PB81-20407?
     The Equilibrium Fluoride Capacity of
     Activated  Alumina.  Determination of the
     Effects  of pH  and.Competing Ions
     Houston  Univ.,  TX
      Prepared  for

      Municipal Environmental Research Lab,
      Cincinnati,  OH /
      Hay  81
I
       Department of Commerce
   National Technical Information Service

-------
                                   TECHNICAL REPORT DATA
                            (Pleae nut Inauctiaa on the reverie before completing)
 1. REPORT NO.
   EPA-600/2-81-082
                             2.
                                  ORD Report
3. RECIPIENT'S ACCESSION NO.
    POT   20A07  5
 4. TITLE AND SUBTITLE
   Equilibrium Floride Capacity  of  Activacted Alumina
                                                          8. REPORT DATE
                                                           May  1981
                                                          0. PERFORMING ORGANIZATION CODE
 7. AUTHORISE
   Gurinderjit Singh
   Dennis A. Clifford
                                                          8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
   Environmental Engineering Program
   University of Houston
   Houston, Texas 77004
                                                          10. PROGRAM ELEMENT NO.

                                                            BNC1A	
                                                          11. CONTRACT/GRANT NO.
                                                             CR80607
 12. SPONSORING AGENCY NAME AND AOORCSS
   Municipal Environmental  Research Laboratory - Cin. ,011
   Office of Research and Development
   U.S. Environmental Protection Agency
   Cincinnati, Ohio 45268
                                                          13. TYPE OF REPORT ANO PERIOD COVERED

                                                            F-lnal  7/7g _ T?/Qn	
                                                           SH
14. SPONSORING AGENCY'COC

  EPA/BOO/14
IS. SUPPLEMENTARY NOTES

  Project Officer - Thomas J. Sorg
                                         513/684-7370
 IS. ABSTRACT
      This report describes  research on the aetermi-atlon of the equilibrium fluoride
 adsorption capacity of small columns of acid pretreatad activated alumina  (Alcoa F-l
 grade).  The experiraent&l observations verified the expectation that  fluoride is very
 favorably adsorbed in preference  tc the common anions: sulfate, chloride and bicarbo-
 nate.  However, the adsorption  capacities were found to be four to  five times higher
 than what has been reported in  the  early literature for municipal defluoridation   ^
 processes.
                                                                                     /
      Fluoride adsorption capacity is significantly affected and is  decreased with
 an increase in pH beyond seven.   The alumina selectivity sequence determined by
 experiments was the same as has been reported in the early literature,
  OF~>SO^=>C1~>HC02~.  Although  fluoride anions are preferred over sulfate ions, the
 sulfate ions compete significantly  at the levels found in ground water supplies.
 Experiments with high ionic strength ranging up to 56 millimoles per  liter (5600 ppm
 as CaC03) indicate that the total adsorption capacity increases slightly.   Fluoride
 adsorption capacity decreases only  slightly with the very significant increases in the
 concentrations of the other anions.  These equilibrium data will prove useful in
 utilizing the maximum adsorption  capacity of activated alumina in municipal defluorida'
 tion processes*	•   - .		
 7.  .
                               KEY WORDS ANO DOCUMENT ANALYSIS
                  DESCRIPTORS
                                             b.lDENTIFIERS/OPEN ENDED TERMS  C. COSATI Field/Group
   Water Treatment
   Water 'Supply
   Adsorption
   Fluoride
                                              Activated Alumina
                                              Bench-Scale Treatment
                    13B
 3. DISTRIBUTION STATEMENT

  Release to Public
                                             19. SECURITY CLASS (This Report)
                                              Unclassified
             21. NO. OF PAGES
               7/
                                             20 SECURITY CLASS
                                               Unclassified
                                                                        22. PRICE
EPA form 222O-I (*-73]

-------
                                                 EPA  600/2-81-082
                                                 May  1981
THE EQUILIBRIUM FLUORIDE CAPACITY OF ACTIVATED ALUMINA

           Determination of the Effects of
                pH and Competing Ions
                          by
                  Gurinderjit Singh
                  Dennis A. CL ifford
          Environmental Engineering Program
              The University of Houston
                Houston, Texas  77004
                "Grant No. R806073010
                   Project Officer

                    Thomas J. Sorg
           Drinking Water Research Division
     Municipal Environmental Research Laboratory
               Cincinnati, Ohio  45268
     MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
          OFFICE OF RESEARCH AND DEVELOPMENT
         U.S. ENVIRONMENTAL PROTECTION AGENCY
               CINCINNATI, OHIO  45268

-------
                                  DISCLAIMER
     This report: has been reviewed by the Municipal Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for publica-
tion.  Approval does not signify that the contents necessarily reflect the
views and policies of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
                                       ii

-------
                                   FOKEWORD

     The U.S. Environmental Protection Agency was created because of Increas-
ing public and government concern about the dangers of pollution to the health
and welfare of the American people.  Noxious air, foul water, and spoiled land
are tragic testimonies to the deterioration of our natural environment.  The
complexity of that environment and the interplay of its components .require a
concentrated and integrated attack on the problem.

     Research and development is that necessary first step in problem solution;
and it involves defining the problem, measuring its Impact, and searching for
solutions.  The Municipal Environmental Research Laboratory develops new and
improved technology and systems to prevent, treat, and manage wastewater and
solid and hazardous waste pollutant discharges from municipal and community
sources, to preserve and treat public drinking water supplies, and to minimize
the adverse economic, social, health, and aesthetic effects of pollution.
This publication is one of the products of that research and provides a most
vital communications link between the researcher and the user community.

     Pursuant to the provisions of the Safe Drinking Water Act of 1974 (Public
Law 93-523), many public water supplies in the U.S. may eventually have to be
treated to remove excess fluoride.  The research described in this report
establishes the maximum adsorption capacity of activated alumina for the
removal of fluoride ions as influenced by pH and competing ions, viz., sulfate,
chloride, and bicarbonate.
                                       Francis T. Mayo, Director
                                       Municipal Environmental Research
                                       Laboratory
                                      iii

-------
                                   ABSTRACT
     This report describes research on the determination of the equilibrium
fluoride adsorption capacity of small columns of acid-pretreated activated
alumina (Alcoa F-l grade).  Mini-columns containing 1 g alumina adsorbent
were equilibrated over a period of 6 to 10 days with continuous flows of
constant pH fluoride solution.  The adsorbed ions were then eluted with 1%
sodium hydroxide solution over a period of A hours.  The alkaline regener-
ants were analyzed to determine the quantities of ions "adsorbed" (ion
exchanged) during the exhaustion step.  As expected, fluoride was very
favorably adsorbed compared to the common anions:  sulfate, chloride, and
bicarbonate.  .However, the equilibrium fluoride adsorption capacity was
found to be twice as great as that reported in the recent literature for
municipal defluoridation processes.

     Fluoride adsorption capacity is significantly affected and is decreased
with an increase in pH beyond 7.  The alumina selectivity sequence determined
by experiments was the same as has been reported in the early literature:
OH~>F~>SO^>C1~>HC03.   Although fluoride ions are preferred over sulfate ions,
the sulfate ions compete significantly at the levels found in groundwater sup-
plies.  Fluoride adsorption capacity decreases only slightly with the very
significant increases (up to 56 millimoles/L) in the concentrations of the
other anions.

     The laboratory research on fluoride removal was done in support of field
studies on the removal of fluoride using activated alumina columns in a mobile
pilot plant designed and constructed at the University of Houston.  The Mobile
Drinking Water Treatment Research Facility was completed on April 9, 1980,
and transported to its first field location—Taylor, Texas, a small community
(population 13,000) with a high-fluoride (3.0 mg/1) groundwater supply.  The
mobile facility containing activated alumina and ion exchange columns and
reverse osmosis and electrodialysis units may be transported to any U.S.
community with i\n inorganic contaminant problem.

     This report .Is submitted in partial fulfillment of Grant Number R8 06073-
010 by the University of Houston under the sponsorship of the U.S. Environ-
mental Protection Agency.  A second project report covering the design, con-
struction, and operation of the Mobile Drinking Water Treatment Research Facil-
ity will be published at a later date.
                                       iv

-------
                               CONTENTS

Foreward	ill

Abstract	iv

Figures  . . 	 ......................  vl

Tables	vii

Abbreviations and Symbols  	  ix

Acknowledgment 	   x

   1.  Introduction  	   1
   2.  Conclusions   	   6
   3.  Theory	   7
   4.  Methods and Materials	12
   5.  Results	21
           Effects of pH	26
           Effects of sulfate Ions	  26
           Chloride effects  	  29
           High ionic strength effects 	  29
   6.  Discussion	35
           Effluent concentration histories  	  36

References	44

Appendices .. 	 ..................  47

Mini-column Effluent Concentration Histories 	  51

Glossary	58

-------
                               FIGURES
Number                                                             Iage

  1A   Mini-column for capacity determination 	   14
  IB   Alumina-conditioning column  	   14
  2    Typical alunina mini-column set-up for
         adsorption run	 .  .   15
  3    Fluoride adsorption isotherm pH - 5.0  	   22
  4    Fluoride adsorption isotherm pH - 6.0  	   23
  5    Fluoride adsorption Isotherm pH - 7.3	   24
  6    Fluoride adsorption Isotherm pH - 8.0  	   25
  7    pH effect on fluoride adsciption 	   27
  8    Effects of ch.ioride and sulfate on
         fluoride adsorption Isotherms  	   28
  9    Mini-column effluent concentration histories
         pH - 6.0, Effect of SO^" cone, on column kinetics  ...   30
 10    Mini-column effluent concentration histories
         pH - 6.0, Effect of CI~ cone, on column kinetics ....   31
 11    Mini-column effluent concentration histories,
         pH - 7.1, Effect of ionic strength on column kinetics   .   32
 12    Anlon adsorption Isotherms, pH - 7.1	   33
 13    An ion adsorption Isotherms, pH - 7.1	   34
 14    Langmulr isotherm for pH 5, pH 6	   38
 15    Langmulr isotherm for pH 7, pH 8	   39
 16    Freundlich isotherm for pH 5, pH 6	   40
 17    Freundlich isotherm for pH 7, pH 8	   41
 18    Effect of pH on time for 90% adsorption Equilibrium  ...   42
                                 vl

-------
                                 FIGURES

                                APPENDIX

Number                                                                Page

  Al    Mini-column effluent concentration histories,
          pH - 5.0	5i.

  A2    Mini-column effluent concentration histories,
          pH - 5.0	52

  A3    Mini-column effluent concentration histories,
          pH - 6.0	53

  A4    Mini-column effluent concentration histories,
          pH - 6.0	54

  A5    Mini-column effluent concentration histories,
          pH - 7.0	t  .  55

  A6    Mini-column effluent concentration histories,
          pH - 8.0	56

  A7    Mini-column effluent concentration histories,
          pH - 7.1	57
                                  vii

-------
                                  TABLES

Number                                                                 Page

  1    Maximum contaminant levels for fluoride 	 	   3

  2    Full-scale packed bed defluoridation plants 	 4-5

  3    Iso-electrlc point of activated alumina 	 ...  10

  4    Typical properties and specifications of activated
         alumina, F-l type	13

  5    Column data for determination of the effects of pH
         on fluoride adsorption capacity 	  18

  6    Column data for determination of the effects of
         sulfate and chloride on fluoride adsorption capacity  ....  19

  7    Column data for determination of the effects of ionic
         strength on fluoride adsorption capacity  	  20

  8    Fluoride adsorption isotherm data at 22°C	37


                                 APPENDIX

  1A   Time spent in hours for throughout T • 1    	49
                                  viii

-------
                    LIST OF ABBREVIATIONS AND SYMBOLS

 g
a       Separation factor between ions B and A, dimensionless

BV      Bed volume

b.      Langnuir constant related to adsorption energy, 1/mg

C       Liouid chase concentration, mg anion/1

EBCT    Empty Bed Contact Time

1       Ionic strength, moles/liter

K.       Freundlich constant, mg/g

M       Moles

n       Freundlich constant, g/1

Q       Rate of flow through column, ml/min

Q'      Langtnuir ultimate solid phase adsorption capacity constant, tng/g

q       Solid phase concentration, mg anions adsorbed/gm Al

r       Correlation coefficient

t -     Time in hours to reach 90 percent equilibrium of exhausting
           mini-column

T       Throughput, i.e., ratio of the total amount of fluoride which
           comes in contact with the adsorption column during a speci-
           fied amount of time to the total adsorptive capacity of the
           column

T .     Throughput at 90 percent approach to equilibrium
  . y
TIC     Total 'inorganic Carbon

TOC     Total Organic Carbon
                                   ix

-------
                                ACKNOWLEDGMENTS
     Mr. Tom Sorg, Project Officer, is thanked and recognized for his helpful
cooperation throughout the course of this work and the concurrent work on the
design and construction of the Mobile Pilot Plant to be used for field research
on the removal of fluoride and other ions.

     Eric Rosenblum is acknowledged for his competent editing efforts on the
report; JoAnn Wardin for her excellent and efficient typing; and  Sumeet Singh
for her valuable assistance in drafting all of the figures.

-------
                                   SECTION 1

                                 INTRODUCTION
EFFECTS OF FLUORIDE IN DRINKING WATER

     Chronic effects of fluoride on the skeletal system have been documented
in several geographic areas where drinking water contains excessive fluoride.
Tliis form of chronic fluorosis was first described in India and has since been
reported in Ceylon, China, South Africa, Japan, Saudia Arabia, the United
States, Canada, and Europe [V.H.O., 1970].

     Mottled tooth enamel is the most delicate index of chronic fluorosis in
humans.  It is epidemic in areas where the drinking water contains 2 ppm F~
or more, and there is a precise relationship between the severity of mottling
and the concentration of fluoride in the drinking water [Sognnaes, 1953].  In
a 1953 report on the medical aspects of fluorosis in Bartlett, Texas, where
the drinking water contained 8 ppm of fluoride [Shimkin, et al., 1953], a
little difference was discovered between the "high fluoride" group and the
control group with respect to the number and types of specific disease symp-
toms elicited.  However, the high fluoride group manifested a higher incidence
of mottled enamel and an Increased bone density of the spine and pelvis.  They
also experienced a certain hrittleness and blotching of the fingernails,
hypertrophic changes in the spine and pelvis, and lenticular opacities in the
eye.

     Dental tissues (including the skeleton) accumulate fluoride most rapidly
during formation and mineralization [National Research Council, 1971].  During
tooth formation, the cells of the dental tissues—particularly ameloblasts—
are very sensitive to fluoride.  At relatively low doses (e.g., 2 ppm of fluo-
ride in the water) small spots of discoloration may form in the tooth surface.
These spots, or "mottling," vary in color from paper white to dark brown, the
brown stains usually accumulating when exposure to fluoride persists after the
tooth has erupted.  At higher doses, the cells may be affected and the tooth
structure severely altered, so that the normally smooth surface becomes corru-
gated.

     The effects of fluoride in drinking water on animals are analogous to
those in human beings.  The ingestion of excessive amounts of fluoride by
young farm animals for long periods causes striking changes in the teeth
such as brown discoloratlons, pitting, and marked wear of the entire tooth.
The exposure of adult farm animals with completely erupted teeth to excessive
amounts of fluoride for long periods will produce chronic symptions of fluoro-
sis, except that the teeth will remain smooth.  Bones and joints of animals

-------
may become enlarged, and extra bone growths or extroses may appear in differ-
ent parts of the skeleton.  Some animals may become stiff, and Intermittent
lameness may occur [Greenwood, 1956].
MAXIMUM CONTAMINANT LIMITS FOR FLUORIDE

     The U.S. Public Health Service (USPHS) Drinking Water Standards of 1962
set a mandatory limit for fluoride based on the annual average of maximum
daily ambient temperature.  The fluoride-temperature relationship has been
established on the premise that the consumption of water increases with tem-
perature; therefore, the maximum allowable fluoride levels should decrease
with increased temperature.  The U.S. Environmental Protection Agency (U.S.
EPA) subsequently adopted these standards [Federal Register, 1975] (Table 1)
when promulgating the National Interim Primary Drinking Water Regulation
(NIPDWR) pursuant to the Safe Drinking Water Act (PL 93-523).

     In May 1972, the National Institute of Dental Research estimated that
community public water supply systems serving approximately 4.2 million people
exceeded the recommended maximum contaminant levels established for fluoride.
A summary published in January 1980 revealed that in the state of Texas, over
400 community water supplies exceeded their maximum NIPDWR limit of 1.6 mg/1
of fluoride.
FLUORIDE REMOVAL PROCESSES

     Several method? of defluoridation have been utilized for municipal and
residential fluoride removal.  These methods can be broadly categorized into
two kinds of processes:  (1) precipitation methods and (2) packed bed adsorp-
tion methous.

     In the first category, chemicals like lime (alone or in combination with
magnesium from dolomite), magnesium sulfate, magnesia, or calcium phosphate
are added to precipitate fluoride or to form hydroxide precipitates onto which
fluoride will adsorb and settle out of solution.  Addition of other chemicals
—such as bentonites, fuller's earth, diatomaceous earth, silica gel, bauxite,
sodium silicate, sodium aluminate, and ferric salts—has also been attempted,
but these materials require a very low pH (less than 3.0) to effect fluoride
removal in reasonable quantities.

     When both hardness and fluoride removal are desired and the water con-
tains sufficient magnesium, softening is the most feasible process.  On the
other hand, fluoride removal with alum has been found to require three to five
times the usual alum dose of 50 ppm normally required for water clarification.
At LaCrosse, Kansas, raw water was treated with 225 ppm of alum to reduce the
fluoride content from 3.6 to 1.5 ppm [Gulp and Stoltenberg, 1958].  Maier
[1953] reported that water treatment at Bartlett would have required 900 ppn
alum to remove 7 ppm fluoride, and further dosage would have raised the sul-
fate content to a very undesirable level.  Thus the total costs of alum treat-
ment of groundwater, including sludge disposal, would likely be prohibitive
for most small communities.  Furthermore, when fluoride is the only contaminant

-------
           TABLE 1.  MAXIMUM CONTAMINANT LEVELS (MCLs) FOR FLUORIDE
         (From National Interim Primary Drinking Water Standards, 1975)

                  Average Maximum Daily Temperature
Temperature
°F
53.7 and Below
53.8 to 58.3
58.4 to 63.8
63.9 to 70.6
70.7 to 79.2
79.3 to 90.5
Temperature
°C
12.0 and Below
12.1 to 14.6
14.7 to 17.6
17.7 to 21.4
21.5 to 26.2
26.3 to 32.5
MCL
mg/1
2.4
2.2
2.0
1.8 .
1.6
1.4
whose removal from drinking water is required, neither alum nor lime addition
may prove economical.

     When fluoride alone must be removed from drinking water, fluoride adsorp-
tion on activated alumina, bone char or trlcalclum phosphate has historically
been considered the most reasonable, cost-effective removal method [Clifford,
1978].  However, an unanswered question remains as to which media is superior
overall for fluoride removal when capacity, service life, regenerant require-
ments, and media cost are taken into consideration.

     Much confusion exists in the literature regarding the fluoride capacities
of the various media.  The reason for this confusion car. be attributed to the
fact that the fluoride capacity of an adsorbent depends rather heavily on (1)
the pH of the feedwater and (2) the regeneration history of the adsorbent.  In
addition, competing anions and cations and the total ionic strength of the
feedwater are expected to influence the fluoride adsorption or ion exchange
process.  In general, one or TPore of these significant parameters have been
overlooked by earlier investigators [Fink and Lindsay, 1936; Swope and Hess,
1937; Savinelli and Black, 1958].

     Clifford [1978] has collected and summarized data from municipal defluor-
idation installations dating back to 1937.  Thia data are reproduced below as
Table 2.

-------
                            TABLE 2.  FULL-SCALE PACKED-BED DEFLUORIDATION PLANTS
            Process,
         Location, Date,
            (Reference)
  Gross
Installed
Capacity
   MGD
Process Description and Operation
   Tri Calcium Phosphate       0.304
     Climax, CO, 1937
   (Warnsley & Jones, 1947)
   Bone Char                   0.30
     Briton, SD, 1948
   (Maier, 1953)
   Activated Alumina           0.58
     Bartlett, TX, 1952
*•  (Kaier, 1953, 1970)
   Bone Char                   0.095
     Ft. Irwin, CA, 1954
   (Harmon, 1965)
   Tri Calcium Phosphate       0.144
     Apple Valley, CA, 1961
   (Harmon, 1965)
   Activated Alumina           0.72
     Elsinor, CA, 1960
   (Harmon, 1965)
            Regeneration with NaOH then CO 2 neutralization
            Regeneration operation considered complicated and time consuming
            F~ removal operates "esceptionally well"
            No mention of attrition losses

            Bone char replaced original "Fluorex," Ca,(PO,)2, media
            Fluorax losses were 42 percent/year
            Regeneration with NaOH, then C02 neutralization
            In 1960,  this was the only F~ plant still using bone char
            Plant abondoned in 1971

            Regeneration with NaOH then H2S04 neutralization
            F~ level - 0.5 - 2.0 ppm in plant effluent
            Alkalinity, hardness, pH, S0$~, and F~ all variable in effluent
            F~ concentration in raw water determines F~ capacity of alumina
            Plant abandoned in 1977
            Regeneration with NaOH,  neutralization with
            Semi-automatic regeneration triggered by manual grab samples
            F- removal capacity decreases with each regeneration
            Media had to be replaced annually

            Regeneration with NaOH,  then CC>2 neutralization
            Spent regenerant disposal by evaporation pond
            Ca3(P04>2 fines in treated water
            Initially high attrition losses, later insignificant

            Regeneration with NaOH then H2S04 neutralization
            "Fortunately" treatment  didn't remove H_S
            Pretreatnent = lower pH  from 9.9 to 7.4
            Alumina "cementation" problem encountered
            Abandoned after only a few years of operation
            Replaced by a low F~ supply

-------
                         TABLE 2.  FULL-SCALE PACKED-BED DEFLUORIDATION PLANTS
                                              (Continued)

Process,
Location, Date
(Reference)
Gross
Installed
Capacity
MGD



Process Description and Operation
Activated Alumina
  Gila Bend, AZ, 1978
(Rubel and Woosley, 1979)
0.75      Raw water pH adjusted to pH 5.5
          Regeneration done in two steps
          Upflow and downflow rinses with 1% NaOH solution followed by
             neutralization with 2.5 pH adjusted raw water
          Removal capacity observed 2000 grains/ft3

-------
                                    SECTION 2

                                   CONCLUSIONS

    From these equilibrium experiments it can be concluded that the adsorption
of fluoride onto activated alumina la a function of pH, and that the optimum
pH for adsorption ranges between.5.0 and 6.0.  While early literature reported
maximum fluoride adsorption capacities from 0.95 to 2.3 mg fluorid-2/gm alumina,
at these optltmm conditions, fluoride adsorption capacity reached 10.1 mg
fluoride/gm alumina at 6 ppm fluoride in the water.  The pure-fluoride equil-
ibrum capacities found in these experiments are 20 to 50% higher than the best
column capacities reported in the literature.  The differences are thought to
be due to kinetic limitations and sulfate competition in actual defluoridatlon
processes utilizing packed beds of activated alumina.


    The experimentally determined selectivity sequence was identical to that
reported in the literature, viz, (in order cf preference):

                        OH~ » F- » S02~ » Cl~ > HCOT
                                       4              J
Fluoride adsorption capacity was not significantly reduced by substantial in-
creases in total dissolved solids concentration, except when sulfate was pres-
ent.  Sulfate competition from (from 250 mg/1 So|~) will reduce the equilibrium
fluoride adsorption capacity of activated alumina by as much as 25% when pH *
7.1 and F~ = 5.7 mg/1.

Indications are that the rate of fluoride adsorption on alumina increases as
pH increases especially in the pH range of 6 - 8.

-------
                                   SECTION 3

                                    THEORY
ACTIVATED ALUMINA

     "Activated alumina" is the common name for gamma aluminum oxide  (y-Al^O.),
a porous adsorbent with a moderately high BCT surface area  (150 - 300 m2/gj.
It adsorbs liquids and gases without changing form.  I* is manufactured by
low temperature dehydration at 300 - 700°C of hydrous aluminum oxide, and dif-
fers from high temperature alpha aluulna (a-Al203> in that it readily takes up
water and dissolves in acids (both forms are soluble in strong alkalis).  Acti-
vated aluminas are used as adsorbents in a variety of processes, principally
to dehydrate organic liquids and gases.  Gamma activated alumina was one of
the earliest adsorbents used in inorganic ion chromatography [Kubll, 1947].
It was then known to be quite specific for hydroxide, fluoride, phosphate,
and silicate.
ADSORPTION

     Activated alumina is usually  grouped  with silica, magnesia, and molecular
sieves as a "polar  adsorbent."  Its crystal structure is that of defect spinel
(MgAl204> and contains cation lattice discontinuities giving rise to localized
areas of negative charge.  Since electroneutrallty considerations mandate the
balancing presence of areas of positive charge, sites for both cation and anlon
edsorption are available [Clifford et al.,  1978],  In an expijiment of nitro-
benzene adsorption on activated alumina, Hesse and Saut&r hypothesized that
both the mechanism of "van der Waals" adsorption and ion exchange adsorption
(i.e., salt transformation through the exchange of one ion for another) occur,
and that nitrobenzene is bound on surface locations which are not affected by
pH change.  Kubli (1947] reported that the most important factors affecting
ion exchange adsorption are the nature of the pretreatment and the quantity of
mostly inorganic acids or bases in rhs alumina, while such factors as grain
sire and drying temperature are of minor importance.  He reported that the
anion adsorption behavior of alumina treated with the base, Na2C03, was due to
the precipitation of metal carbonates or hydroxides and that anion adsorption
behavior depended on pretreatment with an acid such at HC1 or HC.IO^.  He pro-
posed the following mechanism*, for anion exchange on alumina.              *

-------
Adsorption-Neutralization of Acid 4X

                      Al-0^^^
                            J>A1'X + H,0
                      Al-0 ---
                               H  + X
                      Exchange of Preferred Ion Y  for X
The above mechanism can take place when the solubility of

                                  Al-0>
                                  Al-0
                                         Al'Y
is lower than
                                 Al-0,

-------
REGENERATION

     According to Kubll, the elution of adsorbed ions from th'e column  (regen-
eration) can be accomplished by three different means:

     1.  By alkali elution (reversal of the column)

     2.  By elution with anlons which form less soluble basic aluminum salts

     3.  By higher concentration of the anions originally bound to the column

     Kubli also expanded the selectivity sequence of anion adsorption on alum-
ina originally established by Schwalb and Dattler [1937].  The sequence of
Important aninns in the expanded series is as follows:

                OH~ » P0?~ » F~ » SO?" » Cl~ » NO"
                        44            3

     Clifford et al [1978] have suggested a simplified series of chemical reac-
tions to explain the ion exchange adsorption of fluoride and the subsequent
regeneration of the packed bed of fluoride-exhausted alumina:

                         SIMPLIFIED PICTURE OF ALUMINA
                     ADSORPTION AND REGENERATION REACTIONS

                1.  NEUTRAL ALUMIKA

                         Alumina + HOH 	>•  Alumina -HOH

                2.  ACIDIFICATION

                         Alumina-HOH + HC1 	* Alumina-HC1 + HOH

                3.  ION EXCHANGE IN ACIDIC SOLUTION

                         Alumina-HC1 -f NaF 	>• Alumina-HF + NaCl

                4.  REGENERATION

                         Alumina «HF + 2NaOK 	>• Alumina-NaOH + NaF + HOH

                5.  ACIDIFICATION

                         Alumina-NaOr. + 2HC1 	» Alumina-HC1 + NaCl + HOH

-------
CHARGE DEPENDENCE

     Many researchers in the fields of soil and minerology have studied the
presence of electric charges on the oxide minerals suspended in aqueous solu-
tions.  These charges comprise an important consideration when determining the
extent to which the solids to which they are attached can function as adsor-
bents in ion exchange columns [Amphlett, 1964; Bolt, 1965; Van Olphen, 1963].
The iscelectric point IEP (s) and the zero point of charge (ZFC) are convenient
references for predicting charge dependence behaviors [Parks, 1967].

     Parks defined the ZPC as the pH at which solid surface charge from all
sources is zero.  The IEP(s) is a ZPC arising from the interaction of the
solid with H+ and OH~, i.e., the interaction of the solid with water.  For
other species, the ZPC will vary with the ionic composition of the system.
Data on pHx£ps and pHgpc for most minerals [Parks, 1965] are available else-
where.  Values for the pHtEPs of activated alumina have been determined [Choi,
1979] and are presented in Table 3i aging in DI water raises the pHlEPs.

     In general, the sign of the surface charge Itself will determine the
anionic or cationlc adsorption behavior.  Thus aluminum oxide may be either
a cation exchanger (in alkaline medium) or an anlon exchanger (in acid medium),
and within a certain pH range, can act as an amphoteric ion exchanger [Zhabrovs,
1961].  While surface charge is by no means the only factor responsible for
electrolyte adsorption, the extent of adsorption decreases rapidly when the
sign of the oxide's surface charge is changed to that of the sorbing species.
However, Umland [1959] pointed out that in cases where the removal of ions
cannot be entirely attributed to either ion exchange or molecular sorption, it
may be caused by the precipitation of sparingly soluble bases and basic salts.

     By comparison, adsorption caused by electrostatic attraction alone  can
be considered not-specific or non-selective.   Adsorption can also arise from


               TABLE 3.  ISOELECTRIC POINT OF ACTIVATED ALUMINA

        Treatment                                           pHIEPs

        a.  DI water wash, stored in DI
            water for two days                               6.2

        b.  DI water wash, stored in DI
            water for seven days                             5.8

        c.  DI water wash, 2N-HC1 wash,.
            followed by DI wash, stored
            in DI watet. for seven days                       7.3

        d.  DI water wash, 2N-HC1 wash,
            followed by DI wash, stored
            in DI water for ten days                         8.9
                                      10

-------
electrostatic attraction augmented by hydrogen bonding, coordinate bonding, or
London Van der Waals bonding.  Adsorption under the combined influence of ionic
and non-ionic bonding is called "specific adsorption," which can occur even
when the surface is uncharged.

     At ZFC, anlon and cation adsorption is minimal [Parks, 1967].  Specific
adsorption reverses the sign of the surface charge and shifts the ZPC [Kings-
ton, 1972].
                                       11

-------
                                   SECTION 4

                             METHODS AND MATERIALS
OBJECTIVES

     Adsorption capacity of activated alumina,for the fluoride Ion Is reported
to be affected by pH [Fink and Lindsay, 1936], fluoride concentration [Maier,
1953], and the relative concentration of competing Ions, especially sulfate
[Kubll, 1947; Umland, 1956; Hess and Sauter, 1947; Rubel, 1979).  Nevertheless,
the literature reveals that most studies of fluoride adsorption by alumina have
been conducted either at uncontrolled pH values or in the presence of competing
buffers, thus complicating interpretation of the results.  In this report,
experiments have been carried out specifically to determine, both quantita-
tively and qualitatively, the effect of numerous factors on fluoride concen-
tration, competing ions, Including su.'.fate, chloride, bicarbonate, and high
ionic strength.


PROCEDURES

Conditioning the Alumina

     1.  A 200-ml sample of F-l grade Alcoa activated alumina of mesh size
26-48, Lot 4630-18, was rinsed with water and >ec&nted several times until
all the suspended and fine particles were removed.  (Typical properties of
this product are shown in Table 4.)

     2.  The washed sample of alumina was trickled down into a 2.1-mm ID
conditioning column (See Fig. IB) containing deionized water, in such a
way that no air bubbles remained or were entrapped in the settled alumina.

     3.  The column was eluted with tap water (which had been adjusted to pH
3) until the effluent pH dropped down to pH 5.  The elutlon was concluded
with tap water whose pH had been adjusted to exactly 5.0 with 0.01 N
until equilibrium was attained at the column effluent.
                                  rv,
     4.  The conditioned, acid-washed (pH 5) alumina was dried at IIO'C in
an oven for 48 hours.
Exhausting the Alumina with Standard F~ Solutions

     1.  1.00 g of the dried conditioned alumina was placed in each of ten

                                      12

-------
              TABLE 4.   TYPICAL PROPERTIES AND SPECIFICATIONS OF
                          ACTIVATED ALUMINA,  F-l TYPE
              Constituents and Properties
Content
A1203
Na20
Fe2°3
S102
Loss on ignition (1100°C)
Form
Contact surface area (sq.m./g)
Size (mesh)
Bulk density loose (lb/ft3)
Bulk density packed (lb/ft3)
Specific gravity
92.00Z
0.90Z
0.08Z
0.09S
6.501
Granular
210
26 - 48
52
55
3.3
11 mm ID standard wall pyrex mini-columns.  No air bubbles were entrained in
the settled alumina.  See Fig. LA.

     2.  The mini-columns were equilibrated by passing the standard fluoride
solutions through them at un average flow of 0.5-4.0 ml/nin.   The effluent
fluoride concentration was checked from time to time.  It took from five to
six days of continuous operation to achieve more than 90 percent equilibrium,
i.e., the effluent F~ concentration ^0.90 times influent F~ concentration.
See Fig. 2 for typical mini-column set up.

Regeneration of Fluoride-Spent Alumina

     1.  The equilibrated columns were disconnected and the solution in the
columns drained to a point Just above the surface of the alumina.

     2.  Each column was eluted with 100 ml one percent NaOH (.25 N) during
a period of approximately four hours followed by 500 ml deionized water wash.

     3.  The regenerant solution was collected and diluted to exactly one liter.

NOTE:  It had been previously determined that about 95 percent of the fluoride
on the column could be recovered by this procedure.
                                     13

-------
40
cm
 I
                                   j-— 11 mm ID Std. Wall
                                    ~|~]   Pyrex Tubing
• • •£
. t
,
m urn ft
A
t(
/ 2


                               15
                               cm
                               V
                               t
                             6 cm
                                         -V * U ml

                                         "Activated Alumina to be Exhausted
                                       6— Pyrex Wool

                                       <—3/8" OD Tubing

                                         <—  Pinch Clamp

                                         — Eye Dropper Tip
                                 A.  Mini Column for Capacity  r>**formin.ilion
                   Activated Alumina
                   to be Conditioned

                   21 mm ID Standard Wall Pyrex Tubing
                       Coarse Pyrex Glass

                    10 mm OD

                    7/16" OD x 1/16" Wall, Gum Tubing

                        -  Pinching Clamp


                  10 run OD Tip


              B.   Alumina Conditioning Column

Figure 1.  A.  Mini column for capacity o.at«rtnination.

           B.  Alumina conditioning column.

-------
     Plastic Tubing
Tee Connection
For Influent Sample
     Mini-Column

       Class Wool
   2 Liter Jar
   for Effluent
                                                 5 Cal. Adsorbate
                                                 Solution Jar
                                                 Column Stand
1 gm Pretreated
Alumina
        Figure 2.  Typical Alumina Mini-Column Set Up for
                         Adsorption Run
       (Represents one of ten columns used simultaneously)
                              15

-------
Analysis of Regenerant Solution

  2_ 1.  The regenerant solution was analyzed to determine the amounts of F~,
804 , Cl~, and HCOj adsorbed on the alumina.  The amount of fluoride adsorbed
on the various alumina columns was thus determined as a function of pH, SOJ,
Cl~, and HCO-j concentrations, and ionic strength.


METHODS AND MATERIALS

Feedwaters Used

     pH-adjusted Houston tap water was used as a background solution for con-
ditioning the alumina.  Deionized water was used to make standard solutions
which were passed through the alumina columns for the fluoride capacity
d<» terminations.
Fluoride Source

     Sodium fluoride (reagent grade) was used to make standard fluoride stock
solutions throughout the experiments.


pH Effects

     To study pK effects on adsorption, influent fluoride standard solution
was adjusted with diluted ^SO^ (for lowering the pH) or with diluted NaOH
(for raising the pH).  In addition, sodium carbonate solution was used as a
buffer while preparing the standard fluoride solutions to maintain constant
pH values of 7 and 8.


Standard Fluoride Concentrations Used

     Fluoride concentrations of the standard solutions prepared with deionized
water were diluted to 2.00, 4.00, 6.00, 8.00, and 10.00 mg F~/l.

                                       2—    —
Concentrations of Competing Anions:  SO, , Cl

     To determine effects of competing ions, the fluoride concentration of all
the standard solutions was kept fixed at .3 meq F~/l (5.7 mg F~/l).  The con-
centrations of competing Ions used weze 0.50, J.00, 5.00, 10.0, and 15.0 meq/I
in the influent standard solution.  The pH was fixed at 6.0 by adding 1 1/2
drops of N/20 NaOH for each two-liter solution.


High Ionic Strength Solutions

     The high ionic strength standard solution consisted of Na"*", F~, HCO^, Cl~,
and So|~ ions in deionized water.  Some high ionic strength experiments were

                                     16

-------
run at pH - 7.1 and F~ - 0.3 meq/1 (5.7 ppm) and others were run at pH -  7.1
and F~ * 0.5 meq/1 (9.5 ppm).  To produce the background ionic strength solu-
tions, equivalent mixtures of Cl~, HCOj, and S0j£~ were utilized to yield  solu-
tions of I - 3.8, 8.3, 14.3, 28.3, and 56.3 millimoles/liter.

     See Tables 5, 6, and 7 for a complete analysis of the water for each
experimental run.
                                      17

-------
              TABLE 5.   COLUMN DATA FOR DETERMINATION OF THE EFFECTS OF pH ON FLUORIDE ADSORPTION
                               CAPACITY IN pH ADJUSTED DEIONIZED WATER SOLUTIONS
oo
Fluoride Concentration Other Anions Present;
Run
Not
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30

mg/1
2
2
4
4
6
6
8
8
10
10
2
2
4
4
6
6
8
8
10
10
2
4
6
8
10
2
4
6
8
10
* A "run" IB the
1 meq/1
S0| - 48

meq/1
.105
.105
.210
.210
.316
.316
.421
.421
.526
.526
.105
.105
.210
.210
.316
.316
.421
.421
.526
.526
.105
.210
.316
.421
.526
.105
.210
.316
.421
.526
exhaustion
tng/1 S0$

PH
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
7.3
7.3
7.3
7.3
7.3
8.0
8.0
8.0
8.0
8.0
S0*~
meq/1
.005
.005
.005
.005
.005
.005
.005
.005
.005
.005
.006
.006
.006
.006
.006
.006
.006
.006
.006
.006
2.500
2.438
2.500
2.500
2.250
.625
.625
.500
.687
.812
HC03
meq/1
-
_
_
-
-
_
-
-
-
-
_
-
—
-
-
_
-
-
-
-
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
Q
ml/min
4.18
3.63
2.13
3.09
4.01
2.45
3.54
2.11
3.91
2.67
.68
.93
.44
.77
.46
.39
.41
.36
.39
.76
1.47
1.61
.83
1.44
.54
.86
1.46
1.63
1.39
.83
t.90
hrs
104
87.2
90.5
64.7
24.9
62.6
33.0
51.1
21.3
40.0
244
244
224
137
152
155
131
112
110
53.5
67.7
41.3
52.1
30.1
67.1
53.9
27.2
19.3
26.3
31.9

T.9 m{
6.29
4.47
4.67
4.87
3.31
5.33
5.34
5.00
4.75
5.93
2.44
3.13
2.49
2.58
2.48
2.12
2.60
1.87
2.42
2.33
3.49
3.08
2.50
2.96
2.92
2.28
2.58
2.62
2.65
3.03
Fluoride Capacity

5 F~/g Al
8.27
8.51
9.85
9.85
10.89
10.36
10.49
10.40
10.53
10.80
8.20
8.70
9.63
9.91
10.30
10.30
9.85
10.50
10.70
10.5
3.44
5.20
6.25
7.03
7.54
2.45
3.72
4.35
4.82
5.28

meq F~/g Al
.435
.448
.513
.518
.573
.545
.552
.547
.550
.570
.431
.458
.507
.521
.542
.537
.518
.553
.563
.553
.181
.274
.329
.370
.397
.129
.196
.229
.254
.278
of one single column
- 50 mg/1
as CaC03

1 meq/1
HC03
• 61 mg/1
HC03 - 50
mg/1 as CaC03

-------
TAPLE 6.  COLUMN DATA FOR DETERMINATION OF THE EFFECTS OF SULFATE AND CHLORIDE
                 ON FLUORIDE ADSORPTION CAPACITY AND af AT pH - 6.
Liquid
Phase.Concenfration
Run
No.
15
31
32
33
34
35
36
37
38
39
40
NOTES:
1.
2.
3.
4.
Q
ml/min
.46
.89
.78
.72
1.66
1.57
.88
.99
.59
.84
.62

c.9
hrs
152.0
104.0
96.3
85.4
34.1
36.1
75.7
66.8
111.0
79.1
107.0

T.9
2.48
3.69
2.99
3.85
2.74
3.02
2.21
2.11
2.13
2.19
2.07

S0j£~
meq/1
.006
.5
1.0
5.0
10.0
15.0

-
-
-
-

ci-
meq/1
-
-
-
-
-

.5
1.0
5.0
10.0
15.0

Phase
F-
meq/1
.542
.454
.451
.390
.371
.367
.542
.568
.563
.547
.578

Solid
Concentration
meq/1
.002
.106
.133
.290
.246
.320
ci-
meq/1
0.047
0.003
0.069
0.034
0.073

Total
.544
.560
.584
.680
.617
.687
.589
.571
.632
.581
.651

F
.208
7.14
11.4
22.4
50.2
57.34
19.2
631
136
264
545

A run is the exhaustion of one single column.
For runs 31 - 40, F~ concentration - 5.7 mg/1 (.300 meq/1).
For run 15, F~ concentration - 6.0 mg/1 (0.316 meq/1).
a^ - separation factor of F/C1 or F/SO,. See Appendix for calculation.

-------
                  TABLE 7.  COLUMN DATA FOR DETERMINATION OF THE EFFECTS OF IONIC STRENGTH ON

                                 FLUORIDE ADSOivPTION CAPACITY AT pH - 7.1
           Fluoride
Other Anlons Present
Run
No.
*
41
42
43
44
45
46
47
48
49
50
Concentration
ng
F-/1
5.7
5.7
5.7
5.7
5.7
9.5
9.5
9.5
9.5
5.7
meq
F-/1
.3
.3
.3
.3
.3
.5
.5
.5
.5
.5
PH
7.1
7.1
7.1
7.1
7.1
7.1
7.1
7.1
7.1
7.1
SO?'
meq/1
1.0
2.0
4.0
8.0
16. 0
1.0
2.0
4.0
8.0
16.0
Cl"
meq/1
1.0
2.0
4.0
8.0
16.0
1.0
2.0
4.0
8.0
16.0
HCOo
meq/1
1.0
2.0
4.0
8.0
16.0
1.0
2.0
4.0
8.0
16.0
I
mmolea
1
3.8
8.3
14.3
28.3
56.3
4.0
8.5
14.5
28.5
56.5
Q
ml/min
1.22
.78
1.27
.76
1.11
1.21
.75
.72
.90
1.22
T.9
4.38
3.95
3.63
3.48
3.61
3.48
3.71
3.60
3.24
2.93
C.9
75.06
106.69
56.11
87.68
54.22
43. °7
f ,.45
69.21
48.00
29.97
Fluoride
mgF~/g Al
7.16
7.21
6.76
6.55
5.68
8.73
7.94
7.91
7.62
7.13
Capacity
meqF~/g Al
.377
.379
.356
.345
.299
.459
.418
.416
.401
.375
K>
o
    * A run Is the exhaustion of one single column

-------
                                  SECTION 5

                                   RESULTS
COLUMN EFFLUENT HISTORIES

     Effluent fluoride concentration versus volume of feedwater, I.e., column
effluent histories, were recorded for all the exhaustion experiments and are
shown In Figs. Al - A6.  In the first two sets of early experiments (exhaus-
tion of 20 columns), pairs of columns were run In an Identical manner In
order to check the precision of the revolts.  The repeatability proved satis-
factory.

     The adsorption process for fluoride on activated alumina was observed to
be very slow.  It took over 120 hours to reach equilibrium.  It has been fur-
ther noted that the adsorption rate  Is  relatively fast In the beginning and
slowly decelerates to an asymptotic equilibrium.  The area above the curves
shows the accumulated fluoride adsorbed onto alumina.  The effluent rate was
not kept constant for any of the column pairs; to do so would have Involved
additional monitoring equipment, and the investigation of the kinetics of
adsorption was not the primary intention of these tests.  Nevertheless, the
rate of flow was recorded for each column, and is depicted in column data,
Tables 5 to 7.
EFFECT OF FLUORIDE CONCENTRATION

     Results indicate that solid phase adsorption capacity is dependent upon
the concentration of adsorbate in solution.  Adsorption capacity increases
with adsorbate concentration, while the rate of increase in adsorption capa-
city at lower concentrations is much higher than the rate of Increase at
higher concentrations.

     For the first four sets of experiments in which pH had been kept fixed
at 5, 6, 7, or 8, the influent fluoride concentration for each column was
kept at 2, 4, 6, 8, or 10 mg/1.  Equilibrium fluoride concentration effects
are revealed in adsorption Isotherms (Figs.  3  to  6).  As expected, the
total fluoride adsorption capacity of alumina increased with increasing equi-
librium fluoride concentration.  Regarding the rate of adsorption, however,
Figs. Al - A6 illustrate that fluoride uptake slows down as available adsorp-
tion sites are filled.  As equilibrium is approached, the number and avail-
ability of adsorption sites decrease thus the adsorption rate decreases.

     The adsorption isotherms obtained are of the Langmuir type.  Average

                                      21

-------
      11
      10
 e
 o
o
CO
                                             I
                                             6

                                            mg
8
10
                       Liquid Phase Fluoride Concentration



                    FIGURE 3.   FLUORIDE ADSORPTION ISOTHERM




                     Adsorbate pH * 5.0, temperature - 22°C
                                   22

-------
      11
      10
o
c  <

i  &
o  •-»
*  t?
•H  B

O
=

-------
2   <  6
s   g,
4)   —
o
U
o
3
0)
CD
(0

o.
          '6.2
    00
                     2.0
4.0
      6.0

C   mg F"/l
8.0
10.0
                    Liquid Phase Fluoride Concentration


                 FIGURE 5.   FLUORIDE ADSORPTION ISOTHERM



                  Adsorbate pH - 7.3, temperature - 22*C
                                    24

-------
        7
        6   •
o
u
C



s.


"3
o
en
5   •
    O.



    E
        3   •
  6 _

mg F"
                                                            8
                                                                7.0
                         Liquid  Phase Fluoride Concentration


                      FIGURE 6.    FLUORIDE ADSORPTION ISOTHERM


                       Adsorbate pH • 8.0, temperature • 22°C
                                      25

-------
fluoride adsorption capacity on activated alumina obtained from these experi-
ments was 10.2 mg/g at pH 5 and pH 6, five times greater than what has been
reported in  the literature [Savinelli and Black, 1958] but in general agree-
ment with the results of Choi and Chen [1979] for essentially pure fluoride
solutions.
EFFECTS OF pH

     The effects of pH on the adsorption of fluoride by activated alumina are
shown in Fig. 7.  For a fixed fluoride concentration, adsorption remains
almost constant within the pH range 5.0 to 6.0 after which the adsorption
falls steeply until pH 7.5, then continues to decline more gradually as pH
is increased.

     Experiments could be conducted only between pH 5 and 8.  Below pH 5,
alumina dissolves to form aluminum complexes such as A1F2+, A1F£, AlFo, and
A1F£ (pK 6.16, 5.05, 3.91, and 2.71, respectively).  Hence, below pH 5, the
column would never reach equilibrium until all the adsorbent dissolved to
form llgand compounds.  On the other hand, above pH 8, predominant hydroxyl
ions are preferred over any other ions for adsorption onto alumina beds.

     At pH 5 to 6, in the relative absence of hydroxyl and other competing
anions, conditions were most favorable for fluoride to occupy adsorption sites.
Equilibrium adsorption at pH 5 and 6 has been found to be 9.9 and 10.3 mg F~/g
Al, respectively, for adsorbate concentration of 6 mg F~/l.

     As the pH was increased above 7, the hydroxyl anions, though present only
in small quantity (10~? moles/1), competed for adsorption sites and consider-
ably reduced the adsorption capacity of alumina at equilibrium by 30 percent.


EFFECTS OF S0*~ IONS
                       2-
     Though sulfate (504 ) ions have been found to be less preferred than
fluoride ions in competition for adsorption sites, the presence of sulfate
had a pronounced effect upon fluoride adsorption, as reported in Fig. 8.
As.the sulfate ion adsorbate concentration increased from .5 to 15 meq/1,
SO/,~ ion adsorption increased from .1 to .3 meq S0j~/g Al, while fluoride
adsorption waa reduced from .46 F~/g Al to .38 meq F~/g Al—a reduction of
about 17 percent.  By comparison, It should be noted with respect to the
removal of fluoride from municipal wa£er supplies that where fluoride contam-
inant levels are found in excess, S0£~~ let) concentration generally ranges
around 250 mg/1 or 5.2 n,c?q/l.  Extrapolating from the above data, fluoride
adsorption under normal circumstances could be reduced by up to 26 percent.
                                       26

-------
     .6  .
     .5  -
   «
   8
01
T  O-


31  E
"O      -I

Tl  ^
c
      .2
      .1
         4.0
5.0
6.0
7.0
8.0
                                         pH


                   FIGURE  7.   pH EFFECT ON FLUORIDE ADSORPTION
                                     27

-------
            .5
                                               F~ Adsorption  In Presence of  Cl
OO
     O
a
u
O
09
•O


4)
a
a
x
Ou

•o
     O
     to
            .4
          o
          or
          4)
          e
            .3
                                    F  Adsorption in Presence of  SO,
                                                                    2-
SO ~ Adsorption at 0.3 meq/1 F"
                                                                   meq/1
                                                                           10.0
                          FIGURE  8.
                                                                                          532 mg/1 Cl

                                                                  2-                      720 me/1 SO?"
                                    Liquid  Phase  Concentration  SO,   or Cl                     I


                                 EFFECTS OF CHLORIDE AND SULFATE ON  FLUORIDE ADSORPTION ISOTHERMS
                                      Adsorbate pH  - 6.0,  F~ Cone.  - .3 meq/1 (5.7 rag F /I)

-------
 CHLORIDE EFFECTS

      The effects of the presence of chlorides (.5 meq/1 to 15 meq/1) on fluo-
 ride adsorption was also tested, and the results are depicted in Fdg.  8.   The
 adsorbate acidity was fixed at pH 6 and fluoride concentration was kept con-
 stant at .3 meq/1 throughout the experiment.  As expected, no significant
 change in fluoride adsorption was observed, demonstrating that the fluoride
 ion is far more preferred for adsorption by alumina than chloride ions.  This
 result was predicted by Umland, whose preferential series presumed chloride
 to be much less preferred than fluoride which in turn was only slightly less
 preferred than hydroxide and phosphate for anion adsorption.

      A comparison of Figs. 9 and 10 reveals that, whereas variation in SO^
 concentration affected the fluoride adsorption kinetics to a noticeable extent,
 variation iu chloride concentration did not.  This is contrary to the results
 of Choi and Chen [1979].

      Referring to Table 6, it should be noted that although chloride adsorp-
 tion remained relatively constant with increasing chloride concentration,
 the amount of fluoride adsorbed slightly increased so that the sum of anionic
 species adsorbed onto alumina increased with increasing anionic concentration.
 Furthermore, the total anionic adsorption of the sulfate plus fluoride solu-
 tion was almost the same as that of chloride plus fluoride solution, suggest-
 ing that increasing the concentration of a much less preferred ion might
 increase the adsorption of the more preferred ion.  However, this effect
 needs to be researched in greater detail.
 HIGH IONIC STRENGTH EFFECTS

      Fluoride contaminated groundwater found in nature tends to contain levels
 of dissolved solids, consisting of a variety of cations and anions.  The
 anions are mostly sulfate, chloride, bicarbonates (or carbonates), and nitrate.
 Another set of experiments was conducted to determine the effect of high
 ionic strength adsorbate solution on fluoride removal by activated alumina
 adsorption.  Adsorbate solutions were prepared in deionized water with equiva-
 lents of sulfate, chloride, and bicarbonate anions.  The sum of chloride,
 sulfate, and bicarbonate concentration (I Cl~ + 50^  + HCOp was kept at 3,
. 6, 12, 24, or 48 meq/1.  The ionic strengths of these solutions were 4.0,
 8.5, 14.5^ 22.5, or 56.5 mmoles/1.

      The effect of ionic strength on fluoride adsorption is reported in Figs.
 11,  12&13.  Fluoride adsorption decreased with increasing ionic strength.  An
 analysis of species adsorbed onto the alumina indicated that the reduction in
 fluoride adsorption was caused by the S0£~ Jon species, while chloride and
 bicarbonate had an insignificant effect.
                                       29

-------
   6.0
   4.5
'if
o
§
o
   3.0
w
   1.5
   0.0
3 raeq/1 F  + .5 meq/1
S0?~ In Feed, t 9 -
104 hrs
3 meq/1 F  + 1 meq/1
SO?" In Feed, t
96 hrs
                                                    3 meq/1 F~ + 5 meq/1 SO
                                                    In Feed, t
                                            3 meq/1 F  -I- 10 meq/1 SOT  in Feed,
                   1500        3000       4500         6000

                            BED VOLUMES (BV) OF EFFLUENT
         7500
9000
              FIGURE 9.  MINI-COLUMN EFFLUENT CONCENTRATION HISTORIES, pH - 6.0
                         EFFECT OF S0^~ CONCENTRATION ON COLUMN KINETICS
                        Q - 1.1 ml/mln,  EBCT • .88 mln,  BV - 1.0 ml
                         t _ » time to 90% of equilibrium in hours

-------
~  A.5
60
e
u
§
u
   3.0
w
   1.5
   0.0
          90 hrs
               Feed Solutions Contain
                  .3 meq/1 F- -I- Cl~
                  O  .5 meq/1 Cl~

                  Q   1 meq/1 Cl~
                  &   5 meq/1 Cl~

                  G|   10 meq/1 Cl"

                  V   15 meq/1 Cl"
                 1500
3000
4500
6000
7500
                         BED VOLUMES (BV) OF EFFLUENT
9000
           FIGURE 10.  MINI-COLUMN EFFLUENT CONCENTRATION HISTORIES, pH - 6.0
                     EFFECT OF Cl~ CONCENTRATION ON COLUMN KINETICS
                     Q -0.78ml/mln, EBCT - 1.2 mln, BV - 1.0 ml
                      t _ - time to 90X of equilibrium in hourK

-------
                      6.0
U)
                                                                       .9
                                                                   -  106 hra
         (See Fig. A 7 for 9.5 ppm F )
         5.7 ppm f  Feed Solutions Contain
                                                    O E HC03 + SOT  + Cl  - 3 meq/1,  I  -  .0038  M
                                                        HCO~ + SOT + Cl~ - 6 meq/1,  I  -  .0083  M
                                                    0 E HCO~ -f S0,~ + Cl~ - 12 meq/1, I - .0143 M
                                                    Ki E HCOj + SO|~ -»- Cl  -  24 meq/1,  I  -  .0283  M
                                                  .  0E HC03 f sofr* + Cl",-  48 meq/1.. I  -  .0563  M
                                    1500
    3000       4500        6000
BEN VOLUMES (3V) 0V EFFLUENTS
7500
9000
                               FIGURE  11.  MINI-COLUMN KFFLUiZNT roNCENTRATION HISTORIES,  pH
                                          EFFECT OF  IONIC  STRENGTH ON  COLUMN KINETICS
                                          Q - 1.0 ml/min,  EBCT -  i min, BV - 1.0 ml
                                          t _ • time to 90% of equilibrium in hours
                                         See  Fig.  A 7  for  9.5 ppm F*
                                                 7.1

-------
tit
           o

           M
           O
           O
           v.
           
-------
 o
 +

'V1
 +
o  er
o.
u
o
0)
n)
Ou
O
to
       .7
      .6
       .5



       .4

        0
                     2.0
                     100
                                  4.0
                                  20G
6.0
300
8.0
400
10.0
 500
12.0
 600
14.0
 700
mgCaC03/l
                                                               .2-
                                                                         HCO.
                                        Liquid Phase Cone, of SO ~, Cl~, n^v/

                                       FIGURE  13.  ANIONS ADSORPTION ISOTHERM

                                     Adsorbate pH - 7.1, F  cone. » .3,  .5 meq/1

-------
                                  SECTION 6

                                 DISCUSSION
     The exact phenomenon by which fluoride is removed from water by acti-
vated alumina has yet to be conclusively described.  Many authors [Savinelli
and Black, 1958] regarded fluoride adsorption as an ion exchange phenomenon
while others [Kingston, 1972; Parks, 1967; Muljadi, 1966] considered it typi-
cal of adsorption behavior by metal oxides, an aqueous surface phenomenon.
However, the information derived from these experiments—concerning the depen-
dence of fluoride adsorption upon initial fluoride concentration, pH, com-
peting anions, and ionic strength—may provide a new foundation for future
studies of the kinetics of activated alumina adsorption.


ACTIVATED ALUMINA F~ ADSORPTION CAPACITY

     In general, the literature (most notably, Churchill, 1936; Fink and
Lindsay, 1936; Swope and Hess, 1937; Zabban, 1967; Maier, 1953) reports much
lower capacities of F~ adsorption on activated alumina than were here obtained.

     Savinelli and Black, in their bench scale experiments, showed a capacity
of 8.8 mg/g Al (3400 grains/cu ft*) at pH 5.6 for Ce - 10 ppm F~.  In recent
articles, Yeun C. Wu [1978], and Rubel and Uoosley [1979] reportedly attained
an optimum adsorption capacity for fluoride on activated alumina of 2620
grains/cu ft Al (6.8 mg/g), and 3200 grain/cu ft Al (8.3 mg/g) respectively.
In Rubel and Woosley's pilot scale column experiments, the F~ concentration
was 6.0 mg/1 and the pH was controlled between 5 and 6.  The experiments
reported here indicate that, in the absence of competing ions and when
acidity is strictly controlled between pH 5.0 and pH 6.0, adsorption capacity
can reach as high as 10.1 gm/g Al (3894 grain/ft-*) for an influent F~ concen-
tration of 6 ppm (a high but not uncommon concentration for contaminated under-
ground municipal water supplies).  Tne significantly lower values for the
maximum adsorption capacity reported in the early literature may be attributed
to a number of factors, such as uncontrolled pH, equilibria obtained from batch
reactors, or studies conducted In continuous flow columns which were termin-
ated for practical considerations at some arbitrary fluoride concentration
breakthrough and never allowed to reach equilibrium.  It is not being sug-
gested here that in field operation, one should always expect these maximum
capacities to be achieved.  The equilibrium capacities are offered as a goal
for ideally efficient processes.


*Note:  Above conversions based on packed alumina density of 55 Ib/ft .

                                     35

-------
'*FLUORIDE ADSORPTION ISOTHERMS

      The shapes of the  isotherms describing  fluoride  adsorption  equilibria
 were largely determined by  such factors  ar? the  relative  affinity of  the solute
 for the solid  surface;  the  number  of  sites available  for adsorption;  and
 the interactions of the adsorbed molecules.  The  F~ adsorption isotherms
 obtained are typical of weak base  synthetic  resins.   They satisfy the Lang-
 Inulr mathematical model well compared to the Freundlich  model; the average
 correlation coefficients  r's are .994 and .971, respectively.  Other  constants
 and isotherms  are reported  in Table 8, and Figs.  14 through  17.   In his explan-
 ation of the selectivity  coefficient  for ion exchange, Clifford  [1978]  reiter-
 ated that  the  Donan membrane equilibrium theory,  the  law of  mass action, and
 the Langmuir isotherm assumptions  all yield  the same  equilibria  expressions.
 For binary isotherms,   the  Langmuir multicomponent equilibria theory  yields a
 separation factor constant, a, such  that
                                  b,C,
                          Q'bBcB
 where  Q1  - Langmuir ultimate  solid phase  adsorption  (or  Ion  exchange)  capacity
        b.  • Langmuir constant  related  to adsorption energy

 Taking OH~ and  F~  Ions as  the two components,  it may be  noted  that  the rela-
 tive affinity of alumina for  OH~ ions is  so disproportionately large  that  the
 separation factor  approaches  infinity.  As a result, initial acid condition-
 ing and regeneration of the alumina after exhaustion are essential.


 EFFLUENT  CONCENTRATION HISTORIES (Appendix Figs. Al  - A6)

      Further analysis of effluent concentration histories obtained  for each
 mini-column sheds  light on several important kinetic relationships.

      Considering throughput at 90 percent equilibrium  (T.g - meq F~ ions
 passed/meq F~ capacity on  Al)  and t.g time for 90 percent exhaustion  (both
 of which  are flow  rate dependent for  a given pH and  ionic strength—Tables 5
 through 7), t.g is inversely  proportional to the influent F~ concentration
 (Fig.  18) while T.g throughput is Directly proportional  to the rate of load-
 Ing.   Since a slower flow  rate maximizes  adsorption, while the practial
 limitations of  treatment plants  (such as  reactor size  and effluent  demand)
 press  for a faster flow, a compromise must be  accomplished based upon the
 rate of adsorption, which  requires a  detailed  knowledge  of kinetics.   Ne\ar-
 theless,  these  experimental results do suggest that  the  kinetics of fluoride

                                       36

-------
TABLE 8.  FLUORIDE ADSORPTION ISOTHERM DATA AT 22°C

pH


5




6




7




8



Ce
mgF-/l
2
4
6
8
10
2
4
6
8
10
2
4
6
8
10
2
4
6
8
10
! Correlation K n Correlation
mg/g 1/mg r mg/g g/1 r


11.62 1.320 .9867 7.778 6.683 .543




11.24 1.539 .9899 7.873 7.475 .958




10.71 0.236 .9999 2.527 2.037 .994




7.41 0.239 .9995 1.832 2.122 .992



-------
    .12
    .11
00
B
 41  .10
    .09
                                pH - 5.0
                                1/Q' -  .086 g/mg*
                                1/bjQ*  =  .065 g/1
                     pH - 6.0
                     1/Q'  - .089 g/mg
                     1/biQ1 - .058 g/1
    .08
                   .1
.2
                                   1/C
 .3
1/mg
.5
                  FIGURE  14. LANGMUIR  ISOTHERM  FOR  pH  5,  pH 6

-------
CO
VO
                           .3
                        60
                        O
                           .2
                                                pH - 8.0
                                                i/Q
                                                1/b
                           .1
                                                                                         a
                   pH - 7.0
                   1/Q' - .093 g/mg
                   1/b Q' - .39 g/1
.2
                                                                 .3
                                                       1/C  1/mg
.4
                                    .5
                                       FIGURE  15. LANGMUIR ISOTHERM FOR pH 7, pH 8

-------
 1.0
   .8
eo

00
e
  .6
00
o
•-I
  .A
          pH - 6.0     	

          log K - .986 mg/g

          1/n = .134 1/g
                                                                 891 mg/g
                                                        1/n - .149 1/g
        .3
.4
.5
.8
                 .6          .7


                 Log Ce  mg/1


FIGURE 16.  FREUNDLICH ISOTHERM FOR pH 5, pH 6
                                                                                .9
                                                                        1.0

-------
  1.0
   .8
oo

£
   .6
or
00
  .4
                        pH -  7.0
                        log X =• .403 mg/g
                        1/n - .491  1/g
                                                                       pH - 8.0
                                                                       log K - .263 mg/g
                                                                       1/n - .47 1/g
  .21—V
                              .5
                                                      .7
                                                                  .8
              .6

              log C   mg/1

FIGURE 17.  FREUNDLICH ISOTHERM FOR pH 7, pH 8
.9
                                                                                          1.0

-------
                                        O   O 2
                                                  ppm
                                                4 ppm
                                                6 ppm
                                               10 ppm
                                       Cone, of F~ Feed
FIGURE 18.  EFFECT OF pH ON TIME FOR 90* ADSORPTION EQUILIBRIUM
              T • Throughput for given data point
              T « number shown adjacent to each point
                        A2

-------
adsorption on alumina (26 - 48 mesh particles) are very slow—48 hours to reach
50 percent exhaustion of the column and 120 hours to 90 percent exhaustion.
Also, in the pH range of 6 - 8, rate of F  adsorption increases, as pH increases.

COMPETING IONS

     In concentrations at which they are normally found in drinking water sup-
plies, none of th« anlons tested (S0$~, Cl~, HCOj) except for OH~ significantly
decreased the fluoride adsorption capacity of activated alumina.  However, at
higher ionic strengths,-fluoride adsorption capacity was slightly affected,
primarily because of SO^  competing ions, indicating that adsorption onto
alumina is "fluoride specific."  Tills conclusion is in agreement with the
reported preferential sequence of adsorption onto alumina:

                OH" » F~ » S0*~ » Cl" » HCO~

The exact values for separation factors between any two anlons in the series
may be derived after tests for matched pairs of radicals determine the Lang-
muir mathematical model parameters, i.e., after the binary adsorption isotherms
are established.

-------
                                  REFERENCES
 1.   Activated and Catalytic Alumina, Alcoa Product Data Chemicals, Section
      GB-2A, Aluainum Company of America, July 14, 1969.

 2.   Amphlett, C.B.  Inorganic Ion exchangers (New York, Elsevier Publishing
      Company, 1964), p. 90, Fig. 25.

 3.   Bolt, G.H. and Page, A.L.  "Ion Exchange Equations Based on Double Layer
      Theory," Soil Science, 99, 357 (1965).

 4.   Choi, Won-Wook and Chen, Kenneth Y., "The Removal of Fluoride from Waters
      by Adsorption," Journal AtfHA, Vol. 71, No. 10, p. 562, October, 1979.

 5.   Churchill, II.F., U.S. Patent 2,059,552, November 3, 1936.

 6.   Clifford, Dennis A., "Mobile Pilot-Scale Evaluation of Reverse Osmosis,
      Ion-Exchange, and Activated Alumina Adsorption for the Treatment of Small
      Community Water Supplies," A research proposal submitted to US EPA,
      March, 1978.

 7.   Clifford, D.A.; Matson, J.V.; Kennedy, Ralph, "Activated Alumina:  Redis-
      covered 'Adsorbent1 for Fluoride, Humic Acid, and Silica," Industrial
      Water Engineering, p. 6, December, 1978.

 8.   Clifford, D.A., "Nitrate Removal from Water Supplies by Ion Exchange,"
      US EPA-600/2-78-952, June. 1978, p. 34-37.

 9.   Culp, R.L. and Stoltenberg, H.A., "Fluoride Reduction at LaCrosse, Kansas,'
      J. AWWA, Vol. 50, No. 3, p. 423. March, 1958.

10.   Fink, G.J. and Lindsay, F.K., "Activated Alumina for Removing Fluorides
      from Drinking Water," Ind. and Eng. Chem., Vol. 28, No. 8, p. 947 (1936).

11.   Greenwood, D.A., "Some Effects of inorganic Fluoride on Plants, Animals,
      and Man," Fifteenth Annual Faculty Research Lecture, The Faculty Associa-
      tion, Utah State Agricultural College, Logan, Utah, 1956, pp. 12-23.

12.   Hesse, G. and Sauter, 0., "On the Independence of Exchange Adsorption and
      Van der Waals Adsorption of Aluminum Oxide," Naturwiss., Vol. 34, p. 250
      (1947).
                                      44

-------
13.  Hingston, F.J.; Posner, A.M. ; and Quirk, J.P., "Anion Adsorption by Geo-
      thite and Cilinite:  I.  The Role of r.he Portion in Determining Adsorp-
      tion Envelopes," J. Soil Science, Vol. 23, No. 2, p. 177, June, 1972.

14.   Rubli, H.A., "Contribution to the Knowledge of Anion Separation by Means
      of Adsorption on Alumina," Helv. Chem. Acta, Switzerland, 3:453 (1947).

15.   Maier, F.J., "Defluoridation of Municipal Water Supplies," J. AWWA, Vol.
      45, August, 1953, p. 879.

16.   Muljadi, D.; Posner, A.M.: and Quirk, J.P., "The Mechanism of Phosphate
      Adsorption by Kaolinite,  Gibbsite, and Psuedoboahmite I," J. Soil
      Science, Vol. 17, p. 212, September (1966).

17.   Committee on Biological Effects of Atmospheric Pollutants, Division of
      Medical Sciences National Research Council, Biological Effects of Atmos-
      pheric Pollutants:  Fluorides, National Academy of Sciences, Washington,
      D.C., 1971, pp. 209-214.

18.   Parks, G.A. , "The Isoelectric Points of Solid Oxides, Solid Hydroxides,
      and Aqueous Hydroxo Complex Systems," Chem. Rev., 65:117 (1965).

19.   Rubel, F. , Jr. and Woosley, R.D., "The Removal of Excess Fluoride from
      Drinking Water by Activated Alumina," J. AWWA, 71:1:45 (January, 1979).

20.   Savinelli, E.A. and Black, A.P., "Defluoridation of Water with Activated
      Alumina," J. AWWA 50:1:33 (January, 1958).

21.   Schwalb, G.M. and Dattler, G. , "Inorganic Chroma tography," Agnew. Cheraie,
      Vol. 50, No. 33, p. 691 (August, 1937).

22.   Shimkin, M.B.; Arnold, F.A. , Jr.; Hawkins, J.W. ; and Dean, H.T., "Medical
      Aspects of Fluorosis," American Association Advancement Sciences, 1953.

23.   Sognnaes, R.F., "The Problem of Providing Optimum Fluoride Intake for
      Prevention of Dental Caries," A Report of the Committee on Dental Health,
      Publication 294, November, 1953, p. 4.

24.   Swope, H.G. and Hess, R.H., "Removal of Fluorides from Natural Waters
      by Deflurite," Ind. and Eng. Chem., Vol. 29, No. 4, p. 424 (1937).

25.   Texas Department of Health, "Community Water Systems in Texas Which Exceed
      the MCL for Fluroides as Set by the National Interim Primary Drinking
      Water Regulations, 18 pages, Texas Department of Health, Austin, Texas,
      March, 1977.
26.   Umland, F. , "The Interaction of Electrolyte Solutions and a-Al203:  IV:
      Development of a Formal Ion Exchange Theory for the Adsorption of
      Electrolytes from Awueous Solutions," Z. Electrochem. , V. 60, p. 711
      (1956).

-------
27.   Federal Register, U.S. EPA, "National Interim Primary Drinking Water
      Regulations," Washington, D.C., December 24, 1975.

28.   U.S. Public Health Service, Drinking Water Standards, 1962.

29.   Van Olphen, H., An Introduction to Clay Colloid Chemistry, Interscience,
      New York, 1963.

30.   World Health Organization, "Toxic Effects of Larger Doses of Fluoride,"
      Fluorides and Human Health, WHO Monograph Series No. 59, Geneva:  WHO,
      1970, pp. 238-239.

31.   Zabban, Walter and Jevett, N.W., "The Treatment of Fluoride Wastes,"
      Proc. 22nd Ann. Purdue Industrial Waste Conference, Engrs. Bull. No. 129
      (1967).

32.   Parks, G.A., "Aqueous Surface Chemistry of Oxides and Complex Oxide
      Minerals.  Equilibrium Concepts in Natural Water Systems," Advances in
      Chemistry Series, 67, Robert C. Gould (Editor), American Chemical Soceity
      Publications, Washington, D.C., 1967.

33.   Wu, Yeun C., "Activated Alumina Removes Fluoride I0ns from Water," Water
      and Sewage Works, Vol. 125, No. 6, p. 76, June, 1978.

34.   Zhabrova, G.M. and Eorov, E.V., "Sorption and Ion Exchange on Amphoteric
      Oxides and Hydroxides," Russian Chemical Reviews, Vcl. 30, No. 6, p. 338,
      1961.

-------
                                 APPENDICES
MODIFIED TURBIDIMETRIC METHOD FOR SULFATE DETERMINATION

     Reference:  Standard Methods for the Examination of Water and Waatevater,
14th Edition, p. 344.

     1.  Place 100 ml sample in 300 ml Erlenmeyer flask.
     2.  Add 5 ml conditioning reagent.
     3.  Add one "scoop" (0.2 - 0.3 ml) of reagent-grade barium chloride.
     4.  Shake by hand, swirling occasionally, for one minute.
     5.  Allow 4 additional minutes for turbidity to develop with no additional
         agitation.
     6.  Set 10 ppm sulfate to read "40" on 0 - 100 scale of Hach turbidimeter
         (Model 2100A) using 25 ml sample and no spacer in the reading chamber.
     7.  Read turbidity of all standards and samples after exactly four minutes
         of turbidity development following initial one minute agitation period.
     8.  Plot NTU versus ppm sulfate and read off samples.
     Notes:     a.  Linear range is 2 - 10 ppm sulfate
                b.  Standard typically, 2, 4, 6, 8, and 10 ppm sulfate
                c.  See Standard Methods for Preparation of reagents and
                    standards.
FLUORIDE DETERMINATION PROCEDURE WITH ELECTRODES USING 901 IONALYZER

     Reference:  Orion Analyzer Instruction Manual, Fluoride Electrode Model
96-09.

     1.  Prepare 10 ppm and 2 ppm standard fluoride solutions from the 100
         ppm stock solution.
     2.  Add 25 ml TISAB II or IV to each 25 ml of above standard and samples
         to be tested.
     3.  Set the standard value switch to 10.  Set the slope switch to the
         slope of the electrode, that is, 56.0 mV/decade.
     4.  Place electrode in 50 ml of 10 ppm standard solution.  Press Clear/
         Read mV and wait for a stable reading.  Press Set Concentration.
         Check the slope by placing electrode in 2 ppm standard solution.
     5.  Place the electrode in 50 ml of sample solution, wait for a stable
         reading, and record sample concentration.  Similarly repeat step for
         other samples.
     Notes:     a.  For every change of standard or sample solution, electrode
    •              should be rinsed and fillter dried before placing it in
                    the solution.


                                     47

-------
                b.  TISAB II and TISAB iV ionic strength adjusters were pre-
                    pared as per Orion lonalyzer Instruction Manual, Fluoride
                    Electrodes Model 94-09, Model 96-09.
                c.  TISAB IV was used only for testing NaOH-eluted samples.
                    because they contained high concentrations of Al.

POTENTIOMETRIC TITRATION METHOD FOR CHLORIDE

     Reference:  Clifford, Dennis A., "Nitrate Removal from Water Supplies by
Ion Exchange," EPA 600/2-78-052, June, 1978, p. 250.

     1.  Make sample to be titrated up to approximately 50 ml In ft 150 ml
         beaker with teflon-coated magnetic stirring bar.
     2.  Titrate with .0141 N AgN03 (. nO meq Cl/ml) to +275 mV end point.
         This was previously determined to coincide with the inflection point
         in the ml tltrant adder versus mV plot.  Potential due to increase
         in Ag+ ion was measured using double junction (nitrate-external)
         calomel, reference electrode (Orion 90-02-00) with Ag/AgS solid
         state specific Ion electrode (Orion 94-10A).
     Notes:     a.  Sensitivity is 125 mV/ml titrant added at inflection point
                    for .0141 N AgN03.
                b.  AgN03 standardized against 1000 ppm NaCl.


BICARBONATE DETERMINATION WITH TOC ANALYZER

     Reference:  Beckman Model 95 TOC Analyzer Instruction Manula, for Instru-
ment operations.

     1.  Prepare 10, 20, 40, and 100 ppm NaHC03 standards from freshly pre-
         pared NaHCC>3 stock solution.
     2.  Rinse syringe at least twice by discharging contents elsewhere (not
         back into beaker).  Draw sample up slowly.
     3.  Inject into inroganic channel of TOC analyzer and set recordings to
         standardize.
     4.  Make at least duplicate Injections or repeat until reproducible peaks
         are obtained. .
     5.  Repeat Steps 2, 3, and 4 for all unknown samples and from peak heights
         calculate HCO^ in samples.


TYPICAL CALCULATIONS

     For Run No. 34:

                        F
                       Is
                       xs
                     t.3711110]
                    [.300][.246]
                                     48

-------
       TABLE Al

TIME SPENT IN HOURS FOR
   THROUGHPUT T » 1
Run No.
1
M
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
35
19
20
21
22
23
24
25
Hours
16.49
19.54
19.26
13.28
7.54
11.75
6.17
10.26
4.49
6.74
100.49
77.96
82.39
53.63
62.19
73.37
50.05
60.76
45.73
23.03
19.50
13.48
20.02
10.17
23,27
Run No.
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
A3
44
45
46
47
48
I*
50
Hours
23.74
10.62
7.41
7.22
10.60
28.34
32.12
20.09
12.42
12.98
34.22
31.87
53.01
26.17
51.79
17.16
26.99
15.57
25.20
14.96
12.64
18.57
19.26
14.85
10.24

-------
                       TABLE A2 CONVERSION FACTORS
 1.00  milliequivalent fluoride -   19.00  mg fluoride

 1.00  gram                     -   15.43  grains
 1.00  gram alumina             -    1.067 ml alumina  (See Note  2)
                                   (1.134)       3
•1.00  gram alumina             -   f8'888n2r$s}ft  alumlna
 1.00  mg F~7g alumina          -  409.5 g F~/ft  alumina
                                  (385.3)        _   3
 1.00  meq F /g alumina         -  77ft?  grains F /ft  axumlna
                                  (7320)         _   3
 1.00  meq F/ml alumina         -  8303   grains F /ft  alumina
Notes:   (1)  Alumina  is dilute-H2SO^-treated Alcoa F-l grade
             activated alumina, 26x48 mesh, conditioner ac-
             cording  to the procedure given on page 12.
         (2)  Alumina bulk densities are based on
             treated, loosely packed, conditioned alumina whose
             weight was determined after drying for 48 hrs at
             110°C.   The  experimental bulk density  of 1.00  g/
             1.067 ml corresponds to  58.5  lbs/ft3, i.e., 6.4%
             higher than the 55 Ibs/ft3 (manufacturer's data)
             reported for packed alumina in Table 4,  p. 13.
             The values in parentheses are based on 55 Ibs/ft3.
                                   50

-------
                                 10.00 ppm F  in Feed
          t Q - 21 hrs
           • 7
                                                 o—"—o—o-
                                  6.0 ppm P  in Feed
             t   - 24  hrs
                       ^  _     4.0 ppm F  in Feed


                       «—- t  . - 90 hrs
                            * J
                                  2.0 ppm F  in Feed
5,000      10,000       15,000       20,000


          BED  VOLUMES  (BV) OF EFFLUENT (ml)
                                                   25,000     30,000
FIGURE Al  MINI-COLUMN  EFFLUENT CONCENTRATION HISTORIES, pH «• 5.0

           EFFECT OF CONCENTRATION ON COLUMN KINETICS

                Q - 3.55 ml/min, EBCT -  ,28 rain

     BV - 1.0 ml, t o - time to 90% of equilibrium in hours

-------
in
                                                                              ppm F  in  Influent
                                                          t _ - 64 hrs
                                                                              ppm F   i.i  Infli-ent
                                                                                - 87  hrs
                                    5.000       10,000      15.000     20,000

                                             BED VOLUMK (BV)  OF EFFLUENTS (ml)
25,000     30,000
                               FIGURE A2 MINI-COLUMN EFFLUENT CONCENTRATION HISTORIES, pll =  5.0
                                          EFFECT OF CONCENTRATION ON COLUMN KINETICS
                                        Q - 2.79 ml/min,  EBCT = .35 mln, BV - 1.0 ml
                                          t _  = time to 90% of equilibrium in hours

-------
                                          10.0 ppm F  in Feed
                                           8.0 ppm F  in Feed
     2,500       5,000      7,500       10,000     12,500      15,000


              BED VOLUME (BV) OF EFFLUENT



FIGURE A3  MINI-COLUMN EFFLUENT CONCENTRATION HISTORIES, pH - 6.0

            EFFECT OF CONCENTRATION ON COLUMN KINETICS

           Q - 0.5 ral/min, ETCT - 2 min, BV - 1.0 ml

           t „ = time to 90% of equilibrium in hours
                                               *

-------
                                                10.0 ppm F  In Feed
                                                 8.0 ppm F  in Feed
               t    =  112  hrs
                                                 6.0 ppm F  in Feed
                      t    =  155 hrs
                                                 "0 ppm F  in Feed
                                         9 « 137 hrs
                                     „ _        .   „  ,  t «=    hrs
                                     2.0 ppm F  in Feed   .9
       2,000       4,000      6,000       8,000

                BED VOLUMES (BV)  OF EFFLUENT
10,000     12,000
FIGURE  A4 MINI-COLUMN EFFLUENT CONCENTRATION HISTORIES, pll - 6.0
           EFFECT UF CONCENTRATION ON COLUMN KINETICS
             Q - 0.5 ml/rain, EBCT - 2 min, BV - 1.0 ml
             t 0 • time to 90% of equilibrium in hours
              • V

-------
   10




    9



    8
 oo
 e
u


I  5



'-  4
u


§  '
h.
W
    2
    1




    0
                                 10.0 ppm F  in Feed
t . = 67 hrs
                                                          8.0 ppm F  in Feed
                                                          6.0 ppm F  In Feed
   t   - 52 hrs
                                                          4.0 ppm F  in Feed
           t   « 41 hrs
                        .9
                                                          2.0 ppm F  in Feed


                                                     67 hrs
                 2,000       4,000       6,000       8,000      10,000


                          BED VOLUMi7.S (BV)  OF EFFLUENT
                                                    12,000
             FIGURE  A5  MINI-COLUtfN  EFFLUENT CONCENTRATION HISTORIES, pH - 7.0

                        EFFECT OF CONCENTRATION ON COLUMN  KINETICS


                     Q - 1.78 ml/min,  EBCT - .84  min, BV - 1.0 ral

                       t Q  ~  time  to 90%  of equilibrium in hours
                        . 7

-------
1.1
                                                                            10.0 ppm F   In  Fe«d
                                        t Q - 31 hrs
                                         » *
                                                                             8.0 ppsi  F   in  Feed
                                         t   - 26 hrs
                                                                              6.0 ppm  F  in  Feed
                                          t   = 19 hrs
                                           • y
                                                                             4.0  ppm F  in  Feed
                                                                              2.0  ppm  F  in  Feed
                                 2,000       4,000      6.00C       8.000


                                          BED VOLUMES  (BV) OF EFFLUENT
10.000
12,000
                            FICURE  A6 MINI-COLUMN EFFLUENT CONCENTRATION  HISTORIES,  pH - 8.0

                                        EFFECT OF CONCENTRATION ON COLUMN KINETICS

                                       Q - 1.23 ml/rain, EBCT -  .81 irdn.  BV  -  1.0 ml

                                       t Q *•  time to 90% of equilibrium  in  hours
                                        • 7

-------
                    10.0
In
                     7.5
                  60
                  B
                  O
                  25
                  8
                     5.0
                  u.
                     2.5
                    0.0
     t.g = 68 bra
   t 9 = 69 hrs
t ~ = 48 hra
9.5 ppm F  Feed Solutions Contain

    HCO~ + S0*~ + Cl~ - 3 meq/1,  I
                                             .0040 M

                                             .0085 M
                                                                2-
                                                                2-
                + Cl  =6 meq/1,  I

                + Cl~ = 12 meq/1,  I - .0145 M

                + Cl~ = 24 meq/1,  I - .0285 M
                                                      EHC03 + SO   + Cl  - 48 meq/1, 1 = .0565 M
                                   2000        4000       6000        8000

                                           BED VOLUMES (BV) OF EFFLUENTS
                                     10,000
                                         12,000
                             FIGURE A7   MINI-COLUMN EFFLUENT CONCENTRATION HISTORIES, pH - 7.1
                                        EFFECT OF IONIC STRENGTH ON COLUMN KINETICS
                                        Q = 1 ml/mln, EBCT - 1 mln, BV = 1.0 ml
                                        t „ = time to 90% of equilibrium In hours

-------
                                  GLOSSARY
"as CaCC>3":  Normality  (N) can be converted to calcium carbonate  equivalents.
     There are 50 mg of CaCC>3 per milliequivalent.

bed:  Alumina resin contained in a column.  Water to be defluoridated by
     alumina adsorption process is passed downward through the column

breakthrough:  The point at wbic.i F~ concentration in the effluent  reaches
     the MCL or some predetermine' concentration.

elution:  Application of NaOH soiut'.on to an exhausted alumina bed  so as  to
     "elute" or drive off the adsorbed anions.

exhaustion:  The step in an adsorption cycle in which the undesirable F~  ions
     are removed from the water, and are adsorbed on alumina bed.

ion-exchange:  A physicochemical process in which ions in the water being
     treated replace and are exchanged for ions in a solid phase  (alumina).

isotherm:  A constant temperature plot of resin phase concentration of an
     ion versus the water phase concentration of that ion, at the equilibrium
     slope.

milliequivalent (meq):  1/1000 of a gram equivalent of ion.

regeneration:  Restoration of the activated alumina after exhaustion to
     reclaim its original adsorption capacity.

selectivity:  A measure of the relative affinity for one ion over another
     exhibited by the alumina.  Selectivity is measured by the separation
     factor.

separation factor (binary):  The ratio of the distribution of ions  between
     the solid phase and the liquid phase at equilibrium.

                 A = distribution of A ions between solid & liquid  phase
                aB ° distribution of B ions between solid & liquid  phase

                     VXA     VQT * VCA
                   " yB/XB  "  VQT X VCB

superficial detention time (t):  The time a particle of feed water  spends in
     the empty alumina bed assuming plug flow.  It is calculated as the empty


                                      58

-------
     bed volume divided by the feed flow rate.

throughput:  Ratio of the total amount of fluoride which comes in contact with
     the adsorption column during a specified amount of time to the total
     adsorptive capacity of the column.
                                      59

-------

-------