Removal of Radium from Drinking Water
       (U.S.)  Environmental Protection Agency, Cincinnati, OH
       Aug 92
L
J
                        A.VIIC*

-------
                                   TECHNICAL REPORT DATA
                            If1r*tt rraJ Initniciiont on Iht rei-rnt btfore camp/el
i REPORT NO.
   EPA/600/R-92/164
                                                                   PB92-218320
4. TITLE ANOSUBTlTLE
                   Removal of Radium  from Drinking  Water
                                                            5 REPORT DATE
                                                                     August 1992
                                                            • PERFORMING ORGANIZATION CODE
7 AUTMORIS.
                    p.
                                                            8 PERFORMING ORGANIZATION REPORT NO
9 PERFORMING ORGANIZATION NAME AND ADDRESS
  U.S.  Environmental  Protection Agency
  Drinking  Water Research Division
  26 Martin Luther King Dr.
  Cincinnati.  OH  45268
                                                            10 PROGRAM ELEMENT NO.
                                                            1 ». CONTRACT/GRANT NO.
17. SPONSORING AGENCY NAME ANO ADDRESS
 Risk  Reduction Engineering Laboratory
 Office  of Research and Development
 U.S.  Environmental Protection Agency
 Cincinnati,  OH  45268
                                                            ID TYPE OF REPORT ANO PERIOD COVERED

                                                             Prni'prt Rpnnrt	
                                                            14. SPONSORING AGENCY CODE

                                                             EPA/600/14
IS SUPPLEMENTARY NOTES
                      Project Officer = Richard  P.  Lauch
                                                                (513) 569-7237
16. ABSTRACT
            This report summarizes processes  for  removal  of radium from drinking water.
  Ion  exchange,  including strong acid and weak  acid  resin,  is discussed.  Both processes
  remove  better  than 95 percent of the radium frjim the  water.  Weak acid ion exchange does
  not  add sodium to the water.  Calcium cation  exchange removes radium and can be used
  when hardness  removal is not necessary.
       Iron  removal processes are discussed in  relation to  radium removal.  Iron oxides
  remove  much less than 20 percent of the radium  fronr water under typical conditions.
  Manganese  dioxide removes radium from water when competition for sorption sites and
  clogging of sites is reduced.  Filter sand  that is  rinsed daily with dilute acid will
  remove  radium  from water.
       Manganese dioxide coated filter sorption removes radium but more capacity would
  be desirable.   The radium selective complexer selectively removes radium with
  significant capacity if iron fouling is eliminated.
       Other radium removal processes that are  discussed  include limesoda ash softening,
  reverse osmosis, electrodialysis reversal and potassium permanganate-greensand
  filtration.
       A  brief discussion of process cost is  given.
17.
                                KEY WORDS ANO DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b.lOENTiFIERSi'OPEN ENOEO TERMS
                                                                            COSATi Field,Croup
  Radium,  Removal, Potable water, Water
  Treatment,  Ion Exchanging, Iron,
  Manganese Oxides, Water Softening, Osmosis
  Electrodialysis
                                              Removal  of Radium,
                                              Drinking Water,
                                              Water  Treatment
18. DISTRIBUTION STATEMENT


 RELEASE TO PUBLIC


If A form Z220-I («•». 4-77)   previous COIT.ON 
-------
                                           EPA/600/R-92/164
RQ»VAL OF RADIUM FROM DRINKING WATER           August
                  by
           Richard P. Lauch
   Drinking Water Research Division
Water Engineering Research Laboratory
       Cincinnati, Ohio  45268
WATER ENGINEERING RESEARCH LABORATORY
  OFFICE OF RESEARCH AND DEVELOPMENT
 U.S. ENVIRONMENTAL PROTECTION AGENCY
        CINCINNATI, OHIO 45268

-------
                               DISCLAIMER
     This report has been reviewed in accordance with che U.S.
Environmental Protection Agency's peer and adnlnlstrative review
policies and approved for presentation and publication.   Mention of
trade names or commercial products does not constitute endorsement  or
recommendation for use.
                                  ii

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

     The Water Engineering Research Laboratory Is that component of
EPA'3 Research and Development program concerned with preventing,
creating, and managing municipal and Industrial wastevater discharges;
 '.stablishing practices to control and remove contaminants from drinking
water and to prevent Its deterioration during storage and distribution;
and assessing the nature and controllability of releases of toxic sub-
stances to the air, water, and land from manufacturing processes and
subsequent product uses.  This publication is one of the products of
that research and provides a vital communication link between the
researcher and tne user community.

     Total radium (Ra-226 plus Ra-228) exceeds the drinking water
regulation maximum contaminant level of 5 pCi/L In wellwater supplies
for some communities In the United States.  Occurrence Is most frequent
in the states of Iowa, Illinois, Wisconsin, Virginia, North Carolina,
Florida, Colorado and the New England States.  This report Is a summary
of treatment processes chat will remove radium from drinking water
supplies.  The report emphasis is on the results of four recent research
projects.
                                 Francis T. Mayo,
                                     Director
                       Water Engineering Research Laboratory
                                 ill

-------
                              ABSTRACT
     This report summarizes processes for removal of radium from
drinking water.  Ion exchange. Including strong acid and weak acid
resin, Is discussed.  Both processes remove better than 95 percent of
the radium from the water.  Weak acid Ion exchange does not add sodium
to the water.  Calcium cation exchange removes radium and can be used
vhen hardness removal Is not necessary.

     Iron removal processes are discussed In relation to radium
removal.  Iron oxides remove much less than 20 percent of the radium
from water under typical conditions.  Manganese dioxide removes radium
from water when competition for sorptlon sites and clogging of sites Is
reduced.  Filter sand that is rinsed dally with dilute acid will remove
radium from water.

     Manganese dioxide coated filter sorption removes radium but more
capacity would be desirable.  The radium selective cpraplexer selectively
removes radium with significant capacity if iron fouling is eliminated.

     Other radium removal processes that are discussed Include lime-
soda ash softening, reverse osmosis, electrodlalysls reversal and
potassium permanganate-greensand filtration.

     A brief discussion of process cost is given.
                                  iv

-------
                                CONTENTS

                                                                   Page

Foreword	   Ill
Abstract	    Iv
Figures	   vll
Tables. .............................     x
Unit Conversions	   xil
Acknowledgment	xlll

  1.  Introduction.	     1
  2.  Results and Conclusions	     2
  3.  Recommendations for Future Research	    11
  4.  Ion Exchange (University of Illinois Study)	    13
        Summary of Ion Exchange	    13
        Results and Discussion.	    15
          Definition of Column Utilization and
             Regeneration Efficiency	    15
          Strong Acid Resins	    15
            Steady State Condition	    18
            Regeneration	    23
            Radium Removal	    25
          Weak Acid Resin	    28
            Steady State Condition	    30
            Regenerant dose	    38
            Radium Removal. ... 	 ...........    38
          Calcium Cation Exchange ........ 	    41
  5.  Iron and Manganese Removal (University of Iowa Study) ...    46
        Sunmary of Iron and Manganese Removal	    46
        Testing Procedures. 	    47
        Types of Vater Used	    49
        Results and Discussion	    50
          Sorption to Hydrous Iron Oxides ............    50
            Effect of pH	    50
            Removal Mechanism	    51
            Effect of Iron a«*d Radium Concentration on Sorption .    51
            Effect of Competing Ions.	    56
            Comparison of Sorption to Natural and
             Synthetic Iron Oxides	    57
          Sorption onto Manganese Oxides	    59
            Pure Manganese Oxides 	    59
            Mixed Iron/Manganese Oxides	    61
          Sorption to Filter Sand	    63
            Effect of pH in Absence of Hardness	    63
            Effect of Hardness	    63
          Regeneration of Filter Sand	    70
            Pilot Plant Studies	    70
            Field Studies at Oxford, Iowa (Sand Filter Effluent}.    73
            Field Studies at Oxford, Iowa (Aerator Effluent). . .    73
  6.  Manganese Dioxide Coated Filters
            (N. Carolina State Univ. Study)	    78
        Summary of Manganese Dioxide Coasted Filters	    78
        Manganese Dioxide Coating Process 	    79

-------
                                                                       Page
            Bleed Scream Tests	     79
            Sorptlon of Other Metals Onto Manganese Dioxide 	     84
      7.  Evaluation of Radium Removal and Waste Disposal  System
          for a Small Community Water Supply (Rocky Mountain
          Consultants, Inc., Denver, CO.)	     87
            Sumnary (Iron Removal, Ion Exchange, Waste Disposal).  .  .     87
            Discussion	     88
              General Information 	     88
              Iron Removal.	     88
              Ion Exchange	     91
              Waste Brine Treatment with Radium Selective  Complexer  .     92
      8.  Other Radium Recoval Processes	     95
            Lime & Mme-Soda Softening	     95
            Dow Radium Selective Complexer	     97
              Background	     97
              Test Results.	    100
                Bellvllle Te-as	    100
                South Superior, Wyoming .	    101
                Demonstration Columns ........... 	    101
            Reverse Osmosis	*    104
            Electrodialysls Reversal	    105
            Potassium Permanganate Greensand Filtration 	    107
      9.  Costs	    112

References	 .	    115
                                      vl

-------
                              FIGURES

Number                                                            Page
 1.  Barium effluent concentration from strong acid
     resin exhaustion-regeneration cyclic studies	    19

 2.  Breakthrough curves for the sixth exhaustion
     cycle of the strong acid resin	    22

 3.  Acid regeneration of virgin strong acid resin	    24

 4.  Weak acid resin breakthrough curves for run #7
     of cyclic run 92	    35

 5.  Weak acid resin brakthrough curves for run #6
     of cyclic run 03.	    37

 6.  Barium concentration at the termination of
     weak acid resin cyclic tests.	    37

 7.  Barium breakthrough curves at different
     regenerant dosages	    43

 8.  Regeneration efficiency and column capacity
     at various regenerant dosages 	    43

 9.  Schematic diagram of batch jar test apparatus 	    48

10.  Schematic diagram of pilot plant	    48

11.  The effect of pH on Radium removal in the
     presence of iron oxides	    52

12.  Percent radium removed by iron oxides
     (comparison of precipitation methods) 	    52

13.  Percent radium removed by iron oxides as a
     function of pH	    54

14.  Radium adsorbed on iron vs radium concentration
     remaining in solution 	    54

15.  Percent radium sorbed as a function of Iron
     concentration when no other divalent cations
     are present	    55

16.  Percent radium removed by iron oxides as a function
     of pH (Na-form synthetic water vs Oxford groundwater) ...    58

17.  Percent radium removed by iron oxides as a function
     of pH (Na-fona synthetic,water vs. Eldon groundwater) ...    58
                                  vli

-------
                                                                   Page

18•  Percentage of radium removed by Iron oxides In
     Ma-synthetic and Oxford groundwater 	    60

19*  Percent radium removed by natural and synthetic
     iron oxides in synthetic and Oxford water .........    60

20.  Radium removal in Na-forra synthetic water by
     1 and 5 mg/L MnC>2	    62

21.  Radium removal by MnC>2 in Oxford water	    62

22.  Radium removal In sodium form synthetic water
     by mixed iron/manganese oxides	    64

23.  Radium removal in Oxford water by mixed
     iron/manganese oxides 	    64

24.  Percent radium removed In Na-form synthetic
     water by acid washed sand as a function of pH	    65

25.  The effect of sand concentration on radium removal
     at pH 8.3 - 8.4 in sodium form synthetic water	    67

26.  Radium sorbed on sand vs radium concentration
     remaining in solution 	    67

27.  Percent radium removed on acid washed sand
     as a function of pH .	    68

28.  Percent radium removed as a function of radium
     concentration at pH 7.1 in Oxford water	    68

29.  Comparison of radium adsorption Isotherms (sand)
     developed in El don, Oxford, and synthetic waters	    69

30.  Comparison of radium adsorption isotherms (sand)
     developed in Oxford and 6mMCa +• 6 mM mg-form
     synthetic waters	    69

31.  Radium removal in the DHL pilot plant showing effect
     of a daily pH 1 HC1 rinse	,	    71

32.  Radium removal in the UHL pilot plant showing
     effect of a pH 2 HO. rinse during dally backwash.	    71

33.  Comparison of cumulative radium recovered in pH 1
     and pH 2 regenerant rinses	    72

34.  Radium concentration profiles as a function
     of loading rate	    72
                                  vl 11

-------
                                                                   Page

35.  Radium concentration profiles in Oxford water with
     and without pH 1 regeneration.  Loading rate 3 gpra/ft2.  .  .     74
36.  Radium concentration profiles In Oxford
     vat>?r with and without pH 1 regeneration.
     Loading rate 1.5 gpra/ft*	     74
37.  Radium concentration profiles in Oxford
     water with and without pH 3 regeneration	     75

38.  Influent and effluent Ra-226 concentration profiles.
     Treatment of Oxford aeration tank effluent.  Comparison
     of sulfurlc and hydrochloric acid regeneration at
     two regenerant loading rates	     77

39.  Schematic diagram of process for coating
     woven acrylic cartridge filters with Mn02 .........     80

40.  Percent radium removal as a function of pH
     during water softening	     98
                                  Ix

-------
                                 TABLES




Number






 1.  Efficiency and Column Utilization as a Function

2.
3.
4.

5.
6.

7.
8.

9.

10.

11.
12.
13.
14.

15.
16.
17.

18.
19.
20.
21.



Radium Removal, Second Strong Acid Virgin


Radium Accumulation, Strong Acid Resin Cyclic


Experimental Parameters for the Cyclic Tests
with Weak Acid Resin 	 , 	
Weak Acid Exhaustion - Regeneration Cyclic
Test *1 	
Ions Remaining on the Weak Acid Resin After the




Radium Mass Applied vs. M?as Removed During



The Relationship of Barium Ion Concentration to


Bleed Stream Field Test, Mn02 Filtration (Highland Park) . .


18
18
20

26
27

28
30

3J

32

33
34
34
39

40
41
42

56
82
83
83
86

-------
                           TABLES (continued)

Number                                                             Page
22.  Red»»':i Forest Water Quality Monitoring Data -
     General Treatuent Plant Operation	«...    89

23.  Summary of WQ Data for Regeneration
     Wastewater Through RSC Resin 	    94
•
24.  Results of Lime Softening Pilot Plant Tests
     for Radlura-226 Removal from Well Water	    96

25.  R-226 and Hardness Removals at Water Treatment
     Planto Using Lime-Soda Ash Softening 	    99

26.  Radium Removal Chronology - Bellvllle, Texas ........   102

27.  Radiological Results from the Full Scale
     System - South Superior, Wyoming 	   103

28.  Radium Removal by RO in Sarasota County, FL	   106

29.  Radium-226, Iron, and Manganese Removals by
     Iron and Manganese Removal Processes 	   110

30.  Radium Removal with the KMnO$ - Creensand Process	   Ill

31.  Estimated Costs of Radium Removal Processes	   114
                                  xi

-------
UNIT CONVERSIONS
English
Units
ft
"3
'ft3
gal
gal
gpm
gpm/ft2
In
Ib
lb/ft3
HCD
psl
psl
psl
Multiplied
By
0.3048
28.31625
0.02832
3.7854
3.785xlO"3
3.785
40.7
2.54xlO"2
0.454
16.02
3.78541
70.307
0.0703
703.07
Metric (SI)
Units
m
L
m3
L
m3
L/rain
L/mln/n2
m
kg
kg/n3
MLD
g/cm2
kg/ cm2
kg/m2
      XI1

-------
                            ACKNOWLEDGMENTS


     I thank the following people for performing che research and
writing che pi eject reports on the four EPA projects that are dis-
cussed:

     University of Illinois —

        Principal Investigator - Dr. Vernon L. Snoeylnk

        Research Assistants - Julie L. Pfeffer, David W. Snyder,

           Carl C. Chambers, Anthony G. Myers, Sharon K. Rlchter,

           Candace K. Jongeward Schmidt, Rich F. Manner

    'University of Iowa —

        Principal Investigator - Dr. Richard L. Valentine

        Project Manager - Dr. Roger C. Splinter

        Research Assistants - Thomas M. Nogaj , Jao-Jla Horng,

           Timothy S. Mulhoiland, Jeffrey M. Baker.

     North Carolina State University —

        Principal Investigator - Dr. Edward P. Stahel

        Co-Investigator - Marc Y. Menetrez

     Redhill Forest Homeowner's Association, Hartsel, Colorado and

        Rocky Mountain Consultants, Denver, CO. —

        Principal Investigator - Dr. Kenneth A. Mangel son

        Water Treatment Plant Operator - Eric Poclus

     I thank Richard E. Rozelle, Thomas Boyce, Dr. King Way Ma, Dale
M. Fales, and E. C. Flanagan of the Dow Chemical Company for their
reports and other information on the Radium Selective Complexer.

     Finally, I thank Thomas J* Sorg and Dr. Robert M. Clark for their
technical assistance, James J. Westrick and Rim R. Fox for their technical
reviews, and Gall E. Brookhart for her editorial review.
                                  xlll

-------
                             SECTION 1


                            INTRODUCTION




     Over the past four years Che Drinking Water Research Division,


Water Engineering Research Laboratory, Cincinnati, Ohio, has sponsored
•

four cooperative agreements to perform research on the removal of


radium from drinking water sources.  These cooperative agreements were


with the University of Illinois, University of Iowa, North Carolina


State University and the Redhill Forest Property Owner's Association in


Colorado.  Ion exchange, iron and manganese removal, and sorption onto


manganese dioxide coated acrylic filters were investigated under the


first three projects. Iron and manganese, ion exchange and Dow Chemical


Company's radium selective complexer (used for waste brine) were Inves-


tigated full scale, at the small community of Redhlll Forest.  Most of


this report will be devoted to summarizing the results of these cooper-


ative agreements.


     Other processes that remove radium are lime softening, reverse


osnosis, and electrodialysls.  Results from the literature showing the


effectiveness of these processes are also Included in this report.


     The potassium permanganate greensand filtration process has been


shown to remove radium but the mechanisms for removal have never been


determined.  Some data showing radium removal icross the greensand


process is included in this report.


     The radium selective complexer was investigated on waste brine


under one of the cooperative agreements mentioned above, and these data


along with some of Dow Chemical's results on the radium selective com-


plexer are discussed.

-------
                               SECTION 2.


                        RESULTS AND CONCLUSIONS


ION EXCHANGE


     Studies conducted at the University of Illinois yielded the


following results:
•

 1.  Strong acid ion exchange resin In the sodium or hydrogen fora


     removed 99 percent of the radium and/or barium from the water when


     the resin was regenerated with 4.72 meq (Nad or HCl)/g dry


     resin.  This is typical for softening applications.  At steady


     state conditions, both radium and barium were being removed to


     concentrations much lower than their maximum contaminant levels


     (MCL's) when hardness started to break through at 215 bed volumes


     (BV).  This water had hardness, barium and radium concentrations


     of 200 mg/L as CaC03, 20 mg/L and 20 pCl/L respectively.  The


     strong acid resin has less capacity for radium and barium on water


     with higher hardness and more capacity for radium and barium on


     water with lower hardness.   t«""'


 2.  Strong acid resins have the following preference:  Ra2+> Ba2+>


     Ca2+> Mg2+> Na+./ Because of the selectivity radium and barium


     were much more difficult to remove from the resin during regen-


     atlon than calcium and magnesium.  Economical regeneration removes


     most of the hardness ions, but radium and barium buildup on the


     resin after repeated cycles to the point where equilibrium is


     reached and then radium and barium will begin to breakthrough


     shortly after hardness.


 3.  Regeneration of the sodium form strong acid resin for water with


     200 mg/L (CaOXj) hardness with application of 6.5 Ib NeCl/ft3

-------
      (4.72 meq/g) resin produced 2.4 BV of 16,400 IDS brine per 100 BV



      of product water.  If all of the radium and barlun were removed



      from the product water, then the brine would contain radium and



      barium concentrations that equal the raw water concentration times



      100/2.4.  This assumes that the 2.4 BV of brine includes the con-



     • centrated brine and final rinse.



      Strong acid resins in the hydrogen form would not normally be used



      for drinking water treatment because of the resulting strong min-



      eral acids that would form in the product water.



  5.  Weak acid ion exchange resins only remove cations that are



      balanced equlvalently with alkalinity; they must be regenerated



      with a strong acid such as hydrochloric or sulfuric and they



      exchange hydrogen Ions for divalent cations, thereby forming



      carbon dioxide which must be removed by stripping.  The sodium



      concentration in the drinking water will not Increase when weak



      acid resins are used unless a chemical such as Na2C03 were added



      to Increase alkalinity.



; /6.  The weak acid ion exchange resin removed greater than 95 percent



      of the radium and/or barium when the resin was regenerated with



      8.5 meq HCl/g dry resin.  The regenerant dose of 8.5 meq HCl/g



      dry resin yielded a regeneration efficiency of 93 to 96 percent




      and a column utilization of 70 percent which is very close to



      the optimum operating condition.  At the above steady state



      conditions both radium and barium were being removed to concen-



      trations much lower than their MCL's and barium began to break



      through at 650 BV't< which-was 'a little later than hardness.



      Radium continued to be removed after hardness break through but

-------
    the exact break through point for radium was not determined.



    These tests were performed on water with hardness and alkalinity



    concentrations of approximately 200 and 250 mg/L as CaC03 respec-



    tively.  Radium and barium concentrations were approximately 20



    pCi/L and 20 mg/L respectively.



7.  Weak acid resins have the following preference:  Ra2+> Ba2+> •



    Ca^*> Mg2+> H*./ The weak acid resin was easier to regenerate



    than the strong acid resin and therefore a smaller quantity of



    brine will be required per unit of water treated.  The weak acid



    resin comes to equilibrium after fewer cycles than for the strong



    acid resin because cations including radium and barium are much



    easier to remove from the resin during regeneration.



8.  Regeneration of the weak acid resin, for water with 200 and 250




    mg/L as CaCC>3 hardness and alkalinity respectively at a regenerant



    dose of 8.5 meq HCl/g dry resin, produced 1.2 BV of spent brine



    with a IDS of 19,900 mg/L per 100 BV of product water.  If all of



    the radium and barium were removed from the product water, then



    the brine would contain radium and barium concentrations that



    equal the raw water concentration times 100/1.2. This assumes that



    the 1.2 BV of brine includes the concentrated brine and final



    rinse.



*>.  The cost of using a weak acid resin is greater than that of the



    strong acid resin by about $0.15/1000 gal and $0.08/1000 for a



    0.1-MGD and 1-MGD plant respectively.  The use of a strong acid



    resin costs approximately $1.36/1000 gal and 0.38/1000 gal for a




    0.1-MGD and 1-MGD plant respectively.

-------
10.  Strong acid resin in the calcium form can be used co remove radium



     and barium.  The process would be useful for low hardness waters



     that exceed the radium MCL.  Calcium cation exchange can also be



     used to remove radium from the by-pass water for conventional Ion



     exchange plants.



IRON AND MANGANESE REMOVAL



     The University of Iowa study yielded the following results:



 1.  Sorption of radium onto iron oxides Increases as pH and Iron con-



     centration increase, but typical pH values and iron concentrations



     for natural waters are too low for significant radium removal in



     the presence of other competing cations.  Sorption of radiun to



     existing iron oxides in natural waters is expected to be less than



     20Z.



 2.  The amount of radium sorbed onto iron floes decreased as the con-



     centration of divalent cations, such as barium, calclun and magnes-



     ium Increased.



 3.  A series of experiments that compared radium sorptlon onto natural



     iron oxides with that onto synthetic iron oxides showed no differ-



     ence between the two.



 4.  Tests in sodium synthetic water showed that sorptlon of radium



     onto the iron oxide surface takes place and not incorporation of




     radium into the ferrous ion as oxidation is taking place.  This



     was shown by adding radium in the presence of ferrous Ion before



     aeration and then adding radium after the iron was oxidized.  The



     sane quantity of radium (75%) was removed In both cases*



 5.  Radium removals by freshly precipitated hydrous manganese oxides



     were much greater than removal by iron oxides.

-------
 6.  Radium sorpcion onto 1 mg/L Mn02 Increased from 80 co 90  percent



     as pH Increased from 6.5 Co 9.   This was In synthetic water con-



     taining 300 mg/L calcium as Ca(X>3.   The higher sorptlve capacity



     of manganese dioxide over Iron  oxides Is based on the lower point



     of zero charge for Mn02 In water which takes place at pH  less than



     5.  Therefore, MnC>2 has good sorptlve capacities over the normal



     pH range of natural waters.



 7.  The addition of 5 mg/L pure Mn02 to Oxford water (hardness »



     1300 mg/L as CaC03) removed 90  percent of the radium at pH 7.



     Iron was oxidized and removed prior to Mn02 addition.



 8.  Sorption to Mn02 appears to be  a good radium removal mechanism



     if sorption of ferrous ion and  coating of the manganese oxides



     with Iron oxides can be eliminated.



 9.  Filter sand has a potential capacity to sorb radium at typical




     hardness concentrations If the  capacity is maintained by  periodi-



     cally rinsing the sand with a dilute acid.  Removal efficiencies



     of approximately 80 to 90 percent were achieved in laboratory and



     pilot plant studies when dally  rinsing with a dilute acid (0.1 to



     0.01N) was practiced.  The raw  water radium concentration was




     approximately SO pCl/L.



10.  Radium sorption to filter sand  is not very sensitive to normal pH



     variation encountered in drinking water (pH 6.5 to 8.5).



11.  During laboratory jar tests, percent radium sorption increased as



     sand concentration increased, but sorption was not a function of



     radium concentration.



12.  Pilot scale field studies at Oxford, Iowa (Hardnea? « 1300 mg/L as



     CaCOj), on water with iron removed, showed that sand reduced

-------
     radium from approximately 10 to 1 pCl/L.  The sand filter was




     regenerated once a day with 0.1 N HC1 and the loading rate was 1.5




     gpm/ft2.




13.  Pilot scale field studies at Oxford, Iowa, on water that contained




     0.5 mg/L iron and iron floe, showed that sand only removed 25 to




     35Z of the radium.  This sand filter was also regenerated once a




     day with 0.1 N HCL and the loading rate was 1.5 and 3.0 gpm/ft2.




MANGANESE DIOXIDE COATED FILTERS




     The North Carolina State University study yielded the following




results:




 1.  A process was developed for coating MnO£ onto woven acrylic




     cartridge filters.  Rot KMnO^ solution is circulated through




     the filter and the investigator explains that KMnC>4 is reduced




     by oxidizing the acrylic filter material resulting in Mn02




     that adheres to the filter.  A 10-inch long, 250 gram filter




     contains 56.25 grams Mn02 or 22.5 percent Mn02 by weight.




 2.  On high hardness water, a ten-inch Mn0  filter treated 800 ft
     of water before the effluent radium concentration exceeded



     5 pCi/L.  The influent had a radium concentration of  36 pCi/L,



     hardness concentration of 227 mg/L, as CaCOj,  iron and manganese



     concentration that was insignificant and a pH of 7.8*



 3.  On low hardness water, a ten-inch MnO? filter treated 1000 ft



     of water before the effluent radium concentration exceeded 5



     pd/L.  The Influent had a radium concentration of 13 pCi/L,



     hardness concentration of 23 mg/L as CaC03,  iron and  manganese




     concentrations that were insignificant and a pH of 4.5.

-------
 A.  If a radium removal plant were built using ten-Inch Mn02 filters



     and each filter treated water at a rate of 1 gpra for 7500 gallons



     (1000 ft3), then the filters would have to be changed every 5.2



     days for the plant to stay In compliance with the radium MCL.



     It would be unrealistic to change radioactive filters, that re-
*


     quire special handling, this often, and therefore, I conclude



     that MnC>2 filters are Impractical for radium removal unless they



     can be developed to have much greater capacity*  The final project



     report on MnC>2 filters Is still being prepared and the investi-



     gators may have a different opinion.



 5.  Data from this study also showed that Mn(>2 has significant capacity



     for sorptlon of other metals including iron and manganese ions.



     This agrees with the University of Iowa study.



EVALUATION OF RADIUM REMOVAL AND WASTE DISPOSAL FOR SMALL COMMUNITY



WATER SUPPLY



     The study by Rocky Mountain Consultants at Hartsel, Colorado



yielded the following results:



 1.  Results showed that very little radium Is removed across the



     iron removal process which agrees with the Univerlsty of Iowa



     study.



 2.  Ion exchange removed radium to concentrations below the MCL



     and hardness broke through before radium.  Radium broke through



     between 134 and 179 bed volumes of treated water.  Hardness



     concentration in the source water was approximately 265 mg/L as



     CaC03.



 3*  Greater than 99% of the radium (influent radium concentration



,.-    was approximately 1230 pCi/L) was removed from the waste brine

-------
         by the RSC.  The RSC has created over 1023 BV of waste brine with-




         out radium breakthrough.  The RSC Is truly radium selective be-




         cause the influent and effluent IDS concentrations were approxi-




         mately the sane at 42,000 mg/L.  These conclusions are not final




         because the project will not be completed until October, 1987.




    OTHER RADIUM REMOVAL PROCESSES




     Results obtained from various studies cited in the literature yielded




the following conclusions:




     1.  Lime and lime soda softening have been demonstrated on a full-




         scale plant level to achieve 75 to 96Z removal of radium.




     2.  Data indicate that radium removal, for the lime and lime soda




         softening processes is pH dependent with 84 percent and 94 percent




         radium removals at pH 9.5 and 10.5 respectively.




     3.  The radium selective complexer selectively removed radium from




         drinking water to concentrations below the MCL for extended lengths




         of time as long as iron fouling was eliminated.  Best results were




         for a pilot demonstration in Washington, Iowa where the RSC removed




         greater than 90 percent of the radium for eleven months.  Iron and




         manganese were not a problem at this location.




     4.  Low pressure «200 psi) reverse osmosis membranes will remove




         greater than 95 percent of the divalent cations Including radium.




         High pressure reverse osmosis units that are well maintained will




         give better than 981 radium removal.




  —  5.  Electrod!alysis reversal (EDR) removes about 40 percent of the




         dissolved solids, Including radium, per stage.  Ion profile,




         ion concentration and water temperature affect radium removal




         and therefore EDR stage configuration should be arranged far

-------
     specific waters on which they are used.




 6.  Good radium removal for some iron removal plants  using




     Greensand filtration has been reported.   In my opinion, a.




     properly designed KMntfy - filtration system should remove




     radium.  Ferrous ion and iron floe must  be removed first so they




     will not compete with radium or clog sorptlon sites.   About 1




     mg/L of Mn02 formed by the stoichiometrlc addition of  Mn   and




     KMnC>4 should be all that is needed.




COSTS DISCUSSION




     The water analysis for specific water supplies will determine the




     type of treatment plant that is required and, therefore, cost.




     For example, if the water needs radium removal and softening, then




     ion exchange or lime soda softening would be required.  If the




     water is brackish, then reverse osmosis  or electrodialysis rever-




     sal would be used and so forth.  ,'ome locations may require both




     iron removal and softening.  If the water only requires radium




     removal, then the RSC, KMnO^f lltration  or calcium cation exchange




     would probably be used.  Approximate costs for the different




     treatment processes that remove radium were taken from the liter-




     ature and tabulated in this report.
                                  10

-------
                            SECTION 3.




                 RECOMMENDATIONS FOR FUTURE RESEARCH






1.  Weak acid ion exchange resins should be tested in Che labora-




    tory for radium r-anoval at a few alkalinltles that are equlva-




    lently less than hardness.  Weak acid Ion exchange resins should




    be pilot tested for radium removal at two locations on waters




    of lower equivalent alkalinity than hardness.  The two locations




    could have alkallnlties that are equivalent!y equal to 1/3 and




    2/3 of the hardness.




2.  Calcium cation exchange has very good potential for removing




    radium and not hardness and, therefore, this process should be




    pilot tested for radium removal on waters where hardness removal




    is not desired.




3.  Potassium permanganate - filtration should be studied in the




    laboratory and pilot tested for radium removal.




A.  Sand should be studied in the laboratory to determine optimum




    conditions for radium removal.



5.  Electrodlalysis should be pilot tested on a brackish drinking




    water supply to determine radium removal efficiency and cost.




6.  The radium selective ccmplexer should be studied in the labora-




    tory to determine capacity for radium at low through high IDS




    concentrations.  A cost comparison should be made for the RSC,




    when used on influent vs. waste brine.  Methods for disposal of




    the RSC should be investigated.




7.  A cost study should be performed comparing radium removal with




    calcium cation exchange, KMn04-filtration and the RSC.
                                 11

-------
8.  A paper study should be made Co determine when  water  treatment




    waste containing radlun Is a problem.   Economical  ways  to handle




    this waste should be determined.
                                 12

-------
                              SECTION 4.




                             ION EXCHANGE




                    (University of Illinois Study)1






 SUMMARY OF ION EXCHANGE




      A study was completed at the University of Illinois to develop




 technology that can be used by snail water treatment plants to remove




 hardness, barium, and radium.  Special  emphasis was placed  on finding




 an alternative to strong acid ion exchange (which is used in the sodium




 (Na+) fora)  because of the large amounts of Na+ this process adds to




 treated water.  The primary objective of this study was to  determine




 the applicability of weak acid exchange resin (in the hydrogen (H+)




 form) for removal of hardness, barium,  and radium from the  types of




 groundwacer encountered in northern Illinois.  Ion exchange is particu-




 larly suited to small communities where hardness, barium, and/or radium




 may be a problem.  The capacity of the  weak acid ion exchange resin and




 the regeneration requirements were determined and compared  with those



 of strong acid resins for the same application.




      The strong acid resin in the sodium form was tested with an in-




 fluent water containing approximately 200 mg/L as calcium carbonate




 (CaC03) hardness, 250 mg/L as Ca(X>3 total alkalinity, 20 mg/L barium




 (Ba2+) and 20 plcocurles/llter radium-226 (pCl/L Ra-226).  The optimum




 regeneranC dose was found to be 6.5 Ib  sodium chloride (NaCl)/ft^




 (4.72 meq/g dry resin), which Is typical for softening applications.




•Application through several exhaustion/regeneration cycles  led to




 steady-state performance; approximately 215 bed volumes (BV) of water




 was applied before hardness breakthrough, and both Ba2+ and Ra-226  were




 removed to concentrations much lover than their maximum contaminant






                                   13

-------
levels (MCL's).  Ba2"*" broke through together with the hardness, and, at


a hardness concentration of 40 mg/L (at 225 BV), the 1-rag/L MCL for


Ba2+ had been exceeded.  Radlum-226 removal to less than the MCL of 5


pCl/L continued even after the resin was completely saturated with


hardness.  The regeneration efficiency (equivalents of metal ions re-
»

moved per equivalent of Nad applied) was about 60Z, column utilization


(percent of maximum capacity used per exhaustion cycle) was 602, and


2.4 BV of 16,400 IDS brine was produced per 100 BV of product water.


     fhe weak acid resin in the hydrogen form was tested as an alterna-


tive to the strong acid resin because the latter greatly Increases the


amount of Na+ in the product water.  The weak acid resin has a much


higher total capacity (11.5 meq/g versus 4.8 meq/g) than the strong


acid resin.  The weak acid resin must be followed by C0£ stripping and


ptf adjustment, and the resin does swell significantly as it is converted


from the H*" form.  Also the weak acid resin will only remove cations


that are matched by equal equivalents of alkalinity.  Application of


the same water used to test the strong acid resin was used for several


cycles of exhaustion-regeneration to observe steady-state behavior.


Both Ra-226 and Ba2+ were removed to levels far below the MCL's, and


Ba2"1" broke through about the same time as the hardness, after about 600


BV of product water had been processed.  Radium-226 continued to be


removed even after the resin was saturated with hardness.  The regen-


erant dose of 8.5 meq HCl/g dry resin yielded a column utilization of


70Z, a regeneration efficiency of 93 to 96Z, and 1.2 BV of spent brine


with a IDS of 19,900 o.^L per 100 BV of product water.


     The cost of uslnf the weak acid resin with HCL regenerant was


somewhat higher than that for using the strong acid resin in the Ma*
                                  14

-------
form.  For example, the use of a strong acid resin costs approximately

$1.36/1000 gal for a 0.1-MGD plant and $0.38/1000 gal for a 1-MGD

system.  The use of a weak acid resin will Increase these costs- by

$0.15/1000 gal and $0.08/1000 gal, respectively.

     In addition to the above, a strong acid resin in the calcium form

was found to effectively remove Ba^+ and Ra^* to concentrations below

the MCL.  The calcium form resin does not add sodium to the water and

it would be useful for situations where hardness removal is not desired,

     The following is a more detailed discussion of the University of

Illinois study.

RESULTS AND DISCUSSION

Definition of Column Utilization and Regeneration Efficiency

     The terms column utilization and regeneration efficiency will be

used in the following discussion on strong and weak acid resins and,

therefore, the definitions of these terms is given here.

 I Column Utilization - breakthrough capacity (meq/g) X 100         (1)
                        maximum capacity (meq/g)

where:

Breakthrough Capacity - total equivalents of metal Ions
                        removed from the water during the
                        service cycle per gram of resin

    Maximum Capacity  - maximum capacity of the resin
                        as stated by the manufacturer
                        in equivalents/gram of resin.
                        (11.5 meq/g for Diolite C-433)
                        (4.9 meq/g for Diolite C-20)

           Regeneration
           Efficiency » total equivalents of metal Ions removed during
                        total equivalents of ions applied during regen-

                        regeneratlon X JLOO
                        eraclon        *-                           (2)

Strong Acid Resins
                                  15

-------
     The strong acid resin is usually used in Che sodlun form and



involves a stoichiomecrlc exhaustion as shown in reaction (3).



Service:




2Na+ R~ + X2* + 2A~ —> X2"1^' + 2Na+ + 2A~                         (3)



Regenerate:



X2* R2~ + 2NaCl —> 2Na* R" + X2+ + 2 Cl~                           (A)



where:  R * resin



        X - Ra, Ba, Ca, Mg




        A - HC03, Cl, 1/2 S04



Regeneration of the resin is the reverse of the exhaustion process and



an excess of regenerant is required to force the reverse reaction (4)



to take place because the reactions are favorable in the direction



shown in reaction (3).



     The capacity of strong acid resins depends upon the number of



functional groups on the resin and also upon the amount of regenerant



applied.  If low regenerant doses are used, a lower capacity will



result? because some resin sites are not converted back to the Na+



fora.  Table 1 shows that smaller amounts of NaCl remove more cations



per unit of Nad added (higher regeneration efficiency).  However,



smaller amounts of NaCl used during regeneration result in a lower



resin capacity (column utilization) that can be used for hardness,



barium and radium removal.
                                  17

-------
    TABLE 1.  EFFICIENCY AND COLUMN UTILIZATION AS A FUNCTION OF
              REGENERATION LEVEL, Na+-FORM3

Regeneration Level
(Ib NaCl/ft3 resin)
1
2
3
5
10
15
20
Hardness Re
(Ib CaC03/ft3
Theoretical
0.85
1.70
2.55
4.25
8.50
12.75
17.00
moved
resin)
Actual
0.83
1.32
1.83
2.85
3.90
4.65
5.00
Regeneration
Efficiency
U)
98
78
72
67
46
36 '
29
Column
Utilization
m
14
22
30
47
64
76
82
     A strong acid resin, Duolite C-20 manufactured by Diamond Sham-

rock* was used and Table 2 gives the characteristics of this resin.


         —TABLE 2.  STRONG ACID RESIN CHARACTERISTICS

                                                  p-20

           Matrix'                            Polystyrene
           Crossiinking                       Dlvinylbenzene
           Functional Group                   Sulfonate
           Capacity                           4.9 meq/g
           Mesh Size                          16 x 50
           pH Range                           0-14	



Steady State Condition (strong acid resin) —

     Repeat runs were made using C-20 resin to determine barium and

radium removal at steady state conditions.  Test water had the analysis

given in Table 3.  The result of repeat runs for the strong acid resin

is shown in Figure 1.  Figure 1 gives effluent barium concentrations

when the effluent hardness concentration had broken through at 40 mg/L

as CaCC>3.  The figure shows that there was essentially no difference

between the hydrogen and sodium form resin for barium breakthrough as

*Now Rohn and Haas Co.
                                  18

-------
    I.S
   1.0
   as
                         4.61 Of 4O6ID MCI/ft*
                                         OOM<


                                         «J3lbNoCl/f!»
       MCL
                                        0—Na* fform Column
                                        * —M* Form Column
                                             (HCIKogtntrant)

                                     t   I   i   I   i   I   t  I  i
                            •     IO     12    14

                           NUMBER OF CYCLES
18
Figure  1.   Barium effluent concentration from strong  acid
    resin exhaustion-regeneration cyclic studies.  The
      barium values shown are  those when the effluent
        hardness concentration 1s 40 mg/L as CaCO*.
                                   19

-------
      TABLE 3.  TEST WATER ANALYSIS FOR STRONG ACID RESIN RUNS

          Parameter                            Concentration

             Na (mg/L)                               38.7
             Mg (mg/L as CaC03)                     102.2
             Ca (mg/L as CaC03)                     102.2
             Ba (mg/L)                               19.8
             Ra-226 (pCl/L)                          19.1
       Total Hardness (mg/L as CaCC^)               204.4
       Total Alkalinity (mg/L as CaC03)             251.5
             pH                                     6.5-7.2
shown for duplicate tests for the first five cycles.  A Nad applica-

tion of 6.5 lb/ft? of resin contains 4.72 meq/g of resin the same as

4.06 lb HC1/ ft^.  These results indicate that there was essentially no

difference In the regeneration efficiencies between NaCl and RC1 solu-

tions when equal equivalents of each regenerant were used.  The efflu-

ent barium concentration leveled off at 1.7 to 1.75 mg/L for the cycles

when 6.5 lb of NaCl/ft^ of resin was used for regeneration of the

sodium from resin column.  The effluent barium concentration dropped to

about 1.2 mg/L at the end of each cycle after the regenerant dose was

increased by 50t to 9.75 lb NaCl/ft3 of resin.  By the 19th cycle, 0.86

meq of Ba2+/g was on the resin before regeneration and 0.62 meq/g re-

mained after regeneration.

     The results shown In Figure 1 Indicate that barium (and hence

also radium) does accumulate on the strong acid resin with successive

exhaustion-regeneration cycles.  This accumulation levels off at a

concentration dependent on the total hardness concentration at which

the run is terminated, and on the amount of regenerant applied for each

regeneration.  In the 10th and llth cycles of the sodium form of the

resin, for example, Che divalent cation removal was 2.77 meq/g resin,

which is about 58* of the maximum capacity of the resin.  A large
                                  20

-------
 portion of the remaining 422 of the maximum capacity was occupied  by



 ions that were not removed by the regenerant.



      The regeneration efficiency (equivalents  removed per equivalent of



 regenerant applied), obtained for the 10th and llth cycles with 6.50 Ib



 tlaCl/ft3 of resin was 58 to 59%.  The average  TDS of the 2.4  BV of



 waste brine produced per 100 BV of product water was 16,400 mg/L.   The




 waste brine figure does not include the backwash water which  would be



 used to backwash the resin beds for each regeneration but does  Include



 the rinse water.  No effort was made to minimize the volume of  brine



 produced.



      The regeneration efficienty obtained for  the 17th, 18th, and  19th



 cycles (see figure 1), with application of 9.75 Ib NaCl/ft3 of  resin,



 was 46.OX.  The waste brine produced was 2.07  BV per 100 BVs  of product



 water; the average TDS of the brine was 25,200 mg/L.  The capacity used



 during these cycles was 3.26 raeq/g for a column utilization of  67.9Z.



      The breakthrough curves shown in Figure 2 were determined  during



 the 6th exhaustion of the sodium form resin column, with 6.50 Ib Nad/



 fr* of resin applied as regenerant.  The leakage of barium during  most



 of the cycles was very low, approximately O._0 mg/L, Including  at  the



 beginning of the run.  Very low total hardness leakage was also observed



 for this cycle.  Barium and total hardness broke through together  at




 approximately 225 BV.  Effluent samples taken  during the 9th, 10th,  and



 llth cycles of this column continued to show leakage of less  than  about




.0.15 mg/L of barium during most of each cycle.



      The breakthrough curves determined during the 19th cycle of the



 sodium form resin column, with a regenerant dose of 9.75 Ib Nad/ft^ of



 resin, were very similar to those shown in Figure 2.  The leakage  of






                                   21

-------
  40
                                                             2.0
       TOTAL HARDNESS - Q
               BARIUM - •
x
M
30
   20
   10
                                                                15
                                                                 M
                                                                 J


                                                             1.0
                                     _L
                                              JL
               50         100        150        200
                  VOLUME Of INFLUENT (BED VOLUMES)
                                                                   CD
                                                             0.5
                                                        250
            Figure  2.  Breakthrough curves for the sixth
              exhaustion cycle of the strong add resin.
                                 22

-------
 barium was between 0.10 mg/L and 0.20 mg/L, for che entire run.  Hard-

 ness and barium did breakthrough somewhat later at 260 to 270 BV,

 however, in keeping with the usage of a higher percentage of  the resin

 capacity.

      The leakage shown in Figure 2 indicates that the barium  remaining

 on the resin after the completion of the regeneration/rinse portion of

 the cycle is tightly held by the resin.  The barium is not easily

 displaced by the high concentration of Na+ released in the subsequent

 saturation cycle as hardness Is removed from the Influent.

 Regeneration (strong acid resin) —

      The regeneration of each column was performed upflow, counter-

 current to the direction of flow during the exhaustion run.  The resin

 remained firmly packed In the columns at all times, however.   Effluent

 samples were taken continuously throughout regeneration and analyzed

 for magnesium, calcium, and barium;  the HCl regeneration curves  for the

 acid form of the resin are shown in Figure 3.  The curves show the

 cumulative metal ion removal as a function of the amount of regenerant

 applied.  Calcium and magnesium are removed most effectively  during the

 first portion of the regeneration.  At a regenerant dose equivalent to

 the resin capacity (at line A), 55Z of all ions were removed  (based on

 the total amount eventually removed from the resin during regeneration),

 80Z of the magnesium and 67Z of the calcium were removed.  Barium  was

 the most difficult ion to remove.  Only 15Z of the barium ions were

 removed by this amount of regenerant.  After an amount of regenerant
»
 equal to three times the equivalents of divalent cations removed (line

 C) was applied, essentially all of the calcium and magnesium  were

 removed from the resin - but more than 50Z of the barium remained  on

 the resin.

                                   23

-------
   400
a  300 -
   200 -
   100 =
               900       IOCO        1300

                CUMULATIVE  REGENCRANT APPLIED (miq of HCI)
 Figure  3.  Add  regeneration of virgin strong add resin (2).
   At A, B, C and 0, respectively, the amount of regenerant
   applied Is 1,  2, 3 and 4 times the number of equivalents
                of cations removed by the resin.
                                24

-------
      The regeneration curves for Che sodium chloride  regeneration of


 the column origin-ally in the salt  form are very similar  to  those


 obtained for the hydrochloric acid regeneration and thus they are not


 shown.   No difference in she regeneration ability of  the hydrochloric


 acid versus sodium chloride was observed.

 •
      A  large regenerant  dose would be required  to regain the  full


 capacity of this strong  acid resin.  In municipal softening operations


 resin capacity is often  sacrificed in order to  achieve greater regener-


 ation efficiency.2  Municipalities often operate with a  regenerant dose


 of approximately the capacity of the resin (or  the equivalents of


 divalent ions on the resin) where  about 50Z removal of the  ions Is


 obtained according to the results  shown In Figure 3.  The capacity of


 the resin available in a subsequent production  run is then  limited by


 cations still present on the resin.  The results from two virgin resin


 column  studies also indicate that  with only partial regeneration of a


 resin bed between service runs, as in a municipal softening application,


 barium  will accumulate on the resin.  This condition  is  expected to


 alter the relative amount of water that can be  processed before hard-

 ness and barium breakthrough.


 Radium  Removal (strong acid resin) —


     Table 4 gives the results of a test for radium removal  using virgin


 strong  acid resin in the hydrogen  form. Influent Ra-226  averaged 18.5


 pCi/L for this test. Table 4 gives effluent Ra-226 concentrations along


 with total hardness, barium, and sodium concentrations in the effluent


'samples.  The data show  that Ra-226 removal continued long  after the


 saturation of the column with hardness ions.
                                   25

-------
TABLE 4.  RADIUM REMOVAL, SECOND STRONG ACID VIRGIN RESIN COLUMN TEST
                      (Influent Ra-226 average - 18.5 pCi/L)
Influent
Applied
(BV)
25
380
• 881
1121
1622
Effluent
(pCl/L)
1.4
2.1
2.3
2.0
2.7

Effluent
Na* (mg/L)
0
112
36
46
47
Effluent
Hardness
(mg/L as CaCCh)
0
0
214
200
201

Effluent
Ba2+ (mg/L)
0
0
0
0
1.0
These samples were analyzed In the Environmental Research Laboratory,
University of Illinois, Urbana, IL.
     Influent, effluent, and regeneration samples from several  cycles

of the steady state scdluta form resin column test were analyzed for

radluo-226. Using one liter samples and the radon emanation technique,

the detection limit for Ra-226 is 0.1 pCl/L.  The regenerant dose

(shown in Figure 1) was 6.5 Ib Nad/ft^ of resin for runs 1 through 11

and 9.75 Ib Nad/ft^ of resin for runs 12 through 19.  Results  are

given in table 5.  Excellent radium removal by the column wi s observed

for each of the runs. After the 11 cycles at the lower regenerant dose,

the column still produced product water with an average radium-226

concentration of 0.1 pCi/L.  After eight additional cycles of the column

at the higher regenerant dose, the average radium-226 concentration in

the product water was below 0.1 pCl/L.  Throughout each cycle the

radium-226 concentration in the effluent was quite low, and at  no point

did the concentration exceed the MCL of 5 pCi/L.  After comparing the

data in table 5 to the data in table 4 and figure 1 at approximately

the same hardness breakthrough, one can conclude that the strong acid

resin perfers radium over barlun and preference of the strong acid

resin for cations is Ra > Ba > Ca > Mg > Ma.
                                  26

-------
          TABLE 5.  RADIUM REMOVAL, STRONG ACID RESIN CYCLIC TESTS


Cycle No. 1
* Influent
Point Effluent
Average Effluent
Cycle No. 4
Influent
Point Effluent
Point Effluent
Average Effluent
Cycle No. 6
Influent
Point Effluent
Point Effluent
.Point Effluent
Average Effluent
Cycle No. 11
Influent
Point Effluent
Point Effluent
Point Effluent
Average Effluent
Cycle No. 19
Influent
Point Effluent
Point Effluent
Point Effluent
Poinc Effluent
Average Effluent
Influent Applied
(BV)


420
425


8
264
270


6
164
254
259


27
182
248
255


5
216
261
287
295
Radium-226
(pCi/L)

19.1
0.3
0.1

8.4
0.3
0.2


17.7
<0.1
0.3
0.3
<0.1

18.2
<0.1
<0.1
3.0
0.1

18.4
0.2
<0.1
0.1
0.7
<0.1
Total Hardness
(mg/L as CaCCh)

195
39
2.3

209
1.1
43
2.6

199
1.7
1.2
33
1.6

216
0.8
0.8
23
1.7

200
0.4
0.4
4.6
26
1.6
Cycles 6 through 11 were analyzed in the Environmental Research Labor-
atory, University of Illinois, and cycles 1 through 4 were analyzed in
Che University Hygienic Laboratory, University of Iowa.
                                  27

-------
       Table 6 shows the anount of radium-226 removed by column regener-



  ation.  An accumulation of radium-226 on thi resin was observed,  but



  the accumulation per cycle decreased with each successive cycle.   After



  11 cycles at a regenerant dose of 6.5 Ib NaCl/ft.3  of resin, almost 852



  of the radium-226 adsorbed during the llth cycle was removed.  After
  »


  eight additional cycles at a regenerant dose of 9.75 Ib NaCl/ft^  of



  resin, close to 95Z of the radium adsorbed during  the 19th (and  last)



  cycle was removed by regeneration.  Hence radium accumulates on the



  resin but eventually comes to equilibrium where all of the radium



  adsorbed will be removed during regeneration (assuming that the length



  of service and quantity of regenerant remain constant).



       The small accumulation of radium-226 on the column will not  affect



  resin capacity.  During the entire 19 runs of the  cyclic test,  29,800



  pCi of radium-226 was passed through the column.  Even if all of  the



  radium accumulated, only 2.7 x 10~? meq of radium  would be on the



  column; a negligible amount when considering the resin capacity of



  approximately 5.0 meq/g.





         TABLE 6.  RADIUM ACCUMULATION, STRONG ACID  RESIN CYCLIC TESTS
Run
No.
4
6
11
19
Waste Brine
22&Ra (pCl/L)
274
418
639
846
Waste Brine Produced
(BV/100 BV Product Water)
2.30
2.37
2.41
2.05
Percent 226^ Removed
During Regeneration

56.0
84.8
94.4
— Weak Acid Resins



       Weak acid resins remove only cations chat  are balanced  with alka-



  linity and Involve stoichionetrie exhaustion as shown in reaction (5).
                                    28

-------
Service:
SH^R" + X2* + 2HC03~ —> X2* R2~ + 2C02(aq) + 2H20                  (5)

Regeneraclon:

X2*R2~ + 2HC1 —> 2H+R~ + X2+ + 2C1~                                (6)

where:  R  -  resin
        X  -  Ra, Ba, Ca, Mg

Regeneration takes place as shown by reaction (6).  Regeneration for

weak acid resins is very efficient and requires only about ten percent

above stoichiometric excess regenerant.  Regeneraclon requires applica-

tion of a strong acid such as hydrochloric or sulfurlc.  Sulfurlc acid

is less expensive Chan hydrochloric acid but regenerating with sulfurlc

acid could cause CaSC>4 precipitation within the resin.  Therefore, the

most economical regenerant would be a mixture of dilute hydrochloric

and sulfurlc acids Co eliminate CaS04 precipitation.  Sulfurlc acid

should not be used for regeneration if barium is present because of the

possibility of BaSC>4 precipitation in the resin.

     Carbon dioxide, produced during Che service cycle, must be removed

by stripping.  Reaction (5) will not go to the right unless alkalinity

Is present because pH of Che solution drops below 4.5, if alkalinity is

not present, and the concentration of H* becomes large enough Co cause

Che release of divalent ions from che resin according Co reaction (6).

The addition of a small quantity of Na2C03 is possible to Increase

alkalinity if enough natural alkalinity is not present for Che desired

divalent cation removal.

     Some of the advantages of the weak acid resin include ease of

regeneration and Che consequential smaller quantity of brine per unit

of water treated.  Downing* obtained the best regeneration with 1 Co 4

percenc HC1 solutions.  If H2S04 is used, lower concentrations should


                                  29

-------
 be used to eliminate the possibility of CaSO^ or BaSC>4  precipitation in

 the resin bed.   Dow Chemical U.S.A. (1977)  suggests using 0.119  Ib of

 100 percent l^SO^/eq capacity utilized as a 0.5  to 1 percent H2$04

 solution, but they did not address the concentration of l^SO^ required

 if Ba2+ were present.
 •
      There is a drinking water softening plant in Cheam, England that

 uses the weak acid resin^ for softening.  After  a cost  analysis  of

 alternative softening options Including lime-soda softening, and after

 excluding any processes that released sodium Into the drinking water,

 the authorities of Cheara decided to use the weak acid resin.

      A weak acid resin Duollte C-433 manufactured by Diamond Shamrock

 was used for the University of Illinois study and Table 7 gives  the

 characteristics of this resin.


              TABLE 7.  WEAK ACID RESIN CHARACTERISTICS
                                                 C-433
            Matrix                        Polyacrylic
            Crossiinking                  Dlvinylbenzene
            Functional Group              Carboxylic
            Capacity                      11.5  meq/g
            Mesh Size                     16  x  50
            pH Range                	5 - 14	
 Note that the capacity of the weak acid resin is 11.5  meq/g and  much

 higher than the strong acid resin which had a capacity of 4.9  meq/g.

 Steady State Condition (weak acid resin) —

      Three sets of cyclic studies were performed to determine  steady

•state characteristics that could be expected under full scale  operation.

 A column with new resin was constructed and cyclic tests with  the  weak

 acid resin were performed.  The column parameters and  influent charac-
                                   30

-------
 Cerlscics are given In Cable 8.   Flow races and  ocher parameters  are

 characteristic of full scale field  condiclons.   Note  chat  che  alkalln-
      TABLE 8.   EXPERIMENTAL PARAMETERS FOR THE CYCLIC  TESTS  WITH
                                 WEAK ACID RESIN

                                       Tesc II       Tesc  #2     Tesc  #3
 Column length (cm)                         60.6          60.6         60.6
 Column diameter (cm)                        2.5           2.5          2.5
 Flow race (gpm/ft2:  BV/hr)            4.3;  17.5     4.0;  16.4    4.1;  16.5
 One bed volume (cm3;  g dry welghc)   292;  110.7    292; 110.7   292;  110.7
 Na+ (mg/L)                                24           51           25
 Mg,  (mg/L as CaC03)                       92           98          100
 CaZ+ (mg/L as CaCO,)                       97         102           98
 Ba2* (mg/L)                               19           20           22
 Tocal hardness (mg/L as CaC03)            195         200          198
 Total alkalinity (mg/L as CaC03)          250         250          250
 pH                                   6.8  - 7.0     6.8 - 7.0    6.8  - 7.0
 ley of 250 mg/L as CaC03 is greater Chan hardness  of approximately 200

 mg/L as CaC03 and, therefore, Che weak acid resin  would be  capable of

 removing all of Che divalent cations.  Using Che weak acid  resin  over

 Che strong acid resin would be a good choice if alkalinity  is greacer

 Chan hardness.

      The resales of cyclic cests nos. 1, 2, and 3  are given in  Tables

 9, 11, and 12.  The definitions for column utilization and  regeneration

 efficiency were given In equations (1) and (2). Breakthrough capacity

 for the nine exhaustion cycles In Table 9 varied from 7.9 Co 8.9  meq/g.

 In the first five cycles of cyclic test #1 (Table  9) regeneration

 efficiency varied from 76 to 88Z when 10.4 meq HCl/g were used.  These

•relatively low values of regeneration efficiency were caused by applying
                                   31

-------
                TABLE 9.  WEAK ACID EXHAUSTION-REGENERATION
                           CYCLIC TEST #1
Breakthrough Column Regeneration Terminal
Cy- Capacity Utilization* Efflciencyb»c Hardness*1
cle (meq/g) Z Z (mg/L as CaCCh)
1 8.9
2 8.4
'3 8.4
4 8.8
5 8.0
6 7.9
7 8.2
8 8.0
9 8.3
a Maximum capacity:
b Regenerant applied
c Regenerant applied
77
73
73
77
70
69
71
70
72
11.5 meq/g
in cycles 1-5:
in cycles 6-9:
79
76
81
88
79
88
91
91
95
10.4 meq/g
9.0 meq/g
41
60
42
68
54
44
47
42
47

Terminal
. Barium
• (mg/L)
0.2
0.2
0.1
0.1
0.4
0.1
0.3
0.2
0.2

d  Hardness is all magnesium


more regenerant than was necessary.  After lowering the regenerant to

9.0 meq HCl/g of resin, the efficiency of regeneration Increased to an

average of 91Z over the last four cycles.  For all nine cycles, the

terminal barium concentration did not exceed 0.1 to 0.4 mg/L.  After

the ninth exhaustion-regeneration cycle t:as completed, an additional 5*2

meq HCl/g of resin was applied (1.71 bed volumes of 3.58Z (0.99N) ) to

remove any ions that remained on the resin.  The total mass of magnesium,

calcium, and barium removed during this final regeneration Is listed in

Table 10.  After a cumulative total of 74.6 meq/g of metal Ions were

adsorbed over the cine cycles (avg • 8.3 meq/g/cycle), only 1.2Z

(0.097 meq/g) remained after the last cycle.  This shows that with a

regenerant dose of about 10Z higher than the capacity utilized by

divalent ions, the weak acid resin can be regenerated to essentially

virgin capacity.
                                  32

-------
          TABLE 10.   IONS REMAINING ON THE WEAK ACID RESIN AFTER
                 THE FINAL REGENERATION OF CYCLIC TEST #1
               removed                              0.06  meq/g

               removed                              0.03  meq/g

               removed                              0.007  meq/g
      In cyclic tests #2 (table 11)  a set of six exhaustion-

 regeneration cycles were performed  using a regenerant dose of  7.6  meq

 HCl/g of resin.  This regenerant dose was chosen to determine  column

 performance when some of the metal  ions were not removed from  the

 column during regeneration. The breakthrough capacity decreased  sharply

 between the first and second run.  The column was incompletely regener-

 ated after the first exhaustion run and thus the resin still contained

 some metal ions at the start of the second exhaustion run.  This incom-

 plete regeneration also resulted in a decrease in the percent  column

 utilization (from 75Z to 64%), an increase in regeneration efficiency

 to approximately 10 OZ, and increased Ba^4* leakage at the start and end

 of the exhaustion run.  The final barium concentration in the  effluent

 at breakthrough after six column runs was 1.3 mg/L.

      A seventh column run was also  run as a part of cyclic test  12.

 Whereas, the first six column runs  were completed at hardness  break-

 throughs between 40 and 77 mg/L CaC03 the seventh column run was car-

 ried out to 96 percent saturation with divalent metal ions. Results of

-exhaustion run seven are shown in Figure 4.  Effluent profiles for

 calcium magnesium and barium are very similar to those found in  column

 test with virgin resin (virgin resin breakthrough curves for week  acid
                                   33

-------
 resins  are given in the  full  report,  reference 1)  with the  exception

 that  all  of the  ions appear in the  effluent  sooner.   This was  expected

 because a portion of the exchange capacity was not available due  to

      TABLE 11.   WEAK ACID EXHAUSTION-REGENERATION CYCLIC TEST #2
      Breakthrough     Column     Regeneration      Terminal      Terminal
 Cy-    Capacity    Utilization3  Efficiency**      Hardness       Barium
 cle  (meq/g)           Z               Z	   (mg/L  as  CaCO-Q   (mg/L)
1
2
3
4
5
6
8.7
7.8
7.8
7.6
7,4
7.4
75
68
68
66
64
64
IDS
103
101
101
101
102
51
47
77
70
50
40
0.5
0.8
1.9
1.6
1.3
1.3
 a  Maximum capacity:  11.5  meq/g
 b  Regenerant applied:   7.6  meq HCl/g
        TABLE 12.  WEAK ACID EXHAUSTION-REGENERATION CYCLIC TEST  #3

Cy-
cle
1
2
3
4
5
6
Breakthrough
Capacity
(meq/g)
8.0
7.8
7.8
7.8
7.8
8.2
Column
Utilization3
Z
. 70
68
68
68
68
71
Regeneration
Efficiency*
Z
95
96
95
96
93
94
Terminal
Hardness
(mg/L as CaCO-0
45
42
46
42
38
43
Terminal
Barium
(mg/L)
0.2
0.0
0.1
0.1
0.1
0.2
 a  Maximum capacity:  11.5 meq/g
 b  Regenerant applied:  8.5 meq/g
 incomplete regeneration.  Figure 4  shows that there is hardness  and

 barium leakage at the beginning of  the column run.   Hardness and barium

.in the effluent at the beginning of the column runs were a result of

 metal ions left on the column after the previous regeneration.   Metal

 ions that are not removed during regeneration will  be concentrated at
                                   34

-------
 9*9
5
-8

 8
   200-
I"0
55
g I20r-

I
o  eoj—
<

5  4dr-
              EFFLUENT
            CONCENTRATION

               0-M9«*

               A-Co«»
                                           INFLUENT CALCIUM
                                          INFLUENT MAGNESIUM
100



 80


     o>
 CO  J




 401
    0

 20
                       900               MOO

               VOLUME OF INFLUENT (BED
    Figure  4.  Weak acid resin breakthrough  curves for run #7
       of cyclic run 12.  Regenerant dose was 7.6 meq HCI/g.
                               35

-------
 the effluent end  of  the resin column and  these  ions  will  come  off  the


 resin at the beginning of  the next  cycle.


      Cyclic test  #3  (Table 12) consisted  of  six exhaustion-regeneration


 cycles that were  performed with a regenerant dose  of 8.5  meq HCl/g of


 resin.  This regenerant level was chosen  to  examine  the column perform-
 •

 ance at a level of regenerant intermediate to the  previous  two doses.


 The results are listed in  Table 12.   Breakthrough  ringed  from  7.8  to


 8.2 meq/g and column utilization ranged from 68 percent to 71  prcent.


 There was no sharp decrease in breakthrough  capacity over the  first two


 runs as in cyclic test #2.  This indicates that most of the metal  ions

 were removed during  regeneration and near virgin capacity was  retained


 from cycle to cycle.


      Regeneration efficiency averaged 95Z for the  six-column run,  which


 was slightly lower than that experienced  in  cyclic test #2, in keeping


 with the application of a  greater amount  of  regenerant than metal  ions


 removed.  The waste  brine  volume was 1.2  BV/100 BV of product  water


 with a TDS of 19,900 mg/L.


      The effluent barium concentrations ranged  from  0 to  0.2 mg/L  for


 breakthrough hardness concentrations of 46 mg/L as CaCC>3  or less.


 Barium concentrations in this range are at the  lower end  of the detec-


 tion limits of ion chromatography,  and therefore it  was difficult  to


 measure exact concentrations.  However, it is safe to conclude that


 barium removal was more than acceptable for this set of runs.


      Hardness and barium concentration profiles are  shown in Figure 5
»
 for the sixth exhaustion run of cyclic test  #3. There was only minor


 hardness and barinn  leakage at both the beginning  and end of the run;
                                   36

-------
I
8
            0-BARIUM
            • •HARDNESS
                                             MCL	Vs	
                    I
                   ZOO            400             600

                 VOLUME OF INFLUENT (BED VOLUMES)

       Figure  5.  Weak add resin breakthrough curves for run 16
         of cyclic  run #3.  Regenerant dose was 8.5 mcq HCl/g.
              2.0
          5   1.5
              1,0
              0.9
          Ul
REGENERANT DOSE
   •-9.0me«i/g
   0—7,6m«q/q
   A-8J miq/g
                                                    L
                            246
                           ..NUMBER OF CYCLES
         Figure  6.  Barium concentration at the termination
                  of weak acid resin cyclic tests.
                                      37

-------
thus it appears that a still lower regeneranc dose, perhaps 8.0 meq



HCl/g, would give satisfactory performance.



Regenerant dose —



     The barium removal performance of each of Che cyclic tests is



summarized in Figure 6.  The final barium concentrations are plotted



tor Che last four runs of cyclic test #1 and for the first six runs



each of cyclic tests 92 and #3.



     For regeneranc doses of 8.5 and 9.0 meq HCl/g, the barium concen-



tration levels off considerably below the HCL of 1.0 rag/L.  On the



other hand, when a regeneranc dose of 7.6 meq HCl/g was applied as in



cyclic cesc 92, concentrations of barium were found Co be greater Chan



1.0 mg/L.  The optimum dose for Che water treated in Chese experiments




lies between 7.6 and 8.5 meq/g, and quite likely is about 8 meq/g.



This value will also be a function of the quality of water treated and



probably will vary with Che ratios of barium Co hardness, hardness Co



alkalinity, and hardness Co ocher divalent metal ions. Further research




is needed to establish these functions.



Radium Removal (weak acid resin) —



    The results of a column run using virgin weak acid resin for Ra-226



removal are given in table 13.  The Influent radium concentration



averaged 18.6 pCi/L while Che Influent hardness concentration averaged



193 mg/L as CaCQ$.  Effluent radium concencratlons (Table 13) are below



the MCL of 5 pCl/L.  As with the strong acid resin, there was no signif-



icant increase in Ra-226 concentration long after hardness breakthrough.




     Since radlum-226 removal occurred even after the resin was satura-



ted with hardness -luring virgin resin column tests (Table 13), it



appears that the weak acid resin has a very high selectivity for radium.
                                  38

-------
    TABLE 13.  RADIUM REMOVAL, WEAK ACID VIRGIN RESIN COLUMN TEST
           (average influent radium concentration - 18.6 pCl/L)
Bed Volumes
of Influent
25.6
426.2
671.9
915.5
986.6
1167.1
Effluent Hardness
(mg/L as CaC03)
<1.0
<1.0
<1.0
108
148
198
Effluent Radium
(pCi/L)
2.5
2.8
3.1
3.0
3.2
3.0
NOTE:  These samples were analyzed in the Environmental Research labor-
       atory, University of Illinois, Urbana, IL.
To determine whether the Ra-226 was held so tightly that it could not

be removed during regeneration, the mass of Ra-226 removed during

regeneration was determined and compared to the amount that was applied

(s.ee table 14).  The mass of radium :pplied was determined from the

influent concentration and the volume of influent applied, while the

mass of radium removed was measured by the volume of spent regenerant

brine and rinse and its concentration.  The amount that passed through

the column during the exhaustion cycle was assumed to be negligible.

Table 14 shows that the mass removed is approximately equal to the mass

applied so there Is no indication of significant accumulation of radium

on the weak acid resin.
                                  39

-------
         TABLE u.  RADIUM MASS APPLIED vs. MASS REMOVED DURING
                    REGENERATION OF WEAK ACID RESINS


Cyclic Test
2

3



Cycle
4
6
5
6
Mans Mass
Applied During
(jpCl)
3735
3125
3456
3627
; Removed
Regeneration
(_pCl)
—
3830
3479
3181
I Applied
that was
Removed

123
101
88
NOTE:  Samples were analyzed by the Environmental Research Laboratory,
       University of Illinois and the University Hygienic Laboratory,
       University of Iowa.

     The performance of the weak acid resin for radium-226 removal  was

also measured during the three cyclic column tests (Tables 9,  11  and

12)*  Influent and effluent samples were analyzed for a number of the

runs in each of the three tests; the effluent samples were composites

of the entire exhaustion cycle.  The results are given In Table 15.

Radium removal percentages were comparable for all three cyclic tests,

and varied between 90-982.  Regenerant dose varied from 7.6 to 9.0  meq

HCl/g and seemed to have little effect on radium removal.  Effluent

radium concentrations varied from 0.4-2.0 pCi/L, which are well below

the MCL of 5 pCi/L.  These results are similar to those for the strong

acid resin.
                                  40

-------
      TABLE 15.  RADIUM REMOVAL, WEAK ACID RESIN CYCLIC TESTS
Cyclic Test *
1

2

3

Run 4
3
4
4
6
1
2
Influent
19.3*
16.1*
18.5
16.0
19.1
20.0
Effluent
2.0*
0.8*
0.7

0.4
0.6
Z Removed
90
95
96

98
97
*130 mL samples were analyzed for these samples, all others were
 1000 m,.

 NOTE:  Samples were analyzed by the Environmental Research Laboratory,
        University of Illinois, Urbana, IL.
Calcium Cation Exchange

     Complete removal of hardness from water is often not appropriate

or desirable.  Low-calcium waters are corrosive to some extent and

removing this low quantity of hardness in order to remove barium and/or

radium may be undesirable.  The strong-acid resin has a higher selec-

tivity for barium and radium than calcium and magnesium and thus a

resin in the calcium-form will selectively remove barium and radium

from water..^  Tests for barium removal were run on calcium-fora resin

using water with the influent composition given In Table 16.  The

number of bed volumes that can be processed depends on the quantity of

calcium chloride regenerant applied as shown in Figure 7.  Application

of 4, 6, and 8 equivalents of calclum/L of resin resulted in runs of

500, 900, and 1000 bed volumes respectively, compared with 1,200 bed

volumes for virgin resin when influent barium concentration was ?3
                                                f
mg/L.  The trade off between resin capacity per run and regeneration

efficiency is shown in Figure 8.

-------
                 TABLE 16.  COMPOSITION OF INFLUENT WATER
           Parameter                             Concentration '

    Total hardness                               215 mg as Ca(X>3/L
    Calcium                                      105 ng as CaC03/L
    Magnesium                                     95 mg as CaC03/L
    Barium                                        23 mg/L
    Sodium                                        23 mg/L
    Chloride                                      10 mg/L
    Total Alkalinity                             250 mg as CaCC>3/L
    pH                                             7
     The ability of a calcium-form resin column to remove radium from

water was  evaluated through four exhaustion-regeneration cycles.

The fifth cycle consisted of exhaustion only.  Influent water was

similar to that shown in Table 16, except that 43 pCl radium/I, replaced

the barium.  Exhaustion in 4.7-in (12 cm) column was carried out at

1.35 gpm/ft^ (3.4 m/hr) for 500 bed volumes.  Termination of the run at

500 bed volumes was arbitrary; additional runs are needed to establish

the number of bed volumes that can be processed to radium breakthrough.

Regenerant brine contained 34,000 mg/L  calcium and 4,860 mg/L magnesium.

The brine-loading rate ws 0.29 gpm/ft2 (0.7 m/hr), and the dose was 6

equivalents of calclum/L resin (27.51b CaClj*2H20/ft  rssln).  The

spent brine from each cycle was reclaimed and reused in the next cycle.

Rinse-water volume was 8.5 bed volumes.

     The average radium concentration in the effluent In each of the

five exhaustion runs was <0.5 pCl/L (98.8 percent radium removal).  The

amount of radium placed on the column during exhaustion was 45 pCi/g

dry resin (0.02 uCl/L resin).  Radium capacity at breakthrough was not

determined.


                                 42

-------
a
   OS
            A Virgin Resin
            0 8equittCa/L Rtsin
            « 6 equiv. Co/L Rtsin
            o 4 equiK Ca/L Resin
           200
000    MOO
             400     «00     too    1000
             Bad Volumes of Water Treated

Figure 7.  Barium breakthrough curves at different
    regenerant  dosages.
      KOO
I
S   12
                   461
             Regent on< Dosogt, «q.Co**/L Risin

     Fifurt  8.  RegencraCion efficiency and column
         capacity at  Tarious regenerant dosage*•
                                                         •1
                                                          *<
                                                         .1
4  *
  *"

  1
  jt
  uj

-------
     The spent CaCl2 brine can be reclaimed for reuse.  Addition of 10



percent molar excess solid CaS04 (relative to the barium in the brine)




resulted in reduction of the barium concentration to about 100 mg/L.



The CaS04 dissolved, and BaSCfy precipitated.  Use of a very soluble



sulfate salt such as sodium sulfate was not successful because high



localized concentrations of Sulfate caused CaSO^ as well as BaSO^ to



precipitate.  Reuse of brine was possible after removal of Che precip-



itate by filtration.  The concentration of magnesium in the brine



increased through successive cycles until a plateau value was reached.



This resulted in a column that was partially in the magnesium form at



the beginning of an exhaustion run.  This does not pose a problem



because barium and radium can replace magnesium mere ecslly than cal-



cium.  If the brine contains radium as well as barium, the radium will



also be removed by copreclpltatlon with the BaSCfy.  However, a barium



salt such as BaCL2 will have to be added along with the CaSO^ to the



spent Cad2 brine containing only radium to achieve radium removal.



     Additional research la needed to refine the brine reclamation



process.  In particular, the base procedure to precipitate and separate



barium and radium needs to be established*  A procedure is also needed



to control the precipitation process to ensure that the barium and



radium have been removed and that too much sulfate has not been added.



The brine reclamation process should significantly reduce the brine



disposal problem, but ways of disposing of the precipitate must be



found, and the cost involved in using the process must be established.



     Strong acid resins remove practically all of the hardness from the



water.  Complete hardness removal is not necessary or desirable and
                                  44

-------
therefore blending of treated and untreated water is usually practiced



to prevent pipeline corrosion and reduce costs.  Waters with less than



SO mg/L as CaC03 hardness are considered soft and waters with up to



ISO mg/L as Ca003 hardness are considered moderately hard.  Blending is



usually practiced to obtain a finished water hardness of about 100 mg/L
•                                                             '      •


as CaCOj.  If blending Increases the radium concentration above the



MCL, then calcium cation exchange should be used to remove radium from



the by-pass water.
                                  45

-------
                               SECTION 5

                       IRON AND MANGANESE REMOVAL
                       (University of Iowa Study)?
SUMMARY OF IRON AND MANGANESE REMOVAL

     Previous field studies^*^*^ have shown that anywhere from 10 to

53Z radium removal takes place across iron removal processes that

consist of aeration, detention, and filtration.  The primary objective

of the University of Iowa Study was to determine the variables that

control this Incidental removal of radium and to evaluate the possi-

bility of expolicing these factors to provide an inexpensive means of

removing radium using existing or modified iron removal facilities.

This study investigated how water chemistry Influences Ra-226 sorptlon

to iron and manganese oxides and sorptlon to filter sand.

   .  The potential for significant radium sorptlon to hydrous iron

oxides under typical ground water conditions is limited by too low an

iron concentration in the presence of barium, calcium, and magnesium

which are believed to compete for sorptlon sites.  Sorptlon to iron

oxides Increased with pH but significant sorption to iron oxides at

Iron concentrations typical of natural waters would require excessive

pH values.  Batch studies using freshly precipitated iron oxides,

formed in the laboratory, showed that sorptlon of radium from typical

groundwaters is generally expected to result in maximum removals of

approximately 10 to 20Z and should be much less.

     Radium removals obtained by freshly precipitated hydrous manganese

oxides in batch studies were much greater than those obtained in

systems containing only iron oxides.  Radium removals by manganese

oxides were significantly greater in hardness free waters because of
                                  46

-------
less competition by hardness ions.  In the absence of iron, 1.0 mg/L



MnO£ removed approximately 65Z of Che Ra-226 from Oxford, Iowa water




(hardness - 1300 mg/L as €3003) at pH 7.5.  Sorption to MnC>2 appears to



be a good radium removal mechanism if sorption of ferrous Ion and



coating of the manganese oxides with iron oxides can be eliminated.



     Filter sand has a potential capacity to sorb significant concen-



tration of radium at typical hardness concentrations if the capacity is



maintained by periodically rinsing the sand with a dilute acid.  Re-



moval efficiencies of approximately 80-90Z were achieved in labora-



tory and field studies using a 61 cm (2 ft) deep sand bed at conven-



tional loading rates when daily rinsing with a dilute acid was prac-



ticed.  Radium sorption to filter sand increases with pH but is not



very sensitive to pH in the range of pH 5 to 8 or to hardness concen-



trations above approximately 300 mg/L as CaCt>3.



     The following is a more detailed discussion of the University of



Iowa Study.



TESTING PROCEDURES



     Batch studies, laboratory pilot plant studies and field pilot



plant studies at Oxford, Iowa were performed.



     Batch studies as shown In Figure 9 were conducted at 25°C using 2



liter beakers stirred with a Birds and Phlpps gang stlrrer.  Mixtures



of varying carbon dioxide and air were bubbled through to provide pH



control although pH was also adjusted in several experiments by adding




concentrated sulfuric acid or sodium hydroxide.  Solutions containing



pure Iron oxides were generally prepared by first deaeratlng the water



by purging with nitrogen gas, adjusting pH to approximately 6.S by



bubbling pure carbon dioxide, and then adding ferrous sulfate and
                                  47

-------
                                   Ci*rtef Irtw
                                                       Tito
Figure 9.   Schematic diagram of,batch jar test apparatus
                                         motoring pump
                                            NoOH
                                                     contact tonk
                     doefiiorinotofl
                    "  procott
                       wotor
       Ntefifotd
                 fcockwooh JL^
                  pump  GJ
                                    nttr
•jjflow
HI motor
               to
              wooto
                                                       ftitor food
              I now controller
      Figure 10.   Schematic diagram of pilot plaat

-------
 Ra-226.   The solutions were Chen aerated to approximately pH 8 with

 air ' •> oxidize the Iron.   The pH was then adjusted to the desired value

 by bubbling buffer gas (or acid/base addition)  followed by aging for

 one hour.  Solutions of pure i i  ~anese oxides were prepared similarly

 by using stiochionetric amounts  o.  permanganate to oxidize added manganous

 ion.  Sorption onto acid  washed  &.id dIonized water rinsed 0.5 mm filter

 sand (effective size)  was studied using the same apparatus.

      Laboratory pilot  plant studies were conducted using the iron

 aeration-sand filtration  pilot plant shown in Figure 10.

      The 10.2 cm (4 in) diameter pressure sand filter was filled with

 61 cm (2 ft) depth of  filter sand and equipped for water backwashlng

 and periodic rinsing with dilute acid.  Ra-226 and ferrous iron could

 be added to the water  which was  maintained at pH 6.5 prior to aeration

 by bubbling caron dioxide through it.  Aeration to approximately pH 7.5

 was accomplished in a  tank having a 10 minute hydraulic residence time.

 Filter influent was withdrawn from a small tank used to moderate flow.

 Flow rate through the  filter was controlled using a small constant head

 tank equipped with a flow control float and valve.  Automatic samplers

 were used to obtain discrete and composite samples.  The water supply

 was usually 18*C + 3'C.

      Field studies at  Oxford, Iowa utilized a similar pressure filter

 equipped for rinsing with a dilute acid In addition to water backwash-

 Ing.  Water temperature was usually 15"C ± 3°C.

 TYPES OF WATERS USED
»
      Batch studies were conducted using synthetic groundwaters and

 groundwaters obtained  from Oxford, Iowa and Eldon, Iowa.  A hardness
                                   49

-------
 free synthetic water (referred to as Na-form water) was prepared by

 mixing 5 mM NaHCC>3,  1 oM Na2S04 and 1 mM Nad to deionized water.

 Calcium and/or magnesium sulfate was added to vary hardness.  Finished

 waters were obtained from Oxford and Eldon, Iowa and were filtered

 through a 0.45 micron filter prior to use.  Oxford water had a hardness

 and alkalinity of approximately 1300 mg/L and 300 mg/L as CaC03 respec-

 tively.  Eldon water had a total hardness and alkalinity of approxi-

 mately 300 mg/L and  200 mg/L as CaCC>3 respectively.  Oxford and Eldon

 water contained approximately 10 pCi/L and 50 pCl/L of natural Ra-226

 respectively.  Radium concentrations were Increased by adding addltonal

 Ra-226.  Laboratory pilot plant studies were conducted using the tap

 water obtained at the University of Iowa Hygienic Laboratory which had

 a total hardness of  approximately 150 mg/L as CaC03 and an alkalinity

 of approximately 300 mg/L as CaCC>3.  Field studies at Oxford utilized

 the existing sand filter effluent (iron floe removed) or aeration tank

 effluent (containing iron floe).

 RESULTS AND DISCUSSION

 SorpCion to Hydrous  Iron Oxides

 Effect of pH —

      Figure 11 shows the effect of pH on Ra-226 removal in the presence

 of 10 mg/L iron oxides in Na-form water (no Ca or Mg present).  The

 importance of pH control is clearly demonstrated by an increase in re-

 moval from 30 to 40Z near pH 7 to approximately 802 near pH 8.  The

 shape of the curve shows that radium removal by sorption to iron oxides
•
 is very sensitive in the pH range of 7 to 8 which is typically encoun-

 tered in water supplies.  This may explain some of the differences that

 took place in earlier sampling between treatment plants*  "•
                                   50

-------
Removal Mechanism --




     Figure 12 shows chat surface sorption of radium onto the iron



oxide surface takes place and not incorporation of radium into the



ferrous ion as oxidation is taking place.  In Figure 12 slow precipi-



tation (Slow - pptn), radium was added initially in the presence of



'ferrous ion before aeration and then radium was added after the iron



was oxidized, post-precipitation (Posr.-pptn) and approximately 70 to



75Z radium was removed in both cases.  This clearly shows that the



mechansism of removal is surface sorption onto the iron oxide surface.



In Figure 12 (Fasc-pptn), ferrous Iron was added to an aerated solution



containing radium which resulted in nearly Instantaneous formation of



iron oxides.  This shows that the rate of Iron oxide formation did not




affect the amount of radium adsorbed.  Similar trends are shown over a



wide pH range in Figure 13 which shows no difference between fast and



slow precipitation.  These results show that differences in oxidation



rates encountered in the field would not explain observed differences



In radium adsorption.  Sorption of radium to Iron oxides appears to be



very rapid but changes In iron oxide surface characteristics, such as



coating with hardness or ferrous ion, could occur with time which would



affect adsorption.



Effect of Iron and Radium Concentration on Sorption —



     The relationship between sorbent (hydrous iron oxide concentra-



tion) and sorbate (R-226 concentration) was characterized by a linear



sorption isotherm.  The sorption isotherm can be used here to estimate




the capacity of iron oxides for radium and the equilibrium concentra-



tions of radium that remain.  Isotherms are typically developed in



experiments by varying either and/or both sorbent and sorbate.  Figure




14 developed form sodium from water corresponds to a linear isotherm






                                  51

-------
     I
     •
         100
         •0
         60
     -   40

     I   -
                  S      6      7      a      9      10
        Figure 11.  The effect of  pH  on radium removal in
                         the presence of iron oxides
          no

          •o

          M

          90

          •0

          •0

          40'

          JO
                  • Stewpptn   PMt-pptn    Fxt-pptn

Figure 12.  Percent radium removed  by iron oxides:   Comparison of
            precipitation methods at  pH 7.8 in Na-Form synthetic
                vater.  Fe • 10 mg/L  (as Fe),  Ra •  48 pCi/L.
                                 52

-------
chat gives Che amount of radium remaining In solution on che abscissa

and che amount of radlun sorbed per unit mass of iron present on che

ordinate.  Considerable scatter Is apparent, but sorpcion appears Co

reasonably follow a linear isotherm.  The following equation describes

the linear isotherm given In Figure 1A:

     q - Kd fRa-226]                                                (7)

where

     q      - amount of radium sorbed per unic mass of iron

     Kjj     - distribution coefficeint

                    pCl/mg Fe
     Kd     - 0.23  ~           for Fig. 14
                    pCi Ra-226/L

     Ra-226 • concentration of radium reraanining in solution

If a linear 1 sothern is assumed, then percent radium sorbed Is related

Co total Iron concentration by the following equation:

     X Ra sorbed - 100 KdFe/[l + KdFe]                              (8)

where

     £4 - distribution coefficient

     Fe » total iron concentration

Figure IS is a graph of equation 8 showing che concentration of Iron in

oxidized form that is required Co remove radium under ideal condiclons

(no other divalent cations present).

     Iron concentration in waCer supplies is usually less Chan 5 mg/L

and therefore Figure 15 shows that you would expect less than 532 radium

removal in water that has no hardness.  Hardness concentration is

usually high in well water and radium concentration would be in Che

order of 10"^ g/L, therefore, since hardness exists in much higher con-

cent at ions a;J is competing with radium for sorption sites, very little

little sorption of radium onto iron floe would be expected.

                                  53

-------
                 w

                 BO-

                 10

                 •0-

                 10-
                                      7
                                      7
                       10   10  40  &o  to   ia   10  to  «.o
                                       PH
Figure 13.  Percent  radium removed by iron oxides as a function of pH:
             Comparison of the "fast" vs "slow" method of iron oxide
               formation  in Ma-Form Synthetic Water.  Fe • 10 mg/L
                             (as Fe). Ra - 48 pCi/L.
                   C .-
                   1
                   i
                    »
                         r
                                (• Mutton. pO/l
      Figure  14.   Radium adsorbed on iron v» radium concentration
                               remaining in solution
                                    54

-------
           10    20    30    40   50

            Iron Concentration, Mg/L
Figure 15.  Percent radium sorbed as a function
               of iron concentration when no
            other divalent cations are present
                        55

-------
Effect of Competing Ions —

     The amount of radium sorbed onto Iron floes decreases as the con-

centration of divalent ions such as barium, calcium and magnesium

increases.  Relatively small amounts of barium greatly decrease radium

sorption, as shown in Table 17, obtained in Na-form water at pH 8.1.  A

minimum radium removal of approximately 31% was obtained at approxi-

mately 0.5 mg/L barium as compared to about 74Z removal obtained in its

absence.  Increased removals observed at barium concentrations in excess

of 1 mg/L (Table 17) are believed attributable to co-precipitation with

BaSC>4 which probably formed.  This likelihood is indicated by the

increasing proportion of barium removed from solution as the barium is

increased above 1 mg/L.  Radium sorptioa to iron oxides is expected to

be adversely affected by barium in natural waters unless the barium is

significantly complexed and not available for competitive sorptlon.

         TABLE 17.  THE RELATIONSHIP OF BARIUM ION CONCENTRATION TO

                       RADIUM REMOVAL BY IRON OXIDES*
Dosed Barium
mg/L
0.
0.01
0.1
O.S
1.0
2.0
5.0
Soluble Barium
mg/L
-.
<0.05
0.075
0.315
0.598
0.250
0.340
Removed Barium*
mg/L


0.025
0.185
0.402
1.750
4.660
Radium Removal
percent
74
58
49
31
32
59
72
'All the values are averages, Na-form water, pH - 8.1
*Removed barium - Dosed - Soluble

                                  56

-------
      The effect of calcium and magnesium on sorptlon Co iron oxides Is



 shown in Figures 16 and 17 which show examples of Oxford, Iowa and



 Eldon, Iowa waters with hardness of 1300 mg/L and 300 mg/L as CaCOj,



 respectively.   Radium sorptlon by 10 mg/L of Iron oxides formed- by the



 slow-ppcn method In Oxford and Eldon, Iowa, was significantly less



 above pH 7 than obtained when Na-forra synthetic groundwater was used  as



 shown in Figures 16 and 17.  In these experiments naturally occurring



 radium was augmented to 49 and 78 pCl/L for Oxford and Eldon, Iowa,



 water, respecitvely.  Percent radium removed was similar for both



 natural waters  and was approximately 10 and 157 at pH 7.0 in the Eldon



 and Oxford waters, respectively, and about 20Z at pH 8 for both waters.



 This similarity is in spite of the great differences in total hardness



 (1300 mg/L vs.  300 mg/L as CaC03 for Oxford vs. Eldon water) but is



 expected based  on the observation that above approximately 300 mg/L Ca



 as CaCC>3, calcium does not cause a further significant reduction In



 sorptlon.  The  somewhat better removal In Oxford water obtained at pH 7



 is not believed to be significant In comparison to that obtained in tiie



 Eldon water. Radium removals at Iron concentrations less than 10 mg/L



 are expected to be lower.



 Comparison of Sorptlon to Natural and Synthetic Iron Oxides —



      A series of experiments were conducted to compare radium sorptlon



 to natural Iron oxides with that onto synthetic iron oxides typical of



 those used in this study.  Natural iron oxides were obtained from the



 backwash of the existing filter at the City of Oxford and were washed



•in distilled water six times prior to experimental use.  Experiments



 using natural oxides were conducted at a total radium concentration of



 32 pCl/L which  required augmentation of existing radium in the Oxford



 water.






                                   57

-------
               BO-

               W-

               SO-

               TO-

               M-

               CO-

               4O-

               30-
                           •f
                           •f
                                   7


                                  7
                     24   10  4.0  10
                       •.0  HO
                       PH
10  (.0  tt.O
 Figure 16.   Percent radium removed by iron oxides  as a function of pH:
               Ma-Form synthetic water (48 pCi  Ra/L)  vs Oxford ground-
               water [8.5 mM Ca, A.I mM Mg. 49  (7.5  natural) pCi Ra/L]
               Fe • 10 mg/L (as Fe)
10

M

•0

40 H

90

SO
                   •* I
                   7(
                                                7
                                                 7
                    U  10   4.0   10  10  7.0  10   14  C.O
                                     P"
Figure 17.  Percent radium removed by iron oxides as  a function of pH:
              Na-Form synthetic water (48 pCi Ra/L) vs Eldon ground-
              water [2 mM Ca. 1 mM Mg. 78 (50.8  natural)  pCi Ra/L]
                            Fe • 10 mg/L (as Fe)
                                    58

-------
     Figure 18 shows Che pH dependence of sorption of radium onto



natural iron oxides in both the hardness free synethetic groundwater



and in Oxford waCer.  Much higher removals were obtained in the- syn-




thetic water as compared to those obtained in Oxford water.  This is



similar to results obtained using synthetic Iron oxides.



     Removals obtained using natural iron oxides cannot be directly




compared with removals obtained using 10 ag/L Fe of synthetic iron



oxides because the experiments using natural oxides utilized 7.5 to 8.2



mg/L Fe.  However, all natural oxide experiments can be normalized to



results expected at 10 mg/L through the use of calculated partition



coefficients (Kj values).  Figure 19 Is a comparison of percent radium



removal for the natural oxides at 10 mg/L Fe concentrations with remov-



als obtained using 10 mg/L synthetic oxides.  Normalized results for



both types of iron oxides in synthetic groundwater and Oxford water are



very similar.  Therefore it would appear that results obtained using



the experimental synthetic system can be extrapolated to actual field



conditions.  Again the reader Is cautioned not to over generalize this



conclusion as it does not prove that synthetic and natural oxides will



in general yield similar results.  In fact, these results are quite



un expect ed.



Sorption onto Manganese Oxides



Pure Manganese Oxides —



     Ra-226 sorption to 1 and 5 mg/L manganese oxides (as Mn02) pro-



duced by oxidation of manganous ion from the stoichlonetrie addition of



potassium permanganate in Na-fcrm water was significant over the pH



range of approximately 5.5 to 9.2 (Figure 20).  Average removals were



about 80 - 902 for both concentrations abo"e a pH of about 6.5 and were
                                  59

-------
                                     B
                    4     I     •     7
                                                  •     «
      Figure 18.  Percentage of radium removed by iron oxides in
                       Na-synthecic and Oxford groundvater
                                                 0..°
                   D »» •• to fro
                   o
                                    a
                                    o
                       O

                      •9-
                                O

                            a  o
                                               o
                                                •
Figuz* 19.  Percent radium removed by natural and synthetic iron oxides
               in synthetic and Oxford water.  Results normalized to
                                   10 mg/L Fe.

-------
  not  significantly decreased  by  the presence of 1 and 3 mM calcium (100



  and  300 mg/L as CaC03).  These  removals are much greater than determined



  for  the pure iron oxide system  at comparable molar concentrations and



  were not as sensitive to calcium and pH.  The higher sorptive capacity



  Is expected based on the lower  poi.it of zero charge (pzc) for manganese
                                                       *


  oxides.



       Radium removals in Oxford  water were much greater in the pure man-



  ganese oxide system as shown in Figure 21 when compared to comparable



  iron oxide systems.  Significant removals of 60 - 80Z were observed



  over the pH range of approximately 6.6 to 9 In the presence of 1 mg/L



  Mn02 (same molar concentration  obtained by oxidation of 1.57 mg/L Fe).



  Removals exceeded 802 in the presence of 5 and 10 mg/L Mn02 at compar-



  able pH values.  Dilution with  50Z distilled water also did not cause a



  significant increase in removal in spite of the reduction in hardness



  from approximately 1300 to 650  mg/L as CaC03-



  Mixed Iron/Manganese Oxides —



       A possible radium treatment strategy would involve the use of



  potassium permanganate to oxidize iron (or manganese).  In the presence



  of iron, mixtures of iron and manganese oxides would form*  The pres-



  ence of ferrous ion could inhibit radiun sorption to manganese oxides,



  because of competition with  radium for sorption sites.  Sorption to



  mixed iron/manganese oxides  was studied in several experiments by the



  addition of enough potassium permanganate to cause an oxidation of 90Z



— of the available ferrous Iron.  1 mg/L of Fe2+ would produce 0.42 mg/L



  of Mn02 as Hn02 from KHn04 addition.

-------
                        Do
                       o
8 !••* !•*».•*•


• H>»H| »»•
                                   T

                                  PH
Figure 20.  Radium  removal in Na-Fonn synthetic water by 1 and 5 mg/L

                         Mn02 (as MnOj).  Ra - 32 pCi/L.
                     8
                                              £)
                                     A I«M IB«|. CMM «M>
 Figure 21.  Badiua removal by MnO. in Oxford water.  Ra •  32 pCi/L
                                    62

-------
     Sorpcion of Ra-226 from Na-form water was much greater in che



mixed oxide systems than in the presence of comparable concentrations



of iron alone (Figure 22).  For example, radium removal by 1 mg/L Fe



plus 0.4 mg/L Mn(>2 was generally much greater than that obtained in the



presence of 10 ng/L Fe alone.  Sorption by the mixed oxides in Oxford
•


water was significantly less (Figure 23), than obtained in the Na-form



water.  Removals of approximately 10Z were obtained in the 2.0 ng/L



Fe-0.8 mg/L Mn02 system in Oxford water.  This can be compared with



results in Figure 21 showing that 1 mg/L of pure MnO£ brought about a



60 - 80Z reduction of radium in Oxford water.  Similar conclusions can



be made regarding removals from Oxford water in the 5 mg/L and 2.0 mg/L



Fe-Mn02 system.  Therefore, It appears that Iron does Interfere to some



extent with sorption to manganese oxides if all other factors are the



same.



Sorption to Filter Sand



     During the course of the University of Iowa study the investigator



discovered considerable sorption of radium to filter sand.  The Investi-



gator pursued this phenomena in an effort to see if existing sand



filters at iron removal plants could be optimized to remove radium to a



concentration below the HCL.



Effect of pH in Absence of Hardness —



     Sorption generally increased with pH as shown in Figure 24 for



data obtained using 10 g/L sand and no hardness.  Figure 24 shows that



radium sorption was a weak function of pH above pH 5 and a local mini-



mum may have been observed near pH 6.  This pR dependence is in general



agreement with that of Levin ** regarding increased leaching of radium



from mine tailings below pH 5.  The weak pH dependence between pH S and
                                  63

-------
                                <%     O     £«-

                                •     §    B
                      B

                    •      •      T      •      I
Figure 22.  Radium removal in sodium form  synthetic water  by mixed
              iron/manganese oxides.  Permanganate sufficient to
                oxidize 90Z of the iron added.   Ra •  32  pCi/L
                      ••!••* MO)
               | !•••*••* It«*A MO]
O
0
                                         00
                        §
                                f
Figure  23.   Radium removal in Oxford water by mixed iron/manganese
             oxides.  Permanganate sufficient to oxidize 90Z of the
                          Iron added.  Ra - 32 pCi/L
                                  64

-------
               00

                •0

                •0-

                70-

                •0-

                to-

               40-

               30-
                     10  3.0   4.0   10  10  7.0   tO   t.O  10.0
                                      PH

Figure  24.   Percent radium  removed in Na-Form synthetic water by  acid
                 washed sand as  u function of pH.   Sand - 10 g/L.
                                  Ra - 48 pCi/L
                                     65

-------
 9 suggest that a process based on sorption to filter sand would not  be




 very sensitive to normal pH variation encountered in drinking water  (pH




 6.5 to 8.5).




      Percent Ra-226 sorbed was not a function of radium concentration,




 •but sorption increased as sand concentration increased as shown in




 Figure 25.  The results are consistent with a linear Isotherm charac-




 terized by a distribution coefficient of approximately 0.24 L g"1 at pH




 8.2 in Da-form water as shown in Figure 26.




 Effect of Hardness on Sorption to Filter Sand —




      Sorption to 50 g/L filter sand was depressed from approximately




 80Z in the absence of hardness to approximately 60Z in the presence  of




 either 6mMCa or 6mMMg (600 mg/L as CaC03 for both Ca and Mg)  as shown




 in Figure 27.




      Percent radium removed from Oxford water was not a function of




 total dosed R-226 (Figure 28) which is again consistent with a linear




 isotherm.




      Removals were also comparable in both Oxford and Eldon water as a




 function of sand dosage.  Percent removals determined in Oxford and




 Eldon water as a function of sand concentration were also comparable to




 those obtained in synthetic groundwater containing significant hardness.




 Linear isotherms at approximately pH 7.1 for Oxford, Eldon and syn-




 thetic water containing 6 mM Ca (Figure 29) or at pH 8.7 for Oxford  and




 synthetic water containing 6 mM Ca and 6 mM mg were nearly identical




.(Figure 30) which indicated that behavior in synthetic water closely




 resembles that in real waters if hardness is comparable.
                                   66

-------
Figure 25.  The effect of sand concentration on radium removal at pH
                    8.3 - 8.4 IP  sodium form synthetic water
                               Ra-226 - 48 pCi/L
                    5.
                    i
Figure 26.  Radium sorbed on sand vs radium concentration remaining
              in solution.  Sodium fora synthetic water at pB 3.2
                                67

-------
ttO

•0

n

70



10

40

ao-
                                      7

                                      7
                              7
                              7
7
s
                                     7
                                     7
 Figure 27.
         10   10  4.0  10  40  TO  10   t.O   WO
                         PH

Percent radium removed on  acid washed sand as a function of
 pH:  6 mM Ca-Fora vs 6 mM Mg vs Na-Fonn synthetic water.
             Sand • 50 g/L, Ra • 48  pCi/L
00-
eo-
tO-
ft 70-
fit Rtdkim Ren
888
• • •
g to-
0-


t
I * * • *


                               •0   M  «0   QO
                                  Itooium (pCVU
                                    MO   «0
Figure 28.  Percent radium removed as a function of radium concentration
                 at pH 7.1 in Oxford vater (8.5 mM ca, 4.1 mM Mg).
                                 Sand - 50 g/L
                                     68

-------
                             30     M     40
                                 lUKpCVU
                                                        •0
Figure 29.  Comparison of radium adsorption  isotherms  (Sand) developed  in
             Eldon, Oxford and 6 mM Ca-Form  synthetic  waters at pH  7.1
                      tt     20     JO     40      tO     W
                                 |UL
-------
 Regeneration of Filter Sand



 Pilot Plant Studies —




      Pilot plant studies were conducted to investigate the potential of



 using filter sand as an acid regenerable media in the presence of 2



 •mg/L added Fe and approximately 80 pCi/L of added Ra-226.   A dilute



 acid of 4 empty bed volumes at a loading rate of 3 gpm/ft^, the same as



 the filtration rate was used.  Under these conditons the acid volume



 used was about 0.3Z of the total volume of treated water.   A daily



 rinse using 0.1 N HC1 (pH 1) as a regenerant decreased effluent radium



 concentration from about 50 pCi/L to approximately 10 pd/L (Figure



 31).  Similar results were obtained when a 0.01 N (pH 2) dilute acid



 rinse was used aa shown in Figure 32.  Acid cost for pH 2  would be only



 1/10 of pH 1, a significant savings.



      Cumulative Ra-226 recovery in both 0.1 and 0.01 N HC1 rinse is



 shown in Figure 33 and indicated the efficiency of radium recovery as a



 function of both time and acid concentration.



      Figure 33 also shows that only iron is removed by prior backwash



 and therefore the radium that was recovered must be associated with the



 sand and not the iron.



      Figure 34 shows that a loading rate below 4 gptn/ft2 is necessary



 for maximum radium sorption onto filter sand for water hardness of 400



 mg/L as CaG03>  Another test at 100 mg/L as CaC03 water hardness showed



 that loading rates up to 12 gpm/ft^ did not reduce radium sorption



.significantly.
                                   70

-------
                	 ^^f
                I ~
 4*-71  7VM  M-O2 02<
mpoM* 7Im» tmr** 0«un)
                                                   U« «*•>«•
   Figure 31.   Radium removal in the DHL pilot  plant shoving effect of a
                  daily pH 1 BC1 rinse.  Hardness  • 150 mg/L, as CaCO..
                                 Fe - 2.0 mg/L,  pH » 7.5             J
Figure 32.  Radium removal in the DHL pilot plant showing effect  of a pH 2
              HC1  rinaa during dally backwash.  Hardness • 150 mg/L, as
                            CaC03. Fe • 2.0 mg/L. pH - 7.5
                                      71

-------
                                         NM«:
                                         Acid riltMt «urt*d tt tWM 0
                                         •M rvn <»r 20 mniitttt
                                •  •  10  Q  M  1t  U  20  22  24
                                  Tim* (minutes)
 'igure 33.  Comparison of cumulative radium recovered  in  pH  1 and pH 2
                regenerant.rinses.   Regenerant loading  rate 7.4 n/h
                 (3 gpm/ft ), 4  bed volumes used, EBCT  5 minutes.
                           Regenerant flow rate 1 L/min.
                1
Figure 34.   Radius concentration profiles as a function of loading rate.
                Initial pH 1 rinse.  Total hardness 400 mg/L as CaCO..
                                     72

-------
 Field Studies aC Oxford, Iowa (Sand Filter Effluent) —



      Figure 35 shows Influent/effluent radium concentration profiles



 when 4 bed volumes of 0.1 N HC1 (pH 1) regenerant was used daily for



 two cycles utilizing a water loading rate of 3 gpm/ft2. Effluent radium



 xoncentratlon decreased to approximately 3 pCi/L from an influent value



 of about 8 to 11 pCi/L.  Upon termination of the acid regeneration,



 radium removal decreased significantly.




      Significantly better performance was observed when the loading



 rate was decreased to 1.5 gpm/ft2 as shown in Figure 36 which shows



 effluent radium concentration of only about 1 to 2 pCl/L.



      Radium removal efficiency decreased significantly from an initial



 value of 60Z when pH 2 regenerant was used to 25 to 35Z when pH 3



 regenerant was used (Figure 37).



 One possible explanation of the decreasing efficiency with time could



 be that iron precipitates are gradually coating the sand.   The solu-



 bility of ferric hydroxides also decreases greatly as pH is increased



 fron 1 to 3.



 Field Studies at Oxford, Iowa (Aerator Effluent) —



      An experiment was conducted over a period of approximately 13 days



 which investigated removal of radium by sand sorption from the effluent



 of the aeration tank at the city of Oxford (simultaneous radium and



 iron removal) as a function of type and flow rate of acid  regenerant.




 Influent water flow rate was fixed (1.5 gpm/ft2).  The concentration of



.iron in the influent was approximately 0.5 mg/L and was reduced to



 approximately 0.08 mg/L after passage through the experimental radium



 removal filter.  Radium removal in the presence of significant iron
                                   73

-------
TJ-
I "'
f.
m*
!-
5 4-
CC
J-

S^5

\
.O
...-•'
mu, u* u» .
OirNMOMIM ^..-O"
.-••"""
O WHLf kM*Mi* Mrfy
WM Mt.nk «H» MM « M k
1
                  oo-a
                          M-M      »vn      »»-«W
                             ThMkxmiM (MHUl
Figure 35.  Ra Concentration proflies in Oxford tfater.
   With and without  pH 1 (0.1 N HC1) regeneration.
                Loading rate 3 gpm/ft .
                             i ThM **r*t» Ownnl

Figure 36.  Ra  Concentration profiles in Oxford water.
              thout  pH 1 (0.1  . N HC
                Loading rate 1.5 gpm/
   With and without pH  1  (0.1 .  N HC1^regeneration.
               '   '•      -   -  -    n/f£\
                            74

-------
         4-
                         y»~~tf-
                   • MI o^i.
                                               4
                                               o n
-------
floes was only about 25 - 352 (Figure 38) compared with 80 - 90Z removals



obtained when Che existing sand filter effluent (absence of significant



iron floes) was filtered and periodically regenerated.  Removals were



similar with both dilute sulfuric acid and dilute hydrochloric acid



.rinses, and were similar at 1.5 gpm/fc2 and a 3.0 gpm/ft2 regenerant



loading rates.



     Ths large reduction in performance observed when simultaneous iron



and radium removal was practiced was not expected based on previous



laboratory batch and pilot plant results which indicated little iron



floe interference.  The mechanism by which the presence of iron floes



inhibited radium removal is not clear.  While it may be supposed that



the floes coated the sand thereby blocking radium sorptlon, it must he



noted that roost of the iron floes were probably removed in the first



few inches of sand bed as indicated by visible color changes.  There-



fore, significant amounts of "clean" sand probably existed below the



iron floe removal zone even after one day of operation.  Possibly the



influent may have contained significant (with respect to radium) con-



centrations of unoxidized soluble ferrous iron which effectively com-



peted for sorption sites below the zone in which most of the iron floe



was removed.
                                  76

-------
       ^^^
       i
          4-
  JQ
""

                                                  OlRltR
               «nV(H1MQ
               taw MM •*;•*»
                                           MM • 7.4 MM
                              • UVffttNO
Figure 38.  Influent  and effluent Ra-226 concentration profiles.
           Treatment  of Oxford aeration tank effluent.
    Comparison of  sulfuric and hydrochloric acid regeneration
                at two regenerant loading rates.
                                 77

-------
                             SECTION 6.

                    MANGANESE DIOXIDE COATED FILTERS

               (North Carolina State University Study)12



SUMMARY OF MANGANESE DIOXIDE COATED FILTERS
•                                                              •
     A process was developed for coating woven- acrylic cartridge fil-

ters with manganese dioxide (Mn02).  The coating process consists of

circulating hot potassium permanganate (KMn04> through the filter.  The

investigator surmises that the acrylic filter material is oxidized and

the KMn04 reduced; resulting In MnC>2 that adheres to the filter.  Both

loose fiber and woven filters averaged 22.5 percent Mn02 by weight or

56.25 grams MnO^ in a ten-inch long 250 gram filter element.

     The manganese dioxide coated filters were tested for radium re-

moval from waters with high and low hardness.  Bleed serean testing of

the ten-inch Mn02 filter on high hardness water containing a total

radium concentration of 36.4 pCi/L resulted In treating 800 ft^ of

water before the effluent radium concentration reached 5 pCl/L.  Total

filter capacity was exhausted after treating 5500 ft^ of water.  A

total of 801 nanocurles of total radium were removed or 14.2 nanocuries

per gram of Mn02«  The removal efficiency from the first 800 ft  of

water was 86Z.  Bleed stream testing of the ten-inch Mn02 filter

on low hardness water containing a total radium concentration of J.3.2

pCi/L resulted in treating greater than 1000 ft^ of water before the

effluent concentration reached 5 pCl/L.  Total filter capacity was

exhausted after treating 3000 ft^ of water.  A total of 290 nanocurles

of total radium were removed or 5.2 nanocuries per gram of Mn02-  The

removal efficiency for the first 1000 ft' of water was 66%.       . „
                                  78

-------
      Dilute solutions of the metals cadmium, calcium, cesium, chromium,


 cobalt, iron manganese and sodium were tested to determine their


 sorption onto the Mn02 filter.  Results revealed a positive uptake of


 metals.  The greatest was the uptake of cadmium.  Results showed that


 .no chromium was taken up by the MnC>2 filter.


      Testing confirmed that no organic acrylonitrile or acrylic fiber


 were leached off the filter into the product water.  No manganese was


 detected leaching off the filter into the product water.


 MANGANESE DIOXIDE COATING PROCESS


      The woven acrylic cartridge filters were coated with manganese


 dioxide as shown in Figure 39.  Hot potassium permanganate solution at


 55 to 58°C was pumped through two acrylic cartridge filters in series


 at a flow rate of 8 t 10 GPM.  It takes approximately twelve hours for


 both filters to become saturated with manganese dioxide.  After the


 coating process, the filters were rinsed with tap water for twenty


 minutes in both directions until no excess manganese was leacuing Into


 the rinse water.  The prepared filters were then packaged and sealed in


 plastic bags containing a one mg/L solution of potassium permanganate


 to inhibit biological growth.  The quantity of Mn02 that was coated


 onto the filters was approximately 22 1/2X of the total filter weight


 Including MnC^-  It is the investigator's opinion that the acrylic


 filter surface is oxidized and the KIln04 reduced such that MnO£ adheres


 to the filter and the potassium salts are washed away.


 BLEED STREAM FIELD TESTS (RADIUM SORPTION)
»


      A high hardness high pH water (Highland Park) and a low hardness


 low pH water (Gateway) were tested for radium removal with ten-inch
                                                          i

 woven acrylic filter elements.  The raw water analysis for both waters
                                   79

-------
                        RINSE
                        WATER
                                PRE-
                               FILTER
(I
oo
o
                                                      FILTERS
                                       8  TO 10 GPM
0  0
                                                      RINSE
                                                    TO WASTE
                                                  55 TO 58'C
                                                                                           TANK
                                        Figure  39.   Schematic  diagram of process for coating
                                             woven  acrylic cartridge filters with MmO-.

-------
 is given in Cable 18.

      Bleed screan testing of the MnCb filter in the Highland Park Water

 System resulted In a decrease In total radium, from the Influent

 concentration of 36.4 pCi/L total radium (see Table 19).

 •One ten-inch Mn(>2 fiber filter element treated a volume of approxl-
   »
 mately 800 cubic feet of water, at a flowrate of 3 gallons per minute,

 before the effluent radium concentration increased to a level of 5 pCl/1

 Total breakthrough was reached after approximately 5500 cubic feet of

 water was treated (Table 19).  The single ten-inch filter element of

 approximately 250 grams MnC>2 fiber, contained 56.25 grams of MnC»2, and

 193.75 grams of acrylic fiber.  This accounted for the removal of 801

 nanocuries total radium, or 14.2 nanocuries per gram of KnC>2, at a

 removal efficiency of greater than 86 percent from the first 800 cubic

 feet of water treated.  Total breakthrough occurring after approximately

 5500 cubic feet of water treated was preceeded by the uptake of 2076

 nanocuries total radium, or 36.9 aanocuries per gram of MnC^-

      Bleed stream testing of the Mn02 filter in the Gateway Water Sys-

 tem resulted in a decrease in total radium, from the Influent concen-

 tration of 13.2 pCi/L total radium (see Table 20).  A single ten-inch

 Mn02 fiber filter element treated a volume of approximately 1000 cubic

 feet of water, at a flow rate of 3 gallons per minute, before the

 effluent concentration increased to a level of 5 plcocuries per liter.

 Total radium breakthrough was reached after approximately 3700 cubic

.feet of water was treated (Figure 20).
                                   81

-------
       Table Iff.   RAW WATER ANALYSIS FDR HIGHLAND PARK AND GATEWAY*

Parameter
Total Radium (pCl/L)
Total Hardness (mg/L as CaCOj)
Calcium (mg/L)
Sodlura (og/L)
Alkalinity (mg/L)
Chloride (mg/L)
Total Dissolved Solids (mg/L)
Turbidity
Magnesium (mg/L)
PH
Arsenic (mg/L)
Barium (mg/L)
Cadoium (mg/L)
Chromium (mg/L)
Fluoride (mg/L)
Lead (mg/L)
Mercury (mg/L)
Nitrate (og/L)
Selenium (mg/L)
Silver (mg/L)
Iron (mg/L)
Manganese (mg/L)
Hltfh Harrine**
Highland park
36.4
227
74.6
110.8
124
203
590
0.2
9.8
7.8
< 0.01
0.32
< 0.005
< 0.01
1.04
< 0.03
< 0.0002
< 0.05
< 0.005
< 0.02
< 0.05
< 0.03
Low Hardness
Gatcwav
12.3
23
2.4
5
0
0
54
0.05
4.1
4.5
< 0.01
< 0.10
< 0.005
< 0.01
< 0.10
< 0.03
< 0.0002
5.65
< 0.005
< 0.05
< 0.05
< 0.03
 •Analysis performed by N. C. State Laboratory of  Public  Health






      The single ten-Inch filter element  of approximately 250  grams



 fiber (56.25 grans MnO£) removed 290 nanocurles  total  radium, or 5.2



 nanocuries per gran Mn02 at a removal efficiency  of greater than 66



 percent from the first 1000 cubic feet of  water treated. Total  break



 through occurred after approximately 3000  cubic feet of  water was



 Created, and was proceeded by a total uptake  of 439 nanocurles total



.radium, or 7.8 nanocurles per gran of Mn02«
                                   82

-------
TABLK 19.  BLEED STREAM FIELD TEST, Mno2 FILTRATION (HIGHLAND PARK)*
Total Flow

-------
      Bleed stream testing of Mn02 fiber in the Highland Park Water



 System resulted in the removal of 36.9 nanocurles per gram of MnC>2



 fixed on acrylic fiber base.  Bleed stream testing of the Gateway Water



 System resulted in the removal of 7.8 nanocurles per gram of fixed



 •Mn02«  The water quality parameters for the Highland Park and Gateway



 Water Systems are Hated in Table 18.  The difference in the pH of



 these water sources was the most significant water quality parameter.



 The Highland Park water had a pH of 7.8, while Gateway's water had a



 pH of 4.5.  An Increase In pH increases the ability of NnO£ to sorb



 cations due to the higher adsorption energy.  This correlation of pH



 is evidenced In the Increased level of radium uptake per gram Mn02 of



 the Highland Park system.  The bleed stream test at the Highland Park




 Water System had a measured radium uptake per gram Mn02 of 4.7 times



 greatec than the uptake measured at the Gateway Water System.  The



 higher calcium concentration of the Highland Park water would compete



 with radium for adsorption sites on the Mn02 but it appears that the pH



 difference had a greater affect on radium sorptlon than the increased



 calcium.



 SORPTION OF OTHER METALS ONTO MANGANESE DIOXIDE




      Bench-top testing of the Mn02 fiber was performed for removal of



 radium, cadmium, calcium, cesium, cobalt, Iron, and manganese.



      A 53.5 gallon water sample from the Highland Park Water System,



 containing 36.4 picocuries total radium per liter, was pumped through a




.loose Mn02 fiber filter containing 3.38 grams Mn02«  A level of uptake



 of 7.37 nanocurles total radium per gram Mn02 was achieved.  However.



 Che radium removal efficiency of 100 percent was maintained throughout



 the test, a significant level of uotake while performing at a high
                                   84

-------
level of radium removal.




     Dilute solutions of metals, alone and in combination with other



necals were pumped through a column or columns containing loose MnC>2



fiber.  The breakthrough results are listed In Table 21.  From the



amount of metal uptake on Mn02 a comparison of the adsorption potential



can be nade for each metal and combination of metals.  In each of these



cases a significant amount of metal was removed by adsorption onto



Mn02t showing the potential of competition with radium for sorptlon



sites.  For the cations listed In Table 21, calcium would usually be



present In the highest concentration and the Cable Indicates moderate



preference for calcium as compared to the other cations.
                                  85

-------
TABLE 21.  METAL REMOVAL WITH Hno2 COATED FIBER
Metal
•
Cadmium
Calcium
Calcium
Cesium
Cesium
Cobalt
Sodium
Cesium
Cobalt
Cesium
Sodium
Cobalt
Cobalt
Cobalt
Cobalt
Cesium
Sodium
Cobalt
Cesium
Cobalt
Sodium
Iron
Iron
Manganese
Manganese
Concen-
tration
(mg/L)
2.0
2.2
2.2
10.0
10.0
3.0
1.0
10.0
3.0
10.0
1.0
2.8
4.4
4.4
2.8
10.0
3.5
3.1
10.0
3.5
3.5
4.15
4.15
4.15
4.15
Metal
Mn02 Uptake
(grans) (mg metal)
3.825
5.175
10.575
12.375
12.375
16.2
13.95
14.4
3.375
6.75
11.7
11.025
26.1
3.375
6.67
2.6
7.2
700
. 275
550
2000
2000
1230
2870
1680
440
1000
910
775
2275
1037.5
1245
830
996
Metal 1/2 Break-.' Break-
Uptake through through
(mg/g (liters) (liters)
Mn02)
183.01
53.14
52.01
161.62
161.62
75.93
205.73
116.67
130.37
162.96
77.78
70.29
87.16
307.41
186.66
230.55
138.33
240
62
140
150
165
85
190
250
30
82
200
90
460
45
125
70
170
350
125
250
200
200
123
287
600
100
250
325
250
650
250
300
200
240
                       86

-------
                                SECTION 7

             EVALUATION OP RADIUM REMOVAL  AND UASTE  DISPOSAL    .

                 SYSTEM FOR SMALL COMMUNITY WATER SUPPLY

          (Rocky Mountain Consultants,  Inc., Denver,  Colorado)^
*


 SUMMARY (Iron Removal, Ion Exchange, Waste Disposal)

      A full scale well water treatment plant for a  small  community was

 evaluated for radium removal.   The  plant  consists of Iron removal with

 aeration, filtration and icttling;  hardness and  radium  removal with Ion

 exchange; and radium removal from the  waste brine with  the Dow radium

 selective conplexer (RSC).  The purpose of this  research  was  to deter-

 oine the effectiveness of each  stage  for  radium  removal.  Results

 showed that very little radium  was  removed across the iron removal

 process.  Radium was removed to concentrations below the  MCL with Ion

 exchange and as expected, hardness  broke  through before radium.  Greater

 than 99X of the radium was removed  from the waste brine by the RSC.

 Results given here are not final because  the project will not be com-

 pleted until October, 1987.
                                  87

-------
DISCUSSION



General Information




     Data was obtained Co assess the effectiveness of the overall



treatment plant operation.  The treatment plant consisted of iron



removal through aeration, settling and filtration; hardness and radium



removal through ion exchange and waste brine treatment for radlun



removal through the Dow radium selective complexer.  Arrangements were



made to send the spent complexer resin to a radiological disposal site



in Nevada at a nominal cost.  The complexer resin occupies a volume of



approximately 31 gallons and has been treating the waste brine for six



months, to date, without breakthrough.  This project is still going on



and, therefore, final breakthrough capacity remains to be determined.




This is a full scale treatment plant for a snail community in Colorado.



Iron Removal



     Table 22 shows that the raw water radium concentration varied



from approximately 27 to 35 pCl/L and the effluent from the Water Boy



unit varied from approximately 22 to 35 pCl/L.  This verifies the



information in section 5 which states that very little radium removal



takes place across the iron removal process.  Potassium Permanganate



was added before the Water Boy (iron settling and filtration unit)



but It was not enough to oxidize all the Iron.  Therefore, there was



ferrous ion left that used the adsorption sites on the manganese dioxide



instead of radium.  For this plant, the intention was to remove all



radium with ion exchange and not with iron removal so that waste brine



containing radium could be treated with the Dow complexer.
                                  88

-------
TABLE 22.  REDHILL FOREST WATER QUALITY MONITORING DATA - GENERAL TREATMENT PLANT OPERATION
Sample
Location and
WQ Parameters
I. RAW WATER .
Sample ID
(into plant)
Iron (mg/L)
Manganeae (mg/L)
Sodlua (mg/L)
Hardness (ug/L)
Total Solids (mg/L)
Radium 226 (pCl/L)
III. TREATED WATER
Sample ID


9/10/86
RW910-A


7.3
0.43
16.1
335
—
30+6
IXE910-A



9/10/86
RW910-B


6.3
0.42
8. OS
245
—
2746
IXE910-B



9/12/86
RW912-C


6.49
—
—
336
428
33+7
IXE912-C



9/14/86
RW914-D


5.54
—
—
263
—
31+6
IXE914-D



9/14/86
RW914-E


6.60
—
—
225
—
32+6
IXE914-E

Date

9/16/86
RW916-P


6.11
0.40
10.8
281
—
32+6
IXE916-P



9/18/86
RU918-G


8.26
~
—
316
—
32+7
IXE918-G



9/18/86
RU918-H


6.11
—
—
241
—
30+6
IXE918-H



9/21/86
RW921-I


3.19
~
—
265
—
33+7
IXE921-I



9/21/86
RW921-J


6.71
0.38
6.79
249
340
35+8
IXE921-J

(Effluent Ion Exchange)
Iron (mg/L)
Manganese (mg/L)
Sodium (mg/L)
Hardness (mg/L)
Total Solids (mg/L)
Radium 226 (pCl/L)
0.11
.01
147
10
—
4+2.5
0.20
0.01
121
10
~
3.8+2.6
0.20
—
118
22
432
0.3+1.7
0.27
—
86.4
70
—
3.6+2.6
0.39
• ' —
384
28
—
3.7+2.7
0.28
0.07
137
17
—
0+1.4
0.39
—
95.7
22
—
0.6+1
0.96
—
65.6
108
—
.9 4.7+3.0
1.00
—
33.2
207
—
12+4
1.62
1.12
15.7
238
428
13+4

-------
                TABLE 62.  RBDHILL FOREST WATER QUALITY MONITORING DATA - GENERAL TREATMENT PLANT OPERATION (Contd)
   Sample
Location and
WQ Parameters  •

II. EFFLUENT FROM*
     WATERBOY UNIT
     Sample ID
                                                  Date
 9/10/86   9/10/86   9/12/86   9/14/86   9/14/86   9/16/86   9/18/86   9/18/86   9/21/86   9/21/86

WBE910-A  WBE910-B  WBE912-C  WBE914-D  UBE914-E  UBE916-F  UBE918-G  WBE918-H  WBE921-I  WBE921-J
Iron (ag/L)
Manganese (ag/L)
Sodium (ag/L)
Hardness (ag/L)
Total Solids (ag/L)
Radium 226 (pCl/L)
IV. VOLUME TREATED
SINCE LAST BACKWASH
A. Ion Exch. Sya(Unlt
B. Water Boy Unit
0.48
0.60
—
212
—
22+5


1) 0
0
2.5
0.94
— '
240
—
25+6


10,000
10,000
0.89
—
—
292
—
2846


20,000
20,000
1.27
—
—
252
—
28+6


30,000
30.000
4.27
—
—
246
—
29+6


0
40,000
1.70
1.29
—
307
r-
28+6


10,000
50,000
1.77
—
—
254
—
22+6


20,000
60,000
4.49
—
—
247
—
26+6


30,000
70,000
2.05
—
—
277
—
35+8


40,000
80,000
5.22
1.24
—
261
—
29+7


50,000
90,000
 *Water boy unit includes KMnO^ addition, settling and filtration for iron removal.

 NOTE:  From 9/10/86 through 9/21/86 an intensive monitoring of plant operation was conducted with samples collected
        after every 10,000 gallons were treated.

-------
Ion Exchange


     The purpose of Che Intensive monitoring during Che treatment

cycles given In Table 22 was Co determine approxlnaCely when radium


breakthrough occurs In relation Co hardness breakthrough for the Ion
*

exchange process.  As can be seen from Che data In Table 22, Che raw


water hardness varies with time, and affects Che lengch of Che treat-


ment runs between regeneration cycles for Che ion exchange process.


Table 22 shows that radium begins Co break through when Che accurau-


laced volume of water created In Che Ion exchange process exceeds


30,000 gallons and elevated radium In che effluenc occurred at 40,000


gallons.  Previously, che volume of water created for che Ion exchange


process between backwash operations has been between 50,000 and 60,000


gallons, however, 1C appears, based upon Che above results (Table 22),


Chat back washing should be performed more frequenCly, probably every


30,000 Co 40,000 gallons.


     The resin volume Is approximately 224 gallons and therefore 134


Co 179 bed volumes could be created before regeneration.  Ic was shown


in section 4 of this report that  ation exchange resins prefer radium


over hardness.  The data in Cable 22 also verifies Chac che resin has


preference for radium over hardness, buC table 22 also indicates that


radium breakthrough occurs shortly after hardness breakthrough.  This


is seen in the 9/18/86 Column where hardness increased from 108 to


207 mg/L as CaCfrj and radium-226 Increased from 4.7 to 12 pCl/L.  This


indicates that after many cycles of operation radium accumulates on the


resin and.this explains why radium-breaks through'shortly after hard-


ness.  This was also shown in section 4 with data for virgin resins


compared to data taken after repeated cycles.



                                  91

-------
 Waste Brine Treatment  With Radium Selective  Complexer




      The data presented In Table 23  clearly  shows  that  the RSC  is  truly



 radium selective•   In  general,  the concentration of  the other constitu-



 ents analyzed (i.e., iron, manganese,  sodium,  hardness, and  total



 solids)  are virtually  unchanged in passing through the  RSC tank.



 However, in excess of  99% of the Influent radium is  removed  and compos-



 ited in the RSC.   A check was recently made  of the condition of the  RSC



 In regard to solids buildup in  the Inflow side and minor accumulations



 were observed. The flow rate through  the RSC  has  been, and  is, main-



 tained at about 23 gpm when the system is operating.  This flow rate Is



 equivalent to a bed surface loading  rate of  about  10.5  gpm/ft2. With



 the current bed depth  of 24 Inches,  the average contact time of the



 inflow wastewater  with the RSC Is about 1.4  minutes.



      Most of the radium in the  raw water coming into the plant  is  re-



 moved by the ion exchange process and  is subsequently complexed in the



 RSC when the regeneration waste brine  is treated by  the RSC.  The  RSC



 has a resin volume of  31 gallons and,  therefore, 471 BV (14,600/31) of



 waste brine were treated from September 10 through September 29, 1986.



 The RSC was still  removing over 99Z  of the radium  from  the waste brine



 on December 10, 1986 and over 31,700 gal or  1023 BV  of  waste brine had



 been treated.



      The complexer can either be installed at  the  influent to the  ion



 exchange system to treat the potable water or  on the waste brine.  When



.installed on the waste brine, the complexer  will treat  a much smaller



 volume of water at a much higher dissolved solids  and higher radium



 concentration. One purpose of  this  research is to determine which RSC



 location would be  the  most cost effective and  result in the  easiest
                                   92

-------
handling of Che final radium for disposal.  As mentioned above, 1023 BV




of waste brine has been created.  During this time the 1,463,000 gal-



lons of drinking water has been treated.  If this volume of water were



Created by the 31 gallons of complexer it would equal 47,194 BV.  A



larger vessel and much more RSC would be required to treat the potable



water, therefore, the process would be more expensive if the RSC were



used to treat the total water supply.  The capacity of the RSC when



used on potable water vs. waste brine still needs to be determined.
                                  93

-------
            TABLE 23.  SUMMARY OF WQ DATA FOR REGENERATION WASTE WATER THROUGH  RSC  RESIN
PARAMETERS
Accumulated
Voluae
Treated Sample
Date Gal*. ID
7/10/86
7/10/86*
7/30/86
8/18/86
8/25/86
8/31/86
9/23/86
9/29/86
Averages
Collected
t
0 A( Inflow)
B( Out flow)
0 R( In flow)
B(0ut flow)
2,400 (Inflow)
(Outflow)
6,335 (Inflow)
(Outflow)
7,000 (Inflow)
(Outflow)
9,460 (Inflow)
(Outflow)
13,430 (Inflow)
(Outflow)
14,600 (Inflow)
(Outflow)
for samples collected
on 7/10/86
Inflow
Outflow
Iron Manganeae
•g/L mg/L
2.48
0.98
2.39
1.42
2
1
5
4
8
6
9
8
5
4
7
7
.03
.56
.61
.91
.0
.8
.0
.5
.08
.61
.21
.15
for period
5.64
4.93
23.8
16.7
24.8
23.2
31.
32.
30.
30.
29.
29.
33.
33.

SodlUB
mg/L
11.
13,
600
300
12 ,000
12,100
8
2
6
2
9
7
1
1
21.0
21.0
30.
.1.
5
5
7/10/86
28.7
27.8
11
11
11
11
12
12
12
12
11
11
11
11
.000
.000
,600
,500
,100
,000
,600
,700
,700
,800
.400
,500
Hardness
(CaC03)
•g/L
476
245
494
440
9,850
10.200
10,500
12 ,000
10,200
10,100
11.500
11.600
7.830
7.770
8.350
8.420
through 9/29/86 (7 i
11.710 8,390
11,970 8,620
Sol Ida
•g/L
34,900
34,600
35,600
35,300
41.700
41,800
44,400
44.600
50.300
49.500
54.200
55.200
41.700
42.500
37,600
37.600
lamples)
43.540
43.690
Total
Radium I
pCl/L 1
860+30
16+H
870+30
15+11
1280+40
1.6+1.2
1460+50
7.1+5.1
1260+40
7.8>3.2
1400+40
9.4^3.5
1200+100
6.0+3.2
920+30
4.L+2.4
excluding B
1197
7.4
Z
tadlui
tenoval
98.1
98.3
99.9
99.5
99.4
99.3
99.5
99.6
samples
99.4
Accumulated Plow - 14.600 gals.
RSC resin was replaced on 7/10/86
* Q-11.0 gpm all other samples collected are for a flow of 23 gpm through the RSC tank.

-------
                                SECTION 8




                      OTHER RADIUM REMOVAL PROCESSES





 LIME & LIME-SODA SOFTENING




      The literature** shows chat  lime  and lime  soda ash  softening  have



 been denonstraced on * full-scale plant level to  achieve 75-96  percent



 removal of radium.   Radium carbonate and radium sulfate  are  listed15 as



 being Insoluble  In water,  but  at  the very low concentrations around the



 drinking water regulation  MCL  of  5 pCl/L, which by definition is equal



 to 5 X 10*12 graas of radium per  liter, they would still be  soluble.




 Rad'.um is an alkaline earth metal very similar  to calcium and magnesium



 and would be expected to act like these metals.  Therefore,  the removal



 of radium during softening is  probably due to copreclpltatlon which is



 defined by Lalclnen16 as a phenomenon  where the main  precipitate and



 the impurity come down together.   In this case  hardness  is the  main



 precipitate and  radium is  the  impurity.



      Results of  two tests  performed at the U.S. EPA14 on lime softening



 art given in Table 24.  The radium data, after  dual media filtering.



 indicate that removal is pH dependent  with 84 percent and 94 percent



 radium removals  at  pH 9.5  and  10.5 respectively.   This agrees with the



 copreclpltatlon  theory because a  greater percentage of the hardness was




 also reaoved at  the higher pH  as  shown in Table 24.



      Lime softening depends on the use of lime  and soda  ash  to  change



 the soluble calcium and magnesium compounds into  nearly  Insoluble  com-




 pounds that are  flocculated, settled,  and filtered. Conditions for



-carrying out the precipitation of calcium and magnesium  vary because



 different pH levels are needed for each - about pH 9.5 for maximum



 precipitation of calcium carbonate and pR 10.5  for maximum precipitation






                                    95

-------
        Table 24.  Results of  Lime Softening  Pilot  Plant Tests  for Radlum-226 Removal From (fell Water*'4
Average Rat*
Water
Concentration

Test
Number
11
2**

Date
Started
8/4/75
8/18/75
Lengch
of Tests
hr
102
77
pH of
Treated
Water
9.5
10.5

Radium-
226
pCi/L
4.34
4.82

Hardness
(as Ca003>
mg/L
246
256
Percentage
Radium Removed
Settled Water
First Second
Stage Staget
79
92 93
Percentage Hardness
Removed
Filtered Water Filtered Water
Dual
Media
84
94
GACt t Dual
Media
84 57
95 62
CACtt

57
62
*Illinois groundwater containing naturally occurring  radlum-226
tSecond settling stage followed recarbonatlon that  lowered  pH  to  9.8.
ttFlltrasorb 200, Calgon Corp., Pittsburgh, PA.
{Samples collected only on 8/8, 8/9,  and 8/10
**Samples collected only on 8/20 and  8/21

-------
 of  magnesium hydroxide.   Hence,  if magnesium  concentration  Is low,



 treatment  to a  pH of 9.5  will  be sufficient.   If magnesium  concentra-



 tion la high, excess lime,  to  produce  pH of 10.5 can be  used.  .Soda  ash



,1s  added as  needed to precipitate noncarbonate hardness.  Recarbonatlon



 Is  used to stabilize lime treated water, reducing its  scale forming



 potential.  Carbon dioxide  neutralizes  excess lime  precipitating it  as




 calcium carbonate.  Further recarbonation  converts  carbonate to bicar-



 bonate .



      Results of radium and  hardness  removals  at water  treatment plants



 using lime soda ash softening  are given in Table 25.17   The table shows



 that radium removal varied  from  75 to 96 percent and that lime soda



 softening  is effective for  removing  radium.   The Webster City plant



 was using  lime  only during  August 1974  measurements, but was using



 soda ash during the February 1975 measurements.  Table 25 shows an



 increase in radium and hardness  removal and an increase  in  process



 pH  during  the February 1975 test. The  data indicate that percent



 radium  removal  Increases  as pH increases and  this is shown  in Figure



 40.9



 DOW RADIUM SELECTIVE COMPLEXER17« 18



 Background



      The Radium Selective Cora piexer was invented by Professor N. J.



 Hatch,  New Mexico Institute of Mining and  Technology,  Socorro, New



 Mexico* Professor Hatch  was formerly an employee of Dow Chemical and



 he  remains a consultant to  the Dow Company.



      The Radium Selective Coraplexer  (RSC)  was originally intended for



 removal of soluble radium from uranium  mine waters.  The mining Indus-



 try currently uses barium sulfate precipitation for this purpose.
                                  97

-------
.100,

o
E
|80
(T
   60
     9
        10
Proctss pH
II
     Figure 40.  Percent radium removal as a function
              of pH during water softeningq
                             98

-------
Table 25.  Ra-226 and Hardness Removals at Mater Treatment Plants
                  Using Line-Soda Ash Softening'
Ra-226

City
Elgin

Peru

Webster City
(Aug. 1974)


Webster City
(Feb. 1975)

West Das Koines


Sampling
Point
Well
Filter Eff.
Well
Filter Eff.
Well
Clarlfier ll Eff.
Clarlfler 12 Eff.
Filter Eff.
Well
Clarlfler Eff.
Filter Eff.
Well
Contact Unit Eff.
Filter Eff.

pCi/L
5.6
0.8
5.8
1.1
6.1
1.9
2.6
0.9
7.8
0.6
0.3
9.3
2.6
2.4
Percent
Removal

88

81



85


96


75
Hardness

mg/L
237
102
326
94
507
333
282
262
482
150
106
376
215
190
Percent
Removal

57

71



48


78


50
pH of
Process





10.0
10.1
9.3

11. 0
9.9

10.1-10.4
9.4-9.5

-------
Several tests of the application of the RSC for removal of radium from



these waters have been conducted and one of these studies Is described



by Boyce.1^  A large RSC installation is presently in operation for



uranium mine waters at the Key Lake Mine/Mill in northern Saskatchewan.



     A full scale trial was conducted in Bellvllle, Texas to demon-



strate the RSC technology for drinking water treatment.^  A full scale



system was also Installed for drinking water treatment in South Super-



ior, Wyoming in November 1983 and operated for 13 months.*8  Several



two inch diameter demonstration columns have also been pilot tested for



long periods of time in Missouri and Iowa on potable water systems.



     Results of testing demonstrated that the RSC selectively removed



radium to concentrations below the MCL for extended lengths of time as



long as iron fouling was eliminated.  The investigators concluded that



iron, if present, must be removed prior to treatment with the RSC



system.



Test Results



Bellvllle, Texas17 —



     A full scale trial to demonstrate radium removal began at Bell'



vllle, Texas in February, 1983.  Radium results are given in Table 26.



The table shows that the RSC removed total radium from Influent concen-



trations of 81 to 102.3 pCi/L to effluent concentrations of below the



MCL for a period from February 2, 1983 to April 13, 1983.  The RSC



system was also tested at different loading rates given In Table 26 and



Indications were that the system can operate at a loading rate of up to



10 gpm/ft^ with no reduced radium removal.
                                  100

-------
 Souch Superior, Wyoming1J* —



      A full scale RSC system was tested at South Superior Wyoming.  A



 •even foot high by five foot diameter conventional water conditioning



. vessel was utilized for this system.  The Internal distributors were



 •wrapped with 40 mesh screen to retain the RSC.  The system was operated



 downflow at up to 10 gpra/ft^ and was backwashed twice/week due to



 severe iron particulate problems.  Effluent samples were taken every



 two weeks.  Test results are given in Table 27.  The RSC bed was a good



 filter for participates which could not be completely removed by back-



 washing.  This is partially due to resin density which only allows a



 backwash rate of 5 gpm/ft? compared to a downflow rate of 10 gpm/ft .



 The particulates are driven deep into the bed and aren't effectively



 removed.  This system was operated successfully for eight (Table 27)



 months before particulate iron leakage through the bed carried radium



 into the effluent.  The system averaged greater than 751 radium removal



 from November 3, 1983 to September 24, 1984, even though it was operat-



 ing as an Iron particulate removal system as well as a radium removal




 system.



 Demonstration Columns"' —



      Demonstration columns that were 5 ft high X 2 in diameter, con-



 taining 4 ft depth of RSC were operated in St. Charles County, Missouri,




 Washington, Iowa, Eldon, Iowa, Holstein, Iowa and West Bend, Iowa.  The



 4 ft depth o~ RSC decreases to a 3 ft depth as it converts to the cal-



_clum form.  These syi 
-------
TABLE 26. RADIUM REMOVAL CHRONOLOGY - BELLVILLE, TEXAS17
Date
21 2/83
21 2/83
2/ 3/83
21 3/83
21 7/83
21 7/83
21 8/83
21 8/83
21 9/83
2/17/83
2/17/83

2/17/83
2/23/83

3/2/83

3/9/83
3/16/83

3/16/83
3/23/83

3/30/83

4/6/83

4/13/83

Gross
Alpha
Comments pcl/L
Well 16 86
System Start
Well 16 85
3.7 gpm/ft2
Well 16 124
6.4 gpn/ft2
7.4 gpm/ft2
9.3 gpn/ft2
10.0 gpm/ft2
Well 16 89
.5 mm gal
(5.6 gpm/ft2)
Texas Health Dept.
1.0 mm gal
(5.5 gpm/ft2)
1.5 mm gal
(5.7 gpm/ft2)
Start 24 hr/day oper
4.5 mm gal
(6 gpm/ft2)
Texas Health Dept. 3
7.1 mm gal
(6 gpm/ft2)
9.6 mm gal 5
(11 gpm/ft2)
12.0 mm gal 3
(11 gpm/ft2)
14.4 mm gal
(10.7 gpm/ft2)
.0 + 11.0
< 7.0
.0 + 10.0
< 7.0
.0 + 13.0
< 7.0
< 2.0
< 2.0
< 2.0
.0 + 8.0
< 7.0

< 2.0
< 2.0

< 2.0

.< 2.0
< 2.0

.7 + 1.7
< 7.0

.0 * 3.0
~~
.0 + 2.0

< 2.0

Gross
Beta
pci/L
30.0 + 4.0

-------
     TABLE 27.  RADIOLOGICAL RESULTS FROM FULL SCALE SYSTEM
                       SOUTH SUPERIOR, WYOMING18
Date
6/24/83
7/11/83
7/27/83
8/08/83
8/24/83
9/08/83
11/3/83
12/12/83
12/19/83
12/27/83
1/03/84
1/09/84
1/17/34
1/23/84
2/06/84
2/21/84
3/05/84
3/19/84
4/03/84
4/16/84
4/30/84
5/14/84
5/29/84
6/11/84
6.18/84
7/02/84
7/16/84
7/30/84
8/13.84
8/27/84
9/10/84
9/17/84
9/24/84
Sample
Description
Raw Water
Raw Water
Raw Water
Raw Water
Raw Water
Raw Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Finished Water
Gross
Alpha
pci/L
15+5
26+5
34+2
8.9+T.7
24+2
28+4
3.6+1.1
2.2+0.8
5.1+1.0
1.25}.9
16+3
5.1+2.7
7.5+1.9
8.8+1.4
6.6+2.0
10+3
7.3+2.5
16+3
16+4
8.0+2.2
0.7+2.2
5.3+2.4
4.1+2.0
5.2+3.0
4.0+3.0
1.1+2.2
8.8+3.6
6.5+4.2
3.4+3.8
3.0+3.1
14+4
11.5+5.2
7.1+3.5
Gross
Beta
pci/L
19+4
26+4
20+1
17+3
15+1
26+3
9.1+1.1
20+1
20+1
16+1
15+2
13+2
15+2
15+1
14+2
34+2
12+2
15+2
12+2
13+2
11+4
8.8+1.7
9.9+1.5
11+2
8.9+2.4
14+2
16+3
12+3
4.7+2.9
15+3
19+3
6+i
15.6+2.7
RA226
pci/L
20+1
11+1
16+1
12+1
13+1
13+1
3.9+1.2
0.1+0.1
0.1+0.1
2.2+<).4
3.0+0.4
2.5+0.4
1.0+0.4
2.9+0.4
2.7HH3.4
2.1+0.4
1.4+0.3
2.9+0.5
6.5+1.0
3.8+0.9
0.5+0.1
0.9+0.2
1.6+0.3
1.3+0.3
3.8+0.6
4.0+0.5
5.3+0.8
3.9+0.8
11.4+0.9
6.2+1.0
10+1
6.3+0.6
U.7+1.0
RA228
pcl/L
0.0+4.2
2.2+2.5
0.0+2.4
2.5+3.4
0.9+1.3
1.8+1.4
1.1+2.1
1.4+1.3
0.0+1.9
0.0+1.8
0.0+1.7
1.6+3.6
0.6+1.8
2.0+1.4
0.0+1.6
0.0+1.0
0.6+2.1
0.5+0.6
0.0+2.4
2.6+2.8
0.0+1.3
0.0+1.2
0.9+1.1
1.4+1.2
0.8+1.3
0.7+2.5
3.1+2.2
1.3+3.6
0.0+1.8
1.5+2.7
3.9+3.3
0.8+2.3
2.1+2.5
Results by Core Laboratories, Inc., Casper, Wyoming
                                   103

-------
the four systems in lova were located downstream of an existing iron



removal system and upstream of water softeners (if they existed).



     The RSC unit in Washington, Iowa operated for eleven months with



greater than 90Z radium removal.  Iron was not a problem at this



location and no iron removal treatment was used.



     The RSC unit at Eldon, Iowa leaked radium after two months of



operation because the iron removal system did not remove enough iron



to prevent fouling of the RSC.



     The RSC unit at Holstein, Iowa removed greater than 902 of the



radium for a period of 8 months.  This town had iron removal treatment



that was adequate for the RSC system.



     The RSC unit at West Bend, Iowa removed greater than 90Z of the




radium for a period of ten months.  Adequate iron removal treatment



was also in operation for this town.



     In summary these results show that the RSC is satisfactory for



radium removal if iron is kept from fouling the RSC system.



REVERSE OSMOSIS



     Reverse Osmosis (RO) has been demonstrated to be an effective



process for removal of radium from drinking water.  Table 28 gives



Information on eight reverse osmosis systems in Sarasota County,



Florida^O that were tested for radium removal.  The type and perform-




ance of the eight systems varied, but they all lowered the radium



concentration below the MCL.  The groundwater throughout Sarasota




County, Florida is generally poor, containing high concentrations of



hardness, sulfate, sodium, and chloride..as veil as Ra-226. , Therefore,



in addition to solving the radium problem the RO systems are producing



a better overall quality of water.  Another advantage of RO is that it





                                   104

-------
requires a small space as compared tc the space needed for conventional



coagulation or lime softening*  The principal disadvantage of RO Is




high operating cost because of high energy consumption, large quantity



of reject water, and pretreatment requirements.



     New low pressure (< 200 psl) membranes have been developed



recently and they have been shown to remove greater than 95Z of the




divalent cations including radium.  Low pressure membranes will reduce



energy costs.  Also new high pressure RO units that are well-maintained




should give better than 98Z radium removal.



ELECTROD1ALYSIS REVERSAL?!




     Electrodlalysls (ED) is an electrochemical separations process in



which Ions are transferred through membranes from a less concentrated



to a more concentrated solution as a result of the flow of direct



electric current.  Electrodlalysls Reversal (EOR) is a relatively




recent development in which the polarity of the direct current is



changed approximately every 15 minutes.  A compartment which was form-



erly a demineralizing compartment becomes a concentrating compartment



(and vice versa).  It is necessary to switch valves which feed and



collect these two streams shortly after the current is reversed.  It is



also necessary to divert both streams at the time of polarity reversal




for a period of about one minute to purge both compartments before the



demineralizing stream returns to making product water.  The whole



process is completely automatic.




     EDR, therefore, is achieved by aymetrical equal-time operation



in alternating directions so that insoluble or sparingly-soluble sub-



stances which would otherwise coat the membranes will, during the



reversed polarity period, tend to be removed.






                                   105

-------
       TABLE 28.  RADIUM REMOVAL BY RC IN SARASOTA COUNTY, FL.20
Loca- Membrane Operating
tlon Type Press (PSI)
Venice
Sorrento
Shores
Spanish
Lakes
Bay Lakes
Estates
Kings
Gate
Sarasota
Bay
Bay Front
Nokonls
School
Hollow 400
Fiber
Spiral 425
Wound
Spiral 400
Wound
Hollow 400
Fiber
Hollow 400
Fiber
Hollow 400
Fiber
Hollow 200
Fiber
Spiral 200
Wound
Actual
Stages Recovery Ra-226(pCi/L) Removal
No. Z in out Z
2 54 3.4 0.3 91.2
1 39 4.6 0.2 95.7
2 31 10.5 1.2 88.6
2 NA 3.2 0.1 96.9
2 NA 15.7 2.0 87.3
1 50 20.5 0.3 98.5
1 28 12.1 0.6 95.0
1 NA 11.1 0.5 95.5
     An operating EDR unit requires a supply of pressurized feed water




in the range of 70 to 90 psi, a reversing DC power supply, and an array



of demineralization stages in a membrane stack or stacks.  In order to



conserve water, a concentrate recirculatlon pump is used to recirculate



most of the water to the concentrating compartments.  The actual waste



concentrate flow is then simply regulated by a control valve in the



feed line and forces an equal amount of water out to waste.



     EDR is the only membrane process which is symmetrically rever-



sible.  Reversal inhibits the formation of scale and fouling films on



the surface of membranes without addition of chemicals to the feed    :



water, minimizes pretreatment, and maximizes recovery.





                                  106

-------
     A major application of EDR Is the demineralization of brackish     |



waters of several thousand mg/L of salts to potable levels below



500 mg/L.  The Florida waters, mentioned previously under reverse



osmosis, that had high dissolved salts and radium would also be good



candidates for EDR treatment.  EDR systems have found great favor in



remote locations and village water supply applications, where limited



operator skills and availability of chemicals make other processes




less attractive.



     EDR systems remove about 40Z of the dissolved salts per stage.



Therefore, if the total radium concentration in the source water were




40 pCi/L, then five stages would probably be required to remove radium



to a concentration below the MCL.  The specific water analysis is



important for individual cases because the rate of removal of individ-



ual ions will vary depending on such conditions as:  water profile of




ions, water temperature, total ion removal percentage, number of stages,



water recovery rate and current density.



     EDR systems are relatively expensive about the same as reverse



osmosis, but they should be considered for small systems on brackish



waters where little operator attention would be available.



POTASSIUM PERMANGANATE GREENSAND FILTRATION




     The literature shows that radium was removed across KMn04 green-




sand processes; removal was variable; and the exact reasons for radium



removal were not known.*»'  After reviewing this earlier literature



and the manganese dioxide results of the University of Iowa Study



(Section 5) there are very strong indications that a properly controlled



Mn02-flltration process would do a fine job of removing radium.




     Oxidation of Mn*** ions occurring in natural waters and/or reduction
                                  107

-------
of XMn04 leads to Che formation of colloidal MM02>22  This possesses a
negative surface charge over the normal pH range in natural vaters.
Figures 20 through 23 show good sorption of radium by Mn02 above pH S.
The high surface area of the colloidal Mn02 tlong with < ne negative
charge exerts a dominating influence on the concent ratii i of many
elements, e-g  the heavy metals, in the aquatic environuent.**
     Radium concentrations in water are normally in the 10"' mg/L
range.  Iron, manganese and other cations that strongly compete with
radium for adsorption sites on MnOj are present in the 10~* mg/L con-
centration range*  Therefore, £in order to obtain good radium sorption,
it would be necessary to remove as much of the competing cations as
possible and reserve the sorption sites on Mn02 for radium.  This is
possible by installing a good iron manganese removal process before the
Mn(>2 - filtration process or making sure that some MnC>2 is still avail-
able for radium sorption* ,When KMn04 is added during iron removal,
Fe2"1" is oxidized and precipitated as Fe  .  The KMnO^ is reduced to
Mn<>2 which is available for sorption of cations.  Any Fe2* that is
•till present will compete with radium for sorption sites on the
Therefore, a final treatment stage consisting of stoichiometric
oxidation with KMnO^ to have Mn(>2 available for radium adsorption and
final filtration would be required for complete removal of radium.
     This is verified by the data given in table 29.  The table shows
that the best radium removal was obtained at Rersher, IL. where Ra-226
removal ranged from 47 to 56 percent.  It can be seen that the iron
concentraion ranges from 0.1 to 0.4 mg/L and is relatively low compard
to manganese concentration that ranges from 0.41 .to 0.63 mg/L.  Iron/
manganese removal at Harsher consists of aeration, settling, chlorina-
                                   108

-------
tlon, filtration and created vith polyphosphate and potassium pennan~



ganate.23  This data strongly indicates that iron was removed and Mn**"



vas oxidized to MnO£, KMnO^ was also reduced to MnC>2 and, therefore,



sorption sites on the MnO£ were available for radium removal*



     Some selected data, furnished by the State of Virginia Department



of Health in Danville, VA, and presented in table 30, show that three



treatment systems utilizing continuous feed KMnOt, 30-minute contact



time, and manganese greensand all removed radium to concentrations



below the MCL.  This data shows that the KMn04 - greensand processes



can be operated for optimum radium removal.
                                  109

-------
TABLE 29.  RADIUM-226, IRON, AND MANGANESE REMOVALS BY IRON- AND
                   MANGANESE-REMOVAL PROCESSES23
Ra-226


City
Adair, IA
Eldon, IA
Esthervllle, IA
Grinnell, IA
Herscher, IL








Holsteln, IA
Stuart, IA

Raw Water
pH
6.7-6.9
7.8
7.7
7.6
7.6-8.3
7.6-8.3
7.6-8.3
7.6-8.3
7.6-8.3
7.6-8.3
7.6-8.3
7.6-8.3
7.6-8.3
7.4-7.6
7.6-7.9
Raw
Water
pCi/L
13
49
5.7
6.7
14.9
14.5
14.9
14.3
14.0
13.9
13.9
14.1
14.3
13
16
Treated
Water
pCi/L
8
43
5.1
5.7
6.6
6.4
6.9
6.9
6.9
6.8
7.3
6.3
6.5
7
12

Removal
per cent
38
12
11
15
56
56
54
42
51
51
47
55
55
46
25
Raw
Water
mg/L
1.1
2.0
2.0
0.7
0.2
0.4
0.1
0.1
0.1
0.1
0.2
0.1
0.1
1.8
0.94
Iron
Treated
Water
mg/L
0.02
0.3
0.67
0.41
0
0
0
0
0
0
0.1
0
0
0.09
0.03
Manganese

Removal
per cent
98
85
66
42
-
-
-
—
-
—
-
_
-
95
97
Raw
Water
mg/L
0.01
0.01
0.24
0.01
0.47
0.41
0.48
0.39
0.45
0.63
0.44
0.53
0.50
0.15
0.01
Treaced
Water
mg/L
0.01
0.01
0.27
0.01
0.02
0.01
0.01
0
0
0
0.13
0.02
0
0.01
0.01

Removal
per cent
^ .
_
-
—
96
98
98
100
100
100
70
96
100
93
—

-------
      TABLE 30.  RADIUM REMOVAL WITH THE KMnO* - GREENSAND PROCESSt
Location
*1
t2
n
Description
influent
effluent
influent
effluent
influent
effluent
Cross x
pCi/L
6.6
1.5
22.3
3.9
6.2
0.6
Ra-226
pCi/L
3.7
0.6
10.7
0.5
5.3
0.9
Ra-228
pCi/L
3.3
0.0
10.8
0.5
5.9
1.4
•Total Ra
- pCi/L
7.0
0.6
21.5
1.0
11.2
2.3
t Information provided by State of Virginia Department of Health,
  Danville, VA.
                                  Ill

-------
                               SECTION 9



                                 COSTS




     For many communities, a specific process would be selected because



it is desirable to remove radium along with other dissolved solids.



For example, if the water supply has high hardness, then ion exchange



or lime softening would be selected to remove both radium and hardness*



A community that has a brackish water supply with radium would select



reverse osmosis or electrodlalysis.  If the water supply is of excel-



lent quality, except for radium, then the RSC, calcium cation exchange



or KMn04 with filtration may be selected.  Calcium cation exchange and




KMn04 with filtration have never been used exclusively for radium



removal and pilot testing would be required before full scale instal-



lation.  If KMn04 with filtration is used on water of otherwise excel-




lent quality, then It would be necessary to add manganous ion and the



oxidation of the manganous ion and reduction of KMnO£ would form the



Mn02 that is necessary to sorb radium.  Only a small quantity of



manganous ion (about one mg/L) and KMn04 added stoichiometrlcally is



all that would be required for water that has a low concentration of



dissolved solids.  The spent Mn02 would precipitate on the filter.



Water supplies frequently contain both iron and hardness, In which case



both iron and hardness removal processes would be required.  It was



shown in section 5 that conventional iron removal consisting of aera-



tion and filtration would not be expected to remove more than 10 to 20Z




of the radium present.  There are possibilities for optimizing an iron



removal process such as backwashlng the sand with dilute acid or adding



extra KMn04 for better radium removal.  These modifications to the iron



removal process would be specific to a water supply and pilot testing






                                  112

-------
of the modification on each specific water supply should be performed



before full scale application.



     The use of the RSC discussed in section 7 of this report is an



example of a possible cost savings.  In this case the RSC was used to



treat the waste brine instead of the entire water supply.  Therefore, a



much smaller quantity of complexer (5 ft^ instead of 30 ft^) and a



smaller reaction vessel was required.  This approach will also Involve



a smaller quantity of complexer with easier handling for final disposal.



     Table 26, compiled from the literature,24,25,26 compares approxi-



mate costs of the different processes for removing radium.  The table



does not include disposal costs.  As mentioned, each specific water



supply that needs radium removal should be considered separately.



Objective pilot testing on each water supply would be required to



•elect the most economical process or processes for that water source.
                                  113

-------
    TABLE 26.  ESTIMATED COSTS OP RADIUM REMOVAL PROCESSES24' 25»-
                    100.000 gal/day
Process
 const.    O&M      Total
   $      $/yr.  $/1000 gal.
                                     500.000 gal/day
         const.    O&M      Total
           $      $/yr.  $/1000 gal
Ion
Exchange*
Reverse
Osmosis§

Electro-
dialysis**

Dow
•ComplexerH-
 99,900   17,100      0.80    164,100    27,100    0.26
Lime Soda     121,200   16,900
Softening'

Iron &        157,000    9,500
Manganeset
194,900   43,600


221,300   36,900


 99,900    5,700
KMn04 with    122,800   10,000
Filtration**
                      0.85    200,200    24,200    0.26
                      0.76    274,000    21,700    0.30
1.82    684,900   169,400    1.37
1.74    737,600   158,100    1.36
0.48    164,100     9,000    0.15
                      0.66    212,700    19,900    0.24
* Costs brought up to 1986 with ENR index.26  Capital cost amortized at
  10Z for 20 yrs.  Haste disposal costs not included, pretreatment and
  post treatment costs not included.

+ For strong acid ion exchange resin.

t Single stage

t Spray aeration plus pressure sand filtration, chlorine, KMnO^ and
  polymer included.

$ For 2000 mg/L TDS, low pressure membrane.

** For 2000 mg/L TDS

•H- Used ion exchange construction coat and 1/3 of O&M cost..

it KMn04 feed system to form Mn02 followed by pressure sand filtration.
                                   114

-------
                              REFERENCES






1.  Snoeylnk, V. L., Pfeffer, J. L., Snyder, D. V., Chambers., C. C.,




    Barium and Radium Removal from Groundwater by Ion Exchange.



    EPA-600/2-84-093, U.S. Environmental Protection Agency,



    Cincinnati, Ohio, 1984, 115 pp.  .




2.  Bowers, E. Wter Quality and Treatment, A Handbook of Public Water



    Supplies.  AWWA Ed., McGraw-Hall, New York (1971).




3.  Weber, W. J.  Pbysicochemical Processes for Water Quality Control*



    Wiley-Interscience, New York (1972).




4.  Downing, D. G., Quinn, R. A., and Bettacchi, R. J. Dealkaliniza-



    tion by Weak Acid Cation Exchange. Ini. Water Engineering, 32,



    July, 1968.



5.  Simmonds, H. K., Smith, A. J., and Gardner, E. R.  The New



    Cheam Works of the Sutton District Water Company.  Advanced




    copy presented to the South Eastern Section of the Institution



    of Water Engineers and Scientists, England.



6.  Myers, A. G., Snoeyink, V. L., Snyder, D. W., Removing Barium



    and Radium Through Calcium Cation Exchange.  Journal American



    Water Work Association, 77:5:60 (May 1985).



7.  Valentine, R* L., Splinter, R. C., Hulholland, T. S., Baker,




    J. M., Nogau, T. M., Horng, J. J.  A Study of Possible Economical



    Ways of Removing Radium from Drinking Water.  CR-810575,  U.S.



    Environmental Protection Agency, Cincinnati, Ohio, 1987,  104 pp.



    (In process)



8.  Schliekelman, R. J.  Determination of Radium Removal Efficiencies



    in Iowa Water Supply Treatment Processes.  Technical Not* ORP/
                                 115

-------
     TAD-76-1, U.S. Environmental Protection Agency, Office of Radiation




     Programs, Washington, D.C. (Apr. 1976)



 9.  Brink, W. L., Schllekelman, R. J., Bennett, D. L., Bell,- C. R.,




     and Markwood, I. M.  Radium-Removal Efficiencies in Water-



     Treatment Processes.  Journal American Water Works Association



     70:1:31 (Jan. 1978).




10.  Bennett, D. L.  The Efficiency of Water Treatment Processes in



     Radium Removal.  Journal American Water Works Association.



     70:12:698 (Dec. 1978).



11.  Levin, 0. M., Ryan, R. K., and Strong, K. P., Leaching of Radium



     from Uranium Tailings.  Management, Stabilization and Environ-



     mental Impact of Uranium Mill Tailings. NEA and OECD (1978).



12.  Stahel, E. P., Menetrez, M.Y.  Removal of Radium from North




     Carolina's Drinking Water.  CR-811119, U.S. Environmental



     Protection Agency, Cincinnati, Ohio, 1987 (In process).




13.  Hangelson, K. A., Evaluation of Radium Removal and Waste



     Disposal System for Small Community Water Supply.  CR-812691,




     U.S. Environmental Protection Agency, Cincinnati, Ohio.  1987.



     (In process)



14.  Sorg, T. J., Logsdon, G. S., Treatment Technology to Meet the



     Interim Primary Drinking Water Regulations for Inorganics:




     Part 5.  Journal American Water Works Association, 72:7:411



     (July 1980)




15.  Lange, N. A., Lange's Handbook of Chemistry.  Twelfth Edition,



     McGrav Hill, Inc.  1979.
                                  116

-------
16.  Laitinen, H. A., Chemical Analysis, An Advanced Text and



     Reference.  McGraw-Hill Book Company, p. 165, 1960.



17.  Rozelle, R. E., Ma, K. W., A New Putable Water Radium/Radon



     Removal System.  Proceedings AWWA Seminar on Control of



     Inorganic Contaminants, Las Vegas, Nevada (June 5-9, 1983).



18.  Rozelle, R. E., Fales, D. M., Flanagan, E. C., Potable Water



     Radium Removal Update on Tests in Missouri, Iowa and Wyoming



     by Dow Chemical Company.  Dow Chemical Co., Midland, MI.



     (1985).




19.  Boyce, T. D., Boom, S., Removal of Soluble Radium from Uranium



     Minewaters by a Selective Complexer.  Society of Mining



     Engineers of AIME, 82-23 (February, 1982).



20.  Sorg, T.S., Forbes, R. W., and Chambers, D. S.  Removal of



     Radium-226 From Sarasota County, FL., Drinking Water by



     Reverse Osmosis.  Journal American Water Works Association




     12:4:230  (April, 1980).



21.  EDR Electro Dialysis Reversal, Ionics, Inc., Watertown, MA,



     Bulletin No. 121-E, March, 1984.




22.  Gray, M., Heavy Metal Adsorbent Role for Mn02*  Word Water,



     p 44, March, 1980.




23.  Bennett, D. L., Bell, C. R., Markwood, I. M., Determination of



     Radium Removal Efficiencies in Illinois Water Supply Treatment




     Processes.  ORP/TAD-76-2, Radlochemistry and Nuclear Engineering



     Branch, U.S. Environmental Protection Agency, Cincinnati, OH,



     May, 1976.



24.  Hansen, S. P., Gumerman, R. C., Gulp, R. L., Estimating Water



     Treatment Costs, Volume 3.  Cost Curves Applicable to 2,500 gpd
                                  117

-------
     to 1 mgd Treatment Plants.  EPA-600/2-79-162C, U.S. Environmental



     Protection Agency, Cincinnati, Ohio, August, 1979, 178 pp.



25.  Gumerman, R. C., Harris, B. £., Hansen, S. P., Estimation of



     Small System Water Treatment Coats.  BPA-600/2-84-184a, U.S.



     Environmental Protection Agency, Cincinnati, OH, 1983, 505 pp.



26.  ENR, January 8, 1987, Market Trends,  p 34.
                                 113

-------

-------
                   Reproduced by NTIS
                   National Technical Information Service
                   U.S. Department of Commerce
                   Springfield,  VA  22161
        13
     »H  ^

 °  § "S    o>  >
                   This report was printed specifically for you
                   order from our collection of more than 1.5
                   million technical reports.
For economy and efficiency, NTIS does not maintain stock of its vast
collection of technical reports.  Rather, most documents are printed for
each order.  Your copy is the best possible reproduction available from
our master archive. If you have any questions concerning this document
or any order you placed with NTIS, please call our Customer Services
Department at (703)487-4660.

Always think of NTIS when you want:
• Access to the technical, scientific, and engineering results'generated
by the ongoing  multibillion dollar R&D program of the U.S. Government.
• R&D results from Japan, West Germany, Great Britain, and some 20
other countries, most of it reported in English.

NTIS also operates two centers that can provide you with valuable
information:
• The Federal Computer Products Center - offers software and
datafiles produced by Federal agencies.
• The Center for the Utilization of Federal Technology - gives you
access to the best of Federal technologies and laboratory resources.

For more information about NTIS, send for our FREE NTIS Products
and Services Catalog which describes how you can access this U.S. and
foreign Government technology. Call (703)487-4650 or send this
sheet to NTIS, U.S. Department of Commerce, Springfield, VA 22161.
Ask for catalog, PR-827.
                   Name	
                   Address.
                   Telephone
                                - Your Source to U.S. and Foreign Government
                                   Research and Technology.

-------