PB86-158607
    Arsenic  (3)  Oxidation and
    Removal  from Drinking Water
    Houston Univ.,  TX
                                          LIBRARY
                                          U. 5.
                                          ^EWYQRK, N.Y.  10007
    Prepared  for

    Environmental Protection Agency,  Cincinnati, OH
    Feb 86
U.S. DepvtmeRt of Coiranerce
Natimal Techrecd tofarmatwn Service

-------
                                                PB86-156607


                                          EPA/600/2-86/021
                                          February 1986
     ARSENIC(III) OXIDATION AND REMOVAL
            FROM DRINKING WATER
                     by
     Phyllis Frank and Dennis Clifford
           University of Houston
           Houston, Texas  77004
Cooperative Research Agreement No. CE-807939
              Project Officer

               Thomas J. Sorg
      Drinking Water Research Division
   Water Engineering Research Laboratory
          Cincinnati, Ohio  45268
   WATER ENGINEERING RESEARCH LABORATORY
     OFFICE OF RESEARCH AND DEVELOPMENT
    U.S. ENVIRONMENTAL PROTECTION AGENCY
          CINCINNATI, OHIO  45268
               NATIONAI  TECHNICAL
               INFORMATION SERVICE
                  JitUPARKim Of COJMIRCt
                    murjuin •» niM

-------
                                  TECHNICAL REPORT DATA
                           (Plcatf lead Instructions on thr reitrtt Injure cumplctmgl
 . REPORT NO.
  EPA/600/2-86/021
                             2.
                                                           3 RECIPIENT'S A
4 TITLE AND SUBTITLE

 Arsenic (III) Oxidation  and Removal from Drinking Water
                    5 REPOST DATE
                      February 1986
                                                           B. PERFORMING ORGANIZATION CODE
7 AUTHORIS)
 Phyllis Frank
 Dennis Clifford
                                                           B PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
 Dept. of Civil Engineering
 University of Houston,  4800 Calhoun
 Houston. TX  77004
                                                           10 PROGRAM ELEMENT NO
                    tl CONTRACTSGHANT NO
                                                               CR807939
12 SPONSORING AGENCY NAME AND ADDRESS
 Water Engineering Research  Laboratory - Cincinnati, OH
 Office of Research and  Development
 U.S. Environmental Protection  Agency
 Cincinnati, OH 45268
                                                           (3. TYPE OF REPORT AND PERIOD COVERED
                    14. SPONSOPING AGCNCV CODE


                        EPA/600/14
IS. SUPPLEMENTARY NOTES

  Thomas J. Sorg, Project  Officer
313-569-7370
16 ABSTRACT„. The oxidative  pretreannent of As(III) using chlorine  and  oxygen was studied
 following quantification of  As(III)  and As(V) removals by activated alumina columns.
 Activated alumina  removed  100 ug/L As(V) from a typical grounduater at pH 6.0 much
 more effectively than As (IT. I).   Approximately 23,500 bed volumes  of water were treated
 before As(V) reached the 0.05-mg/L maximum contaminant level  (MCL), whereas only 300
 bed volumes could  be treated before  As(IIT) reached that level.
                                                               ,'
      Variables affecting the oxidation of As(IlI) by chlorine include the pH, chloride
 concentration, other ions, chloramine formation and TOC.  In  artificial grcundwater
 containing 100 ug/L As(III), a 1.0 mg/L chlorine dose oxidized 95 percent of the
 As(III) to As(V) in less than 5 seconds.  Increasing chloride concentration slowed the
 reaction slightly, but not significantly for water treatment.  The counterion (sodium
 or calcium) did not appear to affect the extent of reaction in the artificial ground-
 water or in chloride solutions up to 0.010 M.  The presence of 5  mg/L TOC substantially
 slowed the oxidation kinetics of 100 ug/L As(IlI) by 1.0 mg/L chlorine dosage. Sparging
 1 hr with oxygen did not oxidize 100 ug/L As(III) in artificial groundwater, while 100
 ug/L As(III) in deionized  water was  approximately 14 percent  oxidized.  Monochloramine
 was found to be capablo  of oxidizing 40 percent of the initial 100 ug/L As(III) in the
 pH range of 6.5 to 10.5.
17.
                               KEY WORDS AND DOCUMENT XNALVSlS
                  DESCRIPTORS
                                             b IDENTIFIFRS/OPEN ENOED TERMS
                                                                        C COSATI I'iclli/Gioup
13. DISTRIBUTION STATEMENT

      Release to Public
       19 SECURITY CLASS ITIllI Rrporl)

          Unclassified	
                                                                            O OF PAGES
84
                                             2O SECURITY CLASS nh'tpffe
                                                 Unclassified
EPA Form 2220-1 (R«». 4-77)   POEVIO.O EDITION is OBSOLETE

-------
                                DISCLAIMER


     Although the information described In this article has been funded  wholly
or In part  by the  United States  Environmental  Protection Agency through
assistance agreement number CR-807939  to the University of Houston, it has not
been subjected to the Agency's required peer and administrative revlev and
therefore does not necessarily reflect the views of the Agency and no official
endorsement  should be inferred.
                                     11

-------
                                 FOREWORD
     The U.S. Environn* —'ta 1 Protection Agency is charged by Congress with
protecting  the  Natlo. s land, air, and water systems.  Under a mandate  of
national environmental  laws,  the agency strives  to  formulate  and  implement
actions  leading to a  compatible balance  between human activities  and the
ability of natural  systems to support and nurture  life.  The Clean Water Act,
the Safe Drinking Water  Act,  and the Toxics Substances Control  Act  are three
of the major congressional laws  that  provide the  framework for  restoring and
maintaining  the  integrity of our Nation's w&ter, for preserving and  enhancing
the water we drink,  and  for  protecting the  environment from toxic substances.
These  laws  direct  the EPA to  perform research to define our environmental
problems, measure the impacts,  and search for solutions.

     The Water Engineering  Research Laboratory  is that  component of EPA's
Research and Development program concerned with preventing,  treating, and
managing municipal and industrial wastewater  discharges;  establishing
practices to control  and remove contaminants from  drinking  water  and  to
prevent  its deterioration during storage and distribution;  and  assessing the
nature and cortrollability of  releases of toxic substances to the air, water,
and land from manufacturing  processes and  subsequent product uses.  This
publication is  one of the products of that  research and  provides a vital
communication link between the  researcher and the user  community.

     This research is part of a concerted research  effort at the  University  of
Houston to examine  the removal of inorganic  contaminants from drinking water
sources.   In particular,  this research examines the oxidation and removal  of
naturally occuring arsenic concentrations in ground water.


                                    Francis T. Mayo,  Director
                                    Water  Engineering Research Laboratory
                                    111

-------
                                 ABSTRACT
     The  concentration of arsenic In drinking water  is  regulated because of
the known toxicity of arsenic.  A survey of unit  processes for water  treatnent
and our previous research reveal  that pentavalent arsenic is  more effectively
removed  from water than trivalent arsenic:  Thus, following quantification of
As(III)  and Ae(V) removals by activated alumina  columns, the oxidative
pretreatment of trivalent arsenic using chlorine and oxygen was studied.

     Arsenic(V) at a concentration of  0.100 mg/L was much more effectively
removed  from a typical groundwater by  activated  alumina at pH 6.0 than
As(III).   Approximately 23,500 bed volumes (BV> could be treated  before  As(V)
reached  the  0.05 mg/L MCL whereas only  300 BV  could be  treated before As(III)
reached  the  MCL.

     Variables affecting the oxidation  of  As(III) by chlorine include the  pH,
chloride concentration,  other ions,  chloramine  formation and  TOC.  In
artificial  ground water containing no amirDnla or TOC vlth  100 ppb As(III)
present  initially and  1.0 mg/L chlorine dosage,  the  reaction  reached 95
percent  completion in less  than  our shortest  possible observation time of 5
seconds.  Thus with 1.0 tig/L chlorine dosage  the As(lll) oxidation rate vas
greater than 20 mlcrograms/L second.  The extent of oxidation at 30 seconds
was insensitive  to pH in  the range 6.5 to 9.5, with decreasing extent of
reaction outside  this range.   Increasing chloride  concentration slowed  the
reaction  slightly,  although  this effect  is not  significant for  water
treatment.  The counterlon (sodium or calcium) did not  appear to effect the
extent of reaction in the artificial goundwater or in chloride solutions  up to
0.010 molar.  Honochloranlne is capable of oxidizing 40 percent of the initial
100 ppb  As(III)  in the pH  range  6.5  to  10.5.  The presence of  5  mg/L TOC
substantially slowed the oxidation kinetics of 100 ppb As(IZl) by 1.0 mg/L
chlorine dosage.  Although the reaction reached 50  percent  completion in less
than 30  seconds. It did not  reach 80 percent completion until  approximately 30
minutes.

     Sparging  1 hour with oxygen  did not oxidize 100 ppb  As(III)  in artificial
ground water,  while  100  ppb As(lII) in  DI  water  was approximately 14 percent
oxidized.  However,  capped samples of As(III)  in  DI water   and  artificial
ground water were  completely oxidized after 61 days on  the shelf with air In
Che head space.

     This study  is part  of a comprehensive research effort on arsenic removal
froa drinking water  being carried  out by researchers  at the University of
Houston.   The  related work includes field studies  on As(IIl)/As(V) rrooval in
San Ysld.ro,  New Mexico and Ae(III) removal in  Hanford,  California.   Previous


                                      iv

-------
related laboratory studies include work on  the  removal of As(V) by activated
alumina and anion exchange resins,  and development of an analytical method for
arsenic speciation.

     This report was submitted in fulfi 1 latent of Crart No. CR-807939 by the
University  of Hourton under  the sponsorship of the U.S.  Environmental
Protection Agency.  This  report covers the period May 1982 to Hay 1984, and
this work  was completed as of December 1984.

-------
                                CONTENTS
Abstract	iv
Figures	ix
Tables	xi
Abbreviations  and  Symbols	xii
Acknowledgments  	  ............... 	  xiii

1.  Introduction 	  1
    General  Chemistry 	  1
    Occurrence   	  2
    Toxlcity  	  3
    Arsenic Removal  	  4
         Removal  by Coc&ulation	'•
         Removal  during  Filtration  	  7
         Removal  by Reverse Osmosis 	  8
         Removal  by Electrodialysis 	  8
         Removal  by Adsorption  	  9
         Removal  by Ion  Exchange   	10
    Summary	•	..11

2.  Conclusions	;... .12

3.  Recommendations   ......  	  .....13

4.  Theory	**
    Aqueous  Arsenic Chemistry	1*
    Oxidation  Studies 	17
    Aqueous  Chlorine  Chemistry  	17
         Free  Chlorine   	19
         Combined Chlorine	19
         Chlorine  Oxidant Decay 	20
    Reaction Thermodynamics	'	20
    Reaction  Mechanisms 	22
    Oxygen  Oxidation	24

5.  Experiments	25
    Preliminary Column  Studies  	25
         Column Experiment  Design	25
    Oxidation  Studies	26
         Objectives	26
         Reagents	26
         Procedure  Development for  Chlorine Oxidation	27
                                      vii

-------
              Ammonia  quenching  .................... 28
              Thioleulfate quenching  ................. 29
              Sodium Sulfite quenching   ................ 29
              DPD quenching ...................... 30
         Chlorine  Oxldant  Experiments ................. 32
         Oxygen  Oxidant  Experioente ............. ....  .33
6.  Results  and Discussion ................ ...... .36
     Results  and Discussion of Column Studies   ............. 34
     Chlorine  Oxldant Studies  ..................... 62
         Kinetics  ........................... *2
         pH  Effect   .......................... *5
         Effect of Chloride Concentration ............... *7
         Effect  of  Counterion ..................... 69
         Effect of  Chloramines   .................... *'
         Effect of pH on  Chloramint Oxidation ............. 52
         Effect of TOC  ........................ 54
     Oxygen  Oxidant  Studies   ...................... 56
         Sparging  Tests  ........................ 56
         Shelf Life Experiments .................... 56
References
                                                                  .58
Appendices	63
     1.  Equations for arsenic Eh-pH diagram	63
     2.  Equations for chlorine Fh-pH diagram	65
     3.  Chlorine Oxidant Decay Reactions	6*
     4.  Conditions Procedure for Activated Alumina	67
     5.  Arsenic Analysis	68
     6.  Change in Absorbance	69
                                     viii

-------
                                FIGURES


                                                                Page

         pH-concentration diagram for solubility of AsjO^	IS

        Eh-pB diagram cf arsenic-water system at 25 C	16

         Eh-pH diagram of chlorine-water system at 25 C	18

        Calculated free energy of As(III)  oxidation by  chlorine
        as a function of pH and  chloride concentration	23

 5      Effluent history from an activated alumina colunn with
        As(III) in flouridated feedwater	35

 6      Effluent history from an activated alumina column with
        As(V) in flouridated feedwater	36

 7      Comparison of arsenite  and arsenate breakthrough curves
        from activated alumina columns	37

 8      pC-pH diagram for  arsenious acid and  its anion	39

 9      pC-pH diagram for arsenic  acid and its an ions	40

10      Comparison of flouride  adsorption by activated  alumina
        from columns fed As(III) or  As(V)	41

11      Arsenic  and flouride  effluent  histories from activated
        alumina column at San Ytidro, NM	43

12      Kinetics of reaction  in artificial ground water with
        DPD  quenching	44

13      Effect of pH on  extent  of  reaction In artificial
        groundwater	46

14      Comparison of effect  of chloride ion with different
        eounterion upon extent of As(IZI) oxidation	48

15      Comparison of kinetics  of  As(III) oxidation in  presence
        of different counterions	50
                                    ix

-------
16      Kinetics of chloranine oxidation	-  -51

17      Effect  of  pH upon Ae(III)  oxidation  by ch lor amines	53

18      Compcrlaon of kinetics of Ao(III)  oxidation in
        artificial ground water and aged  tap water	55

-------
TABLES
Number
1
2
3

4
5
6
7
8

9
10

11

12

13
14

Competition of feedvater to activate4 Alumina columns

Kinetics of AsdlZ} Oxidation in DJ water with no


Results of screening of sodium aulflte quenching . . .


Final concentration of DPD and phosphate used in


Composition of backgound vater used in oxidation

Composition of calcic background vater used in

Comparieon of adsorption capacities from laboratory



Page
. . .25
. . .27

. . .28
. . .29
. . .29
. . .30
. . .31

. . .31
. . .32

. . .32




. . .56
. . . .57
    xl

-------
                             ABBREVIATIONS AND SYMBOLS
BV        Bed volume; volume occupied by adsorbent  in a column, or an
          equal volume of solution passed through the column.

C         Concentration of solute in liquid.

CQ        Initial concentration.

DPD       N'N diethyl-p-phenylene diamlne  oxalate.  used to quench the
          oxidation reaction with chlorine.

EBCT      Empty bed contact time;  the time required for a volume of
          aolution equal to the bed  volume of adsorbent to pass through
          a column.

G         Cibbs free energy.

C°        Standard Cibbs free energy.

Cf°       Standard Gibbs free energy of formation.

GFAA      Graphite furnace atomic adsorption.

K         Equilibrium constant.

K         Acidity constant; equilibrium constant  for the  deprotonation
          reaction for an acid.

MCL       Maximum contaminant limit.

meq       Mil liequlvalent;  one-thousandth part of standard gram
          equivalent.

NPOC1     Non-purgeable chlorinated organlcs; the non-volatile fraction
          of the chlorinated organlcs formed in water upon chlorlnation.

pKa       Negative logarithm of acidity constant.

THM       TrihaLomethanes.

TOC       Total organic carbon; a measure of the  organics present In
          water.
                                       xii

-------
                             ACKNOWLEDGMENTS
     We are particularly grateful  to  the  following Individuals for their
important  contributions to  this study:

     Mr. Ton Sorg, the U. S. EPA project officer, for his continuing support,
guidance,  and cooperation during this and  related studies, and

     Dr. Arup Sengupta  for his interest, helpful suggestions and encouragement
during  the  conduct  of Che laboratory studies  and the preparation  of the
report.
                                     xili

-------
                                SECTION  1

                               INTRODUCTION'
     This study on arsenic oxidation and removal  from drinking voter was
undertaken  at  the  University of Houston as part  of  a comprehensive  research
effort on arsenic  removal  from drinking water.   Related field studies  using
Che Ul'/EPA Mobile Drinking Water Treatment Research Facility have been carried
out in San Ysldro, New Mexico and are continuing in Hanford, California.
Related laboratory studies have been done  on  establishing the As(V)  capacity
of activated  alumina,  regenerating arsenic-spent  alumina, establishing the
fundamentals  of As(V)  uptake by  ion-exchange  resins, and  developing an
analytical method for  the separation and determination of As(IIl) and  As(V).

     Arsenic is a ubiquitous, naturally occurring, semi-metallic element.  The
global  average concentration of arsenic in the earth's crust is  1.8 ppm.
Certain geologic formations are relatively rich in arsenic, whereas,  other
formations  contain little  or none  of the  element.   Arsenic  in  various
inorganic or organic forms may be present in the air,  soil,  or water.  Its
presence  may  arise  from  natural phenomena  such  as the  weathering or
dissolution of arsenic-bearing rocks or from human activities such as  the
manufacture and application  of  arsenical   pesticides,  the  combustion of
arsenic-bearing coal and oil, and the  smelting of various ores, commonly
copper- and  lead-bearing ores.

    Specifically,  arsenic may enter surface waters such  as rivers,  lakes,  and
estuaries  in   runoff  from agricultural fields  treated  with  arsenical
pesticides, In various industrial wastes, and  in runoff from natural rock
formations  that contain  arsenic.  Similarly, arsenic  may enter ground  water by
a variety of routes.  Inorganic arsenic  may enter the ground water naturally
from the dissolution of arsenic-bearing  geologic formations within a  given
aquifer.  Ground waters  may also be contaminated by  leachate from  improperly
landfilled fly ash from copper  and  lead smelters, wastes from arsenic refining
and pesticide manufacture, tailings from ore processing, and other similar
activities.  Concern has been  expressed about future  activities such as  in-
situ  coal   gasification,  which  may result In the  leaching of  various
uncharacterized organic arsenlcals along with  other retort  products  153).
However, most arsenic-bearing ground waters currently arise as a natural
phenomenon.

GENERAL CHEMISTRY

    Arsenic is stable In four  oxidation states  (+5, +3, 0, -3) at the  redox
potentials  (Eh  conditions) occurring in most aquatic  systems.  The pentavalent

-------
and trlvalent states  of  arsenic are those most  commonly encountered In
solution [21].  The pentavalent form appears in solution as an anion of the
triprotic  arsenic  acid,  and  the  trivalenf  form  appears as arsenious acid or
its anion.  The dissociation of arsenic 8               (2)

                  •=  H+ + AsOA3'     pKa = 11.5               (3)

The dissociation of arsenious  acid (trivalent arsenic) can be written as:

          As (OH) 3 =  AsO(OH)2" + H+    pKfl - 9.2              (4)

Thus in the pH range of 6 to 8 pentavalent arsenic occurs  as  either  a singly-
or doubly-charged  anion,  whereastr iva lent arsenic appears as the nonionlc
arsenious acid.  Since  these species are  interconvertable,  arsenate should
predominate in surface waters, and arsenite  should be the predominant species
in anaerobic ground waters. Ferguson and Gavis  [21] reported that  the rate of
oxidation of arsenlte by oyxgen is very slow  at natural pH values.

OCCURRENCE

    The natural  occurrence of arsenic in ground  water  is  worldwide.  Various
areas of the world must use arsenic-bearing  ground waters  as a  drinking water
supply.  Antofagasta,  Chile  [5], Nova Scotia  [25],  and  certain regions of
Taiwan are three well-known  examples.  Various areas of the United States,
most notably  the southwest and the northwest regions,  have a significant
number of arsenic-bearing groundwaters,  some of which are used as drinking
water supplies  [12].

    Arsenic concentrations found in groundwaters  vary.   Examples  from well-
known problem  areas  are  0.75  mg/L total As  In  Antofagasta,  Chile,  0.85 mg/L
and 1.1  mg/L total As  in Taiwan,  and 0.63 and 8.0 mg/L total  As  in  Nova Scotia
[33].   In  an  AWWA  survey of inorganic contaminants,  the total arsenic
concentration  ranged  from 0.052  to 0.19 mg/L among  reported  wells  that
exceeded the maximum contaminant  limit  (MCL) in use for potable water in the
United States [12].  The modal concentration occurred in the 0.075 to 0.100
mg/L concentration bracket. The  total  arsenic concentration found in surface
waters such as lakes and rivers  can often exceed the concentration found in
ground vater.  The highest values, often greater than 3 mg/L, are usually
associated  with  human  activity [17].  Naturally contaminated arsenic-bearing
ground waters usually have high pH and high bicarbonate  [17].  This phenomenon
presumably is  a  result of the chemical characteristics of the arsenic-bearing
rock in the aquifer.

    The valence state  of arsenic  in ground water  varies with location;
however, data  on the speciation of arsenic-bearing ground  water  is relatively
scarce.  A survey of  various  amenic containing welle around the  world

-------
Indicated that the arsenlte to arsenate ratio varied from 0.007 to 1.3  (33].
These values correspond  to  0.7 percent and 56.5  percent of  the total  arsenic
in  the  -4-3  oxidation state.   Drinking water  wells  sampled  in Hanford,
California,  in February 1984 had a  range of total arsenic concentration  of 30
to 90 ppb, all of which (> 95 percent) was trivalent arsenic |16j.   These well
waters also contained sulfides, indicating reducing conditions within the
aquifer.  Ground  water drawn from an infiltration gallery 4 m (13 ft)  deep in
San Ysldro,  New Mexico contained  80  ppb total arsenic, of which 40 percent was
trivalent  arsenic [16].  Well water from a 13-m (43-ft) depth  in this saire
area had a  much  higher total  arsenic concentration  (190 ppb),  which was
primarily trivalent arsenic.  This  water also contained other reduced species,
namely sulfide,  Fe(II), and Mn(II). From the above reports, it is apparent
that both As(III) and As(V) are important species in  arsenic-contaminated
ground water supplies.

    In a survey  of individual wells at homes in Nova Scotia,  Subramanian et
al. [64] reported that 36 out of 50 samples exceeded 50 ppb total arsenic.
Calculations made from their reported data show that 22  percent of the  samples
had no detectable As(III).  The  arsenic(III)  in  the remainder of the  samples
ranged from 0.8 to 91 percent.  Hire of these samples contained  As(IIl) in
concentrations exceeding 50 ppb.   In fact,  six of the samples had  As(III)
concentrations greater than 100 ppb, witi. fhree samples exceeding  300 ppb
As(IU).

TOXICITY

    The toxicity  of arsenic  is widely  known.  Arsenic poisoning  way  be  acute
or chronic.   Both cases have been fairly well studied — the former  because of
the popularity of arsenic as a poison, and the latter because arsenic  t.as  the
principle ingredient  in Fowler's solution used to treat akin  disorders  and  a
common ingredient in various general health tonics.  Ingestion  of  a  large
dosage of arsenic results in severe  capillary damage that may  ultimately lead
to circulatory failure and death.   The symptoms  of chronic arsenic poisoning
include  diarrhea, skin pigmentation, hyperkeratosis, circumscribed edema,
nausea and  loss  of appetite.  Arsenic exerts  its effect by reacting with
cellular sulfhydryl groups.  Thus the sulfhydryl enzymes (especially pyruvate
oxidase)  essential  to cellular metabolism are  Inhibited [31].

    The toxicity  of arsenic  depends  on its valence state.  Arsenite reportedly
is 60 times more  toxic than arsenate,  primarily because arsenite inactivates
the sulfhydryl enzymes.   Arsenate exhibits toxicity when it is  reduced to
arsenite by  the body  [31].    Plants and  animals  detoxify  arsenate  by
methylation [47]; and one bacterial  species is capable  of detoxifying arsenite
by «*:.ididation to arsenate.   Bacteria are known  to be capable of arsenate
efflux through the phosphate transport system  in  the presence of phosphate
[55, 73].   Enhanced  penetration of the  cellular  membrane by the non-ionic
trivalent arsenic may contribute to its greater  toxicity, but this effect  has
not been measured.  Increased  toxicity because of  penetration by non-ionized
HOC1 and NH3 are,  however, well known,  possibly similar examples.

    In addition to the toxic effects,  it has also been claimed that arsenic is

-------
a carcinogen.   However,  this claim is highly controversial.  Epidemiological
studies of increased lung cancers attributed to  arsenic exposure among smelter
workers  in Tacoma, Washington, have  been criticized on the grounds that
exposure to SO? and other agents were not excluded  [19].   The increased
incidence of skin cancers attributed to arsenic  ingestion in drinking water  in
Taiwan has also been questioned because  that water also contained ergometrlne
and other fluorescent  alkaloids  that cannot be discounted as carcinogens  [24,
56].  Other instances of arsenic ingestion by means of drinking water have
failed to produce an increased incidence of cancers  (56].   Furthermore,
experiments to induce cancer in laboratory animals have failed  to demonstrate
that arsenic acts as a carcinogen [24].  Those experiments where  cancers have
been Induced have such low animal survival rates  « 5 percent)  as to make the
results  meaningless  [24].  Therefore,  it has not been demonstrated that
arsenic is a carcinogen  [24, 44].

    In fact, some current research indicates that arsenic may be  an  essential
trace element.  Arsenic  in the form of organic  arsenicals has  been  used as a
feed  additive  for poultry and  swine to  promote  weight gain (2).   Such
observations have promoted research into the nutritional value of arsenic.
Recent research Indicates  that arsenic may  be  an essential trace element  in
chickens and  goats [44,  69] with a  relationship between zinc and arsenic
metabolism.  However, the metabolic studies seen to deal only with the +5
oxidation state.  Although the toxcity of arsenic  is well-known,  its exact
role and influence in human metabolism remains  unknown--  and fraught with
controversy.

    Despite  the lack of  knowledge  about the consequences  of the  ingestion of
very small amounts of arsenic, the ultimate toxicity of arsenic is well
recognized.   It is for  this  reason that the  concentration  of arsenic in
foodstuffs  and water is regulated.   The  USEPA National Interim Primary
Drinking  Water Regulations  set a Haximun  Contaminant  Limit  (MCL) for arsenic
at 0.05 mg/L.

ARSENIC REMOVAL

    Trace amounts of arsenic are usually successfully  removed by  conventional
water treatment methods [46].   For a number of surface water sources with
significant  levels of arsenic, arsenic occurrence  in  the finished  water was
relatively  infrequent  after conventional water treatment  [43].   Experimental
studies have  shown that arsenic may be removed by coagulation, filtration,
reverse osmosis, ion-exchange,  electrodialysis, and adsorption  onto  activated
alumina.   In most of these processes it is  pentavalent arsenic  that is removed
most effectively. Only the  removal  of  inorganic arsenic will  be  discussed
here because organic species such as monosodium methyl  arsenic (MSHA) do not
appear to be a  problem in ground water used for  potable water supply.

Removal by Coagulation

    In the environment,  arsenic  can  be  removed  from  solution  by
coprecipitation with metal ion precipitation, most notably the iron oxides
[21].   In treatment plants and experimental  studies,  arsenic has  been removed

-------
with varying degrees of success by coagulation and precipitation,  primarily
with aluminum and iron hydroxides.  This process is  usually  followed by
filtration.  Lime softening has also been studied for arsenic removal.  Of  the
various coagulants,  ferric chloride has proved  the most successful.   Arsenic
removal  by  coagulation  appears  to be  dependent upon  initial  arsenic
concentration, dosage of coagulant, pH, and  the  valence of the arsenic species
as will be discussed below.

Aluii--

    In early alum precipitation  experiments by  Buswe 11  [7], some removal of
high  initial arsenic  concentrations  (25  mg/L)  at  pH 6-7  was achieved.
Rosehart  and Lee  [51] found alum to be moderately successful in removing high
concentrations of  arsenic (132 mg/L) from gold mine wastes.   Addition  of alum
to these waters  in an  Al/As ratio of 4:1 at  pH 7-8 achieved a maximum 90
percent removal  of As(V) to a  final concentration of 13.2 mg/L.  The  removal
of lower concentrations of As(III) was successful.  Under the sane optimal
dosage and pH conditions, 500 ppb As(III)  was  reduced  95 percent to 25 ppb.
However,  Shen [57] achieved only a 32 percent  removal of 1 mg/L creenic with
20 mg/L alum at  pH 6.8.  Although the valence state was  unspecified, the
ground water presumably contained  As(III) as it  also contained  siilfides and
ammonia,  indicating reducing conditions.  This observation agrees with what is
known about  the  poor removal  of *.s(III) by activated alumina--dehydrogenated
A1(OH)3.

    Nillson [45]  also  investigated arsenic removal by alum.  At an  initial
concentration of 21 mg/L As(V)  at pH 6.5-7,  a 94 percent  removal  war  achieved
with a 10 mg/L alum dosage.  However,  these same conditions  failed  to remove
any of 23 mg/L As(III).  Under these same conditions of pH and alum dosage, 71
percent removal  of an initial concentration of 4.2 mg/L As(V) was achieved.
An examination of these Hata  indicates that As(V)  is much better removed by
alum than As(III).  And, although data is usually presented as percent  removal
which tends to obscure the point, as the initial  concentration of arsenic
increases, the arsenic remaining after coagulation and precipitation also
increases.

    The extensive research by  Logsdon  [42]  supports these  trends.  Alum
coagulation resulted  in a  5-15  percent removal of As(III) which was
considerably less  than  the  removal of As(V) Gulledge  and  O'Connor [27]
obtained  under the  same conditions.  Gulledge and O'Connor  found  that  greater
than 90 percent removal of arsenic(V)  could be acheived  with 30 mg/L dosage of
alum at pH 5-7.  Logsdon [60] also found that coagulation with 30 mg/L alum
could achieve the arsenic MCL with areenic(V) concentrations of 1.5 mg/L or
less, but the MCL was not  obtainable  when  the  arsenic(III)  concentration was
0.1  mg/L  or  higher unless As(IIl) was oxidized  to As(V)  prior  to coagulatf.cn
and precipitation.

    The optimal pH and  alum dosage for arsenic  removal  appear  to be specific
to each water.  Despite  this observation,  certain general trends are  followed.
Pentavalent arsenic is  more effectively removed than trivalent arsenic under
the same  optimal conditions.   This result is an Important consideration in

-------
drinking  water  treatment when a certain  HCL needs  to  be achieved.  Also, the
higher the initial concentration of arsenic,  the higher the concentration
remaining after alum coagulation and precipitation.

Iron—

    Numerous coagulation and precipitation experiments with some form of iron,
principally either ferric  or t'errous sulfate,  or  feme  chloride, have  been
performed — often in conjunction with alum precipitation experiments.  One
study by  Pierce  and Moore  [48] obtained  very high  arsenic removal rates  with
amorphous  Fe(OH>3.  Ferric hydroxide at 4.45 mg/L removed  119 mg/g at  0.53
Bg/L As(V) and 37.5 mg/g at 0.83 mg/L As(III).  Buswe 11  [7]  observed an 84
percent  removal  of 25  mg/L As(V) after  coagulation,  precipitation, and
filtration at pH  5-6 with 250 mg/L of "Ferrisul" ferric sulfate.   As(III),
only after oxidation to As(V) with calcium hypochlorite, was removed to 4  mg/L
under the same conditions.   Cherkinskli [10] reported  a  96 percent removal of
36.2 mg/L arsenious oxide with a ferrous  sulfate dosage of 750 ng/1.   Rosehart
and Lee  151] obtained 94  percent removal of  132 mg/L  As(V) at pH 8  with
ferrous sulfate at a Fe/As  ratio of 1.5,  but failed  to  remove any of 0.5  mg/L
As(IIl) under these same conditions.   Using ferric chloride  as the  coagulant
at a higher Fe/As ratio of 4.0, 90 percent removal of  132 mg/LAs(V) at pH  9
and 95 percent removal  of 0.5 mg/L As(III) at  pH 8 was obtained.  With  both
ferrous  sulfate and ferric chloride, removal of  As(III) was more difficult
than removal of As(V).  Gull edge and O'Connor 127]  demonstrated that  ferric
sulfate  doses of 30  mg/L at pH 5-8 would remove  0.05 mg/L As(V) almost
completely.   Slien tested  both ferrous sulfate  and  ferric chloride.  He
preferred the ferric chloride in his  treatment plant design because he
achieved greater removal (82 percent  versus 24  percent) of 1.0 mg/L jrsenic at
pH  6.7  and 20 mg/L coagulant.  Although the valence was not stated,  arsenite
is presumed to  be the  predominant species because  the water was anaerobic and
contained other reduced  species.  Shen also obtained 98.7 percent removal of
0.60 mg/L arsenic  after 20 mg/L chlorination —indicating  oxidation.   Sorg and
Logsdon  [60]  achieved  high  removal rates (81 percent  or better) of As(V) with
ferric sulfate at pH 6.7 to 8.  However, As(lII) was removed only at  very lev
initial  arsenic concentrations.

    Pierce and  Moore [48] concluded that the adsorption of arsenic onto ferric
hydroxide was not solely  because of electrostatic processes, but included
specific adsorption  or the formation of a  chemical bond.  Furthermore,  their
results showed that arsenate adsorption was 20 tines faster than arsenite
adsorption.  They concluded that for concentrations  found  in natural  waters
the optimum pU for arsenite removal  is  7 and for As(V) removal pH  4  is
optimal.  While other researchers,  as  discussed  above,  were able to remove
both states at  other pU values,  As(V) was always removed more effectively then
As(IIl).   This  result  is important when attempting to remove arsenic to meet
the MCL.

Lime softening-

    Lime softening removes  arsenic probably by co-precipitation of calcium
arsenate on  the hydrous magnesium oxide floe. Lime softening followed  by

-------
filtration removed arsenic  from water in Argentina [66].  Hard, turbid waters
were  Created with lime and recarbonated  to  adjust pH.   However,  further
details were not  given  in the reference.

    Cherkinekil  (10| studied arsenic removal by lime softening along with
coagulation by ferrous eulfate.  A dosage of 2250 mg/L of calcium oxide was
required  to remove 96  percent of 362 mg/L areenious oxide.  However, pH and
the alkalinity, important variables in softening, were not  given.

    Various researchers  have described arsenic  removal by lime softening from
mining vaste  streams.  Lagultton  [36| wrote that arsenic  valence,  pH,  calcium
activity  and phosphate  concentration  are important  variables for naximlzirig
arsenic removal.   Koeehart and Lee [51|  attained removal of 123 mg/L As(V) to
6.6 mg/L at pH 12 with  CaO.  The optimal  Ca/As ratio was 9.8.  Under these
same optimal conditions, a  95 percent  removal of 0.50 Bg/L As(III) to 0.025
mg/L was  attained.

    In his investigation on drinking water  treatment, Ntllson [45] was unable
to  remove by lime precipitation  any  of 23  mg/L As(IIl) at pH  9.5  after
pretreatment of  tap water vith 6 mg/L phosphate  as P.  In tap water with 21
mg/L As(V), treatment with  lime  at  pH 9.5 removed arsenic to 9.7 ng/L.  At an
arsenic concentration of 4.2 tng/L Ae(V) in "mechanically  treated" municipal
wastewater,  lice treatment at  pH  9.5 yielded a minimum of 0.1 ng/L for a
maximum removal of 98 percent.  The  nature of the rtechanical treatment was not
specified.  Shen  [57] obtained only a 20 percent removal of  1.0 mg/L arsenic
after a dosage of 20 mg/L lime.   The  pK of the raw water was 6.8, but the pH
attained was  not  given.  Also, as noted before,  the valence  of the arsenic was
not given  but was  presumably As(III).   Scrg and  Logsdon 160] conducted pilot
plant  lint softening  teats at  pK  9.5  and  11.3,  the latter tests Included
recarbonation and second stage settling.  Lime  softening at pH 9.5 resulted in
10 percent removal of  O.A8 ng/L As(IIl) and 49 percent renrval of 0.42 ng/L
As(V).   At pH 11.3, 63 percent removal of 0.34 mg/L As(III) was attained after
first-stage treatment and 69 percent overall removal after the  second-stage.
Ai. this pH, 98  percent removal of 0.58  mg/L As(V) was  attained after first-
stage treatment.  Second-stage settling  did not improve removal.   Arsenic(III)
removal was always  poorer than arsenic  (V) removal.  They concluded that the
are«nic HCL can be attained only if  0.35  ng/L or  less AsfV) IB  present
initial ly, and If As(III) concentrations are Leas than 0.1 mg/L, preoxidation
La not necessary  to achieve  the HCL when eoftenlng is performed  at  pH  10.7 or
above.

    Hone of the researchers have examined all the variables that  Laguitton
theorized  were  Important in maximazing arsenic removal.  However, the work to
date shows that  the removal of arsenic   to meet an HCL  depends significantly
upon the valence  state  and  the pH.  Higher pR appears to increase removal and
As(V) is more effectively removed than As(III).

Removal during Filtration

    Much of r-.he filtration data has been collected in conjunction vith arsenic
co-precipitation experiments.  Shen [57] studied filtration vith sand and

-------
anthracite.  At slow filtration rate (2-4 vr/m  day),  a  sand column 48 cm long
removed 90 to 95 percent of 1  mg/'L arsenic, vhile  an anthracite column 70 cm
long was needed to achieve 95  percent  removal.   Slow sand filtration yielded
better arsenic removals than rapid sand filtration (168 -  194 m3/m2/day);
however,  the  filter rune  proved  too short  to  be useful  For practical
application.   Sorg and Logsdon (60) also studied filtration after coagulation
and precipitation.  Increased removals were  obtained  by following coagulation
with dual media filtration.   Filtration with  granular  activated carbon (GAC)
yielded only slightly higher removals  than filtration with dual media.

    Filtration studies  like those above were conducted as polishing  steps.
The mechanism of action can be  attributed at  least  partially to adsorption.
Sand is known to absorb both As(III)  and  As(V)  differentially  (26).  Elution
profiles of 0.180 mg/L As(V) and As(IIl) from sand columns show that As(lll)
elutes much faster r.han  As(V) under oxidizing and neutral conditions,   and at
approximately the  same time under strongly reducing conditions.   This behavior
Is attributed  to differential adsorption.

 Removal by_ Reverse Osmosis

    Preliminary studies  by  Fox  {23] showed reverse osmosis to  be effective in
removing arsenite  and arsenate  from water.   Both the DuPont  (aramfc*) and the
Osmonlcs (cellulose  acetate) membranes  removed 0.750  mg/L below the detection
limit, probably 0.005  mg/L.  Pentavalent arsenic was removed to met the MCL
when the Initial  arsenic concentration  was 4 ng/L for the aramid membrane and
6 og/L for the cellulose acetate membrane,   A maximum removal  of 80 percent
was achieved for trivalent arsenic.   Studies by Fox  [22]  for the EPA in
Fairbanks,  Alaska  and Eugene,  Oregon  show  96  to 98.6 percent removal of 0.42
to 0.46  mg/L arsenic.  The speiiation  of arsenic  was  not  available.  Point of
use research directed by Sorg  [59] at  the Drinking Water Research Division
EPA, studied arsenic removal by a household reverse osmosis unit.  Pentavalent
arsenic was generally better removed than trivalent  arsenic. Removals of
arsenite from well water varied from 43 percent to 81  percent while removals
for arsenate  were 97 to 99 percent.   Chlorinatlon of  the  Ae(III) bearing
waters increased  removal in two out of three  cases.  The removal of Aa(V) by
reverse osmosis is more effective as expected  because As(V) exists as an anion
while  As(lII) is  non-ionic in the pH range of interest,  and as  such more
easily penetrates  the membrane.

Remova 1 by Electrodlalyeis

    Electrodialyals  (ED)  is a  process  that would be  expected  to remove As(V)
preferentially over As(IIl)  since the basis  of  removal depends upon the
existence of charged  species.  Since arsenate exists  as an anion for pH above
2.2,  while  Aa(III) exists predominantly in its non-ionized form below pH 9.2,
As(V) should be preferentially  removed over As(III)  below pH 9.  Reversible
electrodialysis (£DR) was studied for arsenic  removal in  New Mexico  using the
University of  Houston/EPA  Mobile  Drinking Water Treatment Research facility
(16).  Water from  San Ysidro, NM Well Mo.  4 containing 190 ppb total arsenic
which was predominantly  As(III) was treated by EDR.   Most of the arsenic (140
ppb) remained  In the product indicating poor removal of As(lII).   However,


                                       8

-------
more research is needed  in  this area.

Remova\ by Adsorption

    Adsorption by a variety of sorbents has  been widely  investigated because
of the inherent advantages of packed  beds.  Two of the most commonly used
adsorbents are activated carbon  and activated  alumina  although  a variety of
other adsorbents have  been  studied and utilized  in water treatment.

    Gupta and Chen  [28]  investigated the adsorption of arsenic  by activated
alumina,  activated bauxite, and  activated carbon.  These studies  indicated
that activated  carbon was  ineffective  in removing arsenic from water.   This
conclusion was generally supported by Sorg and Logsdon  (60].

    On the other hand, activated alumina can successfully remove arsenic from
water.   Bel lack [3]  removed arsenic from ground  water in Falion,  Nevada using
activated alumina (F-l grade Alcoe, 28 X 48 mesh) in  18.5 mm-ID glass columns
filled with 25  gm at an optimum flov rate of 2.5  to 3.0 gpm/sq.  ft.   Initial
arsenic concentration was 0.106  mg/L and the runs were  terminated when the
effluent arsenic concentration reached 0.01  mg/L.

    The previously mentioned studies by Gupta and Chen [28]  indicated that
arsenic was strongly absorbed by activated alumina.  At  0.300 mg/1 As(V) final
concentration,  the equilibrium adsorption capacity was  approximately constant
at 7.5 mg As(V)/g alumina for varicu, solution salinities  from fresh water
through 0.67 N NaCl  and  "Fe(OH)3-precipUated"  sea water.  However, at a low
equilibrium concentration for 0.50 mg/L As(V), the  capacity varied from  1.0 mg
Ae(V)/g alumina for sea water to 4.2 mg As(V)/'g alumina for fresh water.  No
explanation of this  observation was given.   He also observed that the rate of
adsorption decreased win Increasing  salinity, and that the As(V) capacity of
alumina was a strong function of pH.

    However, these experiments by  Gupta  and Chen indicated that the adsorption
of As(III) onto alumina was not pH dependent in the range 4 to 9. Between pH
4 and  the optimum  pH 9, a nearly constant 0.2 mg As(III)/g  alumina  was
absorbed  at equilibrium with fresh water.  This adsorption  capacity  is
substantially less than  that  for As(V).

    At  arsenic concentrations of about 4 mg/L As(V) and  1 mg/L As(III),  silica
at concentrations of  70 mg/L and higher provided significant  competition.
This result is what  would be expected since silica,  especially  the ionic form,
is a highly preferred  ion by alumina [13].

    Further investigation on arsenic adsorption  onto alumina was performed by
Rosenblum end Clifford [52].  They found that arsenic adsorption  onto alumina
was optimized at pH  6.0  in column tests with an  Influent  concentration of 1.0
og/L As(V)  in nrtifical groundwater.  Arsenic adsorption was found  to  be
significantly reduced  by anion  competition.   At a liquid-phase equilibrium
concentration of 1 mg/L As(V), adsorption of arsenic was reduced by over 50
percent in the presence of  15 meq/L SO^2" und  20 percent In tl.e presence of
15 meq/L Cl" compared to deionlzed water.   An As(V) adsorption capacity of

-------
16.1  mg As(V)/g alumina at 1.0 mg As(V)/L was obtained from minlcolumn tests.
Column tests  on  As(III)  adsorption resulted in much higher than expected
maximun adsorption capacities.  Upon examination  of the  arsenic effluent
histories  they concluded that oxidation of As(IIl)  had occured during  the run
probably  as a result of mlcroblal action.   Further studies  on As(IlI)
adsorption need to  be  conducted.

Removal by_ Ion Exchange

    Calmon  (9] investigated the use of cation and anion exchangers for the
removal of 68 mg/L arsenic  from "field waters".  Not suprisingly,  cation
exchangers  were  ineffective; however,  removals of 55  to 82 percent were
obtained with various anion exchange resins  in the chloride form.  Removals of
99 and 100 percent were obtained  in column tests.  Other pertinent  information
was not given  in  the reference cited.

    Shen [57]  conducted a column test with the weak-base resin lonac A-260.
After 6.6  bed volumes  of  synthetic water containing 1 mg/L As(III), only 21
percent removal was occuring.  However, well water treated in the  same manner
resulted  in the  complete removal  of arsenic.   These results are hard to
definitively explain without  feedwater pH data, arsenic speciation  in the well
water,  and the concentrations  of competing anions such as sulfate in both test
waters.  However,  at acidic pH which weak-base resins require  in order to act
as ion  exchangers, As(III)  is non-ionic  and would not be  expected to be
significantly removed.  The well water which presumably contained at least
some As(III) because of the presence of other reduced species may have been
oxidised upon  standing prior  to ion exchange treatment.

    Lee and Rosehart [72] investigated several  ion-exchange resins.  They
obtained removal  of arsenic with various resinei  however,  lack of  information
on the arsenic speciation and the presence  of other ions severely   limits the
usefulness of  their results.

    Horng  and Clifford [32] studied the removal of arsenate by strong-base ion
exchange.   A study  of  five anion-exchange  resins (Amberlite IRA-900, lonac
ASB-1,  lonac ASB-2,  Dowex 11, Amberlite  IRA-458) resulted  in  the  conclusion
that Amberlite IRA-458,  an  acrylic-amine,  microporous resin  with  quaternary
fuctionaiity,  was  the most arsenate selective resin tested.   All of the
strong-base anion  exchange  resins tested at  arsenate  and  carbonate
concentrations  likely  to  be  encountered  in ground water  gave the following
anion selectivity sequence:

               S042' > HAs042", C032' > Cl" > H2As04~, HCOj"

However, because arsenate is  less preferred than sulfate, chrcmatographic
elution of arsenic  occurred.    This chromatographic peak phenomena  will occur
for all species less preferred than  the most preferred species present.  This
effect is  one very serious drawback to the use  of  strong-base ion-exchange for
the removal of arsenate.

    Studies on strong-base ion-exchange resins  indicate that As(IIl) is not


                                     10

-------
removed at neutral and acidic pH.  In fact, Amber lite IRA-458 was used to
separate As(III) from As(V) by Clifford et  al.(15].  Weak-base ion exchange
resins, which must operate at acidic pH, also do not remove As(lll).   It is
not suprisin^  that ion exchange  is  ineffective for the removal of trivalent
arsenic as it  exists predominantly as the non-ionized arsenious acid (pK^ •=
9.22) at the pH values exchangers are operated.  For this reason, while ion
exchange can remove pentavalent arsenic in the pH range of interest (6-9),
trivalent arsenic is removed ineffectively  if it is removed at all.

SUMMARY

     In the treatment processes discussed, pentavalent  arsenic  is, without
exception,  more  effectively removed  from water than  trivalent arsenic.  Thus
oxidative pretreatment of trivalent  arsenic will improve the arsenic removal
from waters containing As(III).
                                     11

-------
                                 SECTION 2

                                CONCLUSIONS


     Arsenic(V) is much more effectively removed by activated alumina than
trivalent arsenic.   An Alcoa F-l  activated alumina co lupin operated at pH 6.0
rcitoved 18 g As(III)/ m3 alumina, whereas 1610 g As(V)/m3 alumina was removed
when Che effluent  concentration of arsenic  reached 0.05  mg/L.

     The oxidation of ppb levels of arsenlc(lll) by ppm  levels of chlorine is
very rapid  in aqueous solution.  The oxidation of  100 ppb of As(IlI) by 1.0
mg/L free chlorine in artificial  ground water reached completion  (95 percent)
within 5 seconds.  Thus for water treatment purposes,  contact times do not
need to be  particularly long.   The extent of oxidation of  100 ppb As(III) by
1.0 mg/L  free chlorine in artificial  ground water was insensitive to pH after
30 seconds of reaction time in the range of 6.5 to 9.5.   The oxidation of 100
ppb As(IIl) by 1.0 mg/L free chlorine reached 90 percent completion in the
presence  of 0.01  M chloride.   The counterion (sodium or calcium)  had no
significant effect  on either the kinetics or extent of oxidation.  For most
groundwaters,  oxioative  pretreatment  with  free  chlorine may  be  effective in
enhancing arsenic  removal.

     In artificial  ground water, 100 ppb As(III) was  oxidized  by 1.0 mg/L
nonociiloramlne in  the presence of excess  ammonia.  Only  a third of the  As(III)
was oxidized,  however.  This reaction  was also insensitive to pH  in the range
6.5 to 10.5.  Although oxidation by  monochloramine can occur,  the As(III)
remaining may  exceed the MCL.

     Arsenlc(III) oxidation by  1.0 mg/L chlorine in aged tap water (TOC - 5
mg/L)  was considerably slowed.   The reaction did not  reach 95  percent
completion until 60 minutes.   The extent of reaction was the  same as that for
oxidation in artificial ground water.   Presumably, oxidation was slowed by the
chlorine demand of the  TOC present  in the water.  Chlorine contact times that
are sufficient will  be  Important  for efficient water  treatment.

     Although  very slow,  oxygen  is capable of arsenic oxidation.   Sparging for
60  minutes with  oxygen  oxidized only 3 percent  of  100  ppb  As(III) in
artificial ground water,  but 14 percent of  100 ppb As(III) in DI water was
oxidized.   However,  samples of As(III) in DI water and artificial  ground water
were completely oxidized after 61 days of  sitting in bottles with air in the
head space.
                                      12

-------
                                SECTION 3

                             RECOMMENDATIONS
LAB STUDIES

     Further studies will prove useful  for producing optimum arsenlc(III)
oxidation  by  chloramine for subsequent  arsenic  removal.   The  limitations  of
this treatment need to be examined experimentally.  The extent of As(III)
oxidation by nonochloranine may vary with  the monochloramine concentration and
the amount of excess ammonia present.  Experiments under the conditions used in
this work with different monochloramine  dosages (i.e.  0.5,  2.0, 5.0,  10.0
mg/L) and  different  amounts of excess ammonia  should help determine whether
free chlorine or combined chlorine is the  actual oxidant.   Also, further study
of longer  reaction times may  indicate which  is  the  true oxidant.

PILOT STUDIES

     Pilot studies  of arsenic(III) oxidation  by chlorination of actual
arsenic-contaminated  waters may be beneficial.   The  presence of  reduced
species  such  as eu If ides,  Fe(II),  and Mn(Il) may adversely affect the ease of
arsenic(III) oxidation.  Furthermore, species such  as copper are  known to
catalyze the destruction of HOC1, which may adversely affect the extent of
reaction at  a given chlorire dosage.   Although  laboratory studies  look
promising, it may be difficult to obtain good oxidation  in the field at the
low chlorine dosages such as are required  for electrodlalysis  and  reverse
osmosis membranes.

     Further investigation of whether mechanical aeration  or  ponding can
provide adequate oxidation for subsequent  arsenic removal  may a?.so be useful.
The  advantage  of these methods is that they are  relatively low  cost and
require little maintenance.
                                     13

-------
                                SECTION 4

                                  THEORY
AQUEOUS ARSENIC CHEMISTRY


    Arsenic occurs in the +3, +5, 0, and -3 oxidation states; however,  under
the pH and Eh  (redox  potential)  conditions found in natural  waters,  only the
+3 and +5 states are  stable {21}.  In aqueous systems, pentavalent arsenic
occurs as  arsenic acid or  its anions.  Trivalent  arsenic  may  occur  as
arsenious acid,  its anion,  or arsenious anhydride.   The  latter species  is  an
anphoteric  soluble  solid.   Figure  1  shows  the  influence of  pH on the
solubility of As2O^ in aqueous  solutions.   However,  as shown is Figure  1,  at
the concentrations generally encountered in natural  ground waters,  arsenious
anhydride dissolves completely  with the formation of primarily undissociated
arsenious acid—^AsOj.  Thus, at the concentrations of interest, trivalent
Inorganic aqueous arsenic occurs as arsenious  acid or its anion.

    Arsenic acid, H^AsO^, is a triprotic acid with pKj • 2.2, pK2  •  6.98, and
pK-j • 11.5, and appears in solution as either the acid or the corresponding
anion.   Monoprotic  arsenious  acid (pK B 9.2)  may be  written  as  HAs02«
However, strong evidence suggests  that HAs02 and its anion do not actually
appear in  solution  [58].   Arsenious acid appears to be As(OH)3  with the
following three possible anionic  species  depending  upon the basicity of the
solution:  AsO(OH)2~,  As02(OH)  , and AsO^3".  However, the only disocciation
that occurs  in  the range of  interest is:

         As(OH)3  = H+  +  AsO(OH)2~    pK = 9.2               (1)

Arsenic  is not known to form ligands with water; the species As(H2O)  n*  is
probably not formed.  Although As(III)-Cl bonds are known to occur,  these
bonds  only occur  at high UC1  concentrations  [58].  Disproportionation
reactions are not known to occur with either  of the acids.  Thus  the arsenic
oxyions are the principle species of environmental interest.

    Figure 2 shows  the regions  of thermodynamic stability of the relavent
trivalent and pentavalent  arsenic species at various  pH and Eh values. The
equations used  to construct  the  diagram after the method  of  Stumm and  Morgan
(63] are  presented In Appendix 1.  Aa Figure 2 shows, three  different arsenic
oxidation half reactions  are possible in the  pH  range 4  to  )1.   Thus the
oxidation potential of  arsenic half reactions is a function of pH.  Figure 2
also  shows that under aerobic  conditions,  puntavalent arsenic is the
thermodynamics 1ly stable species.   Therefore, oxygen is capable,  at  least
theoretically,  of oxidizing As(lII)  to As(V).

                                     14

-------
I CM
o
V)
 CM
O
(0

I
v»x
+
^*

O
w
 O)
 O
 1

 0

-1

•2

•3

•4

-5

-6
     -7
         V
                   As2O3(Solid)
AsO
                           Saturation
                              Point
                           HAsO,
     I
     I
     |               As2O3  Solution^

     |  CT = 100 ppb   ln P"re Water


     I
     I  .        i	i
                                                           AsO.
       -2      0       2       4       6       8      10

        Figure 1. The Influence of pH on the solubility of arsenlous anhydride.
                                                             12

-------
fc. — T
20
16
12
pE 8

4
0
-4
-8
*x»
• H3AsO4
\
•
AsO+
^



\|

XS"^^

H,A»0.-
H*,0\



«




\


HAsOj"
"^•^ HO Nv
* ^
^^^,

V
w

•
-
'v. •
"^
•
AsO3'

x^

1000
900
800
500

200
100
0


    -20    2     4     6    8    10   12    14
                          PH
Figure 2.  pE - pH diagram for arsenic-water system at 25°C.
                       16

-------
 OXIDATION  STUDIES

     Pourbaix [49] reported that the oxidation of arsenlc(III) occured only
 under conditions of considerable  overpotential.  Oxidizing agents listed wore
 the halogens or their oxygen compounds, chromic acid, nitric acid, hydrogen
 peroxide, or permanganate.  Johnson and  Bruckenstein [35] studied As(Ill)
 oxidation  by  Ig  at pH 8.2  to 9.2  and discovered the rate expression  to  be:

              -d AB(III) -  klI3"][H2As03']  + kjjlj]   (H2As03"]             <2>
                 dt
 where [As(III)] ° [HjAsOj]  + [H2As03~}.

 Roebuck [58] reported the rate of oxidation in acidic media to be
              d As(III) -  kf[An(III)Hl-^.]
                  dt          [I]2 [H+]
Liebhafsky has proposed a me chant em Co  explain this behavior.  Studies on
Asflll) oxidation by bromine  showed  the reaction  to be zero order in bromine
158].   Other studies  on oxidation of  As(III) In strong  HC1 solutions
demonstrated  that the kinetics shoved a complicated dependence upon HC1  1 58].

    Chlorine  has  been used as an oxidant  by Shen  [57], Sorg and Logs don [60],
and other  researchers in drinking water treatment; however,  studies of
kinetics or  the effects of pH and ions in solution have not been reported.
And although As(V)  is the thermodynamical ly stable  species under aerobic
conditions,  the  oxidation of As(III) by oxygen  is reported to  occur very
slowly at neutral pH [21].

AQUEOUS CHLORINE CHEMISTRY

    Chlorine  is a common disinfectant and oxidizing agent for w&ter treatment.
Consequently, the aqueous chemistry  of chlorine has been extensively studied
and reviewed.  The various clilorine  speciec can  participate in the following
types  of reactions:   hydrolysis,  redox,  chloramination,  substitution,
addition,  atom exchange, radical  oxidation, photochemical  decay,  metal
catalyst decomposition, and  self decomposition.   All  of these  reactions can
occur under  natural water conditions.   However, only the environmentally
Significant reactions will  be considered for simplicity.

    Figure 3 shows the  regions of thermodynamlc stability of  the most
Significant chlorine species.  The equations  used  to construct the figure are
given  in /VP^udix 2.  As Figure 3 shows, chloride is the  stable species in
vater under all conditions.  A comparison of Figures  2 and 3 indicates that
considerable  overpotential exists for tne oxidation of As(III).  At a typical


                                      17

-------
   30
   20
   10
                        HOC)
       CI2(aq)
H20
                                        ocr
                                   cr
          H2O
-20
                                  8      10     12
                    24       6



                            PH




Figure 3.  The pE - pH diagram for chlorine in water, 25°C.
                        18

-------
ground water pH of 8.0,  the border of  thermodynanic stability for Aa(III)
occurs at Eh • -100 volts while the border of thermodynamlc  stability for the
chlorine species OC1" occurs at Eh  « 1200 volts.

Free Chlorine

    Chlorine gas dissolves instaneously in water by rapid and reversible
hydrolysis.

          H20 + C12 -  HOC1  + H+ + Cl"                          (4)

Increasing  H  or Cl" can cause Cl2  to  predominate £t equilibrium thereby
decreasing  hypochlorons acid (HOC!) formation.  This behavior is found to
occur in sea water  [29].

    Hypochlorous acid is a weak acid which dissociates as follows!


         HOC1 =   H+ + OC1"    pK - 7.5                       (5)

     At pH < 7.5, HOC1 is the dominant species while at pH  > 7.5 OC1" ir. the
dominant  species.  Each  of these species has a different  redox potential.
Free chlorine species (HOC1, OC1~,  C12,  H2OC1+, Cl+) are capable of oxidizing
I" and Br~  to HOI  and  HOBr resulting  in the formation of bromlnatsd and
lodinated trihalomethanes (THM) in chlorinated  natural waters.  Since chlorine
species  are  so reactive,  other constituents  of natural water such as ammonia
and organics effect the performance of  chlorine as an oxidizing agent for a
target species.

Combined  Chlorine

    The  tern "combined chlorine" refers  to the halogenated nitrogen compounds
such as moTiochlorcamine,  dichloroamine, trichloroamine, and organic haloa^ine.
The formation of chloroamines from ammonia occurs  in a stepwise manner:

          HOC1 + NH3  «=  NH,C1 + K20      fast               (6)

          HOC1 + NH2C1  - NUCL2 + H20    slow               (7)

          HOC1 + KHC12   -  NC13 + H20    slow               (6)

     The  chloramines obtained depend upon  pH, reaction time, the relative
concentration of HOCl and NH^  , and temperature  [18].  At pH greater than S
and a molar  ratio of HOCl to NH^-N  of  Itl or  less, monochlorcaine Is observed.
Usually  nitrogen trichloride is  the  only species detected at  pH  less  than
three [18, 29].  Under certain conditions,  the formation of chloramines is
reversible;  however, the  equilibriun state probably is rarely attained because
the reverse  reaction is so slow in comparison to the forward reaction.   The
mono- and dichloramines retain some disinfecting power, although much  less
than an equivalent  concentration of free chlorine.  Free  chlorine reacts  with
organic amines to form the  organic haloamineit RNHC1, RNC12> R2NC1, but not


                                    19

-------
much Is known about  the mechanisms of these reactions.

Chlorine Oxidant Decay

    Hall  [29] describes  chlorine oxldant decay as any reaction involving
conversion of chlorine  Into constituents that are not delectable by the
analytical methods for  free chlorine.   Decay reactions are summarized in
Appendix 3; however,  only those  reactions with significance  to  this  research
will be discussed.

    The formation of total organic halogen compounds (TOX) that can be further
hydrolyzed to yield non-haloform and THM  is perhaps the  most  important
chlorine decay reaction   [43].   Fleischaker  and Randtke  [20]  report  that the
substitution and addition reactions to  form non-purgable chlorinated  organlcs
(MPOC1) are faster  than the oxidation of organic matter  by  chlorine and
perhaps faster than  the hydrolysis of chlorinated intermediate  byproducts Co
form chloromethanes.

    Another important set of reactions are those associated with the oxidation
of ammonia-nitrogen and the  phenomenon of breakpoint for residual  chlorine.
When the molar ratio of chlorine to nitrogen exceeds 111, the unstable species
NHClj is formed which subsequently decays  to N2 and  NO," resulting in a loss
of ammonia-nitrogen.   However,  if the molar  ratio exceeds 1.5, NC13 is  formed
resulting in a preservation of ammonia-nitrogen (18, 29].  These  reactions
occur during chloramination prior to  the breakpoint wnere free chlorine begins
to exist.

    Other reactions of lesser importance in oxidant decay  in water treatment
are the oxidation of inorganic  ions and the eelf-decomposition reactions.
Lister  [39]  reported that Fe(Il) and Mn(II) were  oxidized by OC1" with no
catalytic activity.  Other inorganic ions are susceptible  to attack; these
ions include nitrite, sulfite, selenite and arsenite.  Nickel, cobalt, and
copper ions catalyze the decomposition of OC1" to 0, and Cl~. Mechanisms of
Inorganic  ion oxidation will be  discussed  later.    Under certain conditions
the self-decomposition of OC1~ and HOC1 to form  02, CIO.,  CIO*'",  CIO," and Cl"
occurs  [29].

REACTION THERMODYNAMICS

    A study of thermodynamics shows whether a reaction is possible and how
potentially energetic it  is.   Since the pK values of hypochlorous acid,
arsenious  acid, and  arsenic acid are known,  four reactions  are possible
depending upon the pH.  The following  half  reaction potentials are  from the
ttiennodynamic daca of La timer [37], Si lien [54], and Pourbalx  [49].

                         HOC1 + H+ + 2 e"  " el" + H20                    (9)

                                       E • 1.49 volts

                       OC1" + H20 + 2 e"  « Cl" + 2 OH"                 (10)

                                       E  -  0.90 volts

                                    20

-------
                            HAs02 + 2 H20 « H2As04~ + 3 H+ + 2  e~         (11)

                                        E = 0.666 volts

                          HAs02 + 2 H20 = HAsOA2' + 4 H+ + 2 e"        (12)

                                       E - 0.881 volts

                            As02' + 2 H20 - HAs042" + 3 H+ + 2  c"         (13)

                                        E = 0.609 volts

The four reactions  based  upon  the possible speciation combinations  of arsenic
and chlorine are given below.

               a.  2.2 <  pH < 6.98

                   HAs02  + HOC1  + H20 - H2As04" + 2 H+ + Cl"              (14)

                                   E - 2.156 volts

                                 AC- -416.2 KJ  (-99.43 kcaJ)

                                   K - 7.869 X 1072

               b.  6.98 < pH < 7.5

                   HAs02  + HOC1  + H20 = HAsOA2" + Cl" + 3 H+              (15)

                                   E o 2.371 volts

                                 AC - -457.72 kJ  (-109.35 kcal)

                                   K " 1.474 X 1080

               c.  7.5 <  pH < 9.2

                   HAs02  + OC1"  + H20 - HAs042" + Cl" + 2 H+              (16)

                                   E = 1.781 volts

                                       -343.82 kJ  (-82.14 kcal)

                                   R - 1.659 X 1060
                                      21

-------
               d.   9.2 < pH < 11.5

                   As02~ + OC1" + H20 - HAs042~ + Cl" + H+               (17)

                                   E » 1.509 voles

                                 AC - -291.34 kJ (-69.60 kcal)

                                   K - 1.063 X 1051

 These calculations show that the oxidation of arsenic  is potentially verv
 energetic at all  pH values from 2.2 to 11.5 and  that these  reactions posess
 very large equilibrium constants, K.   At concentrations normally encountered,
 the reaction is even more favorable. For a chlorine concentration of 1.0
 mg/L,  chloride  concentration of 73 mg/L,  an Ae(III) concentration of  100 ppb,
 an As(V) concentiation  of 0.1 ppb, and at pH equal  to 8.0,  the  potential is
 1.72  volts.

    A itethod  to see the theoretical sensitivity  of  the oxidation reaction by
 chlorine  Is  to examine the reaction potentials  instead  of  the  standard
 reaction potentials.   The following assumptions  were made in order  to solve
 the equationsi  [As(JII)] - [As(V>]  , 1.0 mg/L chlorine  as  either HOC1 of OC1"
 ,  and no attempt  to account for the differences  in corcentration of the
 species as the pH approached  the pK  values.   Figure 4  shows  the  reaction
 potentials and  G  values as a function of pll for  different Cl" concentrations.
 As the  pH  increases,  the reaction potentials and  consequently  the energies of
 reaction decrease.

 REACTION MECHANISMS

    Redox  reactions are electron transfer reactions,  but  these  electron
 transfers  are not  accomplished in the same manner for all  reactions.  Taube
 [65]  suggests classifying  redox reactions into the following categories:
 inner-sphere activated complex,  outer-sphere activated complex, and a bridge
 activated  complex.  For the  inner-sphere  mechanism,  electron  transfer occurs
 within  a  single primary  bond system.  While for the outer-sphere mechanism,
 electron transfer occurs from one primary bond system to another.  In this
 case,  both the oxidizing and  the reducing  agents  are  inert to substitution and
 the electron transfer takes  place between ions of similar geometry.  For the
 bridged type reaction,  at  least one reactant undergoes  rapid  substitution
 thereby forming a ligand which bridges the reacting species.  The bridging
 ligand may facilitate electron transfer in two ways termed "chemical" and
 "resonance".   The "chemical"  mechanism is a process in which either the
 oxidizing  ion is strong enough to oxidize  o- the  reducing ion id  strong enough
 to reduce  the  shared ligand  whereby the electron deficit  or  excess  is passed
 to the  reducing or oxidizing  ion, respectively.   In  other  words, the electron
 hops  from  the  reducing center to a specific bound state  in the  ligand to the
 oxidizing center where it remains.  In  "resonance" transfer the electron never
 occupies a well bound state on the ligand. For both types of transfer, the
 rate of electron transfer is sensitive  to the  nature of the ligand.  This
.ligand determines the height ot the  energy barrier that the electron must


                                     22

-------
 X
 (E
(U
                                Assumptions

                                [AS] = [AS(V>]

                                CI2 • 1 mg/L
         0.01 M Cl
-100
                                                    -75
-50
                                                    -25
                                                           o
      0      24     6     8     10    12


                          PH


 Figure 4.  The calculated reaction potential  of the oxidation of
          As(III) by 1.0 mg/L  chlorine dosaae as a function of
          pH.
                          23

-------
penetrate;  thereby determining the  probabllty of transfer.

    HOC1 nay act as an clectrophillc teagent with either oxygen or chlorine as
r.he  center of  reaction.   J.  E.   Draley  [18]  suggests  that HOC1  is  an
electrophlle in which the chlorine atom partially assumes the characteristics
of Cl+ and combines with an electron from solution.  Either slmultfiieoucly or
subsequently,  hydroxyl ion  is split off.  The transfer of  Cl+ Is facilitated
by a negative charge, the basic part of the molecule.  The same type of attack
is thought to  occur at amlno-nitrogen.

    On the other hand,  the Cl  atom can have such a greater attraction for
electrons that  it may be displaced directly as chloride ion.   This type of
attack seems  to occur with  inorganic ions.   Anbar and  Taube [1]  proposed such
a mechanism for  the oxidation of nitrite by aqueous chlorine.   Work by Lister
[40] supported  this  mechanism.  Halperin and Taube  [30] proposed a similar
mechanism for the oxidation of sulfite.  The oxidation of Fe(II) and Mn(II)
and the reaction with hydrogen peroxide are thought to have similar transfers
with the displacement of either Cl"  or OH" as  initial steps.  However, the
exact nature of  these initial steps has not  been elucidated.

    The solvent can play a  key role in  a redox reaction.  The hydration of an
ion may  significantly  affect  its suceptibility to redox reactions.   The
hydratlon of an ion may either stabilize or destabilize  the intermediates.
Stabilizing the intermediate lowers the energy barrier thereby facilitating
the reaction.

OXYGEN OXIDATION

    Theoretically oxygen should be capable of oxidizing trlvalent arsenic (see
Figure  2).  However, this  reaction  does not appear to  occur except  very
slowly.   Latimer  [37]  proposed that  direct oxidation by 02 proceeds through
the formation of the intermediate, hydrogen peroxide. This reaction has a
very low potential.   Thus, this step becomes the rate determining step for the
direct oxidation by oxygen via this mechanism.  Furthermore,  pH  would also be
an important variable as it determines  arsenic speci-.tion and, therefore, the
half reaction potentials.

-------
                                SECTION 5

                               EXPERIMENTS
PRELIMINARY COLUMN STUDIES

    The objective  of Che preliminary column studies waff  to verify  the
reportedly efficient  removal  of  pentavalent  arsenic  and  the  non-removal  of
trivalent arsenic  on activated alumina.  Since some ground waters  in  the
United States contain both excess arsenic and  fluoride, and activated alumina
is known as a successful adsorbent for fluoride, these column  studies were
conducted  on water that contained both fluoride  and arsenic.

Column Experiment Design

    Two column runs  were conducted  simultaneously.   Each  column was
constructed from 1/4" ID plexiglass tubing with stainless steel Swagelocli
fittings  at each end.  Each column was loaded with 5 mL of conditioned Alcoa
F-l activated alumina (U. S. standard mesh size 24  X 48).  The conditioning
procedure  for activated alumina is given in Appendix 4.  Mini-pumps from
Milton Roy Model # NS1-33R were used to control  the flowiate at 1.7 mL/min  for
an EBCT of 2.9  minutes.   These  pumps provided a steady flow rate over  the
maximum length of the entire column run-- 58 days.

    The column feedwater was  a synthetic  ground water designed to resemble
ground water that had been acidified to  pH  6.0 with sulfuric acid.  This
pretreatment creates a sulfate enriched,  bicarbonate-free water.   An arsenic
concentration of 0.100 mg/L was chosen  because it is twice the MCL and not
uncommon  in arsenic  contaminated ground waters (e.g.  Hanford, CA and San
Ysidro, NM).  The name line of reasoning applies to the choice  of 3 og/L as
the fluoride concentration.  The composition of  the column feedwater is given
in Table 1.

     	          TABLE 1.  COMPOSITION OF COLUMN  FEEDWATER
        SO.
mg/L

 71

384

 3.0
Cations   mg/L     meq/L

         210       9.16

  Ca'T     20       1.0
                                            fl+

                                            .2+
                                     25

-------
The Influent  to  each column differed only in the valence state of  the arsenic;
one column received 0.100 ng/L As(V),  while the other received 0.100  mg/L
AB(IH).   The feedwater was made from reagent grade sodium fluoride,  sodium
chloride,  calcium chloride,  and sodium,  aulfate.  The pH was  adjusted  to  pH 6
with eulfurlc acid.  For the  As(V), column the feed  water was made in 40- liter
batches.   For the  As(lII) coHmn, the feedwater with the exception of arsenic
addition  was also made in 'iO-liter  batches.  In order to avoid arsenic
oxidation  during the long run,  freshly prepared trivalent  arsenic was added to
a 4-liter aliquot each morning for  the  day's  feedstock.

    An ISCO  Model 0  1850 fraction  collector was used  to collect  25.5-mL
samples every 15  minutes.  Various 25.5-mL samples were analyzed for arsenic
content with  a Perkin-Elmer Model 5000  Atomic  Adsorption Spectrophotometer
with a graphite furnace, Zemnan background correction, and a HGA-400 Furnace
Temperature  Programmer.  An electrotieless discharge arsenic  lamp was  used as
the source lamp.  Details of this analysis are given in Appendix 5.  Flouride
analyses were conducted using an Orion,  fluoride Ion-selective electrode Model
# 94-09-00 with  an Orion 50i digital pH/ion analyzer.

OXIDATION STUDIES

Objectives

    The oxidation studies were undertaken with the fol lowing objectives in
nindi

     (1)  to  examine the  kinetics of arsenic(IIl) oxidation by chlorine.

     (2)  to  examine  the effect of pH  on the  oxidation  of  arsenic(III) by
chloiine.

     (3)  to  examine the effect of anions and cations upon  the  oxidation of
arsenic(lll)  by  chlorine.

     (4)  to  examine  the effect of chloramlnes upon the oxidation of
arsenic(III).

     (5) to  examine the effect of TOC upon the oxidation of arsenic(III) by
chlorine.

     (6)   to Investigate the oxidation  of arsenic(III)  by  aerobic conditions.

Reagents

    All stock arsenic(V) solutions were prepared according to Standard  Methods
 |62) using  reagent grade KKjAsO^.  All stock arsenic(III)  solutions were
prepared  according tc  Standard Methods  (62] using reagent grade As203.   All
arsenic(III) solutions wore prepared immediately  preceding their use.   All
other  solutions with the exception of chlorine oxldant were prepared using the
appropriate  reagent grade chemicals.  Stock chlorine oxldant was  prepared froa


                                     26

-------
a commercial bleach (Chlorox)  containing sodium hypochlorite.   The  titer of
this solution was checked  immediately preceding use.  Analysis showed no
arsenic  to  be present  in  tl-e bleach.   The stock  chlorine solution  was
standardized according to Standard Methods [62].

Procedure Development for Chlorine Oxidation

    Ultimately  a quenching agent  was found to be  necessary because  the
oxidation reaction could not be monitored on a continuous basis. No method
was available for  instantaneous arsenic  speciation and  quantification.
Methods  available for analytical  arsenic(IIl) separation from arsenic(V)
include  ion exchange,  pH  controlled arsine  generation,  and selective
extraction.  Each of these methods requires subsequent analysis for total
arsenic  in  various  fractions.  The method chosen for  arsenic speciation,  ion
exchange separation,  does  not provide instantaneous separation.  The method
chosen for speciation is  based  upon the work  of Clifford et al.  (15].   Five mL
of pretreated, chloride-form IRA-458 (finer than  35  mesh) in a mini-column at
a flow rate of 10 mL/min was used to sepcrate Ac(IIl) from As(V) in a 100 mL
sample.  Under these conditions, the  fastest possible speciation  time was
greater  than  10 minutes.   Verification experiments with known  arsenic
standards showed good results as seen in Table 2.  Arsenic was  analyzed by
GFAA as  described earlier.

     	  TABLE 2.  VERIFICATION OF ARSENIC SEPARATION
         Sample                           Z Expected Recovery
                                          As(III)     As(V)

     100 ppb As(III)                         97         — .

     50 ppb As(III)/50 ppb As(V)             102         98

     100 ppb As(V)                           —         98
    An indirect method to monitor  the oxidation  reaction of  AsdII}, by
chlorine is to measure the disappearance  of chlorine.  In addition to the
inherent disadvantage of an indirect method, two other disadvantages arose:

     (1) Arsenic(III) can be the only  species present that exerts  a  chlorine
demand.

     (2)  The colorimetric  methods  for free chlorine determination proved to
be not sensitive enough.

In experiments with 1.0 mg/L chlorine  and 100 ppb As(III)  in 0.01 M NaCl,  the
DPD colorimetric method [62] showed no significant  difference between a blank
of 1.0 mg/L chlorine  and  the  oxidation samples after 1,  2, 5,  10  minutes


                                     27

-------
contact tine.  Furthermore, the adsorbance of a given sample changed over a
tine span of  several minutes (see  Appendix  6).  The color change  was  visible;
the color changed from  an  initial reddish color to a deep purple color  often
within 60 minutes.

    As continuous monitoring of the  reaction proved unfeasible,  oxidation was
conducted  in batch reactors each sampled at a  different time.  The  bacch
reactors were 2-L polyethylene  beakers filled  with  1-L of solution.   In
oxidation tests on 100 ppb As(III) in DT water by  1.1 mg/L chlorine, samples
taken at 1, 5, IS,  30 minutes then separated (10 minutes) had no detectable
arsenic(lll)  remaining (see Table  3).

     	TABLE 3.  OXIDATION IN PI WATER WITH  NO  QUENCHING	
          Time

           1  minute

           5

          IS

          30
    Therefore,  a quenching agent was needed because the reaction  time  of  the
oxidation reaction  was  faster  than the ininimum separation  time of 10 minutes.
Several criteria weie required of the quenching agent.

     (1) The  quenching agent must react with chlorine instantaneously in a
mannner to prevent  further oxidation of As(III).

     (2) The  quenching agent cannot oxidize As(III) nor reduce  As(V).

     (3) The  quenching agent must not deleteriously effect the separation
procedure.

     (4) The  quenching agent should not substantially  interfere  with arsenic
analysis by GFAA.

Ammonia Quenching--

    Ammonia,  at first glance,  appeared to be a good candidate for a quenching
agent.  It reacts very fast with chlorine to form monochloroamines and  was
unlikely to oxidize As(III) or reduce As(V), or  interfere with  the arsenic
separation or analysis procedures.  However, previous success  in  using  the
addition of ammonia  and the  subsequent formation of chlorocmines as  a
quenching agent for the oxidation of Se(IV) [4]  and  Cr(III)  [14]  by chlorine
did not  follow  for  the oxidation of  As(III).   Oxidation of 100 ppb As(III) in


                                     28

-------
DI water occured  in  the initial  presence  of  1.0 mg/L chlorine and a 20-molar
excess of NHjCadded as NH^-C1).  These results are in Table 4.

                    TABLE 4.   SCREENING OF AMMONIA QUENCHING

Sample
control
control
blank
Time
30 sec
1 min
1 min
As(III) Remaining
43
39
98

Thiosulfate Quenching—

    Thiosulfate, a common dechlorinating agent, was also considered for use as
a quenching agent.   However, the reaction of chlorine with thiosulfate was
reported to occur in a somewhat slow stepwise fashion [71).  In view of the
apparent  ease of oxidation of  As(III),  thiosulfate was  abandoned  as  a
potential  quenching agent without performing any  experiments.   Furthermore,
thiosulfate proved to  be unsuccessful in the quenching of Se(IV) oxidation by
chlorine apparently  by interfering with the separation by ion exchange [4].

Scdium Sulfite Quenching--

    Sodium sulfite is  another common dechlorinating agent.   It is  reported to
react instantaneously with chlorine [71].   However, experiments indicated that
As(V) was reduced in the presence of a three times excess of sodium sulfite
based  upon a 1.0  mg/L C12  dosage. The  results of  these experiments are
summarized in Table  5.

     	TABLE 5.   SODIUM SULFITE RESULTS	


                Sample                        As(IIl) Recovered,  ppb


     100 ppb  As(V)                                      18

     100 ppb  As(III) with 1.0 mg/L C12                   98

     100 ppb  As(V) with 1.0 mg/L C\2                    14

     50 ppb As(III)/50 ppb As(V)                        61
                                      29

-------
DPD Quenching —

    DPD, N,N-diethy 1-p-pheny lene dlamine  oxalate. which is used  in  the
colorimetric t»;st  for  chlorine,  is reported  to react instanteously with free
chlorine under the  proper conditions.   The color change associated with the
DPD reaction  with chlorine can  be seen immediately  upon addition  of  DPD
reagent under the  proper conditions.  This reaction is one-to-one, follows
Beer's  Law and results in  the destruction of free chlorine.  Furthermore,  DPD
is unlikely to oxidize Ap(III)  or reduce As(V).  Therefore,  DPD was a  likely
candidate as a quenching  jgent.

    A standard DPD solution (62]  has a pH of approximately 2 because of the
addition of  su If uric coid  to promote dissolution.  In order  to keep the pH in
the range 6.2 to  6.5  a phosphate buffer  is  used in the procedure for  the
determination  of chlorine.  Initial experiments  with this  quenching  method
were conducted using a standard  DPD solution and a  standard phosphate buffer
solution [62], 5 mL each per 100 mL of sample.  However, these experiments
could not be  analyzed successfully  for  arsenic.   The peak  height  absorbance
was drastically  reduced.   Suspicion that the phosphate  buffer was primarily
responsible  was confirmed by addition of 5 mL of phosphate buffer to 100 mL of
arsenic standard.  In order to reduce  interference, the amount of phosphate
buffer added was reduced 10 fold and the  pH  adjusted to the correct range with
1 N NaOH (or,  in the case of extremely high initial pH, with 1 N  HC1).   Jar
tests were conducted to determine  the amount oi NaOH solution  to be added for
each initial pH.   Once these amounts were determined (see Table 6), they were
used consistent Ij.   This  procedure allowed the arsenic analysis by CFAA to be
performed satisfactorily.  As shown in Table 7, DPD appeared to  quench the
oxidation reaction  and neither oxidize  As(lII)  nor reduce Ae(V).N

                  TABLE 6.  AMOUNTS OF NaOH ADDED FOR
                    Initial pH                  IN NaOH, mL
5.5
6.5
7.5
8.5
11.5
12.5
1.5
1.5
1.5
1.0
0.0
1.5





(1 N HC1)
                                     30

-------
                         TABLE 7.   DATA ON DPP QUENCHl.'C
              Sample                       As(III) Remaining, ppb
     100  ppb As(III)                                97

     100  ppb As(III)                                98
         with C12 and quenching agents

     SO ppb As(III)/50 ppb As(V)                    51
         with C12 and quenching agents
    For a dosage of  1.0 mg/L chlorine In an oxidation test, the quenching
dosage of DPD was at 4.3 molar excess.  Further details are given in Table 8.

     	TABLE 8.   REAGENTS	

     ReagentConcentration

                                  meq/L            mol/L


                                  0.34            0.34

                                  0.34            0.17
        **

     C12                        2.8 X 10'5      2.8  X  10'5
     DPD
                                1.2 X 10'*      1.2  X  10'*
The DPD was added under conditions of extremely turbulent nixing in order to
achieve  rapid  dispersion.  The  color  change that  DPD underwent proved
beneficial in monitoring the  effeciency of rapid mixing.

    Mixing tests were conducted  to determine if As(III) oxidation occured
during mixing.   No detectable oxidation took place after 60 minutes of rapid
mixing.   These  results  are  presented in Table 9.
                                      31

-------
                          TABLE 9.  MIXING TEST DATA
     Stirring Speed       Time of Mtxlng.niin  	Percent AsCllI) Recovered
low
tied
high
high
high
high
1
1
1
2
5
10
98
98
98
97
97
98
Chlorine Oxtdant Experiments

    An artlfical ground veeer, composition given in Table 10, was used as
background  water for  the  oxidation  test  except  where otherwise noted.

     	TABLE 10.  BACKGROUND WATER	

     Anlons      meq/L      mg/L         Cations    meq/L      mg/L
soA2'
HC03"
Cl"
2.0
6.0
2.0
96
366
71
Ma* 10.0 230



The pH was adjusted, if needed, by dropwise addition of either 0.1 n HC1 or
1.0 M NaOK  solution.  All  pH measurements were made with an  electrode
calibrated on using  two buffers bracketing  the expected pH value.   An initial
arsenic(HI) concentration  of  100  ppb  was chosen because it is twice the MCL
and is not an uncommon concentration in arsenic contaminated ground waters.  A
chlorine dosage of 1.0 mg/L was chosen for  the batch  tests  because  there was
low  chlorine  demand in  the artifical  ground water and the  dosage was
sufficient for reaction completion.  The batch chlorine oxidation  tests were
conducted in 2-L polyethylene beakers  with 1-L sample  size.  A propeller mixer
provided rapid mixing  for  the addition  of stock  chlorine solution and
quenching agent.  The shaft and  impeller were plastic  coated.   All tioe
increments were measured with a laboratory timer.   A  100-mL pipette was used
to collect the sample which was immediately separated.  An initial  sample and
the  As(III) fraction  were  stored  in  a  60-mL  and  125-mL respectively
polyethylene bottles  for subsequent GFAA  analysis.   An occasional As(V)
fraction was collected as a  spot  check of separation recovery; otherwise,
Ae(V; was determined by difference.
                                     32

-------
Oxygen Oxidant Tests

    Some oxygen oxidation tests were  conducted on various background waters
containing 100 ppb As(III).  For al 1 these experiments, a 500 mL sample In a
Pyrex gas vanhlng bottle. Model  f 31760 was bubbled with Oj at a flow rate  of
300 to 40C nL/minute for 60 minutes.  One sample to serve  as  a blank was
bubbled with nitrogen under  these sane conditions.

    In another experiment, orsenic(Ill) containing  sampled vere allowed to sit
in 125 ml polyethylene bottles with  air in the headspace  for 61 days.   Samples
vere 200 ppb As(lII)  in DI water at  pH 6.0, 7.0, and 8.0;  SO ppb As(V)/ SO ppb
As(III) in DI water at pH 6.0,  7.0,  and 8.0; and 50 ppb As(III)/ 50  ppbAs(V)
in artificial groundwater with  1 raeq Ca2+  (composition given  in Table 11)  «t
pH 7.3 and 8.3.

     	TABLE 11.  BACKGROUND WATER WITH CALCIUM"	


     An ions      neq/L     tag/L        Cations     meq/L.    ng/L


     S^2"        2.0         96         Na+        9.0       207

     HC03"        6.0        366         Ca2+       1.0'        20

     «'         2.0         71
                                      33

-------
                                SECTION 6

                           RESULTS AND DISCUSSION
COLUMN STUDIES

    The effluent history from the As(IIl)  column  is shown in Figure 5, and the
effluent history  from the As(V) column  is  shown in Figure 6.   As  Figure 5
shows,  effluent trivalent arsenic  concentration  reaches the MCL much sooner
than fluoride  (300 and 1600 bed  volumes,  respectively),  indicating  that
fluoride is a more preferred species than trivalent  arsenic.  This result is
expected because, at the  influent  pH of 6.0,  trivalent  arsenic  is
predominantly  non-ionic arsenious acid.  As Figure 6 indicates,  effluent
arsenic(V) reached  the  MCL significantly  later than fluoride breakthrough,
i.e. 23,400 and 1550 bed volumes, respectively.  This behavior was  expected
because  arsenate  is  a more  preferred  species than  fluoride  and  on  a
milliequivalent basis  the As concentration is  much  lower than fluoride--
0.00135  meq/L compared  to 0.158  meq/L.

    As shown in Figure 7,   a comparison of the arsenite versus  arsenate
breakthrough curves, As(III) reached the MCL almost immediately at 300 bed
volumes(0.6 days),  while As(V) did not reach the  MCL until  23,400 bed volumes
(48 days).  Thus, the presence of  pentavalent arsenic  results in column runs
nearly 80 times longer than trivalent arsenic.   Therefore,  pre-oxidation of
Ad(III)-containing  waters is essential  for efficient treatment using activated
alumina.

    Even though trivalent arsenic  appeared in the effluent  very  quickly, some
arsenic (III) was  removed.   In fact,  at 90 percent of equilibrium,  a  mass
balance showed  that 0.344 mg As(III) had been removed yielding an approximate
mass loading of 0.078  mg As(III)/  g alumina.   Speciation of  the  influent and
the effluent showed the arsenic to be total ly As(III).  Therfore,  it is the
trivalent  species that is actually absorbed onto the alumina, assuming no
oxidation of As(III) occured.

    The comparatively poor removal of  As(III) versus As(V)  is not suprising in
light of alumina surface chemistry and of arsenic speciation.  Alumina is an
aophoteric  ion  exchanger, capable  of both anionic and cationic  exchange  (34].
In acidic solutions,  surface-bonded  protons are electrically balanced by
anions adsorbed.  In  basic  solutions, cations Are  electrically balanced by
surface-bonded anions.    Hydroxide  ions are highly preferred as counterions to
the adsorbed  protons in neutral and slightly acidic solutions.   Other anions
such as fluoride and arsenate must  compete  with hydroxide for  ion-exchange
sites.


                                     34

-------
in
                                              Time,Days
                                         7   8  9  10  11 12  13 14  15 16  17 18
   100
    90
    80
    70
£  60
a
•"  50
<
    40
    30
    20
    10
      0
Run Number 1
Influent Composition
                                                                  70.9 mg/L, 2.0 meq/L
                                                                  384 mg/L, 8.0 meq/L
                                                                  3.0 mg/L. 0.16 meq/L
                                                                  100 ppb
                                                                  211 mg/L, 9.16 meq/L
                                                                  20.0 mg/L, 1.0 meq/L
                                                                  6.0
                                                                  2,9 mln.
                                                                                 9
                                          1000  Bed Volumes
     Figure 5.  Arsenlte, As(III), and fluoride breakthrough curves from a minicolumn
               containing 28 X 48 mesh, F-l activated  alumina.

-------
                                                      Days
                                 10
                                     20
                                             30
                               40
91
       JQ
       a
       a
100
 90
 80
 70
 60
 50
 40
 30
 20
 10
   0
OJ
E
                                   Run Number 2
                                   Influent Composition
70.9 mg/L, 2.0 meq/L
384 mg/L, 8.0 meq/L
3.0 mg/L, 0.16 meq/L
100 ppb
211 mg/L, 9.16 meq/L
20.0 mg/L, 1.0 meq/L
                                     6
                                 8
                                10    12   14    16
                                1000 Bed Volumes
                           18    20   22   24    26
         Figure 6.  Arsenate,  A$(V),  and fluoride breakthrough curves from a minlcolumn containing 28 X 48
                   mesh, F-l  activated alumina.

-------
  100
               10
              —I—
20
Days

    30
40
 50
—I—
         ••	As(lll). Arsenlte. Run No. 1
                                  As(v). Araenate. Run No. 2
                        10
          15
            20
           25
                       1000 Bed Volumes
Figure 7.  Comparison of As(III) and As(V) breakthrough curves
          from minicolumns of activated  alumina.  CT = 100 ppb.
          EBC1 = 2.9 min. PH - 6.0
                      37

-------
    In fact,  a generally accepted selectivity sequence  for anlon adsorption
(or Ion exchange) on alumina exists  [6. 67, 68).   This sequence is presented
below:

          OH"  > HP042", HAsO^2" > F~  > Si(OH)30' > CrO^2' > Cl"  >

                       N03" > MnOA~  > C104~ > CH3COO"

At the column  influent pH of 6.0, trivalent  arsenic exists predominantly
(1000:1)  as  the non-ionic  species--arsenious acid.   However, as  Figure 8
shows, at the influent  conditions  (pH = 6.0, 0.100 mg/L As(III)), the hydroxyl
ion concentration is ten tines  greater than the arsenite anlon concentration.
An arsenic mass balance on Run 1  showed  that 0.344 ng As(III) was adsorbed;
however,  only 0.0001 ng As(III) was available as the  anion of arsenious acid
to be adsorbed.  Therefore, it is unlikely that ion  exchange is  the only
mechanism of adsorption of As(lll)  because more arsenic  was adsorbed than
existed as  the  anion.   Furthermore, at pH  6.0 the arsenite anion  always  exists
In concentrations less  than the concentration of the  more  preferred hydroxyl
ion.

    Unlike  trivalent arsenic, pentavalent arsenic exists predominantly  as the
singly-charged anion, HjAsO^".  As Figure 9 shows,  at  pH 6.0,  this species
concentration is  100 times  that of the highly preferred hydroxyl ion.   Even
the doubly-charged  anion, HAs042", exists at over  10  times the  concentration
of  the preferred hydroxyl ion.  Thus,  based  on concentration,  As(V) can
compete favorably for the available adsorption sites.

    As indicated  in  Figure  10,  fluoride adsorption was little  affected  by the
difference  In  adsorption  of As(III) as  compared  to As(V).    The  effluent
concentration of  fluoride reached  the MCL at 1600 bed volumes  for the  column
with As(III) and at 1500 bed volumes for the column with As(V).   Although
As(V) is more  preferred than  fluoride and As(III)  is  less  preferred, the
difference in arsenic adsorption is unlikely to effect  fluoride adsorption.
This situation  occurs  because the  molar ratio of F~/As  is  117.  Thus, arsenic
vould typically occupy only one site for every  117 sites  occupied by  fluoride,
if the selectivities were equal.

    Studies  on  arsenic  and  fluoride  removal were conducted in San Ysidro,  NM
using the University of Houston/EPA  Mobile  Drinking Water Treatment  Research
Facility [16],   The fluoride and  total arsenic effluent histories  from that
work are  presented in Figure 11.  Their column was 1" ID  filled with Alcoa F-l
activated alumina (mesh 24  X 48).  A  comparison of results from  the  field and
laboratory data are presented in Table 12.   Although the bed volumes  to the
fluoride MCL were greater in the field study, the capacity to the fluoride MCL
was almost exactly  the  saca as  the capacity obtained  in   laboratory tests.
                                      38

-------
VO
       o

       O>
       O
           -10
                                                 PK--9.2
- -10
                                                       10     12     14
                                           PH
        Figure 8.  Log concentration vs. pH diagram for arsenlous acid, As(III), system.
                  Total As(III) concentration is 100 ppb.

-------
                 PK,
O

 O)
 o
                                       PH
 Figure 9.   Log concentration vs.  pH diagram for arsenic acid,  As(Y)j system.
            Total As(V) concentration Is  100 ppb.

-------
 D)
 E
i
                           Days
                       4       6
     8
10
      =  2  meq/L
      =  8  meq/L
        3.0 mg/L
        1.0 meq/L
pH    -  6.0
• As(v) = 100 ug/L
• As(lll)=100 ug/L
                   1000  Bed Volumes
 Figure 10.  Comparison of fluoride breakthrough curves from mini-
          columns of activated alumina.  CT = 3.0 ir,g/L EBCT = 2.9 min
                         41

-------
                TABLE 12.  COMPARISON OF LABORATORY AND FIELD DATA
                         FOR FLUORIDE AND ARSENIC REMOVAL	
                                 UH Laboratory  Test        San Ysidro

     	As(III)	As(V)	Field Test

     BV  to As MCL                  300        23400           8760

     Arsenic capacity, g/cubic n     18          1610            575

     BV  to F~ MCL                 1600          1550           2520

     Fluoride capacity, g/cubic m  4190          4280           4160
     Notesi  The San Ysidro water contained 48 ppb As(V), 37. ppb As(III), and
            2.0 ppm fluoride.

    Although the  influent arsenic concentration differed between  the field
study and  the  laboratory study, the arsenic  removed per  m  of  alumina in the
field was qualitatively in the range expected from tlie laboratory studies
(i.e. greater than 18 g/m3 and less than 1610 g/nr).  However, the shape of
the  arsenic breakthrough  curve  in  the field  study  (see Figure 11} was
suprisingly  sharp  because  a  much  earlier  breakthrough of   trivalent arsenic
vas expected. By way of explanation, some oxidation of As(IIl) to As(V) nay
have occured in the  field  column,  and  the trivalent arsenic concentration of
the  influent  in  the  field  study  was only 0.032 mg/L--one third the
concentration  in the  laboratory study.  Furthermore,  the pentavalent arsenic
breakthrough curve in Figure 6 indicated early breakthrough of  A»(V).  This
effect was  presumably because of the shorter EBCT of  the lab  column, its
shallow bed depth, and the  fact that the adsorption zone was a large fraction
of the bed depth.

CHLORINE OXIDANT STUDIES

Kinetics

    The results of the kinetic experiments on the oxidation  of  arsenic(III) by
chlorine in  artificial ground water with sodium as  the  cation are  presented in
Figure 12.   The Figure  12  indicates that the reaction has reached completion
by the 5-second data point. However, five  seconds is the fastest  possible
reproducible quenching time  using this method.   Under  the experimental
conditions,  the data imply that -d[As(IIT)]/dt  £  2.66 X 10~7  mol/L sec.
Johnson and  Bruckenstein [35),  in  their study on the  oxidation of As(III) by
electrogenerated  iodine at  pH 8.2 to 9.2, reported a maximum overall rate -
d(As(III)]/dt of approximately  7 mol/L sec.  Chlorine oxidant Is considered a
more powerful oxidizing agent than iodine since it is capable of oxidizing
iodine.   Assuming the  oxidation of  As(III)  by  chlorine follows  a rate


                                     42

-------
    90

    80

    70

    60
.o
a  50
   40

   30

   20

   10

    0
3.6
3.4
3.2
3.0
2.8
2.6
2.4
2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
                                   Time, Hours
                   100  200  300  400  500 600  700  600
          10
  Time. Days
15     20    25
30
35
   San Yaldfo. New Mexico
   AA, 26x48. Run No. 2
   April 12-Uay 8. T984
   Bod Volume • 400 ml     EBCT  • 9 mln.
   Flow Rata  -80 ml/tnin   pH,,,d - 8.0 -  0.1
Total As Adsorbed • 210 mg(or 230 mg to 50 pob Aa)
Total F" Adsorbed • 2025 mg(or 1663 mg to  1.4 pom F~)
                             .Co • 80 ppb
                                 [40% As(lll)]
                                10

                                9

                                8

                                7

                                6
                                                                    pH
                                4567
                              1000 Bed Volumes
                                       8
 Figure 11.  Total arsenic and fluoride breakthrough  from 1
            column in San Ysidro. KM.
                                                 inch
                              43

-------
1

O)
£
!
-------
expression of the form

         - d[As(III» / dt  -  k  [AidlD]" [Cl2]n                       (1)

where [As(III)]  - [H,AsO3l + lH2As03~],  and  [C12J - [HOClJ + [OCl'J, the only
variables  available for manipulation are  (As(III)]  and IC12).  In American
water treatment, chlorine dosages less than 1.0 mg/L are not often practiced.
Increasing  the  arsenic  concentration 10A times to 1000 mg/L,  an  unrealistic
ground water  concentration,  would  only allow  the determination of
-d(As(III)]/dt  <  2.66  X  10"3  mol/L  sec.   For  these  reasons,  further
experiments to determine the  kinetics were abandoned.

    From equilibrium calculations,  the oxidation  of As(III) to As(V) by
chlorine should  go tc completion;  however, as shown in Figure 12, the  reaction
only goes  to approximately 95  percent completion.   Although at pH 8.0
equilibrium calcula.'->ns predict  the As(V)/As(III)  to  be 2.34 X  10" at  these
experimental  conditions, the  reaction appears to reach  completion at an
As(V)/As(lII) ratio of approximately  19.    The leakage of 1 ppb As(V) during
the  analytical  separation procedure would  automatically  give  a  false
As(V)/As(III) ratio  of 100;  the measured As(III) concentrations shown in
Figure 12 may be attributable  to leakage  of  As(V).    The apparent 95 percent
complete reaction is  surely sufficient  for water treatment practice.  However,
other factors may effect the extent  of reaction  including the  pH, and the
presence of other ions.

jpjl Effect

     As the pH increases, the  speciatlon of both arsenic and chlorine changes,
and the free energy  of reaction decreases.  The experimentally  observed effect
of pH on the  oxidation of As(III) by chlorine  is shown in Figure 13.  In the
neutral  pH range of 6.5 to  9.5,  the  pH does not  significantly effect the
extent of  reaction.   It is  not until pH values greater than  10.5 that the
oxidation reaction is affected adversely.   By pH 11.5  the   G  of reaction has
been approximately  halved from the AC of reaction at pH 6.5.  The relative
insensitivity of the oxidation reaction  of As(III)  to As(V) by chlorine in
the  neutral pH range  may  be attributed  to the fact  that  As(III) exists
predominantly as arsenious acid  in this pH range and that the changes in  pH do
not  significantly affect the activation of  the species.  Furthermore,  from the
data it appears that for pH values less than 10.5 that the AC of  reaction is
sufficient  with  adequate overpotential for the  reaction to occur.   At pH 11.5,
the overpotential necessary for  complete reaction may not be available.

    The  slight  decrease in reaction extent at pH 5.5 can be  explained by
examining the oxidation equation for that pH range:
                  HOC1 -
                                  + 2 H+ + Cl"                           (2)
Even though  the AC of  reaction (2) increases  with decreasing  pH,   this
oxidation reaction is acid producing in the 2.2  to 7.0 pH range.   As the  pH
decreases, the H* concentration increases representing an accumulation  of
products.  Another contributing factor may  be  the  increase in Cl" as a result

                                      45

-------
    70


    60
J3
Ct
a  50
co
c
•5  40
E

c  30
01
<  20
    10
            Initial Conditions
            CI2 Dosage "1.0 mg/L
            As(lll) = 100 ug/L
            Cl~   • 2.0 meq/L (7 t mg/L)
            SO|" • 2.0 meq/L (96 mg/L)
            HCO:
• 6.0 meq/L (366 mg/L)
- 10.0 meq/L (230 mg/L)
               6
         8     9    10    11    12    13

             PH
Figure 13.  The effect of pH on As(III) oxidation by chlorine
           with DPO quenching after 30 sec.
                          46

-------
of acidification from pH  8.3 with  KC1.   The effect  of background  Cl~ is
theoretically Important and will now be discussed.

Effect of Chloride Concentration

    Figure 14 shove the effect of Cl~ on the extent of reaction with either
sodium or calcium as the counterion for  chloride.  The data point at 0.1 M
NaCl is falsely high because the high Cl" concentration  drove some As(V) off
the ion-exchange resin  during separation,  and caused a  falsely  high  As(III)
value.   The curve was  drawn to an estimated actual concentration.  This
estimation was  made on  the  basis of blank separation of 50 ppb As(IIl)/50 ppb
As(V) in 0.1  M  NaCl.

     As Figure  14 shows, increasing Cl"  concentrations in the  bacVground water
decrease the extent of oxidation  reaction. Although some breakthrough of Ae(V)
nay have  occured during separation,  a  real  As(III)  concentration remains at
the 0.02 M data point. The open square represents As(V) leakage from a blank
separation  of  100 ppb As(V) in  0.02 M  Na Cl.   Since chloride  represents a
product of the oxidation,  this  behavior is qualitatively expected.  In DI
water, no As(III) is detected.  However, at 0.01 M Cl", As(IIl) Is detected
after oxidation.    At pH 6.0 and {Cl "] » 0.01 M,  equilibrium calculations
predict the  As(V)/As(III) ratio to be 2.22 X 1085.   Three possibilities exist
to explain this behavior.   The first explanation is experimental error in the
separation.  As(V) leakage during separation nay account for the traces of
arsenlc(III) found after oxidation at salt concentrations up Co 0.01 molar.
The second explanation  is  that the published  thermodynamlc data  are in error.
Some error in the published thermodynanic data is  likely.  The calculated pK
values using  Gj from.  La timer (37]  for arsenic acid are 3.6, 7.26, and 12.47
while the measured values  are  reported  as 2.2,  7.0, and 12.5.  And, Latimer Is
ultimately the basic  source of  the  thermodynamic data published for arsenic.
And although errors in the published  thermodynamic literature  aie possible, it
is unlikely  to account for the complete difference between the predicted and
measured result,  especially since the errors  in the thermodynamic  data  seem to
be predominantly for the  reactions at  lower  pH values.   Chloride ion  appears
to exert substantial influence on the  extent of oxidation at high chloride
concentration,  although this influence is of Insignificant practical value at
the chloride concentrations encountered in drinking water  treatccnt.

    Furthermore,  these  data indicate that the increase in As(III)  remaining at
lower pH values nay be  in part (approximately 5 ppb)  because of the added Cl"
during pre-acidiflcation with HC1.  However,  the trend of decreasing extent of
reaction for pH decreasing  below pH 6.5  still  appears to hold.

    The  effect of Cl~  concentration  upon  arsenic(III) oxidation  Is  very
different than  its effect  upon Cr(III) oxidation.  The oxidation  of  Cr(IlI) by
chlorine  is much slower than  the oxidation of As(lll);  the Cr(IIl)  oxidation
occurs  over several hours at higher chlorine dosages  [14].   Furthermore,
increasing  chloride concentration appears to enhance the reaction, probably
because the  chloro-Cr(HI)  complex is more susceptible to  oxidation than the
Insoluble alternative—Cr(OH)3.
                                     47

-------
a.
a
di
c
'c
'3
a>
CC
Initial Conditions
As(lll) • 100 ug/L
CI2 Dosage = 1.0 mg/L
pH ° 6.0
A  Na * Counter Ion to Cl~
  Ca  Counter Ion to Cl
100
 90
 80
 70
 60
 50
 40
 30
 20
 10
        0            0.001           O.C1             0.1
                         Cl~ Concentration, molar
Figure 14.   The effect of chloride concentration on As(III) remaining after
           oxidation using 1.0 mg/L chlorine dosage.

-------
Effect of Counterion

    Figure 14 also shows a comparison between  the extent of  reaction when
sodium is the only cation or when calcium  is the only cation to  chloride.  At
chloride concentrations encountered in ground  water (0.0001 M to 0.01  M),  no
effect can be attributed to the cation.  No real  ligand effect can be seen.

    Figure 15 shows the kinetics of reaction are not measurably  changed  in the
presence of  1.0  meq/L Ca2"*".   A comparison  of the  rate of  oxidation  in
artificial ground water that  contains 10 meq/L  sodium versus one that contains
9 meq/L sodium and 1 meq/1 calcium shows no measurable difference.  If the
presence of calcium effects the rate of As(III) oxidation  to As(V), it is
probably of no  practical significance for treatment processes.

Effect of Chloramtnes

    Figure 16  shows  the kinetics of oxidation of  0.100  mg/L As(III)  in
artificial ground water containing a dosage of  1.0 mg/L chlorine with a 10
molar excess  of  ammonium chloride  at  pH 8.3. These  conditions produce
monochloramine.   This figure has two salient points.  First, the reaction
appears to reach completion at an As(V)/As(III) ratio of approximately two
thirds.  And, secondly, the reaction appears  to have slower kinetics than the
oxidation  of As(IIl)  by  free chlorine.   As  expected, analysis  of the
chloramine species, showed only the presence of NH2C1.   The following reaction
can be written:
               NH2C1+ U3Ae03 + H20 = NH4+ + HAsO^" + Cl" + 2 H+       (3)

Using the  free energies  of formation given in Appendix 7,  this reaction has a
standard free  energy of  -82.4 kcal/mol and an equilibrium constant  of  2.57 X
1060.   Thus,  as  written,  the  above  reaction  is  possible.   linger  the
experimental  conditions, thermodynamic calculations predict an As(V)/As(IlI)
ratio of 5.66 X 10 7® at equilibrium.  The observed data does not match *he
predicted  value  at all.   The thermodynamic data available may be somewhat in
error as discussed previously.  However,  these  errors are not likely to
account  for  the total difference between observed and predicted values.

    One  possible explanation is that, although the above reaction  is  possible,
it does  not  actually occur because the energy barrier of activation is too
high.   In  this case,  oxidation is accomplished by analytically undetectable
amounts (less than 0.1 mg/L) of free chlorine  remaining in solution.  The
formation of monochloroamine is  a reversible  reaction;  however,  the
equilibrium  occurs in  the direction  of  monochloramine  formation.   Given the
apparent ease of oxidation of As(III) by free chlorine oxidant, very  small
amounts of free chlorine may be capable of partially oxidizing As(III), as
will  be quantified later.  The hydrolysis reaction of monochloramine to
produce  free chlorine has only recently been studied  [29].  Kinetic  data show
this reaction to be slow.  Current kinetic data indicates that,  at 60 minutes,
7 percent  of the monochloramine should be hydrolyzed.  However,  this result
was not  observed experimentally.
                                     49

-------



O)
c
'E
CO
E
or
Q
o
CO




100«
90
80
70
• %^
60
50
40

30
20
10
i
1
Initial Conditions
100 ppb As(lll)
1.0 mg/L Chlorine
Cl~ s 2.0 meq/L
SO2" = 6.0 meq/L
HCOg = 2.0 meq/L
pH = 8.3
A Ca2+= 1 meq/L
Na+ = 9 meq/L


Dosage
(70 mg/L)
(288 mg/L)
(122 mg/L)



® Na+ = 10 meq/L
•
»
•
I 	 i 	 1 	 1 	


A
I 1 .
                                           8
10
                          Time, Minutes
Figure 15.  A comparison of the kinetics of oxidation of 100 ppb
           As(III) by 1.0 mg/L chlorine dosage in the presence of
           different counterions.
                        50

-------


.0
a
a

en
"c
'15
E
a
oc
mz

-------
    Studies on Se(IV)  (4] and  Cr(III)  [14]  showed that they could not be
oxidized by chloranines.  The following reaction  analogous to reaction  (3),
can be written:

                 NH2C1 + Se032' + H20 = NH4+  + Cl' + SeO^2'              (4)

This reaction  has a standard free energy  of reaction of  -25.74 kcal/mol  with
an equilibrium constant of 7.43 X 10  .  It  is potentially less energetic  than
the arsenic  reaction.  No noticible selenium  oxidation occurred in the
presence of monochloramine.  However, these factors do not necessarily imply
that nonochl'oramine  is the oxidizing species.  The oxidation  kinetics of
selenium(IV)  are much slower than the kinetics of arsenic(IIl) oxidation,
implying that  the  activation  energy  of arsenic(III)  is  much  lower than  that
for Se(IV).  As a result  the trace amounts of  free chlorine may not be capable
of detectable  Se(IV) oxidation.

Effect of pH on Chloramlne Oxidation

    As Figure 17  shows, the  solution pH does not appear to substantially
effect the extent of  reaction  in the  pH range 6.5  to  10.5.  This pH
sensitivity of the  chloramine  reaction is  very similar  to that  obtained for
oxidation  of  As(III)  by free chlorine  (See  Figure  13) where  a  somewhat
decreased rate is observed  at  pH  5.5.  The monochloramine formation  reaction
should be dependent upon pU because of the  different opecies of  chlorine and
ammonia present at different pH values.  The reaction of monochloramine to
dichloramine  is not considered here because the reaction only  occurs at pH
less  than  6.0.  And this reaction occurs  only slowly.  Calculations-  from
kinetic data show that after 1 hour only  7  percent of the monochloramine has
been converted and at the 1 minute  quenching time  used in  this experiment  less
than 0.1 percent  of  the monochloramine should  be converted to  dichloramine at
the lowest pH  examined.  The following reactions are possible!

          a.   pH < 7.5

              HOC1 + NH4+ - NH2C1 + H20 + H+                             (5)

                     AC - -2.62 kcal/mol

                       K - 83.3

          b.   7.5 < pH < 9.3

              OC1" + NH4* = KH2C1 + H20


                         - -12.92 kcal/mol

                       K - 2.97 X  109
                                     52

-------


.0
Q.
a
•
c
'ca
V
QC
=
(0




100
90
80

70
60
50
40
30

20
10

Initial Conditions
As(lll) = 100 ug/L
Cl~ =71 mg/L, 2 meq/L
SOj" 3 96 mg/L, 2 meq/L
HCOT » 366 mg/L, 6 meq/L
A , *
X Na =230 mg/L. 10 meq/L
\
^ A A A
-
-

•
-
i . i . i . i . i . i . i
                                      8
9
10     11
                                    PH

Figure  17.  The effect of pH on the oxidation of  100 ppb As(III) by
           1.0 mg/L monochloramine and excess aranonla.
                            53

-------
          c.  9.3 < pH

              OCI" + NH3 - NH2C1 + OH"                                  (7)

                     AC - -6.44 kcal/mol

                       K - 5.27 X 10*

Reaction (6)  implies that  In the pH  range  7.5 to  9.3  the  ratio  of
monochloramine to chlorine IB constant.  At  experimental conditions,  this
ratio  is  equal to  7.5  X  105  for the pH  range 7.5  to 9.3  using the given
thermodynamic  data.  At  the  level  of total chlorine used in the experiment,
the free chlorine concentration  using the  given thermodynanic  data  would  be
3.76 X  10"11 molar.  This concentration  is not nearly  enough  to  oxidize the
amount of arsenic observed to be oxidized experimentally  if  the oxidation
occurs  on a  stolchiometrlc basis.  However,  thes* thermodynanic  data bay be in
error;  the values  for K  and consequently the monochloramine to chlorine ratio
nay be  in error by 2 orders of  magnitude, at least.  Reactions  (5) and (7) are
explicitly dependent upon pH.   The monochloramine to chlorine ratio depends
upon pH  for pH  values  less  than 7.5 or greater  than 9.3.  At  pH  6.0,  the
monochloramine to chlorine ratio is  2.1 X  10*; and at pH 10.3 it is 6.7 X 10*.
Outside  the pH  range  of 7.5 to 9.3,  the monochloramine  to chlorine ratio
decreases implying that  the chlorine concentration  increases.  But the data do
not reflect  an increased extent of  reaction.   However,  as mentioned earlier,
the extent of  oxidation  by chlorine decreased  outside the  pH range 6.5 to 9.5.
It is,  therefore,  still possible for the  chlorine in solution to be the actual
oxidizing agent.

Effect  of TOC

    The kinetics  of As(III) oxidation in aged Houston, TX tap water with 1.0
mg/L chlorine dosage  are shown in Figure IB.  Although tie  kinetics  were
appreciably  slowed in this water,  the reaction  reached 95 percent completion
as did  the oxidation of  100 ppb As(III) in artificial  ground water by  1.0 mg/L
chlorine.   However, the reaction  in the aged tap water  toc-k  60 minutes to
reach  completion while the oxidation reaction in artificial ground water
reached  completion within 5  sec.  The  oxidation of As(III)  in artificial
ground water was  at least 720 times  faster than  the reaction in tap water.
Presumably the chlorine  demand exerted by the  5 mg/L TOC in the  aged tap water
was responsible for slowing the kinetics of reaction.  Hypochlorous acid may
undergo the  following reactions with organic constituents in  water:  oxidation
reactions, substitution to form either N-chlorlnated organics or C-chlorinated
organics, or addition reactions.  In general, at  least part of the chlorine
demand  of a water  is exerted  Immediately, with  other Blower reactions occuring
at longer chlorine contact times [29]. In fact,  the oxidation may be parallel
reactions with oxidation by free chlorine occuring  immediately, and oxidation
by combined  chlorine--either chloramlnes  or organic chloramines--occuring at a
slower rate  for the  second stage.
                                       54

-------
 O>
 CO
 E
 0)
 CC
 CO
<
A Aged Tap Water
  TOG * 5 mg/L
  pH  =7.5 mg/L
9 Synthetic Ground Water
  pH - 8.0
            5  10  15  20  25 30 35 40 45 50 55 60

                         Time. Minutes
Figure 18.  Comparison of the kinetics of oxidation of 100 ppb
           As(III) by 1.0 mg/L chlorine  dosage in aged tap water
           and artificial ground wa-.er.
                       55

-------
OXYGEN OXIDANT STUDIES

Sparging Tests

    The results of the  oxygen sparging  tests are presented in Table  13.
A sample  of  100 ppb As(III)  in  synthetic  ground water, pH 8.3,  was bubbled
with N2 at a  flow rate of 360 mL/min.   Even though  glass  surfaces  reportedly
adsorb arsenic  133],  ninety  seven  percent  of  the  arsenic(HI) was  recovered
after  the test.  In  the  oxidation  tests  in artificial ground  water, very
little, if any, oxidation occurred; whereas,  in  the DI water, somewhat  greater
oxidation occurred.   Because  Cl~ inhibited As(III) oxidation by chlorine,  it
may also  stabilize As(III)  with  respect  to oxygen oxidation; however,  effects
of the other  ions in solution cannot be  Ignored.  The presence of 1.0 mg/L
Fe(III) appeared to enhance  the oxidation reaction by oxygen slightly.

     	TABLE 13.  The Results of 0; Oxidation Batch Tests8	

                                  Arsenic(III)  Oxidized
     Conditions          Synthetic Croundwaterb         DI Water

     pH 6.0                        3~Z14 Z

     pB 7.5                        5 1

     pH 8.3                        5 I

     1.0 mg/L Fe(II)                                        16 1

     1.0 mg/L Fe(TII)               8 Z (pH 7.3)            28 Z (pH 6)


          a) All tests were of 60 minutes duration with 100 ppb  As(III)
     initially.

          b) Composition:   2  meq/1 Cl~, 6 meq/L  HC03~, 2 meq/L SO^"2;
                           1  meq/L Ca, 9 meq/L Na

          c) 02 flow  rate ° 325-400 mL/min

Despite highly favorable thermodynamic  conditions, the oxidation of As(III)
appears to occur only very  slowly.  It is possible, therefore, that the actual
oxidizing species is  not  02, but H202.   The  mechanism was  postulated  to
account for the very  slow kinetics  of many oxidation reactions  by oxygen.  The
formation of hydrogen peroxide Intermediate  Is  the rate limiting step in this
mechanism [37].

As(III) She If-Life Experiments

     Various  standard  arsenic solutions were stored in polyethylene  bottles
and  left  on a shelf at room temperature  and  away from direct sunlight.  After
61 days,  the  As(III) fraction was separated at a flow rate of 5 mL/min.  The


                                      56

-------
results of the shelf-life experiments are given in Table 14.

     	TABLE U. As(IIl) REMAINING AFTER 61 DAYS	

                             ppb Aa(IIl) Remaining after 61 days
      Initial    	PI Water	    Synthetic Groundwater
        pH      50 ppb As
-------
                                    REFERENCES
 1.  Anbar, M. and H. Taube.  The  Exchange  of  Hypochlorite and Hypobromite
     Ions  with  Water..  JACS  80:1073-1079.

 2.  Anderson, C. E.   Arsenlcals  as Feed Additives for Poultry and Swine
     in Arsenic—Industrial, Bionedical, Environmental Perspectives.
     Van Nostrand Reinhold,  1983 pp. 89-98.

 3.  Bel lack, E.  Arsenic Removal  from Potable Waters.  JAWWA July,1971
     pp. 454-458.

 4.  Boegel, J.  V. and 0. Clifford.  Oxidation and removal  of Selenium
     by Ion Exchange.   EPA In  preparation.

 5.  Borgono,  J.  M.   Epldemiologlcal Study of Arsenlclfim in City of
     Antofagasta.  In:  Proceedings of University of Missouri Fifth
     Conference on Trace Substances in Environmental Health, 1972.

 6.  Brattebo, Kelge. Phosphorus  Removal from Wastevater by Fixed-Bed
     Adsorption on Granular  Activated Alumina.   Ph.D. Thesis,
     University of Trondheim, Norwegian Institute of Technology, 1983.

 7.  Buswell, A. M.  et al.  Water Problems in Analysis and Treatment.
     JAWWA Oct.   1943.

 8.  Cachaza, J. M.,  J. Casado, A. Caetro, M.  A. Lopez Quintela. Kinetics
     of Oxidation of  Nitrite  by Hypochlorite  Ions in Aqueous Basic
     Solution.  Can.  J  Chem. 54:   3401-3406.

 9.  Calmon, C.   Comment.  JAWWA  August, 1973 pp. 569-570.

10.  Cherkinskii, S.  et al.  Purification of Arsenious Wastewaters.  WPA.
     Vol.  14.

11.  Chronic /rsenic  Poisoning in the Lake Region (Mexico).  WPA 38: 430.

12.  Clifford, D. A., T. Sorg, P. Frank.  A Survey of  Inorganic Contami-
     nants.  Procedings 1983  AWWA Seminar  on  Inorganic Contaminant
     Removal, AlfWA,  Denver, CO.

13.  Clifford, D. A.,  J.  V. Matson  and R. Kennedy.  Activated Alumina
     Rediscovered 'Adsorbent' for Flourlde, Humic  Acids,  and Silica.
     Industrial  Water Engineering.  Dec,  1978, pp. 6-12.


                                      58

-------
14.   Clifford. 'J.  and J.  Chau.   Fate  of Cr(III)  in Chlorinated Water.
     EPA, In preparation.

15.   Clifford. D., L. Ceber, S. Chow.  ArBenic(IIl)/Arsenic(V)
     Separation by Chloride-Form Ion-Exchange Reeins.  XI AWWA
     WQTC  Proceedings,  Norfolk, VA. Dec 1983.

16.   Clifford, D.  and C. C. Lin.  Arsenic Removal Pilot Studies in San
     Ysidro, NM using Activated Alumina, Ion Exchange, Reverse Osmosis,
     and Electrodialysis.  In preparation.

17.   Committee On  Medical  and Biologic  Effects of Environmental
     Pollutants.  Arsenic.  National  Academy of  Science,  1977.

18.   Draley,  J. E. Aqueous  Chemistry of Chlorine.   Ini  Water
     Chlorlnation Volume  1.  R. J.  Jolley,  ed.  Ann Arbor Science
     Publishers,  1980.

19.   Enter line, P. E. and  C. M. Marsh.  Mortality Among Workers Exposed
     to Arsenic and Other Substances  in a Copper Smelter.  In:  Arsenic
     Industrial, Biomedical and Environmental Perspectives, Van Mostrand
     Reinhold, 1983  pp.   226-244.

20.   Fleischaker,  S.  J. and S.  J. Randtke.  Formation of Organic Chlorine
     In Public Water Supplies.   JAWWA March, 1983 pp.  134-138.

21.   Ferguson, J.  F. and J. Gavls.  A  Review of the Arsenic Cycle in
     Natural Waters.  Water Research 61  1259-1274.

22.   Fox,  K.  R.  Arsenic Removal:   Point-of-use  Study. Unpublished Draft
     EPA, 1980.

23.   Fox,  K. R.  Removal  of  Arsenic and Selenium from Drinking Water by
     Reverse Osmosis.  Master's Thesis, University  of Cincinnati.

24.   Furst,  A.  A  New Look at Arsenic Carcinogenesls.   Ini Arsenic--
     Industrial, Biomedical, Environmental  Perspectives.  Van Nostrand
     Reinhold,  1983  pp.  151-165.

25.   Grantham, D. A.  and J. F. Jones.  Arsenic Contamination of Water
     Wells In Nova Scotia.  JAWWA  Dec,  1977  pp. 653-b57.

26.   Culens,  J., D. R. Champ, and R. E.  Jackson.   Influence of Redox
     Environments on the  Mobility  of Arsenic in Ground Water.  Ini
     Chemical Modeling In Aqueous  Systems.

27.   Cullege, J. H. and J. O'Connor.  Removal of As(V) from Water by
     Adsorption on Aluminium and Ferric Hydroxides.   JAWWA August,1973
     pp. 548-552.

28.   Gupta,  S. and K. Chen.  Arsenic  Removal by Adsorption.   JWPCF


                                       59

-------
     March. 1978  pp. 493-506.

29.   Hall, L. W. et al.  Power Plant Chlorination:  A Biological  and
     Chemical Assessment.   Ann Arbor  Science  Publishers, Electric Power
     Research Institute,  1981.
30.   Halperin,  J.  and  H. Taube.  The Transfer of Oxygen  Atoms  In
     Oxidation-Reduction Reactions III.  The  reaction  of  Halogenates with
     Sulfite in Aqueous Solution.  JACS 74:  375-379.

31.   Harvey,  S. C.  Heavy Metals.   In:  Pharmacological  Bases  of
     Therapeutics, Goodman and Oilman,  eds.  MacMillian.

32.   Horng,  L.  L.  and D. Clifford.  Arsenic(V) Removal  by Ion Exchange.
     EPA, In Press.

33.   Irgolic, K.  J.  Speciation of Arsenic Compounds  In Water Supplies.
     EPA 600/S1-82-010 November,  1982.

34.   James,  R.  0.  Surface  lonization and  Complexation at  the Colloid/
     Aqueous Electrolyte Interface.  In:  Adsorption  of Inorganics at
     Solid-Liquid Interfaces.  M. Anderson and A. Rubin, eds.,  Ann Arbor
     Science, 1981  pp. 219-261.

35.  Johnson, D. C. and S. Bruckenstein.  A Study of  the Kinetics of the
     Oxidation of As(III)  by Electrogenerated  Iodine  In Alkaline Media.
     JACS 90(24):  6592-6596.

36.   Laguitton,  D.  Arsenic  Removal From Gold-Mine Waste Waters:  Basic
     Chemistry of the  Lime  Addition Method.  CIM Bulletin Sept,  1976
     pp. 105-109.

37.   Latimer, W. M.  Oxidation Potentials.  Prentice-Hall,  Inc.,  1952.

38.   Lee, J.  and R. Rosehart.   Arsenic Removal  by  Sorption Processes from
     Wastewaters.  Canadian Mining Institute Bulletin, Nov, 1972
     pp. 33-37.

39.   Lister,  M. W. Decomposition of Sodium  Hypochloritei   The Catalyzed
     Reaction.   Canadian Journal  of Chemistry 34:  479-488.

40.   Lister, M. W. and P.  Rosenblum.   Kinetics  of  the Reaction of
     Hypochlorite and Nitrite. Canadian J of Chen. 39:   1645-1653.

41.   Lister, M. W. and P. Rosenblum.  Rates  of Reaction of Hypochlorite
     Ions with Sulphite and Iodide Ions.  Can J Chem 41i   3013-3020.

42.  Logsdon, G. S. and  J.  Symons.  Removal  of Trace  Inorganics  by
     Drinking Water Unit Processes.  AIChE meeting June 3-6,  1973.
                                        60

-------
43.   National Research Council.  Drinking Water and Health. National
     Academy Press,  1980.

44.   National Research Council.  Drinking Water and Health.  National
     Academy Press,  1983.

45.   Nillson, R.  Removal  of Metals by Chtcical Treatment of Municipal
     Wastewater.   Water Research. 5:   51-60.

46.   O'Connor, J. T.  Trace Metals  versus Conventional Water Treatment
     Systems.  Water and Sewage Works.  Oct,  1974.

47.   Peoples, S.  A.   The  Metabolism of Arsenic in Man and Animals.  In:
     Arsenic — Industrial, Biomedical, Environmental Perspectives, Van
     Nostrand Reinhold, 1983   pp.  125-133.

48.   Pierce,  M.  L.  and  C.  Moore.  Adsorption  of Arsenlte and Arsenate on
     Amorphous Iron Hydroxide.  Water  Research  16:  1247-1253.

49.   Pourbalx, M. and J. Van  Muylder.  Arsenic.   In:  Atlas of Electro-
     chemical Equilibria  in Aqueous Solution, M. Pourbaix, ed.,  1980.

50.   Richardson,  C.  W.  et  al.  Arsenic  Concentrations in  Surface Runoff
     from Small Watersheds  in Texas.   J Environ Qual  7(2):  189-192.

51.   Rosehart,  R. and J. Lee.   Effective  Methods of Arsenic Removal from
     Gold Mine Wastes.  Canadian Mining  Journal.  June, 1972.   pp.  53-57.

52.   Rosenblum,  E.  and  D.  Clifford.   The  Equilibrium Arsenic Capacity of
     Activated Alumina.   EPA  600/S2-83-107 Feb 1984 NTS PB 84110527.

53.   Schraufnagel,  R. A.   Arsenic in Energy  Sources:  A  Future  Supply or
     an Environmental  Problem?.  In:  Arsenic — Industrial, Biomedical,
     Environmental  Perspectives, Van Nostrand Reinhold, 1983  pp.  17-41.

54.   Si lien, L.  G.  and  A.  E. Martell.  Stability Constants of Metal Ion
     Complexes.   The Chemical  Society, Burlington  House,  London,  1964.

55.   Silver, S.  and H. Nakahara.   Bacterial Resistance to Arsenic
     Compounds.   In:  Arsenic--Indu8.trial, Biomcdical, Environmental
     Perspectives,  Van NOBtrand Reinhold,  1983  pp. 190-199.

56.   Shen,  Y.  S. and C. S. Chen.  Relation Between  Arsenic Blackfoot
     Disease and the Pollution of  Drinking Water by Arsenic in Taiwan.
     JWPCF 36(3)1964.

57.   Shen,  Y. S.   Study of As Removal  from Drinking Water.  JAWWA
     August, 1973   pp. 543-548.

58.   Smith,  J.  D.   Arsenic,  Antimony and Bismuth.  In:   Comprehensive
     Inorganic Chemistry Volume 2,  Pergamon  Press,  1973 pp.  580-683.

                                       61

-------
59.   Sorg, T.  and W.  Bruce.  Arsenic  Removal  From Drinking Water by a
     Household Reverse Osmosis System.   Unpublished Draft of Research
     Progress Report, Aug 1979, EPA  Cincinnati, OH.

60.   Sorg, T.  J.  and  G. Logsdon.  Treatment Technology to  Meet  the  1PDWR
     (II).  JAWWA, July 1978  pp. 379-393.

61.   Southvick,  I. W. et al.  An Epidemiological  Study of Arsenic in
     Drinking Water in Mil lard County, Utah.  In:  Arsenic—Industrial,
     Biomedical, Environmental Perspectives, Van Nostrand Reinhoid, 1983
     pp. 210-225.

62.   Standard Methods for the Examination of Water and Wastewater,  APHA,
     AWWA and WPCF,  15 edition, 1979.

63.   Stumm, W. and J. J.  Morgan.   Aquatic Chemistry.  Wiley Interscience,
     New York, 1980.

64.   Subramanian, K.  S.  et al.   Determination of Arsenic(III) in Some
     Ground Water  Samples.  Atomic Spectroscopy  5(4)« 192-194.

65.   Taube, H.  Electron Transfer Reactions of Complex Ions in Solutions,
     Academic Press  New York,  1970.

66.   Trelles, R.  A.  Treatment of Arsenical  Waters with  Lime (Argentina).
     WPA  22:  125.

67.   Trussell, R. R., A. Truscell and P.  Kreft.   Selenium Removal from
     Ground Water Using Activated Alumina.   EPA 600/J2-80-153, EPA,
     Cincinnati,  Ohio.

68.   Umland, F.   Uber die  Wechselwirkung von  Elektrolytlosungen und
     A1203. 4. Mitteilungi  Entwlklung einer formalen
     lonenaustauschertheorie fur die Adsorption von Elektrolyten aus
     Wassrigen Losung.  Z.  fur Elektrochemie,  60(1956):   711-721.

69.  Uthus, E. 0., W. E.  Co.-natzer and F. H. Nielsen.  Concequences of
     Arsenic  Deprivation  t.n Lab oratory  Animals.   In:  Arsenic—
     Industrial,  Biomedic.-il, Environmental Perspectives, Van Nostrand
     Reinhoid,  1983.  pp!73-189.

70.  Weber, W. J.  Physicochemical Processes.  Wiley Interscience,  1972.

71.  White, G. C. Handbook of Chlorination.  Van Nostrand Reinhoid,  1972.

72.  Young, K.  Geology.  Boston, Houghton-Mifflin, 1975.

73.  Zingaro, R. A.  Biochemistry of Arsenic:  Recent Developments.   In:
     Arsenic—Industrial, Biomedical, Environmental Perspectives, Van
     Nostrand Reinhoid,  1983  pp.  327-347.
                                       62

-------
                          pH = 6.98
                                 APPENDIX 1


     Equations used for construction of Eh-pH diagram for arsenic and water
system:

Dissociation reaction boundaries

[H2As04~]
	                  pH = 2.2
lH3As04l


lHAsQ42"l

[H2As04'J


[As043-J
	                  pH » 11.50
[HAs042'J


[As02-]
	                  pH = 9.22
[HAs02]


Oxidation-reduction reaction boundaries


                         pE = 9.45 - pH
[HAs02I


        I
                         pE « 16.55 - 2 pH
[H3As04J
	                 pE = 9.31 -  1.5 pH
[AsO+J
                                      63

-------
'HAsOj]
[HAs02]
                          pE = 11.27 - 1.5 pH
                          pE «• 14.91 - 2 pH
                          pE = 10.30 - 1.5 pH
                                       64

-------
                                  APPENDIX 2


     The equations used for the construction of the Eh-pll diagram for chlorine
and water are as follows.

Dissociation reaction boundary

loci']
	                    pH = 7.5
[HOC1]


Oxidation-reduction reaction boundaries

[HOC1]
                          pE » 26.9 -  pH
pE = 23.6
IC12]'5


IC12]'5
-
[Cl ']


[HOC1]
-                   pE » 25.25 - 0.5 pH
[CD
loci"]
-                    pE - 28.9 -  pH
                                       65

-------
                                  APPENDIX 3






                       Chlorine Oxidant Decay Reactions




Breakpoint reaction




     2 NHC12 + H20 = N2 + 3 H* -r 3 Cl" + HOC1




     4 HOC1 + NH3 " N03" 4 4 Cl" + 5 H+ + H20






Oxidation of Carbon




     CH2OH(CHOH)ACHO + HOC1 - CH2OH(CHOH)4COOC 4 H+ 4 Cl"



     R-CmiH2-COOH + HOC1 " R-CHO + NH3 +C02 + H+ + Cl'






Organic Substitution




     C6H5OH + 3 HOC1 •= C6H2C13OH + 3 HjO






Disproportionation




     3 HOC1 - 3 IT* + 2 Cl" + C103"






Decomposition to Oxygen




     2 HOC1 = 2 H+ + 2 Cl" + 02






Inorganic Oxidation




     N02" + HOC1 - N03" + H* + Cl"




     Mn2+ ••• HOC1 + H20 = Hn02(s) + 3 H+ + Cl"
                                       66

-------
                                  APPENDIX 4






                 Conditioning Procedure for Activated Alumina




Bed Volume » 50 ml of Alcoa F-l type alumina, mesh 28 X 48




     Rinse                        Bed Volumes




     1 I NaOH                        10




     DI Water                        10




     2 * H2S04                       10




     DI Water                        10



     1 Z NaOH                        10




     DI Water                        10




     2 I H2COA                       10



     DI Water                       100    pH «• 4.5






Air dried overnight, then dried 24 hr at 105 C
                                       67

-------
                             APPENDIX 5
                          Arsenic Analysis
Step
Dry
Char
Atomize
Temperature
110 C
250 C
2700 C
Ramp Time
5 sec
5 sec
1 sec
Hold Time
25 sec
20 sec
10 sec
Wavelength: 197.7 nm
Gas Flow:   Interrupt Argon
Low Slit Width:  0.7 nm

Sample Volume:  20x'-L
Matrix Modifier:  20 s>L 1000 mg/L NiN03 as Ni in 2 % HNO3
                               68

-------
Sample


Blank

1.0 mg/L Cl, dosage
100 ppb As(III) with
1.0 ng/L C12
     APPENDIX 6


Change In Abserbance

  Initial Absorbance


           0

         0.35

         0.30

         0.28
Absorbance after
   2 minutes

      0

    0.4!>

    0.38

    0.40
Background water 0.01 M MaCl
                              69

-------