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.
------- |