Ferric (III) Chloride as<br />

Coagulant ........................................................................................................ 39<br />

Effect of Phosphate in NSFI Water <strong>with</strong> Silica Using Alum as Coagulant .......... 40<br />

Effect of Phosphate in NSFI Water <strong>with</strong> Silica Using <strong>Zirconium</strong> (IV)<br />

Chloride as Coagulant ..................................................................................... 41<br />

Effect of Phosphate in NSFI Water <strong>with</strong> Silica Using <strong>Titanium</strong> (IV)<br />

Chloride as Coagulant ..................................................................................... 42<br />

Effect of Phosphate in NSFI Water <strong>with</strong> Silica Using <strong>Titanium</strong> (III)<br />

Chloride as Coagulant ..................................................................................... 43<br />

Effect of Vanadate in NSFI Challenge Water Containing Silica <strong>and</strong> Phosphate ...........45<br />

Effect of Vanadate in NSFI Water <strong>with</strong> Silica <strong>and</strong> Phosphate Using Ferric<br />

(III) Chloride as Coagulant ............................................................................. 45<br />

Effect of Vanadate in NSFI Water <strong>with</strong> Silica <strong>and</strong> Phosphate Using Alum as<br />

Coagulant ........................................................................................................ 46<br />

Effect of Vanadate in NSFI Water <strong>with</strong> Silica <strong>and</strong> Phosphate Using<br />

<strong>Zirconium</strong> (IV) Chloride as Coagulant ........................................................... 47<br />

Effect of Vanadate in NSFI Water <strong>with</strong> Silica <strong>and</strong> Phosphate Using<br />

<strong>Titanium</strong> (IV) Chloride as Coagulant ............................................................. 48<br />

Effect of Vanadate in NSFI Water <strong>with</strong> Silica <strong>and</strong> Phosphate Using<br />

<strong>Titanium</strong> (III) Chloride as Coagulant ............................................................. 49<br />

Individual Effects of Silica, Phosphate <strong>and</strong> Vanadate <strong>with</strong> Ferric (III) Chloride as<br />

Coagulant in NSFI Challenge Water Without Competing Ions................................50<br />

Effect of Silica in the Absence of Phosphate <strong>and</strong> Vanadate ................................. 51<br />

Effect of Phosphate in the Absence of Silica <strong>and</strong> Vanadate ................................. 52<br />

Effect of Vanadate in the Absence of Silica <strong>and</strong> Phosphate ................................. 53<br />

vi


Summarizing the Effects of Competing Ions on the <strong>Arsenic</strong> Adsorption Capacity<br />

in NSFI Challenge Water ..........................................................................................55<br />

Individual Effects of Competing Ions on the <strong>Arsenic</strong> Adsorption Capacity of<br />

Ferric (III) Hydroxide ...............................................................................................61<br />

Conclusions .....................................................................................................................63<br />

CHAPTER 5 TOXICITY CHARACTERISTIC STUDIES ......................................................... 65<br />

Introduction .....................................................................................................................65<br />

Preliminary Toxicity Characteristic Evaluation ..............................................................65<br />

Experimental Study Procedure ............................................................................. 66<br />

Toxicity Characteristic Study Results .............................................................................67<br />

Experimental Evaluation of the Adsorption of High Concentrations of <strong>Arsenic</strong><br />

onto Metal Hydroxides .............................................................................................67<br />

Conclusions .....................................................................................................................68<br />

CHAPTER 6 ARSENIC REMOVAL WITH FeCl 3 COAGULATION: COMPARISON OF<br />

EXPERIMENTAL RESULTS WITH MINEQL+ 4.50 CHEMICAL EQUILIBRIUM<br />

MODELING PROGRAM PREDICTIONS ........................................................................... 69<br />

Introduction .....................................................................................................................69<br />

Mineql+ Program Modeling Procedure ..........................................................................69<br />

As(III) <strong>and</strong> As(V) Adsorption Isotherms Based on Mineql+ Modeling .........................72<br />

Comparison of Model-predicted vs Experimentally Observed <strong>Arsenic</strong> Adsorption<br />

onto Fe(OH) 3 (s) ........................................................................................................76<br />

Conclusions .....................................................................................................................79<br />

CHAPTER 7 SUMMARY AND CONCLUSIONS ..................................................................... 81<br />

SUMMARY ....................................................................................................................81<br />

Conclusions .....................................................................................................................81<br />

Recommendations ...........................................................................................................85<br />

REFERENCES ............................................................................................................................. 87<br />

ABBREVIATIONS ...................................................................................................................... 91<br />

vii


viii


TABLES<br />

Table 2.1 Composition of NSFI 53 challenge water........................................................................8<br />

Table 2.2 Composition of reagents in preparation of coagulant stock for coagulation study ........ 8<br />

Table 2.3 Experimental conditions for the determination of As(III) <strong>and</strong> As(Tot) by FI-HG-<br />

AAS.......................................................................................................................... 10<br />

Table 3.1 Equilibrium constants for some important arsenic containing acids ............................ 13<br />

Table 3.2 Comparison of chemical costs for treating one million gallons of NSFI-53 water<br />

to reduce As(V)/As(III) from 50 to < 10 μg/L ......................................................... 27<br />

Table 4.1 Composition of species of some important acids present in pH 6.5-8.5 ...................... 32<br />

Table 5.1 Percent dry solids in the liquid wastes after 3-hr settling ............................................. 67<br />

Table 5.2 <strong>Arsenic</strong> concentrations in the TCLP- or WET-defined extract (liquid waste after<br />

filtration) .................................................................................................................. 68<br />

Table 5.3 Comparison of expected <strong>and</strong> measured arsenic adsorption capacities of metal<br />

hydroxides for all coagulants ................................................................................... 68<br />

Table 6.1 Composition of NSFI-53 challenge water .................................................................... 70<br />

Table 6.2 Mineql+ titration parameters ........................................................................................ 70<br />

Table 6.3 As(V) adsorption capacity of Fe(OH) 3 based on Mineql+ program ............................. 71<br />

ix


FIGURES<br />

Figure 2.1 <strong>Coagulation</strong>-filtration procedure ................................................................................... 9<br />

Figure 3.1 Mechanism of arsenate lig<strong>and</strong> exchange on the surface of metal oxyhydroxides ...... 14<br />

Figure 3.2 Removal efficiency of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

ferric chloride dose <strong>and</strong> pH ................................................................................................... 15<br />

Figure 3.3 Removal efficiency of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

alum dose <strong>and</strong> pH .................................................................................................................. 16<br />

Figure 3.4 Removal efficiency of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

zirconium(IV) chloride dose <strong>and</strong> pH .................................................................................... 16<br />

Figure 3.5 Removal efficiency of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

titanium (IV) chloride dose <strong>and</strong> pH ...................................................................................... 17<br />

Figure 3.6 Removal efficiency of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

titanium (III) chloride dose <strong>and</strong> pH ...................................................................................... 18<br />

Figure 3.7 Removal/Conversion efficiency of As(III) in NSFI challenge water as a function of<br />

titanium(III) chloride dose <strong>and</strong> pH ....................................................................................... 19<br />

Figure 3.8 Removal efficiency of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

titanium(IV) oxychloride dose <strong>and</strong> pH ................................................................................. 20<br />

Figure 3.9 Removal efficiency of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

zirconium(IV) oxychloride coagulant dose <strong>and</strong> pH .............................................................. 20<br />

Figure 3.10 Comparison of As(V) adsorption isotherms for the in-situ formed oxyhydroxides of<br />

Fe(III), Al(III), Zr(IV), Ti(IV), <strong>and</strong> Ti(III) at pH 6.5 ........................................................... 21<br />

Figure 3.11 Comparison of As(V) adsorption isotherms for the in-situ formed oxyhydroxides of<br />

Fe(III), Al(III), Zr(IV), Ti(IV), <strong>and</strong> Ti(III) at pH 7.5 ........................................................... 21<br />

Figure 3.12 Comparison of As(V) adsorption isotherms for the in-situ formed oxyhydroxides of<br />

Fe(III), Al(III), Zr(IV), Ti(IV), <strong>and</strong> Ti(III) at pH 8.5 ........................................................... 22<br />

Figure 3.13 Comparison of As(III) adsorption isotherms for the in-situ formed oxyhydroxides of<br />

Fe(III), Al(III), Zr(IV), Ti(IV), <strong>and</strong> Ti(III) at pH 6.5 ........................................................... 22<br />

Figure 3.14 Comparison of As(III) adsorption isotherms for the in-situ formed oxyhydroxides of<br />

Fe(III), Al(III), Zr(IV), Ti(IV), <strong>and</strong> Ti(III) at pH 7.5 ........................................................... 23<br />

xi


Figure 3.15 Comparison of As(III) adsorption isotherms for the in-situ formed oxyhydroxides of<br />

Fe(III), Al(III), Zr(IV), Ti(IV), <strong>and</strong> Ti(III) at pH 8.5 ........................................................... 23<br />

Figure 3.16 Comparison of mass As(V) adsorbed per mass of Fe(III), Al(III), Ti(III), Ti(IV), or<br />

Zr(IV) coagulant for an equilibrium concentration of 10 μg/L as a function of pH ............. 24<br />

Figure 3.17 Comparison of mass As(III) adsorbed per mass of Fe(III), Al(III), Ti(III), Ti(IV), or<br />

Zr(IV) coagulant for an equilibrium concentration of 10 μg/L as a function of pH ............. 25<br />

Figure 3.18 Comparison of moles As(V) adsorbed per mol of Fe(III), Al(III), Ti(III), Ti(IV), or<br />

Zr(IV) coagulant for an equilibrium concentration of 10 μg/L as a function of pH ............. 26<br />

Figure 3.19 Comparison of moles As(III) adsorbed per mol of Fe(III), Al(III), Ti(III), Ti(IV), or<br />

Zr(IV) coagulant for an equilibrium concentration of 10 μg/L as a function of pH ............. 26<br />

Figure 3.20 As(III) <strong>and</strong> As(Tot) removed or converted during Ti(III) coagulation as a function of<br />

flocculation <strong>and</strong> settling times .............................................................................................. 29<br />

Figure 4.1 Effect of silica on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed Fe(III)<br />

hydroxide in NSFI water <strong>with</strong> phosphate ............................................................................. 33<br />

Figure 4.2 Effect of silica on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed Fe(III)<br />

hydroxide in NSFI water <strong>with</strong> phosphate ............................................................................. 33<br />

Figure 4.3 Effect of silica on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed Al(III)<br />

hydroxide in NSFI water <strong>with</strong> phosphate ............................................................................. 34<br />

Figure 4.4 Effect of silica on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed Zr(IV)<br />

hydroxide in NSFI water <strong>with</strong> phosphate ............................................................................. 35<br />

Figure 4.5 Effect of silica on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed Zr(IV)<br />

hydroxide in NSFI water <strong>with</strong> phosphate ............................................................................. 35<br />

Figure 4.6 Effect of silica on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed Ti(IV)<br />

hydroxide in NSFI water <strong>with</strong> phosphate ............................................................................. 36<br />

Figure 4.7 Effect of silica on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed Ti(IV)<br />

hydroxide in NSFI water <strong>with</strong> phosphate ............................................................................. 37<br />

Figure 4.8 Effect of silica on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed Ti(III)<br />

hydroxide in NSFI water <strong>with</strong> phosphate ............................................................................. 37<br />

Figure 4.9 Effect of silica on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed Ti(III)<br />

hydroxide in NSFI water <strong>with</strong> phosphate ............................................................................. 38<br />

xii


Figure 4.10 Effect of phosphate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Fe(III) hydroxides in NSFI water <strong>with</strong> silica ........................................................................ 39<br />

Figure 4.11 Effect of phosphate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Fe(III) hydroxides in NSFI water <strong>with</strong> silica ........................................................................ 40<br />

Figure 4.12 Effect of phosphate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Al(III) hydroxides in NSFI water <strong>with</strong> silica ........................................................................ 40<br />

Figure 4.13 Effect of phosphate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Zr(IV) hydroxide in NSFI water <strong>with</strong> silica ......................................................................... 41<br />

Figure 4.14 Effect of phosphate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Zr(IV) hydroxides in NSFI water <strong>with</strong> silica ........................................................................ 41<br />

Figure 4.15 Effect of phosphate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Ti(IV) hydroxides in NSFI water <strong>with</strong> silica ........................................................................ 42<br />

Figure 4.16 Effect of phosphate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Ti(IV) hydroxides in NSFI water <strong>with</strong> silica ........................................................................ 43<br />

Figure 4.17 Effect of phosphate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Ti(III) hydroxides in NSFI water <strong>with</strong> silica ........................................................................ 44<br />

Figure 4.18 Effect of phosphate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Ti(III) hydroxides in NSFI water <strong>with</strong> silica ........................................................................ 44<br />

Figure 4.19 Effect of vanadate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Fe(III) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate ................................................ 45<br />

Figure 4.20 Effect of vanadate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Fe(III) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate ................................................ 46<br />

Figure 4.21 Effect of vanadate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Al(III) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate ................................................ 46<br />

Figure 4.22 Effect of vanadate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Zr(IV) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate ................................................ 47<br />

Figure 4.23 Effect of vanadate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Zr(IV) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate ................................................ 47<br />

Figure 4.24 Effect of vanadate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Ti(IV) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate ................................................ 48<br />

xiii


Figure 4.25 Effect of vanadate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Ti(IV) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate ................................................ 49<br />

Figure 4.26 Effect of vanadate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Ti(III) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate ................................................ 50<br />

Figure 4.27 Effect of vanadate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Ti(III) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate ................................................ 50<br />

Figure 4.28 Effect of silica on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed Fe(III)<br />

hydroxides in the absence of phosphate <strong>and</strong> vanadate .......................................................... 51<br />

Figure 4.29 Effect of silica on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed Fe(III)<br />

hydroxides in the absence of phosphate <strong>and</strong> vanadate .......................................................... 52<br />

Figure 4.30 Effect of phosphate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Fe(III) hydroxides in the absence of silica <strong>and</strong> vanadate ...................................................... 53<br />

Figure 4.31 Effect of phosphate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Fe(III) hydroxides in the absence of silica <strong>and</strong> vanadate ...................................................... 53<br />

Figure 4.32 Effect of vanadate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Fe(III) hydroxides in the absence of silica <strong>and</strong> phosphate .................................................... 54<br />

Figure 4.33 Effect of vanadate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Fe(III) hydroxides in the absence of silica <strong>and</strong> phosphate .................................................... 55<br />

Figure 4.34 Comparison of arsenic(V) adsorption capacities of in-situ formed Fe(III) hydroxides<br />

for an equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5 ....................... 56<br />

Figure 4.35 Comparison of arsenic(III) adsorption capacities of in-situ formed Fe(III) hydroxides<br />

for an equilibrium concentration of 10 μg/L As(III) in the pH range 6.5-8.5 ...................... 56<br />

Figure 4.36 Comparison of arsenic(V) adsorption capacities of in-situ formed Al(III) hydroxides<br />

for an equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5 ....................... 57<br />

Figure 4.37 Comparison of arsenic(V) adsorption capacities of in-situ formed Zr(IV) hydroxides<br />

for an equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5 ....................... 58<br />

Figure 4.38 Comparison of arsenic(III) adsorption capacities of in-situ formed Zr(IV) hydroxides<br />

for an equilibrium concentration of 10 μg/L As(III) in the pH range 6.5-8.5 ...................... 58<br />

Figure 4.39 Comparison of arsenic(V) adsorption capacities of in-situ formed Ti(IV) hydroxides<br />

for an equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5 ....................... 59<br />

xiv


Figure 4.40 Comparison of arsenic(III) adsorption capacities of in-situ formed Ti(IV) hydroxides<br />

for an equilibrium concentration of 10 μg/L As(III) in the pH range 6.5-8.5 ...................... 60<br />

Figure 4.41 Comparison of arsenic(V) adsorption capacities of in-situ formed Ti(III) hydroxides<br />

for an equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5 ....................... 60<br />

Figure 4.42 Comparison of arsenic(III) adsorption capacities of in-situ formed Ti(III) hydroxides<br />

for an equilibrium concentration of 10 μg/L As(III) in the pH range 6.5-8.5 ...................... 61<br />

Figure 4.43 Comparison of arsenic adsorption capacities of in-situ-formed Fe(III) hydroxides for<br />

an equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5 ............................. 62<br />

Figure 4.44 Comparison of arsenic adsorption capacities of in-situ formed Fe(III) hydroxides for<br />

an equilibrium concentration of 10 μg/L As(III) in the pH range 6.5-8.5 ............................ 62<br />

Figure 5.1 Preliminary determination of percent solids ................................................................ 66<br />

Figure 6.1 Modeling of As(V) adsorption isotherm onto Fe(OH) 3 (s) at pH 7.5 .......................... 72<br />

Figure 6.2 Mineql+-predicted effect of silica, phosphate, <strong>and</strong> vanadate on As(V) adsorption on<br />

the in-situ-formed Fe(III) hydroxide at pH 6.5 ..................................................................... 73<br />

Figure 6.3 Mineql+-predicted effect of silica, phosphate, <strong>and</strong> vanadate on As(V) adsorption on<br />

the in-situ-formed Fe(III) hydroxide at pH 7.5 ..................................................................... 73<br />

Figure 6.4 Mineql+-predicted effect of silica, phosphate, <strong>and</strong> vanadate on As(V) adsorption on<br />

the in-situ-formed Fe(III) hydroxide at pH 8.5 ..................................................................... 74<br />

Figure 6.5 Mineql+-predicted effect of silica, phosphate, <strong>and</strong> vanadate on As(III) adsorption on<br />

the in-situ-formed Fe(III) hydroxide at pH 6.5 ..................................................................... 74<br />

Figure 6.6 Mineql+-predicted effect of silica, phosphate, <strong>and</strong> vanadate on As(III) adsorption on<br />

the in-situ-formed Fe(III) hydroxide at pH 7.5 ..................................................................... 75<br />

Figure 6.7 Mineql+-predicted effect of silica, phosphate, <strong>and</strong> vanadate on As(III) adsorption on<br />

the in-situ-formed Fe(III) hydroxide at pH 8.5 ..................................................................... 75<br />

Figure 6.8 Comparison of As(V) adsorption capacities of Fe(OH) 3 for an equilibrium<br />

concentration of 10 μg/L at pH 6.5 ....................................................................................... 76<br />

Figure 6.9 Comparison of As(V) adsorption capacities of Fe(OH) 3 for an equilibrium<br />

concentration of 10 μg/L at pH 7.5 ....................................................................................... 77<br />

Figure 6.10 Comparison of As(V) adsorption capacities of Fe(OH) 3 for an equilibrium<br />

concentration of 10 μg/L at pH 8.5 ....................................................................................... 77<br />

xv


Figure 6.11 Comparison of As(III) adsorption capacities of Fe(OH) 3 for an equilibrium<br />

concentration of 10 μg/L at pH 6.5 ....................................................................................... 78<br />

Figure 6.12 Comparison of As(III) adsorption capacities of Fe(OH) 3 for an equilibrium<br />

concentration of 10 μg/L at pH 7.5 ....................................................................................... 78<br />

Figure 6.13 Comparison of As(III) adsorption capacities of Fe(OH) 3 for an equilibrium<br />

concentration of 10 μg/L at pH 8.5 ....................................................................................... 79<br />

xvi


FOREWORD<br />

The Awwa Research Foundation is a nonprofit corporation that is dedicated to the<br />

implementation of a research effort to help utilities respond to regulatory requirements <strong>and</strong><br />

traditional high-priority concerns of the industry. The research agenda is developed through a<br />

process of consultation <strong>with</strong> subscribers <strong>and</strong> drinking water professionals. Under the umbrella of<br />

a strategic Research Plan, the Research Advisory Council prioritizes the suggested projects based<br />

upon current <strong>and</strong> future needs, applicability, <strong>and</strong> past work; the recommendations are forwarded<br />

to the Board of Trustees for final selection. The foundation also sponsors research projects<br />

through the unsolicited proposal process; the Collaborative Research, Research Applications, <strong>and</strong><br />

Tailored Collaboration programs; <strong>and</strong> various joint research efforts <strong>with</strong> organizations such as<br />

the U.S. Environmental Protection Agency, the U.S. Bureau of Reclamation, <strong>and</strong> the U.S.<br />

Department of Energy.<br />

This publication is a result of one of these sponsored studies, <strong>and</strong> it is hoped that its<br />

findings will be applied in communities throughout the world. The following report serves not<br />

only as a means of communicating the results of the water industry’s centralized research<br />

program but also as a tool to enlist the further support of the nonmember utilities <strong>and</strong> individuals.<br />

Projects are managed closely from their inception to the final report by the foundation’s<br />

staff <strong>and</strong> large cadre of volunteers who willingly contribute their time <strong>and</strong> expertise. The<br />

foundation serves a planning <strong>and</strong> management function <strong>and</strong> awards contracts to other institutions<br />

such as water utilities, universities, <strong>and</strong> engineering firms. The funding for this research effort<br />

comes primarily from the Subscription Program, through which water utilities subscribe to the<br />

research program <strong>and</strong> make an annual payment proportionate to the volume of water they deliver<br />

<strong>and</strong> consultants <strong>and</strong> manufacturers subscribe based on their annual billings. The program offers<br />

a cost-effective <strong>and</strong> fair method for funding research in the public interest.<br />

A broad spectrum of water supply issues is addressed by the foundation’s research<br />

agenda: resources, treatment <strong>and</strong> operations, distribution <strong>and</strong> storage, water quality <strong>and</strong> analysis,<br />

toxicology, economics, <strong>and</strong> management. The ultimate purpose of the coordinated effort is to<br />

assist water suppliers to provide the highest possible quality of water economically <strong>and</strong> reliably.<br />

The true benefits are realized when the results are implemented at the utility level. The<br />

foundation’s trustees are pleased to offer this publication as a contribution towards that end.<br />

David Ragner<br />

Chair, Board of Trustees<br />

Awwa Research Foundation<br />

Robert C. Renner<br />

Executive Director<br />

Awwa Research Foundation<br />

xvii


xviii


ACKNOWLEDGMENTS<br />

This report is the product of a collaborative effort between the members of the <strong>Arsenic</strong><br />

Water Technology Partnership <strong>and</strong> was made possible by funds from Congress <strong>and</strong> the drinking<br />

water community. A special thanks to U.S. Senator Pete Domenici for his support <strong>and</strong> assistance<br />

in helping to bring low-cost, energy efficient solutions to remove arsenic from drinking water.<br />

The authors kindly thank the following persons for their cooperation <strong>and</strong> participation in<br />

this project:<br />

1. Albert Ilges, AwwaRF <strong>Arsenic</strong> Program Project Manager, for his helpful <strong>and</strong> responsive<br />

attitude <strong>and</strong> his technical <strong>and</strong> administrative guidance throughout the course of the<br />

project.<br />

2. Project Advisory Committee members Dr. Abbas Ghassemi <strong>and</strong> Dr. Malcolm Siegel for<br />

reviewing the reports <strong>and</strong> for their helpful suggestions <strong>and</strong> observations throughout the<br />

course of the project.<br />

3. Dr. Issam Najm, President of Water Quality <strong>and</strong> Treatment Solutions, Inc. for his help<br />

<strong>with</strong> proposal preparation <strong>and</strong> for his technical advice on field applications of coagulation<br />

filtration for arsenic removal.<br />

The authors are grateful for the funding provided by the Awwa Research Foundation <strong>and</strong> the inkind<br />

matching funding provided by the University of Houston, <strong>and</strong> Water Quality <strong>and</strong> Treatment<br />

Solutions, Inc.<br />

xix


EXECUTIVE SUMMARY<br />

<strong>Coagulation</strong>-Filtration (C-F) <strong>with</strong> Fe salts is an effective <strong>and</strong> widely used technology<br />

for treating drinking water to remove arsenic. For small systems, however, adsorption onto<br />

granular ferric hydroxide (GFH) or granular ferric oxide (GFO) is currently the process of<br />

choice. In a comparison of C-F vs. adsorption, the arsenic loading, mg of As/g Fe, on the<br />

coagulant was higher than on the solid media, which is unexpected in view of the fact that the<br />

coagulant is in equilibrium <strong>with</strong> the effluent arsenic concentration (


4. Study the effect of competing ions (silica, phosphate, <strong>and</strong> vanadate) in NSFI-53 challenge<br />

water on As(V) <strong>and</strong> As(III) removal during coagulation <strong>with</strong> ferric, aluminum, titanium,<br />

<strong>and</strong> zirconium salts,<br />

5. Quantify the individual effects of silica, phosphate <strong>and</strong> vanadate during coagulation <strong>with</strong><br />

FeCl 3 ,<br />

6. Establish the toxicity characteristics of the sludges produced during coagulation <strong>with</strong><br />

ferric, aluminum, titanium, <strong>and</strong> zirconium salts, <strong>and</strong> compare the results <strong>with</strong> the<br />

regulatory limits of the TCLP <strong>and</strong> WET tests, <strong>and</strong><br />

7. Determine if the adsorption of As(V) <strong>and</strong> As(III) onto Fe(OH) 3 could be accurately<br />

modeled using the Mineql+ water chemistry equilibrium model, <strong>and</strong> compare the model<br />

results <strong>with</strong> experimental results.<br />

APPROACH<br />

Extensive jar testing of the effectiveness of the aluminum, titanium, <strong>and</strong> zirconium<br />

coagulants for arsenic removal in comparison <strong>with</strong> FeCl 3 was conducted. The effects of pH, <strong>and</strong><br />

competing ions such as silicate, phosphate, <strong>and</strong> vanadate on As(III) <strong>and</strong> As(V) removal were<br />

determined. An existing surface complexation model (Mineql+) was applied to quantitatively<br />

predict the influence of these parameters on arsenic adsorption. Finally, recommendations are<br />

made as to the optimum coagulation-filtration processes suitable for representative small<br />

community ground waters contaminated <strong>with</strong> arsenic. An outline of the comprehensive,<br />

systematic study of arsenic III/V removal is given below:<br />

1. NSFI challenge water (NSFI St<strong>and</strong>ard-53) spiked <strong>with</strong> 50 µg/L of As(III) or As(V) was<br />

used as the background water for the isotherm tests. The coagulant stocks were prepared<br />

from the following salts: zirconium(IV) chloride (ZrCl 4 ), zirconium(IV) oxychloride<br />

(ZrOCl 2 ), titanium(IV) chloride (TiCl 4 ,), titanium(IV) oxychloride (TiOCl 2 ), titanium(III)<br />

chloride (TiCl 3 ), alum, <strong>and</strong> ferric chloride (FeCl 3 ).<br />

2. <strong>Coagulation</strong>-filtration experiments were performed using jar tests. The predetermined<br />

coagulant dose <strong>and</strong> the base (NaOH) required to reach the desired equilibrium pH (6.5,<br />

7.5, or 8.5) were added at the start of the experiment. The jar test procedure consisted of<br />

1 min rapid mix at 100 rpm followed by 20 min of slow mixing at 20 rpm. After 1-hr of<br />

settling, the supernatant was filtered through a 0.2-μm filter, <strong>and</strong> the filtrate was then<br />

preserved for arsenic analysis.<br />

3. Experiments that assessed the effect of the competing ions in the challenge water were<br />

conducted using all five coagulant metals (Fe(III), Al(III), Ti(IV), Ti(III), <strong>and</strong> Zr(IV)) at<br />

pH 6.5, 7.5, <strong>and</strong> 8.5, spiked <strong>with</strong> two levels of silica, phosphate, <strong>and</strong> vanadate.<br />

4. Experiments that assessed the individual effect of the competing ions were also<br />

conducted using challenge water <strong>with</strong> Fe(III) as coagulant at three pHs spiked <strong>with</strong> two<br />

levels of silica, phosphate, <strong>and</strong> vanadate in the absence of competing ions.<br />

5. Toxicity characteristics of the sludge produced <strong>with</strong> Fe(III), Al(III), Ti(IV), Ti(III), <strong>and</strong><br />

Zr(IV) coagulant were studied according to TCLP <strong>and</strong> WET procedures <strong>and</strong> checked if<br />

the toxicity was <strong>with</strong>in regulatory limits.<br />

6. The adsorption of arsenic (As(V) <strong>and</strong> As(III)) onto Fe(OH) 3 (s) was modeled using the<br />

Mineql+ water chemistry equilibrium model, <strong>and</strong> the model results were compared <strong>with</strong><br />

coagulation experiment results.<br />

xxii


CONCLUSIONS<br />

The comparative study of the removal of arsenic in challenge water <strong>with</strong> the innovative<br />

coagulants (TiCl 4 , TiCl 3 , TiOCl 2 , ZrCl 4 , <strong>and</strong> ZrOCl 2 ) compared to commonly used coagulants<br />

(FeCl 3 , alum) resulted in the following conclusions:<br />

1. The percent removal of As(V) was highly pH dependent in the NSFI challenge water,<br />

<strong>and</strong> the removal increased <strong>with</strong> decreasing pH for all coagulants tested: FeCl 3 , alum,<br />

TiCl 4 , TiCl 3 , TiOCl 2 , ZrCl 4 , <strong>and</strong> ZrOCl 2 . In particular, the adsorption capacity of As(V)<br />

<strong>with</strong> zirconium salts decreased significantly <strong>with</strong> increasing pH.<br />

2. The percent removal efficiency of As(III) was independent of pH for FeCl 3 , TiCl 4 , ZrCl 4 ,<br />

<strong>and</strong> ZrOCl 2 , <strong>and</strong> it decreased <strong>with</strong> increasing pH for TiCl 3 <strong>and</strong> increased <strong>with</strong> increasing<br />

pH for TiOCl 2 . The removal of As(III) by alum was insignificant.<br />

3. At all doses, the removal efficiency of As(V) was significantly greater than As(III) at all<br />

pHs <strong>with</strong> all seven coagulants tested: FeCl 3 , alum, TiCl 4 , TiCl 3 , TiOCl 2 , ZrCl 4 , <strong>and</strong><br />

ZrOCl 2 .<br />

4. When comparing arsenic adsorption isotherms for all the coagulants, the highest As(V)<br />

loadings on a coagulant on a mass basis (mg As(V)/g metal) were observed <strong>with</strong> FeCl 3 ,<br />

which performed better than aluminum, titanium <strong>and</strong> zirconium salts at pHs of 6.5 <strong>and</strong><br />

7.5. However, at pH 8.5, As(V) loadings on FeCl 3 were approximately the same as TiCl 3<br />

at equilibrium As(V) ≤ 10 μg/L.<br />

5. When comparing adsorption isotherms, the highest As(V) loading on any coagulant on a<br />

molar basis or a mass basis was observed for ferric chloride at all three pHs, <strong>and</strong> the<br />

As(V) loading on iron was significantly greater than aluminum.<br />

6. The ranking of As(V) adsorption capacities of the iron, aluminum, titanium, <strong>and</strong><br />

zirconium salts for an equilibrium concentration of 10 μg/L was as follows:<br />

• pH 6.5: FeCl 3 > Alum > ZrOCl 2 ≈ ZrCl 4 > TiCl 4 > TiCl 3 > TiOCl 2<br />

• pH 7.5: FeCl 3 >> TiCl 3 > TiCl 4 > Alum ≈ ZrOCl 2 > ZrCl 4 > TiOCl 2<br />

• pH 8.5: FeCl 3 ≈ TiCl 3 > TiCl 4 > Alum > ZrCl 4 ≈ TiOCl 2 > ZrOCl 2<br />

7. FeCl 3 was a far better coagulant than alum for As(V) removal on a mass basis <strong>and</strong> on a<br />

molar basis.<br />

8. When comparing arsenic adsorption isotherms for all the coagulants, the highest As(III)<br />

loading on a coagulant (mg As(III)/g metal) was observed <strong>with</strong> titanium(III) chloride,<br />

which performed better than ferric, titanium(IV), <strong>and</strong> zirconium salts at pHs of 6.5 <strong>and</strong><br />

7.5. However Ti(III) had similar adsorption capacity to that of Fe(III) <strong>and</strong> Ti(IV)<br />

coagulants at pH 8.5. Alum did not have any adsorption capacity for As(III). TiCl 4<br />

exhibited similar removal efficiency to that of FeCl 3 , <strong>and</strong> TiOCl 2 offered similar removal<br />

xxiii


efficiency to FeCl 3 at pH 8.5. <strong>Zirconium</strong> salts did not have good adsorption capacity for<br />

As(III). Thus, it appears that on an mg metal/L basis, TiCl 3 could be a better coagulant for<br />

As(III) removal in coagulation-filtration processes.<br />

9. On a mass basis, the ranking of As(III) adsorption capacities of iron, aluminum, titanium<br />

<strong>and</strong> zirconium salts for an equilibrium concentration of 10 μg/L was as follows:<br />

• pH 6.5: TiCl 3 > FeCl 3 ≈ TiCl 4 > TiOCl 2 >> ZrCl 4 ≈ ZrOCl 2 >> Alum<br />

• pH 7.5: TiCl 3 > FeCl 3 ≈ TiCl 4 > TiOCl 2 >> ZrCl 4 ≈ ZrOCl 2 >> Alum<br />

• pH 8.5: TiCl 3 ≈ FeCl 3 ≈ TiCl 4 ≈ TiOCl 2 >> ZrCl 4 ≈ ZrOCl 2 >> Alum<br />

10. The highest As(III) removal/loading on the coagulant on a molar basis was observed for<br />

titanium(III) chloride at pH 6.5 <strong>and</strong> 7.5, while at pH 8.5, ferric chloride had the highest<br />

molar adsorption capacity. Note: As(III) was oxidized to As(V) by Ti(III).<br />

11. When comparing chemical costs for FeCl 3 , alum, TiCl 4 , ZrOCl 2 , <strong>and</strong> TiOCl 2 coagulation<br />

to remove As(V) or As(III), the most economical was FeCl 3 . Of the common coagulants,<br />

alum was found to be 4-8 times more expensive than ferric chloride for As(V) removal.<br />

The chemical cost of ferric chloride coagulation was calculated to be more than 5 to 20<br />

times higher for As(III) treatment compared <strong>with</strong> As(V).<br />

12. There was experimental evidence that the high removal efficiency of As(III) by TiCl 3 <strong>and</strong><br />

the unusual As(III) behavior of increasing removal <strong>with</strong> decreasing pH was due to<br />

oxidation of As(III) to As(V) by H 2 O 2 , which based on the literature, formed from Ti(III)<br />

hydrolysis in the NSFI challenge water that contained some dissolved oxygen. In spite of<br />

its partial oxidation, the experimentally observed removal of As(V) oxidized from As(III)<br />

was far less than the removal of a similar starting concentration of As(V), because (a) the<br />

floc was already formed when it contacted As(V), <strong>and</strong> (b) the As(III) oxidation continued<br />

for many hours during which the Ti(OH) 3 (s) formed had settled <strong>and</strong> was not in contact<br />

<strong>with</strong> the As(V) formed.<br />

The studies on effect of competing ions (silica, phosphate <strong>and</strong> vanadate) on arsenic<br />

adsorption in NSFI challenge water <strong>with</strong> FeCl 3 , alum, TiCl 4 , TiCl 3 , <strong>and</strong> ZrCl 4 as coagulants<br />

resulted in the following conclusions:<br />

13. In the NSFI challenge water <strong>with</strong> phosphate, silica significantly reduced the adsorption of<br />

arsenic, presumably by competing for adsorption sites. The As(V) <strong>and</strong> As(III) removal<br />

efficiencies in the absence of silica were higher than in the presence of silica for all<br />

coagulants tested (FeCl 3 , TiCl 4 , TiCl 3 , <strong>and</strong> ZrCl 4 ). With alum as coagulant, silica<br />

significantly affected As(V) adsorption at pH 6.5-7.5, while it had no significant effect at<br />

pH 8.5. In the NSFI challenge water <strong>with</strong>out silica, the adsorption of As(III) increased<br />

<strong>with</strong> increasing pH for TiCl 4 , FeCl 3 <strong>and</strong> ZrCl 4 , whereas pH did not significantly affect<br />

As(III) adsorption on these coagulants in the st<strong>and</strong>ard challenge water <strong>with</strong> silica present.<br />

The detrimental effect of silica on As(III) removal increased <strong>with</strong> increasing pH for all<br />

coagulants except TiCl 3 .<br />

xxiv


14. In the NSFI challenge water <strong>with</strong> silica, phosphate was found to reduce the adsorption of<br />

As(V) significantly at pH 6.5 <strong>and</strong> 7.5, whereas it had a lesser effect at pH 8.5 <strong>with</strong> the<br />

coagulants tested (FeCl 3 , TiCl 3 , TiCl 4 , <strong>and</strong> ZrCl 4 ). However, <strong>with</strong> alum as coagulant,<br />

phosphate did not affect As(V) adsorption. In contrast to As(V) adsorption, the presence<br />

of phosphate did not significantly affect the adsorption of As(III) <strong>with</strong> the coagulants<br />

tested (FeCl 3 , TiCl 3 , TiCl 4 , <strong>and</strong> ZrCl 4 ).<br />

15. In the NSFI challenge water <strong>with</strong> silica <strong>and</strong> phosphate, vanadate did not significantly<br />

affect the adsorption of As(V) <strong>with</strong> the coagulants tested (FeCl 3 , alum, TiCl 3 , TiCl 4 , <strong>and</strong><br />

ZrCl 4 ). Similarly. in the NSFI challenge water <strong>with</strong> silica <strong>and</strong> phosphate, vanadate did<br />

not significantly affect the adsorption of As(III) <strong>with</strong> the coagulants tested (FeCl 3 , TiCl 3 ,<br />

TiCl 4 , <strong>and</strong> ZrCl 4 )<br />

The studies on the individual effects of competing ions (silica, phosphate <strong>and</strong> vanadate)<br />

on arsenic adsorption <strong>with</strong> FeCl 3 as coagulant resulted in the following conclusions:<br />

16. In the absence of other competing ions, it was found that silica, phosphate, <strong>and</strong> vanadate<br />

exhibited significant competitive effects on the adsorption of As(V) <strong>and</strong> As(III).<br />

17. In the absence of phosphate <strong>and</strong> vanadate, silica exhibited a significant effect on the<br />

adsorption of As(V) <strong>and</strong> As(III) <strong>and</strong> the effect increased <strong>with</strong> increasing pH.<br />

18. In the absence of silica <strong>and</strong> vanadate, phosphate exhibited a significant effect on the<br />

adsorption of As(V) at all pHs <strong>and</strong> a minor effect on the adsorption of As(III) at pH 7.5<br />

<strong>and</strong> 8.5.<br />

19. In the absence of silica <strong>and</strong> phosphate, vanadate exhibited a significant effect on the<br />

adsorption of As(V) <strong>and</strong> had a minor effect on the adsorption of As(III) at pH 7.5 <strong>and</strong> 8.5<br />

during FeCl 3 coagulation.<br />

20. Based on the FeCl 3 experimental results <strong>with</strong> <strong>and</strong> <strong>with</strong>out multiple competing<br />

contaminants, the following inferences were made:<br />

• The presence of silica significantly reduced the magnitude of the phosphate effect<br />

on As(V) adsorption at pH 7.5 <strong>and</strong> 8.5.<br />

• The presence of silica reduced the effect of phosphate in the case of As(III)<br />

adsorption at all pHs.<br />

• The combined presence of silica <strong>and</strong> phosphate reduced the effect of vanadate at<br />

all pHs in the case of As(V) <strong>and</strong> As(III) removal.<br />

The toxicity characteristics of the sludges produced during coagulation <strong>with</strong> FeCl 3 , alum,<br />

TiCl 3 , TiCl 4 , <strong>and</strong> ZrCl 4 led to the following conclusions:<br />

21. The adsorbent loading of As(V) was independent of initial arsenic concentration <strong>with</strong> the<br />

coagulants tested (FeCl 3 , alum, TiCl 3 , TiCl 4 , <strong>and</strong> ZrCl 4 ).<br />

22. Following coagulation <strong>and</strong> 3-hr settling, <strong>with</strong> the five coagulants tested (FeCl 3 , alum,<br />

xxv


TiCl 3 , TiCl 4 , <strong>and</strong> ZrCl 4 ), the liquid wastes obtained were found to contain less than 0.5%<br />

dry solids, <strong>and</strong> according to the TCLP regulations, did not require extraction. The liquid<br />

filtrate, which is considered to be the extract in such cases, easily passed the TCLP <strong>and</strong><br />

WET test regulatory limits of 5 mg/L.<br />

The attempt to model As(V)/As(III) adsorption during coagulation <strong>with</strong> FeCl 3 using the<br />

chemical equilibrium modeling program, Mineql+, led to the following conclusions:<br />

23. The unmodified Mineql+ chemical equilibrium modeling program could not simulate<br />

As(III) or As(V) adsorption onto Fe(OH) 3 (s) during FeCl 3 coagulation.<br />

24. The model predicts significantly less adsorption onto Fe(OH) 3 (s) compared <strong>with</strong><br />

experimental results, <strong>and</strong> the model also does not take into account the effect of silica,<br />

phosphate <strong>and</strong> vanadate on As(V)/As(III) adsorption in most cases.<br />

RECOMMENDATIONS<br />

The main purpose of this project was to determine the technical <strong>and</strong> economic feasibility<br />

of using aluminum, zirconium <strong>and</strong> titanium salts in comparison <strong>with</strong> ferric salt as coagulants for<br />

arsenic removal in coagulation-filtration processes. Based on adsorption <strong>and</strong> economic<br />

comparisons, this work showed that FeCl 3 was clearly superior to the other coagulants tested.<br />

Although Ti(III) had the highest removal efficiency for As(III), its chemical cost was not<br />

available <strong>and</strong> is expected to be higher considering the cost of Ti(IV) salts. A chemical cost<br />

comparison of commonly used coagulants showed that alum did not remove As(III) <strong>and</strong> was 4-6<br />

times more expensive than FeCl 3 for As(V) removal. Taking the detailed conclusions above into<br />

consideration, the following recommendations are made for application of coagulation for<br />

arsenic removal in small systems:<br />

1. FeCl 3 should be considered as the preferred coagulant for As(V) <strong>and</strong> As(III) removal at<br />

all pHs <strong>and</strong> background water compositions.<br />

2. Alum could be considered as a coagulant for As(V) removal at pH ≤6.5 where its As(V)<br />

capacity is closer to that of FeCl 3 .<br />

3. As(III) should be pre-oxidized to As(V) for cost-effective treatment <strong>with</strong> alum <strong>and</strong> FeCl 3<br />

coagulants.<br />

4. <strong>Zirconium</strong> <strong>and</strong>/or titanium could be considered as alternative coagulants to alum, if their<br />

prices drop significantly.<br />

xxvi


CHAPTER 1<br />

INTRODUCTION<br />

ARSENIC OCCURRENCE, HEALTH EFFECTS AND REGULATIONS<br />

Inorganic arsenic is considered as a human carcinogen <strong>with</strong> multiple sites of attack.<br />

Epidemiological studies have demonstrated the higher risks of skin, bladder, lung, liver <strong>and</strong><br />

kidney cancer along <strong>with</strong> other non-cancerous health effects that result from continued<br />

consumption of elevated levels of arsenic in drinking water (Chen et al. 1988, Guha Mazumder<br />

et al. 1998, Ferreccio et al. 2000). Due to the elevated health risk, based on the analysis of the<br />

United States Environmental Protection Agency (USEPA) <strong>and</strong> two independent reports by the<br />

National Research Council (NRC), USEPA has reduced the maximum contamination level<br />

(MCL) of arsenic in drinking water from 50 to 10 µg/L (USEPA 2001).<br />

<strong>Arsenic</strong> is a widely distributed element to which humans are exposed through ingestion<br />

of water <strong>and</strong> food, <strong>and</strong> by inhalation. People in many countries around the world are exposed to<br />

elevated levels of arsenic in their drinking water. The International Agency for Research on<br />

Cancer as well as the United States Environmental Protection Agency (USEPA) has designated<br />

arsenic as a Group A "known" human carcinogen. There are numerous reports in the literature,<br />

based on past <strong>and</strong> ongoing experience in various countries in Asia <strong>and</strong> South America<br />

concerning the high risk of skin, bladder, lung, liver, <strong>and</strong> kidney cancer along <strong>with</strong> other noncancerous<br />

ailments that result from continued consumption of elevated levels of arsenic in<br />

drinking water. Indeed, arsenic is the only major demonstrated human carcinogen where the<br />

principle route of human exposure is through drinking water. The most devastating incidence of<br />

arsenic poisoning has been reported in Bangladesh, <strong>and</strong> West Bengal-India (Chowdhury et al.<br />

2000, Chowdhury et al. 2000a, Das et al. 1995). Reportedly, more than 6 million people in West<br />

Bengal, India, <strong>and</strong> more than 70 million people in Bangladesh are drinking ground water<br />

containing elevated levels (50 µg/L <strong>and</strong> above) of arsenic (Chakraborti et al. 2002). In<br />

Bangladesh out of an estimated 6-11 million shallow tube wells, approximately 27% are<br />

contaminated <strong>with</strong> arsenic above 50 µg/L (Kinniburgh <strong>and</strong> Kosmus 2002). Here in the USA,<br />

according to the US Geological Survey (USGS) <strong>and</strong> the USEPA, high concentrations of arsenic<br />

are widespread in Western, Midwestern <strong>and</strong> Northwestern United States<br />

(http://co.water.usgs.gov/trace/arsenic). The EPA has reported that more than 5.5% of the total<br />

water supply systems in USA contain arsenic at a level greater than the present limit (10 µg/L).<br />

Higher levels of arsenic tend to be found more often in groundwater than in surface water<br />

sources. About 4,100 of the nation’s 54,000 Community Water Supplies <strong>and</strong> 1,100 of the 20,000<br />

Non-Transient Non-Community Water Supplies exceed the current 10 µg/L limit. According to<br />

the Natural Resources Defense Council (NRDC 2000), over 34 million Americans drink water<br />

that increases the risk of arsenic-related cancer.<br />

IRON COAGULATION-FILTRATION AND ADSORPTION PROCESSES<br />

The most frequently used technology for removing arsenic from drinking water is<br />

precipitation/coprecipitation followed by some form of settling <strong>and</strong>/or filtration. The process is<br />

commonly referred to as coagulation-filtration. This technology is capable of treating a wide<br />

range of influent concentrations to the revised MCL for arsenic. According to USEPA more than<br />

52% of the identified applications of arsenic treatment technologies for water are based on<br />

1


coagulation. Many studies have been done to examine the efficiency of arsenic removal using<br />

coagulation <strong>with</strong> ferric <strong>and</strong> aluminum salts (Hering et al. 1996, Gulledge et al. 1973, Edwards<br />

1994). Coagulant type <strong>and</strong> dosage, pH, composition of the water, <strong>and</strong> contaminant type have<br />

significant effects on removal efficiency. It is well accepted that removal of arsenate [As(V)] is<br />

much better than arsenite [As(III)] (Clifford et al. 1990) <strong>and</strong> that silica interferes <strong>with</strong> arsenic<br />

removal at higher pH (Tong 1997). For better removal efficiency, pre-oxidation is necessary for<br />

As(III) removal.<br />

Small-scale systems <strong>and</strong> point-of-entry (POE) systems often use adsorption or ionexchange<br />

(IX) in packed-bed processes for arsenic removal, especially when arsenic is the only<br />

contaminant to be removed. As <strong>with</strong> iron <strong>and</strong> aluminum coagulation, numerous factors including<br />

arsenic oxidation state (III or V), pH, competing anions, media particle size, <strong>and</strong> empty bed<br />

contact time (EBCT) significantly affect arsenic removal by adsorption <strong>and</strong> IX. However, the<br />

major factors limiting the use of adsorption <strong>and</strong> IX processes include the high cost of the media,<br />

the complexity of regeneration, backwash <strong>and</strong> spent-regenerant disposal <strong>and</strong> spent media<br />

disposal. Due to high arsenic concentrations in the backwash waters <strong>and</strong> spent regenerant, direct<br />

discharge to a sanitary sewer is usually not acceptable. Therefore, the spent regenerants may<br />

need to be further treated by a precipitation/coagulation process to produce a sludge that can be<br />

thickened <strong>and</strong> dried prior to disposal. On the other h<strong>and</strong>, the arsenic-laden sludge produced by<br />

coagulation processes does not require an additional precipitation step, <strong>and</strong> can be directly dried<br />

<strong>and</strong> l<strong>and</strong> filled in hazardous or non-hazardous l<strong>and</strong> fills depending on arsenic concentration.<br />

BACKGROUND OF THE STUDY<br />

Significance of In-situ Formed Hydroxides in <strong>Arsenic</strong> Removal<br />

Previous studies conducted by researchers at the University of Houston (Clifford et al.<br />

1997, Ghurye et al. 2004) reported that <strong>Coagulation</strong>-Microfiltration <strong>with</strong> FeCl 3 was effective <strong>and</strong><br />

economical compared <strong>with</strong> adsorption <strong>and</strong> ion exchange. It was observed that pH <strong>and</strong> ferric dose<br />

were the most important variables controlling arsenic removal. Experiments were also performed<br />

that compared hydrolyzed in-situ Fe(OH) 3 (s) <strong>with</strong> preformed ferric hydroxide. Based on arsenic<br />

uptake by the iron oxide/hydroxide surfaces, preformed Fe(OH) 3 (s) was not as effective as ferric<br />

hydroxide formed-in-place (hydrolyzed in situ) by adding FeCl 3 to water <strong>and</strong> mixing it well. The<br />

reason for the higher arsenic capacity on the formed-in-place coagulant is that the arsenate<br />

anions are sorbed by surface complexation onto the short-chain polymers (oligomers) of<br />

Fe x (OH) y z+ as they are forming into Fe(OH) 3 (s) floc particles, which can be filtered. Preformed<br />

Fe(OH) 3 <strong>and</strong> granular ferric oxide/hydroxide (GFO/GFH) media simply do not have the<br />

available surface area in comparison to the oligomers <strong>and</strong> polymers of Fe(OH) 3 (s) that are<br />

formed during Fe 3+ hydrolysis in coagulation processes. This leads to much higher arsenic<br />

loading on the coagulant <strong>and</strong> a smaller volume of Fe(OH) 3 (s) solids to be wasted in coagulation<br />

compared <strong>with</strong> adsorption onto GFH/GFO. When the column performance of GFH is compared<br />

<strong>with</strong> coagulated Fe(OH) 3 (s), an even greater difference in As(V) loading was observed.<br />

Alum as Coagulant for <strong>Arsenic</strong> Removal<br />

Alum is most widely used coagulant for water treatment in the USA, <strong>and</strong> it has been<br />

reported to be equally as effective as ferric iron for arsenic adsorption on a molar basis.<br />

2


However, this reported equality of iron <strong>and</strong> alum is questionable, <strong>and</strong> has not been shown in<br />

direct coagulant comparisons. Alum, like ferric chloride, is an effective coagulant at pH < 6.5,<br />

because it carries a strong cationic charge <strong>with</strong>in this pH region. At pH > 6.5, alum is only<br />

weakly cationic, <strong>and</strong> becomes much less effective for adsorbing anions. Sorg <strong>and</strong> Logsdon<br />

(1978) demonstrated that arsenic removal <strong>with</strong> alum coagulation is most effective at pH 5 to 7.<br />

Edwards (1994) reported that ferric <strong>and</strong> alum coagulation are equally effective on a molar basis<br />

<strong>and</strong> at significant coagulant dosages, As(V) removal was similar for both alum <strong>and</strong> ferric<br />

coagulants at pH 7.6 or lower. Ahmed <strong>and</strong> Rahaman (2000) also reported effective removal <strong>with</strong><br />

alum coagulation in the pH range of 7.2-7.5. McNeil & Edwards (1997) reported that aluminum<br />

<strong>and</strong> iron flocs have equal adsorption capacity, but the aluminum hydroxide flocs <strong>with</strong> sorbed<br />

arsenic can pass through 0.45-μm filters <strong>and</strong> decrease apparent arsenic removal by alum<br />

coagulation. If microfiltration pore size influences arsenic removal, then arsenic removal by<br />

alum coagulation followed by 0.2-μm membrane filtration as used in this research should<br />

produce improved arsenic removal. As will be seen, this was not the case.<br />

<strong>Zirconium</strong> Oxide/Hydroxide Adsorbent Media for <strong>Arsenic</strong> Removal<br />

<strong>Zirconium</strong>-based adsorbents for arsenic removal have appeared in the literature <strong>and</strong> in the<br />

market place. The tests have verified that zirconium-based dry powder adsorbent media are<br />

effective for arsenic removal from drinking water <strong>and</strong> have significant advantages over the<br />

highest quality alumina. <strong>Zirconium</strong> based adsorbents have been reported to have high capacity<br />

for arsenic (Manna et al, 1999; Suzuki et al, 2000; Zhu et al, 2001; Clarke et al, 2002; Daus et al,<br />

2004). Similarly hydrated titanium oxide/hydroxide powders <strong>and</strong> nanocrystalline TiO 2 have also<br />

been shown to have high adsorption capacity for As(III) <strong>and</strong> As(V) in the pH 6.5 – 8.6 range<br />

(Bissen et al, 2001; Lee et al, 2002; Dutta et al, 2004; Bang et al, 2005; Ferguson et al, 2005;<br />

Meng et al, 2005; Pena et al, 2005). Thus, in addition to iron-based media, zirconium <strong>and</strong><br />

titanium based media are now being used for arsenic removal.<br />

Basis of the Project<br />

It has been shown that hydrolyzed in situ Fe(III) coagulants have a significant arseniccapacity<br />

advantage over preformed coagulants <strong>and</strong> granular media such as GFH <strong>and</strong> GFO. It has<br />

also been shown that zirconium oxide/hydroxide <strong>and</strong> titanium oxide/hydroxide powders <strong>and</strong> fine<br />

crystals are good-excellent adsorbents for arsenic. Because of the higher arsenic capacity of<br />

coagulants <strong>and</strong> the growing popularity of zirconium <strong>and</strong> titanium oxyhydroxides for arsenic<br />

removal a comparative study of iron vs. titanium <strong>and</strong> zirconium coagulants seemed warranted.<br />

The higher-arsenic-loading advantage of in-situ formed iron hydroxides during coagulation<br />

compared <strong>with</strong> sorption onto iron hydrous oxide/hydroxides was also expected to occur <strong>with</strong><br />

alternative metal coagulants such as zirconium <strong>and</strong> titanium. Alum was also studied because of<br />

(a) its wide used as a coagulant for water treatment, <strong>and</strong> (b) its reportedly equal effectiveness for<br />

As(V) removal on a molar basis compared <strong>with</strong> iron-based coagulant. The suggested equal<br />

effectiveness of alum <strong>and</strong> iron coagulants for arsenic removal has been called into question by<br />

recent comparisons of iron- <strong>and</strong> aluminum-based sorbents (Wu, 2001) that demonstrate that ironbased<br />

sorbents have far greater As(V) capacity compared <strong>with</strong> activated alumina. Thus it is<br />

logical to do a comparative study of the soluble aluminum, zirconium <strong>and</strong> titanium salts as<br />

coagulants in comparison <strong>with</strong> iron salt for As(III) <strong>and</strong> As(V) removal. As is the case <strong>with</strong><br />

3


Fe(OH) 3 (s), the arsenic-adsorption capacity of Al, Zr <strong>and</strong> Ti oxides/hydroxides will depend on<br />

the presence of hydroxyl groups. These groups are highly reactive, <strong>and</strong> the loss of hydroxyl<br />

groups leads to a loss of surface complexation sites such that the media becomes less efficient.<br />

Hydrolyzed-in-place (coagulated) Al, Zr <strong>and</strong> Ti hydroxides will have a greater number of<br />

available hydroxyl groups compared <strong>with</strong> the crystalline or powder compounds. So, the<br />

adsorption capacity of crystalline or powder materials will be lower than the freshly prepared Al,<br />

Zr <strong>and</strong> Ti amorphous oxides/hydroxides. Additional disadvantages of granular <strong>and</strong> crystalline<br />

media such as poor kinetics of adsorption <strong>and</strong> the dependence of adsorption capacity on crystal<br />

structure can be overcome in the coagulation process.<br />

Although zirconium <strong>and</strong> titanium may be considered rare <strong>and</strong> expensive elements by<br />

professionals in the water utility industry, such is not the case as titanium <strong>and</strong> zirconium are the<br />

10th <strong>and</strong> the 20th most-abundant chemical in the earth’s crust, respectively. In fact, titanium is<br />

more common than chlorine, phosphorus, manganese, sulfur, <strong>and</strong> chromium. <strong>Zirconium</strong> is nearly<br />

as abundant as fluorine <strong>and</strong> nitrogen (Masterson et al. 1973). As already pointed out, zirconium<strong>and</strong><br />

titanium-based media are already being marketed for arsenic removal from drinking water in<br />

small systems <strong>and</strong> POU/POE systems. With regard to coagulation, zirconium <strong>and</strong> titanium<br />

chemicals are available as water hydrolysable salts including TiCl 4 , TiCl 3 , TiOCl 2 , ZrCl 4 , <strong>and</strong><br />

ZrOCl 2 which could be used for studies. Aqueous solutions of these chemicals are also available<br />

commercially.<br />

OBJECTIVES<br />

The objective of this work was to determine the technical <strong>and</strong> economic feasibility of<br />

using aluminum, zirconium <strong>and</strong> titanium salts in comparison <strong>with</strong> ferric salt as coagulants for<br />

arsenic removal in coagulation filtration processes. To attain the main objective, the following<br />

specific tasks were carried out:<br />

1. Study the removal efficiency of As(V) <strong>and</strong> As(III) in NSFI-53 challenge water <strong>with</strong><br />

aluminum, titanium <strong>and</strong> zirconium coagulants <strong>and</strong> compare the removal efficiency <strong>with</strong><br />

FeCl 3 .<br />

2. Study the effect of pH on As(III) <strong>and</strong> As(V) removal <strong>with</strong> each of the coagulants.<br />

3. Study the effect of competing ions such as silicate, phosphate, <strong>and</strong> vanadate on As (III)<br />

<strong>and</strong> As(V) removal in NSFI challenge water.<br />

4. Study the individual effect of competing ions such as silicate, phosphate, <strong>and</strong> vanadate on<br />

As(III) <strong>and</strong> As(V) removal.<br />

5. Test the toxicity characteristics of the sludge produced <strong>with</strong> each coagulant metal <strong>and</strong><br />

check <strong>with</strong> TCLP <strong>and</strong> WET sludge test regulatory limits.<br />

6. Apply readily available surface complexation models to quantitatively predict the arsenic<br />

adsorption <strong>and</strong> compare the model results <strong>with</strong> experimental results.<br />

ANTICIPATED PRACTICAL BENEFITS OF THE PROJECT<br />

Alum <strong>and</strong> ferric chloride are the most commonly used coagulants in drinking water<br />

treatment. Alum is the coagulant preferred by many water utility managers <strong>and</strong> operators because<br />

of industry experience <strong>with</strong> alum, as well as its effectiveness, availability, purity, <strong>and</strong> cost. The<br />

controversies surrounding the reportedly equal-arsenic-removal effectiveness of alum <strong>and</strong> iron<br />

4


coagulants on a molar basis will be resolved. The water supply community will benefit from the<br />

results of this innovative coagulant research by having access to potentially more effective<br />

zirconium <strong>and</strong> titanium coagulants for arsenic removal. Although these metals are not well<br />

known in water treatment, they are 10 th <strong>and</strong> 20 th in occurrence in the earth’s crust, i.e., very<br />

common <strong>and</strong> potentially low cost, <strong>and</strong> non-toxic by nature. Prior to carrying out this research it<br />

was expected that zirconium <strong>and</strong>/or titanium coagulants would have the following advantages<br />

over ferric coagulants: (As will be seen, most of these advantages could not be demonstrated.)<br />

(1) Higher As(III) <strong>and</strong> As(V) loadings on the Zr(OH) 4 (s), Ti(OH) 3 (s), Ti(OH) 4 (s)<br />

precipitates compared <strong>with</strong> Fe(OH) 3 (s). This means less solid waste to send to l<strong>and</strong>fills.<br />

(2) Lower costs for coagulation <strong>and</strong> filtration resulting from lower doses of coagulants.<br />

(3) Less sensitivity to pH <strong>and</strong> interference by competing ions including: hydroxide,<br />

silicate, vanadate, <strong>and</strong> phosphate.<br />

(4) Potentially simpler process design by the elimination of the As(III) oxidation step for<br />

coagulants that adsorb As(III) as well as or better than As(V). The claim of better<br />

As(III) adsorption has been made for titanium oxide/hydroxide solid media.<br />

(5) Less sensitivity to reducing (low E H ) conditions in the l<strong>and</strong>fills that ultimately receive<br />

the arsenic-contaminated sludge. Fe(OH) 3 (s) is susceptible to reduction to Fe(II) <strong>and</strong><br />

As(V) <strong>and</strong> dissolution of the precipitate <strong>with</strong> the subsequent release of arsenic. It is<br />

anticipated that Zr(IV), Ti(IV) <strong>and</strong> Ti(III) hydroxides will not be easily reduced, <strong>and</strong><br />

even if they are, the resulting hydroxides will remain insoluble <strong>and</strong> prevent the release<br />

of adsorbed arsenic III <strong>and</strong> V. It is also expected that these innovative hydroxides will<br />

fare better in passing the California WET procedure for arsenic-contaminated sludge.<br />

(6) Faster-settling sludge in the case of denser Zr(OH) 4 (s) in the event settling is used as a<br />

pretreatment for filtration.<br />

(7) Less-staining potential of white as-opposed-to the red-brown sludge that is produced by<br />

ferric coagulation.<br />

5


REAGENTS AND STOCKS<br />

CHAPTER 2<br />

MATERIALS AND METHODS<br />

All reagents used were of analytical reagent grade. Primary st<strong>and</strong>ards of 100 mg As/L of<br />

each species were prepared from arsenic trioxide (As 2 O 3 ) for As(III) <strong>and</strong> sodium arsenate for<br />

As(V) (Na 2 HAsO 4 ) both from Sigma Chemical Co, Mo. The As(V) <strong>and</strong> As(III) stock solutions<br />

(100 mg/L) were prepared <strong>and</strong> stored in bottles. The stock solutions were then used for spiking<br />

the arsenic into the challenge water. Working st<strong>and</strong>ard solutions were prepared daily <strong>with</strong> proper<br />

dilution.<br />

To prepare challenge water, the following salts were used NaNO 3 , NaHCO 3 ,<br />

Na 2 HPO 4·H 2 O, NaF, Na 2 SiO 3·9H 2 O, MgSO 4·7H 2 O, CaCl 2·2H 2 O, <strong>and</strong> NaVO 3 . All the chemicals<br />

were purchased from Sigma Chemical Co or EM Science. Concentrated stocks were prepared<br />

from these salts <strong>and</strong> used for the preparation of challenge water on the day of experiment.<br />

The NSFI challenge water <strong>and</strong> the coagulants were prepared on the day of arsenic<br />

analysis. Citric/citrate buffer solution was prepared using 2 M citric acid <strong>and</strong> pH was adjusted to<br />

5.0 using NaOH. A 4-mg/mL solution of L-cysteine (Sigma Chemical Co, Mo.) in HCl solution<br />

was used to reduce As(V) to As(III). Sodium tetrahydroborate (EM Science, Germany) solutions<br />

were prepared fresh daily, <strong>and</strong> were supplemented <strong>with</strong> sodium hydroxide.<br />

PREPARATION OF NSFI-53 CHALLENGE WATER<br />

For all coagulation studies, NSFI-53 challenge water (hereafter referred to as challenge<br />

water) whose composition is given in Table 2.1 was prepared <strong>and</strong> used as the background water<br />

for the isotherm tests This water contains realistic concentrations of background contaminants<br />

such as silica, sulfate, phosphate, fluoride, <strong>and</strong> hardness, which are known to affect the arsenic<br />

capacity of adsorbents. Stability of the challenge water was not an issue because it was prepared<br />

on the day of experiment, <strong>and</strong> studies conducted at University of Houston showed that the<br />

challenge water was stable (did not precipitate calcium) for at least a week (Tripp, 2000;<br />

Swaminathan, 2005). The pH of the synthetic groundwater was adjusted by using dilute HCl or<br />

NaOH solution. The required amount of As(III) or As(V) was spiked to the challenge water for<br />

the coagulation studies.<br />

PREPARATION OF COAGULANT STOCK<br />

The coagulants used in this study were reagent-grade anhydrous ferric chloride (FeCl 3 ),<br />

alum (Al 2 (SO 4 ) 3·18H 2 O), anhydrous zirconium (IV) chloride (ZrCl 4 ), titanium (IV) chloride<br />

solution (TiCl 4 ), titanium (III) chloride solution (TiCl 3 ), titanium (IV) oxychloride (TiOCl 2 ), <strong>and</strong><br />

zirconium (IV) oxychloride (ZrOCl 2 ). Table 2.2 lists the approximate amount of salts used for<br />

the preparation of 1000 mg/L of the concentrated stock. Stock solutions were prepared for a<br />

concentration of 1000 mg/L <strong>with</strong> DI water except in the case of TiCl 4 . A solution of 1000 mg/L<br />

titanium (IV) chloride was prepared by adding TiCl 4 solution to 1-L solution of 0.25 N<br />

hydrochloric acid solution to avoid the precipitation of meta-titanic acid in the stock solution.<br />

The stock solutions were prepared on the day of the experiment <strong>and</strong> the exact concentration of<br />

the coagulant metals in the stock were then analyzed using ICP or AAS.<br />

7


Table 2.1<br />

Composition of NSFI 53 challenge water<br />

Cation MW meq/L mg/L Anion MW meq/L mg/L<br />

Ca 2+ 40.1 2.00 40.1<br />

-<br />

HCO 3 61.0 3.00 183.<br />

Mg 2+ 24.3 1.04 12.6<br />

2-<br />

SO 4 96.1 1.04 50.0<br />

Na + 23.0 3.86 88.9 Cl - 35.5 2.00 71.0<br />

NO 3 -N 14.0 0.143 2.00<br />

F - 19.0 0.053 1.00<br />

PO 4 -P 31.0 0.001 0.04<br />

SiO 3 -SiO 2 60.1 0.67 20.0<br />

As(III)/(V) 74.9 0.05<br />

Σ = 6.90 142 Σ = 6.90 327<br />

Table 2.2<br />

Composition of reagents in preparation of coagulant stock for coagulation study<br />

Reagents<br />

Target concentration in<br />

stock<br />

Amount of reagent added<br />

Solid FeCl 3 1000 mg/L of Fe 3+ 2.9 g/L in DIW<br />

Solid Alum 1000 mg/L of Al 3+ 12.34 g/L in DIW<br />

TiCl 4 , Soln.* 1000 mg/L of Ti 4+ 2.3 ml/L in 0.25N HCl<br />

Solid ZrCl 4 1000 mg/L of Zr 4+ 2.555 g/L in DIW<br />

TiCl 3 , Soln.* 1000 mg/L of Ti 3+ 26.83 ml/L in DIW<br />

TiOCl 2 , Soln.* 1000 mg/L of Ti 4+ 2.3 ml/L in DIW<br />

Solid<br />

ZrOCl 2 .8H 2 O<br />

1000 mg/L of Zr 4+ 3.54 g/L in DIW<br />

* Note: The Ti concentrations of the original solutions were verified by flame AAS analysis.<br />

COAGULATION-FILTRATION PROCEDURE<br />

NSFI 53 challenge water spiked <strong>with</strong> 50 µg/L of As(III) or As(V) was prepared, <strong>and</strong> the<br />

pH of the challenge water was adjusted to the desired pH in the 6.5 to 8.5 range by using dilute<br />

HCl <strong>and</strong> NaOH solutions. Since the coagulants used were highly acidic, they reduced the pH of<br />

the challenge water after dosing, thus it was necessary to bring the challenge water to the initial<br />

pH after addition of coagulant. Pre-titrations of the challenge water dosed <strong>with</strong> varying amounts<br />

of coagulant were performed to determine the amount of sodium hydroxide (NaOH) solution<br />

required to achieve the desired equilibrium pHs at the different dosages.<br />

The coagulation study was conducted using batch-scale jar tests. A six-position flat-blade<br />

stirring apparatus manufactured by Phipps <strong>and</strong> Bird, USA was used. 2-L square glass jars were<br />

8


used for each experiment. Before each experiment, 1-L of challenge water adjusted to a specific<br />

pH (6.5, 7.5, or 8.5) was added to each jar. The coagulant dose used for As(V) <strong>and</strong> As(III)<br />

removal was in the range of 0 to 10 mg/L of the metal. With the stirring apparatus preset at the<br />

rapid mix speed of 100 rpm (G = 100 s -1 ), the stirring apparatus was started <strong>and</strong> a predetermined<br />

amount of coagulant was added to each jar. Immediately after addition of coagulant, the<br />

predetermined amount of NaOH solution was added to restore the challenge water to the initial<br />

pH. After rapid mixing for 1 min, the jars were stirred slowly at 20 rpm (G = 14 s -1 ) for 20 min,<br />

<strong>and</strong> then the precipitate was allowed to settle for 1 hour. During slow mixing, the pH of<br />

challenge water was checked <strong>and</strong> adjusted to the initial pH by adding more NaOH, if necessary.<br />

The final pH was measured after the settling period, <strong>and</strong> was typically <strong>with</strong>in ± 0.15 units of the<br />

initial pH. Figure 2.1 shows the steps involved in the coagulation-filtration process.<br />

PRESERVATION AND SPECIATION<br />

After settling, a supernatant water sample was collected using a 30-mL polyethylene<br />

syringe, <strong>and</strong> then the water sample was filtered through a 0.2-μm pore size filter. Samples (30-<br />

mL) were collected in 50-mL centrifuge tubes <strong>and</strong> acidified <strong>with</strong> 30 μL of concentrated nitric<br />

acid to measure total arsenic (final conc. 0.1%, pH < 2). All samples were analyzed <strong>with</strong>in a<br />

week. For arsenic(III) measurement, 10-mL filtered sample was collected in a 15-mL centrifuge<br />

tube containing 0.432 mL of 2 M acetic acid <strong>and</strong> 0.134 mL of 0.1 M EDTA as required by the<br />

recently developed field speciation method for inorganic arsenic speciation (Clifford et al. 2004).<br />

Rapid mix at 100 rpm for 1-min then<br />

flocculate at 20 rpm for 20-min<br />

Settle for 1 hr<br />

Jar test apparatus<br />

Analyze<br />

sample by<br />

FI-HG-AAS<br />

Preserve <strong>with</strong><br />

HNO 3 for total<br />

<strong>Arsenic</strong> analysis<br />

Preserve <strong>with</strong><br />

EDTA-Acetic acid for<br />

<strong>Arsenic</strong>(III) analysis<br />

Filter supernatant<br />

(0.2 μm filter)<br />

Figure 2.1 <strong>Coagulation</strong>-filtration procedure<br />

9


Table 2.3<br />

Experimental conditions for the determination of As(III) <strong>and</strong> As(Tot) by FI-HG-AAS<br />

Parameters Perkin-Elmer (Zeeman 5000)<br />

Lamp<br />

8 W EDL<br />

Wavelength<br />

193.7 nm<br />

Slit<br />

0.7 nm(low)<br />

Sample volume 500 µL<br />

HCl concentration<br />

0.02 M<br />

Citric/citrate buffer (pH 5.0) 2.0 M<br />

HCl/citrate buffer flow rate 8 ml/min<br />

NaBH 4 concentration (for As(Tot)) 0.4% in 0.2% NaOH<br />

NaBH 4 concentration [for As(III)] 0.2% in 0.05% NaOH<br />

NaBH 4 flow rate<br />

5 ml/min<br />

Carrier gas<br />

Argon<br />

Carrier gas flow rate<br />

60-70 ml/min<br />

Quartz cell temp<br />

~900°C, Electrically heated<br />

INSTRUMENTATION<br />

The arsenic analyses were carried out by flow injection hydride-generation atomic<br />

absorption spectroscopy (FI-HG-AAS) using a Perkin-Elmer (Model Zeeman 5000) atomic<br />

absorption spectrometer (AAS) equipped <strong>with</strong> an electrodeless discharge lamp (EDL) operated at<br />

8W from an external power supply. The AAS was coupled <strong>with</strong> a Perkin-Elmer Flow Injection<br />

(FIAS-100) unit for hydride generation for the determinations of As(III) <strong>and</strong> As(Tot). Detail<br />

descriptions of the instrumental conditions are given in Table 2.3. The samples for As(Tot) <strong>and</strong><br />

speciated As(III) were analyzed by FI-HG-AAS.<br />

Total iron, aluminum, <strong>and</strong> titanium analyses were performed by flame atomic absorbance<br />

spectrometry (Flame AAS). Samples were analyzed by direct aspiration into an air-acetylene<br />

flame in the case of iron analysis <strong>and</strong> into a nitrous oxide-acetylene flame in the case of titanium<br />

<strong>and</strong> aluminum analyses. <strong>Zirconium</strong> analysis was performed using an Inductively Coupled<br />

Plasma-Mass Spectrometer (ICP-MS). The stock solutions were diluted <strong>and</strong> acidified before<br />

analysis, <strong>and</strong> the calibration st<strong>and</strong>ards were prepared in the same matrix as that of the sample.<br />

TOTAL ARSENIC AND ARSENIC AND ARSENIC(III) ANALYSIS USING FI- HG-AAS<br />

To determine total arsenic (As(Tot) = As(III) + As(V)), samples were treated <strong>with</strong> L-<br />

cysteine in 2 M HCl. The samples were kept for 15 minutes at room temperature for the<br />

reduction of As(V) to As(III) <strong>and</strong> then diluted to such a volume as to maintain the concentrations<br />

of L-cysteine, <strong>and</strong> acid to 4 mg/mL, <strong>and</strong> 0.02 M, respectively. The arsenic concentration was<br />

measured by FI-HG-AAS against arsenic st<strong>and</strong>ards prepared as samples. Five st<strong>and</strong>ards were<br />

prepared in the range of 0 to 6 μg/L arsenic, <strong>and</strong> the st<strong>and</strong>ards were periodically checked during<br />

analysis. Samples were diluted to the range of 1 to 6 μg/L arsenic. The arsenic concentration of<br />

the sample was then obtained from the calibration curve, derived by analysis of the absorbance<br />

10


of the five different arsenic st<strong>and</strong>ards. Reduced arsenic sample was injected by means of a rotary<br />

valve fitted <strong>with</strong> a 500µL sample loop into the stream of 0.02 M HCl solution which was flowing<br />

at 8 mL/min. The injected sample together <strong>with</strong> carrier solution, met subsequently <strong>with</strong> a<br />

continuous stream of 0.4% sodium tetrahydroborate in 0.2% NaOH flowing at 5 mL/min. After<br />

mixing <strong>with</strong> sodium tetrahydroborate, the generated hydride (AsH 3 ) subsequently entered into<br />

the gas-liquid separator. Inside this apparatus a continuous flow of argon carrier gas (70-80<br />

mL/min) carried the hydride to the quartz tube fitted on an electrically heated heater at 900 ºC.<br />

<strong>Arsenic</strong> adsorption was measured at 193.7 nm. At least triplicate measurements were<br />

made for each sample <strong>and</strong> st<strong>and</strong>ard. For the determination of As(III) in the presence of As(V),<br />

the samples were not treated <strong>with</strong> L-cysteine <strong>and</strong> the carrier HCl solution was replaced by 2 M<br />

citric/citrate buffer of pH 5.0. Arsine was generated using 0.2% sodium tetrahydroborate in<br />

0.05% NaOH. Under this condition only As(III) generates AsH 3 <strong>and</strong> As(V) concentration is not<br />

measured. The As(V) was then calculated from the difference of As(Total) <strong>and</strong> As(III).<br />

FREUNDLICH ISOTHERM<br />

Equilibrium isotherms are constant temperature plots of the mass of contaminant<br />

adsorbed per unit mass of adsorbent versus the concentration of the contaminant. In order to use<br />

isotherms to estimate the mass adsorbed, equilibrium must be reached between the sorbent <strong>and</strong><br />

the sorbate. Although the Freundlich isotherm equation has a theoretical basis <strong>and</strong> can be derived<br />

assuming a Boltzmann distribution of adsorption site energies <strong>and</strong> multi-layer adsorption, it is<br />

used here only as an empirical curve fitting technique. The Freundlich equation is shown in<br />

Equation 2.1.<br />

q e<br />

KC e<br />

1/ n<br />

= (2.1)<br />

where q e = Mass of arsenic adsorbed per mass of the adsorbent, μg/mg<br />

K = Freundlich constant indicative of adsorption capacity of adsorbent, L/mg<br />

C e = Equilibrium concentration of arsenic in the liquid phase, μg/L<br />

1/n = Freundlich exponent, a constant<br />

Plotting the Freundlich isotherms for various pH <strong>and</strong> competing-ion conditions allows a<br />

quick comparison of the effectiveness of arsenic removal under the various pH <strong>and</strong> competing<br />

ion conditions.<br />

11


CHAPTER 3<br />

As(III) VS. As(V) REMOVAL<br />

INTRODUCTION<br />

It has been found in several studies of arsenic occurrence that only, inorganic arsenite<br />

(As(III)), <strong>and</strong> arsenate (As(V)) are important in ground water supplies (Andrae 1977, Irgolic<br />

1982, etc). As(V) is the thermodynamically stable species <strong>and</strong> As(III) is the metastable species.<br />

In the natural pH range of 6 to 9, As(III) is mostly present as H 3 AsO 3 , while As(V) exists either<br />

as H 2 AsO 4 - or HAsO 4 2- . Table 3.1 presents the equilibrium constants (pK a ’s) of arsenic <strong>and</strong><br />

arsenious acids. As(V), the oxidized form of arsenic is more easily removed from water by ion<br />

exchange, adsorption, <strong>and</strong> membrane processes compared <strong>with</strong> As(III), which, in general,<br />

requires pre-oxidation.<br />

Various mechanisms explain contaminant removal in the coagulation process. The major<br />

mechanisms are as follows: (1) compression of the double layer, (2) adsorption to produce<br />

charge neutralization, (3) enmeshment in precipitates, (4) adsorption to permit interparticle<br />

bridging, (5) surface precipitation, (6) lig<strong>and</strong> exchange-surface complexation, <strong>and</strong> (7) hydrogen<br />

bonding.<br />

In coagulation treatment, trace inorganic contaminants, are removed by the sorption onto<br />

the surfaces of freshly formed metal hydroxides/oxyhydroxides (M x (OH) y + ), which are formed<br />

upon metal coagulant addition <strong>and</strong> hydrolysis reaction. Stumm (1992) reported anion removal by<br />

metal oxyhydroxide surfaces by the process of lig<strong>and</strong> exchange (Figure 3.1). During lig<strong>and</strong><br />

exchange, the lig<strong>and</strong> H 2 AsO 4 - replaces two hydroxides <strong>and</strong> forms a bidendate surface complex<br />

on the metal oxyhydroxide surface. This reaction is facilitated by low pH <strong>and</strong> excess hydrogen<br />

ions, which consume the hydroxides released. Silicate, phosphate, <strong>and</strong> vanadate, three common<br />

competing anions that could compete for adsorption sites are shown in the figure. The surface<br />

complex formed between the metal <strong>and</strong> the arsenate anion is very strong <strong>and</strong> not easily reversed<br />

except at high pH when the reaction is reversed due to the presence of hydroxide ions, which are<br />

lig<strong>and</strong>s highly preferred by the central-atom metals.<br />

Table 3.1<br />

Equilibrium constants for arsenic <strong>and</strong> arsenious acids<br />

Concentration of species<br />

<strong>Arsenic</strong> pK a values Species<br />

(μmol/L)<br />

pH 6.5 pH 7.5 pH 8.5<br />

<strong>Arsenic</strong> acid,<br />

C T = 0.67 μM (50 μg/L)<br />

Arsenous acid,<br />

C T = 0.67 μM (50 μg/L)<br />

Source: Schecher 1998.<br />

pK a1 = 2.26<br />

pK a2 = 6.79<br />

pK a3 = 11.29<br />

pK a1 = 9.23<br />

pK a2 = 12.1<br />

H 2 AsO 4<br />

-<br />

HAsO 4<br />

2-<br />

0.5 0.15 0.02<br />

0.17 0.52 0.65<br />

H 3 AsO 3 0.67 0.66 0.57<br />

H 2 AsO 3<br />

-<br />

0.001 0.01 0.09<br />

13


Figure 3.1 Mechanism of arsenate lig<strong>and</strong> exchange on the surface of metal oxyhydroxides<br />

In addition to producing competing hydroxide ions, an increase in pH decreases the<br />

fraction of positively charged adsorption sites on the metal hydroxide surface resulting in lesser<br />

adsorption of negatively charged As(V) species. With decrease in pH, the concentration of<br />

monovalent As(V) (H 2 AsO 4 - ) increases significantly from 0.02 to 0.5 μΜ from pH 8.5 to 6.5<br />

(Table 3.1) which would result in greater removals of As(V) due to the process of lig<strong>and</strong><br />

exchange at low pH. Also, as can be seen in Table 3.1, <strong>with</strong> increasing pH, the concentration of<br />

monovalent As(III) (H 2 AsO 3 - ) increases which would result in greater removals of As(III) due to<br />

the process of lig<strong>and</strong> exchange at high pH. This trend of increasing As(III) removal <strong>with</strong><br />

increasing pH is offset by the increasing competition from hydroxide ions at the higher pH.<br />

<strong>Iron</strong> coagulation followed by flocculation <strong>and</strong> filtration (Cheng et.al. 1994) or iron<br />

coagulation-filtration <strong>with</strong>out flocculation (Ghurye et.al. 2004) have been shown to be effective<br />

technologies for removing arsenic from drinking water. The arsenic removal efficiency depends<br />

on the pH, As(III/V) speciation, <strong>and</strong> coagulant dose. Among the coagulants, ferric chloride <strong>and</strong><br />

alum are the one most commonly used due to widespread use, availability, low cost, <strong>and</strong> high<br />

capacity for arsenic. As mentioned, the purpose of the research was to test the effectiveness of<br />

Al(III), Ti(IV), Ti(III), <strong>and</strong> Zr(IV) salts for coagulation-filtration in comparison <strong>with</strong> Fe(III) as<br />

the benchmark coagulant for As(V) <strong>and</strong> As(III) removal.<br />

ARSENIC REMOVAL STUDY PROCEDURE<br />

Challenge water spiked <strong>with</strong> 50 µg/L of As(III) or As(V) was used in the coagulation<br />

experiments, which were performed at pH 6.5, 7.5, <strong>and</strong> 8.5 <strong>with</strong> the following coagulants: FeCl 3 ,<br />

alum, TiCl 3 , TiCl 4 , TiOCl 2 , ZrCl 4 , <strong>and</strong> ZrOCl 2 .<br />

<strong>Arsenic</strong> Removal Using Ferric (III) Chloride<br />

The removal efficiencies of As(V) <strong>and</strong> As(III) in NSFI challenge water using ferric<br />

chloride as a function of coagulant dose <strong>and</strong> pH are shown in Figure 3.2, which indicates that<br />

14


emoval efficiency of As(V) was pH dependent, whereas As(III) removal efficiency was<br />

independent of pH. With increase in pH, the removal efficiency of As(V) decreased. The<br />

percentage removals of As(V) <strong>with</strong> 1 mg/L Fe(III) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 90, 76,<br />

<strong>and</strong> 46%, respectively. At 2 mg/L Fe(III) dosage, approximately 26% of As(III) was removed<br />

irrespective of pH. It can be observed from Figure 3.2 that the removal efficiency of As(V) was<br />

much higher than As(III) at all pH values.<br />

<strong>Arsenic</strong> Removal Using Alum<br />

The removal efficiencies of As(V) <strong>and</strong> As(III) in NSFI challenge water using alum as a<br />

function of coagulant dose <strong>and</strong> pH are shown in Figure 3.3, which indicates that removal<br />

efficiency of As(V) was pH dependent, <strong>and</strong> no significant As(III) removal was observed at any<br />

pH. With increase in pH, the removal efficiency of As(V) decreased. The percentage removals of<br />

As(V) <strong>with</strong> 1 mg/L Al(III) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 85, 49, <strong>and</strong> 20%, respectively,<br />

which were less than the removal for the same Fe(III) dose. The removal efficiency of As(V)<br />

during Al(III) coagulation was more influenced by pH, <strong>and</strong> the removals decreased significantly<br />

<strong>with</strong> increase in pH. When alum was used as coagulant, no significant As(III) removal was<br />

observed at any dose or pH. The removal efficiencies of As(V) <strong>and</strong> As(III) <strong>with</strong> alum were lower<br />

than <strong>with</strong> ferric chloride.<br />

% removal of As(V) & As(III)<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

pH 6.5-As(V)<br />

pH 7.5-As(V)<br />

pH 8.5-As(V)<br />

pH 6.5-As(III)<br />

pH 7.5-As(III)<br />

pH 8.5-As(III)<br />

0 2 4 6 8 10 12<br />

Dose of FeCl 3 as Fe(III) (mg/L)<br />

Figure 3.2 Removal efficiency of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

ferric chloride dose <strong>and</strong> pH<br />

15


% removal of As(V) & As(III)<br />

100<br />

75<br />

50<br />

25<br />

0<br />

pH 6.5-As(V)<br />

pH 7.5-As(V)<br />

pH 8.5-As(V)<br />

pH 6.5-As(III)<br />

pH 7.5-As(III)<br />

pH 8.5-As(III)<br />

0 2 4 6 8 10<br />

Dose of Alum as Al(III) (mg/L)<br />

Figure 3.3 Removal efficiency of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

alum dose <strong>and</strong> pH<br />

% removal of As(V) & As(III)<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

pH 6.5-As(V)<br />

pH 7.5-As(V)<br />

pH 8.5-As(V)<br />

pH 6.5-As(III)<br />

pH 7.5-As(III)<br />

pH 8.5-As(III)<br />

0 2 4 6 8 10 12 14<br />

Dose of ZrCl 4 as Zr(IV) (mg/L)<br />

Figure 3.4 Removal efficiency of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

zirconium(IV) chloride dose <strong>and</strong> pH<br />

<strong>Arsenic</strong> Removal Using <strong>Zirconium</strong> (IV) Chloride<br />

The removal efficiencies of As(V) <strong>and</strong> As(III) in NSFI challenge water using<br />

zirconium(IV) chloride as a function of coagulant dose <strong>and</strong> pH are shown in Figure 3.4, which<br />

demonstrates that As(V) removal was highly pH dependent, whereas As(III) removal was<br />

independent of pH. The removal efficiency of As(V) during Zr(IV) coagulation was more<br />

influenced by pH than was Fe(III) coagulation. With increase in pH, the removal efficiency of<br />

As(V) decreased. The percentage removal of As(V) <strong>with</strong> 1 mg/L Zr(IV) dosage at pHs 6.5, 7.5,<br />

<strong>and</strong> 8.5 were 62, 28, <strong>and</strong> 8%, respectively, which were significantly less than the removal for the<br />

16


same Fe(III) dose. With 2 mg/L Zr(IV) dosage, approximately 8% of As(III) was removed<br />

irrespective of pH, which was significantly less than the removal for the same Fe(III) dose. So<br />

the removal efficiencies of As(V) <strong>and</strong> As(III) <strong>with</strong> zirconium(IV) chloride were lower than <strong>with</strong><br />

ferric chloride. It can be observed from Figure 3.4 that the removal efficiency of As(V) was<br />

much higher than As(III) at all pH values.<br />

<strong>Arsenic</strong> Removal Using <strong>Titanium</strong> (IV) Chloride<br />

The removal efficiencies of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

coagulant dose <strong>and</strong> pH titanium (IV) chloride are shown in Figure 3.5, which demonstrates that<br />

As(V) removal was highly pH dependent, whereas As(III) removal was independent of pH. With<br />

increase in pH, the removal efficiency of As(V) decreased. The percent removals of As(V) <strong>with</strong><br />

1 mg/L Ti(IV) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 54, 40, <strong>and</strong> 26%, respectively, which were<br />

significantly less than the removal for the same Fe(III) dose. With 2 mg/L Ti(IV) dosage,<br />

approximately 26% of As(III) was removed irrespective of pH, which was the same <strong>with</strong> Fe(III).<br />

So the removal efficiencies of As(V) in the presence of titanium (IV) chloride were lower than in<br />

the presence of ferric chloride, but the removal efficiency of As(III) was similar to that of ferric<br />

chloride. It can be observed from Figure 3.5 that the removal efficiency of As(V) was much<br />

higher than As(III) at all pH values.<br />

% removal of As(V) & As(III)<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

pH 6.5-As(V)<br />

pH 7.5-As(V)<br />

pH 8.5-As(V)<br />

pH 6.5-As(III)<br />

pH 7.5-As(III)<br />

pH 8.5-As(III)<br />

0 2 4 6 8 10 12 14 16<br />

Dose of TiCl 4 as Ti(IV) (mg/L)<br />

Figure 3.5 Removal efficiency of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

titanium (IV) chloride dose <strong>and</strong> pH<br />

17


<strong>Arsenic</strong> Removal Using <strong>Titanium</strong> (III) Chloride<br />

The removal efficiencies of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

coagulant dose <strong>and</strong> pH using titanium (III) chloride are shown in Figure 3.6, which demonstrates<br />

that both As(V) <strong>and</strong> As(III) removal were highly pH dependent. As was observed <strong>with</strong> the other<br />

coagulants, the removal efficiency of As(V) decreased <strong>with</strong> increasing pH. The percent removals<br />

of As(V) <strong>with</strong> 1 mg/L Ti(III) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 49, 37, <strong>and</strong> 30%, respectively,<br />

which were significantly less than the removals for the same Fe(III) dose.<br />

The removal efficiency of As(III) by Ti(III) increased <strong>with</strong> decreasing pH specifically for<br />

As(III) removal. The removal efficiencies were 42, 32, <strong>and</strong> 27% at pH 6.5, 7.5 <strong>and</strong> 8.5,<br />

respectively for a dose of 2 mg/L Ti(III). Thus the removal efficiency <strong>with</strong> titanium (III) chloride<br />

was higher than ferric chloride, which had a removal efficiency of 26% <strong>and</strong> was pH independent<br />

for all dosages.<br />

An interesting observation made during the titanium (III) chloride coagulation tests was<br />

that, based on speciation of the sample after coagulation As(III) appeared to be oxidized to<br />

As(V). The percent change or removal efficiency of As(III) species was found to be much higher<br />

than the removal efficiency of As(Total), which is shown in Figure 3.7. With a Ti(III) dose of 2<br />

mg/L, the arsenic removal efficiencies calculated as total arsenic were 42, 32, <strong>and</strong> 27% at pH<br />

6.5, 7.5, <strong>and</strong> 8.5, respectively, whereas the As(III) that was removed or converted was<br />

approximately 70% in all cases. This was possible presumably As(III) was efficiently oxidized<br />

by Ti(III) to As(V), whereas the As(V) produced was not as efficiently removed by adsorption<br />

onto the hydrolyzed Ti(III) oxyhydroxide. Further oxidation studies were carried out to elucidate<br />

the mechanism of As(III) oxidation during Ti(III) coagulation, <strong>and</strong> the results are discussed later<br />

in this chapter.<br />

In summary, the removal efficiency of As(V) during titanium(III) chloride coagulation<br />

was lower than during ferric chloride coagulation while the removal efficiency of As(III) during<br />

titanium(III) chloride coagulation was higher than that of ferric chloride.<br />

% removal of As(V) & As(III)<br />

100<br />

75<br />

50<br />

25<br />

0<br />

pH 6.5-As(V)<br />

pH 7.5-As(V)<br />

pH 8.5-As(V)<br />

pH 6.5-As(III)<br />

pH 7.5-As(III)<br />

pH 8.5-As(III)<br />

0 3 6 9 12<br />

Dose of TiCl 3 as Ti(III) (mg/L)<br />

Figure 3.6 Removal efficiency of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

titanium (III) chloride dose <strong>and</strong> pH<br />

18


100<br />

As(III) change/removal in %<br />

80<br />

60<br />

40<br />

20<br />

0<br />

pH 6.5, As(III) removed+oxidized<br />

pH 7.5, As(III) removed+oxidized<br />

pH 8.5, As(III) removed+oxidized<br />

pH 6.5, As(Total) removed<br />

pH 7.5, As(Total) removed<br />

pH 8.5, As(Total) removed<br />

0 2 4 6 8 10 12<br />

Dose of TiCl 3 as Ti(III) (mg/L)<br />

Figure 3.7 Removal/Conversion efficiency of As(III) in NSFI challenge water as a function<br />

of titanium(III) chloride dose <strong>and</strong> pH<br />

<strong>Arsenic</strong> Removal Using <strong>Titanium</strong> (IV) Oxychloride<br />

The removal efficiencies of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

titanium (IV) oxychloride coagulant dose <strong>and</strong> pH are shown in Figure 3.8, which demonstrates<br />

that both As(V) <strong>and</strong> As(III) removal were pH dependent. The percent removals of As(V) <strong>with</strong> 2<br />

mg/L Ti(IV) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 54, 37, <strong>and</strong> 29%, respectively, which were<br />

significantly less than the removals for the same Fe(III) dose. The removal efficiency of As(III)<br />

was also pH dependent; it increased <strong>with</strong> increasing pH. The removal efficiencies of As(III) were<br />

16, 20, <strong>and</strong> 26% at pH 6.5, 7.5 <strong>and</strong> 8.5, respectively for a dose of 2 mg/L Ti(IV), which shows<br />

that the removal efficiency was less than that observed for ferric chloride at pH 6.5 <strong>and</strong> 7.5 while<br />

it was the same at pH 8.5. So the removal efficiencies of As(V) by titanium (IV) oxychloride<br />

coagulation were lower than those for ferric chloride coagulation, while the removal efficiencies<br />

of As(III) were less than or equal to those of ferric chloride. It can be observed from Figure 3.8<br />

that the removal efficiency of As(V) was higher than As(III) at all pH values.<br />

<strong>Arsenic</strong> Removal Using <strong>Zirconium</strong> (IV) Oxychloride<br />

The removal efficiencies of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

zirconium (IV) oxychloride dose <strong>and</strong> pH are shown in Figure 3.9, which demonstrates that As(V)<br />

removal was highly pH dependent whereas As(III) removal was independent of pH. The removal<br />

efficiency of As(V) by zirconium (IV) oxychloride was more influenced by pH than was As(V)<br />

removal by ferric chloride. The percent removals of As(V) <strong>with</strong> 2 mg/L Zr(IV) dosage at pHs<br />

6.5, 7.5, <strong>and</strong> 8.5 were 94, 59, <strong>and</strong> 23%, respectively, which were significantly less than the<br />

removals for the same Fe(III) doses. With 2 mg/L Zr(IV) dosage, approximately 8% of As(III)<br />

was removed irrespective of pH. So the removal efficiencies of As(V) <strong>and</strong> As(III) <strong>with</strong><br />

zirconium(IV) oxychloride were lower than <strong>with</strong> ferric chloride. It can be observed from Figure<br />

3.9 that the removal efficiency of As(V) was much higher than As(III) at all pH values.<br />

19


% removal of As(V) & As(III)<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

pH 6.5-As(V)<br />

pH 7.5-As(V)<br />

pH 8.5-As(V)<br />

pH 8.5-As(III)<br />

pH 7.5-As(III)<br />

pH 6.5-As(III)<br />

0 2 4 6 8 10 12<br />

Dose of TiOCl 2 as Ti(IV) (mg/L)<br />

Figure 3.8 Removal efficiency of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

titanium(IV) oxychloride dose <strong>and</strong> pH<br />

% removal of As(V) & As(III)<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

pH 6.5-As(V)<br />

pH 7.5-As(V)<br />

pH 8.5-As(V)<br />

pH 6.5-As(III)<br />

pH 7.5-As(III)<br />

pH 8.5-As(III)<br />

0 2 4 6 8 10 12<br />

Dose of ZrOCl 2 as Zr(IV) (mg/L)<br />

Figure 3.9 Removal efficiency of As(V) <strong>and</strong> As(III) in NSFI challenge water as a function of<br />

zirconium(IV) oxychloride coagulant dose <strong>and</strong> pH<br />

ARSENIC ADSORPTION ISOTHERMS<br />

Graphs showing percent arsenic removal as a function of coagulant dose <strong>and</strong> pH are of<br />

practical importance, but it is the equilibrium adsorption isotherm that is generally used to<br />

compare the effectiveness of coagulants. The adsorption isotherms of As(V) on the in-situ<br />

formed hydroxides of Fe(III), Al(III), Zr(IV), Ti(III), Ti(IV) at pH 6.5, 7.5, <strong>and</strong> 8.5 are shown in<br />

Figures 3.10, 3.11, <strong>and</strong> 3.12, respectively. The most favorable As(V) adsorption isotherm was<br />

observed <strong>with</strong> ferric chloride, which performed better than other coagulants except at pH 8.5<br />

20


where it was similar to that of TiCl 3 . At pH 8.5, the adsorption of As(V) on the oxyhydroxides of<br />

TiCl 3 were greater or equal to that of ferric hydroxide at equilibrium As(V) concentration less<br />

than 10 μg/L. Alum performed less efficiently in comparison <strong>with</strong> FeCl 3 for As(V) sorption at all<br />

pHs. Alum performed better than zirconium <strong>and</strong> titanium coagulants at pH 6.5, while it was<br />

similar to or poorer at pH 7.5 <strong>and</strong> 8.5.<br />

qe (μg As(V)/mg of metal)<br />

90<br />

75<br />

60<br />

45<br />

30<br />

15<br />

0<br />

FeCl3<br />

ZrOCl2<br />

ZrCl4<br />

TiCl4<br />

Alum<br />

TiCl3<br />

TiOCl2<br />

0 5 10 15 20 25 30<br />

C e (μg As(V)/L)<br />

Figure 3.10 Comparison of As(V) adsorption isotherms for the in-situ formed<br />

oxyhydroxides of Fe(III), Al(III), Zr(IV), Ti(IV), <strong>and</strong> Ti(III) at pH 6.5<br />

qe (μg As(V)/mg of metal)<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

FeCl3<br />

TiCl3<br />

TiCl4<br />

Alum<br />

ZrOCl2<br />

ZrCl4<br />

TiOCl2<br />

0 5 10 15 20 25 30 35 40<br />

C e (μg As(V)/L)<br />

Figure 3.11 Comparison of As(V) adsorption isotherms for the in-situ formed<br />

oxyhydroxides of Fe(III), Al(III), Zr(IV), Ti(IV), <strong>and</strong> Ti(III) at pH 7.5<br />

21


q e (μg As(V)/mg of metal)<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

FeCl3<br />

TiCl3<br />

TiCl4<br />

Alum<br />

TiOCl2<br />

ZrOCl2<br />

ZrCl4<br />

0 5 10 15 20 25 30 35 40 45 50<br />

C e (μg As(V)/L)<br />

Figure 3.12 Comparison of As(V) adsorption isotherms for the in-situ formed<br />

oxyhydroxides of Fe(III), Al(III), Zr(IV), Ti(IV), <strong>and</strong> Ti(III) at pH 8.5<br />

The As(III) adsorption isotherms for the in-situ formed oxyhydroxides of Fe(III), Al(III),<br />

Zr(IV), Ti(IV), Ti(III), at pH 6.5, 7.5, <strong>and</strong> 8.5 are shown in Figures 3.13, 3.14, <strong>and</strong> 3.15,<br />

respectively. The highest As(III) loading was observed <strong>with</strong> the oxyhydroxide of TiCl 3 , which<br />

performed better than other coagulants at pHs 6.5 <strong>and</strong> 7.5. But at pH 8.5, the highest As(III)<br />

loading was observed <strong>with</strong> oxyhydroxides of TiCl 3 , FeCl 3 , TiCl 4 , <strong>and</strong> TiOCl 2 which had similar<br />

adsorption isotherms. The oxyhydroxide of TiCl 4 exhibited similar loading to that of FeCl 3 at all<br />

pHs, but the oxyhydroxide of TiOCl 2 exhibited similar loading only at pH of 8.5. Alum had<br />

virtually no measurable capacity for As(III) at any pH.<br />

qe (μg As(III)/mg of metal)<br />

12<br />

9<br />

6<br />

3<br />

0<br />

TiCl3<br />

FeCl3<br />

TiCl4<br />

TiOCl2<br />

ZrCl4<br />

ZrOCl2<br />

Alum<br />

0 10 20 30 40 50 60<br />

C e (μg As(III)/L)<br />

Figure 3.13 Comparison of As(III) adsorption isotherms for the in-situ formed<br />

oxyhydroxides of Fe(III), Al(III), Zr(IV), Ti(IV), <strong>and</strong> Ti(III) at pH 6.5<br />

22


qe (μg As(III)/mg of metal)<br />

10<br />

8<br />

6<br />

4<br />

2<br />

0<br />

v<br />

TiCl3<br />

FeCl3<br />

TiCl4<br />

TiOCl2<br />

ZrCl4<br />

ZrOCl2<br />

Alum<br />

0 10 20 30 40 50 60<br />

C e (μg As(III)/L)<br />

Figure 3.14 Comparison of As(III) adsorption isotherms for the in-situ formed<br />

oxyhydroxides of Fe(III), Al(III), Zr(IV), Ti(IV), <strong>and</strong> Ti(III) at pH 7.5<br />

qe (μg As(III)/mg of metal)<br />

10<br />

8<br />

6<br />

4<br />

2<br />

0<br />

TiCl3<br />

FeCl3<br />

TiOCl2<br />

TiCl4<br />

ZrCl4<br />

ZrOCl2<br />

Alum<br />

0 10 20 30 40 50 60<br />

C e (μg As(III)/L)<br />

Figure 3.15 Comparison of As(III) adsorption isotherms for the in-situ formed<br />

oxyhydroxides of Fe(III), Al(III), Zr(IV), Ti(IV), <strong>and</strong> Ti(III) at pH 8.5<br />

23


COMPARISON OF ADSORPTION CAPACITIES<br />

The best-fit isotherm equations (power curves) were used to calculate the adsorption<br />

capacities of the metal hydroxides for an equilibrium concentration of 10 μg /L As(V) or 10 μg/L<br />

As(III).<br />

<strong>Arsenic</strong> Adsorption Comparison on a Mass Basis<br />

The comparisons of As(V) <strong>and</strong> As(III) adsorption capacities of aluminum, titanium <strong>and</strong><br />

zirconium salts <strong>with</strong> ferric chloride on a mass basis for an equilibrium concentration of 10 μg/L<br />

in NSFI challenge water are shown in Figures 3.16 <strong>and</strong> 3.17. As was observed in the isotherm<br />

comparisons, As(V) capacity was typically 3 to 20 times greater than As(III) depending on the<br />

coagulant <strong>and</strong> pH (Al <strong>and</strong> Zr coagulants were not considered due to negligible As(III)<br />

adsorption). Also, for all coagulants, As(V) adsorption capacity decreased significantly as<br />

equilibrium pH increased from 6.5 to 8.5 as observed in Figure 3.16. Ferric chloride was found<br />

to have the highest adsorption capacity at pH 6.5 <strong>and</strong> 7.5, while at pH 8.5 TiCl 3 had similar<br />

adsorption capacity. Alum was found to have As(V) adsorption capacities higher than the<br />

titanium- <strong>and</strong> zirconium-based coagulants at pH 6.5, while at pH 7.5 <strong>and</strong> 8.5 the adsorption<br />

capacities were similar or less. The order of As(V) adsorption capacities of different coagulants<br />

at three different pHs based on an equilibrium concentration of 10 μg/L was as follows:<br />

pH 6.5: FeCl 3 > Alum > ZrOCl 2 ≈ ZrCl 4 > TiCl 4 > TiCl 3 > TiOCl 2<br />

pH 7.5: FeCl 3 >> TiCl 3 > TiCl 4 > Alum ≈ ZrOCl 2 > ZrCl 4 > TiOCl 2<br />

pH 8.5: FeCl 3 ≈ TiCl 3 > TiCl 4 > Alum > ZrCl 4 ≈ TiOCl 2 > ZrOCl 2<br />

70<br />

qe (μg As(V)/mg of metal)<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

FeCl3; 59.6<br />

TiCl3; 20.0<br />

TiCl4; 25.1<br />

Alum; 40.9<br />

ZrCl4; 28.1<br />

ZrOCl2; 30.1<br />

TiOCl2; 9.3<br />

FeCl3, 37.3<br />

TiCl3; 16.1<br />

TiCl4; 13.9<br />

Alum; 13<br />

ZrCl4; 9.89<br />

ZrOCl2; 11.2<br />

TiOCl2; 5.94<br />

FeCl3; 14.3<br />

TiCl3; 14.0<br />

TiCl4; 10.7<br />

Alum; 7.4<br />

pH 6.5 pH 7.5 pH 8.5<br />

Figure 3.16 Comparison of mass As(V) adsorbed per mass of Fe(III), Al(III), Ti(III),<br />

Ti(IV), or Zr(IV) coagulant for an equilibrium concentration of 10 μg/L as a function of<br />

pH<br />

ZrCl4; 5.02<br />

ZrOCl2; 4.14<br />

TiOCl2; 4.78<br />

24


7<br />

qe (μg As(III)/mg of metal)<br />

6<br />

5<br />

4<br />

3<br />

2<br />

1<br />

0<br />

TiCl3; 6.13<br />

FeCl3; 3.00<br />

TiCl4; 2.74<br />

TiOCl2; 1.44<br />

ZrCl4; 0.29<br />

ZrOCl2; 0.03<br />

Alum; 0.00<br />

TiCl3; 4.00<br />

FeCl3; 2.96<br />

TiCl4; 2.86<br />

TiOCl2; 2.00<br />

ZrCl4; 0.76<br />

ZrOCl2; 0.10<br />

Alum; 0.00<br />

TiCl3; 3.00<br />

FeCl3; 3.12<br />

TiCl4; 2.74<br />

TiOCl2; 2.71<br />

ZrCl4; 0.51<br />

pH 6.5 pH 7.5 pH 8.5<br />

ZrOCl2; 0.04<br />

Alum; 0.00<br />

Figure 3.17 Comparison of mass As(III) adsorbed per mass of Fe(III), Al(III), Ti(III),<br />

Ti(IV), or Zr(IV) coagulant for an equilibrium concentration of 10 μg/L as a function of<br />

pH<br />

A similar comparison of As(III) adsorption (Figure 3.17) showed that the As(III)<br />

adsorption capacities (a) were independent of pH for FeCl 3 , alum, TiCl 4 , ZrCl 4 , <strong>and</strong> ZrOCl 2 , (b)<br />

decreased <strong>with</strong> increasing pH for TiCl 3 <strong>and</strong> (c) increased <strong>with</strong> increasing pH for TiOCl 2 . The<br />

As(III) adsorption capacity <strong>with</strong> TiCl 3 was high compared <strong>with</strong> other coagulants at pHs 6.5 <strong>and</strong><br />

7.5, while it was same as FeCl 3 <strong>and</strong> Ti(IV) salts at pH 8.5. The As(III) adsorption capacity of<br />

TiCl 4 was the same as that of FeCl 3 at all three pHs. The As(III) adsorption capacities of<br />

zirconium salts were significantly less than titanium <strong>and</strong> ferric salts, while alum did not have any<br />

measurable As(III) adsorption capacity. Thus, it appears that based on a mg metal/L basis <strong>and</strong><br />

ignoring costs, Ti(III) could be the coagulant of choice for As(III) removal. The order of As(III)<br />

adsorption capacities of different coagulants at three different pHs based on an equilibrium<br />

concentration of 10 μg/L was as follows:<br />

pH 6.5: TiCl 3 > FeCl 3 ≈ TiCl 4 > TiOCl 2 >> ZrCl 4 ≈ ZrOCl 2 >> Alum<br />

pH 7.5: TiCl 3 > FeCl 3 ≈ TiCl 4 > TiOCl 2 >> ZrCl 4 ≈ ZrOCl 2 >> Alum<br />

pH 8.5: TiCl 3 ≈ FeCl 3 ≈ TiCl 4 ≈ TiOCl 2 >> ZrCl 4 ≈ ZrOCl 2 >> Alum<br />

<strong>Arsenic</strong> Adsorption Comparison on a Molar Basis<br />

The comparisons of As(V) <strong>and</strong> As(III) adsorption capacities of aluminum, titanium <strong>and</strong><br />

zirconium salts <strong>with</strong> ferric chloride on a molar basis for an equilibrium concentration of 10 μg/L<br />

in NSFI challenge water are shown in Figures 3.18 <strong>and</strong> 3.19. Similar to what was observed in the<br />

mass comparison, the highest As(V) loading on a molar basis (mmol As(V)/mol coagulant metal)<br />

was observed for ferric chloride at all three pHs. The highest As(III) loading on the coagulant on<br />

a molar basis was observed for titanium(III) chloride at pH 6.5 <strong>and</strong> 7.5, while at pH 8.5, ferric<br />

25


chloride had the highest molar adsorption capacity. Although aluminum was reported to have<br />

equal adsorption capacity (mmol As(V)/mol metal) to that of iron(III) (Edwards, 1994; McNeil<br />

& Edwards, 1997), our study found that ferric chloride had significantly higher adsorption<br />

capacity than alum at all three pHs. Finally, on a molar basis, alum was found to have As(V)<br />

adsorption capacities less than or equal to titanium <strong>and</strong> zirconium based coagulants.<br />

50<br />

qe (mmol As(V)/mol metal)<br />

40<br />

30<br />

20<br />

10<br />

0<br />

FeCl3; 46.7<br />

ZrCl4; 34.2<br />

ZrOCl2; 36.6<br />

TiCl3; 12.8<br />

TiCl4; 16.1<br />

TiOCl2; 6.0<br />

Alum; 14.7<br />

FeCl3; 27.8<br />

ZrCl4; 12.0<br />

ZrOCl2; 13.6<br />

TiCl3; 10.3<br />

TiCl4; 8.9<br />

TiOCl2; 3.8<br />

Alum; 4.7<br />

FeCl3; 10.7<br />

ZrCl4; 6.1<br />

ZrOCl2; 5.0<br />

TiCl3; 8.9<br />

pH 6.5 pH 7.5 pH 8.5<br />

TiCl4; 6.8<br />

TiOCl2; 3.1<br />

Figure 3.18 Comparison of moles As(V) adsorbed per mol of Fe(III), Al(III), Ti(III), Ti(IV),<br />

or Zr(IV) coagulant for an equilibrium concentration of 10 μg/L as a function of pH<br />

Alum; 2.6<br />

4<br />

q e (mmol As(III)/mol metal)<br />

3<br />

2<br />

1<br />

0<br />

TiCl 3 ; 3.91<br />

FeCl 3 ; 2.24<br />

TiCl 4 ; 1.75<br />

TiOCl 2 ; 0.92<br />

ZrCl 4 ; 0.35<br />

ZrOCl 2 ; 0.04<br />

Alum; 0.00<br />

TiCl 3 ; 2.55<br />

FeCl 3 ; 2.21<br />

TiCl4; 1.83<br />

TiOCl 2 ; 1.28<br />

ZrCl 4 ; 0.93<br />

ZrOCl 2 ; 0.12<br />

Alum; 0.00<br />

TiCl 3 ; 1.92<br />

FeCl 3 ; 2.33<br />

TiCl 4 ; 1.75<br />

TiOCl 2 ; 1.73<br />

pH 6.5 pH 7.5 pH 8.5<br />

Figure 3.19 Comparison of moles As(III) adsorbed per mol of Fe(III), Al(III), Ti(III),<br />

Ti(IV), or Zr(IV) coagulant for an equilibrium concentration of 10 μg/L as a function of<br />

pH<br />

ZrCl 4 ; 0.62<br />

ZrOCl 2 ; 0.04<br />

Alum; 0.00<br />

26


COMPARISON OF COAGULANT COSTS<br />

The determination of the economic feasibility of using these salts as coagulants for<br />

arsenic removal was studied by comparing the costs of the industrial grade coagulants: FeCl 3<br />

(Thatcher Company, Chemical Market Reporter), alum (Thatcher Company, General Chemicals),<br />

ZrOCl 2 (MEI), TiOCl 2 (Millennium Chemicals), <strong>and</strong> TiCl 4 (Millennium Chemicals).<br />

Unfortunately the chemical costs of industrial-grade TiCl 3 , <strong>and</strong> ZrCl 4 were not available, <strong>and</strong> so<br />

for these chemicals, cost comparison could not be made. The coagulant cost comparisons were<br />

based on reducing As(V) or As(III) from 0.05 to 0.01 mg/L in the NSFI challenge water (Table<br />

3.2). The calculations indicated that compared <strong>with</strong> alum, ZrOCl 2 , TiCl 4 <strong>and</strong> TiOCl 2 , ferric<br />

chloride was the most cost effective coagulant ($/million gallons water treated) at pH 6.5-8.5 for<br />

both As(V) <strong>and</strong> As(III) (Table 3.2). When comparing the common coagulants alum <strong>and</strong> ferric<br />

chloride for As(V) removal, chemical costs for alum were found to be 4-8 times greater than<br />

ferric chloride depending on the pH.<br />

Table 3.2<br />

Comparison of chemical costs for treating one million gallons of NSFI-53 water to reduce<br />

As(V)/As(III) from 50 to < 10 μg/L<br />

Coagulant<br />

Cost for treating As(V)<br />

from 0.05 to 0.01 mg/L<br />

($/Mgal)<br />

Cost for treating As(III)<br />

from 0.05 to 0.01 mg/L<br />

($/Mgal)<br />

FeCl 3 2.46 51.2<br />

Alum 10.5 > 120,000*<br />

pH 6.5 TiCl 4 40.0 366<br />

TiOCl 2 109 709<br />

ZrOCl 2 116 120,000<br />

FeCl 3 4.12 52.0<br />

Alum 32.9 > 120,000*<br />

pH 7.5 TiCl 4 72.1 351<br />

TiOCl 2 172 510<br />

ZrOCl 2 311 36,500<br />

FeCl 3 10.7 49.3<br />

Alum 58.4 > 120,000*<br />

pH 8.5 TiCl 4 93.9 367<br />

TiOCl 2 213 377<br />

ZrOCl 2 843 99,400<br />

* There was no measurable adsorption of As(III) onto alum, thus precise cost estimates<br />

could not be made.<br />

27


When comparing the chemical costs of FeCl 3 coagulation to remove As(III) vs As(V)<br />

from the NSFI challenge water, As(III) removal was found to be 5-20 times more costly than<br />

As(V) removal depending on the pH. Thus, As(III) should be oxidized to As(V) for costeffective<br />

treatment. Although Ti(III) was found to have the highest capacity for As(III) a<br />

comparison of treatment cost could not be made <strong>with</strong> ferric chloride because of the unavailability<br />

of costs for industrial-grade Ti(III) coagulant. However, due to the very high coagulant costs for<br />

As(V) <strong>and</strong> As(III) removal using TiCl 4 <strong>and</strong> TiOCl 2 , it is unlikely that TiCl 3 would be cost<br />

competitive <strong>with</strong> FeCl 3 for As(III) removal.<br />

OXIDATION STUDY OF ARSENIC(III) IN THE PRESENCE OF TITANIUM(III)<br />

CHLORIDE<br />

During the removal of As(III) using TiCl 3 , it was observed that the μg/L As(III)<br />

remaining was far less than that of the As(Tot) remaining, which suggested that in the presence<br />

of Ti(III), As(III) was oxidized to As(V). To verify <strong>and</strong> characterize the Ti(III)-facilitated<br />

oxidation of As(III) to As(V) jar tests were conducted at pH 7.5 for a dose of 2 mg/L Ti(III).<br />

Flocculation (0, 5, 10, 20, <strong>and</strong> 30 min) <strong>and</strong> settling (0, 1, <strong>and</strong> 2 hr) times were varied in order to<br />

monitor the As(III) to As(V) oxidation rate.<br />

TiCl 3 Oxidation Study Results<br />

The results of the study on oxidation of arsenic(III) <strong>with</strong> TiCl 3 are shown in Figure 3.20,<br />

which demonstrates that the oxidation of As(III) to As(V) by Ti(III) proceeded slowly, i.e., it<br />

continued for several hours during which the conversion increased <strong>with</strong> time. From Figure 3.20 it<br />

can be seen that, for a flocculation time of 20 minutes, the As(Tot) removals were 36, 37, <strong>and</strong><br />

39% for 0, 1, <strong>and</strong> 2 hours of settling, i.e., As(Tot) removal was approximately the same<br />

regardless of the settling time. However, for this same 20-min flocculation time, As(III) removed<br />

or converted to As(V) was 71, 64, <strong>and</strong> 51% after 0, 1, <strong>and</strong> 2 hours of settling, i.e., coagulated<br />

samples that were settled for two hours showed greater As(III) oxidation than coagulated<br />

samples settled for one hour which in turn showed greater oxidation than coagulated samples that<br />

were not settled. Thus it was concluded that even though the As(III) continued to get oxidized to<br />

As(V) during the settling stage, arsenic removal was restricted to the time period during<br />

coagulation. The reason is that during settling <strong>and</strong> after the floc had settled, As(III) oxidation<br />

continued, but the new As(V) that was formed was not adsorbed onto the Ti(OH) 3 (s) because of<br />

severe mass transfer limitations between As(V) in the bulk solution <strong>and</strong> Ti(OH) 3 (s) on the<br />

bottom of the beaker. Another factor to consider is that, based on prior experience <strong>with</strong><br />

preformed metal oxyhydroxide flocs, the already-formed floc could not adsorb as much As(V)<br />

(or As(III)) as was adsorbed by in-situ formed floc during rapid mixing.<br />

Flocculation time in the range of 0-30 minutes did have an effect on arsenic removal in<br />

the samples that were not settled. During slow mixing of the floc, additional As(III) was oxidized<br />

to As(V) <strong>and</strong> adsorbed onto the floc particles as they were slowly flocculated.<br />

28


As removal/conversion (%)<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

As(III) removal/conversion after Settling=2hr<br />

As(III) removal/conversion after Settling=1hr<br />

As(III) removal/conversion after Settling=0hr<br />

Total As removal after Settling=2hr<br />

Total As removal after Settling=1hr<br />

Total As removal after Settling=0hr<br />

0 5 10 15 20 25 30 35 40<br />

Flocculation Time (min)<br />

Figure 3.20 As(III) <strong>and</strong> As(Tot) removed or converted during Ti(III) coagulation as a<br />

function of flocculation <strong>and</strong> settling times<br />

Proposed Mechanism for the Oxidation of <strong>Arsenic</strong>(III) in Presence of <strong>Titanium</strong>(III)<br />

The oxidation of hydrolyzed titanium(III) by means of molecular oxygen consumes more<br />

oxygen than is necessary to convert Ti(III) to Ti(IV), <strong>and</strong> the extra oxygen requirement has been<br />

attributed to the formation of H 2 O 2 during Ti(III) oxidation in water (Remy, 1956).<br />

TiCl 3 + H 2 O Ti(OH) 3 + 3HCl (3.1)<br />

Ti(OH) 3 + ½O 2 + H 2 O Ti(OH) 4 + ½ H 2 O 2 (3.2)<br />

If the hydrogen peroxide formed is used up as it is formed, it is found that for each atom<br />

of Ti(III) oxidized, ½ molecule of oxygen is taken up, <strong>and</strong> ½ molecule of H 2 O 2 is formed.<br />

Hydrogen peroxide thus formed can promote the oxidation of As(III) to As(V), but studies have<br />

shown that peroxide oxidation of As(III) is very slow at pH < 9.0 (Pettine et al, 1999). However,<br />

the rate has been shown to be faster in the presence of metals, particularly Fe(II) <strong>and</strong> copper, in<br />

the near-neutral pH range (Pettine et al, 2000; Hug et al, 2003; Bissen et al, 2003; Voegeline et<br />

al, 2003). If titanium (III) behaves in a similar catalytic manner to Fe(II), one might expect an<br />

enhanced rate of As(III) oxidation by peroxide, as was observed in this work.<br />

Thus, the results from these tests show that the oxidation of As(III) by Ti(III) is a slow<br />

step, i.e., it continues for several hours during which the As(III) conversion to As(V) increases<br />

<strong>with</strong> time, but As(V) is not removed by the floc, because the floc is not in contact <strong>with</strong> the<br />

supernatant solution. Furthermore pre-formed floc has been shown to adsorb less arsenic than<br />

floc formed in-situ during rapid mixing.<br />

29


SUMMARY AND CONCLUSIONS<br />

The following commercially available, hydrolysable Fe, Al, Zr <strong>and</strong> Ti salts were studied:<br />

FeCl 3 , alum, TiCl 4 , TiCl 3 , TiOCl 2 , ZrCl 4 , <strong>and</strong> ZrOCl 2 . Variable-coagulant-dose jar tests were<br />

conducted at pH 6.5, 7.5, <strong>and</strong> 8.5 using the st<strong>and</strong>ard NSFI challenge water spiked <strong>with</strong> 50 μg/L<br />

As(III) or As(V).<br />

The percent removal of As(V) was highly pH dependent <strong>and</strong> the removal increased <strong>with</strong><br />

decreasing pH for all coagulants tested. In particular, the adsorption capacity of As(V) <strong>with</strong><br />

zirconium salts decreased significantly <strong>with</strong> increasing pH. When comparing arsenic adsorption<br />

isotherms for all the coagulants, the highest As(V) loadings on a coagulant on a mass basis (mg<br />

As(V)/g metal) were observed <strong>with</strong> FeCl 3 , which performed better than aluminum, titanium, <strong>and</strong><br />

zirconium salts at pHs of 6.5 <strong>and</strong> 7.5. However, at pH 8.5, As(V) loadings on ferric chloride<br />

were approximately the same as TiCl 3 at equilibrium As(V) ≤ 10 μg/L. When comparing<br />

adsorption isotherms, the highest As(V) loading on any coagulant on a molar basis or a mass<br />

basis was observed for ferric chloride at all three pHs, <strong>and</strong> the As(V) loading on iron was<br />

significantly greater than aluminum. Regardless of the basis of comparison, ferric chloride was a<br />

far better coagulant than alum for As(V) removal.<br />

The percent removal efficiency of As(III) was independent of pH for FeCl 3 , TiCl 4 , ZrCl 4 ,<br />

<strong>and</strong> ZrOCl 2 , <strong>and</strong> it decreased <strong>with</strong> increasing pH for TiCl 3 <strong>and</strong> increased <strong>with</strong> increasing pH for<br />

TiOCl 2 . The removal of As(III) by alum was insignificant. When comparing arsenic adsorption<br />

isotherms for all the coagulants, the highest As(III) loading on a coagulant (mg As(III)/g metal)<br />

was observed <strong>with</strong> TiCl 3 , which performed better than ferric, titanium(IV), <strong>and</strong> zirconium salts at<br />

pHs of 6.5 <strong>and</strong> 7.5. However Ti(III) had similar adsorption capacity to that of Fe(III) <strong>and</strong> Ti(IV)<br />

coagulants at pH 8.5. Alum did not have any adsorption capacity for As(III). The highest As(III)<br />

loading on the coagulant on a molar basis was observed for titanium(III) chloride at pH 6.5 <strong>and</strong><br />

7.5, while at pH 8.5, ferric chloride had the highest molar adsorption capacity. At all doses, the<br />

removal efficiency of As(V) was significantly greater than As(III) at all pHs <strong>with</strong> all seven<br />

coagulants tested: FeCl 3 , alum, TiCl 4 , TiCl 3 , TiOCl 2 , ZrCl 4 , <strong>and</strong> ZrOCl 2 .<br />

When comparing chemical costs for FeCl 3 , alum, ZrOCl 2, TiCl 4 , or TiOCl 2 coagulation to<br />

remove As(V) or As(III) the most economical was FeCl 3 . Of the common coagulants, alum was<br />

found to be 4-8 times more expensive than ferric chloride. The chemical cost of ferric chloride<br />

coagulation was calculated to be more than 5 to 20 times higher for As(III) treatment compared<br />

<strong>with</strong> As(V).<br />

There was experimental evidence that the high removal efficiency of As(III) by TiCl 3 <strong>and</strong><br />

the unusual As(III) behavior of increasing removal <strong>with</strong> decreasing pH was due to oxidation of<br />

As(III) to As(V) by H 2 O 2 , which based on the literature, was formed from Ti(III) hydrolysis in<br />

the challenge water, which contained some dissolved oxygen. In spite of its partial oxidation, the<br />

experimentally observed removal of As(V) oxidized from As(III) was far less than the removal<br />

of a similar starting concentration of As(V), because (a) the floc was already formed when it<br />

contacted As(V), <strong>and</strong> (b) the As(III) oxidation continued for many hours during which the<br />

Ti(OH) 3 formed had settled <strong>and</strong> was not in contact <strong>with</strong> As(V) formed.<br />

30


CHAPTER 4<br />

EFFECT OF COMPETING IONS ON ARSENIC (III)/(V) ADSORPTION<br />

INTRODUCTION<br />

The adsorption of arsenic occurs by the process of lig<strong>and</strong> exchange as discussed in<br />

Chapter 3 (Figure 3.1). However the adsorption is influenced by the presence of competing<br />

anions in source water such as bicarbonate, sulfate, silica, phosphate, <strong>and</strong> vanadate. These anions<br />

are expected to sorb onto the metal hydroxides <strong>and</strong> competition between these substances for<br />

adsorption sites will significantly interfere <strong>with</strong> As removal. This competition has the potential to<br />

reduce the overall effectiveness of arsenic removal from source water.<br />

The NSFI challenge water under study contains silica <strong>and</strong> phosphate, which will compete<br />

<strong>with</strong> arsenates for adsorption sites on the oxyhydroxide surfaces because they form strong<br />

surface complexes <strong>with</strong> iron, aluminum <strong>and</strong> similar metal oxides (Hingston 1981). Vanadate is<br />

also expected to affect the adsorption of arsenic <strong>and</strong> reduce the adsorption capacity of arsenic<br />

onto the coagulant hydroxides. Table 4.1 gives the equilibrium constants (pKa’s) for arsenic<br />

containing acids, silicic acid, phosphoric acid <strong>and</strong> vanadic acid.<br />

Soluble silica in water exists as silicic acid (H 4 SiO 4 ) which dissociates to H 3 SiO - 4 (pK a1 =<br />

9.84). So at pH 6.5-8.0, silica exists almost entirely as uncharged H 4 SiO 4 , but at pH 8.5-9.0, the<br />

-<br />

fraction of silica existing as anionic silicate, H 3 SiO 4 is significant. Both neutral H 4 SiO 4 <strong>and</strong><br />

anionic H 3 SiO - 4 can act as lig<strong>and</strong>s <strong>and</strong> form surface complexes, but the anions are stronger<br />

lig<strong>and</strong>s. Although the neutral species dominates at pH 6.5-8.5, the concentration of silica is so<br />

high (712 μM) that the monovalent silicate anion concentration (H 3 SiO - 4 ) is very high in<br />

comparison to the arsenic species at pHs 7.5 <strong>and</strong> 8.5 <strong>and</strong> is nearly equal at pH 6.5 (Table 4.1).<br />

Because of the high concentration of silicate anion (H 3 SiO - 4 ), compared <strong>with</strong> the concentration<br />

of arsenic, silica is expected to compete significantly <strong>with</strong> arsenic for adsorption sites at pH 6.5-<br />

8.5.<br />

Phosphoric acid has similar pK a values to those of arsenic acid, <strong>and</strong> phosphate exists as<br />

monovalent <strong>and</strong> divalent anions (H 2 PO - 4 <strong>and</strong> as HPO 2- 4 ) at pH 6.5-8.5, which are similar in<br />

chemical behavior to H 2 AsO - 4 <strong>and</strong> HAsO 2- -<br />

4 . Both H 2 PO 4 <strong>and</strong> HPO 2- 4 can have lig<strong>and</strong> exchange<br />

reactions <strong>with</strong> the hydroxides formed <strong>and</strong> will compete for adsorption sites <strong>with</strong> the arsenates.<br />

Since the concentration of phosphorous (1.29 μM) is higher than that of arsenic (0.67 μM),<br />

phosphate was expected to compete strongly <strong>with</strong> arsenic for adsorption sites.<br />

Vanadate in water is actually a mixture of mono <strong>and</strong> oligovanadates, the composition of<br />

which depends on the pH, the overall vanadium concentration, <strong>and</strong> the ionic strength. Vanadic<br />

acid has similar pK a values to those of arsenic <strong>and</strong> phosphoric acids, <strong>and</strong> exists as H 2 VO - 4 <strong>and</strong><br />

2- -<br />

2-<br />

HVO 4 anions at pH 6.5-8.5 which are similar to H 2 AsO 4 <strong>and</strong> HAsO 4 anions. Since the<br />

concentration of vanadium (0.98 μΜ) is higher than that of arsenic (0.67 μM), <strong>and</strong> since<br />

vanadate is chemically similar to phosphate, it was expected that vanadate would compete for<br />

adsorption sites <strong>with</strong> arsenic.<br />

Experiments were conducted using NSFI challenge water at three pHs spiked <strong>with</strong> two<br />

levels of silica, phosphate, <strong>and</strong> vanadate in order to assess the combined effects of the competing<br />

ions under study. The competing ions were also studied separately; i.e., when silicate was<br />

present, phosphate <strong>and</strong> vanadate were absent so as to be able to isolate the influence of each<br />

contaminant.<br />

31


Table 4.1<br />

Composition of species of some important acids present in pH 6.5-8.5<br />

Concentration of species (μmol/L) in<br />

Acid<br />

NSFI challenge water<br />

pKa pH 6.5 pH 7.5 pH 8.5<br />

-<br />

pK<br />

<strong>Arsenic</strong> acid (0.67 μmol/L * 1 =2.22 H 2 AsO 4 0.50 0.15 0.02<br />

2-<br />

pK 2 =6.98 HAsO 4 0.17 0.52 0.65<br />

50 μg/L as As<br />

pK 3 =11.5<br />

-<br />

pK<br />

Phosphoric acid (1.29 μmol/L)* 1 =2.16 H 2 PO 4 1.07 0.43 0.06<br />

2-<br />

pK 2 =7.2 HPO 4 0.22 0.86 1.23<br />

40 μg/L as P<br />

pK 3 =12.4<br />

Arsenous acid (0.67 μmol/L)* pK 1 =9.23 H 3 AsO 3 0.67 0.66 0.57<br />

50 μg/L as As pK 2 =12.1<br />

-<br />

H 2 AsO 3 0.001 0.01 0.09<br />

Silicic acid (333 μmol/L)*<br />

20 mg/L as SiO 2<br />

Vanadic acid (0.98 μmol/L)**<br />

50 μg/L as V<br />

*Source: Schecher 1998.<br />

** Source: Sigel 1995.<br />

pK 1 =9.84 Si(OH) 4 712 709 681<br />

pK 2 =13.2<br />

-<br />

H 3 SiO 4 0.33 3.24 31<br />

pK 1 =7.91 H 3 VO 3 0.94 0.71 0.20<br />

pK 2 =13.4<br />

-<br />

H 2 VO 3 0.04 0.27 0.78<br />

The effects of silicate, phosphate <strong>and</strong> vanadate in NSFI challenge water were studied<br />

<strong>with</strong> five coagulants: FeCl 3 , alum, TiCl 3 , TiCl 4 <strong>and</strong> ZrCl 4 . The coagulants TiOCl 2 <strong>and</strong> ZrOCl 2<br />

were dropped from further studies because they did not appear to offer any advantage over TiCl 4<br />

<strong>and</strong> ZrCl 4 , respectively. The effect of competing ions on As(III) adsorption using alum as<br />

coagulant was also not carried out since it was found that aluminum does not have any<br />

significant As(III) adsorption.<br />

EFFECT OF SILICA IN NSFI CHALLENGE WATER CONTAINING PHOSPHATE<br />

<strong>Coagulation</strong> experiments were first performed using the st<strong>and</strong>ard challenge water <strong>with</strong><br />

silica (SiO 2 = 20 mg/L) <strong>and</strong> <strong>with</strong>out silica (SiO 2 = 0 mg/L) in the presence of phosphate. The<br />

objective was to study the effect of silica in the challenge water on arsenic removal in the pH<br />

range of 6.5 to 8.5 using all five coagulants tested: FeCl 3 , alum, ZrCl 4 , TiCl 4 , <strong>and</strong> TiCl 3 .<br />

Effect of Silica in NSFI Water <strong>with</strong> Phosphate Using Ferric (III) Chloride as Coagulant<br />

Based on jar tests in the presence <strong>and</strong> absence of silica, it was found that silica does have<br />

a significant impact on the adsorption of arsenic. The removal percentages of As(V) at 1 mg/L<br />

Fe(III) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 90, 76, <strong>and</strong> 46%, respectively, in the presence of<br />

silica. In the absence of silica, the removal percentages at the same pHs were 94, 85, <strong>and</strong> 76%,<br />

respectively for the same doses, which shows that removal efficiency was higher in the absence<br />

of silica .This can be observed in Figure 4.1 through the more favorable adsorption isotherms in<br />

the absence of silica at all three pH’s. It was somewhat surprising that silica had such a<br />

significant effect at pH 6.5, where 99.95% of the silica is present as non-ionic H 4 SiO 4 .<br />

32


q e (μg As(V)/mg Fe(III))<br />

140<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

pH6.5, SiO2=0mg/L<br />

pH7.5, SiO2=0mg/L<br />

pH6.5, SiO2=20mg/L<br />

pH8.5, SiO2=0mg/L<br />

pH7.5, SiO2=20mg/L<br />

pH8.5, SiO2=20mg/L<br />

0 10 20 30 40 50<br />

C e (μg As(V)/L)<br />

Figure 4.1 Effect of silica on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Fe(III) hydroxide in NSFI water <strong>with</strong> phosphate<br />

qe (μg As(III)/mg of Fe(III))<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

pH 8.5, SiO2=0mg/L<br />

pH 7.5, SiO2=0mg/L<br />

pH 6.5, SiO2=0mg/L<br />

pH 6.5, SiO2=20mg/L<br />

pH 7.5, SiO2=20mg/L<br />

pH 8.5, SiO2=20mg/L<br />

0 5 10 15 20 25 30 35 40 45<br />

C e (μg As(III)/L)<br />

Figure 4.2 Effect of silica on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Fe(III) hydroxide in NSFI water <strong>with</strong> phosphate<br />

The As(III) adsorption isotherms shown in Figure 4.2 illustrate the strong negative<br />

influence of silica on As(III) adsorption onto Fe(III) hydroxide in the 6.5-8.5 pH range. The<br />

adsorption behavior of As(III) was similar to that observed for the As(V) in the absence of silica<br />

where the removal efficiency was higher than <strong>with</strong> silica present. In the absence of silica, the<br />

adsorption of As(III) was found to be pH dependent, <strong>and</strong> the As(III) removal efficiency<br />

increased <strong>with</strong> increasing pH, which was not observed in the presence of silica were the removal<br />

was the same regardless of pH. At 2 mg/L Fe(III) dosage, approximately 26% of As(III) was<br />

removed irrespective of pH in the presence of silica. But in the absence of silica, higher removal<br />

33


efficiencies of 41, 53, <strong>and</strong> 71% were obtained at pHs 6.5, 7.5, <strong>and</strong> 8.5, respectively. Thus, the<br />

As(III) removal efficiency was highly pH dependent in the absence of silica <strong>and</strong> the removal<br />

efficiencies increased <strong>with</strong> increasing pH.<br />

Effect of Silica in NSFI Water <strong>with</strong> Phosphate Using Alum as Coagulant<br />

Based on jar tests in the presence <strong>and</strong> absence of silica, it was found that silica does have<br />

a effect on the adsorption of As(V) at low pH during coagulation <strong>with</strong> alum. The removal<br />

percentages of As(V) at 1 mg/L Al(III) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 85, 49, <strong>and</strong> 20%,<br />

respectively, in the presence of silica. In the absence of silica, the removal percentages at the<br />

same pHs were 95, 73, <strong>and</strong> 12%, respectively for the same doses, which shows that removal<br />

efficiency was higher in the absence of silica at pH 6.5 <strong>and</strong> 7.5. However there was no effect of<br />

silica on As(V) adsorption at pH 8.5, but in fact the adsorption decreased in the absence of silica<br />

at pH 8.5. This can be observed in Figure 4.3 through the more favorable adsorption isotherms in<br />

the absence of silica at pH 6.5 <strong>and</strong> 7.5. It was somewhat surprising that silica had so litter effect<br />

at pH 8.5, where silica is present as ionic species compared to low pH where silica is non-ionic.<br />

Effect of Silica in NSFI Water <strong>with</strong> Phosphate Using <strong>Zirconium</strong> (IV) Chloride as<br />

Coagulant<br />

Based on jar tests in the presence <strong>and</strong> absence of silica, it was found that silica does have<br />

a significant impact on the adsorption of arsenic during coagulation <strong>with</strong> ZrCl 4 . The removal<br />

percentages of As(V) at 1 mg/L Zr(IV) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 62, 28, <strong>and</strong> 8%,<br />

respectively, in the presence of silica, while in the absence of silica, the removals were 83, 56,<br />

<strong>and</strong> 30%, respectively for the same dose <strong>and</strong> pHs. Thus, removal efficiency was higher in the<br />

absence of silica. The adsorption isotherms in Figure 4.4 clearly indicate the greater As(V)<br />

adsorption in the absence of silica at all three pH values.<br />

q e (μg As(V)/mg Al(III))<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

pH 6.5, SiO2=0mg/L<br />

pH 6.5, SiO2=20mg/L<br />

pH 7.5, SiO2=0mg/L<br />

pH 7.5, SiO2=20mg/L<br />

pH 8.5, SiO2=20mg/L<br />

pH 8.5, SiO2=0mg/L<br />

0 10 20 30 40 50<br />

C e (μg As(V)/L)<br />

Figure 4.3 Effect of silica on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Al(III) hydroxide in NSFI water <strong>with</strong> phosphate<br />

34


q e (μg As(V)/mg Zr(IV))<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

pH 6.5, SiO2=0mg/L<br />

pH 6.5, SiO2=20mg/L<br />

pH 7.5, SiO2=0mg/L<br />

pH7.5, SiO2=20mg/L<br />

pH 8.5, SiO2=0mg/L<br />

pH8.5, SiO2=20mg/L<br />

0 10 20 30 40 50<br />

C e (μg As(V)/L)<br />

Figure 4.4 Effect of silica on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Zr(IV) hydroxide in NSFI water <strong>with</strong> phosphate<br />

q e (μg As(III)/mg Zr(IV))<br />

7<br />

6<br />

5<br />

4<br />

3<br />

2<br />

1<br />

0<br />

pH8.5, SiO2=0mg/L<br />

pH7.5, SiO2=0mg/L<br />

pH6.5, SiO2=0mg/L<br />

pH6.5, SiO2=20mg/L<br />

pH7.5, SiO2=20mg/L<br />

pH8.5, SiO2=20mg/L<br />

0 10 20 30 40 50 60<br />

C e (μg As(III)/L)<br />

Figure 4.5 Effect of silica on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Zr(IV) hydroxide in NSFI water <strong>with</strong> phosphate<br />

The As(III) adsorption isotherms shown in Figure 4.5 illustrate the negative influence of<br />

silica on arsenic adsorption onto Zr(IV) hydroxide at the three pHs tested: 6.5, 7.5, <strong>and</strong> 8.5. The<br />

As(III) adsorption behavior was similar to that observed for the adsorption of As(V) where the<br />

removal efficiency was higher in the absence of silica. In the absence of silica, the adsorption of<br />

As(III) was found to be pH dependent, <strong>and</strong> the removal efficiency of As(III) increased <strong>with</strong><br />

increasing pH, which was not observed in the presence of silica were the removal was the same<br />

regardless of pH. At 2 mg/L Zr(IV) dosage, approximately 8% of As(III) was removed<br />

irrespective of pH in the presence of silica. But in the absence of silica, removal efficiencies of<br />

35


16, 19, <strong>and</strong> 24% were obtained at pHs 6.5, 7.5, <strong>and</strong> 8.5, respectively. Thus, the As(III) removal<br />

efficiency was highly pH dependent in the absence of silica <strong>and</strong> the removal efficiencies<br />

increased <strong>with</strong> increasing pH.<br />

Effect of Silica in NSFI Water <strong>with</strong> Phosphate Using <strong>Titanium</strong> (IV) Chloride as Coagulant<br />

Based on jar tests in the presence <strong>and</strong> absence of silica, it was found that silica does have<br />

a significant impact on the adsorption of arsenic during coagulation <strong>with</strong> TiCl 4 . The As(V)<br />

removals at 1 mg/L Ti(IV) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 54, 40, <strong>and</strong> 26%, respectively, in<br />

the presence of silica (20 mg/L as SiO 2 ), while in the absence of silica, the removals were 73, 57,<br />

<strong>and</strong> 42%, respectively for the same dose <strong>and</strong> pHs. Clearly, the As(V) removal efficiency was<br />

higher in the absence of silica as shown by the isotherms in Figure 4.6.<br />

In the absence of silica, the adsorption of As(III) was pH dependent, <strong>and</strong> its removal<br />

efficiency increased <strong>with</strong> increasing pH, which was not observed in the presence of silica were<br />

the removal was the same regardless of pH. For example, at 2 mg/L Ti(IV) dosage, As(III)<br />

removal in the presence of silica was 26%, irrespective of pH. But, in the absence of silica,<br />

removal efficiencies of 49, 51, <strong>and</strong> 53% were obtained at pHs 6.5, 7.5, <strong>and</strong> 8.5, respectively. The<br />

As(III) adsorption isotherms shown in Figure 4.7 illustrate the negative influence of silica on<br />

arsenic adsorption onto Ti(IV) hydroxide at all three pHs.<br />

qe (μg As(V)/mg Ti(IV))<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

pH6.5, SiO2=0mg/L<br />

pH7.5, SiO2=0mg/L<br />

pH6.5, SiO2=20mg/L<br />

pH8.5, SiO2=0mg/L<br />

pH7.5, SiO2=20mg/L<br />

pH8.5, SiO2=20mg/L<br />

0 5 10 15 20 25 30 35 40<br />

C e (μg As(V)/L)<br />

Figure 4.6 Effect of silica on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Ti(IV) hydroxide in NSFI water <strong>with</strong> phosphate<br />

36


q e (μg As(III)/mg Ti(IV))<br />

14<br />

12<br />

10<br />

8<br />

6<br />

4<br />

2<br />

0<br />

pH8.5, SiO2=0mg/L<br />

pH7.5, SiO2=0mg/L<br />

pH6.5, SiO2=0mg/L<br />

pH6.5, SiO2=20mg/L<br />

pH7.5, SiO2=20mg/L<br />

pH8.5, SiO2=20mg/L<br />

0 5 10 15 20 25 30 35 40<br />

C e (μg As(III)/L)<br />

Figure 4.7 Effect of silica on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Ti(IV) hydroxide in NSFI water <strong>with</strong> phosphate<br />

Effect of Silica in NSFI Water <strong>with</strong> Phosphate Using <strong>Titanium</strong> (III) Chloride as Coagulant<br />

Based on jar tests in the presence <strong>and</strong> absence of silica, it was found that silica does have<br />

a significant impact on the adsorption of arsenic during coagulation <strong>with</strong> TiCl 3 . The removals of<br />

As(V) at 1 mg/L Ti(III) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 49, 37, <strong>and</strong> 30%, respectively, in<br />

the presence of silica, while in the absence of silica, higher removals of 77, 58, <strong>and</strong> 48%,<br />

respectively, were observed for the same dose <strong>and</strong> pHs. Thus, As(V) removal by TiCl 3 was<br />

higher in the absence of silica. This can be observed in Figure 4.8 through the more favorable<br />

adsorption isotherms in the absence of silica.<br />

qe (μg As(V)/mg Ti(III))<br />

45<br />

40<br />

35<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

pH6.5, SiO2=0mg/L<br />

pH7.5, SiO2=0mg/L<br />

pH6.5, SiO2=20mg/L<br />

pH8.5, SiO2=0mg/L<br />

pH7.5, SiO2=20mg/L<br />

pH8.5, SiO2=20mg/L<br />

0 5 10 15 20 25 30 35 40 45<br />

C e (μg As(V)/L)<br />

Figure 4.8 Effect of silica on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Ti(III) hydroxide in NSFI water <strong>with</strong> phosphate<br />

37


qe (μg As(III)/mg Ti(III))<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

pH6.5, SiO2=0mg/L<br />

pH7.5, SiO2=0mg/L<br />

pH8.5, SiO2=0mg/L<br />

pH6.5, SiO2=20mg/L<br />

pH7.5, SiO2=20mg/L<br />

pH8.5, SiO2=20mg/L<br />

0 5 10 15 20 25 30 35 40 45<br />

C e (μg As(III)/L)<br />

Figure 4.9 Effect of silica on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Ti(III) hydroxide in NSFI water <strong>with</strong> phosphate<br />

Although As(III) adsorption was lower, its behavior <strong>with</strong> silica absent was similar to that<br />

observed for the adsorption of As(V). In the presence <strong>and</strong> absence of silica, the adsorption of<br />

As(III) increased <strong>with</strong> decreasing pH. For example, in the presence of silica <strong>and</strong> <strong>with</strong> 2 mg/L<br />

Ti(III) dosage, As(III) removals of 42, 32, <strong>and</strong> 27% were obtained at pHs 6.5, 7.5, <strong>and</strong> 8.5,<br />

respectively. In the absence of silica, As(III) removals of 76, 61, <strong>and</strong> 51% were obtained at the<br />

same dose <strong>and</strong> pHs. As was observed earlier, the As(III) concentration remaining after Ti(III)<br />

coagulation was far less than that of the total arsenic. This was due to oxidation of As(III) to<br />

As(V) in the presence of Ti(III) during coagulation, as was described in Chapter 3. Thus, the<br />

As(III) removal efficiency was pH dependent both in the presence <strong>and</strong> absence of silica <strong>and</strong> the<br />

As(III) removal was higher in the absence of silica. The As(III) adsorption isotherms shown in<br />

Figure 4.9 illustrate the negative influence of silica on arsenic adsorption onto Ti(III) hydroxide<br />

at all three pHs.<br />

Thus, in general it was found that silica does reduce the adsorption of arsenic<br />

(As(V)/As(III)) <strong>and</strong> the effect increases <strong>with</strong> increasing pH <strong>with</strong> the following coagulants tested:<br />

FeCl 3 , ZrCl 4 , TiCl 4 , <strong>and</strong> TiCl 3 . In the case of alum as coagulant, silica did have a significant<br />

effect on As(V) adsorption at pH 6.5-7.5, while it had little if any effect at pH 8.5.<br />

EFFECT OF PHOSPHATE IN NSFI CHALLENGE WATER CONTAINING SILICA<br />

<strong>Coagulation</strong> experiments were first performed using challenge water <strong>with</strong> phosphate<br />

(PO 4 -P = 40 μg/L) <strong>and</strong> <strong>with</strong>out phosphate (PO 4 -P = 0 μg/L) in the presence of silica. The overall<br />

objective was to study the effect of phosphate in the NSFI challenge water containing silica on<br />

arsenic removal in the pH range of 6.5 to 8.5 using five coagulants: FeCl 3 , alum, ZrCl 4 , TiCl 4 ,<br />

<strong>and</strong> TiCl 3 .<br />

38


Effect of Phosphate in NSFI Water <strong>with</strong> Silica Using Ferric (III) Chloride as Coagulant<br />

For FeCl 3 coagulation, the presence of phosphate in NSFI water <strong>with</strong> silica had a<br />

significant effect at pH 6.5-7.5 on the adsorption of As(V) <strong>and</strong> no significant effect on the<br />

adsorption of As(III). The removals of As(V) in the presence of phosphate at 1 mg/L Fe(III)<br />

dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 90, 76, <strong>and</strong> 46%, respectively. In the absence of phosphate,<br />

the removals at the same pHs were 98, 88, <strong>and</strong> 51%, respectively for the same dose, which<br />

shows that the decrease in removal efficiency was significant in the presence of phosphate at pHs<br />

6.5 <strong>and</strong> 7.5 <strong>and</strong> was less significant at pH 8.5. This can be observed in Figure 4.10 through the<br />

more favorable adsorption isotherms in the absence of phosphate at pHs of 6.5 <strong>and</strong> 7.5 as<br />

compared <strong>with</strong> the influence of phosphate on the adsorption isotherms at pH 8.5.<br />

The presence or absence of phosphate in the NSFI water <strong>with</strong> silica had no significant<br />

effect on As(III) adsorption, which was found to be pH independent using FeCl 3 as coagulant. At<br />

2 mg/L Fe(III) dosage, approximately 26% of As(III) was removed irrespective of pH in the<br />

presence of phosphate. In the absence of phosphate, approximately 29% of As(III) was removed<br />

irrespective of pH. The As(III) adsorption isotherms shown in Figure 4.11 illustrate no<br />

significant influence of phosphate or pH on As(III) adsorption onto Fe(III) hydroxide at the three<br />

pHs tested: 6.5, 7.5, <strong>and</strong> 8.5.<br />

qe (μg As(V)/mg Fe(III))<br />

180<br />

160<br />

140<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

pH6.5, PO4-P=0ug/L<br />

pH7.5, PO4-P=0ug/L<br />

pH6.5, PO4-P=40ug/L<br />

pH7.5, PO4-P=40ug/L<br />

pH8.5, PO4-P=0ug/L<br />

pH8.5, PO4-P=40ug/L<br />

0 5 10 15 20 25 30<br />

C e (μg As(V)/L)<br />

Figure 4.10 Effect of phosphate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Fe(III) hydroxides in NSFI water <strong>with</strong> silica<br />

39


q e (μ g As(III)/mg of Fe(III))<br />

9<br />

8<br />

7<br />

6<br />

5<br />

4<br />

3<br />

2<br />

1<br />

0<br />

pH8.5, PO4-P=0ug/L<br />

pH7.5, PO4-P=0ug/L<br />

pH6.5, PO4-P=0ug/L<br />

pH8.5, PO4-P=40ug/L<br />

pH7.5, PO4-P=40ug/L<br />

pH6.5, PO4-P=40ug/L<br />

0 10 20 30 40 50<br />

C e (μg As(III)/L)<br />

Figure 4.11 Effect of phosphate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Fe(III) hydroxides in NSFI water <strong>with</strong> silica<br />

Effect of Phosphate in NSFI Water <strong>with</strong> Silica Using Alum as Coagulant<br />

When silica was present in the NSFI water, phosphate had no significant effect on the<br />

adsorption of As(V) during coagulation <strong>with</strong> alum. The removals of As(V) in the presence of<br />

phosphate at 1 mg/L Al(III) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 85, 49, <strong>and</strong> 20%, respectively.<br />

In the absence of phosphate, the removals at the same pHs were 88, 53, <strong>and</strong> 19%, respectively<br />

for the same dose. This can be observed in Figure 4.12 through similar adsorption isotherms in<br />

the presence <strong>and</strong> absence of phosphate at all pH 6.5, 7.5, <strong>and</strong> 8.5.<br />

qe (μg As(V)/mg Al(III))<br />

90<br />

75<br />

60<br />

45<br />

30<br />

15<br />

0<br />

pH 6.5, PO4-P=40ug/L<br />

pH 6.5, PO4-P=0ug/L<br />

pH 7.5, PO4-P=40ug/L<br />

pH 7.5, PO4-P=0ug/L<br />

pH 8.5, PO4-P=40ug/L<br />

pH 8.5, PO4-P=0ug/L<br />

0 10 20 30 40 50<br />

C e (μg As(V)/L)<br />

Figure 4.12 Effect of phosphate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Al(III) hydroxides in NSFI water <strong>with</strong> silica<br />

40


Effect of Phosphate in NSFI Water <strong>with</strong> Silica Using <strong>Zirconium</strong> (IV) Chloride as<br />

Coagulant<br />

When silica was present in the NSFI water, phosphate had a significant effect at pH 6.5-<br />

7.5 on the adsorption of As(V) <strong>and</strong> no significant effect on the adsorption of As(III) during<br />

coagulation <strong>with</strong> ZrCl 4 . The removals of As(V) in the presence of phosphate at 1 mg/L Zr(IV)<br />

dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 62, 28, <strong>and</strong> 8%, respectively. In the absence of phosphate,<br />

the removals at the same pHs were 79, 41, <strong>and</strong> 16%, respectively for the same dose, which<br />

shows that the decrease in removal efficiency was significant in the presence of phosphate at pHs<br />

6.5 <strong>and</strong> 7.5 <strong>and</strong> was less significant at pH 8.5. This can be observed in Figure 4.13 through the<br />

qe (μg As(V)/mg Zr(IV))<br />

50<br />

45<br />

40<br />

35<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

pH6.5, PO4-P=0ug/L<br />

pH6.5, PO4-P=40ug/L<br />

pH7.5, PO4-P=0ug/L<br />

pH7.5, PO4-P=40ug/L<br />

pH8.5, PO4-P=0ug/L<br />

pH8.5, PO4-P=40ug/L<br />

0 10 20 30 40 50<br />

C e (μg As(V)/L)<br />

Figure 4.13 Effect of phosphate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Zr(IV) hydroxide in NSFI water <strong>with</strong> silica<br />

qe (μg As(III)/mg Zr(IV))<br />

9<br />

8<br />

7<br />

6<br />

5<br />

4<br />

3<br />

2<br />

1<br />

0<br />

pH8.5, PO4-P=0ug/L<br />

pH7.5, PO4-P=0ug/L<br />

pH6.5, PO4-P=0ug/L<br />

pH8.5, PO4-P=40ug/L<br />

pH7.5, PO4-P=40ug/L<br />

pH6.5, PO4-P=40ug/L<br />

0 10 20 30 40 50 60<br />

C e (μg As(III)/L)<br />

Figure 4.14 Effect of phosphate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Zr(IV) hydroxides in NSFI water <strong>with</strong> silica<br />

41


more favorable adsorption isotherms in the absence of phosphate at pHs of 6.5 <strong>and</strong> 7.5 as<br />

compared <strong>with</strong> the influence of phosphate on the adsorption isotherms at pH 8.5.<br />

When silica was present in the NSFI water, phosphate did not significantly effect As(III)<br />

adsorption, which was independent of pH using ZrCl 4 coagulant. At 2 mg/L Zr(IV) dosage,<br />

approximately 8% of As(III) was removed irrespective of pH in the presence of phosphate. In the<br />

absence of phosphate, approximately 7% of As(III) was removed irrespective of pH in the<br />

presence of phosphate. The As(III) adsorption isotherms in Figure 4.14 demonstrate very little<br />

As(III) adsorption <strong>and</strong> no significant effect of phosphate or pH during Zr(IV) coagulation at the<br />

three pHs tested: 6.5, 7.5, <strong>and</strong> 8.5.<br />

Effect of Phosphate in NSFI Water <strong>with</strong> Silica Using <strong>Titanium</strong> (IV) Chloride as Coagulant<br />

When silica was present in the NSFI water, phosphate had a significant effect at pH 6.5-<br />

7.5 on the adsorption of As(V) <strong>and</strong> no significant effect on the adsorption of As(III) during<br />

coagulation <strong>with</strong> TiCl 4 . The removals of As(V) at 1 mg/L Ti(IV) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5<br />

were 54, 40, <strong>and</strong> 26%, respectively, in the presence of phosphate. In the absence of phosphate,<br />

the removals at the same pHs were 65, 47, <strong>and</strong> 30%, respectively for the same dose, which<br />

shows that the decrease in removal efficiency was significant in the presence of phosphate at pHs<br />

6.5 <strong>and</strong> 7.5 <strong>and</strong> was less significant at pH 8.5. This can be observed in Figure 4.15 through the<br />

more favorable adsorption isotherms in the absence of phosphate at pHs of 6.5 <strong>and</strong> 7.5 as<br />

compared <strong>with</strong> the lesser influence of phosphate on the adsorption isotherms at pH 8.5.<br />

The presence or absence of phosphate in the NSFI water <strong>with</strong> silica had little effect on<br />

As(III) adsorption, which was found to be pH independent using TiCl 4 as coagulant. At 2 mg/L<br />

Ti(IV) dosage, approximately 26% of As(III) was removed irrespective of pH in the presence of<br />

phosphate. In the absence of phosphate, removal efficiencies of 31, 30, <strong>and</strong> 26% were obtained<br />

at pHs 6.5, 7.5, <strong>and</strong> 8.5, respectively. The As(III) adsorption isotherms shown in Figure 4.16<br />

illustrate little if any influence of phosphate or pH on As(III) adsorption onto Ti (IV) hydroxide<br />

in the 6.5-8.5 pH range.<br />

qe (μg As(V)/mg Ti(IV))<br />

35<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

pH6.5, PO4-P=0ug/L<br />

pH6.5, PO4-P=40ug/L<br />

pH7.5, PO4-P=0ug/L<br />

pH7.5, PO4-P=40ug/L<br />

pH8.5, PO4-P=0ug/L<br />

pH8.5, PO4-P=40ug/L<br />

0 5 10 15 20 25 30 35 40 45<br />

C e (μg As(V)/L)<br />

Figure 4.15 Effect of phosphate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Ti(IV) hydroxides in NSFI water <strong>with</strong> silica<br />

42


qe (μg As(III)/mg Ti(IV))<br />

8<br />

7<br />

6<br />

5<br />

4<br />

3<br />

2<br />

1<br />

0<br />

pH6.5, PO4-P=0ug/L<br />

pH7.5, PO4-P=0ug/L<br />

pH8.5, PO4-P=0ug/L<br />

pH6.5, PO4-P=40ug/L<br />

pH7.5, PO4-P=40ug/L<br />

pH8.5, PO4-P=40ug/L<br />

0 5 10 15 20 25 30 35 40<br />

C e (μg As(III)/L)<br />

Figure 4.16 Effect of phosphate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Ti(IV) hydroxides in NSFI water <strong>with</strong> silica<br />

Effect of Phosphate in NSFI Water <strong>with</strong> Silica Using <strong>Titanium</strong> (III) Chloride as Coagulant<br />

When silica was present in the NSFI water, phosphate had a significant effect at low pH<br />

on the adsorption of As(V) <strong>and</strong> no significant influence on the adsorption of As(III) during<br />

coagulation <strong>with</strong> TiCl 3 . The removals of As(V) in the presence of phosphate at 1 mg/L Ti(III)<br />

dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 49, 37, <strong>and</strong> 30%, respectively. In the absence of phosphate,<br />

the removals at the same pHs were 66, 51, <strong>and</strong> 36%, respectively for the same dose, which<br />

shows that the decrease in removal efficiency was significant in the presence of phosphate at pHs<br />

6.5 <strong>and</strong> 7.5 <strong>and</strong> was less significant at pH 8.5. This can be observed in Figure 4.17 through the<br />

more favorable adsorption isotherms in the absence of phosphate at pHs of 6.5 <strong>and</strong> 7.5 as<br />

compared <strong>with</strong> the influence of phosphate on the adsorption isotherms at pH 8.5.<br />

When silica was present in the NSFI water, phosphate had no significant effect on<br />

As(III) adsorption using TiCl 3 coagulant. In the presence or absence of phosphate, the adsorption<br />

of As(III) was found to be pH dependent, <strong>and</strong> the removal efficiency of As(III) increased <strong>with</strong><br />

decreasing pH. At 2 mg/L Ti(III) dosage, As(III) removal efficiencies of 42, 32, <strong>and</strong> 27% were<br />

obtained at pHs 6.5, 7.5, <strong>and</strong> 8.5, respectively in the presence of phosphate, while in the absence<br />

of phosphate, removal efficiencies of 41, 33, <strong>and</strong> 25% were obtained at the same pHs <strong>and</strong> dose,<br />

respectively. Thus, the As(III) removal efficiency was the same <strong>and</strong> was pH dependent both in<br />

the presence <strong>and</strong> absence of phosphate in NSFI water <strong>with</strong> silica. As was observed earlier, the<br />

As(III) concentration remaining after Ti(III) coagulation was far less than that of the total<br />

arsenic. This is due to oxidation of As(III) to As(V) in the presence of Ti(III) during coagulation,<br />

which was described earlier. The As(III) adsorption isotherms shown in Figure 4.18 illustrate no<br />

significant influence of phosphate on As(III) adsorption onto Ti(III) hydroxide at the three pHs<br />

tested: 6.5, 7.5, <strong>and</strong> 8.5.<br />

43


qe (μg As(V)/mg Ti(III))<br />

40<br />

35<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

pH6.5, PO4-P=0ug/L<br />

pH6.5, PO4-P=40ug/L<br />

pH7.5, PO4-P=0ug/L<br />

pH7.5, PO4-P=40ug/L<br />

pH8.5, PO4-P=0ug/L<br />

pH8.5, PO4-P=40ug/L<br />

0 5 10 15 20 25 30 35 40 45<br />

C e (μg As(V)/L)<br />

Figure 4.17 Effect of phosphate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Ti(III) hydroxides in NSFI water <strong>with</strong> silica<br />

12<br />

q e (μg As(III)/mg Ti(III))<br />

10<br />

8<br />

6<br />

4<br />

2<br />

0<br />

pH 6.5, PO4-P=0ug/L<br />

pH 6.5, PO4-P=40ug/L<br />

pH 7.5, PO4-P=0ug/L<br />

pH 7.5, PO4-P=40ug/L<br />

pH 8.5, PO4-P=0ug/L<br />

pH 8.5, PO4-P=40ug/L<br />

0 5 10 15 20 25 30 35 40 45<br />

C e (μg As(III)/L)<br />

Figure 4.18 Effect of phosphate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Ti(III) hydroxides in NSFI water <strong>with</strong> silica<br />

Thus, in NSFI challenge water containing silica, it was found that phosphate significantly<br />

lowers the adsorption of As(V) at pH 6.5-7.5 but does not influence the adsorption of arsenic(III)<br />

in the 6.5-8.5 pH range <strong>with</strong> any of the coagulants tested: FeCl 3 , ZrCl 4 , TiCl 4 <strong>and</strong> TiCl 3 .<br />

However <strong>with</strong> alum as coagulant, their was no significant effect of phosphate on As(V)<br />

adsorption at any pH tested: 6.5, 7.5, <strong>and</strong> 8.5.<br />

44


EFFECT OF VANADATE IN NSFI CHALLENGE WATER CONTAINING SILICA<br />

AND PHOSPHATE<br />

<strong>Coagulation</strong> experiments were performed using the st<strong>and</strong>ard NSFI challenge water<br />

<strong>with</strong>out vanadate (VO 3 -V = 0 μg/L). Then, the st<strong>and</strong>ard challenge water containing silica <strong>and</strong><br />

phosphate was spiked <strong>with</strong> vanadate (VO 3 -V = 50 μg/L) <strong>and</strong> tested again. The objective was to<br />

study the effect of vanadate in the NSFI challenge water on arsenic removal in the pH range of<br />

6.5 to 8.5 using all five coagulants: FeCl 3 , alum, ZrCl 4 , TiCl 4 , <strong>and</strong> TiCl 3 .<br />

Effect of Vanadate in NSFI Water <strong>with</strong> Silica <strong>and</strong> Phosphate Using Ferric (III) Chloride as<br />

Coagulant<br />

Based on jar tests in the presence <strong>and</strong> absence of vanadate, it was found that vanadate<br />

does not significantly influence the adsorption of As(V) or As(III) in the st<strong>and</strong>ard NSFI<br />

challenge water containing silica <strong>and</strong> phosphate during coagulation <strong>with</strong> FeCl 3 . The removals of<br />

As(V) at 1 mg/L Fe(III) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 90, 76, <strong>and</strong> 46%, respectively, in<br />

the absence of vanadate. In the presence of vanadate, the removals at the same pHs were 89, 74,<br />

<strong>and</strong> 37%, respectively for the same dose. The As(V) adsorption isotherms shown in Figure 4.19<br />

illustrate no significant influence of vanadate on As(V) adsorption onto Fe(III) hydroxide at the<br />

three pHs tested: 6.5, 7.5, <strong>and</strong> 8.5.<br />

The presence or absence of vanadate in the NSFI water <strong>with</strong> silica <strong>and</strong> phosphate had no<br />

significant effect on As(III) adsorption <strong>and</strong> was found to be pH independent using FeCl 3 as<br />

coagulant. At 2 mg/L Fe(III) dosage, approximately 25% of As(III) was removed irrespective of<br />

pH in the presence of vanadate. In the absence of vanadate, approximately 26% of As(III) was<br />

removed irrespective of pH. Thus, the As(III) removal efficiency was not pH dependent, nor was<br />

it significantly affected by the presence of vanadate in NSFI water containing silica <strong>and</strong><br />

phosphate. The As(III) adsorption isotherms shown in Figure 4.20 illustrate no significant<br />

influence of vanadate or pH on As(III) adsorption onto Fe(III) hydroxide at the three pHs tested:<br />

6.5, 7.5, <strong>and</strong> 8.5.<br />

q e (μg As(V)/mg Fe(III))<br />

80<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

pH6.5, VO3-V=0ug/L<br />

pH6.5, VO3-V=50ug/L<br />

pH7.5, VO3-V=0ug/L<br />

pH7.5, VO3-V=50ug/L<br />

pH8.5, VO3-V=0ug/L<br />

pH8.5, VO3-V=50ug/L<br />

0 5 10 15 20 25 30 35<br />

C e (μg As(V)/L)<br />

Figure 4.19 Effect of vanadate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Fe(III) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate<br />

45


q e (μg As(III)/mg Fe(III))<br />

8<br />

7<br />

6<br />

5<br />

4<br />

3<br />

2<br />

1<br />

0<br />

pH8.5, VO3-V=0ug/L<br />

pH7.5, VO3-V=0ug/L<br />

pH6.5, VO3-V=0ug/L<br />

pH8.5, VO3-V=50ug/L<br />

pH7.5, VO3-V=50ug/L<br />

pH6.5, VO3-V=50ug/L<br />

0 5 10 15 20 25 30 35 40 45<br />

Figure 4.20 Effect of vanadate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Fe(III) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate<br />

Effect of Vanadate in NSFI Water <strong>with</strong> Silica <strong>and</strong> Phosphate Using Alum as Coagulant<br />

Based on jar tests in the presence <strong>and</strong> absence of vanadate, it was found that vanadate<br />

does not significantly influence the adsorption of As(V) in the st<strong>and</strong>ard NSFI challenge water<br />

containing silica <strong>and</strong> phosphate during coagulation <strong>with</strong> alum. The removals of As(V) at 1 mg/L<br />

Al(III) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 85, 49, <strong>and</strong> 20%, respectively, in the absence of<br />

vanadate. In the presence of vanadate, the removals at the same pHs were 82, 54, <strong>and</strong> 15%,<br />

respectively for the same dose. Thus there was not a significant decrease in removal efficiency in<br />

the presence of vanadate in NSFI water containing silica <strong>and</strong> phosphate. The As(V) adsorption<br />

isotherms shown in Figure 4.21 illustrate no significant influence of vanadate on As(V)<br />

adsorption onto Al(III) hydroxide at the three pHs tested: 6.5, 7.5, <strong>and</strong> 8.5.<br />

q e (μg As(V)/mg Al(III))<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

pH 6.5, VO3-V=0ug/L<br />

pH 6.5, VO3-V=50ug/L<br />

pH 7.5, VO3-V=0ug/L<br />

pH 7.5, VO3-V=50ug/L<br />

pH 8.5, VO3-V=0ug/L<br />

pH 8.5, VO3-V=50ug/L<br />

0 10 20 30 40 50<br />

C e (μg As(V)/L)<br />

Figure 4.21 Effect of vanadate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Al(III) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate<br />

46


Effect of Vanadate in NSFI Water <strong>with</strong> Silica <strong>and</strong> Phosphate Using <strong>Zirconium</strong> (IV)<br />

Chloride as Coagulant<br />

Based on ZrCl 4 coagulation jar tests in the presence <strong>and</strong> absence of vanadate, it was<br />

found that vanadate does not significantly influence the adsorption of As(V) or As(III) in NSFI<br />

challenge water containing silica <strong>and</strong> phosphate. The removals of As(V) at 1 mg/L Zr(IV)<br />

dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 62, 28, <strong>and</strong> 8%, respectively, in the absence of vanadate. In<br />

the presence of vanadate, the removals at the same pHs were 59, 28, <strong>and</strong> 8%, respectively for the<br />

same dose. Thus there was no significant effect of vanadate in NSFI water containing silica <strong>and</strong><br />

phosphate. The As(V) adsorption isotherms shown in Figure 4.22 illustrate insignificant effect of<br />

vanadate on As(V) adsorption onto Zr(IV) hydroxide at the three pHs tested: 6.5, 7.5, <strong>and</strong> 8.5.<br />

qe (μg As(V)/mg Zr(IV))<br />

40<br />

35<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

pH6.5, VO3-V=0ug/L<br />

pH6.5, VO3-V=50ug/L<br />

pH7.5, VO3-V=0ug/L<br />

pH7.5, VO3-V=50ug/L<br />

pH8.5, VO3-V=0ug/L<br />

pH8.5, VO3-V=50ug/L<br />

0 10 20 30 40 50<br />

C e (μg As(V)/L)<br />

Figure 4.22 Effect of vanadate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Zr(IV) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate<br />

q e (μg As(III)/mg Zr(IV))<br />

7<br />

6<br />

5<br />

4<br />

3<br />

2<br />

1<br />

0<br />

pH6.5, VO3-V=0ug/L<br />

pH7.5, VO3-V=0ug/L<br />

pH8.5, VO3-V=0ug/L<br />

pH6.5, VO3-V=50ug/L<br />

pH7.5, VO3-V=50ug/L<br />

pH8.5, VO3-V=50ug/L<br />

0 10 20 30 40 50 60<br />

C e (μg As(III)/L)<br />

Figure 4.23 Effect of vanadate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Zr(IV) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate<br />

47


The presence or absence of vanadate in the NSFI water <strong>with</strong> silica <strong>and</strong> phosphate had no<br />

significant effect on As(III) adsorption <strong>and</strong> was pH independent using ZrCl 4 as coagulant. At 2<br />

mg/L Zr(IV) dosage, approximately 6% of As(III) was removed irrespective of pH in the<br />

presence of vanadate. In the absence of vanadate, approximately 8% of As(III) was removed<br />

irrespective of pH. Thus, the As(III) removal efficiency was not pH dependent, nor was it<br />

significantly affected by the presence of vanadate in NSFI water containing silica <strong>and</strong> phosphate.<br />

The As(III) adsorption isotherms shown in Figure 4.23 illustrate no significant effect of vanadate<br />

or pH on As(III) adsorption onto Zr(IV) hydroxide at the three pHs tested: 6.5, 7.5, <strong>and</strong> 8.5.<br />

Effect of Vanadate in NSFI Water <strong>with</strong> Silica <strong>and</strong> Phosphate Using <strong>Titanium</strong> (IV) Chloride<br />

as Coagulant<br />

Based on jar tests in the presence <strong>and</strong> absence of vanadate, it was found that vanadate<br />

does not significantly affect the adsorption of As(V) or As(III) in challenge water containing<br />

silica <strong>and</strong> phosphate during coagulation <strong>with</strong> TiCl 4 . The removals of As(V) at 1 mg/L Ti(IV)<br />

dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 54, 40, <strong>and</strong> 26%, respectively, in the absence of vanadate. In<br />

the presence of vanadate, the removals at the same pHs were 54, 39, <strong>and</strong> 25%, respectively for<br />

the same dose. Thus, there was no effect of vanadate in NSFI water containing silica <strong>and</strong><br />

phosphate. The As(V) adsorption isotherms shown in Figure 4.24 illustrate insignificant effect of<br />

vanadate on As(V) adsorption onto Ti(IV) hydroxide at the three pHs tested: 6.5, 7.5, <strong>and</strong> 8.5.<br />

When coagulating <strong>with</strong> TiCl 4 , the presence or absence of vanadate in the NSFI water<br />

<strong>with</strong> silica <strong>and</strong> phosphate did not significantly affect As(III) adsorption, which was independent<br />

of pH. At 2 mg/L Ti(IV) dosage, approximately 26% of As(III) was removed irrespective of pH<br />

in the presence of vanadate, while in the absence of vanadate, approximately 26% of As(III) was<br />

removed, again, irrespective of pH. The As(III) adsorption isotherms shown in Figure 4.25<br />

illustrate no significant influence of vanadate or pH on As(III) adsorption onto Ti(IV) hydroxide<br />

at the three pHs tested: 6.5, 7.5, <strong>and</strong> 8.5.<br />

q e (μg As(V)/mg Ti(IV))<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

pH6.5, VO3-V=0ug/L<br />

pH6.5, VO3-V=50ug/L<br />

pH7.5, VO3-V=0ug/L<br />

pH7.5, VO3-V=50ug/L<br />

pH8.5, VO3-V=0ug/L<br />

pH8.5, VO3-V=50ug/L<br />

0 5 10 15 20 25 30 35 40<br />

C e (μg As(V)/L)<br />

Figure 4.24 Effect of vanadate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Ti(IV) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate<br />

48


qe (μg As(III)/mg Ti(IV))<br />

7<br />

6<br />

5<br />

4<br />

3<br />

2<br />

1<br />

0<br />

pH6.5, VO3-V=0ug/L<br />

pH7.5, VO3-V=0ug/L<br />

pH8.5, VO3-V=0ug/L<br />

pH6.5, VO3-V=50ug/L<br />

pH7.5, VO3-V=50ug/L<br />

pH8.5, VO3-V=50ug/L<br />

0 5 10 15 20 25 30 35 40<br />

C e (μg As(III)/L)<br />

Figure 4.25 Effect of vanadate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Ti(IV) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate<br />

Effect of Vanadate in NSFI Water <strong>with</strong> Silica <strong>and</strong> Phosphate Using <strong>Titanium</strong> (III) Chloride<br />

as Coagulant<br />

Based on TiCl 3 coagulation jar tests in the presence <strong>and</strong> absence of vanadate, it was<br />

found that vanadate does not significantly influence the adsorption of As(V) or As(III) in NSFI<br />

challenge water containing silica <strong>and</strong> phosphate. The removals of As(V) at 2 mg/L Ti(III) dosage<br />

at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 49, 37, <strong>and</strong> 30%, respectively in the absence of vanadate, while in<br />

the presence of vanadate, the removals were 47, 35, <strong>and</strong> 28%, respectively for the same dose <strong>and</strong><br />

pHs. Although As(V) adsorption decreased <strong>with</strong> increasing pH, the As(V) adsorption isotherms<br />

shown in Figure 4.26 illustrate no significant influence of vanadate on As(V) adsorption onto<br />

Ti(III) hydroxide at the three pHs tested: 6.5, 7.5, <strong>and</strong> 8.5.<br />

Based on TiCl 3 coagulation tests, the presence or absence of vanadate in the NSFI water<br />

<strong>with</strong> silica <strong>and</strong> phosphate was found to have no significant effect on As(III) adsorption. In the<br />

presence <strong>and</strong> absence of vanadate, the adsorption of As(III) was found to be pH dependent, <strong>and</strong><br />

the removal of As(III) increased <strong>with</strong> decreasing pH. At 2 mg/L Ti(III) dosage, As(III) removals<br />

of 42, 32, <strong>and</strong> 27% were obtained at pHs 6.5, 7.5, <strong>and</strong> 8.5, respectively in the absence of<br />

vanadate, while in the presence of vanadate, removal efficiencies of 42, 30, <strong>and</strong> 25% were<br />

obtained at the same pHs <strong>and</strong> dose, respectively. Thus, the As(III) removal was the same <strong>and</strong> pH<br />

dependent both in the presence <strong>and</strong> absence of vanadate in NSFI water containing silica <strong>and</strong><br />

phosphate. The As(III) adsorption isotherms shown in Figure 4.27 illustrate no significant effect<br />

of vanadate on As(III) adsorption onto Ti(III) hydroxide at the three pHs tested: 6.5, 7.5, <strong>and</strong> 8.5.<br />

Thus, for the five coagulants tested (FeCl 3 , alum, ZrCl 4 , TiCl 4 , <strong>and</strong> TiCl 3 ) in NSFI<br />

challenge water containing silica <strong>and</strong> phosphate, it was found that vanadate does not affect the<br />

adsorption of arsenic(V) in the 6.5-8.5 pH range. Similarly, vanadate did not affect the<br />

adsorption of arsenic(III) in NSFI challenge water containing silica <strong>and</strong> phosphate in the same<br />

pH range <strong>with</strong> the four coagulants tested (FeCl 3 , , ZrCl 4 , TiCl 4 , <strong>and</strong> TiCl 3 ).<br />

49


qe (μg As(V)/mg Ti(III))<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

pH6.5, VO3-V=0ug/L<br />

pH6.5, VO3-V=50ug/L<br />

pH7.5, VO3-V=0ug/L<br />

pH7.5, VO3-V=50ug/L<br />

pH8.5, VO3-V=0ug/L<br />

pH8.5, VO3-V=50ug/L<br />

0 5 10 15 20 25 30 35 40 45<br />

C e (μg As(V)/L)<br />

Figure 4.26 Effect of vanadate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Ti(III) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate<br />

qe (μg As(III)/mg Ti(III))<br />

11<br />

10<br />

9<br />

8<br />

7<br />

6<br />

5<br />

4<br />

3<br />

2<br />

1<br />

0<br />

pH6.5, VO3-V=0ug/L<br />

pH6.5, VO3-V=50ug/L<br />

pH7.5, VO3-V=0ug/L<br />

pH7.5, VO3-V=50ug/L<br />

pH8.5, VO3-V=0ug/L<br />

pH8.5, VO3-V=50ug/L<br />

0 5 10 15 20 25 30 35 40 45<br />

C e (μg As(III)/L)<br />

Figure 4.27 Effect of vanadate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Ti(III) hydroxides in NSFI water <strong>with</strong> silica <strong>and</strong> phosphate<br />

INDIVIDUAL EFFECTS OF SILICA, PHOSPHATE AND VANADATE WITH FERRIC<br />

(III) CHLORIDE AS COAGULANT IN NSFI CHALLENGE WATER WITHOUT<br />

COMPETING IONS<br />

Although not included of the original research plan, limited coagulation experiments<br />

were performed <strong>with</strong> ferric chloride to determine the effects of silicate, phosphate, <strong>and</strong> vanadate<br />

ions in the absence of one another. By contrast, the previously reported effects of silica,<br />

phosphate, or vanadate had been determined in the presence of other competing ions, e.g., the<br />

50


effect of silica (0 <strong>and</strong> 20 mg/L SiO 2 ) had been determined using NSFI challenge water<br />

containing phosphate, <strong>and</strong> the effect of phosphate (0 <strong>and</strong> 40 μg P/L) had been determined in the<br />

challenge water containing silica. So, to study the individual effect of silica on arsenic adsorption<br />

in the absence of competing ions, experiments were performed using NSFI water <strong>with</strong> silica<br />

(SiO2 = 20 mg/L) <strong>and</strong> <strong>with</strong>out silica (SiO2 = 0 mg/L), in the absence of phosphate <strong>and</strong> vanadate<br />

in the pH range of 6.5 to 8.5 <strong>with</strong> ferric chloride as coagulant. Similarly, the effect of phosphate<br />

on arsenic removal in the absence of silica <strong>and</strong> vanadate, <strong>and</strong> the effect of vanadate in the<br />

absence of silica <strong>and</strong> phosphate were studied in the same pH range <strong>with</strong> FeCl 3 .<br />

Effect of Silica in the Absence of Phosphate <strong>and</strong> Vanadate<br />

Based on FeCl 3 coagulant tests using challenge water <strong>with</strong>out phosphate, it was found<br />

that silica interferes <strong>with</strong> the adsorption of As(V) <strong>and</strong> As(III) <strong>and</strong> the effect increases <strong>with</strong><br />

increasing pH. For example, the removals of As(V) in the presence of silica at 0.5 mg/L Fe(III)<br />

dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 92, 77, <strong>and</strong> 30%, respectively. In the absence of silica, the<br />

corresponding removals were 96, 93, <strong>and</strong> 82% at the same Fe(III) dose. Thus, silica significantly<br />

interfered <strong>with</strong> As(V) adsorption at pHs 7.5 <strong>and</strong> 8.5; while somewhat less silica interference was<br />

observed at pH 6.5. This can be observed in Figure 4.28 through the more favorable adsorption<br />

isotherms in the absence of silica at pHs of 7.5 <strong>and</strong> 8.5 as compared <strong>with</strong> the influence of silica<br />

on the adsorption isotherms at pH 6.5.<br />

Silica also interferes <strong>with</strong> the adsorption of As(III) onto Fe (III) hydroxide in the absence<br />

of phosphate <strong>and</strong> vanadate, <strong>and</strong> the effect of silica increases <strong>with</strong> increasing pH. In the absence<br />

of silica, the adsorption of As(III) was found to be pH dependent, <strong>and</strong> the removal efficiency of<br />

As(III) increased <strong>with</strong> increasing pH, which was not observed in the presence of silica where the<br />

removal was independent of pH. At 2 mg/L Fe(III) dosage, approximately 29% of As(III) was<br />

removed irrespective of pH in the presence of silica. In the absence of silica, removal efficiencies<br />

of 48, 64, <strong>and</strong> 78% were obtained at pHs 6.5, 7.5, <strong>and</strong> 8.5, respectively for the same dose. The<br />

As(III) adsorption isotherms shown in Figure 4.29 illustrate the significant effect of silica on<br />

As(III) adsorption onto Fe(III) hydroxide at all pHs <strong>and</strong> the effect increasing <strong>with</strong> increase in pH.<br />

q e (μg As(V)/mg Fe(III))<br />

180<br />

160<br />

140<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

pH 6.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH 6.5, SiO2=20mg/L,P=0ug/L,V=0ug/L<br />

pH 7.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH 8.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH 7.5, SiO2=20mg/L,P=0ug/L,V=0ug/L<br />

pH 8.5, SiO2=20mg/L,P=0ug/L,V=0ug/L<br />

0 5 10 15 20 25 30<br />

C e (μg As(V)/L)<br />

Figure 4.28 Effect of silica on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Fe(III) hydroxides in the absence of phosphate <strong>and</strong> vanadate<br />

51


q e (μg As(III)/mg Fe(III))<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

pH8.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH7.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH6.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH8.5, SiO2=20mg/L,P=0ug/L,V=0ug/L<br />

pH7.5, SiO2=20mg/L,P=0ug/L,V=0ug/L<br />

pH6.5, SiO2=20mg/L,P=0ug/L,V=0ug/L<br />

0 10 20 30 40 50<br />

C e (μg As(III)/L)<br />

Figure 4.29 Effect of silica on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ formed<br />

Fe(III) hydroxides in the absence of phosphate <strong>and</strong> vanadate<br />

Effect of Phosphate in the Absence of Silica <strong>and</strong> Vanadate<br />

In the absence of silica <strong>and</strong> vanadate, phosphate significantly interfered As(V) adsorption<br />

<strong>and</strong> had <strong>and</strong> a minor effect on the adsorption of As(III) during coagulation <strong>with</strong> FeCl 3 . The<br />

removals of As(V) in the presence of phosphate at 0.5 mg/L Fe(III) dosage at pHs 6.5, 7.5, <strong>and</strong><br />

8.5 were 80, 71, <strong>and</strong> 57%, respectively. In the absence of phosphate, the corresponding removals<br />

were 96, 93, <strong>and</strong> 82%, which shows that the decrease in removal efficiency was significant in the<br />

presence of phosphate. This can be observed in Figure 4.30 through the more favorable<br />

adsorption isotherms in the absence of phosphate at all three pHs tested: 6.5, 7.5, <strong>and</strong> 8.5.<br />

Phosphate had a minor effect on the adsorption of As(III) onto Fe(III) hydroxide in the<br />

absence of silica <strong>and</strong> vanadate at pHs of 7.5 <strong>and</strong> 8.5. In the presence <strong>and</strong> absence of phosphate,<br />

the adsorption of As(III) was found to be pH dependent, <strong>and</strong> the removal of As(III) increased<br />

<strong>with</strong> increasing pH. The removals of As(III) in the presence of phosphate at 2 mg/L Fe(III)<br />

dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 41, 53, <strong>and</strong> 71%, respectively. In the absence of phosphate,<br />

the removals at the same pHs were 48, 64, <strong>and</strong> 78%, respectively for the same dose, which<br />

shows a minor decrease in removal in the presence of phosphate. Thus, As(III) removal was pH<br />

dependent <strong>and</strong> was affected by the presence of phosphate. This can be observed in Figure 4.31<br />

through the more favorable adsorption isotherms in the absence of phosphate at pHs of 7.5 <strong>and</strong><br />

8.5 as compared <strong>with</strong> the lesser influence of phosphate on the adsorption isotherms at pH 6.5.<br />

52


q e (μg As(V)/mg Fe(III))<br />

200<br />

160<br />

120<br />

80<br />

40<br />

0<br />

pH6.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH7.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH8.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH6.5, SiO2=0mg/L,P=50ug/L,V=0ug/L<br />

pH7.5, SiO2=0mg/L,P=50ug/L,V=0ug/L<br />

pH8.5, SiO2=0mg/L,P=50ug/L,V=0ug/L<br />

0 5 10 15 20 25 30 35 40 45<br />

C e (μg As(V)/L)<br />

Figure 4.30 Effect of phosphate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Fe(III) hydroxides in the absence of silica <strong>and</strong> vanadate<br />

q e (μg As(III)/mg Fe(III))<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

pH8.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH8.5, SiO2=0mg/L,P=40ug/L,V=0ug/L<br />

pH7.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH7.5, SiO2=0mg/L,P=40ug/L,V=0ug/L<br />

pH6.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH6.5, SiO2=0mg/L,P=40ug/L,V=0ug/L<br />

0 5 10 15 20 25 30 35<br />

C e (μg As(III)/L)<br />

Figure 4.31 Effect of phosphate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Fe(III) hydroxides in the absence of silica <strong>and</strong> vanadate<br />

Effect of Vanadate in the Absence of Silica <strong>and</strong> Phosphate<br />

In the absence of silica <strong>and</strong> phosphate, vanadate exhibited a significant effect on the<br />

adsorption of As(V) <strong>and</strong> minor effect on the adsorption of As(III) during coagulation <strong>with</strong> FeCl 3 .<br />

This vanadate effect was not observed in the presence of silica <strong>and</strong> phosphate. The removals of<br />

As(V) in the presence of vanadate at 0.5 mg/L Fe(III) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 91,<br />

86, <strong>and</strong> 66%, respectively. In the absence of vanadate, the corresponding removals were 96, 93,<br />

53


<strong>and</strong> 82%. This can be observed in Figure 4.32 through the more favorable adsorption isotherms<br />

in the absence of vanadate at the three pHs tested: 6.5, 7.5, <strong>and</strong> 8.5.<br />

Vanadate had a minor effect on the adsorption of As(III) onto Fe (III) hydroxide in the<br />

absence of silica <strong>and</strong> phosphate at pHs of 7.5 <strong>and</strong> 8.5. In the presence <strong>and</strong> absence of vanadate,<br />

the adsorption of As(III) was found to be pH dependent, <strong>and</strong> the removal efficiency of As(III)<br />

increased <strong>with</strong> increasing pH. The removals of As(III) in the presence of vanadate at 2 mg/L<br />

Fe(III) dosage at pHs 6.5, 7.5, <strong>and</strong> 8.5 were 46, 58, <strong>and</strong> 72%, respectively. In the absence of<br />

vanadate, the corresponding removals were 48, 64, <strong>and</strong> 78%, which shows a minor decrease in<br />

removal efficiency in the presence of vanadate at pHs of 7.5 <strong>and</strong> 8.5. Thus, the As(III) removal<br />

efficiency was pH dependent <strong>and</strong> was affected by the presence of vanadate. This can be observed<br />

in Figure 4.33 through the more favorable adsorption isotherms in the absence of vanadate at<br />

pHs of 7.5 <strong>and</strong> 8.5 as compared <strong>with</strong> the lesser influence of vanadate on the adsorption isotherms<br />

at pH 6.5.<br />

q e (μg As(V)/mg Fe(III))<br />

180<br />

150<br />

120<br />

90<br />

60<br />

30<br />

pH6.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH7.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH6.5, SiO2=0mg/L,P=0ug/L,V=50ug/L<br />

pH7.5, SiO2=0mg/L,P=0ug/L,V=50ug/L<br />

pH8.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH8.5, SiO2=0mg/L,P=0ug/L,V=50ug/L<br />

0<br />

0 5 10 15 20 25 30<br />

C e (μg As(V)/L)<br />

Figure 4.32 Effect of vanadate on arsenic(V) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Fe(III) hydroxides in the absence of silica <strong>and</strong> phosphate<br />

54


qe (μg As(III)/mg Fe(III))<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

pH8.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH8.5, SiO2=0mg/L,P=0ug/L,V=50ug/L<br />

pH7.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH7.5, SiO2=0mg/L,P=0ug/L,V=50ug/L<br />

pH6.5, SiO2=0mg/L,P=0ug/L,V=0ug/L<br />

pH6.5, SiO2=0mg/L,P=0ug/L,V=50ug/L<br />

0 5 10 15 20 25 30 35<br />

C e (μg As(III)/L)<br />

Figure 4.33 Effect of vanadate on arsenic(III) adsorption at different pHs <strong>with</strong> in-situ<br />

formed Fe(III) hydroxides in the absence of silica <strong>and</strong> phosphate<br />

SUMMARIZING THE EFFECTS OF COMPETING IONS ON THE ARSENIC<br />

ADSORPTION CAPACITY IN NSFI CHALLENGE WATER<br />

Figure 4.34 compares the As(V) adsorption capacities of ferric (III) hydroxides for an<br />

equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5 in the presence <strong>and</strong> absence<br />

of silica, phosphate, <strong>and</strong> vanadate in NSFI challenge water. Based on the experiments, the<br />

presence of silica significantly reduced the adsorption capacity of As(V) onto Fe(OH) 3 .<br />

Phosphate also significantly lowered the adsorption of As(V) at pHs of 6.5 <strong>and</strong> 7.5 but had a<br />

lesser effect at pH 8.5. Vanadate did not significantly affect the adsorption of As(V) onto<br />

Fe(OH) 3 in the presence of silica <strong>and</strong> phosphate.<br />

Figure 4.35 compares the As(III) adsorption capacities of ferric (III) hydroxides for an<br />

equilibrium concentration of 10 μg/L As(III) in the pH range 6.5-8.5 in the presence <strong>and</strong> absence<br />

of silica, phosphate, <strong>and</strong> vanadate in NSFI challenge water. The presence of silica significantly<br />

reduced the adsorption capacity of As(III) <strong>and</strong> the effect increased <strong>with</strong> increasing pH. However,<br />

there was no significant competition of phosphate in the adsorption of As(III) onto Fe(OH) 3 in<br />

the presence of silica. Vanadate also did not significantly affect the adsorption of As(III) onto<br />

Fe(OH) 3 in the presence of silica <strong>and</strong> phosphate.<br />

Figure 4.36 compares the As(V) adsorption capacities of aluminum (III) hydroxides for<br />

an equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5 in the presence <strong>and</strong><br />

absence of silica, phosphate, <strong>and</strong> vanadate in NSFI challenge water. Based on the experiments,<br />

the presence of silica significantly reduced the adsorption capacity of As(V) at pH 6.5 <strong>and</strong> 7.5,<br />

while at pH 8.5 there was no significant effect of silica on As(V) adsorption. Phosphate did not<br />

affect As(V) adsorption onto Al(OH) 3 in the presence of silica at any pH. Vanadate also did not<br />

significantly affect the adsorption of As(V) onto Al(OH) 3 in the presence of silica <strong>and</strong> phosphate<br />

at pH 6.5-8.5.<br />

55


q e (mg As(V)/g Fe(III))<br />

160<br />

140<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

NSFI*<br />

NSFI <strong>with</strong>out silica<br />

NSFI <strong>with</strong>out phosphate<br />

NSFI <strong>with</strong> vanadate<br />

pH 6.5 pH 7.5 pH 8.5<br />

NSFI* 62.6 37.3 14.3<br />

NSFI <strong>with</strong>out silica 83.2 63.8 43.1<br />

NSFI <strong>with</strong>out phosphate 151 67.9 16.6<br />

NSFI <strong>with</strong> vanadate 61.3 32.8 11.9<br />

*NSFI-53 challenge water contains 20 mg/L SiO 2 , 40 μg/L PO 4 -P, <strong>and</strong> 0 μg/L VO 3 -V.<br />

Figure 4.34 Comparison of arsenic(V) adsorption capacities of in-situ formed Fe(III)<br />

hydroxides for an equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5<br />

qe (mg As(III)/g Fe(III))<br />

16<br />

14<br />

12<br />

10<br />

8<br />

6<br />

4<br />

2<br />

0<br />

NSFI*<br />

NSFI <strong>with</strong>out silica<br />

NSFI <strong>with</strong>out phosphate<br />

NSFI <strong>with</strong> vanadate<br />

pH 6.5 pH 7.5 pH 8.5<br />

NSFI* 3 2.96 3.12<br />

NSFI <strong>with</strong>out silica 4.95 7.23 13.4<br />

NSFI <strong>with</strong>out phosphate 2.2 2.52 2.76<br />

NSFI <strong>with</strong> vanadate 2.53 2.52 2.65<br />

*NSFI-53 challenge water contains 20 mg/L SiO 2 , 40 μg/L PO 4 -P, <strong>and</strong> 0 μg/L VO 3 -V.<br />

Figure 4.35 Comparison of arsenic(III) adsorption capacities of in-situ formed Fe(III)<br />

hydroxides for an equilibrium concentration of 10 μg/L As(III) in the pH range 6.5-8.5<br />

56


qe (mg As(V)/mg Al(III))<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

NSFI*<br />

NSFI <strong>with</strong>out silica<br />

NSFI <strong>with</strong>out phosphate<br />

NSFI <strong>with</strong> vanadate<br />

pH 6.5 pH 7.5 pH 8.5<br />

NSFI* 40.90 13.05 7.36<br />

NSFI <strong>with</strong>out silica 60.98 30.56 8.39<br />

NSFI <strong>with</strong>out phosphate 43.67 14.80 8.03<br />

NSFI <strong>with</strong> vanadate 43.26 14.48 7.36<br />

*NSFI-53 challenge water contains 20 mg/L SiO 2 , 40 μg/L PO 4 -P, <strong>and</strong> 0 μg/L VO 3 -V.<br />

Figure 4.36 Comparison of arsenic(V) adsorption capacities of in-situ formed Al(III)<br />

hydroxides for an equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5<br />

Figure 4.37 compares the As(V) adsorption capacities of zirconium (IV) hydroxides for<br />

an equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5 in the presence <strong>and</strong><br />

absence of silica, phosphate, <strong>and</strong> vanadate in NSFI challenge water. Silica significantly lowered<br />

the adsorption capacity of As(V) onto Zr(OH) 4 . Phosphate also significantly lowered the<br />

adsorption of As(V) at pHs of 6.5 <strong>and</strong> 7.5 but had a lesser effect at pH 8.5. Vanadate did not<br />

significantly affect the adsorption of As(V) onto Zr(OH) 4 in the presence of silica <strong>and</strong> phosphate.<br />

Figure 4.38 compares the As(III) adsorption capacities of zirconium (IV) hydroxides for<br />

an equilibrium concentration of 10 μg/L As(III) in the pH range 6.5-8.5 in the presence <strong>and</strong><br />

absence of silica, phosphate, <strong>and</strong> vanadate in NSFI challenge water. It can be observed that the<br />

As(III) capacities are extremely low compared <strong>with</strong> As(V) adsorption capacities for Zr(IV). The<br />

presence of silica significantly reduced the adsorption capacity of As(III) onto Zr(OH) 4 <strong>and</strong> the<br />

effect increased <strong>with</strong> increasing pH. Even though Figure 4.38 shows that adsorption capacity in<br />

the absence of phosphate is higher than in the presence of phosphate, it is not true as adsorption<br />

capacities were calculated <strong>with</strong> the Freundlich isotherm equation. But it can be seen from Figure<br />

4.14, the isotherms were similar regardless of the presence or absence of phosphate in the<br />

equilibrium concentration range of 35 to 50 μg/L, which was the experimental range. So there<br />

was no significant competition of phosphate in the adsorption of As(III) onto Zr(OH) 4 in the<br />

presence of silica. Vanadate also did not significantly affect the adsorption of As(III) onto<br />

Zr(OH) 4 in the presence of silica <strong>and</strong> phosphate.<br />

Figure 4.39 compares the As(V) adsorption capacities of titanium (IV) hydroxides for an<br />

equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5 in the presence <strong>and</strong> absence<br />

of silica, phosphate, <strong>and</strong> vanadate in NSFI challenge water. Silica significantly lowers the<br />

adsorption capacity of As(V) onto Ti(OH) 4 . Phosphate also significantly lowered the adsorption<br />

57


of As(V) at pHs of 6.5 <strong>and</strong> 7.5, but had a lesser effect at pH 8.5. Vanadate did not significantly<br />

affect the adsorption of As(V) onto Ti(OH) 4 in the presence of silica <strong>and</strong> phosphate.<br />

q e (mg As(V)/g Zr(IV))<br />

50<br />

45<br />

40<br />

35<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

NSFI*<br />

NSFI <strong>with</strong>out silica<br />

NSFI <strong>with</strong>out phosphate<br />

NSFI <strong>with</strong> vanadate<br />

pH 6.5 pH 7.5 pH 8.5<br />

NSFI* 28.1 9.89 5.02<br />

NSFI <strong>with</strong>out silica 46.8 23.9 11.8<br />

NSFI <strong>with</strong>out phosphate 41.7 13.6 5.28<br />

NSFI <strong>with</strong> vanadate 24.2 9.84 3.95<br />

*NSFI-53 challenge water contains 20 mg/L SiO 2 , 40 μg/L PO 4 -P, <strong>and</strong> 0 μg/L VO 3 -V.<br />

Figure 4.37 Comparison of arsenic(V) adsorption capacities of in-situ formed Zr(IV)<br />

hydroxides for an equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5<br />

q e (mg As(III)/g Zr(IV))<br />

3<br />

2<br />

1<br />

NSFI*<br />

NSFI <strong>with</strong>out silica<br />

NSFI <strong>with</strong>out phosphate<br />

NSFI <strong>with</strong> vanadate<br />

0<br />

pH 6.5 pH 7.5 pH 8.5<br />

NSFI* 0.29 0.76 0.51<br />

NSFI <strong>with</strong>out silica 0.72 1.11 2.75<br />

NSFI <strong>with</strong>out phosphate 1.12 1.07 1.81<br />

NSFI <strong>with</strong> vanadate 0.84 0.85 0.71<br />

*NSFI-53 challenge water contains 20 mg/L SiO 2 , 40 μg/L PO 4 -P, <strong>and</strong> 0 μg/L VO 3 -V.<br />

Figure 4.38 Comparison of arsenic(III) adsorption capacities of in-situ formed Zr(IV)<br />

hydroxides for an equilibrium concentration of 10 μg/L As(III) in the pH range 6.5-8.5<br />

58


q e (mg As(V)/g Ti(IV))<br />

45<br />

40<br />

35<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

NSFI*<br />

NSFI <strong>with</strong>out silica<br />

NSFI <strong>with</strong>out phosphate<br />

NSFI <strong>with</strong> vanadate<br />

pH 6.5 pH 7.5 pH 8.5<br />

NSFI* 25.1 13.9 10.7<br />

NSFI <strong>with</strong>out silica 40.9 24.1 15.9<br />

NSFI <strong>with</strong>out phosphate 30.1 19.2 13.3<br />

NSFI <strong>with</strong> vanadate 22 13.9 9.75<br />

*NSFI-53 challenge water contains 20 mg/L SiO 2 , 40 μg/L PO 4 -P, <strong>and</strong> 0 μg/L VO 3 -V.<br />

Figure 4.39 Comparison of arsenic(V) adsorption capacities of in-situ formed Ti(IV)<br />

hydroxides for an equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5<br />

Figure 4.40 compares the As(III) adsorption capacities of titanium (IV) hydroxides for an<br />

equilibrium concentration of 10 μg/L As(III) in the pH range 6.5-8.5 in the presence <strong>and</strong> absence<br />

of silica, phosphate, <strong>and</strong> vanadate in NSFI challenge water. The presence of silica dramatically<br />

reduced the adsorption capacity of As(III) onto Ti(OH) 4 at all three pH values. However there<br />

was no significant competition of phosphate in the adsorption of As(III) onto Ti(OH) 4 in the<br />

presence of silica. Vanadate also did not significantly affect the adsorption of As(III) onto<br />

Ti(OH) 4 in the presence of silica <strong>and</strong> phosphate.<br />

Figure 4.41 compares the As(V) adsorption capacities of titanium (III) hydroxides for an<br />

equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5 in the presence <strong>and</strong> absence<br />

of silica, phosphate, <strong>and</strong> vanadate in NSFI challenge water. Silica <strong>and</strong> phosphate significantly<br />

lower the adsorption capacity of As(V) onto Ti(OH) 3 at pH 6.5 <strong>and</strong> 7.5, but have a lesser<br />

interference at pH 8.5. Vanadate did not significantly affect the adsorption of As(V) onto<br />

Ti(OH) 3 in the presence of silica <strong>and</strong> phosphate.<br />

Figure 4.42 compares the As(III) adsorption capacities of titanium (III) hydroxides for an<br />

equilibrium concentration of 10 μg/L As(III) in the pH 6.5-8.5 range in the presence <strong>and</strong> absence<br />

of silica, phosphate, <strong>and</strong> vanadate in NSFI challenge water. The presence of silica significantly<br />

reduced the adsorption capacity of As(III) onto Ti(OH) 3 at all pHs. Phosphate interference was<br />

less that that of silica in the adsorption of As(III) onto Ti(OH) 3 . Phosphate <strong>and</strong> Vanadate did not<br />

significantly affect the adsorption of As(III) onto Ti(OH) 3 in the NSFI challenge water.<br />

59


q e (mg As(III)/g Ti(IV))<br />

12<br />

10<br />

8<br />

6<br />

4<br />

2<br />

NSFI*<br />

NSFI <strong>with</strong>out silica<br />

NSFI <strong>with</strong>out phosphate<br />

NSFI <strong>with</strong> vanadate<br />

0<br />

pH 6.5 pH 7.5 pH 8.5<br />

NSFI* 2.74 2.86 2.74<br />

NSFI <strong>with</strong>out silica 9.96 9.25 8.28<br />

NSFI <strong>with</strong>out phosphate 2.97 2.95 2.65<br />

NSFI <strong>with</strong> vanadate 3.04 2.77 2.4<br />

*NSFI-53 challenge water contains 20 mg/L SiO 2 , 40 μg/L PO 4 -P, <strong>and</strong> 0 μg/L VO 3 -V.<br />

Figure 4.40 Comparison of arsenic(III) adsorption capacities of in-situ formed Ti(IV)<br />

hydroxides for an equilibrium concentration of 10 μg/L As(III) in the pH range 6.5-8.5<br />

qe (mg As(V)/g Ti(III))<br />

40<br />

35<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

NSFI*<br />

NSFI <strong>with</strong>out silica<br />

NSFI <strong>with</strong>out phosphate<br />

NSFI <strong>with</strong> vanadate<br />

pH 6.5 pH 7.5 pH 8.5<br />

NSFI* 20 16.2 14.2<br />

NSFI <strong>with</strong>out silica 37 25.4 19.7<br />

NSFI <strong>with</strong>out phosphate 24.2 20.9 15.1<br />

NSFI <strong>with</strong> vanadate 18.3 14.7 12.9<br />

*NSFI-53 challenge water contains 20 mg/L SiO 2 , 40 μg/L PO 4 -P, <strong>and</strong> 0 μg/L VO 3 -V.<br />

Figure 4.41 Comparison of arsenic(V) adsorption capacities of in-situ formed Ti(III)<br />

hydroxides for an equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5<br />

60


qe (mg As(III)/g Ti(III))<br />

20<br />

15<br />

10<br />

5<br />

0<br />

NSFI*<br />

NSFI <strong>with</strong>out silica<br />

NSFI <strong>with</strong>out phosphate<br />

NSFI <strong>with</strong> vanadate<br />

pH 6.5 pH 7.5 pH 8.5<br />

NSFI* 6.13 4 2.36<br />

NSFI <strong>with</strong>out silica 17.5 10.7 7.93<br />

NSFI <strong>with</strong>out phosphate 6.45 4.2 2.4<br />

NSFI <strong>with</strong> vanadate 5.63 4.06 2.76<br />

*NSFI-53 challenge water contains 20 mg/L SiO 2 , 40 μg/L PO 4 -P, <strong>and</strong> 0 μg/L VO 3 -V.<br />

Figure 4.42 Comparison of arsenic(III) adsorption capacities of in-situ formed Ti(III)<br />

hydroxides for an equilibrium concentration of 10 μg/L As(III) in the pH range 6.5-8.5<br />

INDIVIDUAL EFFECTS OF COMPETING IONS ON THE ARSENIC ADSORPTION<br />

CAPACITY OF FERRIC (III) HYDROXIDE<br />

Figure 4.43 compares the individual effects of silica, phosphate <strong>and</strong> vanadate on the<br />

As(V) adsorption capacities of ferric (III) hydroxides for an equilibrium concentration of 10<br />

μg/L As(V) in the pH range 6.5-8.5. In the absence of phosphate <strong>and</strong> vanadate, silica<br />

significantly reduced the adsorption capacity of As(V) onto Fe(OH) 3 <strong>and</strong> the effect increased<br />

significantly <strong>with</strong> increasing pH. Phosphate significantly lowered the adsorption of As(V) at all<br />

pHs in the absence of silica <strong>and</strong> vanadate. Vanadate also significantly affected the adsorption of<br />

As(V) in the absence of silica <strong>and</strong> phosphate at all pHs during coagulation <strong>with</strong> FeCl 3 .<br />

Figure 4.44 compares the individual effects of silica, phosphate <strong>and</strong> vanadate on the<br />

As(III) adsorption capacities of ferric (III) hydroxides for an equilibrium concentration of 10<br />

μg/L As(III) in the pH range 6.5-8.5. In the absence of phosphate <strong>and</strong> vanadate, silica<br />

significantly reduced the adsorption capacity of As(III) onto Fe(OH) 3 <strong>and</strong> the effect increased<br />

<strong>with</strong> increasing pH. Phosphate <strong>and</strong> vanadate had a minor effect on As(III) adsorption onto<br />

Fe(OH) 3 , <strong>and</strong> the effect increased <strong>with</strong> increasing pH..<br />

61


q e (mg As(V)/g Fe(III))<br />

200<br />

180<br />

160<br />

140<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

NSFI <strong>with</strong> no competing ion*<br />

SiO2=20mg/L<br />

PO4-P=40ug/L<br />

VO3-V=50ug/L<br />

pH 6.5 pH 7.5 pH 8.5<br />

NSFI <strong>with</strong> no competing ion* 178 152 88.5<br />

SiO2=20mg/L 151 67.9 16.6<br />

PO4-P=40ug/L 83.2 63.8 43<br />

VO3-V=50ug/L 112 102 53.8<br />

*NSFI <strong>with</strong> no competing ion contains 0 mg/L SiO 2 , 0 μg/L PO 4 -P, <strong>and</strong> 0 μg/L VO 3 -V.<br />

Figure 4.43 Comparison of arsenic adsorption capacities of in-situ-formed Fe(III)<br />

hydroxides for an equilibrium concentration of 10 μg/L As(V) in the pH range 6.5-8.5<br />

q e (mg As(III)/g Fe(III))<br />

20<br />

18<br />

16<br />

14<br />

12<br />

10<br />

8<br />

6<br />

4<br />

2<br />

0<br />

NSFI <strong>with</strong> no competing ion*<br />

Si=20mg/L<br />

PO4-P=40ug/L<br />

VO3-V=50ug/L<br />

pH 6.5 pH 7.5 pH 8.5<br />

NSFI <strong>with</strong> no competing ion* 5.03 9.22 17.3<br />

Si=20mg/L 2.2 2.52 2.76<br />

PO4-P=40ug/L 4.95 7.23 13.4<br />

VO3-V=50ug/L 4.1 7.9 14.4<br />

*NSFI <strong>with</strong> no competing ion contains 0 mg/L SiO 2 , 0 μg/L PO 4 -P, <strong>and</strong> 0 μg/L VO 3 -V.<br />

Figure 4.44 Comparison of arsenic adsorption capacities of in-situ formed Fe(III)<br />

hydroxides for an equilibrium concentration of 10 μg/L As(III) in the pH range 6.5-8.5<br />

62


CONCLUSIONS<br />

In the NSFI challenge water <strong>with</strong> phosphate, silica significantly reduced the adsorption of<br />

arsenic by competing for adsorption sites. So the As(V) <strong>and</strong> As(III) removal efficiencies in the<br />

absence of silica were higher than in the presence of silica <strong>with</strong> the four coagulants tested: FeCl 3 ,<br />

TiCl 4 , TiCl 3 , <strong>and</strong> ZrCl 4 . With alum as coagulant, silica did affect As(V) adsorption at pH 6.5 <strong>and</strong><br />

7.5, while at pH 8.5, it had no effect. The effect of silica at 20 mg/L increased <strong>with</strong> increasing pH<br />

in the case of As(III) removal <strong>with</strong> all coagulants (FeCl 3 , TiCl 4 , <strong>and</strong> ZrCl 4 ) except TiCl 3 . In the<br />

NSFI challenge water <strong>with</strong> silica, phosphate at 40 μg P/L was found to reduce the adsorption of<br />

As(V) significantly at pH 6.5 <strong>and</strong> 7.5, whereas it had a lesser effect at pH 8.5 <strong>with</strong> all four<br />

coagulants tested: FeCl 3 , TiCl 4 , TiCl 3 , <strong>and</strong> ZrCl 4 . With alum as coagulant, phosphate did not<br />

have a significant effect on As(V) adsorption in the 6.5-8.5 pH range. In contrast to its effect on<br />

As(V) adsorption, the presence of phosphate did not have a significant effect on the adsorption of<br />

As(III) for all four of the coagulants tested (FeCl 3 , TiCl 4 , TiCl 3 , <strong>and</strong> ZrCl 4 ). In the NSFI<br />

challenge water <strong>with</strong> silica <strong>and</strong> phosphate, vanadate at 50 μg V/L did not significantly affect the<br />

adsorption of As(V) or As(III) <strong>with</strong> all the coagulants tested including FeCl 3 , alum, TiCl 4 , TiCl 3 ,<br />

<strong>and</strong> ZrCl 4 .<br />

In the absence of other competing ions <strong>with</strong> ferric chloride as coagulant, it was found that<br />

silica, phosphate, <strong>and</strong> vanadate exhibited significant competitive effects on the adsorption of<br />

As(V) <strong>and</strong> As(III). In the absence of phosphate <strong>and</strong> vanadate, silica exhibited a significant effect<br />

on the adsorption of As(V) <strong>and</strong> As(III) <strong>and</strong> the effect increased <strong>with</strong> increasing pH. In the<br />

absence of silica <strong>and</strong> vanadate, phosphate exhibited a significant effect on the adsorption of<br />

As(V) at all pHs <strong>and</strong> a minor effect on the adsorption of As(III) at pH 7.5 <strong>and</strong> 8.5 during<br />

coagulation <strong>with</strong> FeCl 3 . Similarly, in the absence of silica <strong>and</strong> phosphate, vanadate exhibited a<br />

significant effect on the adsorption of As(V) <strong>and</strong> a minor effect on the adsorption of As(III) at<br />

pH 7.5 <strong>and</strong> 8.5 during coagulation <strong>with</strong> FeCl 3 .<br />

So, as expected, the competing ions under study did affect the adsorption of arsenic in<br />

most of the cases. Silica did compete <strong>with</strong> As(V) <strong>and</strong> As(III) for adsorption sites <strong>and</strong> affected the<br />

adsorption capacity regardless of the presence of competing ions. Phosphate did compete <strong>with</strong><br />

As(V) <strong>and</strong> As(III) adsorption in the absence of competing ions. However, the phosphate effect<br />

was suppressed in the presence of silica: the phosphate interference on As(V) adsorption<br />

decreased <strong>with</strong> increasing pH <strong>and</strong> had no effect on As(III) adsorption. The effect of vanadate on<br />

arsenic adsorption was observed in the absence of competing ions while no effect was observed<br />

in the presence of competing ions which suggests that the presence of silica <strong>and</strong> phosphate<br />

suppressed the effect of vanadate at all pHs in the case of As(V) <strong>and</strong> As(III) removal.<br />

63


CHAPTER 5<br />

TOXICITY CHARACTERISTIC STUDIES<br />

INTRODUCTION<br />

The Toxicity Characteristic (TC) is used to determine whether a solid waste is classified<br />

as a hazardous waste due to its toxicity. The TC of waste is established by determining the levels<br />

of the contaminant in the TC extract of the waste. Under USEPA regulations, the toxicity<br />

characteristic is assessed by the Toxicity Characteristic Leaching Procedure (TCLP), which<br />

measures the leachability of toxic contaminants from the wastes utilizing a buffered acetic acid<br />

solution as an extraction fluid. In addition to using the TCLP to determine hazardous waste<br />

status, California employs a method known as the Waste Extraction Test (WET), which employs<br />

buffered citric acid as a solvent <strong>and</strong> tends to leach more metals than the TCLP test. Both the tests<br />

have a common preliminary-evaluation-of-solids procedure, which is followed by a leaching<br />

procedure which differentiates them if needed. The TCLP <strong>and</strong> WET test regulatory limit for<br />

arsenic is 5.0 mg/L.<br />

PRELIMINARY TOXICITY CHARACTERISTIC EVALUATION<br />

A preliminary TC evaluation is performed on a waste <strong>and</strong> includes: (1) determination of<br />

the percent solids; (2) determination of whether the waste contains insignificant solids <strong>and</strong> is,<br />

therefore, its own extract after filtration; (3) determination of whether the solid portion of the<br />

waste requires particle size reduction; <strong>and</strong> (4) determination of which of two extraction fluids are<br />

to be used for the extraction of the waste.<br />

Figure 5.1 provides a flow chart which delineates preliminary determination of percent<br />

solids. The first step is to take the waste <strong>and</strong> pass it through a 0.6 to 0.8 μm filter <strong>and</strong> determine<br />

the percent dry solids, <strong>and</strong>, depending on the dry solids, determine if further studies are needed<br />

as follows:<br />

i) For liquid wastes (i.e., those containing less than 0.5% dry solid material), the waste,<br />

after filtration through a 0.6 to 0.8 μm glass fiber filter is defined as the TCLP extract.<br />

ii) For wastes containing greater than or equal to 0.5% solids, the liquid, if any, is<br />

separated from the solid phase <strong>and</strong> stored for later analysis; the particle size of the<br />

solid phase is reduced, if necessary. The solid phase is extracted <strong>with</strong> an amount of<br />

extraction fluid equal to 20 times the weight of the solid phase. The composition of<br />

the extraction fluid employed is a function of the alkalinity of the solid phase of the<br />

waste. Following extraction, the liquid extract is separated from the solid phase by<br />

filtration through a 0.6 to 0.8 μm glass fiber filter.<br />

iii) If compatible (i.e., multiple phases will not form on combination), the initial liquid<br />

phase of the waste is added to the liquid extract, <strong>and</strong> these are analyzed together. If<br />

incompatible, the liquids are analyzed separately <strong>and</strong> the results are mathematically<br />

combined to yield a volume-weighted average concentration.<br />

65


Sample of<br />

waste<br />

Solids are < 0.5%.<br />

Discard solids.<br />

Filtrate = TCLP<br />

extract<br />

< 0.5% Determine % ≥ 0.5%<br />

dry solids in<br />

the waste<br />

Leachate<br />

studies<br />

needed<br />

100%<br />

Examine solids.<br />

Leachate studies<br />

needed<br />

Figure 5.1 Preliminary determination of percent solids<br />

Experimental Study Procedure<br />

The preliminary-determination-of-percent-solids studies were conducted at pH 7.5 <strong>with</strong><br />

an initial concentration of 500 μg/L As(V) in the case of ferric chloride <strong>and</strong> 250 μg/L As(V) <strong>with</strong><br />

all other coagulants (Alum, TiCl 3 , TiCl 4 , <strong>and</strong> ZrCl 4 ). The higher (500 or 250 μg/L As(V)) initial<br />

arsenic concentrations were used because the small amount of coagulant needed in the case of<br />

initial concentration of 50 μg/L As(V) would not yield enough sludge for further testing. The<br />

coagulant doses required to achieve equilibrium arsenic concentrations ≤10 μg/L in the challenge<br />

water were calculated based on the isotherms obtained previously for the four coagulants tested:<br />

FeCl 3 , Alum, ZrCl 4 , TiCl 4 , <strong>and</strong> TiCl 3 . The jar tests were then conducted as before <strong>with</strong> the<br />

calculated amount of metal being dosed. After rapid mixing <strong>and</strong> flocculation, the waste was<br />

allowed to settle in an Imhoff cone for 3 hours. The volume of settled sludge was noted, <strong>and</strong> the<br />

initial weight of the waste was calculated assuming a density of 1.0 g/mL. The waste was<br />

vacuum filtered through a 0.7-μm glass fiber filter. The residue on the filter was defined as the<br />

solid phase of the waste, <strong>and</strong> the liquid phase was defined as the filtrate. The filter <strong>with</strong> the solid<br />

phase was then removed from the filtration apparatus <strong>and</strong> dried at 110 ° C until two successive<br />

weighings yielded the same value <strong>with</strong>in ± 1%. The dry weight of the solid was then calculated,<br />

which was used to calculate the percent dry solids using Equation 5.1.<br />

Weight of dry waste<br />

Percent dry solids = *100<br />

(5.1)<br />

Initial weight of waste<br />

66


According to the Toxicity Characteristic Leaching procedure (TCLP) test method 1311<br />

(USEPA SW 846) <strong>and</strong> California Waste Extraction Test (WET) procedure (California Code of<br />

Regulations, Title 22, Division 4.5, Chapter 11, Article 5), liquid wastes containing less than<br />

0.5% dry solids do not require extraction. The liquid waste, after filtration, is defined as the<br />

extract. The filtered extract was analyzed <strong>and</strong> the resulting concentration compared directly to<br />

the appropriate regulatory concentration.<br />

TOXICITY CHARACTERISTIC STUDY RESULTS<br />

As explained earlier (TCLP <strong>and</strong> WET procedures), the percent dry solids in the 3-hr<br />

settled sludges were calculated for each of the coagulants tested <strong>and</strong> the values are given in Table<br />

5.1.<br />

It can be observed that the percent dry solids results were less than 0.5% for all<br />

coagulants tested. So, the liquid wastes (containing less than 0.5% dry solids) did not require<br />

extraction. The liquid waste, after filtration, is defined as the extract or filtrate. It was expected<br />

that the arsenic concentration in the filtrate would be less than or equal to 10 μg/L, since it is in<br />

equilibrium <strong>with</strong> the effluent concentration. As expected the measured arsenic concentration as<br />

reported in Table 5.2 were ≤ 10 μg/L, which easily passed the TCLP <strong>and</strong> WET test regulatory<br />

limit of 5 mg As/L (5,000 μg As/L).<br />

EXPERIMENTAL EVALUATION OF THE ADSORPTION OF HIGH<br />

CONCENTRATIONS OF ARSENIC ONTO METAL HYDROXIDES<br />

The initial arsenic concentrations in these extraction studies were 500 or 250 μg/L As(V),<br />

which were much higher than the initial concentration of arsenic (50 μg/L) used in previous jar<br />

tests. Thus, it was necessary to determine if the higher initial arsenic concentrations would<br />

influence the equilibrium adsorption capacity of arsenic on the metal hydroxides. The highinitial-concentration<br />

arsenic loadings were determined by dissolving the metal hydroxide<br />

precipitates <strong>and</strong> measuring the arsenic released. These arsenic loadings were compared <strong>with</strong> the<br />

calculated loadings based on previously determined arsenic isotherms beginning <strong>with</strong> challenge<br />

water containing 50 μg /L As(V). Specifically, the high-initial-arsenic-concentration loading of<br />

As(V) onto precipitated solids was determined by dissolving the dry solids obtained in the<br />

leachate studies into 1:1 nitric acid. The acid solution was then analyzed for arsenic<br />

concentration, which was then converted to an As(V) loading onto the known amount of<br />

precipitated metal hydroxides. The comparison of the adsorption capacities based on the<br />

calculated <strong>and</strong> experimental work is given in Table 5.3, where it can be seen that the loadings are<br />

almost the same. Therefore, it was concluded that the As(V) adsorption capacity was the same,<br />

regardless of the initial concentration of arsenic when the final As(V) concentration was ≤ 10<br />

μg/L.<br />

Table 5.1<br />

Percent dry solids in the liquid wastes after 3-hr settling<br />

Coagulant FeCl 3 Alum ZrCl 4 TiCl 4 TiCl 3<br />

% dry solids 0.27 ± 0.02% 0.31 ± 0.04% 0.4 ± 0.05% 0.22 ± 0.02% 0.21 ± 0.02%<br />

67


Table 5.2<br />

<strong>Arsenic</strong> concentrations in the TCLP- or WET-defined extract (liquid waste after filtration)<br />

Coagulant FeCl 3 Alum ZrCl 4 TiCl 4 TiCl 3<br />

<strong>Arsenic</strong><br />

concentration in<br />

extract (μg/L)<br />

3.39 ± 0.24 13.45 ± 0.68 3.02 ± 1.23 4.09 ± 0.5 11.2 ± 0.89<br />

Table 5.3<br />

Comparison of expected <strong>and</strong> measured arsenic adsorption capacities of metal hydroxides<br />

for all coagulants at 10 μg /L As(V) equilibrium concentration.<br />

Coagulant<br />

Theoretical adsorption<br />

capacity (μg As/mg<br />

metal)<br />

Experimental<br />

adsorption capacity<br />

based on filtrate (μg<br />

As/mg metal)<br />

Experimental<br />

adsorption capacity<br />

based on arsenic<br />

adsorbed on solid (μg<br />

As/mg metal)<br />

Initial<br />

concentration<br />

of As(V)<br />

FeCl 3 Alum ZrCl 4 TiCl 4 TiCl 3<br />

50 μg/L 37.3 13 10 14 16.2<br />

250 or 500<br />

μg/L<br />

250 or 500<br />

μg/L<br />

37.3 ±<br />

0.02<br />

35.6 ±<br />

0.65<br />

10.81 ±<br />

0.03<br />

12.51 ±<br />

1.2<br />

10.2 ±<br />

0.05<br />

8.79 ±<br />

0.7<br />

14.1 ±<br />

0.03<br />

13.3 ±<br />

0.52<br />

16.4 ±<br />

0.06<br />

15.1 ±<br />

0.84<br />

CONCLUSIONS<br />

Following coagulation <strong>and</strong> 3-hr settling, <strong>with</strong> all five coagulants tested (FeCl 3 , Alum,<br />

TiCl 3 , TiCl 4 , <strong>and</strong> ZrCl 4 ), the liquid wastes obtained were found to contain less than 0.5% dry<br />

solids, <strong>and</strong> therefore did not require extraction. The liquid filtrate, which is the extract in such<br />

cases, easily passed the TCLP <strong>and</strong> WET test regulatory limit of 5.0 mg/L. It was also found that<br />

the adsorption capacity of As(V) remained the same regardless of initial arsenic concentration<br />

<strong>with</strong> all four of the coagulants tested. Thus the adsorption capacity of arsenic at a final arsenic<br />

concentration of 10 μg/L was found to be independent of initial arsenic concentration in the<br />

range of 50-500 μg/L As(V).<br />

68


CHAPTER 6<br />

ARSENIC REMOVAL WITH FeCl 3 COAGULATION: COMPARISON OF<br />

EXPERIMENTAL RESULTS WITH MINEQL+ 4.50 CHEMICAL<br />

EQUILIBRIUM MODELING PROGRAM PREDICTIONS<br />

INTRODUCTION<br />

The scope of work for this project included a requirement to model the effects of arsenic<br />

concentration <strong>and</strong> background ion concentrations on the adsorption of As(III) <strong>and</strong> As(V) onto the<br />

coagulants tested using commercially available software. The objective of the modeling research<br />

described in this chapter was to determine if the adsorption of arsenic (As(V) <strong>and</strong> As(III)) onto<br />

Fe(OH) 3 could be accurately modeled using the Mineql+ Chemical Equilibrium Modeling<br />

Program. To determine the accuracy of the Mineql+ predictions, modeling results were<br />

compared <strong>with</strong> experimental results. Fe(III) was the only coagulant modeled because the<br />

unmodified Mineql+ program is not capable of modeling adsorption onto zirconium or titanium<br />

oxyhydroxides.<br />

MINEQL+ is a widely used chemical equilibrium modeling program capable of<br />

calculating aqueous-, solid-, <strong>and</strong> gas-phase concentrations of chemical species derived from<br />

input components. An extensive thermodynamic database is included in the model. MINEQL+ is<br />

a powerful <strong>and</strong> easy-to-use chemical equilibrium modeling system that can be used to perform<br />

calculations in low temperature (0-50 o C) <strong>and</strong> low-to-moderate ionic strength (


Table 6.1<br />

Composition of NSFI-53 challenge water<br />

Ion Concentration (mg/L) MW Concentration (M)<br />

Ca 2+ 40.1 40.1 1.000E-3<br />

Mg 2+ 12.6 24.3 5.185E-4<br />

Na + 88.87 23.0 3.864E-3<br />

-<br />

HCO 3 183 61.0 3.000E-3<br />

2-<br />

SO 4 50 96.1 5.203E-4<br />

Cl - 71 35.5 2.000E-3<br />

NO 3 -N 2.0 14.0 1.429E-4<br />

F - 1.0 19.0 5.263E-5<br />

PO 4 -P 0.04 31.0 1.290E-6<br />

SiO 3 -SiO 2 20 60.1 3.328E-4<br />

As(III)/(V) 0.05 74.9 6.676E-7<br />

Table 6.2<br />

Mineql+ titration parameters<br />

Parameter Start End Points<br />

Total conc. of Fe 3+ 8.952E-6 M 8.952E-5 M 10<br />

Total conc. of Cl - 2.852E-3 M 3.094E-3 M 10<br />

Total conc. of Na + 3.891E-3 M 4.133E-3 M 10<br />

concentration of Cl - in the NSFI challenge water plus the amount of Cl - added <strong>with</strong> Fe 3+ , which<br />

would be three times the Fe 3+ concentration. In the “pH wizard”, the calculation type was set to<br />

“pH is supplied by the user” <strong>and</strong> the “pH value was set at pH 7.5”. Although the coagulation jar<br />

tests were done in the atmosphere, the coagulated solutions were not actually in equilibrium <strong>with</strong><br />

the atmosphere. Therefore, the calculation type was set to “Closed to the atmosphere” <strong>and</strong> the<br />

“Total CO 3 ” concentration was then set in the “CO 2 wizard”. The “Ionic Strength Corrections”<br />

was set to “Calculated” <strong>and</strong> “2-layer FeOH” was selected in the “Adsorption model section” in<br />

the “Run Time Manager”. The calculation was then run.<br />

From <strong>with</strong>in the “Output Manager”, “Report on Ion balance sample QA/QC” was opened<br />

from the “Special Reports Output Type”. The ion balance for this system showed a value of<br />

8.272E-4 which corresponds to the amount of acid (HCl) anion that was not accounted for in the<br />

system at fixed pH 7.5. So, going back to calculation wizard, the required concentration of Cl -<br />

was added to the initial concentration. The pH calculation type was changed to “pH calculated by<br />

Mineql+” <strong>and</strong> the “Base pH calculation on Electroneutrality” was selected in the “pH wizard”.<br />

After the initial conditions were set, the addition of FeCl 3 was started. The “Multirun Option”<br />

was selected <strong>and</strong> the type of calculation was set to “Titration” <strong>and</strong> the variables Fe 3+ , Cl - <strong>and</strong> Na +<br />

were selected. It was necessary that a 3:1 ratio of Cl - to Fe 3+ be maintained, <strong>and</strong> also that that we<br />

would have to add an equivalent amount of Na + for every Cl - added to simulate NaOH<br />

neutralization of HCl resulting from the hydrolysis Fe 3+ to produce Fe(OH) 3 (s). A summary of<br />

the titration parameters that were selected are shown in Table 6.2.<br />

70


The problem was then rerun, <strong>and</strong> from the output, the pH of the system was checked for<br />

each addition of Fe 3+ , <strong>and</strong> was found to be around “pH 7.5 ± 0.01” in the “Alkalinity Summary<br />

Special Reports”. The ion balance was also checked from the “Report on Ion balance sample<br />

QA/QC” <strong>and</strong> was found to “pass” the ion balance. From the “Output Manager”, the “Component<br />

Groups” was selected <strong>and</strong> the “Data Object” was set to “AsO 4 3- ”. The “Obs × Variables” was set<br />

to “Total adsorbed AsO 4 3- ”, which gave the concentration of AsO 4 3- that was adsorbed for each<br />

addition of Fe 3+ . The values are tabulated <strong>with</strong> the corresponding concentration of Fe 3+ that was<br />

added in Table 6.3. The adsorption capacity then was calculated as follows.<br />

Total adsorbed AsO<br />

3 −<br />

Adsorption capacity ( μ g As / mg Fe)<br />

=<br />

4<br />

(6.1)<br />

Concentrat ion of Fe<br />

3 +<br />

Table 6.3 gives the As(V) adsorption capacity of Fe(OH) 3 (s) calculated at pH 7.5 <strong>with</strong><br />

NSFI-53 challenge water from the concentration of arsenic adsorbed for every addition of Fe 3+ in<br />

the range of 0.5-5.0 mg/L. The As(V) adsorption isotherm, i.e., the As(V) adsorption capacity<br />

(μg As(V)/mg Fe(III)) vs. the final concentration of arsenic in challenge water was plotted <strong>and</strong> is<br />

shown in Figure 6.1. A representative As(V) adsorption capacity was calculated for 10 μg/L<br />

As(V) in the challenge water. The adsorption capacity thus calculated was then compared <strong>with</strong><br />

the experimental result.<br />

Titration<br />

#<br />

Table 6.3<br />

As(V) adsorption capacity of Fe(OH) 3 based on Mineql+ program<br />

<strong>Arsenic</strong> <strong>Arsenic</strong> Concentration<br />

Concentration Concentration<br />

of Fe 3+ (M) of Fe 3+ adsorbed adsorbed of <strong>Arsenic</strong> in<br />

(mg/L)<br />

(M) (μg/L) water, C e (μg/L)<br />

Adsorption<br />

capacity, Q e (μg<br />

As(V)/mg Fe)<br />

Initial 0 0 0 0 50.00 0<br />

1 8.95E-6 0.5 2.63E-7 19.7 30.3 39.41<br />

2 1.79E-5 1 4.74E-7 35.5 14.5 35.52<br />

3 2.69E-5 1.5 6.00E-7 44.94 5.06 29.97<br />

4 3.58E-5 2 6.45E-7 48.31 1.69 24.16<br />

5 4.48E-5 2.5 6.58E-7 49.28 0.72 19.72<br />

6 5.37E-5 3 6.62E-7 49.58 0.42 16.53<br />

7 6.27E-5 3.5 6.64E-7 49.73 0.27 14.21<br />

8 7.16E-5 4 6.65E-7 49.81 0.19 12.46<br />

9 8.06E-5 4.5 6.66E-7 49.88 0.12 11.09<br />

10 8.95E-5 5 6.66E-7 49.88 0.12 9.98<br />

71


q e (μg As(V)/mg of Fe(III))<br />

50<br />

45<br />

40<br />

35<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

NSFI-53 water<br />

Mineql+ Modeling Results<br />

0 5 10 15 20 25 30 35<br />

C e (μg As(V)/L)<br />

Figure 6.1 Modeling of As(V) adsorption isotherm onto Fe(OH) 3 (s) at pH 7.5<br />

Similarly the Mineql+ modeling program was used to calculate the adsorption capacity of<br />

As(V) or As(III) for the following compositions of challenge water at pH 6.5, 7.5, <strong>and</strong> 8.5.<br />

• NSFI-53 challenge water<br />

• NSFI-53 <strong>with</strong>out phosphate (i.e., P = 0 μg/L, Si = 20 mg/L)<br />

• NSFI-53 <strong>with</strong> vanadate (i.e., VO 3 -V = 50 μg/L, Si = 20 mg/L, PO 4 -P = 40 μg/L)<br />

• NSFI-53 <strong>with</strong> vanadate <strong>and</strong> <strong>with</strong>out silica or phosphate (i.e., V = 50 μg/L, Si = 0 mg/L<br />

<strong>and</strong> P = 0 μg/L )<br />

• NSFI-53 <strong>with</strong>out silica (i.e., Si = 0 mg/L, PO 4 -P = 40 μg/L)<br />

• NSFI-53 <strong>with</strong>out silica or phosphate (i.e., Si = 0 mg/L <strong>and</strong> PO 4 -P = 0 μg/L)<br />

The Fe 3+ concentration ranges used in the calculations were 8.952e-6 M (0.5 mg/L) to<br />

8.952e-5 (5 mg/L) in the case of As(V) adsorption <strong>and</strong> 1.791e-5 M (1 mg/L) to 1.791e-4 (10<br />

mg/L) in the case of As(III) adsorption.<br />

As(III) AND As(V) ADSORPTION ISOTHERMS BASED ON MINEQL+ MODELING<br />

The arsenic adsorbed for each dose of FeCl 3 was calculated by Mineql+ for the 18 pH,<br />

silica, phosphate, <strong>and</strong> vanadate conditions described above for As(V) <strong>and</strong> As(III). The resulting<br />

theoretical isotherms are compared in Figures 6.2-6.7.<br />

Figure 6.2 shows the As(V) adsorption isotherms onto Fe(III) hydroxide at pH 6.5 for all<br />

the six cases modeled. Only two different isotherms resulted from the six cases modeled. It can<br />

be seen that the adsorption isotherms for the three cases which had phosphate (lower curves)<br />

were different from other cases which did not have phosphate (upper curves). So presence of<br />

phosphate significantly reduced the adsorption capacity at pH 6.5. But the presence of neither<br />

silica nor vanadate had any effect on As(V) adsorption capacity according to the Mineql+<br />

program. Figure 6.3 shows the As(V) adsorption isotherms onto Fe(III) hydroxide at pH 7.5 for<br />

all the six cases modeled. It can be seen that the adsorption isotherms are similar in all cases. So<br />

72


presence of silica, phosphate, <strong>and</strong> vanadate did not have any effect on the adsorption capacity at<br />

pH 7.5 according to the Mineql + program. Figure 6.4 shows the As(V) adsorption isotherms<br />

onto Fe(III) hydroxide at pH 8.5 for all the six cases modeled. It can be seen that the adsorption<br />

isotherms are similar in all cases. So presence of silica, phosphate, <strong>and</strong> vanadate did not have any<br />

effect on the adsorption capacity at pH 8.5 according to the Mineql + program.<br />

q e (μg As(V)/mg Fe(III))<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

NSFI-53 <strong>with</strong> no silica<br />

or phosphate<br />

NSFI-53 <strong>with</strong>out phosphate<br />

NSFI-53 <strong>with</strong><br />

vanadate<br />

NSFI-53<br />

NSFI-53 <strong>with</strong> vanadate <strong>and</strong><br />

<strong>with</strong>out silica or phosphate<br />

NSFI-53 <strong>with</strong>out silica<br />

Mineql+ Modeling Results<br />

0 5 10 15 20 25 30<br />

C e (μg As(V)/L)<br />

Figure 6.2 Mineql+-predicted effect of silica, phosphate, <strong>and</strong> vanadate on As(V) adsorption<br />

on the in-situ-formed Fe(III) hydroxide at pH 6.5<br />

q e (μg As(V)/mg Fe(III))<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

NSFI-53 <strong>with</strong><br />

vanadate<br />

NSFI-53<br />

<strong>with</strong>out<br />

phosphate<br />

NSFI-53 <strong>with</strong> no<br />

silica or phosphate<br />

NSFI-53 <strong>with</strong>out silica<br />

NSFI-53<br />

NSFI-53 <strong>with</strong> vanadate<br />

<strong>and</strong> <strong>with</strong>out silica or<br />

phosphate<br />

Mineql+ Modeling Results<br />

0 5 10 15 20 25 30 35<br />

C e (μg As(V)/L)<br />

Figure 6.3 Mineql+-predicted effect of silica, phosphate, <strong>and</strong> vanadate on As(V) adsorption<br />

on the in-situ-formed Fe(III) hydroxide at pH 7.5<br />

73


q e (μg As(V)/mg Fe(III))<br />

35<br />

28<br />

21<br />

14<br />

7<br />

0<br />

NSFI-53 <strong>with</strong><br />

vanadate<br />

NSFI-53 <strong>with</strong><br />

vanadate <strong>and</strong> <strong>with</strong>out<br />

silica or phosphate<br />

NSFI-53 <strong>with</strong>out<br />

phosphate<br />

NSFI-53 <strong>with</strong> no<br />

silica or phosphate<br />

NSFI-53 <strong>with</strong>out silica<br />

NSFI-53<br />

Mineql+ Modeling Results<br />

0 5 10 15 20 25 30 35 40<br />

C e (μg As(V)/L)<br />

Figure 6.4 Mineql+-predicted effect of silica, phosphate, <strong>and</strong> vanadate on As(V) adsorption<br />

on the in-situ-formed Fe(III) hydroxide at pH 8.5<br />

q e (μg As(III)/mg Fe(III))<br />

18<br />

16<br />

14<br />

12<br />

10<br />

8<br />

6<br />

4<br />

2<br />

0<br />

NSFI <strong>with</strong> no silica<br />

or phosphate<br />

NSFI-53 <strong>with</strong> vanadate<br />

<strong>and</strong> <strong>with</strong>out silica or<br />

phosphate<br />

NSFI-53 <strong>with</strong><br />

vanadate<br />

NSFI-53 <strong>with</strong>out<br />

phosphate<br />

NSFI-53 <strong>with</strong>out<br />

silica<br />

NSFI-53<br />

Mineql+ Modeling Results<br />

0 5 10 15 20 25 30 35 40<br />

C e (μg As(III)/L)<br />

Figure 6.5 Mineql+-predicted effect of silica, phosphate, <strong>and</strong> vanadate on As(III)<br />

adsorption on the in-situ-formed Fe(III) hydroxide at pH 6.5<br />

Figure 6.5 shows the As(III) adsorption isotherms onto Fe(III) hydroxide at pH 6.5 for all<br />

the six cases modeled. Only two different isotherms resulted from the six cases modeled. It can<br />

be seen that the adsorption isotherms for the three cases which had phosphate (lower curves) are<br />

different from other cases which did not have phosphate (upper curves). So, the presence of<br />

phosphate reduced the adsorption capacity at pH 6.5. But the presence of neither silica nor<br />

vanadate had any effect on the adsorption capacity according to the Mineql + program. Figure<br />

6.6 shows the As(III) adsorption isotherms onto Fe(III) hydroxide at pH 7.5 for all the six cases<br />

74


modeled. It can be seen that the adsorption isotherms are similar in all cases. So neither the<br />

presence of silica, phosphate, nor vanadate had any effect on the adsorption capacity at pH 7.5<br />

according to the Mineql + program. Figure 6.7 shows the As(III) adsorption isotherms onto<br />

Fe(III) hydroxide at pH 8.5 for all the six cases modeled. It can be seen that the adsorption<br />

isotherms are similar in all cases. So, neither the presence of silica, phosphate, nor vanadate had<br />

any effect on the adsorption capacity at pH 8.5 according to the Mineql + program.<br />

q e (μg As(III)/mg Fe(III))<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

NSFI-53 <strong>with</strong> no silica<br />

or phosphate<br />

NSFI-53 <strong>with</strong>out<br />

phosphate<br />

NSFI-53 <strong>with</strong><br />

vanadate<br />

NSFI-53 <strong>with</strong>out silica<br />

NSFI-53<br />

NSFI-53 <strong>with</strong> vanadate <strong>and</strong><br />

<strong>with</strong>out silica or phosphate<br />

Mineql+ Modeling Results<br />

0 5 10 15 20 25 30 35<br />

C e (μg As(III)/L)<br />

Figure 6.6 Mineql+-predicted effect of silica, phosphate, <strong>and</strong> vanadate on As(III)<br />

adsorption on the in-situ-formed Fe(III) hydroxide at pH 7.5<br />

q e (μg As(III)/mg Fe(III))<br />

20<br />

15<br />

10<br />

5<br />

0<br />

NSFI-53 <strong>with</strong> vanadate<br />

<strong>and</strong> <strong>with</strong>out silica or<br />

phosphate<br />

NSFI-53 <strong>with</strong><br />

vanadate<br />

NSFI-53 <strong>with</strong> no<br />

silica or phosphate<br />

NSFI-53 <strong>with</strong>out silica<br />

NSFI-53 <strong>with</strong>out<br />

phosphate<br />

NSFI-53<br />

Mineql+ Modeling Results<br />

0 5 10 15 20 25 30 35<br />

C e (μg As(III)/L)<br />

Figure 6.7 Mineql+-predicted effect of silica, phosphate, <strong>and</strong> vanadate on As(III)<br />

adsorption on the in-situ-formed Fe(III) hydroxide at pH 8.5<br />

75


COMPARISON OF MODEL-PREDICTED VS EXPERIMENTALLY OBSERVED<br />

ARSENIC ADSORPTION ONTO Fe(OH) 3 (S)<br />

The adsorption isotherms previously plotted were used to calculate the adsorption<br />

capacity for an equilibrium concentration of 10 μg/L As(V) or As(III) as explained earlier. The<br />

results were then compared (Figures 6.8-6.13) <strong>with</strong> the experimental results obtained during<br />

coagulation experiments.<br />

Figures 6.8, 6.9, <strong>and</strong> 6.10 compare the As(V) adsorption capacities of Fe(OH) 3 by model<br />

<strong>and</strong> experiment at pH 6.5, 7.5, <strong>and</strong> 8.5, respectively. For different compositions of challenge<br />

water, the modeling results differed significantly from the experimental results, which was<br />

expected in light of the adsorption isotherms generated from the model (Figs. 6.2, 6.3, <strong>and</strong> 6.4).<br />

The Mineql+ program did not show any effect of silica, phosphate, or vanadate except in the<br />

presence of phosphate at pH 6.5. But from experiment results, the presence of silica, phosphate,<br />

<strong>and</strong> vanadate had effect on the adsorption of As(V), which was not predicted by the unmodified<br />

Mineql+ 4.50 model.<br />

q e (mg As(V)/g Fe(III))<br />

200<br />

160<br />

120<br />

80<br />

40<br />

Model<br />

Experiment<br />

0<br />

NSFI-53<br />

NSFI-53<br />

<strong>with</strong>out<br />

phosphate<br />

NSFI-53 <strong>with</strong><br />

vanadate<br />

NSFI-53 <strong>with</strong><br />

vanadate &<br />

<strong>with</strong>out silica<br />

or phosphate<br />

NSFI-53<br />

<strong>with</strong>out silica<br />

NSFI-53 <strong>with</strong><br />

no silica or<br />

phosphate<br />

Figure 6.8 Comparison of As(V) adsorption capacities of Fe(OH) 3 for an equilibrium<br />

concentration of 10 μg/L at pH 6.5<br />

76


q e (mg As(V)/g Fe(III))<br />

160<br />

120<br />

80<br />

40<br />

Model<br />

Experiment<br />

0<br />

NSFI-53<br />

NSFI-53<br />

<strong>with</strong>out<br />

phosphate<br />

NSFI-53 <strong>with</strong><br />

vanadate<br />

NSFI-53 <strong>with</strong><br />

vanadate &<br />

<strong>with</strong>out silica<br />

or phosphate<br />

NSFI-53<br />

<strong>with</strong>out silica<br />

NSFI-53 <strong>with</strong><br />

no silica or<br />

phosphate<br />

Figure 6.9 Comparison of As(V) adsorption capacities of Fe(OH) 3 for an equilibrium<br />

concentration of 10 μg/L at pH 7.5<br />

100<br />

q e (mg As(V)/g Fe(III))<br />

80<br />

60<br />

40<br />

20<br />

Model<br />

Experiment<br />

0<br />

NSFI-53<br />

NSFI-53<br />

<strong>with</strong>out<br />

phosphate<br />

NSFI-53 <strong>with</strong><br />

vanadate<br />

NSFI-53 <strong>with</strong><br />

vanadate &<br />

<strong>with</strong>out silica<br />

or phosphate<br />

NSFI-53<br />

<strong>with</strong>out silica<br />

NSFI-53 <strong>with</strong><br />

no silica or<br />

phosphate<br />

Figure 6.10 Comparison of As(V) adsorption capacities of Fe(OH) 3 for an equilibrium<br />

concentration of 10 μg/L at pH 8.5<br />

A similar comparison between model-predicted <strong>and</strong> experimentally observed As(III)<br />

adsorption was made <strong>and</strong> the results are shown in Figures 6.11-6.13. Figures 6.11, 6.12, <strong>and</strong> 6.13<br />

have compared the As(III) adsorption capacities of Fe(OH) 3 by model <strong>and</strong> experiment at pH 6.5,<br />

7.5, <strong>and</strong> 8.5, respectively. For different compositions of challenge water, the modeling results<br />

differed significantly from the experimental results, which was expected in light of the<br />

77


adsorption isotherms generated from the model (Figs. 6.5, 6.6, <strong>and</strong> 6.7). The Mineql+ program<br />

did not show any effect of silica, phosphate, or vanadate except in the presence of phosphate at<br />

pH 6.5. But from experiments, the presence of silica, phosphate, <strong>and</strong> vanadate did significantly<br />

affect the adsorption of As(III) which was not predicted by the Mineql+ model.<br />

q e (mg As(III)/g Fe(III))<br />

7<br />

6<br />

5<br />

4<br />

3<br />

2<br />

1<br />

Model<br />

Experiment<br />

0<br />

NSFI-53<br />

NSFI-53<br />

<strong>with</strong>out<br />

phosphate<br />

NSFI-53 <strong>with</strong><br />

vanadate<br />

NSFI-53 <strong>with</strong><br />

vanadate &<br />

<strong>with</strong>out silica<br />

or phosphate<br />

NSFI-53<br />

<strong>with</strong>out silica<br />

NSFI-53 <strong>with</strong><br />

no silica or<br />

phosphate<br />

Figure 6.11 Comparison of As(III) adsorption capacities of Fe(OH) 3 for an equilibrium<br />

concentration of 10 μg/L at pH 6.5<br />

10<br />

q e (mg As(III)/g Fe(III))<br />

8<br />

6<br />

4<br />

2<br />

Model<br />

Experiment<br />

0<br />

NSFI-53<br />

NSFI-53<br />

<strong>with</strong>out<br />

phosphate<br />

NSFI-53 <strong>with</strong><br />

vanadate<br />

NSFI-53 <strong>with</strong><br />

vanadate &<br />

<strong>with</strong>out silica<br />

or phosphate<br />

NSFI-53<br />

<strong>with</strong>out silica<br />

NSFI-53 <strong>with</strong><br />

no silica or<br />

phosphate<br />

Figure 6.12 Comparison of As(III) adsorption capacities of Fe(OH) 3 for an equilibrium<br />

concentration of 10 μg/L at pH 7.5<br />

78


q e (mg As(III)/g Fe(III))<br />

20<br />

16<br />

12<br />

8<br />

4<br />

0<br />

NSFI-53<br />

NSFI-53<br />

<strong>with</strong>out<br />

phosphate<br />

NSFI-53 <strong>with</strong><br />

vanadate<br />

NSFI-53 <strong>with</strong><br />

vanadate &<br />

<strong>with</strong>out silica<br />

or phosphate<br />

Model<br />

Experiment<br />

NSFI-53<br />

<strong>with</strong>out silica<br />

NSFI-53 <strong>with</strong><br />

no silica or<br />

phosphate<br />

Figure 6.13 Comparison of As(III) adsorption capacities of Fe(OH) 3 for an equilibrium<br />

concentration of 10 μg/L at pH 8.5<br />

CONCLUSIONS<br />

The Mineql+ (version 4.50) program used for modeling the adsorption of arsenic onto<br />

Fe(OH) 3 during coagulation gave results significantly different from the experimental results.<br />

The reasons for the variation could be due to the following.<br />

1. Co-precipitation of arsenic onto iron solids is not included.<br />

2. According to the model, there is no adsorption of silica onto Fe(OH) 3 , whereas silica<br />

adsorption does occur experimentally.<br />

3. According to the model, there is no adsorption of vanadate onto Fe(OH) 3 , whereas<br />

vanadate adsorption does occur experimentally.<br />

4. According to the model, phosphate is precipitated at pH greater than 6.5 <strong>and</strong> so does not<br />

get adsorbed onto Fe(OH) 3 . So phosphate does not affect the adsorption of arsenic at pH<br />

7.5 <strong>and</strong> 8.5 according to model, whereas phosphate adsorption does occur at pH 7.5 <strong>and</strong><br />

8.5 experimentally.<br />

Thus, the unmodified Mineql+ model (version 4.50), although widely used for chemical<br />

equilibrium modeling, cannot be used to accurately model arsenic adsorption onto Fe(OH) 3<br />

during coagulation experiments.<br />

79


CHAPTER 7<br />

SUMMARY AND CONCLUSIONS<br />

SUMMARY<br />

The objectives of the proposed work were to determine the technical <strong>and</strong> economic<br />

feasibility of using aluminum, zirconium <strong>and</strong> titanium salts in comparison <strong>with</strong> ferric salts as<br />

coagulants for arsenic removal in coagulation-filtration processes. The project objectives were<br />

attained during an experimental <strong>and</strong> modeling study comprising seven phases:<br />

1. study the removal efficiencies of As(V) <strong>and</strong> As(III) during coagulation <strong>with</strong> alum, TiCl 4 ,<br />

TiCl 3 , TiOCl 2 , ZrCl 4 , ZrOCl 2 salts in comparison <strong>with</strong> FeCl 3 ,<br />

2. establish the effect of pH on As(V) <strong>and</strong> As(III) removals using these seven coagulants,<br />

3. compare the As(III)/As(V) adsorption on a molar, mass <strong>and</strong> cost basis for these seven<br />

coagulants <strong>and</strong> recommend the best coagulant,<br />

4. study the effect of competing ions (silica, phosphate <strong>and</strong> vanadate) in NSFI-53 challenge<br />

water on As(V) <strong>and</strong> As(III) removal during coagulation <strong>with</strong> FeCl 3 , alum, TiCl 4 , TiCl 3 ,<br />

<strong>and</strong> ZrCl 4 ,<br />

5. quantify the individual effects of silica, phosphate <strong>and</strong> vanadate during coagulation <strong>with</strong><br />

FeCl 3 ,<br />

6. establish the toxicity characteristics of the sludges produced during coagulation <strong>with</strong><br />

FeCl 3 , alum, TiCl 4 , TiCl 3 , <strong>and</strong> ZrCl 4 , <strong>and</strong> compare the results <strong>with</strong> the regulatory limits of<br />

the TCLP <strong>and</strong> WET tests, <strong>and</strong><br />

7. determine if the adsorption of As(V) <strong>and</strong> As(III) onto Fe(OH) 3 could be accurately<br />

modeled using the MINEQL+ (version 4.50) water chemistry equilibrium model <strong>and</strong><br />

compare the model results <strong>with</strong> experimental results.<br />

CONCLUSIONS<br />

The study of the removal of arsenic in NSFI-53 challenge water <strong>with</strong> the innovative<br />

coagulants (TiCl 4 , TiCl 3 , TiOCl 2 , ZrCl 4 , <strong>and</strong> ZrOCl 2 ) compared to commonly used coagulants<br />

(FeCl 3 <strong>and</strong> alum) resulted in the following conclusions:<br />

1. The percent removal of As(V) was highly pH dependent in the NSFI-53 challenge water,<br />

<strong>and</strong> the removal increased <strong>with</strong> decreasing pH for all coagulants tested: FeCl 3 , alum,<br />

TiCl 4 , TiCl 3 , TiOCl 2 , ZrCl 4 , <strong>and</strong> ZrOCl 2 . In particular, the adsorption capacity of As(V)<br />

<strong>with</strong> zirconium salts decreased significantly <strong>with</strong> increasing pH.<br />

2. The percent removal efficiency of As(III) was independent of pH for FeCl 3 , TiCl 4 , ZrCl 4 ,<br />

<strong>and</strong> ZrOCl 2 , <strong>and</strong> it decreased <strong>with</strong> increasing pH for TiCl 3 <strong>and</strong> increased <strong>with</strong> increasing<br />

pH for TiOCl 2 . The removal of As(III) by alum was insignificant..<br />

3. At all doses, the removal efficiency of As(V) was significantly greater than As(III) at pH<br />

6.5, 7.5, <strong>and</strong> 8.5 <strong>with</strong> all seven coagulants tested: FeCl 3 , alum, TiCl 4 , TiCl 3 , TiOCl 2 ,<br />

ZrCl 4 , <strong>and</strong> ZrOCl 2 .<br />

81


4. When comparing arsenic adsorption isotherms for all the coagulants, the highest As(V)<br />

loadings on a coagulant on a mass basis (mg As(V)/g metal) were observed <strong>with</strong> FeCl 3 ,<br />

which performed better than aluminum, titanium <strong>and</strong> zirconium salts at pHs of 6.5 <strong>and</strong><br />

7.5. However, at pH 8.5, As(V) loadings on FeCl 3 were approximately the same as TiCl 3<br />

at equilibrium As(V) ≤ 10 μg/L.<br />

5. When comparing adsorption isotherms, the highest As(V) loading on any coagulant on a<br />

molar basis or a mass basis was observed for ferric chloride at all three pHs, <strong>and</strong> the<br />

As(V) loading on iron was significantly greater than aluminum.<br />

6. On a mass basis, the comparison of As(V) adsorption capacities of titanium <strong>and</strong><br />

zirconium salts <strong>with</strong> FeCl 3 for an equilibrium concentration of 10 μg/L was as follows:<br />

• pH 6.5: FeCl 3 > Alum > ZrOCl 2 ≈ ZrCl 4 > TiCl 4 > TiCl 3 > TiOCl 2<br />

• pH 7.5: FeCl 3 >> TiCl 3 > TiCl 4 > Alum ≈ ZrOCl 2 > ZrCl 4 > TiOCl 2<br />

• pH 8.5: FeCl 3 ≈ TiCl 3 > TiCl 4 > Alum > ZrCl 4 ≈ TiOCl 2 > ZrOCl 2<br />

7. Regardless of the basis of comparison (mass or molar), FeCl 3 was a far better coagulant<br />

than alum for As(V) removal.<br />

8. When comparing arsenic adsorption isotherms for all the coagulants, the highest As(III)<br />

loading on a coagulant (mg As(III)/g metal) was observed <strong>with</strong> titanium(III) chloride,<br />

which performed better than ferric, titanium(IV), <strong>and</strong> zirconium salts at pHs of 6.5 <strong>and</strong><br />

7.5. However Ti(III) had similar adsorption capacity to that of Fe(III) <strong>and</strong> Ti(IV)<br />

coagulants at pH 8.5. Alum did not have any adsorption capacity for As(III). TiCl 4<br />

exhibited similar removal efficiency to that of FeCl 3 , <strong>and</strong> TiOCl 2 offered similar removal<br />

efficiency to FeCl 3 at pH 8.5. <strong>Zirconium</strong> salts did not have good adsorption capacity for<br />

As(III). Thus, it appears that on an mg metal/L basis, TiCl 3 could be a better coagulant for<br />

As(III) removal in coagulation-filtration processes.<br />

9. On a mass basis, the comparison of As(III) adsorption capacities of titanium <strong>and</strong><br />

zirconium salts <strong>with</strong> ferric chloride for an equilibrium concentration of 10 μg/L was as<br />

follows:<br />

• pH 6.5: TiCl 3 > TiCl 4 ≈ FeCl 3 > TiOCl 2 >> ZrCl 4 ≈ ZrOCl 2 >> Alum<br />

• pH 7.5: TiCl 3 > TiCl 4 ≈ FeCl 3 > TiOCl 2 >> ZrCl 4 ≈ ZrOCl 2 >> Alum<br />

• pH 8.5: TiCl 3 ≈ TiCl 4 ≈ FeCl 3 ≈ TiOCl 2 >> ZrCl 4 ≈ ZrOCl 2 >> Alum<br />

10. On a molar basis, the highest As(III) loading on a coagulant was observed for<br />

titanium(III) chloride at pH 6.5 <strong>and</strong> 7.5, while at pH 8.5, ferric chloride <strong>and</strong> had the<br />

highest molar adsorption capacity.<br />

11. When comparing chemical costs for FeCl 3 , alum, TiCl 4 , ZrOCl 2 , <strong>and</strong> TiOCl 2 coagulation<br />

to remove As(V) or As(III) the most economical was FeCl 3 . Of the common coagulants,<br />

alum was found to be 4-8 times more expensive than ferric chloride for As(V). The<br />

82


chemical cost of ferric chloride coagulation was calculated to be more than 5 to 20 times<br />

higher for As(III) treatment compared <strong>with</strong> As(V).<br />

12. There was experimental evidence that the high removal efficiency of As(III) by TiCl 3 <strong>and</strong><br />

the unusual As(III) behavior of increasing removal <strong>with</strong> decreasing pH was due to<br />

oxidation of As(III) to As(V) by H 2 O 2 , which based on the literature, formed from Ti(III)<br />

hydrolysis in the NSFI challenge water, which contained some dissolved oxygen. In spite<br />

of its partial oxidation, the experimentally observed removal of As(V) oxidized from<br />

As(III) was far less than the removal of a similar starting concentration of As(V), because<br />

(a) the floc was already formed when it contacted As(V), <strong>and</strong> (b) the As(III) oxidation<br />

continued for many hours during which the Ti(OH) 3 formed had settled <strong>and</strong> was not in<br />

contact <strong>with</strong> As(V) formed.<br />

The studies on effect of competing ions (silica, phosphate <strong>and</strong> vanadate) on arsenic<br />

adsorption in NSFI-53 challenge water <strong>with</strong> FeCl 3 , alum, TiCl 4 , TiCl 3 , <strong>and</strong> ZrCl 4 as coagulants<br />

resulted in the following conclusions:<br />

13. In the NSFI-53 challenge water <strong>with</strong> phosphate, silica significantly reduced the<br />

adsorption of arsenic presumably by competing for adsorption sites. The As(V) <strong>and</strong><br />

As(III) removal efficiencies in the absence of silica were higher than in the presence of<br />

silica for all coagulants tested (FeCl 3 , TiCl 4 , TiCl 3 , <strong>and</strong> ZrCl 4 ). With alum as coagulant,<br />

silica significantly affected As(V) adsorption at pH 6.5-7.5, while it had no significant<br />

effect at pH 8.5. In the NSFI-53 challenge water <strong>with</strong>out silica, the adsorption of As(III)<br />

increased <strong>with</strong> increasing pH for TiCl 4 , FeCl 3 <strong>and</strong> ZrCl 4 , whereas pH did not significantly<br />

affect As(III) adsorption on these coagulants in the st<strong>and</strong>ard challenge water <strong>with</strong> silica<br />

present. The detrimental effect of silica on As(III) removal increased <strong>with</strong> increasing pH<br />

for all coagulants except TiCl 3 .<br />

14. In the NSFI-53 challenge water <strong>with</strong> silica, phosphate was found to reduce the adsorption<br />

of As(V) significantly at pH 6.5 <strong>and</strong> 7.5, whereas it had a lesser effect at pH 8.5 <strong>with</strong> all<br />

coagulants tested (FeCl 3 , TiCl 3 , TiCl 4 , <strong>and</strong> ZrCl 4 ). However <strong>with</strong> alum as coagulant,<br />

phosphate did not affect As(V) adsorption. In contrast to As(V) adsorption, the presence<br />

of phosphate did not significantly affect the adsorption of As(III) <strong>with</strong> the coagulants<br />

tested (FeCl 3 , TiCl 3 , TiCl 4 , <strong>and</strong> ZrCl 4 ).<br />

15. In the NSFI-53 challenge water <strong>with</strong> silica <strong>and</strong> phosphate, vanadate did not significantly<br />

affect the adsorption of As(V) <strong>with</strong> all coagulants tested (FeCl 3 , alum, TiCl 3 , TiCl 4 , <strong>and</strong><br />

ZrCl 4 ). Similarly, in the NSFI-53 challenge water <strong>with</strong> silica <strong>and</strong> phosphate, vanadate<br />

did not significantly affect the adsorption of As(III) <strong>with</strong> all coagulants tested (FeCl 3 ,<br />

TiCl 3 , TiCl 4 , <strong>and</strong> ZrCl 4 )<br />

The studies on individual effect of competing ions (silica, phosphate <strong>and</strong> vanadate) on<br />

arsenic adsorption <strong>with</strong> FeCl 3 as coagulant resulted in the following conclusions:<br />

16. In the absence of other competing ions, it was found that silica, phosphate, <strong>and</strong><br />

vanadate exhibited significant competitive effects on the adsorption of As(V) <strong>and</strong> As(III).<br />

83


17. In the absence of phosphate <strong>and</strong> vanadate, silica significantly decreased the adsorption<br />

of As(V) <strong>and</strong> As(III), <strong>and</strong> the effect increased <strong>with</strong> increasing pH.<br />

18. In the absence of silica <strong>and</strong> vanadate, phosphate significantly decreased the adsorption<br />

of As(V) at all pHs, <strong>and</strong> a had minor effect on the adsorption of As(III) at pH 7.5 <strong>and</strong> 8.5.<br />

19. In the absence of silica <strong>and</strong> phosphate, vanadate significantly decreased the adsorption<br />

of As(V), <strong>and</strong> had a minor effect on the adsorption of As(III) at pH 7.5 <strong>and</strong> 8.5 during<br />

FeCl 3 coagulation.<br />

20. Based on the FeCl 3 experimental results <strong>with</strong> <strong>and</strong> <strong>with</strong>out multiple competing<br />

contaminants the following inferences were made:<br />

• The presence of silica significantly reduced the magnitude of the phosphate effect<br />

on As(V) adsorption at pH 7.5 <strong>and</strong> 8.5.<br />

• The presence of silica reduced the effect of phosphate in the case of As(III)<br />

adsorption at all pHs.<br />

• The combined presence of silica <strong>and</strong> phosphate reduced the effect of vanadate at<br />

all pHs in the case of As(V) <strong>and</strong> As(III) removal.<br />

The toxicity characteristics of the sludges produced during coagulation <strong>with</strong> FeCl 3 , alum,<br />

TiCl 3 , TiCl 4 , <strong>and</strong> ZrCl 4 led to the following conclusions:<br />

21. The adsorption capacity of As(V) was independent of initial arsenic concentration <strong>with</strong><br />

all the coagulants tested (FeCl 3 , alum, TiCl 3 , TiCl 4 , <strong>and</strong> ZrCl 4 ).<br />

22. Following coagulation <strong>and</strong> 3-hr settling, <strong>with</strong> all four coagulants tested (FeCl 3 , alum,<br />

TiCl 3 , TiCl 4 , <strong>and</strong> ZrCl 4 ), the liquid wastes obtained were found to contain less than 0.5%<br />

dry solids, <strong>and</strong> according to the TCLP regulations, did not require extraction. The liquid<br />

filtrate, which is considered to be the extract in such cases, easily passed the TCLP <strong>and</strong><br />

WET test regulatory limit of 5 mg/L arsenic.<br />

The attempt to model As(V)/As(III) adsorption during coagulation <strong>with</strong> FeCl 3 using<br />

Mineql+ (version 4.50) program led to the following conclusions:<br />

23. The unmodified Mineql+ chemical equilibrium modeling program could not simulate<br />

As(III) or As(V) adsorption onto Fe(OH) 3 during FeCl 3 coagulation.<br />

24. The model predicts significantly less adsorption onto Fe(OH) 3 compared <strong>with</strong><br />

experimental results, <strong>and</strong> the model also does not take into account the effect of silica,<br />

phosphate <strong>and</strong> vanadate on As(V)/As(III) adsorption in most cases.<br />

84


RECOMMENDATIONS<br />

The main purpose of this project was to determine the technical <strong>and</strong> economic feasibility<br />

of using aluminum, zirconium <strong>and</strong> titanium salts in comparison <strong>with</strong> ferric salt as coagulants for<br />

arsenic removal in coagulation-filtration processes. Based on adsorption <strong>and</strong> economic<br />

comparisons, this work showed that FeCl 3 was clearly superior to the other coagulants tested.<br />

Although Ti(III) had the highest removal efficiency for As(III), its chemical cost was not<br />

available <strong>and</strong> is expected to be higher considering the cost of Ti(IV) salts. A chemical cost<br />

comparison of commonly used coagulants showed that alum was 4-6 times more expensive than<br />

FeCl 3 for As(V) removal. Taking the detailed conclusions above into consideration, the<br />

following recommendations are made for application of coagulation for arsenic removal in<br />

drinking water treatment systems:<br />

1. FeCl 3 should be considered as the preferred coagulant for As(V) <strong>and</strong> As(III) removal at<br />

all pHs <strong>and</strong> background water compositions.<br />

2. Alum could be considered as a coagulant for As(V) removal at pH ≤6.5 where its As(V)<br />

capacity is closer to that of FeCl 3 .<br />

3. As(III) should be pre-oxidized to As(V) for cost-effective treatment <strong>with</strong> alum <strong>and</strong> FeCl 3<br />

coagulants.<br />

4. <strong>Zirconium</strong> <strong>and</strong>/or titanium could be considered as alternative coagulants to alum, if their<br />

prices drop significantly.<br />

85


REFERENCES<br />

Ahmed, M.F., <strong>and</strong> M.M. Rahaman. 2000. Water Supply <strong>and</strong> Sanitation – Low Income Urban<br />

Communities, International Training Network (ITN) Centre, BUET.<br />

Andrae, M.O. 1977. Determination of <strong>Arsenic</strong> Species in Natural Waters. Analytical Chemistry,<br />

49(6):820.<br />

Bang, S., M. Patel, L. Lippincott, <strong>and</strong> X. Meng. 2005. Removal of <strong>Arsenic</strong> from Groundwater by<br />

Granular <strong>Titanium</strong> dioxide Adsorbent. Chemosphere, 60(3):389-397.<br />

Bissen, M., M. Vieillard-Baron, A.J. Schindelin, <strong>and</strong> F.H. Frimmel. 2001. TiO 2 -catalyzed<br />

Photooxidation of Arsenite to Arsenate in Aqueous Samples. Chemosphere, 44(4):751-<br />

757.<br />

Bissen, M., <strong>and</strong> F.H. Frimmel. 2003. <strong>Arsenic</strong>: Oxidation of <strong>Arsenic</strong> <strong>and</strong> its Removal in Water<br />

Treatment. Acta hydrochimica et hydrobiologica, 31(2):97-107.<br />

Chakraborti, D., M.M. Rahman, K. Paul, U.K. Chowdhury, M.K. Sengupta, D. Lodh, C.R.<br />

Ch<strong>and</strong>a, K.C. Saha, <strong>and</strong> S.C. Mukherjee. 2002. <strong>Arsenic</strong> Calamity in the Indian<br />

Subcontinent What Lessons Have Been Learned Talanta, 58(1):3-22.<br />

Chen, C.J., M.M. Wu, S.S. Lee, J.D. Wang, S.H. Cheng, <strong>and</strong> H.Y. Wu. 1988. Atherogenicity <strong>and</strong><br />

Carcinogenicity of High-<strong>Arsenic</strong> Artesian Well Water: Multiple Risk Factors <strong>and</strong> Related<br />

Malignant Neoplasms of Black Foot Disease. Arteriosclerosis, 8(5):452-460.<br />

Cheng, R.C., S. Liang, H. Wang, <strong>and</strong> M.D. Beuhler. 1994. Enhanced <strong>Coagulation</strong> for <strong>Arsenic</strong><br />

Removal. Jour. AWWA, 86(9):78-90.<br />

Chowdhury, U.K., B.K. Biswas, T. Roychowdhury, B.K. M<strong>and</strong>al, G. Samanta, G.K. Basu, C.R.<br />

Ch<strong>and</strong>a, D. Lodh, K.C Saha, D. Chakraborti, S.C. Mukherjee, S. Roy, S. Kabir, <strong>and</strong> Q.<br />

Quamruzzaman. 2000. <strong>Arsenic</strong> Groundwater Contamination <strong>and</strong> Sufferings of People in<br />

West Bengal-India <strong>and</strong> Bangladesh, In Trace Elements in Man <strong>and</strong> Animals. Edited by<br />

A.M. Roussel, R.A. Anderson, <strong>and</strong> A.E. Faviers. New York: Kluwer Academic/Plenum<br />

Publishers.<br />

Chowdhury, U.K., B.K. Biswas, T. Roychowdhury, G. Samanta, B.K. M<strong>and</strong>al, G.K. Basu, C.R.<br />

Ch<strong>and</strong>a, D. Lodh, K.C. Saha, S.K. Mukherjee, S. Roy, S. Kabir, Q.Quamruzzaman, <strong>and</strong><br />

D. Chakraborti. 2000a. Groundwater <strong>Arsenic</strong> Contamination in Bangladesh <strong>and</strong> West<br />

Bengal-India. Environ Health Perspec., 108(5):393-397.<br />

Clarke, S.R., R.J. Clarke, R. Murdock, C.J. Butler, <strong>and</strong> S. Mohanta. 2002. Water Treatment<br />

Method <strong>and</strong> Apparatus. U.S. Patent 6,383,395.<br />

Clifford D. A. 1990. Ion exchange <strong>and</strong> inorganic adsorption, In Water Quality <strong>and</strong> Treatment:a<br />

H<strong>and</strong>book of Community Water Supplies. 4th ed. pp.561-639. Edited by F.W. Pontius.<br />

New York: McGraw-Hill, Inc. (See also 5 th ed, 1999).<br />

Clifford, D.A., G. Ghurye, A. Tripp, <strong>and</strong> J. Tong. 1997. Final Report: Phases 1, 2, <strong>and</strong> 3 City of<br />

Albuquerque <strong>Arsenic</strong> study, Field study on <strong>Arsenic</strong> Removal in Albuquerque, New<br />

Mexico Using the University of Houston/EPA Mobile Drinking Water Treatment<br />

Research Facility, Dept. of Civil <strong>and</strong> Environmental Engineering, University of Houston,<br />

Texas.<br />

Clifford, D.A., S. Karori, G, Ghurye, <strong>and</strong> G. Samanta. 2004. Field Speciation Method for<br />

<strong>Arsenic</strong> Inorganic Species. Denver, Colo.: American Water Works Association Research<br />

Foundation.<br />

87


Das, D., A. Chatterjee, B.K. M<strong>and</strong>al, G. Samanta, D. Chakraborti, <strong>and</strong> B. Ch<strong>and</strong>a. 1995. <strong>Arsenic</strong><br />

Concentration in Drinking Water, Hair, Nails, Urine, Skin-Scale <strong>and</strong> Liver Tissue<br />

(Biopsy) of the Affected People. Analyst, 120(3):917-924.<br />

Daus, B., R. Wennrich, <strong>and</strong> H. Weiss. 2004. Sorption Materials for <strong>Arsenic</strong> Removal from Eater:<br />

a Comparative Study. Water Research, 38(12):2948-2954.<br />

Dutta, P. K., A.K. Ray, V.K. Sharma <strong>and</strong> F.J. Millero. 2004. Adsorption of Arsenate <strong>and</strong><br />

Arsenite on <strong>Titanium</strong> dioxide Suspensions. Jour. Colloid Interface Sci., 278(2):270-275.<br />

Edwards, M. 1994. Chemistry of <strong>Arsenic</strong> Removal during <strong>Coagulation</strong> <strong>and</strong> Fe-Mn Oxidation.<br />

Jour. AWWA, 86(9):64-77.<br />

Ferguson, M.A., M.R. Hoffmann, <strong>and</strong> J.G. Hering. 2005. TiO 2 -Photocatalyzed As(III) Oxidation<br />

in Aqueous Suspensions: Reaction Kinetics <strong>and</strong> Effects of Adsorption. Environ. Sci.<br />

Technol., 39(6):1880-1886.<br />

Ferreccio, C., C. González, V. Milosavjlevic, G. Marshall, A.M. Sancha, <strong>and</strong> A.H. Smith. 2000.<br />

Lung Cancer <strong>and</strong> <strong>Arsenic</strong> Concentrations in Drinking Water in Chile. Epidemiol,<br />

11(6):673-679.<br />

Ghurye, G., D.A. Clifford, <strong>and</strong> A.R. Tripp. 2004. Pilot Study of <strong>Coagulation</strong> Microfiltration for<br />

<strong>Arsenic</strong> Removal from Groundwater. Jour. AWWA, 96(4):143-152.<br />

Guha Mazumder., D.N, R. Haque, N. Ghosh, B.K. De, A. Santra, D. Chakraborty, <strong>and</strong> A.H.<br />

Smith. 1998. <strong>Arsenic</strong> Levels in Drinking Water <strong>and</strong> the Prevalence of Skin Lesions in<br />

West Bengal, India. International Journal of Epidemiology, 27(5): 871-877.<br />

Gulledge, J.H., <strong>and</strong> J.T. O’Connor. 1973. Removal of <strong>Arsenic</strong>(V) from Water by Adsorption on<br />

<strong>Aluminum</strong> <strong>and</strong> Ferric Hydroxides. Jour. AWWA, 65(8):548:552.<br />

Hering, Janet G., Chen, Pen-Yuan, Wilkie, Jennifer A., <strong>and</strong> Sun Liang. 1996. <strong>Arsenic</strong> Removal<br />

by Ferric Chloride. Jour. AWWA, 88(4):155-167.<br />

Hingston, F.J. 1981. A Review of Anion Adsorption, In Adsorption of Inorganics at Solid-<br />

LiquidInterfaces. Edited by M.A. Anderson <strong>and</strong> A.J. Rubin. Ann Arbor, MI.: Ann Arbor Sci.<br />

Publishers.<br />

Hug, S.J., <strong>and</strong> O. Leupin. 2003. <strong>Iron</strong>-Catalyzed Oxidation of <strong>Arsenic</strong>(III) by Oxygen <strong>and</strong> by<br />

Hydrogen Peroxide: pH-dependent Formation of Oxidants in the Fenton Reaction.<br />

Environ. Sci. Technol., 37(12):2734-2742.<br />

Irgolic, K. 1982. Speciation of <strong>Arsenic</strong> in Water Supplies. EPA 600/1-82-010. USEPA,<br />

Washington.<br />

Kinniburgh, D.G., <strong>and</strong> W. Kosmus. 2002. <strong>Arsenic</strong> Contamination in Groundwater: Some<br />

Analytical Considerations. Talanta, 58(1):165-180.<br />

Lee H., <strong>and</strong> W. Choi. 2002. Photocatalytic Oxidation of Arsenite in TiO 2 Suspension: Kinetics<br />

<strong>and</strong> Mechanisms. Environ. Sci. Technol., 36(17):3872-3878.<br />

Manna, B.R., S.C. Bhat., M. Dasgupta., U.C. Ghosh. 1999. Studies on Removal of <strong>Arsenic</strong> from<br />

Water Using Hydrated <strong>Zirconium</strong> Oxide. Chem. Environ. Res., 8(1,2):51–56.<br />

Masterson, W.L., <strong>and</strong> E.J. Slowinski. 1973. Chemical Principles. 3rd ed. Philadelphia, Pa.: W.B.<br />

Saunders Co.<br />

McNeill, L.S., <strong>and</strong> M. Edwards. 1997. Predicting <strong>Arsenic</strong> Removal during Metal Hydroxide<br />

Precipitation. Jour. AWWA, 89(1):75-86.<br />

Meng, X.G., M. Dadachov, G.P. Korfiatis, <strong>and</strong> C. Christodoulatos. 2005. Methods of Preparing a<br />

Surface-activated <strong>Titanium</strong> Oxide Product <strong>and</strong> of using same in Water Treatment<br />

Processes. U.S. Patent 6,919,029.<br />

NRC (National Research Council, Subcommittee on <strong>Arsenic</strong> in Drinking Water). 1999. <strong>Arsenic</strong><br />

in Drinking Water. Washington, DC: National Academy Press.<br />

88


NRDC (National Resources Defense Council). 2000. <strong>Arsenic</strong> <strong>and</strong> old Laws. A Scientific <strong>and</strong><br />

Public Health Analysis of <strong>Arsenic</strong> Occurrence in Drinking Water, its Health Effects, <strong>and</strong><br />

EPA’s Outdated <strong>Arsenic</strong> tap Water St<strong>and</strong>ard.<br />

(www.nrdc.org/water/drinking/arsenic/chap1.asp).<br />

Pena, M., G.P. Korfiatis, M. Patel, L. Lippincott, <strong>and</strong> X. Meng. 2005. Adsorption of As(V) <strong>and</strong><br />

As(III) by Nanocrystalline <strong>Titanium</strong> dioxide. Water Research, 39(11):2327-2337.<br />

Pettine, M., L. Campanella, <strong>and</strong> F.J. Millero. 1999. Arsenite Oxidation by H 2 O 2 in Aqueous<br />

Solutions. Geochim. Cosmochim. Acta, 63(18):2727-2735.<br />

Pettine, M., <strong>and</strong> F.J. Millero. 2000. Effect of Metals on the Oxidation of As(III) <strong>with</strong> H 2 O 2 .<br />

Marine Chemistry, 70(1-3):223-234.<br />

Remy, H. 1956. Treatise on Inorganic Chemistry, Volume II. New York: Elsevier Publishing<br />

Company.<br />

Schecher, W., <strong>and</strong> McAvoy. 1998. Mineql+ v. 4.5: A Chemical Equilibrium Modeling System.<br />

Hallowell, Md.: Environmental Research Software.<br />

Sigel, H., <strong>and</strong> A. Sigel. 1995. Vanadium <strong>and</strong> its Role in Life. New York: M. Dekker.<br />

Sorg, T.J., <strong>and</strong> G.S. Logsdon. 1978. Treatment Technology to Meet the Interim Primary<br />

Drinking Water Regulations for Inorganics. Jour. AWWA, 70(2):105-112.<br />

Stumm, W. 1992. Chemistry of the Solid-Water interface. NewYork: Wiley-Interscience.<br />

Suzuki, T.M., J.O. Bomani, H. Matsunaga, <strong>and</strong> T. Yokoyama. 2000. Preparation of Porous Resin<br />

Loaded <strong>with</strong> Crystalline Hydrous <strong>Zirconium</strong> Oxide <strong>and</strong> Its Application to the Removal of<br />

<strong>Arsenic</strong>. React. Funct. Polym., 43(1,2):165–172.<br />

Swaminathan, P. 2005. Comparison of arsenic(V) adsorption equilibria <strong>and</strong> kinetics of iron,<br />

zirconium, titanium, <strong>and</strong> lanthanum-based adsorbents. M.S thesis, Dept. of Civil <strong>and</strong><br />

Environmental Engineering, University of Houston, Houston, Texas.<br />

Tong, J. 1997. Development of an <strong>Iron</strong>(III)-<strong>Coagulation</strong>-Microfiltration Process for <strong>Arsenic</strong><br />

Removal from Groundwater. M.S thesis, Dept. of Civil <strong>and</strong> Environmental Engineering,<br />

University of Houston, Houston, Texas.<br />

Tripp, A.R. 2001. Selectivity Considerations in Modeling the Treatment of Perchlorate Using Ion<br />

Exchange Processes. Ph.D. diss., Dept. of Civil <strong>and</strong> Environmental Engineering,<br />

University of Houston, Houston, Texas.<br />

U.S. EPA (United States Environmental Protection Agency). 2001. Quick Reference Guide to<br />

<strong>Arsenic</strong> <strong>and</strong> Clarifications to Compliance <strong>and</strong> New Source Monitoring Rule<br />

(http://www.epa.gov/OGWDW/arsenic/pdfs/quickguide.pdf).<br />

Voegelin, A., <strong>and</strong> S.J. Hug. 2003. Catalyzed Oxidation of <strong>Arsenic</strong>(III) by Hydrogen Peroxide on<br />

the Surface of Ferrihydrite: An in Situ ATR-FTIR Study. Environ. Sci. Technol.,<br />

37(5):972-978.<br />

Zhu, X., <strong>and</strong> A. Jyo. 2001. Removal of <strong>Arsenic</strong>(V) by <strong>Zirconium</strong>-loaded Phosphoric acid<br />

Chelating Resin. Separation Science & Technology, 36(14):3175-3189.<br />

89


ABBREVIATIONS<br />

AAS<br />

As<br />

AWWA<br />

AwwaRF<br />

Atomic Absorption Spectrometer<br />

<strong>Arsenic</strong><br />

American Water Works Association<br />

Awwa Research Foundation<br />

C e<br />

Equilibrium concentration (μg As/L)<br />

Co.<br />

company<br />

°C degree Celcius<br />

DIW<br />

$/Mgal<br />

ed.<br />

EDL<br />

FI-HG-AAS<br />

g<br />

g/mL<br />

hr<br />

ICP-MS<br />

IX<br />

Jour<br />

Kg/L<br />

L<br />

M<br />

MCL<br />

min<br />

mg/g<br />

mg/L<br />

mL<br />

mM<br />

mmol/mol<br />

μg/L<br />

deionized water<br />

dollars per million gallons<br />

edition<br />

Electrodeless discharge lamp<br />

Flow Injection Hydride Generation Atomic Absorption Spectrometer<br />

gram<br />

gram per milliliter<br />

hour<br />

Inductively Coupled Plasma-Mass Spectrometer<br />

Ion Exchange<br />

Journal<br />

Kilograms/liter<br />

Liter<br />

Molar<br />

Maximum contaminant level<br />

minute<br />

milligram/gram<br />

milligram/liter<br />

milliliter<br />

millimolar<br />

millimole/mole<br />

microgram/liter<br />

91


μg/mg<br />

μL<br />

μm<br />

nm<br />

NSFI<br />

pK a<br />

POU/POE<br />

PO 4 -P<br />

ppb<br />

ppm<br />

microgram/milligram<br />

microliter<br />

micrometer<br />

nanometer<br />

National Sanitation Foundation International<br />

Negative logarithm of an ionization/ equilibrium constant<br />

Point-of-use/Point-of-entry<br />

Phosphate as Phosphorous<br />

parts per billion<br />

parts per million<br />

q e<br />

Adsorption capacity (μg As/mg metal)<br />

rpm<br />

(s)<br />

sec<br />

TCLP<br />

USEPA<br />

USGS<br />

VO 3 -V<br />

WET<br />

Revolutions per minute<br />

solid<br />

second<br />

Toxicity Characteristic Leaching Procedure<br />

U.S. Environmental Protection Agency<br />

U.S. Geological Survey<br />

Vanadate as Vanadium<br />

Waste Extraction Test<br />

92

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!