United States
Environmental Protection
Agency
Arsenic Treatment Technology
     Evaluation Handbook
      for Small Systems

-------
Office of Water (4606M)
EPA816-R-03-014
July 2003
www.epa.gov/safewater
                                               Printed on Recycled Paper

-------
                                                        Disclaimer
The information in this document has been subjected to the Agency's peer and administrative re-
views and has been approved for publication as an EPA document. Mention of trade names or
commercial products does not constitute an endorsement or recommendation of use.

-------
                                            Executive  Summary
In January 2001, the U.S. Environmental Protection Agency (USEPA) published a final Arsenic
Rule in the Federal Register. This rule established a revised maximum contaminant level (MCL)
for arsenic at 0.010 mg/L. All community and non-transient, non-community (NTNC) water sys-
tems, regardless of size, will be required to achieve compliance with this rule by January 2006.

This technical handbook is intended to help small drinking water systems make treatment decisions
to comply with the revised arsenic rule. A "small" system is defined as a system serving 10,000 or
fewer people.  Average water demand for these size systems is normally less than 1.4  million
gallons per day (MOD).

Provided below is a checklist of activities that should normally take place in order to comply with
the new Arsenic Rule.  Many of the items on this checklist refer to a section in this handbook that
may help in completing the activities.
                           Arsenic Mitigation Checklist
1.  Monitor arsenic concentration at each entry point to the distribution system (see Section 1.3.2).
2.  Determine compliance status. This may require quarterly monitoring.  See Section 1.3.2 for
   details on Arsenic Rule compliance.
3.  Determine if a non-treatment mitigation strategy such as source abandonment or blending can
   be implemented. See Sections 2.1.1 through 2.1.3 for more detail and Decision Tree 1, Non-
   Treatment Alternatives.
4.  Measure water quality parameters. See Section 3.1.1 for more detail on water quality param-
   eters that are used in selecting a treatment method.
       •   Arsenic, Total                       •  Nitrite
       •   Arsenate [As(V)]                     •  Orthophosphate
       •   Arsenite [As(III)]                     •  pH
       •   Chloride                          •  Silica
       •   Fluoride                           •  Sulfate
       •   Iron                              •  Total Dissolved Solids (TDS)
       •   Manganese                         •  Total Organic Carbon (TOC)
       •   Nitrate
5.  Determine the treatment evaluation criteria.  See  Section  3.1.2 for more detail on parameters
   that are used in selecting a treatment method.
       •  Existing Treatment Processes
       •  Target Finished Water Arsenic Concentration
Arsenic Treatment Technology Evaluation Handbook for Small Systems

-------
       •  Technically Based Local Limits (TBLLs) for Arsenic and TDS
       •  Domestic Waste Discharge Method
       •  Land Availability
       •  Labor Commitment
       •  Acceptable Percent Water Loss
       •  Maximum Source Flowrate
       •  Average Source Flowrate
       •  State or primacy agency requirements that are more stringent than those of the USEPA.
6.  Select a mitigation strategy using the decision trees provided in Section 3.2. These trees lead to
    the following mitigation strategies.
       •  Non-Treatment & Treatment Minimization Strategies
          O  Source Abandonment
          O  Seasonal Use
          O  Blending Before Entry to Distribution System
          O  Sidestream Treatment
       •  Enhance Existing Treatment Processes
          O  Enhanced Coagulation/Filtration
          O  Enhanced Lime Softening
          O  Iron/Manganese Filtration
       •  Treatment (Full Stream or Sidestream)
          O  Ion Exchange
          O  Activated Alumina
          O  Iron Based Sorbents
          O  Coagulation-Assisted Microfiltration (CMF)
          O  Coagulation-Assisted Direct Filtration (CADF)
          O  Oxidation/Filtration
       •  Point-of-Use Treatment Program
          O  Activated Alumina
          O  Iron Based Sorbent
          O  Reverse Osmosis
7.  Estimate planning-level capital and operations and maintenance (O&M) costs for the mitiga-
    tion strategy using the costs curves provided in Section 4. Include costs for arsenic removal and
    waste handling.  If this planning level  cost is not within a range that is financially possible,
    consider using different preferences in the decision trees.
8.  Evaluate design considerations for the mitigation strategy. See Section 2.5 for enhancing exist-
    ing treatment processes and Sections 6 through 8 for the design of new treatment processes.
Arsenic Treatment Technology Evaluation Handbook for Small Systems

-------
9.  Pilot the mitigation strategy.  Although not explicitly discussed in this Handbook, piloting the
   mitigation strategy is a normal procedure to optimize treatment variables and avoid implement-
   ing a strategy that will not work for unforeseen reasons. For many small systems, piloting may
   be performed by the vendor and result in a guarantee from the vendor that the system will
   perform.
10. Develop a construction-level cost estimate and plan.
11. Implement the mitigation strategy.
12. Monitor arsenic concentration at each entry point to the distribution system to ensure that the
   arsenic levels are now in compliance with the Arsenic Rule - assumes centralized treatment
   approach, not point-of-use treatment.
Table ES-1 provides a summary of information about the different alternatives for arsenic mitigation
found in this Handbook. Please note that systems are not limited to using these technologies.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                 in

-------
                Table  ES-1.  Arsenic Treatment Technologies  Summary Comparison.
                                                         (Iof2)

Factors

USEPA BAT B
USEPA SSCTB
System Size B-D
SSCT for POU B
POU System Size B-D
Removal Efficiency
Total Water Loss
Pre- Oxidation Required F




Optimal Water
Quality Conditions



Operator Skill Required
Waste Generated

Other Considerations

Centralized Cost
POU Cost
Sorption Processes
Ion Exchange
IX
Yes
Yes
25-10,000
No

95% E
1-2%
Yes



pH 6.5 - 9 E
< 50 m&L SO42' '
< 500 mg/L TDS K
< 0.3 NTU Turbidity


High
Spent Resin, Spent Brine,
Backwash Water
Possible pre & post pH
adjustment.
Pre-filtration required.
Potentially hazardous brine
waste.
Nitrate peaking
Carbonate peaking affects pH.
Medium
-
Activated Alumina A
AA
Yes
Yes
25-10,000
Yes
25-10,000
95% E
1-2%
Yes
pH5.5 - 6 '
pH6- 8.3 L
< 250 rng^ C1- '
< 2 mg/L F- :
< 360 mg/L S042'K
< 30 mg^ Silica M
< 0.5 mg/L Fe+31
< 0.05 mg/L Mn+2 :
< 1,000 mg^ TDS K
< 0.3 NTU Turbidity
Low A
Spent Media, Backwash
Water

Possible pre & post pH
adjustment.
Pre-filtration may be
required.
Modified AA available.

Medium
Medium
Iron Based
Sorbents
IBS
No c
Noc
25-10,000
No c
25-10,000
up to 98% E
1- 2%
Yes ฐ




pH6- 8.5
< 0.3 NTU Turbidity



Low
Spent Media, Backwash
Water

Media may be very
expensive. ฐ
Pre-filtration may be
required.

Medium
Medium
Membrane
Processes
Reverse
Osmosis
RO
Yes
Yes
501-10,000
Yes
25 -10,000
> 95% E
15-75%
Likely11




No Particulates



Medium
Reject Water

High water loss (15-
75% of feed water)

High
Medium
 A Activated alumina is assumed to operate in a non-regenerated mode.
 B USEPA, 2002a.
 c IBS's track record in the US was not established enough to be considered as Best Available Technology (BAT) or Small System Compliance
  Technology (SSCT) at the time the rule was promulgated.
 D Affordable for systems with the given number of people served.
 E USEPA, 2000.
 F Pre-oxidation only required for As(III).
 0 Some iron based sorbents may catalyze the As(III) to As(V) oxidation and therefore would not require a pre-oxidation step.
 H RO will remove As(III), but its efficiency is not consistent and  pre-oxidation will increase removal efficiency.
 1 AwwaRF, 2002.
 1 Kempic, 2002.
 K Wang, 2000.
 L AA can be used economically at higher pHs, but with a significant decrease in the  capacity of the media.
 M Clifford, 2001.
 N Tumalo, 2002.
 0 With increased domestic use, IBS cost will significantly decrease.
Arsenic Treatment Technology Evaluation Handbook for Small Systems


-------
               Table ES-1.  Arsenic Treatment Technologies Summary Comparison.
                                                    (2 of 2)
                                                         Precipitative Processes
Factors
USEPA BAT B
USEPA SSCT B
System Size B-D
SSCT for POU B
POU System Size B-D
Removal Efficiency
Total Water Loss
Pre- Oxidation Required F
Optimal Water
Quality Conditions
Operator Skill Required
Waste Generated
Other Considerations
Centralized Cost
POU Cost
Enhanced Lime
Softening
LS
Yes
No
25-10,000
No

90% E
0%
Yes
pH 10.5 - 11 :
> 5 mg/LFe+31
High
Backwash Water,
Sludge (high volume)
Treated water requires pH
adjustment.
Low1:!
N/A
Enhanced
(Conventional)
Coagulation
Filtration
CF
Yes
No
25-10,000
No

95% (w/ FeCy E
< 90% (w/ Alum) E
0%
Yes
pH5.5 - 8.5 p
High
Backwash Water,
Sludge
Possible pre & post
pH adjustment.
Low<3
N/A
Coagulation-
Assisted
Micro-
Filtration
CMF
No
Yes
500-10,000
No

90% E
5%
Yes
pH5.5 - 8.5 p
High
Backwash Water,
Sludge
Possible pre &
post pH
adjustment.
High
N/A
Coagulation-
Assisted Direct
Filtration
CADF
Yes
Yes
500-10,000
No

90% E
1-2%
Yes
pH5.5 - 8.5 p
High
Backwash Water,
Sludge
Possible pre & post
pH adjustment.
Medium
N/A
Oxidation
Filtration
OxFilt
Yes
Yes
25-10,000
No

50-90% E
1-2%
Yes
pH5.5 - 8.5
>0.3 mg/L Fe
Fe:As Ratio > 20:1
Medium
Backwash Water.
Sludge
None.
Medium
N/A
B USEPA, 2002a.
D Affordable for systems with the given number of people served.
E Depends on arsenic and iron concentrations.
F Pre-oxidation only required for As(III).
1 AwwaRF, 2002.
p Fields, et aL, 2002a.
Q Costs for enhanced LS and enhanced CF are based on modification of an exisitng technology. Most small systems will not have this technology in
place.
Arsenic Treatment Technology Evaluation Handbook for Small Systems

-------
Preceeding Page Blank
                                                                      Contents
     EXECUTIVE SUMMARY	i

     CONTENTS	vii

     LIST OF FIGURES	x

     LIST OF TABLES	JCHI

     LIST OF ACRONYMS AND ABBREVIATIONS	xiv

     EQUATION NOMENCLATURE	xvii

     1.0  BACKGROUND	1
       1.1 Purpose of this Handbook	1
       1.2 How to Use this Handbook	1
       1.3 Regulatory Direction	2
         1.3.1 The Arsenic Rule	2
         1.3.2 Health Effects	3
         1.3.3 Other Drinking Water Regulations	4
         1.3.4 Waste Disposal Regulations	5
       1.4 Arsenic Chemistry	8

     2.0  ARSENIC MITIGATION STRATEGIES	11
       2.1 Description of Arsenic Mitigation Strategies	 11
         2.1.1 Abandonment	 12
         2.1.2 Seasonal Use	12
         2.1.3 Blending	 13
         2.1.4 Treatment	 15
         2.1.5 Sidestream Treatment	 16
       2.2 Pre-Oxidation Processes	18
         2.2.1 Chlorine	 19
         2.2.2 Permanganate	20
         2.2.3 Ozone	21
         2.2.4 Solid Phase Oxidants (Filox-R™)	22
       2.3 Sorption Treatment Processes	23
         2.3.1 Ion Exchange	24
         2.3.2 Activated Alumina	26
         2.3.3 Iron Based Sorbents	29
       2.4 Membrane Treatment Processes	30
       2.5 Precipitation/Filtration Treatment Processes	32
         2.5.1 Enhanced Lime Softening	32
         2.5.2 Conventional Gravity Coagulation/Filtration	33


     Arsenic Treatment Technology Evaluation Handbook for Small Systems                              vii

-------
   2.5.3  Coagulation-Assisted Microfiltration	34
   2.5.5  Oxidation/Filtration	35
  2.6  Point-of-Use Treatment	37

3.0 ARSENIC TREATMENT SELECTION	39
  3.1  Selection Criteria	39
   3.1.1  Source Water Quality	39
   3.1.2  Process Evaluation Basis	41
  3.2  Process Selection Decision Trees	41

4.0 PLANNING-LEVEL TREATMENT COSTS	55
  4.1  Pre-Oxidation System Costs Using Chlorine	57
  4.2  Ion Exchange System Costs	58
  4.3  Activated Alumina System Costs	63
  4.4  Iron Based Sorbent System Costs	68
  4.5  Greensand System Costs	68
  4.6  Coagulation Assisted Microfiltration System Costs	71
  4.7  Coagulation/Filtration System Enhancement Costs	74
  4.8  Lime Softening System Enhancement Costs	76
  4.9  Point-of-Use Reverse Osmosis System Costs	77
  4.10  Point-of-Use Activated Alumina System Costs	79
  4.11  Point-of-Use Iron Based  Sorbent System Costs	80

5.0 PRE-OXIDATION DESIGN CONSIDERATIONS	81
  5.1  Chlorine Pre-Oxidation Design Considerations	81
   5.1.1  Commercial Liquid Hypochlorite	82
   5.1.2  On-Site Hypochlorite Generation	84
  5.2  Permanganate Pre-Oxidation Design Considerations	86
  5.3  Ozone Pre-Oxidation Design Considerations	88
  5.4  Solid Phase Oxidant Pre-Oxidation Design Considerations	90
  5.5  Comparison of Pre-Oxidation Alternatives	93

6.0 SORPTION PROCESS DESIGN CONSIDERATIONS	95
  6.1  Process Flow	95
  6.2  Column Rotation	96
  6.3  Sorption Theory	97
   6.3.1  Non-Regenerated Sorption Processes	98
   6.3.2  Ion Exchange Processes	98
  6.4  Process Design & Operational Parameters	100
  6.5  Column Design	 101
   6.5.1  Column Diameter	 102
   6.5.2  Column Height	 103
  6.6  Media Replacement Frequency	104
  6.7  Regeneration of Ion Exchange Resin	 105
  6.8  Waste Handling Systems	 106


Arsenic Treatment Technology Evaluation Handbook for Small Systems                               viii

-------
7.0 PRESSURIZED MEDIA FILTRATION PROCESS DESIGN CONSIDERATIONS ... 107
  7.1  Process Flow	 107
  7.2  Process Design & Operational Parameters	109
  7.3  Filter Design	 110
   7.3.1  Filter Diameter	 113
   7.3.2  Media Weight	 113
  7.4  Waste Handling System Design	 114
  7.5  Coagulant Addition System Design	 115

8.0 POINT-OF-USE TREATMENT	117
  8.1  Treatment Alternatives	 117
   8.1.1  Adsorption Point-of-Use Treatment	 117
   8.1.2  Reverse Osmosis Point-of-Use Treatment	 119
  8.2  Implementation Considerations	 121
   8.2.1  Program Oversight	 121
   8.2.2  Cost	 122
   8.2.3  Compliance Monitoring	 122
   8.2.4  Mechanical Warnings	 122
   8.2.5  Operations and Maintenance	 122
   8.2.6  Customer Education and Residential Access	123
   8.2.7  Residual Oxidant in Distribution System	 123
   8.2.8  Waste Handling	 123
  8.3  Device Certification	 124

9.0 REFERENCES ..                                                            .. 125
Arsenic Treatment Technology Evaluation Handbook for Small Systems                               ix

-------
                                                        List of Figures
Figure 1-1. Optimal pH Ranges for Arsenic Treatment Technologies	4
Figure 1-2. Flow Diagram for RO POU	8
Figure 1-3. Dissociation of Arsenite [As(III)]	9
Figure 1-4. Dissociation of Arsenate [As(V)]	9
Figure 2-1. Example of Seasonal High Arsenic Source Use	  13
Figure 2-2. Blending	  14
Figure 2-3. Sidestream Treatment	  16
Figure 2-4. Treatment and Blending	  16
Figure 2-5. Sidestream Treatment and Blending	  16
Figure 2-6. Sidestream Treatment	  17
Figure 2-7. Sidestream Treatment for RO	  17
Figure 2-8. Ion Exchange Process Flow Diagram	24
Figure 2-9. Effect of Sulfate on Ion Exchange Performance (Clifford, 1999)	25
Figure 2-10. Activated Alumina Process Flow Diagram	27
Figure 2-11. Effect of pH on Activated Alumina Performance	28
Figure 2-12. RO Membrane Process Flow Diagram	30
Figure 2-13. Two-Stage RO Treatment Process Schematic	31
Figure 2-14. Generic Precipitation/Filtration Process Flow Diagram	32
Figure 3-1. Decision Tree Overview	43
Figure 3-2. Decision Tree 1 - Non-Treatment Alternatives	44
Figure 3-3. Decision Tree 2 - Treatment Selection	45
Figure 3-4. Decision Tree 2a  - Enhanced Coagulation/Filtration	46
Figure 3-5. Decision Tree 2b  - Enhanced Lime Softening	47
Figure 3-6. Decision Tree 2c  - Iron/Manganese Filtration	48
Figure 3-7. Decision Tree 3 - Selecting New Treatment	49
Figure 3-8. Decision Tree 3a  - Ion Exchange Processes	50
Figure 3-9. Decision Tree 3b  - Sorption Processes	51
Figure 3-10. Decision Tree 3c - Filtration and Membrane Processes	52
Figure 4-1. Chlorination Capital Costs	57
Figure 4-2. Chlorination O&M  Costs	58
Figure 4-3. Ion Exchange (<20 mg/L SO42) Capital Costs	59
Figure 4-4. Ion Exchange (<20 mg/L SO42-) O&M Costs	59
Figure 4-5. Ion Exchange (<20 mg/L SO42~) Waste Disposal Capital Costs	60

Arsenic Treatment Technology Evaluation Handbook for Small Systems                                x

-------
Figure 4-6. Ion Exchange (<20 mg/L SO/) Waste Disposal O&M Costs	60
Figure 4-7. Ion Exchange (20-50 mg/L SO/) Capital Costs	61
Figure 4-8. Ion Exchange (20-50 mg/L SO/) O&M Costs	61
Figure 4-9. Ion Exchange (20-50 mg/L SO/) Waste Disposal Capital Costs	62
Figure 4-10. Ion Exchange (20-50 mg/L SO/) Waste Disposal O&M Costs	62
Figure 4-11. Activated Alumina (Natural pH) Capital Costs	63
Figure 4-12. Activated Alumina (Natural pH of 7-8) O&M Costs	64
Figure 4-13. Activated Alumina (Natural pH of 7-8) Waste Disposal O&M Costs	64
Figure 4-14. Activated Alumina (Natural pH of 8-8.3) O&M Costs	65
Figure 4-15. Activated Alumina (Natural pH of 8.0-8.3) Waste Disposal O&M Costs	65
Figure 4-16. Activated Alumina (pH Adjusted to 6.0) Capital Costs	66
Figure 4-17. Activated Alumina (pH adjusted to 6.0 - 23,100 BV) O&M Costs	66
Figure 4-18. Activated Alumina (pH adjusted to 6.0 - 23,100 BV) Waste Disposal O&M
            Costs	67
Figure 4-19. Activated Alumina (pH adjusted to 6.0 - 15,400 BV) O&M Costs	67
Figure 4-20. Activated Alumina (pH adjusted to 6.0 - 15,400 BV) Waste Disposal O&M
            Costs	68
Figure 4-21. Greensand Capital Costs	69
Figure 4-22. Greensand O&M Costs	69
Figure 4-23. Greensand Waste Disposal Capital Costs	70
Figure 4-24. Greensand Waste Disposal O&M Costs	70
Figure 4-25. Coagulation Assisted Microfiltration Capital Costs	71
Figure 4-26. Coagulation Assisted Microfiltration O&M Costs	72
Figure 4-27. Coagulation Assisted Microfiltration (w/ Mechanical Dewatering) Waste Disposal
            Capital Costs	72
Figure 4-28. Coagulation Assisted Microfiltration (w/ Mechanical Dewatering) Waste Disposal
            O&M Costs	73
Figure 4-29. Coagulation Assisted Microfiltration (w/ Non-Mechanical Dewatering) Waste
            Disposal Capital Costs	73
Figure 4-30. Coagulation Assisted Microfiltration (w/ Non-Mechanical Dewatering) Waste
             Disposal O&M Costs	74
Figure 4-31. Coagulation/Filtration System Enhancement Capital Costs	75
Figure 4-32. Coagulation/Filtration System Enhancement O&M Costs	75
Figure 4-33. Lime Softening Enhancement Capital Costs	76
Figure 4-34. Lime Softening Enhancement O&M Costs	77
Figure 4-35. POU Reverse Osmosis Capital Costs	78

Arsenic Treatment Technology Evaluation Handbook for Small Systems                                xi

-------
Figure 4-36. POU Reverse Osmosis O&M Costs	78
Figure 4-37. POU Activated Alumina Capital Costs	79
Figure 4-38. POU Activated Alumina O&M Costs	80
Figure 5-1. Typical Liquid Hypochlorite Process Flow Diagram	83
Figure 5-2. Liquid Hypochlorite System Schematic (USFilter, Wallace & Tiernan)	83
Figure 5-3. Typical On-Site Hypochlorite Generation Process Flow Diagram	85
Figure 5-4. On-Site Hypochlorite Generation System Schematic (USFilter, Wallace &
           Tiernan)	85
Figure 5-5. On-Site Hypochlorite Generation System (Severn Trent Services)	86
Figure 5-6. Typical Permanganate Process Flow Diagram	87
Figure 5-7. Permanganate Dry Feed System (Merrick Industries, Inc.)	88
Figure 5-8. Permanganate Dry Feed System (Acrison, Inc.)	88
Figure 5-9. Typical Ozonation Process Flow Diagram	89
Figure 5-10. Ozone Generator and Contactor (ProMinent)	90
Figure 5-11. Typical  Solid Phase Oxidant Arsenic Oxidation Process Flow Diagram	90
Figure 5-12. Venturi  Air Injector Assembly Schematic (Mazzei)	91
Figure 6-1. Sorption  Treatment Process Flow Diagram w/o pH Adjustment and Regeneration. 95
Figure 6-2. Sorption  Treatment Process Flow Diagram w/ pH Adjustment and Regeneration. . 95
Figure 6-3. Sorption  Column Operation Modes	97
Figure 6-4. Multi-Component Ion Exchange	98
Figure 6-5. Activity of Nitrate and Nitrite During Ion Exchange	99
Figure 6-6. Ion Exchange System (Tonka Equipment Company)	 102
Figure 6-7. Process Flow Diagram for Example Problem	 104
Figure 7-1. Typical Media Filtration Process Flow Diagram	 107
Figure 7-2. Media Filtration Process Flow Modes	 108
Figure 7-3. Schematic of a Vertical Greensand Pressure Filter	 110
Figure 7-4. Hub-Lateral Distribution System (Johnson Screens)	 Ill
Figure 7-5. Header-Lateral Distribution System (Johnson Screens)	 Ill
Figure 7-6. Multiple  Media Filter Setup	 112
Figure 7-7. Pressurized Media Filter (USFilter)	 112
Figure 7-8. Pre-Engineered Arsenic Filtration System (Kinetico)	 113
Figure 7-9. Ferric Chloride  Addition Flow Diagram	 116
Figure 8-1. Point-Of-Use Adsorption Setup (Kinetico)	 118
Figure 8-2. Metered Automatic Cartridge (Kinetico)	 119
Figure 8-3. Point-Of-Use Reverse Osmosis Setup (Kinetico)	 120
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                xii

-------
                                                         List  of Tables
Table ES-l. Arsenic Treatment Technologies Summary Comparison	iv
Table 1-1.  Waste Disposal Options	6
Table 2-1.  Typical Treatment Efficiencies and Water Losses	16
Table 2-2.  Comparison of Oxidizing Agents	 19
Table 2-3.  Water Quality Interferences with AA Adsorption	28
Table 2-4.  Examples of Iron Based Sorbents	29
Table 3-1.  Key Water Quality Parameters to be Monitored	40
Table 3-2.  Other Water Quality Parameters to be Monitored	41
Table 3-3.  Arsenic Treatment Technologies Summary Comparison	53
Table 5-1.  Typical Filox-R™ Design and Operating Parameters	92
Table 5-2.  Comparison of Pre-Oxidation Alternatives	93
Table 6-1.  Typical Sorption Treatment Design and Operating Parameters	 101
Table 7-1.  Typical Greensand Column Design and Operating Parameters	 109
Arsenic Treatment Technology Evaluation Handbook for Small Systems                             xiii

-------
          List of Acronyms  and  Abbreviations
AA
Al
ANSI
As
As(III)
As(V)
AsO33
AsO;3
ASTM
AWWA
AwwaRF
BAT
BV
eft
Ca+2
CaCO3
CADF
CCI
CF
ci-
ci2
CMF
CO32
CT
CWA
DBF
DBPR
DO
EBCT
ENR
F
Fe,Fe+2,
FeCl3
Fe(OH)3
ft
FTW
g
gal
GFH
Fe+
Activated Alumina
Aluminum
American National Standards Institute
Arsenic
Valence +3 Arsenic (found in Arsenite ion, AsO3~3)
Valence +5 Arsenic (found in Arsenate ion, AsO4~3)
Arsenite ion
Arsenate ion
American Society for Testing and Materials
American Water Works Association
American Water Works Association Research Foundation
Best Available Technology
Bed Volume
Cubic feet
Calcium
Calcium Carbonate
Coagulation-Assisted Direct Filtration
Construction Cost Index
Enhanced (Conventional) Coagulation/Filtration
Chloride
Chlorine
C oagul ati on-Assi sted Mi crofiltrati on
Carbonate
Disinfectant Concentration Times Contact Time
Clean Water Act of 1987
Disinfection By-Product
Disinfectants/Disinfection By-Products Rule
Dissolved Oxygen
Empty Bed Contact Time
Engineering News Record
Fluoride
Iron
Ferric Chloride
Ferric Hydroxide
Feet
Filter-To-Waste
Gram
Gallon
Granular Ferric Hydroxide
Gallons per Day
Gallons per Hour
Arsenic Treatment Technology Evaluation Handbook for Small Systems


-------
gpm
hr
H+
H2AsO3
H2As04
H2O
H2S
H2SO4
H3AsO3
H3AsO4
HAAS
HAsO32
HAsO;2
HC1
HOC1
HS
ros
in.
IESWTR
IX
kg
kWh
L
LACSL
lb(s)
LCR
LS
LT IESWTR
MCL
MF
mg
Mg
MOD
min.
mL
mm
Mn, Mn+2
MnO2
MnO4
MSHA
MTZ
N
NaCl
NaOCl
NaOH
Gallons per Minute
Hour
Hydronium, Hydrogen ion
Monovalent Arsenite Ion
Monovalent Arsenate Ion
Water
Hydrogen Sulfide
Sulfuric Acid
Arsenite Molecule
Arsenate Molecule
Haloacetic Acid
Divalent Arsenite Ion
Divalent Arsenate Ion
Hydrochloric Acid
Hypochlorous Acid
Hydrogen Sulfide Ion
Iron Based Sorbents
Inches
Interim Enhanced Surface Water Treatment Rule
Ion Exchange
kilogram
Kilowatt Hour
Liter
Land Application Clean Sludge Limit
Pound(s)
Lead and Copper Rule
Lime Softening
Long Term-1 Enhanced Surface Water Treatment Rule
Maximum Contaminant Level
Micro-Filtration
Milligram
Magnesium
Million Gallons per Day
Minute
Milliliter
Millimeter
Manganese
Manganese Dioxide
Permanganate
Mine Safety and Health Administration
Mass Transfer Zone
Nitrogen
Sodium Chloride
Sodium Hypochlorite
Sodium Hydroxide
Arsenic Treatment Technology Evaluation Handbook for Small Systems


-------
NIOSH
NIPDWRs
NO2
NO3
NOM
NPDES
NPDWRs
NTNC
NTU
ฐ2
03
O&M
OC1
OH
PFLT
pH
P043
POE
POTW
POU
psi
RCRA
RO
Sฐ
SBA
scf
scfm
SDWA
sft
Si(OH)3O
S042
SSCT
SWTR
TBLL
TC
TCLP
IDS
TTHM
TOC
UFC
USEPA
UV
WET
wt%
y
National Institute for Occupational Safety and Health
National Interim Primary Drinking Water Regulations
Nitrite
Nitrate
Natural Organic Matter
National Pollutant Discharge Elimination System
National Primary Drinking Water Regulations
Non-Transient, Non-Community
Nephelometric Turbidity Units
Oxygen
Ozone
Operations and Maintenance
Hypochlorite
Hydroxide
Paint Filter Liquids Test
Negative Log of Hydrogen Ion Concentration
Phosphate, Orthophosphate
Point-of-Entry
Publicly Owned Treatment Works
Point-of-Use
Pounds per Square Inch
Resource Conservation and Recovery Act
Reverse Osmosis
Sulfur, zero valence
Strong Base Anion exchange resin
Standard Cubic Feet
Standard Cubic Feet per Minute
Safe Drinking Water Act
Square Feet
Silicate Ion
Sulfate
Small System Compliance Technologies
Surface Water Treatment Rule
Technically Based Local Limit
Toxicity Characteristic
Toxicity Characteristic Leaching Procedure
Total Dissolved Solids
Total Trihalomethanes
Total Organic Carbon
Uniform Fire Code
United States Environmental Protection Agency
Ultra-Violet
Waste Extraction Test
Weight Percent
Year
Arsenic Treatment Technology Evaluation Handbook for Small Systems


-------
                                  Equation  Nomenclature
Symbol
            Definition
                                                             Units
BV
   i
s~\

C
 As.B
C
  C12
C
 MCL
 RAA
 TDS
CCI
   Current
D
D


D
  ci2
  02
EBCT
F
GBW

HLR
H
h.
i
M .
  Brine
MCI,
            Number of Bed Volumes to Exhaustion
            Arsenic Concentration of Source j
            Arsenic Concentration of Blended Stream
            Chlorine Concentration
            Ferric Chloride Stock Solution Concentration
            Concentration of Species /'  in the Feed Stream
            Arsenic Concentration Entering System During
            Quarter k
            Arsenic MCL
            Permanganate Stock Solution Concentration
            Concentration of Species /'  in the Retentate
            Running Annual Average Arsenic Concentration
            Concentration of Total Dissolved Solids
            Construction Cost Index for the Current Year
            Construction Cost Index for 1998
            Column Diameter
            Ultimate Chlorine Demand
            Ultimate Permanganate Demand
            Ultimate Oxygen Demand
            Ultimate Ozone Demand
            Overall Rejection Rate
            Individual Stage Contaminant Rejection Rate
            Empty Bed Contact Time
            Freeboard Allowance
            Backwash Flux
            Regeneration Flux
            Hydraulic Loading Rate
            Column Height
            Height of Media Layery
            Annual Inflation Rate
            Brine Molarity
            Chlorine Mass Flow
                                                             mg/L
                                                             mg/L
                                                             Ibs Cl2/gal
                                                             wt%
                                                             mg/L
                                                             mg/L

                                                             mg/L
                                                             mg/L
                                                             mg/L
                                                             mg/L
                                                             g/L
ft
mg/L as C12
mg/L as Mn
mg/L
mg/L


minutes

gpm/sft
gpm/sft
gpm/sft
ft
ft

mole/L
lb/dayofC!2
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                                             xvi i

-------
Symbol
M03
n
np
1998
P
Current
Q
Qj
QB
QBW
Qa2
QFeCl3
n
^.MnO4
Qs
t
tfiw
FTW
TP
Jx
V
ww
WJ
Y
Current
Z
P
5C12
5FeCl3
ฐMnO4
งo2
งo3
e
PFeCl3
Pj
a
i
CO
Definition
Ozone Mass Flow
Number of Stages
Number of Parallel Treatment Trains
Year 1998 Cost
Current Cost
Design Flow Rate
Flowrate of Source j
Flowrate of Blended Stream
Backwash Flowrate
Hypochlorite Metering Pump Rate
Ferric Chloride Metering Pump Rate
Permanganate Metering Pump Rate
Flowrate to be Split off and Treated
Storage Time
Backwash Duration
Filter-To-Waste Duration
Regeneration Duration
Storage Volume
Volume of Wastewater
Weight of Mediay
Current Year
Depth of Media
Individual Stage Water Recovery Rate
Chlorine Dose
Ferric Chloride Dose
Permanganate Dose
Oxygen Dose
Ozone Dose
Arsenic Rejection Rate
Density of Ferric Chloride
Bulk Density of Mediay
Safety Margin
Optimal Filter Run Time
Treatment Water Loss
Units
g/hrofO3
-
-
$
$
gpm
gpm
gpm
gpm
gph
mL/min
gph
gpm
days
minutes
minutes
minutes
gal
gal
Ibs
-
ft
%
mg/L as C12
mg/L
mg/L as Mn
mg/L
mg/L
%
kg/L
Ibs/cft
%
hr
H
Arsenic Treatment Technology Evaluation Handbook for Small Systems
xvm

-------
                                                                  Section  1
                                                            Background
1.1    Purpose of this Handbook

This Handbook is intended to serve as a resource for small municipal drinking water systems that
may be affected by provisions of the Arsenic Rule. A "small" system is defined as a system serving
10,000 or fewer people.  Average water demand for these size systems is normally less than 1.4
million gallons per day (MGD). Please note that the USEPA statutes and regulations described in
this document contain legally binding requirements. The recommendations provided in this hand-
book do not substitute for those statutes or regulations, nor is this document a regulation itself. The
approaches described in this handbook  are strictly  voluntary  and do not impose legally-binding
requirements on USEPA,  states, local or tribal governments, or members of the public, and may not
apply to a particular situation based upon the circumstances. Although the USEPA strongly recom-
mends the approach outlined in this document,  state and local decision makers are free to adopt
approaches that differ from this handbook or to evaluate and choose technologies that are not dis-
cussed here.  Interested parties are free to raise questions and objections about the appropriateness
of the application of this  handbook.  Any USEPA decisions regarding a particular system will be
made based on the applicable statutes and regulations.  The USEPA may review and update this
handbook as necessary and appropriate.

1.2    How to Use this Handbook

This Handbook includes  general arsenic treatment information, cost evaluation tools, and design
considerations for specific treatment technologies. The Handbook is organized to enable the utility
to make educated decisions about the most appropriate treatment approach(es) to address arsenic
concerns  prior to getting  involved in detailed design considerations.  The utility should read Sec-
tions 1 through 3 in sequence, then use the treatment selection guidance provided in Section 3 to
determine the most appropriate cost and design considerations sections (Sections 4 through 8).

Section 1 provides background information on the  Arsenic Rule, waste disposal regulation, and
arsenic chemistry that is useful in understanding the remainder of the Handbook.

Section 2 provides descriptions and background information for established arsenic mitigation strat-
egies, with emphasis on those that are most technically and financially suitable for small systems.
The utility should use this section to gather background information on the various arsenic mitiga-
tion strategies and determine the flowrate to be treated if treatment is selected as the mitigation
strategy.

Section 3 describes the considerations required to make an informed treatment method selection.
Decision  trees are provided to guide the utility to the most applicable mitigation or treatment strat-
Arsenic Treatment Technology Evaluation Handbook for Small Systems

-------
egy. The selected process is the one that has the highest chance of achieving the most cost effective
solution for the particular water source, given the parameters used in the decision making process.

Section 4 enables the utility to quickly estimate the planning-level costs for the selected treatment
process.  This section is intended for those utilities that have identified the need to install new
arsenic treatment.  Based on the cost estimate, the utility can then decide if the selected treatment
process is economically feasible.  If it is not, the utility can repeat the decision trees and apply
different preferences.  It is important to recognize that the cost curves provided are for planning-
level considerations only and should not be used as the primary decision-making tools.

Section 5 presents pre-oxidation alternatives and design calculations. This section is relevant to
those utilities that have selected a treatment alternative and do not currently employ oxidation at the
source(s) with arsenic concerns.

Sections 6-8  are intended for those utilities that have identified the need to install new arsenic
treatment technologies.  These sections provide design information on each of the primary arsenic
treatment technologies.

After the selected mitigation strategy has been reviewed, the utility should evaluate the cost and
constraints of the mitigation strategy.  For strategies that involve modification of an existing pro-
cess, a test should be run. For strategies that involve a new process,  a pilot plant test should be run.
After the tests have been performed and the results analyzed, the utility should re-evaluate whether
the strategy will reduce the arsenic concentration below the Maximum Contaminant Level (MCL).

1.3   Regulatory Direction

       1.3.1   The Arsenic Rule
       The former arsenic MCL was 0.05 mg/L, as established under the 1975 National Interim
       Primary Drinking Water Regulations (NIPDWRs). As part of the 1996 Safe Drinking Wa-
       ter Act (SDWA) Amendments, the United States Environmental Protection Agency (USEPA)
       was directed to conduct health effects and cost/benefit research to finalize a new arsenic
       standard.

       In June 2000, the USEPA proposed a revised arsenic MCL of 0.005 mg/L, and requested
       public comment on alternative MCLs of 0.003, 0.010, and 0.020 mg/L.  The USEPA pub-
       lished a final rule in the Federal Register in January 2001 (USEPA, 2001). This rule estab-
       lished a revised arsenic MCL of 0.010 mg/L, and identified the following as Best Available
       Technologies (BATs) for achieving compliance with this regulatory level:

       •   Ion Exchange (IX)
       •   Activated Alumina (AA)
       •   Oxidation/Filtration
       •   Reverse Osmosis (RO)
Arsenic Treatment Technology Evaluation Handbook for Small Systems

-------
       •  Electrodialysis Reversal
       •  Enhanced Coagulation/Filtration
       •  Enhanced Lime Softening

       The following are listed in the final rule and the Implementation Guidance for the Arsenic
       Rule (USEPA, 2002a) as Small System Compliance Technologies (SSCT).

       •  IX
       •  AA, centralized and point-of-use (POU)
       •  RO, centralized and POU
       •  Electrodialysis Reversal
       •  Oxidation/Filtration
       •  Coagulation/Filtration, Enhanced Coagulation/Filtration, and Coagulation-Assisted
          Microfiltration (CMF)
       •  Lime Softening (LS) and Enhanced Lime Softening

       This regulation applies to all community water systems and non-transient, non-community
       (NTNC) water systems, regardless of size.  Please note that systems are not required to use
       these technologies.

       Compliance with the Arsenic Rule will be required by January 2006. The running annual
       average arsenic level must be at or below 0.010 mg/L at each entry point to the distribution
       system. However, POU treatment can be instituted instead of centralized treatment. Ana-
       lytical results for arsenic are rounded to the nearest 0.001 mg/L for reporting and compli-
       ance determination.

       1.3.2   Health Effects
       Motivation to reduce the arsenic MCL is driven by the findings of health effects research.
       Over the past several years, numerous toxicological and epidemiological studies have been
       conducted to ascertain health risks associated with low-level exposure to As(V) ingestion.

       Ingestion of inorganic arsenic can result in both cancer and non-cancer health effects (NRC,
       1999).  Arsenic interferes with a number of essential physiological activities, including the
       actions of enzymes, essential cations, and transcriptional events in cells (NRC, 1999). The
       USEPA has classified arsenic as a Class A human carcinogen.  Chronic exposure to low
       arsenic levels (less than 0.05 mg/L) has been linked to health complications, including
       cancer of the  skin, kidney, lung, and bladder, as well as other diseases of the skin, neurologi-
       cal, and cardiovascular system (USEPA, 2000).

       The primary mode of exposure is ingestion of water containing arsenic. Dermal absorption
       of arsenic is  minimal; therefore, hand washing and bathing do not pose a known risk to
       human health.
Arsenic Treatment Technology Evaluation Handbook for Small Systems

-------
       1.3.3  Other Drinking Water Regulations
       In an attempt to comply with one drinking water regulation, it is possible to compromise
       treatment performance or compliance with other drinking water regulations. Therefore, in
       an effort to conform with the Arsenic Rule, community water systems should be cognizant
       of potential system-wide, regulatory, and operational impacts.  In particular, compliance
       with the following regulations should be considered.

       •  Lead and Copper Rule (LCR)
       •  Surface Water Rules (SWTR, IESWTR, LT1ESWTR)
       •  Disinfectants/Disinfection By-Products Rule (DBPR)

       Many of the  arsenic treatment technologies require pH adjustment for optimization of
       performance.  Figure 11 provides a summary of the optimal pH ranges for several arsenic
       treatment technologies.  Sorption and coagulation processes are particularly sensitive to
       pH, and function most effectively at the lower end of the natural pH range. However, use of
       AA at a natural pH may be a cost effective option for many small water systems.
                                                               Enhanced Lime Softening
1

1

1

1

1

1 1

1 I Enhance
i i i i
5678
	 | Reverse Osmosis

	 | Anion Exchange

Iron Based Sorbents

| Oxidation Filtration

| Conventional Activated Alumina
Enhanced Iron Coagulation
d Aluminum Coagulation
i i i
9 10 11
                                pH
             Figure 1-1.  Optimal pH Ranges for Arsenic Treatment Technologies.

       In addition to affecting arsenic treatability, pH also can have a significant effect on disinfec-
       tion, coagulation, and chemical solubility/precipitation within the distribution system and
       in plumbing systems (LCR).
Arsenic Treatment Technology Evaluation Handbook for Small Systems

-------
       Lead and Copper Rule
       Lead and copper in tap water is primarily due to corrosion of plumbing system components
       within buildings, including copper pipes, lead-based solder used to join segments of copper
       pipe, and faucets made from brass.  Alkalinity and pH play a critical role in providing
       passivation protection from corrosion. In general, the optimum pH range for minimizing
       corrosion of lead and copper is 7.59.0. Therefore, post-treatment pH adjustment is recom-
       mended for many of the treatment techniques provided in Figure 11.

       Surface Water Rules (SWTR, IESTWR, LT1ESWTR)
       Disinfection efficacy is also related to pH if chlorine is used. When pre-chlorinating, the
       biocidal potential of chlorine is enhanced as the pH is reduced.  Therefore, utilities that are
       currently required to meet a CT (disinfectant concentration times contact time) standard
       should receive disinfection benefit from pH reduction at the head of the treatment process.
       Post-treatment pH adjustment for corrosion control should be conducted after CT require-
       ments are met.

       Coagulation and flocculation processes  are also related to pH.  The formation of floe is
       improved as the pH is reduced, and optimized within the range of 5-8. However, iron and
       aluminum-based coagulants also consume  alkalinity, thereby decreasing the buffering ca-
       pacity of the water.

       Disinfectants/Disinfection By-Products Rule (DBPR)
       Chlorine reacts with natural organic matter (NOM) to form halogenated disinfection by-
       products, such as total trihalomethanes (TTHM) and haloacetic acids (HAAS).  Therefore,
       incorporating pre-chlorination to convert As(III) to As(V) could increase the occurrence of
       these regulated chemicals. However, as most arsenic in surface water is already oxidized to
       As(V), chlorination may not be necessary in surface water, where disinfection byproducts
       are of the most concern. The Stage 1 DBPR1 establishes running annual average MCLs of
       0.080 mg/L and 0.060 mg/L for TTHM and HAAS, respectively.

       The Stage 2 DBPR, scheduled for proposal in early 2003, augments the Stage 1 DBPR to
       reduce health risks from DBF exposure.

       1.3.4  Waste Disposal Regulations
       Waste disposal is an important consideration in the treatment selection process.  Arsenic
       removal technologies produce several different types of waste, including sludges,  brine
       streams, backwash slurries, and  spent media.  These  wastes have the potential for being
       classified as hazardous  and  can pose disposal problems.  Table  1-1 provides a summary of
       the available waste disposal options and associated criteria. These are further discussed in
       the following paragraphs. In addition, specific waste disposal considerations for each tech-
       nology are discussed in Section 2.
1 The Stage 1 DBPR became effective for surface water systems and groundwater systems under the direct influence
of surface water serving at least 10,000 people in January 2002. The rule will take effect for all groundwater
systems, and surface water systems and groundwater systems under the direct influence of surface water systems
serving less than 10,000 people in January 2004.

Arsenic Treatment Technology Evaluation Handbook for Small Systems

-------
Table 1-1. Waste Disposal Options.
Waste Type
Liquid
Solid (sludge, media)
Disposal Method
• Direct Discharge to Surface Water (CWA,
• Indirect Discharge to POTW (TBLLs)
• On- Site Sewerage (POU systems)
• Infiltration to Ground Water
• Evaporation Ponds
• Recycle/Reuse
• Ocean Discharge
(Criteria)
NPDES)
• Land Application
• Municpal Solid Waste Landfill (PFLT, TCLP, WET in California)
• Hazardous Waste Landfill (PFLT)
       Liquid waste streams must have lower concentrations than the Toxicity Characteristic (TC)
       in order for the waste to be classified as non-hazardous. The arsenic TC is 5.0mg/L. Those
       liquid waste streams that contain more than 5.0 mg/L of arsenic would therefore be classi-
       fied as a hazardous waste.  Many of the arsenic removal technologies also remove other
       constituents (e.g., chromium). The waste stream must be analyzed for these other sub-
       stances that may be in concentrations above their respective TCs.  Because of Resource
       Conservation and Recovery Act (RCRA) requirements and cost implications, on-site treat-
       ment or off-site disposal of hazardous waste is likely to be infeasible for small water sys-
       tems.  Indirect discharge may be an option since wastes that pass through a sewer system to
       a Publicly Owned Treatment Works (POTW) are exempt from RCRA regulation once the
       waste mixes with wastewater from the sewer.  Utilities considering indirect discharge should
       work with the POTW to determine if the arsenic waste levels would be acceptable to a
       revised Technically Based Local Limit (TBLL). (The TBLL would be revised because the
       arsenic treatment will change the arsenic background at the POTW).

       Solids waste streams are subj ect to the Toxicity  Characteristic Leaching Procedure (TCLP).
       This test  is used to simulate the potential for leaching in a landfill setting.  The TCLP
       leachate must be lower than any of the TC values in order for the waste to be classified as
       non-hazardous.

       There are five realistic methods for the disposal of arsenic waste streams.

       Landfill Disposal
       Historically, municipal solid waste landfills have been commonly used for the disposal of
       non-hazardous solid wastes emanating from treatment processes. However, the hazard po-
       tential of arsenic may limit the feasibility of this alternative.

       Dewatered sludge and spent media can be disposed in a municipal solid waste landfill if the
       waste passes both the Paint Filter Liquids Test  (PFLT) and the TCLP. The PFLT is used to
       verify there is no free liquid residual associated with the waste.  However, if the TCLP
       extract contains arsenic or any other contaminant (e.g., chromium) above the TC, the waste
Arsenic Treatment Technology Evaluation Handbook for Small Systems

-------
       residuals must be disposed in a designated and licensed hazardous waste landfill.  These
       landfills are strictly regulated under RCRA and have extensive monitoring and operational
       guidelines. As such, the costs of disposal are relatively high.  As with municipal solid waste
       landfill disposal, waste sludges must not contain free liquid residuals.

       A critical element of hazardous waste disposal  is the cradle-to-grave concept.  The party
       responsible for generating the hazardous waste retains liability and responsibility for the
       fate and transport of the waste.

       Direct Discharge to Surface Waters
       Direct discharge refers to the disposal of liquid waste streams to nearby surface waters,
       which act to dilute and disperse the waste by-products.  The primary advantage of direct
       discharge is reduced capital and operations and maintenance (O&M) costs due to the elimi-
       nation of residuals treatment.  The feasibility of this disposal method is subject to provi-
       sions of the National Pollution Discharge Elimination System (NPDES) and state anti-deg-
       radation regulation.  The allowable discharge is a function  of the ability of the receiving
       water to assimilate the arsenic without exceeding water  quality criteria established under
       the Clean Water Act (CWA) or state regulation. Different water quality criteria exist de-
       pending on the classification of the receiving water. For specific NPDES conditions and
       limits, the appropriate NPDES permitting agency should be contacted, because the condi-
       tions and limits can vary according to the receiving stream's particular characteristics.

       Indirect Discharge
       The discharge of liquid waste streams to a POTW is a potential disposal alternative.  In this
       case, the waste stream will be subject to TBLLs established regionally by sewer authorities
       as part of the POTW's Industrial Pretreatment Program. TBLLs are established in order to
       protect POTW operation,  assure compliance with NPDES permits, and prevent an unac-
       ceptable level of accumulation of contaminants in the process sludge and biosolids.  The
       arsenic limit is usually on the order of 0.05 to 0.1 mg/L. The TBLLs are computed for each
       POTW to take into account the background levels of contaminants in the municipal waste-
       water. The background level will change because of the drinking water treatment process,
       which may lead to revised TBLLs.  The revised TBLL would be used to determine if the
       liquid waste stream could  be discharged to the POTW.

       Land Application
       Land application of concentrated sludge may be allowed under certain conditions depend-
       ing on the state law and regulations.  Some states  do  not allow land application of solid
       residuals. Sewage sludge  (also called "biosolids") containing <41 mg As/kg biosolids can
       be land-applied with no restrictions. Biosolids with arsenic concentrations between 41 and
       75 mg/kg can be land-applied, but must track arsenic accumulation.  The lifetime arsenic
       accumulation limit is 41 kg As per hectare of land. Federal part 503 land application limits
       are only applicable in states that have adopted these limits for water plant residuals or in
       cases where these residuals are mixed with sewage  sludge.
Arsenic Treatment Technology Evaluation Handbook for Small Systems

-------
       On-Site Sewerage
       Liquid waste streams from RO POU devices should be suitable for disposal in an on-site
       sewerage or septic system.  Figure 1-2 illustrates a typical flow diagram for RO POU water
       treatment and on-site waste disposal.
                                         From Water
                                           Supply
                                                    POU Feed
                         Non -Consumptive
                             Water
Treated Water for
Consumption
                                          To Septic
                                         System and ^	
                                          Drainfield
                          Figure 1-2.  Flow Diagram for RO POU.

       Arsenic is concentrated in the RO retentate during normal process operation.  However,
       eventually this retentate is combined with other domestic wastewater in the septic tank.
       Because the amount of water consumed is small relative to the total flow entering the dwell-
       ing, the concentration of arsenic in the blended wastewater is nearly identical to that in the
       influent stream.

1.4   Arsenic  Chemistry

Arsenic is introduced into the aquatic environment from both natural and manmade sources. Typi-
cally, however, arsenic occurrence in water is caused by the weathering and dissolution of arsenic-
bearing rocks, minerals, and ores. Although arsenic exists in both organic and inorganic forms, the
inorganic forms are more prevalent in water and are considered more toxic. Therefore, the focus of
this Handbook is on inorganic arsenic.

Total inorganic arsenic is the sum of particulate and soluble arsenic.  A 0.45-micron filter can
generally remove particulate arsenic.

Soluble, inorganic arsenic exists in either one of two valence states depending on local  oxidation-
reduction conditions.  Typically groundwater has anoxic conditions and arsenic is found in its ars-
enite or reduced trivalent form  [As(III)].  Surface water generally has aerobic conditions and ar-
senic is found in its arsenate or  oxidized pentavalent form [As(V)].

Both arsenite and arsenate exist in four different species. The speciation of these molecules changes
by dissociation and is pH dependent. The kinetics of dissociation for each are nearly instantaneous.
The pH dependencies of arsenite and arsenate are depicted in Figure 1-3 and Figure 1-4, respec-
Arsenic Treatment Technology Evaluation Handbook for Small Systems

-------
lively.  The species shown in bold are those that are most likely to be removed by the techniques
discussed in this handbook.
         100%
                0
2
10
           468

                        pH

Figure 1-3. Dissociation of Arsenite [As(III)].
12
14
        ,100%
                0
                                                pH

                        Figure 1-4. Dissociation of Arsenate [As(V)].

Chemical speciation is a critical element of arsenic treatability.  Negative surface charges facilitate
removal by adsorption, anion exchange, and co-precipitative processes.  Since the net charge of
arsenite [As(III)] is neutral at natural pH levels (6-9), this form is not easily removed. However, the
net molecular charge of arsenate [As(V)] is negative (-1 or -2) at natural pH levels, enabling it to be
removed with greater efficiency. Conversion to As(V) is a critical element of most arsenic treat-
ment processes.  This  conversion can be accomplished by adding an oxidizing agent such as chlo-
rine or permanganate. Selection of the most appropriate oxidation technology should be based on
several considerations, including cost, integration with existing treatment, disinfection require-
ments, and secondary  effects. This is discussed further in Section 2.2.
Arsenic Treatment Technology Evaluation Handbook for Small Systems

-------
Preceeding Page Blank
                                                                        Section  2

      	Arsenic  Mitigation  Strategies


      2.1   Description of Arsenic Mitigation Strategies

      Problematic arsenic levels in drinking water can be mitigated in several different ways. This Hand-
      book will address the following mitigation strategies:

      •  Abandonment - The total abandonment of the problematic source(s) and subsequent switch to
          other source(s) within the system or purchase from a neighboring system.
      •  Seasonal Use - Switching the problematic source(s) from full-time used to seasonal or peaking
          use only with subsequent blending with other full-time source(s).
      •  Blending - The combination of multiple water sources to produce a stream with an arsenic
          concentration below the MCL.
      •  Sidestream Treatment - The treatment of a portion of the high arsenic water stream and subse-
          quent blending back with the untreated portion of the stream to produce water that meets the
          MCL.
      •  Treatment - The processing of all or part of a water stream to reduce the arsenic concentration
          to below the MCL.
          O  Wellhead Treatment - Treatment is located at the wellhead location before the water is
             mixed with water from  other sources.
          O  Centralized Treatment - Water from several sources is piped to a centralized location for
             treatment before the water enters the distribution system.
          O  POU Treatment - Treatment devices  are located at the Point-Of-Use within the building or
             home and treat only the water intended for direct consumption, typically at a single tap.

      There are three primary categories of available treatment processes.
      •  Sorption Treatment Processes
          O  IX2
          O  AA 2-3
          O  Iron Based Sorbents (IBS)4
      •  Membrane Treatment Processes
          O  RO23
          O  Precipitation/Filtration  Processes
       2 Technologies that have been designated as small system compliance technologies (SSCT) for centralized or
       wellhead treatment.
       3 Technologies that have been designated as SSCT for POU treatment.
       4 Due to limited performance research at the time the rule was promulgated, IBS was not designated as a BAT or a
       SSCTbytheUSEPA.
       Arsenic Treatment Technology Evaluation Handbook for Small Systems                               11

-------
    O  Enhanced Conventional Gravity Coagulation/Filtration 2above
    O  CMF2above
    O  Coagulation- Assisted Direct Filtration (CADF) 2 above
    O  Oxidation/Filtration2above
    O  Enhanced Lime Softening 2above

The selection of the most appropriate mitigation strategy  should be based on feasibility issues,
system constraints, and costs.

       2.1.1  Abandonment
       Perhaps the simplest approach for remedying a high arsenic source is abandonment of the
       high arsenic water source and procurement of a new source that meets the arsenic MCL.
       This option is most realistic for utilities with multiple water sources where at least one
       source can be relied upon for producing water with arsenic below the MCL.  There may,
       however, be other constraints to switching primary sources, such as inadequate treatment
       capacity or water rights.  Many  small systems do not have the flexibility to switch to a
       source with a lower arsenic concentration. In this particular case, the utility has two op-
       tions: (1) locate or install a new source, or (2) purchase water from a nearby system if an
       interconnection exists.  New source installations may or may not be more costly than treat-
       ment. It should also be noted that drilling a new  source may not be the best option if the
       aquifer has consistently high levels of arsenic.  The utility should check with other systems
       in the area before drilling.

       2.1.2  Seasonal Use
       Another option is to  switch a high arsenic water source from full-time production to sea-
       sonal or peaking use only. When used, it would be blended with low arsenic water sources
       before entry to the distribution system.  This is allowed at the federal level, as long as the
       running annual average at the entry point to the distribution system does not exceed the
       MCL. Individual state requirements may preclude this option.

       The running  annual average can be calculated by adding the four most  recent quarterly
       arsenic concentrations together and dividing by 4 as seen  in the following equation.
                                                                                 Eqn.2-1
       Where:
           CRAA  = Running Annual Average Arsenic Concentration Entering System,
           Cj     = Arsenic Concentration Entering System During the Quarter 1,
           C2     = Arsenic Concentration Entering System During the Quarter 2,
           C3     = Arsenic Concentration Entering System During the Quarter 3, and
           C     = Arsenic Concentration Entering System During the Quarter 4.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                 12

-------
       An example of this is shown in Figure 2-1.  Well 1 is the only source for the first and second
       quarters of year and has an arsenic concentration of 0.003 mg/L. Well 2 is used in conjunc-
       tion with Well 1 for the third and fourth quarters of the year and the combined output from
       Wells 1 and 2 has an arsenic concentration of 0.014 mg/L. The running annual average is
       calculated below as 0.009 mg/L and complies with the federal Arsenic Rule.
                        0.003 + 0.003 + 0.014 + 0.014
                   nri^
                = 0.0085 mg/L
                                                                  nno
                                                                 0.009 mg/L
         Quarter 1
           Flow To
         Distribution
           System
Quarter 2
 Flow To
Distribution
  System
Quarter 3
 Flow To
Distribution
  System
Quarter 4
 Flow To
Distribution
  System
                 Figure 2-1. Example of Seasonal High Arsenic Source Use.

       2.1.3  Blending
       The revised arsenic MCL must be met at all entry points to the distribution system. There-
       fore, blending is a viable mitigation strategy for conservative inorganic substances and should
       be considered by systems that utilize multiple sources. Blending involves mixing waters
       from two or more different sources prior to entering the distribution system. The purpose of
       blending is to eliminate the need for treatment.

       Stand-alone blending shown in Figure 2-2 is only a consideration when a water system has
       multiple sources, one (or more) with arsenic levels above the MCL, and one (or more) with
       arsenic levels below the MCL. Also, the wells with low arsenic levels must be reliable on
       a continuous basis.

       Each stream in the blending process should have a flow measurement to insure that the
       streams are blended in a ratio that produces an arsenic concentration that meets the MCL
       requirement. Flow measurement is shown in the  following figure and an "F" inscribed in a
       circle.
Arsenic Treatment Technology Evaluation Handbook for Small Systems


-------
                                        Total Flow
                                    To Distribution System

                                  Figure 2-2. Blending.

       The concentration of the blended stream (CAsB) can be calculated using the following
       formula.
                                      Ql "
                                'As,B
                                           Qi + Q2
                                                                    Eqn. 2-2
       Where:
          C
          C
As,B
          C
As,l

As,2
          Q,
= Arsenic Concentration of Blended Stream (mg/L),
= Arsenic Concentration of Well 1 (mg/L),
= Arsenic Concentration of Well 2 (mg/L),
= Flowrate of Well 1 (gpm), and
= Flowrate of Well 2 (gpm).
       For example, suppose that water from Wells 1 and 2 in Figure 2-2 contain arsenic concen-
       trations of 0.015 mg/L and 0.006 mg/L, respectively, with flowrates of 700 gpm and 2,450
       gpm, respectively.  The blended stream's arsenic concentration will then be 0.008 mg/L,
       which meets the MCL, and is calculated as follows:
                  -As,B
             700 gpm-0.015 mg/L+2,450 gpm-0.006 mg/L
                       700 gpm + 2,450 gpm
                                                               = 0.008 mg/L
       The following equation can be used to determine the required flowrate from the low arsenic
       source (Q2) that, when blended with flow from the high arsenic source (Ql), will produce
       water with an arsenic concentration a safe margin below the MCL.
       Where:
          Qi
          Q2
          c
          c
          c
As,l

            MCL
                               Q2=Qi
                                              "MCL ~ CAM
                                                    'MCL
                                                                    Eqn. 2-3
  Flowrate of Well 1 (gpm) (high arsenic source},
  Flowrate of Well 2 (gpm) (low arsenic source},
  Arsenic Concentration of Well 1 (mg/L) (high arsenic source},
  Arsenic Concentration of Well 2 (mg/L) (low arsenic source},
  Arsenic MCL (mg/L), and
  Safety Margin (% expressed as a decimal).
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                                          14

-------
       For example, suppose that water from Wells 1 and 2 in Figure 2-2 contain arsenic concen-
       trations of 0.015 mg/L and 0.006 mg/L, respectively. Assuming that the utility wants to
       provide a 20% safety margin (i.e., produce water with 0.008 mg/L of arsenic) and the maxi-
       mum flowrate from Well 1 is 700 gpm, the minimum required flowrate from Well 2 is 2,450
       gpm and is calculated as follows:

                                 ([l - 0.2]- 0.010 mg/L - 0.015 mg/L ")
                     Q2=700gpm-	r & -,	— =2,450 gpm
                                 I 0.006 mg/L-[l-0.2]-0.010 mg/L I

       2.1.4 Treatment
       If a treatment method is used to mitigate the arsenic problem from multiple sources, the
       utility will need to decide between wellhead treatment and centralized treatment. Wellhead
       treatment treats the water from each well at or near the wellhead.  For systems with multiple
       wells, this could result in multiple treatment facilities.  If possible, piping high arsenic water
       from multiple sources to a single, centralized treatment facility may be more economical.
       Some factors to take into account would be:

       •  the proximity of wells to be treated to each other,
       •  feasibility of piping the sources to a central location,
       •  availability of land and power at the treatment site(s), and
       •  labor requirements for multiple sites rather than a single site.

       Some treatment processes (e.g., RO) may have significant  water  losses associated with
       them. Water loss is incoming water that does not exit the system as treated water. Water
       losses frequently occur as a stream used to dispose of waste. Typical treatment efficiencies
       and water losses for processes operated under normal conditions are provided in Table 2-1.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                15

-------
                    Table 2-1.  Typical Treatment Efficiencies and Water Losses.
Treatment
Sorption Processes
Ion Exchange
Activated Alumina (Throw- Away Media)
Iron Based Sorbents
Iron and Manganese Removal Processes
Oxidation/Filtration (Greensand)
Membrane Processes
Reverse Osmosis
Precipitative Processes
Coagulation Assisted Microfiltration
Enhanced Coagulation/Filtration
With Alum

With Ferric Chloride
Enhanced Lime Softening
As(V) Removal Efficiency

95% '
95% !
Up to 98% '

50-90% 2

>95% '

90% '

<90% '

95% '
90% '
Water Loss

1-2%
1-2%
1-2%

<2%

15-50% '

5%


1-2%

1-2%
        1 USEPA, 2000.
        2 Depends on arsenic and iron concentrations

       2.1.5   Sidestream Treatment
       The treatment and blending strategies can be combined in a variety of ways as illustrated in
       Figure 2-3 through Figure 2-5.
               Total Flow
          To Distribution System
                                                      Treatment
                                              Blending
        Total Flow
    To Distribution System
        Total Flow
   To Distribution System
         Figure 2-3.  Sidestream
               Treatment.
Figure 2-4.  Treatment and
        Blending.
 Figure 2-5. Sidestream
Treatment and Blending.
       Sidestream treatment, Figure 2-3, involves splitting the source flow, treating one stream,
       and then blending it with the untreated stream prior to distribution. Sidestream treatment is
       feasible when a water source exceeds the revised MCL by a relatively small margin.  This
       approach is viable because most arsenic treatment processes (operated under optimal con-
       ditions) can achieve at least 80% arsenic removal and, in many cases, this high level of
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                      16

-------
       performance is not needed to meet the MCL.  Any utility considering treatment should
       consider sidestream treatment as a method to reduce the overall level and cost of the treat-
       ment required.

       In the simple case of sidestream treatment from a single source, as in Figure 2-6, the flowrate
       of the split stream requiring treatment can be calculated using Equation 2-4. If the treat-
       ment process is RO, as in Figure 2-7, the flowrate of the split stream requiring treatment can
       be calculated using Equation 2-5. RO requires a more complex equation because RO has a
       continuous stream of water lost during operation and this water is not available for blend-
       ing. In all the other treatments, when the water loss occurs (e.g., the backwash of a column
       or filter) no treated water is available for blending  so no blending occurs.  (Under these
       circumstances, the untreated water flow to the distribution is also halted.) For all the treat-
       ment methods, the resulting blended treated flowrate can be calculated using Equation 2-6.
                               Qs=Qi
                            CI-(I-O)-CMCL
                                 eC,
                                                                  Eqn. 2-4
                        n  =n
                              !l
                                        Ci-(l-o)-CMCL
                                                                     Eqn. 2-5
                                      QB=Qi-coQs
                                                                     Eqn. 2-6
       The variable for the equations are:
          Qs
          QR
          c
          a
          CO
          e
MCL
     = Flowrate to Split Off for Treatment (gpm),
     = Flowrate of the Final Blended Stream (gpm),
     = Source 1 Flowrate (gpm),
     = Arsenic Concentration of the Source (mg/L),
     = Arsenic MCL (mg/L),
     = Safety Margin (% expressed as a decimal),
     = Treatment Water Loss (% expressed as a decimal), and
     = Arsenic Rejection Rate (% expressed as a decimal).

           Source
           Water
Treatment
 By-Pass
                                                   Treatment
                                                    By-Pass
                                                                Treatment
                                                                Water Loss
                          Blended
                        Treated Water
                                                  Blended
                                                Treated Water
           Figure 2-6.  Sidestream Treatment.
                                      Figure 2-7.  Sidestream Treatment for RO.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                                            17

-------
       For example, suppose a water utility operates a single well at a maximum flowrate (Qj) of
       500 gpm.  The well water contains (C:) 0.012 mg/L of arsenic. Further, assume that utility
       wants to provide a 20% safety margin (o) on the arsenic MCL (CMCL) of 0.010 mg/L (i.e.,
       produce water with 0.008 mg/L of arsenic).  The utility has selected a RO process that has
       demonstrated an arsenic removal efficiency (e) of 95% at a water loss (co) of 40%. Using
       Equation 2-5, 237 gpm of the well's flow (or approximately 47%) should be split and sent
       to the RO treatment unit.  The final (blended) flowrate  is 405 gpm, as determined using
       Equation 26.

               /       *>(          0.012 mg/L-[1-0.2]-0.010 mg/L         ")
           Qs = (500  gpm)- 	f—-.	^-—	^	f——i	 = 237 gpm
                       ' ^0.012mg/L [l-(l-0.4Xl-0.95)]-0.4-[l-0.2]-0.010 mg/L J

                            QB =(500 gpm)- 0.4 -(237 gpm)=405  gpm

2.2   Pre-Oxidation Processes

Reduced inorganic As(III) (arsenite) should be converted to As(V) (arsenate) to facilitate removal.
This step is critical for achieving optimal performance of all unit processes described in this
Handbook. Conversion to As(V) can be accomplished by providing an oxidizing agent at the head
of any proposed arsenic removal process.  Chlorine, permanganate, ozone, and  Filox-R™5 are
highly effective for this purpose.  Chlorine dioxide and monochloramine are ineffective in oxidiz-
ing As(III).  Ultraviolet (UV) light, by itself, is also ineffective. However, if the water is spiked
with sulfite, UV photo-oxidation shows promise for As(III) conversion (Ghurye and Clifford, 2001).
Based on these considerations, only chlorine, permanganate, ozone, and Filox-R™ are discussed
further in this Handbook.

Table 2-2 provides a summary of the benefits  and drawbacks associated with the use of several
oxidation technologies.  The choice of oxidation method should be based primarily on the arsenic
treatment technology to be employed (as described in Section 3),  and secondarily on factors pro-
vided in Table 2-2. Many small water systems employ chlorine disinfection, either alone or as part
of a larger treatment process. In most of these  instances, the existing chlorination process can be
optimized to provide concurrent As(III) oxidation.
5 Filox is a registered trademark of Matt-Son, Inc., Harrington, IL. Filox-R is a trademark of Matt-Son, Inc.,
 Harrington, IL.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                 18

-------
                          Table 2-2.  Comparison of Oxidizing Agents.
   Oxidant
                  Benefits
                   Drawbacks
Chlorine
    Low relative cost ($0.50/b.)
    Primary disinfection capability
    Secondary disinfectant residual
    MnO2 media regenerant
    Oxidizes arsenic in less than 1 minute
•   Formation of disinfection by-products
•   Membrane fouling
•   Special handling and storage requirements
Permanganate
    Unreactive with membranes
    No formation of disinfection by-products
    MnO2 media regenerant?
    Oxidizes arsenic in less than 1 minute
    High relative cost ($1.35/lb.)
    No primary disinfection capability
    Formation of MnO2 particulates
    Pink Water
    Difficult to handle
    An additional oxidant may be required for
    secondary disinfection
Ozone
    No chemical storage or handling required
    Primary disinfection capability
    No chemical by-products left in water
    Oxidizes arsenic in less than 1 minute in
    the absence of interfering reductants
    Suffide and TOC interfere with conversion and
    increase the required contact time and ozone
    dose for oxidation
    An additional oxidant may be required for
    secondary disinfection
Solid Phase
Oxidants
(FJoxR™)
•   No chemical storage or handling required   •
•   No chemical by-products left in water      •
•   Oxidizes arsenic with an EBCT of 1.5      •
    minutes in the absence of interfering        •
    reductants                             •
    Backwashing required
    Backwash waste is generated
    Requires dissolved oxygen to work
    No primary disinfection capability
    An additional oxidant may be required for
    secondary disinfection
    Iron, manganese, sulfide, and TOC increase the
    contact time and dissolved oxygen
    concentration required for oxidation
       2.2.1  Chlorine
       Issues associated with pre-chlorination are: (1) sensitivity of the treatment process to chlo-
       rine; (2) disinfection by-product (DBF) formation potential; (3) code requirements associ-
       ated with chemical storage and handling;  and (4) operator safety.  Chlorine can be added
       either as a gas or as liquid hypochlorite, although chlorine gas may not be appropriate for
       small systems.  For new chlorine feed installations, these alternatives should be evaluated
       with respect to capital and operating costs, O&M requirements, code restrictions, contain-
       ment requirements, footprint, and safety concerns, among other issues.  Gas feed is typi-
       cally conducted with either 150-pound cylinders or 2,000-pound (1-ton) containers, de-
       pending on  the rate of chlorine consumption.  Small  systems normally use the  150-pound
       cylinders. Liquid hypochlorite can either be generated on-site (0.8% strength), or purchased
       as commercial strength  (5!/4 or 12V2%) liquid hypochlorite.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                                                    19

-------
       The stoichiometric oxidant demand is 0.95 mg of chlorine (as C12) per mg of As(III). The
       oxidation-reduction reaction for chlorine (as hypochlorite) is provided in the following equa-
       tion.

                           H3AsO3 + OC1 -ป H2AsO4 + H+ + Cl                   Eqn. 2-7

       The ability of chlorine to convert As(III) to As(V) was found to be relatively independent of
       pH in the range 6.3 -  8.3. Based on laboratory oxidation studies (Ghurye and Clifford,
       2001), chlorine applied in a stoichiometric excess of 3 times was capable of converting over
       95% of As(III) to As(V) within 42 seconds. Dissolved iron, manganese, and total organic
       carbon (TOC) had no significant effects on the conversion time.  However, sulfide in 1 and
       2 mg/L concentrations increased the conversion time to 60 seconds.

       The stoichiometric oxidant demands and the oxidation-reduction reactions for chlorine (as
       hypochlorite) to oxidize iron, manganese, and sulfide are provided below.

       Stoichiometric ratio for oxidation of Fe(II) is 0.64 mg  C12 per mg Fe2+.

                        2Fe2+ + OC1 +  5H2O -ป 2Fe(OH)3 + Cl + 4FT              Eqn. 2-8

       Stoichiometric ratio for oxidation of Mn(II) is 1.29 mg C12 per mg Mn2+.

                        Mn2+ + OC1 + H2O -ป MnO2 + Cl + 2FT                   Eqn. 2-9

       Stoichiometric ratio for oxidation of sulfide is 2.21 mg C12 per mg HS~.

                               HS  +OC1 --ป Sฐ + C1  +OH                      Eqn. 2-10

       Information on the design of a chlorination system can be found in Section 5.1, Chlorine
       F^re-Oxidation Design  Considerations.

       2.2.2 Permanganate
       Permanganate is a powerful oxidizing agent that is commonly used in iron and manganese
       removal processes. Potassium permanganate exists in solid, granular form and is readily
       soluble in water (60 g/L at room  temperature).  Most applications involve metering of a
       permanganate solution.

       The stoichiometric oxidant demand is 1.06 mg of permanganate per mg of As(III). The
       oxidation-reduction reaction for permanganate is provided in the following equation.

                 3H3AsO3 + 2MnO4 -ป 3H2AsO4 + FT + 2MnO2 + H2O            Eqn. 2-11

       The ability of permanganate to convert As(III) to  As(V) was found to be relatively indepen-
       dent of pH in the range 6.3 - 8.3. Based on laboratory oxidation studies (Ghurye and
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                 20

-------
       Clifford, 2001), permanganate applied in a stoichiometric excess of 3 times was capable of
       converting over 95% of As(III) to As(V) within 36 seconds. Dissolved iron, manganese,
       and TOC had no significant effects on the conversion time.  However, sulfide in 1  and 2
       mg/L concentrations increased the conversion time to 54 seconds.

       The stoichiometric oxidant demands and the oxidation-reduction reactions for permangan-
       ate to oxidize iron, manganese, and sulfide are provided below.

       Stoichiometric ratio for the oxidation of Fe(II) is 0.71 mg MnO4~ per mg Fe2+.

                 3Fe2+ + MnO4 + 7H2O -ป 3Fe(OH)3 + MnO2 + 5FT                Eqn. 2-12

       Stoichiometric ratio for oxidation of Mn(II) is 1.44 mg MnO4~ per mg Mn2+.

                    3Mn2+ + 2MnO4 + 2H2O -ป 5MnO2 + 4FT                    Eqn. 2-13

       Stoichiometric ratio for oxidation of sulfide is 2.48 mg MnO4~ per mg HS~.

                   3HS  + 2MnO4 + 5FT -ป 3Sฐ + 2MnO2 + 4H2O                 Eqn. 2-14

       The use of permanganate has several disadvantages.

       Firstly, it is difficult to handle.  It comes as a powder, is very corrosive, and stains nearly
       everything purple. The second drawback is the formation of manganese parti culates (MnO2).
       To prevent the accumulation of these deposits in the distribution system, they  must  be re-
       moved via filtration.  A third drawback is that, because permanganate is not used as a
       secondary disinfectant, another oxidant may be required for secondary disinfection.  Addi-
       tionally, if a secondary disinfectant is not used in the distribution system when a POU treat-
       ment strategy is implemented, anoxic conditions could develop in the  distribution system
       causing the As(V) to reduce back to As(III).  This would  decrease the effectiveness  of the
       POU devices and increase the cost of the treatment.

       Information on the design of a permanganate system can be found in Section 5.2, Perman-
       ganate Pre-Oxidation Design Considerations.

       2.2.3  Ozone
       Ozone is the most powerful and rapid-acting oxidizer produced. It is created by exposing
       oxygen, either in air or pure oxygen, to high energy such as an electric discharge field (i.e.,
       corona discharge) or to UV radiation.  This causes the oxygen molecules to react to form an
       unstable configuration of three oxygen atoms - the oxygen molecule contains only two.
       Because of its instability, ozone is very reactive and is a very efficient oxidant. The only by-
       product from oxidation with ozone is oxygen, which is dissolved in aqueous systems. But
       because of ozone's highly reactive nature, it will quickly self-react and revert back to oxy-
       gen if in high concentrations or not used within short periods of time. Therefore, if ozone is
       used as an oxidant, it must be produced on site.

Arsenic Treatment Technology Evaluation Handbook for Small Systems                                21

-------
       The stoichiometric oxidant demand is 0.64 mg of ozone per mg of As(III). The oxidation-
       reduction reaction for ozone is provided in the following equation.

                           H3 AsO3 + O3 -ป H2 AsO4 + H+ + O2                   Eqn. 2- 1 5

       The ability of ozone to convert As(III) to As(V) was found to be relatively independent of
       pH in the  range 6.3  - 8.3. Based on laboratory oxidation studies (Ghurye and Clifford,
       2001), ozone applied in a  stoichiometric excess of 3 times was capable of converting over
       95% of As(III) to As(V) within 18 seconds. Dissolved iron, manganese, and TOC had no
       significant effects on the conversion time. However, sulfide in 1 and 2 mg/L concentrations
       increased the conversion time to 54 and 132 seconds, respectively.

       The stoichiometric oxidant demands and the oxidation-reduction reactions for  ozone to
       oxidize iron, manganese, and sulfide are provided below.
                                                                       2+
       Stoichiometric ratio for the oxidation of Fe(II) is 0.43 mg O3 per mg Fe

                        2Fe2+ + O3 + 5H2O -ป 2Fe(OH)3 + O2 + 4FT                Eqn. 2- 1 6

       Stoichiometric ratio for oxidation of Mn(II) is 0.88 mg O3 per mg Mn2+.

                          Mn2+ + O3 + H2O -ป MnO2 + O2 + 2FT                   Eqn. 2- 1 7

       Stoichiometric ratio for oxidation of sulfide is 1.50 mg O3 per mg HS~.

                              HS +O3 + H+^Sฐ + O2 + H2O                    Eqn. 2-18

       The primary drawback to the use of ozone is that, because ozone does not provide a second-
       ary disinfectant, another oxidant may be required for secondary disinfection. Additionally,
       if a secondary disinfectant is not used in the distribution system when a POU treatment
       strategy is implemented, anoxic conditions could develop in the distribution system causing
       the As(V) to reduce back to As(III).  This  would decrease the effectiveness of the POU
       devices and increase the cost of the treatment.

       Information on the design of an ozonation system can be found in Section 5.3, Ozone Pre-
       Oxidation Design Considerations.

       2.2.4  Solid Phase Oxidants  (Filox-R™)
       Filox-R™ is a granular manganese dioxide media that can catalyze the oxidation of As(III)
       to As(V)  (Ghurye and Clifford, 2001).  This media catalytically oxidizes As(III) to As(V)
       using dissolved oxygen in the water. The Filox-R™ media also tends to adsorb some of the
       arsenic. New media can adsorb as much as 26% of the arsenic. Once  the media's capacity
       is exhausted, the media will no longer remove arsenic but will continue to oxidize it.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                22

-------
       The stoichiometric oxidant demand is 0.21 mg of oxygen per mg of As(III). The oxidation-
       reduction reaction for oxygen is provided in the following equation.

                           2H3AsO3 + O2(aq) -ป 2H2AsO4- + 2H+                   Eqn. 2-19

       Based on laboratory oxidation studies (Ghurye and Clifford, 2001) with an empty bed con-
       tact time (EBCT) of 1.5 minutes, Filox-R™ was capable of converting over 98.7% of As(III)
       to As(V). Decreasing the pH from 8.3 to 6.0 increased the conversion to 100%. Dissolved
       iron, manganese, hydrogen sulfide, and total organic carbon (TOC) were found to interfere
       with arsenic oxidation when the dissolved oxygen (DO) concentration was low (0.1 mg/L)
       and the EBCT was low (1.5 minutes). Either increasing the DO concentration (to 8.2 mg/L)
       or increasing the EBCT (to 6 minutes), eliminated the effects of these interfering reduc-
       tants.

       Filox-R™ also has the ability to remove iron, manganese, hydrogen sulfide. The stoichio-
       metric oxidant demands and the oxidation-reduction reactions for oxygen to oxidize iron,
       manganese, and sulfide are provided below.

       Stoichiometric ratio for the oxidation of Fe(II) is 0.43 mg O2 per mg Fe2+.

                           4Fe2+ + 3O2 + 6H2O + 2e" -ป 4Fe(OH)3                 Eqn. 2-20

       Stoichiometric ratio for oxidation of Mn(II) is 0.58 mg O2 per mg Mn2+.

                                  Mn2+ + O2 + 2e- -ป MnO2                      Eqn. 2-21

       Stoichiometric ratio for oxidation of sulfide is 0.48 mg O2 per mg HS~.

                              2HS + O2 + 2FT -ป 2Sฐ + 2H2O                    Eqn. 2-22

       The primary drawback to the use of a solid phase oxidant is that, because the solid phase
       oxidant does not provide a secondary disinfectant,  another oxidant may be required for
       secondary disinfection. Additionally, if a secondary disinfectant is not used in the distribu-
       tion system when a POU treatment strategy is implemented, anoxic conditions  could de-
       velop in the distribution system causing the As(V) to reduce back to As(III).  This would
       decrease the effectiveness of the POU devices and increase the cost of the treatment.

       Information on the design of a solid phase oxidant system can be found in Section 5.4, Solid
       Phase Oxidant Pre-Oxidation Design Considerations.

2.3    Sorption Treatment Processes

The following three forms of sorption treatment are addressed: (1) ion exchange,  (2) adsorption to
AA media, and (3) adsorption on IBS media.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                23

-------
       2.3.1  Ion Exchange
       Ion exchange is a physical-chemical process in which ions are swapped between a solution
       phase and solid resin phase.  The solid resin is typically an elastic three-dimensional hydro-
       carbon network containing a large number of ionizable groups electrostatically bound to the
       resin. These groups are exchanged for ions of similar charge in solution that have a stronger
       exchange affinity (i.e., selectivity) for the resin.  In drinking water treatment, this technol-
       ogy is commonly used for POE softening and nitrate removal.

       Arsenic removal is accomplished by continuously passing water under pressure through
       one or more columns packed with exchange resin.  Figure 2-8 shows a typical process flow
       diagram for ion exchange.  As(V) can be removed through the use of strong-base anion
       exchange resin (SBA) in either chloride or hydroxide form. These resins are insensitive to
       pH in the range 6.5 to 9.0 (USEPA, 2000; reference to Clifford et al., 1998).  The following
       paragraphs discuss factors that affect IX system efficiency and economics.
uxiuam
Raw ^
Water
1
r
Pre-
Oxidation


w

Pre-
Filtration
1
k.
T


Ion
Exchange
^ Treated
Water
r
                                         Backwash    Waste
                                          Waste
Regenerant
                      Figure 2-8.  Ion Exchange Process Flow Diagram.

       The exchange affinity of various ions is a function of the net surface charge.  Therefore, the
       efficiency of the IX process for As(V) removal depends strongly on the solution pH and the
       concentration of other anions, most notably sulfates and nitrates. These and other anions
       compete for sites  on the exchange resin according to the following selectivity sequence
       (Clifford, 1999).

                          SO42 > HAsO42 > NO3-, CO32 > NO2 > Cl

       High levels of total dissolved solids (TDS) can adversely affect the performance of an IX
       system. In general, the IX process is not an economically viable treatment technology if
       source water contains over 500 mg/L of TDS (Wang et al., 2000) or over 50 mg/L of sulfate
       (SO42-) (Kempic, 2002). Figure 2-9 illustrates the effect of sulfate ions on the performance
       of IX media. Although this  relationship will not be exactly the same for all  waters, it does
       provide a general indication of the impact of sulfates on IX treatment.  Additionally, small
       amounts of iron may form a soluble complex with arsenic and carry it out of the column.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                     24

-------
              1,600
           o
           i  1,200
           JS
            a
            s
            3
           I
           •o
            3J
           M
               800
400
                 0
0
                             25
                                               100
125
                                  50         75
                             Sulfate Concentration (mg/L)

   Figure 2-9.  Effect of Sulfate on Ion Exchange Performance (Clifford, 1999).

One of the primary concerns related to IX treatment is the phenomenon known as chro-
matographic peaking, which can cause As(V) and nitrate levels in the treatment effluent to
exceed those in the influent stream. This can occur if sulfates are present in the raw water
and the bed is operated past exhaustion. Because  sulfate is preferentially exchanged, in-
coming sulfate anions may displace previously sorbed As(V) and nitrate.  In most
groundwaters, sulfates are present in concentrations that are orders of magnitude greater
than As(V). Therefore, the level of sulfates is one of the most critical factors to consider for
determining the number of bed volumes that can be treated.  A useful technique for avoid-
ing chromatographic peaking is to perform careful monitoring of the effluent stream during
startup. Then, based on the analysis, determine a setpoint for the total volume treated be-
fore breakthrough occurs. This volumetric setpoint  would then be used to trigger the regen-
eration cycle. Regular monitoring of the column effluent should be continued to insure that
loss  of capacity in the media does not lead to premature breakthrough.  Frequently, the
volumetric setpoint is based on the breakthrough of sulfate.  The kinetics of breakthrough
are rapid; therefore a margin of safety should be provided or a guard column should be used
in series with the IX column.

Hydraulic considerations associated with IX include empty bed contact time (EBCT) and
headloss.  The recommended EBCT range is 15 minutes.  EBCTs as low as 1.5 minutes
have been shown to work in some installations.  The presence of suspended solids in the
feed water could gradually plug the media, thereby increasing headloss and necessitating
more frequent backwashing. Therefore, pre-filtration is recommended if the source water
turbidity exceeds 0.3 NTU.

Another concern is resin fouling. Resin fouling occurs when mica or mineral-scale coat the
resin or when ions bond the active sites and are not removable by the standard regeneration
methods.  This can have a significant effect on the resins capacity as the media becomes
older. Replacement of the media or reconditioning  may be needed after a number of years.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                                       25

-------
       Resin that has been used to exhaustion can be regenerated on-site using a four-step process:
       (1) backwash, (2) regeneration with brine (for chloride-form SB A) or caustic  soda (for
       hydroxide-form SB A), (3) slow water rinse, and (4) fast water rinse. It is recommended that
       small systems use the chloride-form resin due to the easier chemical handling  consider-
       ations for regeneration.

       Single-pass regeneration of anion exchange media typically produces 45 bed volumes of
       brine waste (USEPA,  2000; reference to AwwaRF,  1998).  In a study conducted by the
       USEPA (Wang et al., 2000), dissolved arsenic concentrations in spent brine ranged from
       1.83  mg/L to 38.5 mg/L, with an average value of 16.5 mg/L. It is anticipated that for most
       sources with arsenic levels above 0.010 mg/L and sulfate levels below 50 mg/L, the spent
       regenerant will contain at least 5.0 mg/L of dissolved arsenic.

       Liquid waste streams (less than 0.5% solids) are evaluated directly against the TC to charac-
       terize hazard  potential. Those liquid waste streams that contain more than 5.0 mg/L of
       arsenic would be classified as hazardous waste based on TC.

       Indirect discharge may be an option since wastes that pass through a sewer system to a
       POTW are exempt from RCRA regulation. The  critical factor dictating the feasibility of
       this option will be TBLLs for arsenic and TDS. The background level will change because
       of the drinking water treatment process, which may lead to revised TBLLs.  The revised
       TBLL would be used  to determine if the liquid waste stream could be discharged to the
       POTW.  Water systems that elect to use brine recycle will further concentrate the  dissolved
       arsenic and solids, making it even more unlikely that the stream will meet local TBLLs.

       Because of RCRA requirements and cost implications, off-site disposal of hazardous waste
       or on-site treatment of waste is likely to be infeasible for small water systems.

       Replacement of IX media may be required over time. Based on previous studies, spent IX
       resin does not exceed any TC concentrations,  enabling it to be disposed of in a municipal
       solid waste landfill.  This is true regardless of whether or not the media has been regener-
       ated  prior to conducting the TCLP

       Information on the design of an IX system can be  found in Section 6, Process Design Con-
       siderations.

       2.3.2  Activated Alumina
       Activated alumina is a porous, granular material with ion exchange properties. The media,
       aluminum trioxide, is  prepared through the dehydration of aluminum hydroxide at  high
       temperatures.  AA grains have a typical diameter of 0.3 to 0.6 mm and a high surface area
       for sorption.

       In drinking water treatment, packed-bed AA adsorption is commonly used for removal of
       natural organic matter and fluoride.  The removal of As(V) by AA adsorption can be accom-
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                26

-------
       plished by continuously passing water under pressure through one or more beds packed
       with AA media. Figure 2-10 shows a typical process flow diagram for ion exchange. Dashed
       lines and boxes indicate optional streams and processes. The efficiency and economics of
       the system are contingent upon several factors, as discussed in the following paragraphs.
        Oxidant
               1
         Acid	,
                                   Base	1
                                          i
                                         _*_
      Raw
      Water
Pre-
Oxidation


^P

pH
Adjustment


^p

Pre-
Filtration
i
Backwash


1

w
i
Activated
Alumina


4-

pH
Re-Adjustment

                                                                    Treated
                                                                     Water
                                       Waste
                                               Waste
                                             Regenerant
                  Figure 2-10. Activated Alumina Process Flow Diagram.

       The level of competing ions affects the performance of AA for As(V) removal, although not
       in the same manner nor to the same extent as IX. The following selectivity sequence has
       been  established for AA adsorption (USEPA, 2000):
OH
H2AsO4-
Si(OH)3O > F > HSeO3 > TOC
                                                SO42
                                                                    H3AsO3
       The selectivity of AA towards As(III) is poor, owing to the overall neutral molecular charge
       at pH levels below 9.2. Therefore, pre-oxidation of As(III) to As(V) is critical.  Several
       different studies have established the optimum pH range as 5.5-6.0, and demonstrated greater
       than 98% arsenic removal under these conditions.  AA column runs operated under acidic
       pH conditions are  5 to 20 times longer than under natural pH conditions (6.0-9.0), as de-
       picted in Figure 2-11. However, many small utilities elect to conduct AA treatment under
       natural pH conditions.  In these cases, the savings in capital and chemical costs required for
       pH adjustment and media regeneration offset the costs associated with decreased run length.
=
_o
s
es
H
                20,000
                16,000-
                12,000
              S  8,000
              s
              I
              •a  4,000
              OJ   '
              M
                    0
                                                                          10
                                             Water pH
                Figure 2-11. Effect of pH on Activated Alumina Performance
               (USEPA, 2000; original data from Hathaway and Rubel, 1987).
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                                        27

-------
       Several constituents can interfere with the adsorption process, either by competing for ad-
       sorption sites or clogging the media with particulate matter.  These constituents, and their
       corresponding problematic levels, are summarized in Table 2-3.
                   Table 2-3. Water Quality Interferences with AA Adsorption.
Parameter
Chloride
Fluoride
Silica
Iron
Manganese
Sulfate
Dissolved Organic Carbon
Total Dissolved Solids
Problem Level
250mg/L1
2mg/L1
30mg/L2
0.5 mg/L1
0.05 mg/L '
720mg/L3
4mg/L3
1,000 mg/L 3
        1 AwwaRF (2002)
        2 Clifford (2001)
         Wang, et aL (2000)
       Hydraulic considerations associated with AA adsorption include empty bed contact time
       and headloss. For most types of AA media, the recommended EBCT range is 310 minutes.
       The presence of suspended solids in the feed water could gradually clog the media, thereby
       increasing headloss. Pre-filtration is recommended for sources where the turbidity exceeds
       0.3 NTU.

       The technologies and market for alumina-based adsorptive media continue to expand. There
       are several emerging proprietary media, commonly referred to as modified AA, which con-
       tain alumina in a mixture with other substances such as iron and sulfur. In some instances,
       these media have greater overall adsorptive capacities, enhanced selectivity towards ar-
       senic, and/or greater overall operational flexibility than conventional AA, thus making them
       more cost-effective.  To account for this industry growth, the decision trees in Section 3
       include a treatment alternative known as modified-AA.  If this endpoint is reached, the
       water system should strongly consider more detailed investigation into current, innovative
       media.  It is required by most states that all media used in water treatment be approved
       under NSF Standard 61.

       AA media can either be regenerated on-site or disposed of and replaced with fresh media.
       On-site regeneration of AA media typically produces 37 to 47 bed volumes of caustic soda
       waste (USEPA, 2000).  Because of the high pH of the regeneration process, roughly 2% of
       the AA media dissolves during each regeneration sequence.  Therefore, the waste solution
       typically contains high levels of TDS,  aluminum, and soluble arsenic. In most cases, this
       arsenic level  will exceed the 5.0 mg/L TC, and the waste stream will be classified as a
       hazardous liquid waste.  Backwashing may also be necessary to prevent cementation of the
       media, which can occur as a result of dissolution caused by chemical addition during regen-
       eration. For these reasons, regeneration of AA is likely to be an infeasible option for most
       small water systems.


Arsenic Treatment Technology Evaluation Handbook for Small Systems                                 28

-------
      The alternative for utilities considering AA adsorption is the use of throwaway media, oper-
      ated with or without pH adjustment.  The savings in O&M requirements and residuals dis-
      posal may offset the cost of periodically replacing the media.  For this option, systems must
      provide an equalization basin for backwash water (if applicable) and a staging area to store
      spent media prior to disposal. Throwaway AA media is expected to not exceed any TCs,
      enabling it to be disposed of in a municipal solid waste landfill (Wang et al., 2000).  As an
      added convenience to small systems, media suppliers may offer a media disposal service
      with the purchase of their media.

      Information on the design of an AA system can be found in Section 6, Sorption Process
      Design Considerations.  Information on POU systems can be found in Section  8.1.1,
      Adsoprtion Point-of-Use Treatment.

      2.3.3  Iron Based Sorbents
      Adsorption on IBS is an emerging treatment technique for arsenic. Examples of IBS prod-
      ucts currently available withNSF 61 approval are shown in Table 2-4. The sorption process
      has been described as chemisorption (Selvin et al., 2000), which is typically considered to
      be irreversible. It can be applied in fixed bed pressure columns, similar to those for AA.
      Due to limited performance research at the time the Arsenic Rule was promulgated, it was
      not designated as  a BAT or a SSCT by the USEPA.


                        Table 2-4. Examples of Iron Based Sorbents. *
Product Name
G2
SMI III
GFH
Bayoxide E 33
Company
ADI International
SMI
U.S. Filter/General Filter Products
Bayer AG
Material Type
Modified Iron
Iron/Sulfur
Granular Ferric Hydroxide
Iron Oxide
        Examples are taken from Rubel 2003.
      The few studies conducted with IBS media have revealed that the affinity of this media for
      arsenic is strong under natural pH conditions, relative to AA.  This feature allows IBS to
      treat much higher bed volumes without the need for pH adjustment.  However, similar to
      AA, optimal IBS performance is obtained at lower pH values. The recommended operating
      conditions include an EBCT of 5 minutes and a hydraulic loading rate of 5 gpm/sft.

      Phosphate has been shown to compete aggressively with As(V) for adsorption sites. Each
      0.5 mg/L increase in phosphate above 0.2 mg/L will reduce adsorption capacity by roughly
      30% (Tumalo, 2002).

      In previous studies, exhausted IBS media has not exceeded any TCs,  enabling it to be dis-
      posed  of in a municipal solid waste landfill (MacPhee et al., 2001).  As an added conve-
      nience to small systems, media suppliers may offer a media disposal  service with the pur-
      chase of their media.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                29

-------
       Information on the design of IBS systems can be found in Section 6, Sorption Process
       Design Considerations. Information on POU systems can be found in Section 8.1.1, Ad-
       sorption Point-of-Use Treatment.

2.4   Membrane Treatment Processes

Membrane separation technologies are attractive arsenic treatment processes for small water sys-
tems. They can address numerous water quality problems while maintaining simplicity and ease of
operation.  The molecular weight cut-off of microfiltration (MF) processes necessitates the use of a
coagulation stage to generate arsenic-laden floe and is therefore discussed in Section 2.5.3, Coagu-
lation-Assisted'Microfiltration. However, RO units have a much larger retention spectrum, and can
be used as stand-alone arsenic treatment under most water quality conditions. Figure 2-12 provides
a block flow diagram for a typical RO membrane process. Dashed lines indicate optional streams
and processes.
                      Oxidant

                   Raw _     Pre-         Pre-        RO/NF      Treated
                  Water     Oxidation   1 Filtration     Membrane     Water

                                       Backwash     Retentate
                                        Waste        Waste

                    Figure 2-12.  RO Membrane Process Flow Diagram.

Most RO membranes are made of cellulose acetate or polyamide composites cast into a thin film.
The semi-permeable (non-porous) membrane is then constructed into a cartridge called an  RO
module, typically either hollow-fiber or spiral-wound.

RO is a pressure-driven membrane separation process capable of removing dissolved solutes from
water by means of particle size, dielectric characteristics, and hydrophilicity/hydrophobicity.  Re-
verse osmosis is capable of achieving over 97% removal of As(V) and 92% removal of As(III) in a
single pass (NSF, 200la; NSF 200Ib). As an added benefit, RO also effectively removes several
other constituents from water, including organic carbon, salts, dissolved minerals, and color.  The
treatment process is relatively insensitive to pH. In order to drive water across the membrane
surface against natural osmotic pressure, feed water must be sufficiently pressurized with a booster
pump. For drinking water treatment, typical operating pressures are between 100 and 350  psi.
Water recovery is typically 60 -80%, depending on the desired purity of the treated water. In some
cases, particularly POU applications, RO units are operated at tap water pressures. This results in
a significantly lower water recovery.

Multiple RO units can be applied in series to improve the overall arsenic removal efficiency. Fig-
ure 2-13 illustrates a 2-stage RO treatment process.  The overall rejection rate for a multi-staged
RO treatment process can be calculated as:

Arsenic Treatment Technology Evaluation Handbook for Small Systems                                30

-------
                                                                               Eqn. 2-23
Where:
   E  = Overall Rej ection Rate (Treatment Efficiency),
   Es  = Individual Stage Contaminant Rejection Rate, and
   n  = Number of Stages.
                               Stage 1               Stage 2
                Feed
                Water
                                                      Permeate
                                                                  Combined
                                                                  Retentate
                 Figure 2-13.  Two-Stage RO Treatment Process Schematic.

Membrane fouling can occur in the presence of NOM and various inorganic ions, most notably
calcium, magnesium, silica, sulfate, chloride, and carbonate. These ions can be concentrated (in
the retentate) to concentrations an order of magnitude higher than in raw water. This can lead to the
formation of scale on the membrane surface, which in turn can cause a decline in arsenic rejection
and water recovery. Further, the membrane surface can act as a substrate for biological growth.
Membrane cleaning can restore treatment performance; however, the cleaning process is difficult
and costly. The rate of membrane  fouling depends on the configuration of the module and feed
water quality.  Most RO  modules are designed for cross-flow filtration, which allows water to
permeate the membrane while the retentate flow sweeps rejected salts away from the membrane
surface. In many cases, pre-filtration (commonly through  sand or granular activated carbon) is
worthwhile. This minimizes the loading of salt precipitates and suspended solids on the membrane
surface, thereby extending run length, improving system hydraulics, and reducing O&M require-
ments.

Some membranes, particularly those composed of polyamides, are sensitive to chlorine.  Feed wa-
ter should be dechlorinated (if applicable) in these instances. Another potential concern associated
with RO treatment is the  removal of alkalinity  from water, which in turn could affect corrosion
control within the distribution system. If feasible, this problem can usually be avoided by conduct-
ing sidestream treatment for arsenic removal.

Indirect discharge to a POTW or direct discharge to an on-site sewerage system (for POU systems)
are considered the most viable residuals disposal option.  For those systems considering indirect
discharge, the retentate must meet local TBLLs for arsenic. The arsenic concentration in the retentate
can be calculated using Equation 2-24.
Where:
    C:
                                             1-P
Concentration of Species / in the Retentate (mg/L),
Concentration of Species / in the Feed Stream (mg/L),
                                                                               Eqn. 2-24
Arsenic Treatment Technology Evaluation Handbook for Small Systems


-------
    Es    = Individual Stage Contaminant Rejection Rate (% expressed as a decimal), and
    (3     = Individual Stage Water Recovery Rate (% expressed as a decimal).

It is not anticipated that a small system will use an RO system for centralized treatment because RO
systems for centralized treatment can be expensive. Therefore, no design information on central-
ized systems has been provided in this Handbook. However, information on RO POU systems can
be found in Section 8.1.2, Reverse Osmosis Point-of-Use Treatment.

2.5   Precipitation/Filtration Treatment Processes

•  The following four chemical precipitation processes are addressed:
•  LS,
•  Conventional Gravity Coagulation/Filtration,
•  CMF,
•  CADF, and
•  Oxidation/Filtration.

Figure 2-14 provides a block flow diagram for a generic precipitation/filtration process. Dashed
lines and boxes indicate optional streams and processes.
     Oxidant
Acid/Base --i   Coagulant —,
 Raw
Water
                                                           Acid/Base --
                                                                    PH
                                                                Re- Adjustment
                                                             Treated
                                                              Water
                                                   Backwash
                                                    Waste

             Figure 2-14. Generic Precipitation/Filtration Process Flow Diagram.

       2.5.1   Enhanced Lime Softening
       Lime softening is a chemical-physical treatment process used to remove calcium and mag-
       nesium cations from solution.  The addition of lime increases the pH of solution, thereby
       causing a shift in the carbonate equilibrium and the formation of calcium carbonate and
       magnesium hydroxide precipitates. These precipitates are amenable to removal by clarifi-
       cation and filtration.

       LS solely for arsenic removal is uneconomical and is generally considered cost-prohibitive.
       However, for water systems that use LS to reduce hardness, the process can be enhanced for
       arsenic removal. To remove As(V), additional lime is added to increase the pH above 10.5.
       In this range magnesium hydroxide precipitates and As(V) is removed by co-precipitation
       with it.  As(V) removal by co-precipitation with calcium carbonate (i.e., below a pH of
       10.5) is poor (less than 10%).
Arsenic Treatment Technology Evaluation Handbook for Small Systems


-------
       The amount of waste residual produced by LS is dependent on the hardness removed. While
       the total volume of waste produced from LS is typically higher than that produced by coagu-
       lation/filtration and co-precipitative processes, the arsenic concentration in the sludge is
       generally lower because more solids are produced.  Typical solids concentrations are 1 - 4%
       arsenic.  Prior to disposal, this waste residual will require thickening and dewatering, most
       likely via mechanical devices. Previous studies have indicated that typical lime sludge will
       not exceed TC limits, enabling it to be disposed of in a municipal solid waste landfill (Fields
       et al., 2000a).

       Because LS is unlikely to be installed solely for arsenic removal in small systems, no design
       discussion is provided in this Handbook.

       2.5.2  Conventional Gravity Coagulation/Filtration
       Coagulation is the process of destabilizing the surface charges of colloidal and suspended
       matter to allow for the agglomeration of particles.  This process results in the formation of
       large, dense floe, which is amenable to removal by clarification or filtration.  The most
       widely used coagulants for water treatment are aluminum and ferric salts, which hydrolyze
       to form aluminum and iron hydroxide particulates, respectively.

       Conventional gravity  coagulation/filtration processes use gravity to push water through a
       vertical bed of granular media that retains the floe and are typically used within surface
       water treatment plants. They are less commonly used for treatment of groundwater supplies
       since these sources usually contain much lower concentrations of suspended solids, organic
       carbon, and pathogenic microorganisms. Installation and  operation of a conventional grav-
       ity coagulation/filtration process solely for arsenic  removal is uneconomical.

       Coagulation/filtration processes can be optimized to remove dissolved inorganic As(V) from
       water. The mechanism involves the adsorption of As(V)  to an aluminum or ferric hydrox-
       ide precipitate.  The As(V) becomes entrapped as the particle continues to agglomerate.
       As(III)  is not effectively removed because of its overall  neutral charge under natural pH
       conditions.  Therefore, pre-oxidation is recommended.  The efficiency and economics of
       the system are contingent upon several factors, including the type and dosage of coagulant,
       mixing intensity, and pH. In general, however, optimized coagulation-filtration systems are
       capable of achieving over 90% removal of As(V) and producing water with less than 0.005
       mg/L of As(V).  Influent As(V) levels do not appear to impact the effectiveness of this
       treatment process.

       Iron-based coagulants, including ferric sulfate and ferric chloride,  are more effective at
       removing As(V) than their aluminum-based counterparts. This is because iron hydroxides
       are more stable than aluminum hydroxides in the  pH range 5.5 to 8.5.  A fraction of the
       aluminum remains as a soluble complex, which is incapable of adsorbing As(V) and can
       pass through the filtration stage. The optimal pH ranges for coagulation with aluminum and
       ferric salts are 5 to 7 and 5 to 8, respectively.  At pH values above 7, the removal perfor-
       mance of aluminum-based coagulants drops markedly.  Feed water pH should be adjusted
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                33

-------
       to the appropriate range prior to coagulant addition.  Post-filtration pH adjustment may be
       necessary to optimize corrosion control and comply with other regulatory requirements.

       Several batch studies have demonstrated that As(V) removal is positively related to coagu-
       lant dosage. However, specific dose requirements needed to meet As(V) removal objec-
       tives were contingent upon the source water quality and pH.  Effective coagulant dosage
       ranges were 5-25 mg/L of ferric chloride and as much as 40 mg/L of alum.

       Water intended for indirect discharge will be subject to TBLLs for IDS and arsenic. Dewa-
       tering can be accomplished by gravity thickening, followed by other mechanical or non-
       mechanical techniques. Settling basins can be used to allow settleable solids to drop out of
       solution via gravity, while the supernatant can be decanted and recycled to the process head.
       The solids can be slurried out periodically and passed through a small filter press for dewa-
       tering. The resultant sludge can be disposed of in a municipal solid waste landfill if it meets
       the criteria of the PFLT (no free liquid) and the TCLP. Previous studies have indicated that
       typical coagulation/filtration sludge will not exceed TC limits (Fields et al., 2000a).

       Because  conventional gravity coagulation filtration is unlikely to be installed  solely for
       arsenic removal in small systems, no design discussion is provided in this Handbook.

       2.5.3  Coagulation-Assisted Microfiltration
       Coagulation-Assisted Microfiltration uses the same coagulation process described above.
       However, instead of the granular media filtration step, the water is forced through a semi-
       permeable membrane by a pressure differential.  The membrane retains the As(V) laden
       floe formed in the coagulation step.

       The use of pre-engineered CMF package plants is a realistic possibility for new installations
       where water quality precludes the use of sorption treatment. Due to limited full-scale appli-
       cation, it was not designated as a BAT by the USEPA but was listed as a SSCT in the final
       rule.

       The membrane must be periodically backwashed to dislodge solids and restore hydraulic
       capacity. Backwash water is typically a high-volume,  low solids (less than 1.0%) waste
       stream. The specific amount of solids will depend on several factors, including coagulant
       type, dosage, filter run length, and ambient solids concentration. Two treatment options are
       available for this waste stream: (1) indirect discharge, and (2) dewatering and sludge dis-
       posal (AwwaRF 2000).

       Water intended for indirect discharge will be subject to TBLLs for TDS and arsenic. Dewa-
       tering can be accomplished by gravity thickening, followed by other mechanical or non-
       mechanical techniques. Settling basins can be used to allow settleable solids to drop out of
       solution via gravity, while the supernatant can be decanted and recycled to the process head.
       The solids can be slurried out periodically and passed through a small filter press for dewa-
       tering. The resultant sludge can be disposed of in a municipal solid waste landfill if it meets
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                34

-------
       the criteria of the PFLT (no free liquid) and the TCLP. Previous studies have indicated that
       typical CMF sludge will not exceed TC limits (Fields et al., 2000a).

       Design of a CADF system can be found in Section 7, Pressurized Media Filtration Process
       Design Considerations.

       2.5.5  Oxidation/Filtration
       Oxidation/filtration refers to processes that are designed to remove naturally occurring iron
       and manganese from water. The processes involve the oxidation of the soluble forms of
       iron and manganese to their insoluble forms and then removal by filtration.  If arsenic is
       present in the water, it can be removed via two primary mechanisms: adsorption and co-
       precipitation. First, soluble iron and As(III) are oxidized. The As(V) then adsorbs onto the
       iron hydroxide precipitates that are ultimately filtered out of solution.

       Although some arsenic may be removed by adsorption/co-precipitation with manganese,
       iron is much  more efficient for arsenic removal. The arsenic removal efficiency is strongly
       dependent on the initial iron concentration and the ratio of iron to arsenic. In general, the
       Fe: As mass ratio should be at least 20:1. These conditions customarily result in an arsenic
       removal  efficiency of 80-95%. In some cases, it may be appropriate to add ferric coagulant
       to the beginning of the iron removal process to optimize arsenic removal.

       The effectiveness of arsenic co-precipitation with iron is relatively independent of source
       water  pH in  the range 5.5  to 8.5.  However, high  levels of NOM, orthophosphates, and
       silicates weaken arsenic removal efficiency by competing for sorption sites on iron hydrox-
       ide precipitates (Fields et al., 2000b).

       The common iron/manganese methods consist of (1) air oxidation or chemical oxidation
       followed by media filtration and (2) potassium permanganate oxidation followed by a green-
       sand media filter. The latter process is commonly referred to as the greensand process. The
       greensand process can be operated on an intermittent regeneration (IR) basis  or on a con-
       tinuous feed  (CF) basis. With IR operational procedure, the greensand filter is periodically
       regenerated with potassium permanganate following the back washing of the filter. In the
       CF mode, permanganate or chlorine is continuously added to the feed water ahead of green-
       sand filter.

       In the air/chemical oxidation filtration iron removal process, the iron is oxidized with either
       air (aeration tower) or with an oxidizing chemical, usually chlorine. Because of the limita-
       tions of air to oxidize As(III), chlorine is normally used in order for the process to be effec-
       tive for arsenic removal.  After the water is oxidized, it is filtered with a granular media to
       remove the iron hydroxide precipitates that contain the adsorbed arsenic. The effectiveness
       of the granular media is important because any iron particles that manage to get through the
       filter media will contain some arsenic.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                 35

-------
       The greensand process is a special case of pressurized granular-media filtration where the
       granular media, greensand, catalyzes the oxidation and precipitation of iron and manga-
       nese.  In the greensand process, operated an IR basis, the water is passed through a column
       of greensand media, which adsorbs and catalyzes the oxidation of the iron and manganese.
       In order for greensand to retain its adsorption and catalytic oxidation capabilities, the media
       must be regenerated with permanganate or chlorine.  When operated on an IR basis, the
       greensand filter column is taken offline and the media is soaked in a solution of permanga-
       nate. In the CF mode, permanganate or chlorine is continuously added to the  feed water
       ahead of greensand filter where they provide continuous oxidation of the iron and As(III)
       and regeneration of the greensand. If the arsenic in the ground water is not already oxidized
       to As(V),  it is recommended that CF process using chlorine or permanganate be used to
       provide continuous oxidation of the iron, manganese, and As(III).

       Greensand is manufactured by coating glauconite with manganese dioxide.  Other manga-
       nese dioxide  media are also used  for iron and manganese removal such  as  pyrolucite,
       Pyroloxฎ6, Filox-R™  7, MTMฎ8, BIRMฎ9, and Anthrasand. Greensand and some of the
       other manganese dioxide media (Filox-R™) have been shown to have some arsenic adsorp-
       tive effectiveness in removing arsenic from drinking water (Hanson et al., 1999,  Fields et
       al., 2000b, Ghurye and Clifford, 2001).

       In all  oxidation/filtration processes, the filter media must be periodically backwashed to
       dislodge solids and restore hydraulic capacity.  Backwash water is typically a high-volume,
       low solids (less than 0.1%) waste stream. The specific amount of solids will  depend on
       several factors, including raw water iron levels, coagulant addition (if any), filter run length,
       and background solids concentration.  Two treatment options are available for this waste
       stream: (1) indirect discharge and (2) dewatering and sludge disposal.

       Water intended for indirect discharge will be subject to TBLLs for TDS and arsenic.  The
       background level will change because of the drinking water treatment process, which may
       lead to revised TBLLs. The revised TBLL would be used to determine if the liquid waste
       stream could be discharged  to the  POTW.  Dewatering can be  accomplished by gravity
       thickening, followed by other mechanical or non-mechanical techniques.  Settling basins
       can be used to provide gravity clarification, while the supernatant can be decanted and
       recycled to the process head.  The solids can be slurried out periodically and passed through
       a filter press for dewatering. The resultant sludge can be disposed of in a municipal solid
       waste landfill if it passes the PFLT (no free liquid) and the TCLP Previous studies have
       indicated that typical ferric coagulation-filtration sludge will not exceed TC limits (Fields
       et al.,  2000b).

       Design of an oxidation/filtration system can be found in Section 7, PressurizedMedia Fil-
6 Pyrolox is a trademark of American Materials.
7 Filox is a registered trademark of Matt-Son, Inc., Harrington, IL. Filox-R is a trademark of Matt-Son, Inc.,
 Harrington, IL.
8 MTM is a registered trademark of Clack Corporation, Windsor, WI.
9 Birm is a registered trademark of Clack Corporation, Windsor, WI.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                36

-------
       tration Process Design Considerations.

2.6    Point-of-Use Treatment

Under the final Arsenic Rule, POU devices are approved as SSCTs. However, SDWA requires that
the devices be owned, controlled, and maintained by the public water utility or by an agency under
contract with the water utility. Therefore, the responsibility of operating and maintaining the de-
vices cannot be passed to the customer.

POU devices are particularly attractive for removing contaminants that pose only an ingestion risk,
as is the case with arsenic.  This is because a very small fraction of the total water supplied to a
given household is ultimately consumed. In most cases, the POU unit is plumbed in at the kitchen
sink (the device will have its own faucet).

The primary advantage of employing POU treatment in a small system is reduced capital and treat-
ment costs, relative to centralized treatment. On the downside, however, these programs generally
incur higher administrative and monitoring costs to make sure that all units are functioning  prop-
erly. Previous studies have suggested that POU programs are an economically viable alternative to
centralized treatment for systems serving roughly 50-500 people.

Most POU devices do not address the issue of pre-oxidation. While RO may remove As(III) to
acceptable standards, sorptive processes such as AA or IBS will probably not.  In this case, water
systems may need to conduct centralized chlorine treatment to convert As(III) to As(V). There is
also a concern that even with centralized pre-oxidation, anoxic conditions could exist in the distri-
bution system that allow As(V) to reduce back to As(III).  Depending  on the extent of reduction,
this could be detrimental to a POU program.

The technologies that are most amenable to POU treatment include AA, IBS, and RO. AA and RO
are approved as SSCTs for POU applications (USEPA, 2002a). Although there are no IBS systems
currently approved as SSCTs, there are several media currently being tested.

The primary criteria for selecting an appropriate POU treatment device are arsenic removal perfor-
mance and cost. The unit must be independently certified against NSF/ANSI product standards to
be used for compliance purposes.

More  information on POU technologies and can be found in Section 8, Point-of-Use Treatment,
and in USEPA's soon to be released Guidance for Implementing a Point-of-Use or Point-of-Entry
Treatment Strategy for Compliance with the Safe Drinking Water Act (USEPA, 2002b Draft).
Arsenic Treatment Technology Evaluation Handbook for Small Systems                               37

-------
Preceeding Page Blank
                                                                        Section  3
                                  Arsenic Treatment Selection
        3.1   Selection Criteria

        The task of navigating through the alternative arsenic treatment technologies involves several
        technical considerations. Although nearly all of the unit processes previously presented could be
        used for arsenic reduction at an arbitrary site,  some are more economically viable under specific
        circumstances. Optimization of existing processes is a realistic option for some utilities. Although
        most water systems today have been designed without the goal of arsenic removal, many current
        practices may accomplish incidental removal. Optimization of these processes is a realistic option.
        The utility should coordinate the selection and implementation process with its state drinking water
        program.

              3.1.1  Source Water Quality
              Source water quality dictates the performance of the removal processes identified in Section
              2.  In turn, process performance, associated  O&M requirements, and  residuals disposal
              dictate the economics of a particular treatment approach. Therefore, it is important that
              utilities conduct thorough up-front monitoring of water quality at all active sources to make
              the most informed treatment selection decision.

              Tables 3-1 and 3-2 provide a summary of recommended monitoring parameters and associated
              analytical methods. The parameters are divided into two categories: (1) Key and (2) Other.
              Key parameters are those most critical to evaluating the treatment performance potential of
              various arsenic removal processes. These parameters should be monitored multiple times
              over the course of several weeks or months to capture variability in concentrations.  Other
              parameters should be monitored at least once in  order to optimize the selected arsenic
              treatment method.
        Arsenic Treatment Technology Evaluation Handbook for Small Systems                              39

-------
                      Table 3-1.  Key Water Quality Parameters to be Monitored.
Parameter
Arsenic, Total '
Arsenic {As(III)}
Arsenate (As(V)}
Chloride (Cl') 2
Fluoride (F) '-2
Iron (Fe) 2
Manganese (Mn) 2
Nitrate (NO;) '
Nitrite (NO;) '
Orthophosphate (PO;3) '
pH'<2
Silica1
Sulfete (SO;2) 2
Total Dissolved Solids (TDS) 2
Total Organic Carbon (TOC)
USEPA Method
200.8
200.9


300.0
300.0
200.7
200.9
200.7
200.8
200.9
300.0
353.2
300.0
353.2
365.1
300.0
150.1
150.2
200.7
300.0
375.2

415.1
Standard Method 3
3113 B
3114 B
3500-As B
3500-As B
4110 B
4500-C1-D
4500-C1-B
4110 B
4500-F- B
4500-F C
4500-F D
4500-F- E
3120B
3111 B
3113 B
3120B
3111 B
3113 B
4110 B
4500-NO; F
4500-NO; D
4500-NO; E
4110 B
4500-NO; F
4500-NO; E
4500-NO; B
4500-P F
4500-P E
4110 B
4500-H+ B
4500-SiD
4500-SiE
4500-SiF
4110 B
4500-SO/- F
4500-SO/- C
4500-SO/- D
4500-SO/- E
2540 C

ASTM4
D2972-93B
D2972-93C


D4327-91
D512-89B
D4327-91
D1179-93B


D4327-91
D3867-90A
D3867-90B
D4327-91
D3867-90A
D3867-90B
D515-88A
D4327-91
D1293-95
D859-95
D4327-91
D5 16-90


         1 USEPA Approved Methods for Drinking Water Analysis of Inorganic Chemicals and other parameters.
         2 USEPA Recommended Methods for Secondary Drinking Water Contaminants.
         318th and 19th editions of Standard Methods for the Examination of Water and Wastewater, 1992 and 1995,
          American Water Works Association (AWWA).
         4 Annual Book of ASTM Standards,  1994 and 1996, Vols. 11.01 and 11.02, American Society for Testing
          and Materials (ASTM).
Arsenic Treatment Technology Evaluation Handbook for Small Systems
40

-------
                   Table 3-2.  Other Water Quality Parameters to be Monitored.
Parameter
Alkalinity '
Aluminum (Al) 2
Calcium (Ca^)1
Magnesium (Mg+2) '
Turbidity
Water Hardness
USEPA Method

200.7
200.8
200.9
200.7
200.7
180.1
215.1
242.1
Standard Method 3
2320 B
3120B
3113 B
3111 D
3500-Ca D
3111 B
3120B
3113 B
3120B
3500-MgE


ASTM4
D1067-92B

D511-93A
D511-93B
D511-53B
D511-93B


        1 USEPA Approved Methods for Drinking Water Analysis of Inorganic Chemicals and other parameters.
        2 USEPA Recommended Methods for Secondary Drinking Water Contaminants.
        318th and 19th editions of Standard Methods for the Examination of Water and Wastewater, 1992 and 1995.
         American Water Works Association (AWWA).
        4 Annual Book of ASTM Standards,  1994 and 1996, Vols. 11.01 and 11.02, American Society for Testing
         and Materials (ASTM).	

       3.1.2  Process  Evaluation Basis
       There are several variables, design criteria, and assumptions that should be established prior
       to navigating the decision trees and cost tables. These include the following:
       •  Existing Treatment Processes
       •  Target Finished Water Arsenic Concentration
       •  TBLLs for Arsenic and TDS
       •  Domestic Waste Discharge Method
       •  Land Availability
       •  Labor Commitment
       •  Acceptable Percent Water Loss
       •  Maximum Source Flowrate
       •  Average Source Flowrate
       •  State or primacy agency requirements that are more stringent than those of the USEPA.

3.2   Process Selection Decision Trees

Decision trees are useful tools for narrowing the field of available treatment technologies to those
that are most economical for a particular system.  This is accomplished through a series of input-
output blocks, which direct the utility along the path towards the best technologies. While they do
not always point to a single solution, they allow the utility to rapidly eliminate some technologies
that are cost-prohibitive for a specific application.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
41

-------
It is critical that the utility employ these decision trees, rather than the cost correlation curves,
as the primary tool for selecting an arsenic mitigation strategy. These trees take into account
system-specific conditions and system preferences.

The decision trees  guide the utility to the technologies that are expected to work best for their
particular situation. In some cases, the pathway is contingent upon a water utility's willingness to
impose a particular change in their operating scheme. These decision-making scenarios were pre-
sented only for cases where it may be economically advantageous to make such a change.  How-
ever, there may be other restrictions (i.e., operating labor, space) to making the operational changes
in question. In other cases, there may be more than one equally viable technology. At that point,
the water utility should further evaluate its preferences with respect to costs and labor commit-
ments, and capabilities with respect to residuals disposal and facility expansion.

The decision trees employ the following labeling scheme:
        Question/Decision Box
        Action Box

        Reference Box

The Question/Decision block requests information or utility preference in the form of a yes/no or
multiple-choice question.  The Action Box provides the recommended follow-up action given sys-
tem-specific constraints and preferences.  This box is frequently used as the stopping point for a
particular branch of the decision tree.   The Reference Box simply directs the utility to another
portion of the decision tree.

If a utility reaches an action box pertaining to switching sources, blending, or existing  treatment
optimization, they should refer to Section 2 for more specific information.  If a utility reaches an
action box pertaining to new treatment installation, they should refer to Section 4 for cost informa-
tion and Sections 6-8 for specific design considerations.

The decision trees are intended for use as an iterative tool.  If a utility proceeds to a specific action
box, conducts follow-up cost estimation, process optimization, and/or pilot-testing, and the results
indicate that the selected strategy may be ineffective or too expensive for arsenic removal, the
utility can restart the tree and modify preferences.

The following assumptions were made in the development of the decision trees:

•   Optimization of existing treatment process is economically preferable over new installations.
•   Construction of new conventional gravity coagulation/filtration or LS systems is not appropri-
    ate for the sole purpose of arsenic removal.
•   Small water systems would opt for  disposable adsorptive media rather than conduct on-site
    regeneration.
•   Small systems would choose to not generate hazardous waste for either on-site treatment or off-
    site disposal.

Arsenic Treatment Technology Evaluation Handbook for Small Systems                                 42

-------
               Decision Tree Overview
                    Non -Treatment Alternatives
            a Tree 1  Non-Treatment Alternatives
                         Treatment Selection
            a Tree 2  Treatment Selection
                  Tree 2a  Enhanced Coagulation/Filtration
                  Tree 2b  Enhanced Lime Softening
                  Tree 2c  Iron & Manganese Filtration
                      Selecting New Treatment
            a Tree 3  Selecting New Treatment
                  Tree 3a  Ion Exchange Processes
                  Tree 3b  Sorption Processes
                  Tree 3c  Filtration & Membrane Processes
                     Figure 3-1. Decision Tree Overview.
Arsenic Treatment Technology Evaluation Handbook for Small Systems


-------
                                     Tree  1
               Non-Treatment Alternatives
      Does the Running Annual
     Average Arsenic Concentration
          exceed the MCL?
        Are there one or more
        other sources available
       with arsenic levels below
            the MCL?
               I
         Can these sources be
        operated in lieu of the
      problematic source to meet
        total system demand?
               E
       Can these sources always
      be operated in conjunction
         with the high arsenic
      Can the sources be blended
       in a manner such that the
       arsenic MCL is met at all
      entry points to the system?
I
                  N
                KI
                                                  Consider switching
                                                  problematic source
                                                  to back-up/seasonal
                                                     use. Refer to
                                                     Section 2.1.2
 Would you prefer to
site/install a new source
 before employing or
 modifying treatment?
                                 N
                                                 N
                                                   Consider locating
                                                   or installing a new
                                                   source. Refer to
                                                     Section 2.1.1
                                     Go to Tree 2 - "Treatment
Are there any constraints to
blending, such as distance
between sources, water
quality impacts, water
rights, etc.?
Selection"
Y ^ A
T


/

                               N
                 Figure 3-2.  Decision Tree 1 -Non-Treatment Alternatives.
                                                                Consider using
                                                               blending to meet
                                                                MCL. Refer to
                                                                Section 2.1.3
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                                     44

-------
                                      Tree 2
                       Treatment  Selection
         Is the Arsenic in the
       problematic source water
         primarily as As(III)?
                  N
  Are the problematic
       source(s)
 pre-oxidized with either
  chlorine, potassium
permanganate, or ozone*?
     Are the problematic source(s)
     treated beyond disinfection or
          corrosion control?
       Have previous attempts to
     optimize existing treatment for
     arsenic removal been made and
              failed?
          Identify Existing
            Treatment:
    Coagulation/Filtration
    Lime Softening
    Iron/Manganese Filtration
N
                                                   Treat the problematic source water with
                                                   either chlorine, potassium permanganate,
                                                                or ozone.
                                                         Refer to Sections 2.2 & 5.0
         N
                   Go to Tree 3 -
              'Selecting New Treatment"
                   Go to Tree 2a -
          "Enhanced Coagulation/Filtration'
                  Go to Tree 2b -
              'Enhanced Lime Softening"
                   Go to Tree 2c -
            "Iron & Manganese Filtration'
   *Pre-oxidized refers to the process
     of converting As(III) to As(V)
                     Figure 3-3. Decision Tree 2 - Treatment Selection.

          Use this decision tree only after using Tree 1 "Non-Treatment Alternatives"
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                  45

-------
                                  Tree  2a
         Enhanced Coagulation/Filtration
      Identify coagulant:
      Iron-based
      Aluminum-based

      Polymer •
  Is the current process
  operated at pH < 8.5?
Evaluate increasing
 Fe coagulant dose.
Refer to Section 2.5.2
                                 Are you willing to
                                install pH adjustment
                                    capabilities?
                      N
                                                 Evaluate adjusting pH to 5.5-8.5
                                                 and increasing Fe coagulant dose.
                                                      Refer to Section 2.5.2
                        Is the current
                      process operated
                        atpH<7.0?
                         Evaluate increasing
                         Al coagulant dose.
                        Refer to Section 2.5.2
Are you willing
to install pH
adjustment capabilities?
Y
^

Evaluate adjusting pH to 5-7 and
increasing Al coagulant dose.
Refer to Section 2.5.2
                           I
N
    fm
   T
                      Are you willing to
                    switch to or incorporate
                   an iron-based coagulant?
      Denotes alternate techniques
      that should be investigated.
                              N
                       Evaluate switching to or
                          incorporating an
                        iron-based coagulant.
                        Refer to Section 2.5.2
                                                 Add new treatment technology by
                                                        going to Tree 3 -
                                                   "Selecting New Treatment"
              Figure 3-4. Decision Tree 2a - Enhanced Coagulation/Filtration.

             Use this decision tree only after using Tree 2 "Treatment Selection."
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                  46

-------
                         Tree 2b
            Enhanced  Lime Softening

Is the <
process o
pH<
1
current V
10.5?
N
r
Does the softening ป .
process remove > 10
mg/L (as CaCO3) of
magnesium?

Are you willing to
install pH adjustment
capabilities and can
you handle additional
sludge production?

Y.

N
,
E
N
Are you willing to
add magnesium and
increase the lime
dose?
Y

Y


Evaluate optimizing
existing LS process
by increasing pH
between 10.5-11.
Refer to Section 2.5.1

Add new treatment
technology by going to
Tree 3— "Selecting New
Treatment"
Evaluate optimizing
existing LS process
by adding magnesium.
Refer to Section 2.5.1

Evaluate addition of
iron (up to 5 mg/L).
Refer to Section 2.5.1
            Figure 3-5. Decision Tree 2b - Enhanced Lime Softening.

         Use this decision tree only after using Tree 2 "Treatment Selection''
Arsenic Treatment Technology Evaluation Handbook for Small Systems


-------
                                  Tree 2c
              Iron  &  Manganese  Filtration
       Are filters capable of
     handling an increase in iron
             load?
              I
      Evaluate adding a ferric
        coagulant to optimize
      influent Fe concentration.
        Refer to Section 2.5.5
     Are you willing to install an
      iron feed system, provide
      detention time and mixing,
        and possibly increase
      backwashing frequency?
N
             Is pH < 7.5?
   N
                                                 N
                                            Are you
                                            willing to
                                          adjust the pH
                                            to < 7.5?
       Evaluate adjusting
          pH to < 7.5
      Refer to Section 2.5.5
                             N
Add new treatment
technology by going
   to Tree 3 -
  "Selecting New
   Treatment"
                 Figure 3-6. Decision Tree 2c - Iron/Manganese Filtration.

            Use this decision tree only after using Tree 2 "Treatment Selection."
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                   48

-------
                                TreeS
               Selecting  New Treatment
        Are all of the following water
          quality criteria met at the
           problematic source?
          • SO42- < 50 mg/L
          • NO3- (as N) < 5 mg/L
          • TDS < 500 mg/L
          • pH >6.5 and < 9
    Go to Tree 3a -
'Ion Exchange Processes"
            Is the source water:
            Fe < 0.5 mg/L, and
             Mn < 0.05 mg/L.
                    N
    Go to Tree 3b -
  "Sorption Processes"
                                                 Go to Tree 3c -
                                         "Filtration & Membrane Processes"
                Figure 3-7. Decision Tree 3 - Selecting New Treatment.

           Use this decision tree only after using Tree 2 "Treatment Selection."
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                              49

-------
                                   Tree 3a
                 Ion  Exchange  Processes
       Are you willing to
       install and operate
         brine or caustic
          regeneration
           facilities?
N
 Are your customers
   connected to a
wastewater collection
 system or POTW?
                             Can local TBLLs
                             for As and TDS
                              be met if IX is
                                used? *
               N
                    N
                                           Evaluate Centralized
                                              Ion Exchange
                                            (refer to Section 6).
                     Are you willing to install and operate
                     regenerant waste treatment facilities
                       (settling basins, decant, recycle,
                       mechanical dewatering, etc.) and
                     potentially deal with hazardous waste
                    permitting and environmental liability?
                            N
                                    Go to Tree 3b -
                                  'Sorption Processes'
* This evaluation will be complex because removal of As from drinking water will change the
background As concentration for the TBLL. A revised TBLL would be used to determine if the
brine stream could be discharged to the POTW.  TDS may be the more critical restriction,
especially in the western U.S.
                  Figure 3-8. Decision Tree 3a - Ion Exchange Processes.

           Use this decision tree only after using Tree 3 "SelectingNew Treatment."
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                         50

-------
                                  Tree  3b
                     Sorption  Processes
       Are all of the following
       water quality criteria met
       at the problematic source?
          • Cl- < 250 mg/L
          • F- < 2 mg/L
          • Silica < 30 mg/L
          • SO4-2 < 360 mg/L
          •TDS<  1,000 mg/L
          • TOC < 4 mg/L
N,
      IsPO4-3
-------
                                   Tree  3c
        Filtration  & Membrane  Processes
Is the source water
  pH 5.5-8.5?
        N
Is the source water:
  Fe> 15 mg/L,
 Mn > 15 mg/L, or
  H2S > 5 mg/L?
        Is the source water
          Fe:As Ratio >
             20:1?
                 N
                       Willing to install
                        pH adjustment
                         Equipment?
                                 Evaluate pH
                                 Adjustment
                         N
                               Is the source water
                              Fe:As Ratio > 20:1?
                                       N
                                                       Is service
                                                    population < 500?
                                                            N
                                                     Evaluate POU RO.
                                                      Refer to Section 8
                                                   Evaluate Fe/Mn
                                                 Oxidation/Filtration
                                              (refer to Sections 2.5.5 & 7).
                                                           Evaluate Fe/Mn
                                                        Oxidation/Filtration with
                                                        Iron Coagulant Addition
                                                       (refer to Sections 2.5.5 & 7).
                                               Evaluate Pre-Engineered
                                                Microfiltration (refer to
                                                 Section 2.5.3) or Pre-
                                              Engineered Direct Filtration
                                              (refer to Sections 2.5.4 & 7).
                                                        Evaluate Iron Coagulant
                                                          Addition with Pre-
                                                       Engineered Microfiltration
                                                        (refer to Section 2.5.3) or
                                                         Pre-Engineered Direct
                                                       Filtration (refer to Sections
                                                             2.5.4 & 7).
             Figure 3-10. Decision Tree 3c - Filtration and Membrane Processes.

           Use this decision tree only after using Tree 3 "SelectingNew Treatment."
Arsenic Treatment Technology Evaluation Handbook for Small Systems


-------
Table 3-3 provides a summary of information about the different alternatives for arsenic mitigation
found in this Handbook.

                Table 3-3. Arsenic Treatment Technologies  Summary Comparison.
                                                      (Iof2)

Factors

USEPA BAT B
USEPA SSCTB
System Size B-D
SSCT for POU B
POU System Size B-D
Removal Efficiency
Total Water Loss
Pre- Oxidation Required F




Optimal Water
Quality Conditions



Operator Skill Required
Waste Generated

Other Considerations

Centralized Cost
POU Cost
Sorption Processes
Ion Exchange
IX
Yes
Yes
25-10,000
No

95% E
1-2%
Yes



pH 6.5 - 9 E
< 50 m&L SO42' '
< 500 mg/L TDS K
< 0.3 NTU Turbidity


High
Spent Resin, Spent Brine,
Backwash Water
Possible pre & post pH
adjustment.
Pre-filtration required.
Potentially hazardous brine
waste.
Nitrate peaking
Carbonate peaking affects pH.
Medium
-
Activated Alumina A
AA
Yes
Yes
25-10,000
Yes
25-10,000
95% E
1-2%
Yes
pH5.5 - 6 '
pH6- 8.3 L
< 250 rng^ C1- '
< 2 mg/L F- :
< 360 mg/L S042'K
< 30 mg^ Silica M
< 0.5 mg/L Fe+31
< 0.05 mg/L Mn+2 :
< 1,000 mg^ TDS K
< 0.3 NTU Turbidity
Low A
Spent Media, Backwash
Water

Possible pre & post pH
adjustment.
Pre-filtration may be
required.
Modified AA available.

Medium
Medium
Iron Based
Sorbents
IBS
No c
Noc
25-10,000
No c
25-10,000
up to 98% E
1- 2%
Yes ฐ




pH6- 8.5
< 0.3 NTU Turbidity



Low
Spent Media, Backwash
Water

Media may be very
expensive. ฐ
Pre-filtration may be
required.

Medium
Medium
Membrane
Processes
Reverse
Osmosis
RO
Yes
Yes
501-10,000
Yes
25 -10,000
> 95% E
15-75%
Likely11




No Particulates



Medium
Reject Water

High water loss (15-
75% of feed water)

High
Medium
A Activated alumina is assumed to operate in a non-regenerated mode.
B USEPA, 2002a.
c IBS's track record in the US was not established enough to be considered as Best Available Technology (BAT) or Small System Compliance
  Technology (SSCT) at the time the rule was promulgated.
D Affordable for systems with the given number of people served.
E USEPA, 2000.
F Pre-oxidation only required for As(III).
0 Some iron based sorbents may catalyze the As(III) to As(V) oxidation and therefore would not require a pre-oxidation step.
H RO will remove As(III), but its efficiency is not consistent and  pre-oxidation will increase removal efficiency.
1 AwwaRF, 2002.
1 Kempic, 2002.
K Wang, 2000.
L AA can be used economically at higher pHs, but with a significant decrease in the capacity of the media.
M Clifford, 2001.
N Tumalo, 2002.
0 With increased domestic use, IBS cost will significantly decrease.
Arsenic Treatment Technology Evaluation Handbook for Small Systems


-------
                Table  3-3.  Arsenic Treatment Technologies  Summary Comparison.
                                                    (2 of 2)
                                                         Precipitative Processes
Factors
USEPA BAT B
USEPA SSCT B
System Size B-D
SSCT for POU B
POU System Size B-D
Removal Efficiency
Total Water Loss
Pre- Oxidation Required F
Optimal Water
Quality Conditions
Operator Skill Required
Waste Generated
Other Considerations
Centralized Cost
POU Cost
Enhanced Lime
Softening
LS
Yes
No
25-10,000
No

90% E
0%
Yes
pH 10.5 - 11 :
> 5 mg/LFe+31
High
Backwash Water,
Sludge (high volume)
Treated water requires pH
adjustment.
Low1:!
N/A
Enhanced
(Conventional)
Coagulation
Filtration
CF
Yes
No
25-10,000
No

95% (w/ FeCy E
< 90% (w/ Alum) E
0%
Yes
pH5.5 - 8.5 p
High
Backwash Water,
Sludge
Possible pre & post
pH adjustment.
Low<3
N/A
Coagulation-
Assisted
Micro-
Filtration
CMF
No
Yes
500-10,000
No

90% E
5%
Yes
pH5.5 - 8.5 p
High
Backwash Water,
Sludge
Possible pre &
post pH
adjustment.
High
N/A
Coagulation-
Assisted Direct
Filtration
CADF
Yes
Yes
500-10,000
No

90% E
1-2%
Yes
pH5.5 - 8.5 p
High
Backwash Water,
Sludge
Possible pre & post
pH adjustment.
Medium
N/A
Oxidation
Filtration
OxFilt
Yes
Yes
25-10,000
No

50-90% E
1-2%
Yes
pH5.5 - 8.5
>0.3 mg/L Fe
Fe:As Ratio > 20:1
Medium
Backwash Water.
Sludge
None.
Medium
N/A
 B USEPA, 2002a.
 D Affordable for systems with the given number of people served.
 E Depends on arsenic and iron concentrations.
 F Pre-oxidation only required for As(III).
 1 AwwaRF, 2002.
 p Fields, et aL, 2002a.
 Q Costs for enhanced LS and enhanced CF are based on modification of an exisitng technology. Most small systems will not have this technology in
 place.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
54

-------
                                                                  Section  4
                    Planning-Level Treatment Costs
This section presents information the utility can use to calculate planning-level capital and O&M
costs for the treatment method selected in Section 3. All the charts are from Technologies and
Costs for Removal of Arsenic from Drinking Water (USEPA, 2000). This information will give the
utility only a rough estimate of the selected treatment process costs so that relative costs can be
evaluated. If the costs are too high, the utility is encouraged to re-evaluate the criteria used in the
treatment selection process in Section 3.

It is critical that the utility employ the decision trees in Section 3, rather than the cost corre-
lation curves provided in this section, as the primary tool for selecting an arsenic mitigation
strategy. The trees take into account system-specific conditions and utility preferences.  Compar-
ing planning-level costs without consideration of the technical issues incorporated in the decision
trees may lead the utility to an inappropriate technology.

The cost curves incorporate different mathematical models for different sized systems. Because of
this, there are step changes between the model outputs in some of the charts. If the system being
sized falls at a flowrate that lays on one of these step  changes, the utility is encouraged  to use an
average cost number and then perform a more site specific cost evaluation.

Capital cost charts are based on the maximum flowrate for which the facility was designed (i.e.,
design flowrate). The design flowrate should be higher than the treated flowrate determined in
Section 3. The capital costs include: process costs (including manufactured equipment,  concrete,
steel, electrical and instrumentation, and pipe and valves), construction costs (including site-work
and excavation, subsurface considerations, standby power, contingencies, and interest during con-
struction), engineering costs (including general contractor overhead and profit, engineering fees,
and legal, fiscal, and administrative fees) and the costs associated with retrofitting, permitting, pilot
testing, housing, and system redundancy (where prudent). The capital costs do not include costs
associated with additional contaminants or land.

The O&M costs are based on the average flowrate that the facility is expected to treat.  The O&M
costs are based on the following assumptions:
•   Electricity costs of $0.08/kWh,
•   Diesel fuel costs of $1.25/gallon,
•   Natural gas costs of $0.006/scf,
•   Large systems labor costs of $40/h (or loaded labor costs of $52/h),
•   Loaded labor costs for small systems of $28/h, and
•   Building energy use of 102.6 kWh/sft/y.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                               55

-------
All of the costs presented in the charts are given in year 1998 dollars. To escalate the capital cost of
a technology from 1998 to the present, construction cost indexes (CCI) published by Engineering
News Record (ENR) can be used in the following formula.

                            p      _p    (CCI Current |
                            ^Current - F1998   „„                                     Eqn. 4-1
                                         ^ CU1998  J
Where:
PCurrent      = PreSent COSt,
Pj™      = Year 1 998 Cost (from the charts),
CCL      = Construction Cost Index Value for the current year, and
    Current                                               J
CCI1998    = Construction Cost Index Value for 1998.

For example, a greensand filtration system is designed to handle 1 MGD.  Values taken from the
figures and their equations are:
•   1998 Capital Cost is $587,584 (Figure 4-21)
•   1998 Waste Disposal Capital Cost is $3,955 (Figure 4-23)

The annual average 20-cities ENR CCI for 1998 and for November 2002 are 5,920.44 and 6,578.03,
respectively. Therefore,  the total capital cost for this facility can  be estimated for the year 2002
using Equation 4-1 as follows:
                     PcapiH2002 =($587,584 + $3,955)         = $657,242
The O&M costs presented in the Handbook are 1 998 costs and can be escalated to the current year's
costs using the formula below. This formula can also be used in place of the CCI equation, al-
though it is less accurate.


                            PCurrent = Pl998 (l + i/Yฐ"~-1998)                            Eqn. 4-2
Where:
    Pr    = Current Cost,
     Current
    P1998   = Year 1998 Cost (from the charts),
    i      = Annual rate of inflation (currently ~ 2.5% - 3%), and
    Y^    = Current Year.
      Current

Using the same example of a 1 MGD greensand filtration system, the values taken from the figures
and their equations are:
•   1998 O&M cost is $66,3 14/y (Figure 4-22)
•   1998 Waste Disposal O&M cost is $8,678/y (Figure 4-24)
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                 56

-------
Assuming an annual inflation rate of 2.5%, the O&M can be estimated for the year 2002 using
Equation 4-2 as follows:
                  O&M, 200:
                          = ($66,314 + $8,678) (l + 0.025/2002 1998) = $82,777
Additionally, the USEPA has a cost model for estimating the costs of processes using sorptive
media. This cost model, titled Cost Estimating Program for Arsenic Removal by Small Drinking
Water Facilities, can be found on the USEPA's website.

4.1    Pre-Oxidation System Costs Using Chlorine

Costs presented in the following charts make the following assumptions:
•  A new chlorination system is installed.
•  A dose of 1.5 mg/L of free chlorine is added to the treated flow.
•  Systems use 15% sodium hypochlorite feed stock and are designed to handle dosages as high as
   10 mg/L.

t/3
o
U $10 000
s
ft
a
u
$1,000
0.




































































y = i


y



01


































4,560







^^


7,277









-^ —




0.1






























































•





































y ฃ



* *
s
fT
^- 	





1





,065.8x + 8,494.2
\

.--
/








X
*^
\

^ ^^
w^









),875x + 402



































— f
*










1
                                      Design Flowrate (mgd)
                                       w/o Housing •
w/ Housing
                          Figure 4-1.  Chlorination Capital Costs.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                             57

-------
$10,000
$1,000
0.(


















































^^M
















^^


















































































































































































































y=-lE-10x2+4,239.7x+ 1,161.8

^
M
s
•
s
•
L—
^^^M
^^
^^

*•



)01 0.01 0.1

^















X
















X




y









X







































-0.3928x"+823.4x+ 19,36

















^
' i




*
s
s
s
?

\


/









/







y = 4,920.1x + 916.9








M










\












5














0

--•










1 1
                                      Average Flowrate (mgd)

                          Figure 4-2. Chlorination O&M Costs.

4.2   Ion  Exchange System Costs

Costs presented in the following charts make the following assumptions:
•   A new IX system is installed.
•   Capital Cost Design Assumptions:
    O  Pre-oxidation is required but not included in these costs.
    O  Cost includes a redundant column to allow the system to operate during regeneration.
•   O&M Cost Design Assumptions:
    O  Run length when sulfate is at or below 20 mg/L is 1500 bed volumes (BV).
    O  Run length when the sulfate is between 20 and  50 mg/L is 700 BV.
    O  Labor rate for small systems is $28/hour.  The  loaded labor rate for large  systems is $52/
       hour.
•   Waste is discharged to a POTW (i.e., indirect discharge).
Arsenic Treatment Technology Evaluation Handbook for Small Systems
58

-------
~ $1,000,000
o
U
"3
'a
U $100,000
$10,000
0.

















	 '

01













\
j
^^^















u=













19
^













M
7














6
.^















	 i































=







































































y = -6,278.7x2+ 319,169x + 36,758






.^




0.1

















^































































?






\
	 J^-3
..^^^^
__^^^~




















1






^
















^
















^

















































^ '











1
                                           Design Flowrate (mgd)
                   Figure 4-3.  Ion Exchange (<20 mg/L SO42-) Capital Costs.
   $100,000
o
U
ฐ   $10,000
     $1,000
          ).001
                     y = -67,837x2 + 77,544x + 4,997.4
                                                                y = -I,695.5x2+ 37,350x + 16,532
0.01                  0.1                   1
            Average Flowrate (mgd)
10
                   Figure 4-4. Ion Exchange (<20 mg/L SO42-) O&M Costs.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                              59

-------
  o
 U
  cs
 U
$1,000
0.










































































y = 3,955














---H



01 0.1





k.













































































y- 5,C




1








85








--•








-•s








k.













1
                                         Design Flowrate (mgd)
           Figure 4-5. Ion Exchange (<20 mg/L SO 2) Waste Disposal Capital Costs.
  $100,000
  ,$10,000
^
 o
U
   $1,000
                                       y = 3,429x+375
                             ).01
                                        Average Flowrate (mgd)
10
           Figure 4-6. Ion Exchange (<20 mg/L SO 2) Waste Disposal O&M Costs.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
60

-------
O
U
$1,000,000
$100,000
$10,000
0.















. ,

\
==^





























































t,220x+ 19,546

. —





w^





















































































f
















y -y





^






01 0.1































































,756.1x2+ 373, 847x + 73,320




_r*i*
^^










. — '







\













ฃ
A











^












































^g




^^^^^










1






4*









































^


















































































^ ""'















1
                                           Design Flowrate (mgd)
                  Figure 4-7.  Ion Exchange (20-50 mg/L SO 2) Capital Costs.
 $1,000,000
   $100,000
O
U
    $10,000
     $1,000
          ).001
y = -82,812x2+ 94,848x + 8,449.9
                                                                 y = -1,843.Ix2+ 48,269x + 24,614
    0.01                  0.1
               Average Flowrate (mgd)
10
                  Figure 4-8. Ion Exchange (20-50 mg/L SO42-) O&M Costs.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                                 61

-------
 o
U


1
'ft

U
   $1,000
                                  = 3,955
                                                                             = 21.47x + 5,198
        0.01
       0.1                          1

            Design Flowrate (mgd)
10
          Figure 4-9.  Ion Exchange (20-50 mg/L SO 2) Waste Disposal Capital Costs.
  $100,000
   $10,000
    $1,000
     $100
      $10
        0.001
                                              y = 7,348x
                                             ฑ
0.01
         0.1                   1

Average Flowrate (mgd)
10
          Figure 4-10. Ion Exchange (20-50 mg/L SO/) Waste Disposal O&M Costs.
^4rseซ/'c Treatment Technology Evaluation Handbook for Small Systems
                                                              62

-------
4.3    Activated Alumina System Costs

Costs presented in the following charts make the following assumptions:
•  A new AA system is installed.
•  AA media is disposed of in a non-hazardous landfill rather than regenerated.
•  Four treatment modes are assumed:
   O  No pH adjustment, Natural pH of 7-8, run length is 10,000 BY
   O  No pH adjustment, Natural pH of 8-8.3, run length is 5,200 BY
   O  pH adjusted to 6.0 using hydrochloric acid, run length is 23,100 BY
   O  pH adjusted to 6.0 using sulfuric acid, columns are good for 15,400 BY
•  Capital Cost Design Assumptions:
   O  Redundant column included for operation during media replacement.
   O  Costs for constructing housing for the equipment are included.
   O  Capital costs include both pre- and post-treatment pH adjustment if pH adjustment is used.
•  O&M Cost Design Assumptions:
   O  Power costs are $0.08/kwh.
   O  pH adjustment costs are included.
   O  Labor rate for small systems is $28/hour.  The loaded labor rate for large systems is $527
       hour.
J1U,UUU,UUU
^$1,000,000
-M
ซ:
O
U
"3
^
'ft
&
u $100,000

$10,000
0.
















































































y-515,309x+ 10,214



^^
01
[

J^



^^


**








































































^















^










y















= 4



ft











2(


.












V















8
L














6















x+ 10,21:

^^













I ^















S















<*•





























































s












0.1 1 1
                                       Design Flowrate (mgd)
                Figure 4-11. Activated Alumina (Natural pH) Capital Costs.
     /'c Treatment Technology Evaluation Handbook for Small Systems
63

-------
  $10,000,000
   $1,000,000
    $100,000
     $10,000
      $1,000
           0.001
                           y=188,890x +4,122.7
                                                                   y=190,726x+9,798.2
0.01                  0.1                  1
           Average Flowrate (mgd)
10
               Figure 4-12. Activated Alumina (Natural pH of 7-8) O&M Costs.
  $100,000
   $10,000
    $1,000
     $100
                                             y = 7,568x
                              ).01
                                         Average Flowrate (mgd)
10
        Figure 4-13.  Activated Alumina (Natural pH of 7-8) Waste Disposal O&M Costs.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
64

-------
 $10,000,000
  $1,000,000
3   $100,000


o

     $10,000
      $1,000
          0.001
 0.01                 0.1
            Average Flowrate (mgd)
10
              Figure 4-14. Activated Alumina (Natural pH of 8-8.3) O&M Costs.
 $1,000,000
   $100,000
    $10,000
o
u
     $1,000
      $100
       $10
         0.001
                                           y=14,555x
0.01                  0.1
            Average Flowrate (mgd)
10
      Figure 4-15.  Activated Alumina (Natural pH of 8.0-8.3) Waste Disposal O&M Costs.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                              65

-------
^$1,000,000
0
u
1
'ft
u $100000
$10,000
0.










































































y = 555,826x + 42,103
(


•



01

i^




ป-ซ





^





—






^



















^_























































y-4


^







0.1



































50,803x + 42,101
\
\
^








—









































































































^^^^
f

















^



















S


















^



















s

























































^
















1 1
                                           Design Flowrate (mgd)
              Figure 4-16. Activated Alumina (pH Adjusted to 6.0) Capital Costs.
  $10,000,000
   $1,000,000
    $100,000
     $10,000
      $1,000
           0.001
0.01                 0.1
           Average Flowrate (mgd)
10
         Figure 4-17.  Activated Alumina (pH adjusted to 6.0 - 23,100 BV) O&M Costs.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                            66

-------
 $100,000
  $10,000
   $1,000
o
u
     $100
      $10
       $1
        0.001
                                           y = 3,276x
0.01
         0.1
Average Flowrate (mgd)
10
 Figure 4-18. Activated Alumina (pH adjusted to 6.0 - 23,100 BV) Waste Disposal O&M Costs.
 $10,000,000
  $1,000,000
    $100,000
     $10,000
      $1,000
          0.001
                 ,368x +6,325.7
                                                                    y = 206,113x+13,092
  0.01                 0.1
             Average Flowrate (mgd)
                                                   10
         Figure 4-19.  Activated Alumina (pH adjusted to 6.0 - 15,400 BV) O&M Costs.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                               67

-------
  $100,000
  $10,000
^
   $1,000
    $100
                                            - = 4,915x
                           ).01
                                     Average Flowrate (mgd)
10
 Figure 4-20.  Activated Alumina (pH adjusted to 6.0 - 15,400 BV) Waste Disposal O&M Costs.
4.4   Iron Based  Sorbent System Costs

IBS are relatively new technologies and, as such, costs for IBS treatment systems have not yet been
developed.

4.5   Greensand  System  Costs

Costs presented in the following charts make the following assumptions:
•   A new greensand filtration system is installed.
•   Potassium permanganate feed rate of 10 mg/L (however, chlorination will work also).
•   Hydraulic loading rate of 4 gpm/sft.
•   Backwash flowrate of 10-12 gpm/sft.
•   Backwash waste is  discharged to a POTW (i.e., indirect discharge).
^4rseซ/'c Treatment Technology Evaluation Handbook for Small Systems
68

-------
o
U
$1,000,000
$100000

$10,000
0.


















^

















^

















^^












y












































































587,584 X0'838




























\^



^






01 0.1












S*

















^

















	 1


















^













































































































^


















^

















_i

























































































1 1
                                            Design Flowrate (mgd)
                             Figure 4-21. Greensand Capital Costs.
 $1,000,000
   $100,000
o
u
    $10,000
     $1,000
          ).001
                                   y = 0.0009x2+ 58,921x + 7,392.9
                               0.01                  0.1                    1
                                           Average Flowrate (mgd)
10
                             Figure 4-22.  Greensand O&M Costs.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                                                              69

-------
 O
U

I
'5,
u
   $1,000
        0.01
                                                      \
              0.1
     Design Flowrate (mgd)
                     Figure 4-23.  Greensand Waste Disposal Capital Costs.
   $10,000
                                                                         y = 7,548x+1,130
    $1,000
     $100
         0.001
0.01                         0.1
     Average Flowrate (mgd)
                     Figure 4-24.  Greensand Waste Disposal O&M Costs.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                        70

-------
4.6    Coagulation Assisted Microfiltration System Costs

Costs presented in the following charts make the following assumptions:
•  Ferric chloride dose of 25 mg/L.
•  For Systems Less Than 1 MGD:
   O  Package plants with a hydraulic loading rate of 5 gpm/sft.
   O  Sodium hydroxide dose of 20 mg/L for pH control.
   O  Standard MF.
•  For Systems Larger Than 1 MGD:
   O  Rapid mix for 1 minute.
   O  Flocculation for 20 minutes.
   O  Sedimentation at 1000 gpd/sft using rectangular tanks.
   O  Standard MF.
•  Waste is dewatered before being disposed of in a non-hazardous landfill. Costs are given for
   dewatering performed either mechanically or non-mechanically. Land costs are not included in
   the waste disposal costs.
J>1UU,UUU,UUU
,$10,000,000
*s
•i
5
5
)
ซ
t
•i
s
•* $1,000,000
v — 11 Q^S 4
y -ii,yj:>,4
ft nn nnn


















































































































































































y = -483,591x2+ 2,308,991x + 273,14










y = 2,343, 199x + 228,653
65^+ 4,880,036x + 94,32

^^
\
4

— •










y
















#



— ^c^*
^^^




^=—






,^—







^







^






\.
^







^
































































y=l,030,810x+ 1,067,733






^
^^











^^








— :


+**










































































           0.01
0.1                       1
    Design Flowrate (mgd)
10
              Figure 4-25. Coagulation Assisted Microfiltration Capital Costs.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                 71

-------
  $1,000,000
   $100,000
    $10,000
          0.001
0.01                  0.1
            Average Flowrate (mgd)
10
                Figure 4-26.  Coagulation Assisted Microfiltration O&M Costs.
 $10,000,000
  $1,000,000
 o
U
 a $100,000
 C3
    $10,000













y = -


_ ~~














692,7
i
















93x:

















+ 4

















55

















,3

















2S





























X4






















26,748


















































v-


^










































































y = -I,644.1x2+ 150,402x + 157,575 '



= 211,627x +


^





\








ซ
>































42,043 :

^S^
^^









*










&~
^^




















\



































^











X*











          0.01
        0.1                         1
            Design Flowrate (mgd)
10
 Figure 4-27. Coagulation Assisted Microfiltration (w/ Mechanical Dewatering) Waste Disposal
                                        Capital Costs.
^4rseซ/'c Treatment Technology Evaluation Handbook for Small Systems
                                                              72

-------
  $1,000,000
   $100,000
o
u
    $10,000
     $1,000
                y = -3,476,860x4- 684,838x + 5,839
                                \
                                                               y = -359.7x2
y = 25,165x +36,857
                                                                   S
                              0.01                 0.1                   1

                                         Average Flowrate (mgd)
                                     10
 Figure 4-28.  Coagulation Assisted Microfiltration (w/ Mechanical Dewatering) Waste Disposal

                                        O&M Costs.
$1UU,UUU,UUU
$10,000,000
&
-M
Oft
u $1,000,000
^^
a.
ซ
U
$100,000
$10,000
0.








	 	 '
" y- 3,06
\
	 ^
^^^^










9,36C













X-'
_r^












790













r

>











































* ^














> 1








01 0.1





,749,





352)
i
S
















c +













10













8,













01













7















x




























y= 1,627,970x4- 326,504


^*^








1











\
^=^
^^^^





















































X





















1
                                          Design Flowrate (mgd)
   Figure 4-29. Coagulation Assisted Microfiltration (w/ NonMechanical Dewatering) Waste

                                   Disposal Capital Costs.
Arsenic Treatment Technology Evaluation Handbook for Small Systems


-------
  $1,000,000
    $100,000
 o
u
    $10,000
     $1,000





































y =


































































































































18,812x2+4,686x + 2,124
^^H







• •

• ^^^^^^^^

^^^

^


V




















J



















b



















— •
























17 1
y A








6.966x2+ 60,792x + 28,76(




y=lll,819x- 6,950 :










/







/
— ^-


f







/
/








\
I
>






















^
~ฃ_














x^














\
M
X*








































































































          0.001
0.01                0.1
          Average Flowrate (mgd)
10
   Figure 4-30. Coagulation Assisted Microfiltration (w/ NonMechanical Dewatering) Waste
                                 Disposal O&M Costs.

4.7    Coagulation/Filtration System Enhancement Costs

Costs presented in the following charts make the following assumptions:
•  A coagulation/filtration system is already installed. Costs are only for system enhancement for
   arsenic removal.
•  Assumptions about the Existing Coagulation/Filtration System:
   O  Existing coagulation/filtration system removes 50% of the arsenic without enhancement.
   O  Ferric chloride dose of 25 mg/L.
   O  Polymer dose of 2 mg/L.
   O  Lime dose of 25 mg/L for pH control.
   O  Systems less than 1 MGD are package plants with a hydraulic loading rate of 5 gpm/sft.
   O  Systems Larger Than 1 MGD:
       •  Rapid mix for  1 minute.
       •  Flocculation for 20 minutes.
       •  Sedimentation at 1000 gpd/sft using rectangular tanks.
       •  Dual media gravity filters running at a hydraulic loading rate of 5 gpm/sft.
•  Assumptions for the Enhancement of the Coagulation/Filtration System:
   O  Additional ferric chloride dose of 10 mg/L.
   O  Additional feed system for increased ferric chloride dose.
   O  Additional lime dose of 10 mg/L for pH adjustment.
   O  Additional feed system for increased lime dose.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                        74

-------
^ $100000
0
u
"a
5.
^ $10000
$1,000
0.














:^^^^^^^Bปป








































































































































y =-3,825.4x2+ 14,734x + 7,144.5

•••^^







^^••i






—















-































^^^^^^







M^

	 1







01 0.1













^




















•*••









































•ซ




















—
























































	

f
/
/
/






/
J^ 	
\
\





/












































-b









y = 94,000x - 75,947 "









1































































1
                                        Design Flowrate (mgd)
           Figure 4-31.  Coagulation/Filtration System Enhancement Capital Costs.
$1,UUU,UUU
$100,000
r— \
-<—
2 $10,000
ง
0
$1,000
$100
0.











	 — •














^z



















P


01










,— -
y












=














ฃ










ซ










=










^










, — - —









! — ^
X









X^"

502.3x2 + 7,299x + 220.8






















^ซ















.













































































_^>^r
s^













-^^~
y=l














7,4/













— f
9x













S
-3














,4














33














.3








0.1 1 1
                                        Average Flowrate (mgd)
 Figure 4-32. Coagulation/Filtration System Enhancement O&M Costs.Lime Softening System
                                   Enhancement Costs
Arsenic Treatment Technology Evaluation Handbook for Small Systems
15

-------
4.8   Lime Softening System Enhancement Costs

Costs presented in the following charts make the following assumptions:
•   An LS system is already installed. Costs are only for system enhancement for arsenic removal.
•   Lime dosage of 250 mg/L.
•   Carbon dioxide dosage of 35 mg/L for recarbonation.
•   Assumptions about the Existing LS System:
    O  Systems less than 1 MOD are package plants.
    O  Systems Larger Than 1 MGD:
       •  Rapid mix for 1 minute.
       •  Flocculation for 20 minutes.
       •  Sedimentation at 1500 gpd/sft using circular tanks.
       •  Dual media gravity filters running at a hydraulic loading rate of 5 gpm/sft.
•   Assumptions for the Enhancement of Existing LS System:
    O  Additional lime dose of 50 mg/L.
    O  Additional feed system for increased LS dose.
    O  Additional carbon dioxide dose of 35 mg/L for recarbonation.
    O  Additional feed system for increased carbon dioxide dose.
   $10,000,000
    $1,000,000
^    $100,000
"3
-^
'a
U
      $10,000
       $1,000
            0.01
                      y =-22,974x2+ 48,21 Ix + 7,897.7
                                                                  y=133,486x-100,351
                                    0.1                      1
                                       Design Flowrate (mgd)
10
                 Figure 4-33. Lime Softening Enhancement Capital Costs.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                                                    76

-------
$1,UUU,UUU
$100,000
^
^
-M
ฃ $10,000
ง
=3
0
$1,000
$100
0.







v=2 1<
j ^i1













14 7x


	











+ ^(
\
"
*-^-











) 5'

^











>5^

_











I 4













- 1

^











I

^











Q














'/


L^^






01 0.1





































*


























y =


•o"1











2














9,










>

I










A A COO-.,-
— U.4jZox

4











5x +1,094
^^









•+

3
26,067x+14,95(

_,










i

^










-N
^










"X,*
2



































}

+ *











1
                                        Average Flowrate (mgd)

                  Figure 4-34. Lime Softening Enhancement O&M Costs.

4.9    Point-of-Use Reverse Osmosis System Costs

Costs presented in the following charts make the following assumptions:
•  In an average household, there are 3 individuals using 0.53 gallon each per day for a total of 579
   gallons per year.10
•  Life of POU unit is 5 years.
•  Duration of cost study is 10 years.
•  Cost of water meter and automatic shut-off valve included.
•  No shipping and handling included.
•  If the water is chlorinated, dechlorination may be required. Costs for dechlorination are not
   included.
•  Volume discount schedule—retail for a single unit, 10 percent discount for 10 or more units, 15
   percent discount on more than 100 units.
•  Installation time—1 hour unskilled labor (POU)
•  Minimally skilled labor—$14.50 per hour (population less than 3,300 individuals).
•  Skilled labor—$28 per hour (population greater than 3,300 individuals).
•  O&M costs include maintenance, replacement of pre-filters and membrane cartridges, labora-
   tory sampling and analysis, and administrative costs.
 1 USEPA, 1998.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
77

-------
   $10,000,000-
    $1,000,000-
 o
u
 =.
 cs
u   $100,000'
      $10,000
             10
                                      = 864.7x
        0.9261
 100                        1000
          Households
10000
                       Figure 4-35. POU Reverse Osmosis Capital Costs.
   $1,000,000-
    $100,000
 e
u
     $10,000
      $1,000
            10
                                     = 266.9x
                                             0.9439
100                       1000
          Households
10000
                        Figure 4-36.  POU Reverse Osmosis O&M Costs.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                       78

-------
4.10  Point-of-Use Activated Alumina System Costs

Costs presented in the following charts make the following assumptions:
•  In an average household, there are 3 individuals using 0.53 gallon each per day for a total of 579
   gallons per year. n
•  Life of POU unit is 5 years.
•  Duration of cost study is 10 years.
•  Cost of water meter and automatic shut-off valve included.
•  No shipping and handling included.
•  Volume discount schedule—retail for a single unit, 10 percent discount for 10 or more units, 15
   percent discount on more than 100 units.
•  Installation time—1 hour unskilled labor (POU)
•  Minimally skilled labor—$14.50 per hour (population less than 3,300 individuals).
•  Skilled labor—$28 per hour (population greater than 3,300 individuals).
•  O&M costs include maintenance, replacement of pre-filters and membrane cartridges, labora-
   tory sampling and analysis, and administrative costs.
  $1,000,000-
    $100,000
o
u
=.
a
U
     $10,000
      $1,000
= 296.9x
                                      0.9257
           10
    100                     1000
             Households
10000
                    Figure 4-37. POU Activated Alumina Capital Costs.
  USEPA, 1998.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                       79

-------
J>1U,UUU,UUU
$1,000,000-
S
-4^
K
u $100,000-
^
ซ2
o
$10,000-
ซ 1 nnn -








































^-








































y

_:




















^





- 1

*











n













. 0.937
2.6x
~4^>
*S^












5
















-^





























ซ^














^









































	 ^^
^^^^~
. **^












^^




































































































           10
100                    1000
        Households
10000
                    Figure 4-38. POU Activated Alumina O&M Costs.
4.11  Point-of-Use Iron Based Sorbent System Costs

Iron based sorbents (IBS) are relatively new technologies and, as such, the costs for using small IBS
units in a POU scheme have not been well defined. Costs for an IBS POU system are anticipated to
be similar to those of an AA POU system.
^4rseซ/'c Treatment Technology Evaluation Handbook for Small Systems
                                                80

-------
                                                                Section 5
        Pre-Oxidation  Design  Considerations
The conversion of reduced inorganic As(III) to As(V) is critical for achieving optimal performance
of all unit processes described in this Handbook. Conversion to As(V) can be accomplished by
providing an oxidizing agent at the head of any proposed arsenic removal process. Chlorine and
permanganate are highly effective for this purpose. They oxidize As(III) to As(V) within one
minute in the pH range of 6.3 to 8.3. Ozone rapidly oxidizes As(III) but its effectiveness is signifi-
cantly diminished by the presence of sulfides or TOC. Solid phase oxidants such as Filox-R™ have
also been shown to oxidize As(III). Chlorine dioxide and monochloramine are ineffective in oxi-
dizing As(III). UV light, by itself is also ineffective. However, if the water is spiked with sulfite,
UV photo-oxidation shows promise. Because of these considerations, only chlorine, permangan-
ate, ozone, and solid phase oxidants are discussed in this section.

5.1   Chlorine Pre-Oxidation Design Considerations

The primary applications of chlorine in water treatment include pre-oxidation, primary disinfec-
tion, and secondary disinfection.  Several arsenic removal processes, particularly membranes, are
chlorine sensitive and/or intolerant. In these instances, the utility should consider an alternate
oxidation technology.  If this is the case, but the system already has chlorination capabilities in
place, the process of modifying the existing system to achieve As(III) oxidation is complicated.
One alternative is the application of a pre-chlorination—dechlorination—arsenic removal—re-chlo-
rination treatment setup. However, this alternative may be more costly than integrating a perman-
ganate pre-oxidation system.

Chlorine can be added either as liquid sodium hypochlorite (Equation 5-1) or dissolved gas (Equa-
tion 5-2). In either case, biocidal hypochlorous acid is generated.

                         NaOCl + H2O -ป HOC1 + NaOH                     Eqn. 5-1

                             C12 + H2O -ป HOC1 + HC1                        Eqn. 5-2

The first step in selecting the most appropriate method of chlorination is to determine the chlorine
flow requirements for the particular application. Chlorine demand can be calculated with Equation
5-3.
                                             T~  ~T"                      Eqn. 5-3
                                              d   gal mg I
                                       V                  /
where:
    Mcl2   = Chlorine Mass Flow (Ib/day of C12),



Arsenic Treatment Technology Evaluation Handbook for Small Systems                              81

-------
    Q     = Design Flow Rate (gpm), and
    5cl2   = Chlorine Dose (mg/L as C12).

Careful consideration should be given to the chlorine dose estimate. Most waters contain sub-
stances other than As(III) that exert chlorine demand. In many cases, these substances compete for
chlorine more aggressively than As(III).  Section 2.2.1 lists the chlorine demand for the stoichio-
metric conversion of As(III), Fe2+, Mn2+, and HS~. Chlorine will also react with ammonia and TOC.
Simple chlorine demand bench testing can be used to ascertain the instantaneous and ultimate
chlorine demand of particular water.  The applied chlorine dose should be three times the ultimate
chlorine demand.

                                     5cl2 =3-Dcl2                                 Eqn. 5-4
Where:
    5cl2   = Chlorine Dose (mg/L as C12), and
    Dcl2   = Ultimate Chlorine Demand (mg/L as C12).

Selection of the type of chlorination system should include consideration of capital and operating
costs, O&M requirements, code restrictions, containment requirements, footprint, and safety con-
cerns.  This Handbook will address the following options, which are considered most viable for
small water systems:

•   Commercial liquid hypochlorite feed system
•   On-site hypochlorite generation system

The application of chlorine gas for chlorination is not discussed as it is more hazardous, frequently
more expensive, and frequently less applicable to small  systems.

       5.1.1   Commercial Liquid Hypochlorite
       Liquid sodium hypochlorite can be purchased as a 5Vi% or 12V2%  strength solution. The
       solution must be delivered to the  facility by tanker trucks or in drums on a regular basis.
       The solution is stored on-site in a tank and metered into the system by a small pump. Figure
       5-1 shows a flow diagram for a typical liquid hypochlorite process.  Figure 5-2 is a typical
       flow schematic for a flooded suction hypochlorite metering system.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                82

-------
12-15%
Hypochlorite
Storage Tank
Raw
Water
1 >Lb
Metering
Pump
V
^ ^^^^^^^ Chlorinated
Figure 5-1 . Typical Liquid Hypochlorite Process Flow Diagram.
MAIN
f CONNECTION
PROCESS ^^^ /
LINE " ^J -_ V
i- SUPPLYTANK
1 ^ f
i !PY^
SUCTION
SHUT-OFF VALVE
T STRAINER
Figure 5-2. Liquid Hype
BACKPRESSURE /~ W^-^ /
VALVE A f ^ VALVE
PRESSURE REUEF V-,-, /
VALVE AT .rt~t><^~b.
"l A Tlo y TO ORAIH
TvALVE - T OR SUPPLY
UNION -yf
T^ 15"HIN —1
1 Ml IS-fl 1

KEEP TOMB. 1
DISTANCE DISCHARGE J
DRAIN VALVE
i XT
•' 1 (
UNION — L_J 1 I


OR SUPPLY \-ฑ_J-
ii
CALIBRATION W
CHAMBER A B ^ฃ 1
ป- UNION
O
o-q m | ^
WALWE ' / X^ DRAIN PLUG
UNION -f OR VALVE
^chlorite System Schematic (USFilter, Wallace & Tiernan).
Arsenic Treatment Technology Evaluation Handbook for Small Systems
83

-------
       The flow rate of liquid hypochlorite required to meet chlorine mass flow requirements can
       be approximated by Equation 5-5. This flow rate should be used to size the metering pump,
       as well as provide an estimate of chemical operating costs.
                                         MC1
       Where:
           Qci2   = Hypochlorite Metering Pump Rate (gph) and
           Mci2   = Chlorine Mass Flow (Ib/day of C12).
           CCi2   = Concentration of Chlorine Solution (Ibs Cl2/gal).
                    For 12.5 wt% Sodium Hypochlorite, the concentration is 1.26 Ibs/gal.
                    For 5.25 wt% Sodium Hypochlorite, the concentration is 0.47 Ibs/gal.
                    For 0.008 wt% Sodium Hypochlorite, the concentration is 0.068 Ibs/gal.

       The required capacity of the storage tank is contingent upon the desired frequency of tanker
       truck deliveries.   Tanks are commonly sized to provide 7-21 days of storage. Because
       commercial strength liquid hypochlorite is a Class 1 Liquid Oxidizer, storage of more than
       4,000 pounds represents a non-exempt quantity and requires special precautions. The stor-
       age volume required may be calculated as follows.
                                                                                Eqn
                                                  day J
       Where:
           V     = Storage Volume (gal),
           Qci2   = Hypochlorite Metering Pump Rate (gph), and
           t      = Storage Time (days).

       5.1.2  On-Site Hypochlorite Generation
       On-site generation of sodium hypochlorite is accomplished by adding electricity to a satu-
       rated (32%) brine solution. The strength of hypochlorite produced is 0.8%, which is below
       the hazardous material threshold of 1%.  These systems can be constructed piecewise or
       purchased as pre-engineered units.

       Figure 5-3 shows a typical flow diagram for an on-site hypochlorite generation system. The
       equipment requirements of an on-site generation system, which can be seen in Figure 5-4,
       include a salt saturator, hypochlorite storage tanks, electrolyzers, rectifiers,  controls, and
       hypochlorite metering pumps.  The following material inputs are required per pound of
       chlorine generated: 3.5 Ibs NaCl salt, 15 gallons of water, and 2.5 kWh of electrical energy.

       Figure 5-5 shows an on-site hypochlorite generator that will produce up to 36 Ibs of chlorine
       per day.

Arsenic Treatment Technology Evaluation Handbook for Small Systems                                84

-------
                                      Water
                                     Softener
                                            Pump
                  Raw
                 Water
                                                       	W    Hypochlori
            orite
                                                                   Generator
83
                                                                  Hypochlorite
                                                                  Storage Tank
      N-
Metering
 Pump
               Chlorinated
                 Water
           Figure 5-3.  Typical On-Site Hypochlorite Generation Process Flow Diagram.
                                Sodium                 Hydrogen Discharge
                                Hypochlorite        Supply/Rectifier
                                Solution x
                       Interstage
                       Hydrogen Gas
                       Take-off   \
                       Electrolyzer
                       Bellows Pump
          Sodium Hypochlorite
          Storage Tank

            Sodium Hypochlorite
            Metering Pump
                                                                     '   Blowers
                                                                       primary and Back-up
                                                                         Water Main
                                                                         Water Supply
                                                                      Water Softener
                                                  Salt Saturator
                                  Brine Solution
Figure 5-4.  On-Site Hypochlorite Generation System Schematic (USFilter, Wallace & Tiernan).
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                            85

-------
         Figure 5-5.  On-Site Hypochlorite Generation System (Severn Trent Services).

5.2   Permanganate Pre-Oxidation Design  Considerations

The primary applications of permanganate (MnO4~) in water treatment include preoxidation (par-
ticularly for iron and manganese) and taste and odor control. Potassium permanganate exists in
solid, granular form, but is typically applied as a saturated liquid (60 g/L at room temperature).

Permanganate is not biocidal against drinking water pathogens, so there should be negligible re-
sidual leaving the treatment works. Manganese particulates (MnO2) are produced as a result of
permanganate oxidation reactions. To prevent the accumulation of these deposits in the distribu-
tion system, post-filtration treatment must be applied.

Potassium permanganate is a Class 2 Solid Oxidizer. The storage of more than 250 Ibs necessitates
special hazardous waste precautions.  Potassium  permanganate can be purchased in a variety of
quantities, including 55-lb (25-kg) pails, 110-lb (50-kg) kegs, and 330-lb (150-kg) drums. The
solids can be  stored indefinitely if kept in a covered container and maintained in a cool, dry envi-
ronment. Special handling and safety requirements should be employed when working with solid
potassium permanganate,  including the use of goggles, rubber gloves, and an  approved NIOSH-
MSHA dust and mist respirator.

Careful consideration should be given to the permanganate dose estimate.  Most waters contain
substances other than As(III) that exert oxidant demand. Section 2.2.2 lists the permanganate de-
mand for the  stoichiometric conversion of As(III), Fe2+, Mn2+, and HS~.  Permanganate reacts ag-
gressively with organic materials. Permanganate may also be consumed during the regeneration of
MnO2 media.  The ultimate permanganate demand is the sum of all these factors. The applied dose
should be three times larger than the ultimate permanganate demand.


Arsenic Treatment Technology Evaluation Handbook for Small Systems                                86

-------
>MnO4
                                            MnO4
         Eqn. 5-7
Where:
        4  = Permanganate Dose (mg/L as Mn) and

    DMno4 = Ultimate Permanganate Demand (mg/L as Mn).

The application of potassium permanganate is straightforward.  Permanganate solution is prepared
by loading solid potassium permanganate into a storage silo. A feeder meters the permanganate
into a dry hopper, which allows the solids to be pulled into a water stream where it dissolves.  The
permanganate  solution  is then  stored in a solution tank until  it is metered into the water to be
treated.  This process is shown in the flow diagram in Figure 5-6.  For small  systems looking to
maintain simplicity, manually loading solids into a solution tank filled with water to create batch
quantities of permanganate solution is recommended.

Pre-engineered drum inverters  (Figure 5-7) and dry  feeders (Figure 5-8) are available  in several
different styles, including gravimetric weigh-belt and volumetric (hopper) type.
         Storage
          Silo
       Gravimetric
         Feeder
          Dry
         Hopper
    Raw
   Water
                                                 •Q—
                                             Permanganate
                                             Solution Pump
Oxidized Water
  to Treatment
                 Figure 5-6.  Typical Permanganate Process Flow Diagram.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
               87

-------
  Figure 5-7. Permanganate Dry Feed System      Figure 5-8. Permanganate Dry Feed System
          (Merrick Industries, Inc.).                            (Acrison, Inc.).

The stock solution is then metered into the water system with the use of a small pump. The flow
rate of solution required to meet the dose requirements are contingent upon the strength of the stock
solution, according to Equation 5-8.
                                          UMn04  V.       /
Where:
    QMno4 = Permanganate Metering Pump Rate (gph),
    Q     = Design Flowrate (gpm),

    5Mno4 = Permanganate Dose (mg/L), and

         4 = Permanganate Stock Solution Concentration (mg/L).
                                                                               E
5.3   Ozone Pre-Oxidation Design  Considerations

Ozone can be used in water treatment for disinfection, oxidation, and taste and odor control. Ozone
is a gas and is created either by passing air through an electrical discharge or by irradiating air with
UV light. The UV method is much less expensive, quite reliable, and can produce ozone in a 0. 1%
concentration.

Careful consideration should be given to the ozone dose estimate. Most waters contain substances
other than As(III) that exert oxidant demand. Section 2.2.3 lists the ozone demand for the stoichio-
metric conversion of As(III), Fe2+, Mn2+, and HS~. Ozone will also react with TOC. The ultimate
ozone demand is the sum of all these factors. The applied dose should be three times larger than the
ultimate ozone demand.
Arsenic Treatment Technology Evaluation Handbook for Small Systems

-------
                                                                                 Eqn. 5-9
Where:
    งo3
    D
     03
= Ozone Dose (mg/L) and
= Ultimate Ozone Demand (mg/L).
To create ozone, an air stream is passed through a tube irradiated with UV light. This excites the
oxygen (O2) molecules and causes some of them to form ozone (O3).  The air stream containing
ozone is injected and mixed into the raw water, which then passes into a contactor, which provides
time for the ozone to dissolve into the water.  The mixture then flows into a de-gas separator that
allows the un-dissolved gasses to separate to the top where they leave the separator, pass through a
residual ozone gas destructor, and are off-gassed. The contactor and de-gas separator also provide
time for ozone to oxidize the As(III) into As(V), which, depending on interfering reductants that
may be present, could be as long as 2.2 minutes.  Figure 59 below shows a process flow diagram for
a typical ozonation process.
                                                                • Offgas
                                      Ozone/Air
                                       Recycle
                                                3 "
                                                CD I
                Raw
                Water
                                                 o
                                                 O
                                                          11
                                                       Oxidized Water
                                                        to Treatment
                                     Venturi
                                     Injector
                   Figure 5-9.  Typical Ozonation Process Flow Diagram.

The ozone generator can be sized by taking the flowrate times the ozone dose as shown in the
equation:
                        Mo3 = S03 •

Where:

Mo3   = Ozone Mass Flow (g/h O3)
A,-,     	 flr7*-vป"ป^ T^*-vnzi / -w\ /~r /T \ o ป^ H
5o3    = Ozone Dose (mg/L) and
Q     = Design Flowrate (gpm).

Figure 5-10 shows
                                                 --.
                                                i       i
                                             gal  mg  hr
                                                                     Eqn. 5-10
       owrate (gpm).

       an ozone generator that will produce up to 35 g/h (583 mg/min) of ozone.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                                             89

-------
                 Figure 5-10. Ozone Generator and Contactor (ProMinent).

5.4   Solid Phase Oxidant Pre-Oxidation Design Considerations

Filox-R™ is a solid, granular manganese dioxide media typically used to remove iron and manga-
nese from drinking water. FiloxR™ media has also been shown to effectively catalyze the oxida-
tion of As(III) to As(V) using dissolved oxygen.

For most ground water sources, the dissolved oxygen content will be very low.  Oxygen may need
to be added depending upon the concentrations of interfering reductants. An alternative to adding
oxygen is to increase the empty-bed contact time (EBCT) to overcome the interfering reductants. If
oxygen addition is selected, it can be done by injecting air into the water stream using a venturi air
injector as shown below in process flow diagram Figure 5-11. Figure 5-12 shows the schematic of
an air injection assembly. The water and air are allowed to mix for a short period of time and then
the undissolved gasses are removed from the water by a degassing unit. The oxygenated water then
flows downward through a column of FiloxR™ media.

                                              Offgas
Air 	


^
Raw 1
Water
Yen


r

turi Air
Injector
-. 	 	 ^



~H

Air
Contact





ฃt}
De-
Gas

T










T
Solid
Phase
Oxidant
Column
LJ






                                                                Oxidized Water
                                                                 to Treatment
      Figure 5-11.  Typical Solid Phase Oxidant Arsenic Oxidation Process Flow Diagram.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
90

-------
                                   PRESSURE REGULATING, OR
                                    FLOW CONTROL VALVE
                                           I
                MAIN FLOW

1
JI L8-

INJECTOR
1
L
L


i
                               CHECK
                               VALVE
                                                  . METERING
                                                   VALVE
                                                               TO PROCESS
               Figure 5-12.  Venturi Air Injector Assembly Schematic (Mazzei).

Careful consideration should be given to the dissolved oxygen dose estimate. Most waters contain
substances other than As(III) that exert oxidant demand.  Section 2.2.4 lists the oxygen demand for
the stoichiometric conversion of As(III), Fe2+, Mn2+, and HS~.  FiloxR™ may also catalyze oxida-
tion with TOC. The ultimate ozone demand is the sum of all these factors. The applied dose should
be at least ten times larger than the ultimate oxygen demand as seen in the equation below. Some
test runs by Ghurye and Clifford with interfering reductants used as much as 65 times the stoichio-
metric oxygen demand (Ghurye and Clifford, 2001).
                                                                              Eqn. 5-11
Where:
   80  = Oxygen Dose (mg/L) and
   D0 = Ultimate Oxygen Demand (mg/L).
The EBCT is the other important design criteria for a solid-phase oxidant system.  Tests using
FiloxR™ were successful with EBCTs of 1.5 minutes. If oxygen was not present in 10 to 65 times
the stoichiometric demand, EBCTs of 6 minutes were required.  (Ghurye and Clifford, 2001).

Typical hydraulic loading rates for Filox-R™ systems are  10 to 20 gpm/sft. Given this and the
EBCT, the height of the Filox-R™ bed can be determined using the equation:
                                Z >HLR-EBCT
                                                 eft
                                               7.48 gal
Eqn. 5-12
Where:
   Z     = Depth of Media (ft),
   HLR  = Hydraulic loading rate (gpm/sft), and
   EBCT = (Minimum) Empty Bed Contact Time (min).

Additional typical design and operating parameters for a Filox-R™ system are given in Table 5-1.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
       91

-------
 Table 5-1.  Typical Filox-R™ Design and Operating Parameters.
                    Parameter                                         Value    Units
 Bulk Density Fibx-R™ '	114    bs/cft
 Freeboard '	30-50%	
 Filox-R™ Media '	> 20    in.
 Hydraulic loading rate2                                                  10-20    gpm/sft
 Empty Bed Contact Time 3	1.5-6.0    min.
 Minimum Backwash Flowrate '                                             12-15    gpnVsft
 1 Recommendation by Matt-Son, Inc., Filox-R™ Media, FormNo. FXR-01.
 2 Recommendation by Matt-Son, Inc., Filox-R™ Media, FormNo. FXR-06.
 3 Ghurye and Clifford, 2001.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                    92

-------
5.5   Comparison of Pre-Oxidation Alternatives




Table 5-2 provides a review of issues pertinent to the five pre-oxidation methods previously discussed.
                  Table 5-2.  Comparison of Pre-Oxidation Alternatives.
Criteria
Safety and
Regulatory
Issues



Space
Requirements

Chemical
Characteristics


Chemical
Delivery
Labor

Operation and
Maintenance
Off-Normal
Operation
Community
Relations
Liquid Sodium
Hypochlorite System
• HazMat regulations for
safety and handling apply.
• Potential for corrosive
vapors in the presence of
moisture.
• Emergency response
plan required with local
fire department.
• Secondary containment
required.
• Space requirements are
small, assuming the
Uniform Fire Code
(UFC) exempt criteria
are met.

• 5'/4 or \2l/2ฐ/o sodium
hypochlorite solution.
Degrades over time.
• Decay of solution creates
chlorate byproduct.
• Increases pH of water
slightly.

• Liquid hypochlorite
delivered by tanker truck,
55-gal drum, or 5-gal
pail
• Periodic delivery.
• Dilution procedures.
• Low day-to-day O&M.
Long-term material
maintenance could be a
problem because of
corrosive effects of liquid
hypochlorite.
• A temporary bleach
solution can be mixed in
the storage tank.
• HazMat signage required.
On-Site Hypochlorite
Generation System
• Below 1% threshold for
hazardous classification
• Exempt from HazMat
regulations.

• No secondary
containment requirements.

• Space requirements are
large. There must be
room for salt storage,
brine tanks, hypochlorite
holding tanks, electrolytic
equipment, as well as
instrumentation & control
and power.
• Stable sodium
hypochlorite solution
(0.8%).
• Constant application
concentration.
• Chlorate formation low to
none.
• Increases pH of water
slightly.
• Salt delivered in 50-lb
bags or 2000-lb totes.
• Salt delivery.
• Weekly loading of salt
into brine tank.
• Moderate O&M, mainly
associated with salt
handling. Change
electrode cells every five
years.
• A temporary bleach
solution can be mixed in
the day tank.
• No HazMat regulations.
Hydrogen byproduct
vented to atmosphere.
Permanganate Solution
Feed System
• Solid permanganate
poses dust and
inhalation hazard.



• Space requirements
are small Additional
space may be required
for storage of solid
permanganate.

• Stable permanganate
solution, generally 3-
4%.
• Reacts rapidly with
dissolved organics.


• Solid permanganate
available in 25-kg
pails, 50-kg kegs, and
150-kg drums.
• Load dry feeder.
• Dilution procedures.
• Low day-to-day
O&M for automated
systems.
• Stains everything
purple.
• N/A
• N/A
Ozone Generation
• Poisonous and
reactive gas.



• Space requirements
are small

• Gas.
• Very strong oxidizer.


• N/A
• N/A

• Low day-to-day
O&M.
• N/A
• N/A
Solid Oxidant
System
• None.



• Space
requirements are
small.

• Solid.
• Requires
dissolved oxygen
in the water.


• N/A
• N/A

• Low day-to-day
O&M.
• N/A
• N/A
Arsenic Treatment Technology Evaluation Handbook for Small Systems


-------
Preceeding Page Blank
                                                                       Section 6
        Sorption  Process Design  Considerations
        This section describes the design of sorptive processes, including AA, modified AA, IBS, and IX.
        For reasons previously cited, the discussion about AA, modified-AA, and IBS are restricted to non-
        regenerable applications. Conversely IX is most economically feasible when used in a regenerable
        process.

        6.1   Process Flow

        Despite the availability of several different types of sorptive treatment processes, the overall treat-
        ment approach for each is similar. Pre-treatment can consist of oxidation, to convert As(III) to
        As(V), and pre-filtration stages when turbidity is high, as well as optional pH adjustment and pre-
        filtration backwash.  Next, the water is fed through a column packed with sorptive media.  Post-
        treatment consists of an optional pH re-adjustment stage and some media have an option for regen-
        erating the media. Typically, the entire process is carried out under pressure. Figure 6-1 shows a
        typical sorption treatment process while Figure 6-2 shows the same flow diagram with the optional
        media regeneration and pH adjustment and re-adjustment. Dashed lines and boxes indicate op-
        tional streams and processes.
                        Oxidant
                   Raw 	^   Pre-      _^J   Pre-   [	^  Sorptive  	^ Treated
                  Water       Oxidation      ! Filtration !      Treatment        Water
                                            Backwash
                                              Waste

         Figure 6-1.  Sorption Treatment Process Flow Diagram w/o pH Adjustment and Regeneration.
              Oxidant —i      Acid	1                                  Base —i
                             ,____*____,    	    	        ,_____*_____.
          Raw       Pre-          pH    !_^!  ^re"   !	    Sorptive    , ;     pH         Treated
         Water     Oxidation    \ Adjustment !    | Filtration \   _ Treatment  ^   \ Re-Adjustment T    Water
                                              i
                                           Backwash  Waste     Regenerant
                                            Waste   (IX Only)     (IX Only)


          Figure 6-2. Sorption Treatment Process Flow Diagram w/ pH Adjustment and Regeneration.


        Arsenic Treatment Technology Evaluation Handbook for Small Systems                              95

-------
Pre-filtration is strongly recommended when the source water turbidity is above 0.3 NTU.  Sus-
pended solids in the feed water can clog sorption sites and impair process hydraulics. One prefiltration
option for smaller systems is backwashable cartridge filters.

The performance of AA treatment is highly pH-sensitive. Treatment conducted under acidic condi-
tions (pH 5.5-6.0) can be expected to  produce run lengths 5 to 20 times longer than treatment
conducted under natural pH conditions. As a result, in the decisions trees in Section 3, conven-
tional AA is only recommended over IBS when the pH is naturally low or the system is willing to
adjust the pH below 6.0. In most cases, pH adjustment will require chemical addition of a strong
acid, such as sulfuric (H2SO4) or hydrochloric (HC1) acid. Dose requirements depend on the back-
ground pH and buffering capacity of the water.

6.2   Column Rotation

Sorption processes are conducted using two or more columns in series.  The first column in the
treatment process is referred to as the roughing column, and the last sorption  column is referred to
as the guard column.  Frequently, there  is an additional column on standby. The roughing column
serves as the primary arsenic removal column. The guard column is intended to  capture arsenic
breakthrough as soon as it occurs from the roughing column.

The columns are operated in this manner until arsenic breakthrough of the roughing column occurs,
which is detected by periodic grab samples. Breakthrough is generally defined as the time when the
effluent arsenic concentration is equal to 50% of the feed water arsenic level. However, this num-
ber can be adjusted after piloting or operation to optimize the economics of the process.  At this
point, adsorptive sites on the roughing  column have become saturated and the column should be
taken off-line for media replacement or regeneration, after which it is placed in standby mode to
wait for the next column rotation.  The guard column is then promoted to the roughing column
position and the standby column becomes the guard column in the series.  Figure 6-3 illustrates
how the columns' positions are rotated between roughing, guard, and standby operation modes.

The number of columns to be placed in series depends on the estimated lifetime of each column and
the desired monitoring and media change-out or regeneration frequency. IX processes operating
with sulfate in  the feed may have a  sulfate roughing column at the head of the operation to remove
sulfate before arsenic is removed by the arsenic roughing column.

Typically two parallel process trains are used.  This provides operational redundancy and, by stag-
gering their operation, chromatographic peaking may be reduced.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                96

-------
                               Normal Operation Mode
       Feed Water


A
Roughing
Column

A

Guard
Column
                                                       Standby
                                                       Column
                                                                      Distribution
                                                                        System
       Feed Water
                          Regeneration/Replacement Mode
                          Standby
                          Column
               Waste
                       Regenerant or
                       Replacement
                          Media
                                                                      Distribution
                                                                        System
                      Figure 6-3. Sorption Column Operation Modes.

6.3    Sorption Theory

To understand operation of sorption processes, it is important to understand fundamental ion ex-
change theory. An important consideration in sorption processes is the mass transfer zone (MTZ),
which can be viewed as a wave or a zone of activity (i.e., non-equilibrium between liquid and
media phases) for a particular contaminant. As depicted in Figure 6-4, the MTZ also represents the
front of the exhaustion zone for a particular contaminant.  Exhaustion zones and MTZ waves are
typically considered for the target contaminant (i.e., arsenic) and any species that have a higher
exchange affinity for the media. Arsenic must compete with other anions for exchange sites ac-
cording to the selectivity sequence for the particular media (see Section 2).  Previously sorbed
arsenic can be displaced by anions of higher selectivity. Exhaustion and MTZs order themselves
according to the selectivity sequence, as illustrated in Figure 6-4. Other sorbed contaminants, such
as carbonate (CO32~) and nitrate (NO3~), would be present further down from the As(V) MTZ.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
97

-------
                                                                Resin Loaded (%)
                                    SO42"     HAsO42'
o   o   c    _    _
•sP   -sP   -sP    -sP   -sP   -sP
o^   o^   o^    o^   o^   o^
                                                                           SCO   O
                                                                           o   o

           Exhausted Media
           Partially Loaded Media
           (MTZ)
           Fresh Media
                                                                                 o
                                                                                 
-------
       and low capacity of IX resin generally renders it uneconomical for one-time application and
       disposal.  Instead, periodic regeneration should be applied to restore the exchange capacity
       of the resin.

       Figure 6-4 illustrates a resin-phase loading profile down an IX column for treatment of
       hypothetical natural water containing arsenic and sulfate. As arsenic is exchanged with
       anions on the SBA, the arsenic band  develops and its MTZ moves downward. The same
       phenomenon is true for sulfate ions.  However, because of its higher exchange affinity,
       sulfate anions  displace the arsenic, thereby forcing the arsenic-exhausted region and the
       arsenic MTZ downward further.

       An important consideration in the application of IX treatment is the potential for chromato-
       graphic peaking of nitrate (NO3~) and nitrite  (NO2~).  These contaminants pose an acute
       health risk, and as such are regulated under the SDWA with primary MCLs of 10 mg/L (as
       N) and 1 mg/L (as N), respectively.   According to the  selectivity sequence provided in
       Section 2.3.1, nitrate and nitrite will also replace chloride on exchange sites, although with
       less preference than As(V) or sulfate.  As a result, the region of nitrate and nitrite activity
       will reside further down the column (relative to the activity of sulfate and As(V)), as illus-
       trated in Figure 6-5.  These species will  chromatographically peak before As(V), and this
       peaking could  produce water that does not meet the aforementioned MCLs.  Utilities with
       source water with measurable quantities of nitrite or nitrate  should be aware of this phe-
       nomenon and plan column operation to avoid  this occurrence.
                                                                     Resin Loaded (%)
                         SO42'   HAsO42-
      Exhausted Media
      Partially Loaded Media
      (MTZ)
      Fresh Media
               Figure 6-5. Activity of Nitrate and Nitrite During Ion Exchange.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
99

-------
       The removal of carbonate (CO32~) by IX resin can also lead to a pH drop of 0.5 to 1.0 units,
       particularly at the beginning of a run.  This impact can be minimized by post-treatment
       addition of soda ash or caustic soda or by sequencing the regenerative cycles of parallel
       process trains. Pilot testing is recommended to evaluate the impact on pH for the specific
       water in question.

6.4   Process Design & Operational Parameters

Design and operational parameters for sorption treatment processes vary significantly depending
on the specific technology chosen, and to a lesser degree on the media type. The most appropriate
way to identify the optimal engineering parameters for a particular treatment application is to con-
duct on-site pilot column  studies with the media of interest.

Regenerable IX processes involve three operating modes: (1) Loading; (2) Regeneration; and (3)
Rinsing.  Loading can be conducted with flow in either the downward or upward direction, al-
though the former is more common in water treatment applications.   Once the column is fully
loaded it should be taken  off-line.  The next step is regeneration with concentrated brine for chlo-
ride-based SB A, which can be conducted in either the downward or upward direction.  The latter
case is generally more effective, although care must be taken to prevent fluidization of the media.
Prior to returning the column to service, water rinsing should be conducted to displace regenerant
solution from the column. Slow rate and fast rate rinsing should be conducted in sequence, with
each displacing about 2-3 bed volumes of solution per column.

Table 6-1 details key design and operational parameters for AA, IBS, and IX processes. As de-
scribed in Section 2, non-regenerable AA and IBS process are recommended for small utilities.
Therefore, rinsing and regeneration data is only provided for ion exchange processes.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                               100

-------
Table 6-1. Typical Sorption Treatment
Parameter
Media Bulk Density
Design and
IX
40-44
Operating
AA
40-47
Parameters.
IBS
72-75
Units
Ibs/cft
Minimum Column Layers
Freeboard
Media
90% '-2
36-60 2
50% 3
36-60 3
50%
32-40
-
in
Operating Conditions
Hydraulic Loading Rate
Empty Bed Contact Time
Downflow Pressure Drop 4
Maximum Pressure Differential
8-122
1.5-5
0.7-1.3 2
14
4-9 3
53
0.1 6
5
5-8
5-10
N/A
3.5
gpm/sft
mm.
psi/ft
psi
Backwash Conditions
Backwashing Flow Rate
Backwashing Duration
3-4 2
5-20 2
73
10 3
-
-
gpm/sft
min
Regeneration Conditions '
Brine Strength
Downflow Rate
Regenerant Volume
6-10 2
2-6
20 2
-
-
-
-
-
-
wt%
gpm/sft
gal/eft resin
Rinsing Conditions
Slow Rinse Rate
Fast Rinse Rate
Displacement Requirements
0.4-4
2-20
4-6
-
-

-
-

gpm/sft
gpm/sft
bed volumes
 1 This will be very resin specific. Check with the resin manufacturer before design.
 2Rubel,2001aDraft.
 3 Rubel, 200 Ib Draft.
 4 This depends on temperature, type of media, and hydraulic loading rate.
 5 For strong base anion exchange resin at 70ฐF and 10 gpm/sft.
 6 For AA at 2 gpm/sft.
 N/A - Not Available.
6.5   Column Design

The vessels should be made from typical, well-known materials of construction such as carbon
steel or fiberglass and must be NSF approved.  The vessels should have distribution and collector
systems that provide a uniform distribution of fluids during all phases of the operation. More detail
on these accessories is provided in Section 7.  Also, it is advisable to install sight-glasses in order to
check resin levels.

Columns placed in series are referred to as a treatment train. The utility should evaluate the number
of parallel treatment trains based on the desired redundancy and state design standards. Figure 6-6
shows a commercially available multiple-column IX treatment train.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
101

-------
               Figure 6-6. Ion Exchange System (Tonka Equipment Company).

       6.5.1   Column Diameter
       Once the number of parallel treatment trains has been established, column diameter can be
       calculated based on the recommended hydraulic loading rate of the particular media and the
       design flowrate. Hydraulic loading rate is the flowrate per unit of cross-sectional area and
       is proportional to the linear velocity of the fluid through the bed.  Recommended maximum
       hydraulic loading rates are provided in Table 6-1.  Column diameter (D) can be calculated
       using the equation:
                                  D =
                                         4-Q
                                      71-np -HLR
                                                                               Eqn. 6-1
       Where
          D
          Q
          n
          HLR
: Column Diameter (ft),
: Design Flowrate (gpm),
: Number of Parallel Treatment Trains, and
: Hydraulic loading rate (gpm/sft).
       The benefits of lower hydraulic loading rates include a sharper MTZ and potentially better
       media utilization.  However, a lower hydraulic loading rate also translates into a larger
       column footprint.

       Consider an example where IX will be used to treat a design flowrate of 70 gpm. The utility
       has decided to provide no parallel treatment trains. Based on a recommended maximum
       hydraulic loading rate of 10 gpm/sft, the column diameter should be 3 feet.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                                  102

-------
                                 D
                                      Ti-1-10 gpm/sft
       6.5.2  Column Height
       The depth of sorptive media required can be calculated based on the selected hydraulic
       loading rate and consideration of the minimum empty bed contact time.  Values of EBCT
       are provided in Table 6-1.


                                 Z>HLR-EBCT-   cft                          Fnn 6 2
                                               [ 7.48 gal J
       Where:
          Z     = Depth of Sorptive Media (ft),
          HLR  = Hydraulic loading rate (gpm/sft), and
          EBCT = (Minimum) Empty Bed Contact Time (min).

       Returning to the previous example, suppose the specific resin selected had a minimum
       EBCT of 3 minutes.  The total depth of sorptive media required for the primary treatment
       columns would be 4 feet.
                                                        =4ft
                                     sft          7.48 gal
      The depth of sorptive media (Z) should then be used in conjunction with the column free-
      board to determine column height. For ease of change-out, all columns should be sized
      similarly.

                                        H = Z-(l+F)                             Eqn. 6-3
      Where:
          H  = Column Height (ft),
          Z  = Depth of Sorptive Media (ft), and
          F  = Freeboard Allowance (% expressed as a decimal).

      For the previous example, if the freeboard requirement was 50% of media depth, the col-
      umn height should be 6 feet.

                                     H = (4 ft)-(l + 0.5)= 6ft

      Therefore, for this particular example, the design should include a single treatment train
      consisting of three columns (i.e., roughing, guard, and  standby).  All columns should be 3
      feet in diameter by 6 feet tall and contain 4 feet of media. The process flow diagram for this
      example is provided as Figure 6-7.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                               103

-------
                                Roughing
                                 Column
                                    Guard
                                   Column
Standby
Column
Pre-Oxidized
 Raw Water
                                                         o
                                                                     Treated
                                                                     Water
                  Figure 6-7.  Process Flow Diagram for Example Problem.

       The following constraints should also be considered:

       •  Small column aspect ratios (i.e., H:D <1) can lead to flow maldistribution.
       •  Large column heights can lead to excessive pressure drop.
       •  The available building footprint.
       •  The available building height.

6.6   Media Replacement Frequency

It is advantageous for a utility to obtain a rough estimate of the optimal operating time until media
exhaustion occurs. This is important for establishing an appropriate O&M and monitoring sched-
ule.  The optimal filter run time until media exhaustion can be calculated as:
                            = BV -EBCT
                                             hr
                                           60 min
                                                                              Eqn. 6-4
Where:
    T     = Optimal Filter Run Time (hr),
    BVe   = Number of Bed Volumes to Exhaustion, and
    EBCT = Empty Bed Contact Time (min).

The roughing column should be operated until 50% arsenic breakthrough occurs.  Therefore, the
actual filter run time will be less than the calculated optimal filter run time (t). The deviation will
depend on the efficiency of the sorption/exchange process and the width of the MTZ.

Consider an example where the estimated lifetime of a particular combination of media and raw
water was 1,357 BV.  If the columns are sized to provide an EBCT of 3 minutes, the optimal run
time until media exhaustion is 68 hours.
                    T = (1,357 Bed Volumes)- (3 min)-
                                                     hr
                                                   60 min
                                                  = 68hr
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                                          104

-------
6.7   Regeneration of Ion  Exchange Resin

IX resins are essentially unusable for arsenic removal unless they can be efficiently regenerated.
Because of the high selectivity of SBA for sulfate (SO42~), the exchange capacity would be ex-
hausted within a few days for many natural waters. The cost of the virgin resin is far too great to
dispose of it at that time.

Chloride-based SBA can be regenerated with concentrated brine (1-5 mole/L) in either the upflow
or downflow mode. The more concentrated the regenerant solution, the greater the fraction of the
bed that is regenerated.  It should be noted, however, that regeneration efficiencies are generally
less than 100%. Therefore,  successive runs can be expected to be slightly shorter in duration.

Utilities should consider the size  of the brine holding tank, as they are typically much larger than
the IX columns themselves. Based on previous studies (AwwaRF, 2000), roughly 4 BVs of spent
brine are produced per regeneration. The regeneration duration can be calculated as:


                               t=_?_.f7.48^-1-(4BV)                           Ban 6-5
                                    GR ^    eft J V
Where:
tR = Regeneration Duration (min),
Z = Depth of Sorptive Media (ft/BV), and
GR = Regeneration Flux (gpm/sft).

Following regeneration, this brine can either be disposed of via indirect discharge (assuming local
TBLLs are met) or stored for recycle. In the case of recycle, it may be necessary to add salt to bring
the strength of the brine back to the range 15 mole/L.

For a conventional IX process, spent regenerant will contain arsenic and sulfate in a ratio approxi-
mately corresponding to their relative concentration in the raw water. If the water contains a mod-
erate amount of competing ions, it is possible that the brine waste will contain less than 5.0 mg/L of
arsenic, and thus will  not exceed the TC values.  However, in most instances, the liquid waste
stream will contain more than 5.0  mg/L of arsenic. This will force utilities to consider disposal and
waste treatment options. If indirect discharge to a local POTW is the waste disposal method cho-
sen,  a spent brine holding tank may be required in order to slowly release the spent brine to the
POTW.

Rinsing with water is typically conducted after regeneration to flush out residual brine and prepare
the  column for normal operation. Generally 4 to 6 BV of rinse water are used per regeneration.
This waste may be added with the brine waste being sent to the POTW.

IX resin typically lasts 4-8 years  before chemical and mechanical degradation necessitates media
replacement.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                105

-------
6.8   Waste Handling Systems

This section addresses three types of waste: backwash water from pre-filters, spent regenerant, and
spent media.

The two most probable methods for disposal of backwash water from pre-filters are indirect dis-
posal through a POTW or by settling the solids, recycling supernatant, and sending the solid sludge
to a landfill.

Regarding brine used in IX regeneration, there are two waste disposal options.  Spent brine that
contains less than 5.0 mg/L of arsenic can either be disposed of via indirect discharge or treated on-
site. The feasibility of indirect discharge of regenerant waste will be dictated by local TBLLs for
TDS. The concentration of TDS in the spent regenerant can be approximated as:
                            CTDS = 58.4 -      1 MBrine                             Eqn. 6-6
Where:
    C^g   = Concentration of Total Dissolved Solids (g/L) and
        e = Brine Molarity (mole/L).
When indirect discharge is not an option, the system must deal with the waste on-site. The most
common approach for treating brine waste (containing less than 5.0 mg/L of arsenic) is chemical
precipitation with iron-based salts and subsequent solids thickening. Thickening can be conducted
using a settling basin, or for more rapid results, mechanical dewatering equipment.  The brine
decant can then be sent to an evaporation pond.

Spent brine used in the regeneration of arsenic-laden resin may be classified as hazardous.  There-
fore,  manipulating the chemical form of the waste on-site constitutes treatment of a hazardous
waste, which has extensive permit and cost implications.  As a result, when the brine waste stream
contains over 5.0 mg/L of arsenic, indirect discharge to a POTW is considered the only viable
option for small utilities. When this option is unavailable, on-site regeneration of arsenic-laden
resin  should not be performed. Rather, the resin should be disposed of at a municipal solid waste
landfill and replaced with fresh resin.

The appropriate disposal method for spent resin is dependant on the results of the TCLP, as de-
scribed in Section 1 .
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                106

-------
                                                              Section  7
         Pressurized  Media  Filtration  Process
        	Design Considerations
This section describes the design of a typical pressurized granular-media filtration system includ-
ing sand filtration and iron and manganese oxidation/filtration systems. Although the following
information specifically describes a pressurized greensand filter, it can be applied to any pressur-
ized granular media filtration system.

7.1   Process Flow

In a typical media filtration process, seen in Figure 7-1, the raw water is first put through a pre-
oxidation step. Dashed lines and boxes indicate optional streams and processes. If the preoxidant
is chlorine, potassium permanganate, or ozone, the As(III) and any natural iron will be oxidized to
As(V) and Fe III respectively. If, however, aeration (aeration tower) is used for iron oxidation, the
air oxidation process will not oxidize As(III) to As(V) and the addition of a chemical preoxidant
would be required.  If greensand is being used as the filter media, potassium permanganate and
chlorine also provides the oxidant for the continuous regeneration of the greensand media.
Oxid

ant — i
Pre-
Oxidation

Coagulant 	 .
i
i
^j Coagulant i ^
^| Addition ]


Media
Filtration
1
Backwash
Waste
          Water          ^rev         i i .MH^iimm i         ivieMim          rr*
                                                                    Treated
                                                                    Water
               Figure 7-1. Typical Media Filtration Process Flow Diagram.

It should be noted that, although greensand can be regenerated in either batch or continuous meth-
ods, only the continuous regeneration method has been shown to also oxidize the As(III) to As(V)
so that arsenic can be removed.  Therefore, under most circumstances, only the continuous regen-
eration method is recommended for arsenic removal.

After pre-oxidation, a coagulant addition step may be necessary if the iron concentration or the
Fe: As ratio is low. Next, the water is passed through filters containing granular media before being
sent to the distribution system. Typically, three or more filters are provided in parallel.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                            107

-------
Media filters are operated in three different modes: (1) Filtration; (2) Backwash; and (3) Filter-To-
Waste (FTW).  In the operating mode, all filters are fed in parallel with flow in the downward
direction. The effluent is sent to the distribution system as shown in Figure 7-2.

After some time of operation, solids captured by the filtration media will impede the flow and
increase the differential pressure across the filter.  To restore hydraulic capacity, the filter will have
to be backwashed.  The backwash flow is in the upward direction, which fluidizes the granular
media and washes the accumulated solids out of the filter.  In some instances, air scouring is con-
ducted prior to fluid backwashing.  Air scouring bubbles large volumes of air upward through the
filter.  This assists in breaking apart conglomerates of filtered material, allowing the  subsequent
fluid backwash to more easily remove the captured solids. An air scour also reduces the volume of
backwash waste that is generated.
                                     Filtration Mode
           Flow from
           Pre-Oxidation
                                                                   Distribution
                                                                     System
                                    Backwash Mode
           Flow from
           Pre-Oxidation
                                                                     Waste
                                  Filter-To-Waste Mode
           Flow from
           Pre-Oxidation
                                                                   Distribution
                                                                     System
                                                         Waste
                      Figure 7-2. Media Filtration Process Flow Modes.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
108

-------
After backwashing, the media is allowed to settle and downward flow is reinstated with the filter
effluent going to waste. This re-stratifies the column, setting it up for operation. It also reduces the
amount of particulate matter that gets into the distribution system.  After the FTW mode, the filter
is returned to standard operation.

7.2   Process Design & Operational Parameters

Table 7-1 lists design and operational parameters typical of media filtration systems.

          Table 7-1.  Typical Greens and Column Design and Operating Parameters.
Parameter
Value
Units
Media Bulk Density
Anthracite Media '
Greensand Media 2
Garnet Media 3
Support Gravel 4
50
85
140
100
bs/cft
bs/cft
bs/cft
bs/cft
Column Layers
Freeboard 1>2
Anthracite Media 5
Greensand Media 6
Garnet Media 3
Support Gravel 4
50%
12-24
15-24
4
18-30
of Anthracite and Greensand
in.
in.
in.
in.
Operational Parameters
Hydraulic loading rate 7-8
Max Pressure Differential
3-5
8-10
gpm/sft
psi
Backwash Parameters
Minimum Backwash Flowrate 2
Backwash Duration
Backwash Frequency
Bed Expansion 2
Air Scouring Rate
12
15
1-7
40%
0.8-2.0
gpm/sft
min
days
minimum
scfin/sft
Filter-to-Waste Parameters
FTW Hydraulic loading rate
FTW Duration
3-5
5
gpm/sft
min
 1 Recommendation by Clack Corporation, Anthracite, Form No. 2354.
 2 Recommendation by Clack Corporation, Manganese Greensand, Form No. 2349.
 3 Recommendation by Clack Corporation, Garnet, FormNo. 2355.
 4 Recommendation by Clack Corporation, Filter Sand and Gravel, FormNo. 2352.
 5 Clack Corporation, Anthracite, FormNo. 2354 recommends 10-18 for multimedia filters but may need to be higher
 depending on iron concentrations.
 6 Clack Corporation, Manganese Greensand, FormNo. 2349 recommends 30" but can be lower if used with
 continuous regeneration
 7 Clack Corporation, Manganese Greensand, FormNo. 2349 recommends 3-5 gpm/sft with 8-10 gpm/sft intermittent
 flow possble.
 8 Under some circumstances, continuous flowrates of 10 gpm/sft are possible.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
109

-------
7.3   Filter Design

Typically, granular-media pressure filters have multiple layers of media selected to maintain a coarse-
to-fme grading from the top to bottom of the filter. The coarse, upper layer provides rough filtration
and the bulk of the particulate retention while the fine, lower layer provides superior filtration.  This
scheme allows for longer runs times while maintaining filtration quality. A typical oxidation/filtra-
tion filter is shown in cross-section in Figure 7-3.
             Filter Influent
             Backwash Effluent
             Freeboard
                                                                Distribution Plate
             Anthracite
                                  ••••••••••••••••••••••••••••••
                                                                     Filter Effluent
                                                                     Backwash Influent
             Distribution Laterals
                                                        Distribution Header
                 Figure 7-3. Schematic of a Vertical Greensand Pressure Filter.

In the manganese greensand oxidation/filtration process, the primary layer in the filter is made of a
media that catalyzes iron  and manganese oxidation, promotes its precipitation, and filters out the
precipitate.  For  optimum arsenic removal, continuous chemical preoxidation with either potas-
sium permanganate or chlorine is recommended.  Arsenic is removed by the co-precipitation with
the iron and, to a lesser degree, the manganese. Greensand, glauconite sand coated with a thin layer
of MnO2, is the most common of these types of materials.

Because greensand is very fine (16-60 mesh) it is susceptible to being overloaded with solids. To
reduce the solids loading on the greensand a layer of filter coal such as anthracite is put on top. This
layer also provides an area for the iron floe to coagulate.  Because of anthracite's low density, the
filter coal will naturally stratify as the top layer after backwash.

In order to keep the greensand from being slurried out the under-drain, a layer of filter garnet is
placed below it.  This filter garnet has a particle size of 8-12 mesh and a density almost 50% greater
than the greensand. This  puts the filter garnet below the greensand after stratification.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
110

-------
The bottom layer is support granite, which allows the water to flow easily into the lower distribu-
tion system and exit the filter. Because of its larger size, the support granite is not fluidized during
backwash. Instead, it assists in distributing the backwash flow evenly throughout the filter.

When the media is backwashed, it will expand 30% to 50%. To accommodate this, the filter is
designed with freeboard.  Freeboard is the amount of space in the filter between the upper layer of
media and the upper distribution manifold. The height of this freeboard is dependent on the height
of media but is generally 40-50% of the settled height of the media that undergoes fluidization (i.e.,
anthracite and greensand in the case of a greensand filter).

Every filter will have an upper and lower distribution manifold. The upper manifold distributes the
influent and collects the backwash water. The lower manifold collects filtered water and distributes
backwash water. There are numerous designs for these distribution manifolds. Smaller diameter
filters may have a distribution plate or a hub-lateral design shown in Figure 7-4. Larger diameter
columns may have a header-lateral design, shown in Figure 7-5.  The header-lateral design gives a
more even distribution of the flow, which is much more important for the lower manifold, as flow
distribution directly affects the effectiveness of the backwash.
               Figure 7-4. Hub-Lateral Distribution System (Johnson Screens).
              Figure 7-5. Header-Lateral Distribution System (Johnson Screens).

Typical media filtration installations include several filters in parallel. This allows one to be taken
offline while the others continue to work. It also allows the other filters to provide the backwash
water necessary to backwash a single filter.  Figure 7-6 shows one potential valving arrangement
that allows the use of multiple filters.  Figure 7-7 and Figure 7-8 show pictures of commercially
available pressurized media filters.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                               111

-------
      Inlet
     Water
 Filter-To -Waste
                                                               Backwash
              1           11           11           1,
                        Figure 7-6. Multiple Media Filter Setup.
_ Filtered Water to
Distribution System


Backwash Waste
                     Figure 7-7. Pressurized Media Filter (USFilter).
Arsenic Treatment Technology Evaluation Handbook for Small Systems
            112

-------
              Figure 7-8. Pre-Engineered Arsenic Filtration System (Kinetico).

       7.3.1  Filter Diameter
       The primary design variable for the granular media filters is the hydraulic loading rate. This
       is the flowrate the filters handle per horizontal cross-sectional  area of media.  Typical hy-
       draulic loading rates for greensand filters range between 3 and 5 gpm/sft, although, under
       some circumstances, greensand filters can be successfully operated at hydraulic loading
       rates as high as 10 gpm/sft.  Using this information, the number of filters, and the maximum
       flowrate for which the filters are designed (i.e., design flowrate), the filter diameter can be
       calculated using Equation 7-1.
                                  D =
                                         4-Q
                                      Tt-rip-HLR
                                                                                Eqn. 7-1
      Where:
          D
          Q
          n
          HLR
: Column Diameter (ft),
: Design Flowrate (gpm),
: Number of Parallel Treatment Trains, and
: Hydraulic loading rate (gpm/sft).
      For the example of 3 parallel filters designed to treat a maximum of 300 gpm of water at a
      filter hydraulic loading rate of 5 gpm/sft, the filter diameter should be 5 feet.
                                  D
                    '            \l
                      4-300gpm  |
                    7i • 3 • 5 gpm/sft I
                                                  0.5
= 5ft
       7.3.2  Media Weight
       The weight of each media layer can be calculated using the following equation:
                                       Ji-D'
                                                                                Eqn. 7-2
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                                   113

-------
Where:
   Wj    = Weight of Media Layery (Ibs),
   D     = Column Diameter (ft),
   h      = Height of Media Layery (ft), and
   Pj     = Bulk Density of Mediay (Ibs/cft).

For the previously calculated 5ft filters, using 1 ft of anthracite, 2.5 ft of greensand, 0.25 ft
of filter garnet, and 2 ft of support granite, the media weights per filter are 982 Ibs of anthra-
cite, 4,172 Ibs of greensand, 687 Ibs of filter garnet, and 3,927 Ibs of support granite, respec-
tively.  Typical densities for each of the media can be found in Table 7-1 .
                                             Ibs of greensand (per filter)
7.4   Waste Handling  System Design

Both the backwash water and the FTW water from granular media filtration processes pose dis-
posal issues. The backwash flowrate can be calculated using the equation:


                                 QBw=^-D2-GBw                            Eqn. 7-3
Where:
    QBW   = Backwash flowrate (gpm),
    GBW   = Backwash flux (gpm/sft), and
    D     = Column Diameter (ft).

The FTW flowrate is typically the same as the flowrate used in the filtration mode. Therefore, the
volume of wastewater produced  by the backwash and FTW modes can be calculated using the
equation:
                                              Q

Where:
    V^^  = Volume of Wastewater (gal),
    QBW   = Backwash Flowrate (gpm),
    t,w    = Backwash Duration (min),
    D W                        \    / •>
    Q     = Design flowrate  (gpm),
    np     = Number of Parallel Treatment Trains, and
    t,TW   = Filter-To-Waste Duration (min).
    r 1W                            ^    '

For example, assume the same 3-filter system as before (5-foot diameter, 300 gpm design flowrate,
and 5 gpm/sft hydraulic loading rate) has  a backwash flux of 12  gpm/sft, a backwash time of 15
minutes, and a FTW time of 5  minutes. The required backwash flowrate is then 236 gpm/filter and
the wastewater volume created is 4,040 gallons per backwash per filter.


Arsenic Treatment Technology Evaluation Handbook for Small Systems                              114
                                                                            7.4
                                       np

-------
                        QBW = J'(5ft)2  12-^f ]=236gpm(perfilter)
V 300gPmT5   "^
       Vww =  236-15                     5           4,040
                  -                                          ,    -     -
                  Filter 1 1   Backwash I I  3 Filters 1  Backwash I       Filter • Backwash

The wastewater can be disposed of in several different ways.  The two most probable methods are
indirect disposal through a POTW or by settling the solids and recycling the supernatant and send-
ing the solids to a landfill.

In the indirect discharge through a POTW, a holding tank may be desired to eliminate the surging to
the POTW system.  In the liquid recycle/solids disposal method, a settling tank or basin is required.
The holding basin or tank should be sized to hold at least two backwash/FTW cycles. In the above
example, this leads to an 8,100 gallon tank.

7.5    Coagulant Addition System Design

The efficiency of arsenic co-precipitation to iron floe may vary depending on the concentration of
iron and the iron:arsenic ratio. Optimal performance is obtained with an iron:arsenic mass ratio of
at least 20: 1 .  If the  raw water does not meet these two parameters, iron addition may be required to
provide enhanced coagulation.  Ferric chloride (FeCl3)  is commonly available for use in potable
water systems and can be obtained as a 38wt% liquid.  The volumetric flowrate of ferric chloride
solution required to meet a predetermined dose rate can be calculated with Equation 7-5.


                                    Q'งFeCl3    (       kff-mT ^
                            QFeCl3 = - - - --  0.003785^^                   E    ?_5
                                   CFeci3-PFeCl3  (,       mg-galj
Where:
   Qpeci3 = Ferric Chloride Metering Pump Rate (mL/min),
   Q     = Design flowrate (gpm),
   งFeci3  = Ferric Chloride Dose (mg/L),
   Cpeci3 = Ferric Chloride Stock Solution Concentration (wt%), and
   PFeci3  = Density of Ferric Chloride (kg/L).

For example, if the design flow rate of water to be treated was 300 gpm and the water needed an
additional 1.0 mg/L of iron, a 38-wt% solution of ferric  chloride with a density of 1.42 kg/L could
be added to the water at a rate of 6. 1 mL/min to provide the required iron.
                          = (300 gpm>(l       0003785
                      te^    (0.38)-(l.42kg/L)          mg.gal
Arsenic Treatment Technology Evaluation Handbook for Small Systems                               115

-------
The required storage capacity for the ferric chloride solution can be calculated using Equation 7-6.
Where:
    V
    Q
    ฐFeCl3

    t

    CpeCl3

    PFeCl3
                      CFeCl3-PFeCl3
                                 .  0.00144
                                                      kp-min
                                                       g
                                                       mg-d
                                                                       Eqn. 7-6
Storage Volume (gal),
Design flowrate (gpm),
Dose Rate of Ferric Chloride (mg/L),
Storage Time (days),
Ferric Chloride Stock Solution Concentration (wt%), and
Density of Ferric Chloride (kg/L).
Using the same example and specifying 14 days of ferric chloride storage, the required storage
volume would be 32.6 gallons.
                  y=
         (300gpm)-(l4days)-(l rr
              (0.38)-(l.42kg/L)
                                                      mg-d

A generalized flow diagram for a ferric chloride chemical addition system is shown in Figure 7-9.
The ferric chloride should be stored in a tank made of either fiberglass-reinforced polyester or
rubber-lined steel. A flow meter installed along the main water line is used to pace the addition of
ferric chloride to the water flowrate.  An isolation valve and check valve are used in the connection
to the water line. After the ferric chloride addition, the water is mixed with an inline mixer and the
dosed water is sent to the filters.
                                                                     Water to
                                                                      Filters
water irom rre- r—
Oxidation \L

— Flow
Meter
t
/vv
Static
Mixer




w

.
r
Ferric
Chloride
Storage
                                  Metering
                                   Pump
                                FM|

                                Flow
                               Meter
                     Figure 7-9.  Ferric Chloride Addition Flow Diagram.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                                                                            116

-------
                                                                 Section  8
                                     Point-Of-Use Treatment
POU devices were approved as SSCTs for meeting the revised arsenic MCL. POU devices are
attractive for removing contaminants that pose (solely) an ingestion risk, as is the case with arsenic.
This is because a very small fraction of the total water supplied to a given household is ultimately
consumed. In most cases, the POU unit is plumbed into the kitchen faucet. As such, the kitchen tap
would be the only source from which water should be collected for consumption.

The primary advantage of employing POU treatment in a small system is reduced capital and treat-
ment costs, relative to centralized treatment.  On the downside, however, it is the utility's responsi-
bility to maintain equipment. Therefore, these programs generally incur higher administrative and
monitoring costs to make sure that all units are functioning properly.  POU programs are an  eco-
nomically viable alternative to centralized treatment for systems serving up to 500 people.

Another downside is that the media or membranes used in POU treatment devices may be suscep-
tible to microbial colonization. Higher levels of bacteria have been found in the finished water
produced by some POU treatment devices, particularly those that incorporate an activated carbon
element, than in the corresponding untreated water. Although no  illnesses have been reported as a
result of the use of these treatment devices, the health effects of these bacteria are still unknown.
Therefore, additional monitoring and post-treatment disinfection may be required to ensure  cus-
tomer safety, increasing overall costs.

The primary criteria for selecting an appropriate POU treatment device are arsenic removal perfor-
mance and cost.  Additional considerations include third party certification to NSF/ANSI stan-
dards, appropriate mechanical warning devices, and ease of serviceability.

8.1   Treatment Alternatives

The technologies that are most amenable to POU treatment include column adsorption with  AA,
IBS, or RO with pre-filtration. The decision trees in Section 3 lead to the most appropriate POU
technology among these choices.

      8.1.1  Adsorption Point-of-Use Treatment
      While finished water pH values will likely be much higher than the optimal pH for activated
      alumina (pH 6.0), it can be operated on a disposable basis at higher pH values. Modified
      AA and IBS provide improved treatment capacity  across  a broader pH range, and may be
      preferred depending upon the cartridge replacement frequency selected by the system.  Col-
      umn operation has the advantages of simple operation, low maintenance, low relative cost,
      small under-the-counter footprint, and high treatment capacity.   Additionally, the break-
Arsenic Treatment Technology Evaluation Handbook for Small Systems                              117

-------
       through kinetics of sorption technologies are slow and more readily detected by routine
       monitoring.

       Figure 8-1 shows how POU adsorption equipment is typically connected to kitchen plumb-
       ing.
                                 o
                                 o
                   Figure 8-1.  Point-Of-Use Adsorption Setup (Kinetico).

       Adsorption columns are typically operated to a set volume to prevent arsenic leakage. This
       is accomplished through the use of a metered cartridge that provides flow totalization and
       will automatically shut-off water flow once the unit reaches the prescribed volume limit.
       Figure 8-2 shows a cross-section of one manufacturer's adsorption cartridge.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
118

-------
                                              Turbint
                    Feed Water
                       Inlet
                     Gearing-
                     Mensural
                     Shut-off
                     Assembly
Filtered Water
   Outlet
 Flow Control
 Filter
                                                          Arsenic
                                                          Media
                    Figure 8-2. Metered Automatic Cartridge (Kinetico).

       8.1.2   Reverse Osmosis Point-of-Use Treatment
       RO POU devices are recommended for treating arsenic-rich water containing high levels of
       sulfates, phosphates, or total dissolved solids.  When operating at typical tap pressures, RO
       devices commonly achieve greater than 95% As(V) rejection at a water recovery of  10-
       25%.  Most units are designed with pre- and post-filters.  Pre-filtration through granular
       media is applied to reduce solids loading and extend membrane life.  For chlorine sensitive
       membranes, pre-filtration typically utilizes a  dechlorinating media  such as granular acti-
       vated carbon.  Post-filtration utilizes carbon or arsenic adsorbent media and serves as a final
       guard step.

       Although the cost of RO POU devices is relatively high compared to other possible options,
       the immediate improvement of the overall water quality could make it very attractive to
       customers.  The potential disadvantages associated with RO systems include poor water
       recovery, disposal of the reject stream, and high capital cost.

       The most common types of membranes used for RO applications are cellulose acetate, thin-
       film polyamide composites, and sulfonated polysulfone. The membranes are manufactured
       in various forms, including tubes, sheets, and hollow fibers.  The membrane is then con-
       structed into a cartridge called an RO module, either spiral wound or hollow fiber.
Arsenic Treatment Technology Evaluation Handbook for Small Systems
                            119

-------
       Most RO POU devices operate at tap water pressure, and therefore have relatively poor
       water recoveries. Permeate is sent to a bladder tank large enough to meet on-demand re-
       quirements. Typical production rates range from 5-15 gpd.

       Over time, the membrane surface will require cleaning in order to maintain performance.
       This capability is built in to most RO devices. Depending on the specific design, the water
       source for washing the membrane surface may either be feed water or permeate.

       Figure 8-3 shows how RO POU equipment is typically connected to kitchen plumbing.
              HOUSEHOLD
              WATER IN
                Figure 8-3. Point-Of-Use Reverse Osmosis Setup (Kinetico).
Arsenic Treatment Technology Evaluation Handbook for Small Systems
120

-------
8.2    Implementation Considerations

Amendments to the SDWA in 1996 explicitly allow utilities to install POU treatment devices to
achieve compliance with the NPDWRs. More information on the implementation of a POU pro-
gram can be found in the USEPA's Guidance for Implementing a Point-of-Use or Point-of-Entry
Treatment Strategy for Compliance with the Safe Drinking Water Act (USEPA, 2002b Draft).

The implementation of a centrally managed POU program is very different from application of
centralized treatment. In many cases, the customer's acceptance of the treatment unit is affected by
familiarity with the technology, the need for treatment, the appearance of the unit, and other subj ec-
tive factors.  Many homeowners currently employ some form of POU treatment such as carbon
filtration or water softening. These products are generally used to enhance aesthetic properties of
water, and are therefore used voluntarily.  Under a centrally managed POU treatment program, all
customers would be required to employ treatment devices in their home.  As such, utility staff or
contractors would need access inside individual homes to install treatment devices, make plumbing
modifications, and make periodic O&M checks. The  extent of customer acceptance and potential
for resistance associated with this utility-customer interface are not well known.

       8.2.1   Program Oversight
       POU units must be owned, controlled, and maintained by the public water system or
       by a contractor hired by the public water system to ensure proper operation and main-
       tenance of the device and compliance with MCLs. The utility must retain oversight of
       unit installation, maintenance, and sampling.  While this provision does not require the
       utility to perform all maintenance or management functions - utilities are free to contract
       out these tasks - it does imply that the utility retains final responsibility of the quality and
       quantity of the water provided to the service community and must closely monitor all con-
       tractors. Further, the utility may not delegate its responsibility for the operation and main-
       tenance of POU devices installed as part of a compliance strategy to homeowners.

       The implications of this requirement to the utility are significant.  The utility must decide
       whether it wants to implement the POU program in-house or contract out the necessary
       services.  In one case, the utility would be the main contact with the customer, and utility
       staff would be responsible for installation, monitoring, record-keeping, and O&M activi-
       ties.  This raises several important issues. First, many small utilities often have difficulty
       finding the time and budget to hire, train,  and retain operators. Second, utilities that elect to
       keep the work in-house must provide staff training on installation and O&M procedures.
       Third, the utility should consider the liability implications of entering individuals' homes to
       conduct work.  If the utility decides to contract out the services, the vendor would be the
       main contact with the customer and the utility  would need to monitor the contractor.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                               121

-------
       8.2.2  Cost
       There are a number of cost elements involved in conducting a POU program.  These in-
       clude:

       •  Capital cost of POU devices. The typical cost ranges of RO devices and adsorption
          units are $300-$ 1,000 and $100-$300 each, respectively.
       •  Installation labor.  Installation of each device is anticipated to take 30 to 60 minutes
          assuming no significant plumbing modifications are necessary.
       •  Installation parts
       •  Replacement parts. Carbon-based pre-filters typically cost between $15-50. Newmem-
          branes typically cost about $150.
       •  Water quality analyses. Arsenic can be measured by a commercial laboratory for ap-
          proximately $10-$20 per sample.
       •  O&M labor.

       8.2.3  Compliance Monitoring
       The current approach is that compliance monitoring would be conducted for each and every
       installed POU device, though only one-third within the same year. A representative moni-
       toring approach that requires less frequency monitoring is under evaluation.  States may
       have more stringent monitoring requirements.  Samples can be taken by the utility or the
       contractor.

       8.2.4  Mechanical Warnings
       Each POU treatment device installed as part of a compliance strategy must be  equipped
       with a warning device (e.g.,  alarm, light, etc.) that will alert users when their unit is no
       longer adequately treating their water. Alternatively, units may be equipped with an auto-
       matic shut-off mechanism to meet this requirement. Several communities have implemented
       POU treatment strategies using units equipped with water meters and automatic shut-off
       devices to disable the units after a specified amount of water has been treated to prevent
       contaminant breakthrough.

       8.2.5  Operations and Maintenance
       Periodic maintenance is necessary to ensure that the devices are functioning properly and
       producing tap water in compliance with the arsenic MCL.  O&M activities consist of both
       regular scheduled tasks as well as emergency troubleshooting responses.

       The sorbent  media or RO membrane should be replaced periodically either on  a set fre-
       quency or based on monitoring and tracking use. Both replacement schedules should be
       based on pilot testing  results. The Arsenic Rule also stipulates that the POU device be
       equipped with mechanical warnings to ensure that customers are automatically notified of
       operational problems. Many devices include a programmable indicator that tracks cumula-
       tive water use, and serves as  a convenient visual guide for the remaining life of the POU
Arsenic Treatment Technology Evaluation Handbook for Small Systems                               122

-------
       device. However, it is not recommended that the utility depend solely on the customer for
       POU servicing.  Rather, there should be an established schedule that is made public to the
       community and adhered to.

       8.2.6 Customer Education and Residential Access
       Utilities should attempt to educate the public prior to implementing a POU strategy. This
       education may include public hearings, water bill inserts, posters, or notices in print or on
       radio or TV. When presented with the facts, most people will happily provide the water
       utility with access to ensure their ongoing effectiveness.

       To address the possibility that an individual or a group of individuals may refuse to provide
       utility personnel with the necessary  access, the utility may need to convince the local gov-
       ernment to pass an ordinance guaranteeing water utility personnel access to service treat-
       ment units. To meet the legal responsibility to provide water in compliance with all NPDWRs,
       the utility may also have to pass an ordinance that requires customers to use POU treatment
       units, and that provides the utility with the authority to shut off a customer's water if the
       customer refuses to allow installation and maintenance of, tampers with, bypasses, or re-
       moves the treatment unit.

       To minimize the burden associated with gaining access to individual residences, POU sam-
       pling  should be coordinated with routine maintenance.  Reducing the number  of house
       visits will reduce administrative costs and travel time, resulting in substantial cost savings
       as well reducing the disruption to the residents.

       8.2.7 Residual Oxidant in Distribution System
       In order to effectively use many sorbent type POU devices, the arsenic must be in its As(V)
       form as it  is treated at the tap. RO type POU devices may also work more efficiently if
       arsenic is oxidized.  This may require installation of a POU oxidation unit or centralized
       oxidation.  If anoxic  conditions occur in the distribution system,  there is a potential for
       arsenic to reduce back to the As(III) state. This would drastically decrease the effectiveness
       of most of the sorptive type POU devices.  Therefore, maintaining an adequate residual
       oxidant in  the distribution system is important.

       8.2.8 Waste Handling
       The type of waste produced from a POU device will depend  on the treatment employed.
       RO treatment will produce a continuous liquid  waste stream (i.e., retentate) that should be
       suitable for disposal in an on-site or community sewerage system (see  Section 2.4). Con-
       versely, with column adsorption treatment, the only waste  is exhausted media,  which is
       produced on a periodic basis.

       Because the solid residuals generated by POU units are collected from individual  house-
       holds, these wastes may be exempt from federal regulation as hazardous wastes, regardless
       of their toxicity.  However, state regulations and each state's implementation of federal
       regulation can vary.  In the case of liquid wastes, local wastewater treatment plants may
Arsenic Treatment Technology Evaluation Handbook for Small Systems                                123

-------
       issue their own limits for the disposal of arsenic.  It is anticipated that this waste will not
       exceed the TC characteristics and will be disposable in a municipal landfill. Additionally,
       POU manufacturers or vendors may also provide waste disposal services for the POU de-
       vices.

8.3   Device Certification

To meet the requirements of the SDWA, POU devices installed as part of a compliance strategy
must be certified according the American National Standards Institute (ANSI) standards, if a stan-
dard exists for that type if device. RO POU devices must be certified as per ANSI/NSF 58 (2002)
- Reverse Osmosis Drinking Water Treatment Systems. POU devices utilizing a sorption technol-
ogy such as AA or an IB S must be certified as per ANSI/NSF 53 (2002) - Drinking Water Treatment
Units-Health Effects.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                               124

-------
                                                               Section  9
                                                            References
AwwaRF (2000). Arsenic Treatability Options and Evaluation of Residuals Management Issues,
      Amy, G.L., M. Edwards, P. Brandhuber, L. McNeill, M. Benjamin, F. Vagliasindi, K.
      Carlson, and J. Chwirka. Awwa Research Foundation, Denver, CO.

AwwaRF (2002). Implementation of Arsenic Treatment Systems-Part 1. Process Selection,
      Chowdhury, Z., S. Kommineni, R. Narasimhan, J. Brereton, G. Amy, and S. Sinha. Awwa
      Research Foundation, Denver, CO.

Clifford, Dennis (1999).  Presentation at Arsenic Technical Work Group. Washington, D.C.

Clifford, Dennis (2001).  Arsenic Treatment Technology Demonstration Drinking Water
      Assistance Program for Small Systems, Final Report to the Montana Water Resources
      Center, March 2001.

Fields, Keith, Abraham Chen, and Lili Wang (2000a). Arsenic Removal from Drinking Water by
      Coagulation/Filtration and Lime Softening Plants, EPA 600R00063, Prepared by Battelle
      under contract 68C70008 for U.S.  EPA ORD, June 2000.

Fields, Keith, Abraham Chen, and Lili Wang (2000b). Arsenic Removal from Drinking Water by
      Iron Removal Plants, EPA 600R00086, Prepared by Battelle under contract 68C70008 for
      U.S. EPA ORD, August 2000.

Ghurye,  Ganesh and Dennis Clifford (2001). Laboratory Study on the Oxidation ofAs(III) to
      As(V), EPA 600R01021, Prepared  under contract 8CR311-NAEX for EPA ORD, March
      2001.

Hanson,  Adrian, Tared Bates, Dean Heil, Andrew Bistol (1999). Arsenic Removal from Water
      Using Greensand: Laboratory Scale Batch and Column Tests. New Mexico State
      University, Las Cruces, NM, June  1999.

Kempic, Jeffery (2002), Teleconference on October 29, 2002.

MacPhee, Michael J., Gail E. Charles, and David A Cornwell (2001).  Treatment of Arsenic
      Residuals from Drinking Water Removal Processes, EPA 600R 01033, Prepared by
      Environmental Engineering & Technology, Inc. under contract 8CR613-NTSA for EPA
      ORD, June 2001.

National Research Council (NRC) (1999). Arsenic in Drinking Water. National Academy Press,
      Washington,  D.C.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                              125

-------
NSF International (200 la). Environmental Technology Verification Report: Removal of Arsenic
       in Drinking Water - Hydranautics ESPA2-4040 Reverse Osmosis Membrane Element
       Module, NSF 0120EPADW395, March 2001.

NSF International (200 Ib). Environmental Technology Verification Report: Removal of Arsenic
       in Drinking Water - KOCH Membrane Systems TFC - ULP4 Reverse Osmosis
       Membrane Module, NSF 0125EPADW395, August 2001.

Rubel, Frederick, Jr. Design Manual - Removal of Arsenic from Drinking Water Supplies by Ion
       Exchange, EPA DRAFT.

Rubel, Frederick, Jr.  Design Manual - Removal of Arsenic from Drinking Water Supplies by
       Adsorptive Media, EPA 600-R-03-019, 2003.

Selvin N., Messham G., Simms J., Pearson I, and Hall J. (2000). The Development of Granular
       Ferric Media - Arsenic Removal and Additional Uses in Water Treatment. Proceedings of
       the AWWA Water Quality Technology Conference, Salt Lake City.

Tumalo, Jamie (2002). U.S. Filter. Personal Conversation with Andrew Hill.

United States Environmental Protection Agency (1998).  Variance Technology Findings for
       Contaminants Regulated Before 1996, EPA 815R98003, September 1998.

United States Environmental Protection Agency (2000).  Technologies and Costs for Removal of
       Arsenic from Drinking Water, EPA 815R00028, Prepared by Malcolm Pirnie, Inc. under
       contract 68C60039 for EPA ORD, December 2000.

United States Environmental Protection Agency (2001). Federal Register, Final Arsenic Rule,
       40CFRParts9, 141, and 142.

United States Environmental Protection Agency (2002a).  Implementation Guidance for the
       Arsenic Rule - Drinking Water Regulations for Arsenic and Clarifications to Compliance
       and New Source Contaminants Monitoring, EPA 816K02018, August 2002.

United States Environmental Protection Agency (2002b Draft). Guidance on Implementing a
       Point-of-Use or Point-of-Entry Treatment Strategy for Compliance with the Safe
       Drinking Water Act, EPA xxxx02xxx DRAFT, March 2002.

Wang, Lili, Abraham Chen, and Keith Fields (2000). Arsenic Removal from Drinking Water by
       Ion Exchange and Activated Alumina Plants, EPA 600R00088, Prepared by Battelle
       under contract 68C70008 for U.S. EPA ORD, October 2000.
Arsenic Treatment Technology Evaluation Handbook for Small Systems                               126

-------
Office of Water (4606M)
EPA816-R-03-014
July 2003
www.epa.gov/safewater
                                               Printed on Recycled Paper

-------