EPA/600/R-11/090
September 2011
COSTS OF ARSENIC REMOVAL TECHNOLOGIES FOR
SMALL WATER SYSTEMS:
U.S. EPA ARSENIC REMOVAL TECHNOLOGY DEMONSTRATION PROGRAM
by
Lili Wang
Abraham S.C. Chen
ALSA Tech, LLC
Powell, OH 43065-6082
to
Battelle
Columbus, OH 43201-2693
Contract No. EP-C-05-057
Task Order No. 0019
for
Thomas J. Sorg
Task Order Manager
Water Supply and Water Resources Division
National Risk Management Research Laboratory
Cincinnati, Ohio 45268
National Risk Management Research Laboratory
Office of Research and Development
United States Environmental Protection Agency
Cincinnati, Ohio 45268
-------
DISCLAIMER
The work reported in this document is funded by the United States Environmental Protection Agency
(EPA) under Task Order (TO) 0019 of Contract EP-C-05-057 to Battelle. It has been subjected to the
Agency's peer and administrative reviews and has been approved for publication as an EPA document.
Any opinions expressed in this paper are those of the author(s) and do not, necessarily, reflect the official
positions and policies of the EPA. Any mention of products or trade names does not constitute
recommendation for use by the EPA.
-------
FOREWORD
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the Nation's
land, air, and water resources. Under a mandate of national environmental laws, the Agency strives to
formulate and implement actions leading to a compatible balance between human activities and the ability
of natural systems to support and nurture life. To meet this mandate, EPA's research program is
providing data and technical support for solving environmental problems today and building a science
knowledge base necessary to manage our ecological resources wisely, understand how pollutants affect
our health, and prevent or reduce environmental risks in the future.
The National Risk Management Research Laboratory (NRMRL) is the Agency's center for investigation
of technological and management approaches for preventing and reducing risks from pollution that
threaten human health and the environment. The focus of the Laboratory's research program is on
methods and their cost-effectiveness for prevention and control of pollution to air, land, water, and
subsurface resources; protection of water quality in public water systems; remediation of contaminated
sites, sediments and groundwater; prevention and control of indoor air pollution; and restoration of
ecosystems. NRMRL collaborates with both public and private sector partners to foster technologies that
reduce the cost of compliance and to anticipate emerging problems. NRMRL's research provides
solutions to environmental problems by: developing and promoting technologies that protect and improve
the environment; advancing scientific and engineering information to support regulatory and policy
decisions; and providing the technical support and information transfer to ensure implementation of
environmental regulations and strategies at the national, state, and community levels.
This publication has been produced as part of the Laboratory's strategic long-term research plan. It is
published and made available by EPA's Office of Research and Development to assist the user
community and to link researchers with their clients.
Sally Gutierrez, Director
National Risk Management Research Laboratory
in
-------
EXECUTIVE SUMMARY
As part of the Arsenic Rule Implementation Research Program, between July 2003 and July 2011, the
U.S. Environmental Protection Agency (EPA) conducted 50 full-scale demonstration projects on
treatment systems removing arsenic from drinking water in 26 states throughout the U.S. The projects
were conducted to evaluate the performance, reliability, and cost of arsenic removal technologies selected
for demonstration and to determine their effects on water quality in distribution systems. A key objective
was to collect cost and performance data that might be used by small water systems, engineering firms,
and state agencies to make informed decisions on selecting appropriate arsenic treatment technologies to
achieve the revised arsenic maximum contaminant level (MCL) of 10 |og/L. While results from each
demonstration are documented in individual technology performance evaluation reports, this report
summarizes cost data across all demonstrations grouped by the technology type. For each type of
technologies, a brief overview of demonstration sites, demonstration technologies, system designs and
configurations, and system operations was provided to assist in understanding relevant cost data.
The arsenic demonstration program was divided into three rounds of projects: Round 1 (12 projects),
Round 2 (28 projects), and Round 2a (10 projects). Treatment systems selected for demonstration
included 28 adsorptive media (AM) systems, 18 iron removal (IR) and coagulation/filtration (CF) systems
(including four using IR pretreatment followed by AM), two ion exchange (IX) systems, and one each
reverse osmosis (RO), point-of-use (POU) RO, POU AM, and system/process modification. Among the
50 locations, 42 were community water systems (CWS) and eight were non-transient non-community
water systems (NTNCWS).
The capital cost of each treatment system was broken down into three components - equipment, site
engineering, and installation, and was divided by its design capacity in gallons per minute (gpm) or
gallons per day (gpd) for comparison among systems. The unit capital cost expressed per 1,000 gal of
water treated was also compared based on a 7% interest rate, a 20-year return period, and the system's
maximum (assuming 100 % utilization rate) and average annual production rates. Factors affecting the
capital cost included system flowrate, vessel design, material of construction, media type and quantity,
pre- and/or post-treatment requirements, and level of instrumentation and controls.
The operation and maintenance (O&M) cost for each treatment system was categorized into media
replacement (AM systems only), chemical consumption, electricity, and labor. O&M costs might be
affected by source water quality and other technology-specific factors, such as arsenic adsorptive
capacities for AM technologies. Building construction and residual handling and disposal were outside of
the scope of this program so their costs are not included in this report (except for spent media disposal
cost).
Costs of AM Technology
Nine different AM products were used by 28 systems: three iron-based media, either ferric oxide (ARM
200 and E33) or ferric hydroxide (GFH®); four iron-modified media, either alumina-based (A/I Complex
2000 and AAFS50), silica-based (G2®), or resin-based (ArsenXnp); one titanium oxide-based media
(Adsorbsia™ GTO™); and one zirconium oxide-based media (Isolux™). All of the media have NSF
Standard 61 certification for use in drinking water applications.
Design flowrates of the AM systems ranged from 10 to 640 gpm. Total capital investment costs for the
systems ranged from $14,000 to $305,000 and varied by flowrate, system design, material of
construction, monitoring equipment, and specific site conditions. Normalized costs ranged from $477 to
$6,171 per gpm or from $0.33 to $4.29 per gpd of design capacity. Unit costs of total capital investments
IV
-------
ranged from $0.09 to $1.11 per 1,000 gal of water treated (assuming a 100% utilization rate). Generally,
the unit cost decreased as the size of the treatment system increased. Equipment costs for the treatment
systems ranged from $8,640 to $218,000, representing an average of 70% of the total capital investment
cost. Site engineering costs for the treatment systems ranged from $1,800 to $50,659, accounting for
14% of the total capital investment (on average). Installation costs for the treatment systems ranged from
$2,610 to $61,209, which accounted for 12 to 34% of the total capital investment (or 16% on average).
System performance was evaluated for a period of 14 to 45 months with more extensive sampling and
analysis conducted during the first 12 to 18 months and less thereafter. Spent media were replaced for 15
systems (or 54% of the AM systems), thus providing ample data for the O&M cost. The remaining 46%
systems did not replace media because they had not reached 10-|a,g/L arsenic breakthrough. The media
replacement cost was the majority (79%) of the O&M cost. Media replacement costs varied widely from
$0.30 to $22.05 per 1,000 gal of water treated due to large variations in media cost and media life. Media
costs ranged from $40/ft3 to $678/ft3, depending on the media type and quantity. Affected by media type,
raw water quality, and process condition, lengths of media life to 10-|o,g/L arsenic breakthrough varied
from 3,100 to 80,000 bed volumes (BV).
Chemicals required for system operation at some of the AM sites included carbon dioxide (CO2) and/or
acid/base for pH adjustment and chlorine for pre-oxidation and disinfection. Their costs varied from
negligible to $0.61 per 1,000 gal of water treated. Five sites used CO2 for pH adjustment and their costs
ranged from $0.11 to $0.41 per 1,000 gal of water. Electricity costs for the treatment systems (not
including pumping from wells to treatment plants or re-pumping to distribution systems) ranged from
zero to $0.16 (or $0.03 on average) per 1,000 gal of water treated. Routine, non-demonstration related
labor activities consumed only 10 to 30 min a day, one or several days a week at most of the sites. At a
labor rate of $18.2 to $37.5/hr (averaging $22.4/hr), labor costs per 1,000 gal of water treated varied
significantly from $0.45 to $3.10 for NTNCWS and from $0.03 to $2.36 for CWS, due largely to
variations in annual water production rates at the AM sites. A NTNCWS often had a lower demand and a
lower utilization rate than a CWS. Therefore, the labor cost (per 1,000 gal of water treated) of a small
NTNCWS tended to be higher than that of a large CWS.
Costs of IR/CF Technology
The 18 IR/CF systems demonstrated include 10 IR systems (two requiring supplemental iron addition),
four IR/AM systems, and four CF systems. Each demonstration study was conducted for a period of 12
to 15 months, except at two sites where more extensive studies were performed to troubleshoot system
performance issues. Filter media used included silica sand/anthracite, GreensandPlus™, Birm®, Filox™,
AD26 (AdEdge), AD GS+ (AdEdge), Macrolite® (Kinetico), and Electromedia® I (Filtronics). All media
have NSF Standard 61 certification for use in drinking water applications.
Design flowrates of the IR/CF systems ranged from 20 to 770 gpm. Total capital investment costs ranged
from $55,423 to $427,407, and varied by flowrate, system design (e.g., use contact tank or not), material
of construction, monitoring equipment, and specific site conditions. Normalized costs ranged from $555
to $3,177 per gpm or $0.39 to $2.21 per gpd. Unit costs of the total capital investment ranged from $0.10
to $0.57 per 1,000 gal of water treated (assuming 100% utilization rate). Similar to the AM systems, the
unit costs of the IR/CF systems generally decreased with increasing sizes of the treatment systems.
Equipment costs for the treatment systems ranged from $19,790 to $296,430, representing an average of
60% of the total capital investment. Site engineering costs ranged from $3,850 to $53,435, accounting for
15% of the total capital investment (on average). Installation costs ranged from $12,410 to $132,039,
which accounted for 14 to 36% of the total capital investment (or 25% on average).
-------
Total O&M costs, including the costs for chemical supplies, electricity consumption, and labor, ranged
from $0.07 to $2.90 per 1,000 gal of water treated. Chemicals used for IR/CF system operation included
chlorine, KMnO4, or NaMnO4 for oxidation and disinfection and an iron salt for coagulation. Overall
chemical costs ranged from zero to $0.37 per 1,000 gal of water treated, equivalent to zero to 57% (19%
on average) of the total O&M cost. Iron addition was used at six sites at a dosage of 0.5 to 2.2 mg/L (as
Fe), either as a coagulant or to augment the natural iron for arsenic removal. Costs of iron addition
ranged from $0.01 to $0.07 per 1,000 gal of water treated.
Incremental electricity costs ranged from zero to $0.39 and averaged $0.07 per 1,000 gal of water treated.
Electricity accounted for an average of 19% of the total O&M cost. The routine, non-demonstration
related labor activities consumed only 10 to 30 min per day and 3.4 hr per week (on average). At an
average labor rate of $22.6/hr, labor costs per 1,000 gal of water treated varied from $0.04 to $2.57,
accounting for 18 to 95% (61% on average) of the total O&M cost. A small NTNCWS often had a higher
labor cost (per 1,000 gal of water treated) than a large CWS due to its lower production rate.
Costs of Other Technologies
Other arsenic removal technologies in the demonstration program included IX, RO, POU, and system/
process modification, each being demonstrated at one or two sites. Two IX systems, each at a design
flowrate of 250 and 540 gpm, used a strong base anionic (SBA) exchange resin to remove both arsenic
and nitrate from source water. The capital investment cost of the 250-gpm system was $286,388, which
included $173,195 for equipment, $35,619 for site engineering, and $77,574 for installation, equivalent to
61%, 12%, and 27% of the total capital cost, respectively. The capital investment cost of the 540-gpm
system was $395,434, which included $260,194 for equipment, $49,840 for site engineering, and $85,400
for installation, equivalent to 66%, 13%, and 22% of the total capital cost, respectively. The normalized
capital cost was $l,146/gpm ($0.80/gpd) for the 250-gpm system and $732/gpm ($0.51/gpd) for the 540-
gpm system. Unit costs were $0.21 and $0.13 per 1,000 gal of treated water (100 % utilization rate),
respectively. Total O&M costs were $0.62 and $0.35 per 1,000 gal of water treated, respectively. Salt
was a major operating cost for the IX systems, accounting for 80% of the total O&M cost. Optimizing
salt loading for system regeneration and adding more salt storage capacities to allow for full truckload
delivery could reduce the salt cost. Electricity costs were $0.08 and $0.03/1,000 gal of water treated,
respectively. Labor costs were $0.05 and $0.03/1,000 gal of water treated, respectively. The electricity
and labor costs accounted for 20% of the total O&M cost.
An innovative approach using POE RO coupled with a dual plumbing distribution system was
demonstrated at one NTNCWS as a low cost alternative to achieve simultaneous compliance with the
arsenic and antimony MCLs. With installation of a dual distribution system, only a portion of raw water
needed to be treated for potable use (i.e., kitchen sinks, water fountains, etc.). Therefore, a smaller RO
system could be used to meet the potable water demand, thus reducing the capital and O&M costs. The
capital investment for the system was $20,452, including $8,600 for the dual plumbing system and
$11,942 for a 1,200-gpd RO system. The normalized cost was $17.12/gpd or $4.43/1,000 gal of water
treated. The total annual O&M cost was $1,404, including $351 for repairs, $376 for electricity
consumption, and $666 for labor cost. The annual cost was $12.89/1,000 gal of permeate water produced.
Nine POU RO units were demonstrated at a CWS with nine participating residences to remove arsenic,
nitrate, and uranium from source water. Water softeners were used for pre-treatment. The cost of each
RO unit was $1,220, including $1,025 for equipment and $195 for installation. The cost of each water
softener was $2,395, including $1,585 for equipment and $810 for installation. The one-year O&M cost
included $115 for the salt supply and $86.50 for pre-and post-filter replacement, totaling $201.50 or $17
per month.
VI
-------
Eight POU cartridges containing ARM 200 media were evaluated either under a sink or inside a drinking
water fountain in different buildings at a university. Upon completion of initial testing, 48 POU E33
cartridges were installed by the school. The cost of each POU ARM 200 and E33 cartridge was $152 and
$215, respectively. Although the cost of the E33 cartridge was 40% higher than that of the ARM 200
cartridge, E33 media was capable of producing up to 3,000 gal of permeate, almost three times higher
than that by ARM 200 media.
Cost Comparison
Capital investment costs for smaller AM and IR/CF systems (with a design flowrate of < 100 gpm) varied
extensively but mean values of the investment for these two technology types were comparable. Capital
investment costs for large AM systems (i.e., >100 gpm) generally were higher than those for IR/CF
systems with similar sizes. For example, average normalized and unit costs for the large AM systems
were 25% and 26%, respectively, lower than those for the large IR/CF systems. IX capital investment
costs were comparable to the IR/CF costs. The large IR/CF and IX systems were more expensive than the
large AM systems because of the use of ancillary equipment and controls, such as contact tanks and iron
addition systems for IR/CF and salt saturators and salt supply systems for IX.
The AM systems had a higher O&M cost than the IR/CF and IX systems, due mainly to media
replacement, which accounted for 79% of the total O&M cost. The lower O&M cost is a significant
advantage of IR/CF over AM as long as the facility can handle IR/CF and IX residuals at a low cost.
Because the O&M cost did not include residuals disposal cost, a key factor in selecting a treatment
technology for arsenic removal, direction comparisons among different technologies would be less
accurate.
The cost for salt constituted a large portion of the O&M cost for IX. Chemical costs for pH adjustment,
(supplemental) iron addition, and pre-oxidation/disinfection was insignificant. The cost for incremental
electricity to overcome headless across filter beds and to power system controls and/or chemical feed
pumps was also insignificant for any of the three technologies. Based on the average weekly labor hours
reported by operators, the AM systems required the least amount of time to operate and maintain.
Although subject to individual operators' opinions, the AM systems required less operator attention and
were easier to operate than the IR/CF and IX systems.
vn
-------
CONTENTS
DISCLAIMER ii
FOREWORD iii
EXECUTIVE SUMMARY iv
FIGURES ix
TABLES x
ABBREVIATIONS AND ACRONYMS xii
ACKNOWLEDGEMENTS xiv
1.0 INTRODUCTION 1
1.1 Purpose and Scope 1
1.2 Background 1
2.0 ADSORPTIVE MEDIA SYSTEMS 6
2.1 Overview of AM Demonstration Sites 7
2.2 Overview of AM Demonstration Technologies 7
2.3 AM System Design and Configuration 8
2.3.1 System Flowrate 9
2.3.2 Tank Design 9
2.3.3 Media Type and Volume 11
2.3.4 Pre-and Post-Treatment 17
2.3.5 Instrumentation and Controls 17
2.4 AM System Capital Investment Costs 19
2.4.1 Total Capital Investment Costs 19
2.4.2 Equipment Cost 25
2.4.3 Site Engineering Cost 25
2.4.4 Installation Cost 25
2.5 AM System O&M Costs 25
2.5.1 Media Replacement Cost 28
2.5.2 Chemical Cost 34
2.5.3 Electricity Cost 34
2.5.4 Labor Cost 34
3.0 IRON REMOVAL/COAGULATION/FILTRATION SYSTEMS 36
3.1 Overview of IR/C/F Demonstration Sites 36
3.2 Overview of IR/CF Demonstration Technologies 38
3.3 IR/CF System Design and Configuration 39
3.3.1 System Flowrate 39
3.3.2 Contact/Detention Tank 39
3.3.3 Filter Design 46
3.3.4 Instrumentation and Controls 46
3.4 IR/CF System Capital Investment Costs 46
3.4.1 Total Capital Investment Costs 46
3.4.2 Equipment Cost 51
3.4.3 Site Engineering Cost 51
3.4.4 Installation Cost 54
3.5 IR/CF System O&M Cost 54
3.5.1 Chemical Cost 54
3.5.2 Electricity Cost 54
3.5.3 Labor Cost 54
Vlll
-------
4.0 OTHER ARSENIC TREATMENT TECHNOLOGIES 57
4.1 Overview of Demonstration Sites 57
4.2 IX Demonstration Systems 59
4.2.1 IX System Design and Configuration 59
4.2.2 IX System Capital Investment Costs 61
4.2.3 IX System O&M Costs 62
4.3 RO Demonstration System 63
4.3.1 RO System Design and Configuration 63
4.3.2 RO System Capital Investment Cost 64
4.3.3 RO System O&M Cost 65
4.4 POU RO Demonstration Units 65
4.4.1 POU RO Unit Design and Configuration 65
4.4.2 POU RO Costs 65
4.5 POU AM Demonstration Units 67
4.5.1 POU AM Cartridge Design and Configuration 67
4.5.2 POU AM Cartridge Costs 67
5.0 COST SUMMARY 69
5.1 Total Capital Investment Costs of Treatment Technologies 69
5.2 O&M Cost of Treatment Technologies 75
6.0 REFERENCES 78
FIGURES
Figure 1-1. Locations of 50 Arsenic Demonstration Projects 2
Figure 2-1A. 20-gpm Adsorbsia™ GTO™ Media System by Siemens 14
Figure 2-1B. 14-gpm As/I Complex 2000 Media System by ATS 14
Figure 2-1C. 40-gpm G2® Media Arsenic Adsorption System by ADI 15
Figure 2-1D. 150-gpm Isolux™-302M Media Arsenic Adsorption System by MEI 15
Figure 2-1E. 160-gpm E33 Media Arsenic Adsorption System by AdEdge 16
Figure 2-1F. 450-gpm E33 Media Arsenic Adsorption System by Severn Trent Services 16
Figure 2-2. Carbon Dioxide Gas Flow Control System for pH Adjustment 18
Figure 2-3. Total Capital Investment Costs of Smaller AM Systems (< 100 gpm) 21
Figure 2-4. Total Capital Investment Costs of Larger AM Systems (=100 gpm) 21
Figure 2-5. AM Treatment System Components at VV by Kinetico 22
Figure 2-6. Backwash Recycling System at VV 22
Figure 2-7. Smaller AM System Capital Investment Costs per gpd of Design Capacity
(<100gpm) 23
Figure 2-8. Larger AM System Capital Investment Costs per gpd of Design Capacity (>100
gpm) 24
Figure 2-9. AM System Unit Costs per 1,000 gal of Water Treated as a Function of Utilization
Rates 24
Figure 2-10. Equipment Costs of Smaller AM Systems (<100 gpm) 27
Figure 2-11. Equipment Costs of Larger AM Systems (=100 gpm) 27
Figure 2-12. E33 Media Loading 28
Figure 2-13. Media Replacement Costs of Various AM 32
Figure 2-14. Media Replacement Costs of 13 E33 Systems 32
Figure 2-15. Hypothetic Media Replacement Cost Curves 33
Figure 3-1A. 20-gpm Macrolite® Pressure Filtration System by Kinetico 42
-------
Figure 3-1B. 35-gpm Birm®/Filox™ and Adsorbsia GTO™ System by Filter Tech 42
Figure 3-1C. 140-gpm Macrolite® Pressure Filtration System by Kinetico 43
Figure 3-1D. 250-gpm AD26/E33 Filtration System by AdEdge 43
Figure 3-1E. 340-gpm AERALATER Filtration System by Siemens 44
Figure 3-1F. 550-gpm Electromedia® I Filtration System by Filtronics 44
Figure 3-2. Chlorine and Iron Addition Systems 45
Figure 3-3. Total Capital Investment Costs of Smaller IR/CF Systems (<100 gpm) 48
Figure 3-4. Total Capital Investment Costs of Larger IR/CF Systems (>100 gpm) 48
Figure 3-5. Smaller IR/CF System Capital Investment Costs per gpd of Design Capacity
(<100gpm) 50
Figure 3-6. Larger IR/CF System Capital Investment Costs per gpd of Design Capacity
(>100gpm) 50
Figure 3 -7. IR/CF System Unit Capital Investment Costs as a Function of Utilization Rates 51
Figure 3-8. Equipment Costs of Smaller IR/CF Systems (<100 gpm) 53
Figure 3-9. Equipment Costs of Larger IR/CF Systems (MOO gpm) 53
Figure 4-1. Photograph of IX-248-As/N System at Fruitland, ID 60
Figure 4-2. EPRO-1,200 RO Unit 63
Figure 4-3. Under-the-Sink RO Plus Deluxe Unit 66
Figure 4-4. POU AM Units Installed Under a Sink (top) and Inside a Drinking Water Fountain
(bottom) 68
Figure 5 -1. Total Capital Investment Costs of Smaller AM and IR/CF Systems (< 100 gpm) 71
Figure 5 -2. Total Capital Investment Costs of Larger AM, IR/CF, and IX Systems (> 100 gpm) 71
Figure 5 -3. Total Capital Investment Costs per gpd of Design Capacity (< 100 gpm) 72
Figure 5 -4. Total Capital Investment Cost per gpd of Design Capacity (> 100 gpm) 72
Figure 5-5. Equipment Costs as a Percentage of Total Capital Investment Cost 73
Figure 5 -6. Engineering Costs as a Percentage of Total Capital Investment Costs 74
Figure 5-7. Installation/Startup Costs as a Percentage of Total Capital Investment Costs 74
Figure 5-8. Smaller System (<100 gpm) Total O&M Costs per 1,000 gal of Water Treated 76
Figure 5-9. Larger System (=100 gpm) Total O&M Costs per 1,000 gal of Water Treated 76
TABLES
Table 1-1. Summary of 50 Arsenic Removal Demonstration Locations, Technologies, and
Source Water Quality 3
Table 1-2. Number of Demonstration Systems for Each Type of Arsenic Removal Technology 5
Table 2-1. Summary of AM Demonstration Locations, Technologies, and Study Durations 6
Table 2-2. Summary of AM Demonstration Sites 8
Table 2-3. Summary of AM Site Source Water Quality 9
Table 2-4. Properties of AM Used for EPA Demonstration Projects 10
Table 2-5. Summary of AM System Design and Components 12
Table 2-6. EBCT vs. Media Type and Tank Configuration 17
Table 2-7. Total Capital Investment Costs for AM Systems 20
Table 2-8. Summary of Equipment, Site Engineering, and Installation Costs of AM Systems 26
Table 2-9. O&M Costs for AM Systems with Media Replacement 29
Table 2-10. Breakdowns of Media Replacement Costs 30
Table 2-11. Replacement Costs of Various Types of AM 31
Table 2-12. Costs of pH Controls for AM Systems 34
Table 3 -1. Summary of IR/C/F Demonstration Locations, Technologies, and Study Durations 36
Table 3-2. Summary of IR/CF Demonstration Sites 37
Table 3-3. Summary of IR/CF Site Source Water Quality 38
-------
Table 3 -4. Characteristics of Filtration Media Used in EPA Demonstration Proj ects 40
Table 3-5. Summary of IR/CF System Design and Components 41
Table 3-6. Filtration Rates of Different Filter Media 46
Table 3-7. Capital Investment Costs for IR/CF Systems 47
Table 3-8. Summary of Equipment, Site Engineering, and Installation Costs of IR/CF Systems 52
Table 3-9. O&M Costs for IR/CF Systems 55
Table 3-10. Cost of Iron Addition for IR/CF Systems 56
Table 4-1. Summary of IX, RO, and POU Demonstration Locations, Technologies, and Study
Durations 57
Table 4-2. Summary of IX, RO, and POU Demonstration Sites 58
Table 4-3. Summary of IX, RO and POU Site Source Water Quality 58
Table 4-4. Properties of IX Resins Used for EPA Demonstration Projects 59
Table 4-5. Summary of IX System Design and Components 60
Table 4-6. Total Capital Investment Costs for IX Systems 61
Table 4-7. Summary of Equipment, Site Engineering, and Installation Costs of IX Systems 61
Table 4-8. O&M Costs for IX Systems 62
Table 4-9. Design Specifications of EPRO-1,200 RO System 64
Table 4-10. RO System Capital Investment Cost 64
Table 4-11. Kinetico RO Plus Deluxe Unit Performance Specifications 66
Table 4-12. Design Specifications of Kinetico and AdEdge POU AM Cartridges 67
Table 5-1. Summary of Total Capital Investment Costs 70
Table 5-2. Summary of O&M Costs 75
XI
-------
ABBREVIATIONS AND ACRONYMS
AM adsorptive media (process)
As arsenic
ASME American Society of Mechanical Engineers
ATS Aquatic Treatment Systems
BV bed volume
CF coagulation/filtration (process)
CO2 carbon dioxide
CRF capital recovery factor
CS carbon steel
CWS community water system
EBCT empty bed contact time
EPA Environmental Protection Agency
Fe iron
FRP fiberglass reinforced plastic
gpd gallons per day
gpm gallons per minute
HDPE high-density polyethylene
hp horsepower
IR iron removal (process)
IX ion exchange (process)
KMnO4 potassium permanganate
MCL maximum contaminant level
MEI Magnesium Elektron, Inc.
MG million gallons
N/A not available
NTNCWS non-transient non-community water system
NSF NSF International
O&M operations and maintenance
OIP operator's interface panel
ORD Office of Research and Development
PE Professional Engineer
PLC programmable logic controller
POE point of entry
POU point of use
PVC polyvinyl chloride
xn
-------
RO reverse osmosis
SBA strong based anionic
SDWA Safe Drinking Water Act
SMCL secondary maximum contaminant level
SS stainless steel
STMGID South Truckee Meadows General Improvement District
STS Severn Trent Services
TCLP Toxicity Characteristic Leaching Procedure
TDS total dissolved solid
THM trihalomethane
TOC total organic carbon
xin
-------
ACKNOWLEDGEMENTS
This report was prepared by ALSA Tech with input from Thomas J. Sorg, EPA's Task Order Manager.
The authors wish to acknowledge the 50 host facilities and the vendors who participated in this
demonstration program and provided cost information. We also acknowledge the Battelle study leads for
their diligence in collecting the cost and performance data during the demonstration studies.
xiv
-------
1.0 INTRODUCTION
1.1 Purpose and Scope
Between July 2003 and July 2011, the U. S. Environmental Protection Agency (EPA) conducted 50 full-
scale demonstration projects on treatment systems removing arsenic from drinking water in 26 states
throughout the U.S. These demonstration projects evaluated the efficiency and effectiveness of the
treatment systems in meeting the new arsenic maximum contaminant level (MCL) of 0.010 mg/L
(10 ng/L). One of the major objectives of the demonstration program was to determine the cost-
effectiveness of the technologies by collecting cost data associated with the 50 systems, including capital
investment costs for equipment, site engineering, and installation, and operation and maintenance (O&M)
costs.
This report summarizes the capital investment and O&M costs associated with the demonstration
systems. Background information on demonstration sites, demonstration technologies, system designs
and configurations is also provided to support the cost data. Building construction and residuals disposal
were outside the scope of the program so their costs were not included. However, residuals disposal
options and costs could affect the technology selection (EPA, 2000; Cornwell and Roth, 2011). Detailed
information on the system performance and cost data can be found in individual final reports posted on
the EPA Web site at http://www.epa.gov/ORD/NRMRL/wswrd/dw/arsenic/index.html.
1.2 Background
The Safe Drinking Water Act (SOWA) mandates that EPA identify and regulate drinking water
contaminants that may have adverse human health effects and that are known or anticipated to occur in
public water supply systems. In 1975, under the SDWA, EPA established a MCL for arsenic (As) at
0.05 mg/L. Amended in 1996, the SDWA required that EPA develop an arsenic research strategy and
publish a proposal to revise the arsenic MCL by January 2000. On January 18, 2001, EPA finalized the
arsenic MCL at 0.01 mg/L (EPA, 2001). In order to clarify the implementation of the original rule, EPA
revised the rule text on March 25, 2003, to express the MCL as 0.010 mg/L (10 (ig/L) (EPA, 2003). The
final rule required all community and non-transient, non-community water systems to comply with the
new standard by January 23, 2006.
In October 2001, EPA announced an initiative for additional research and development of cost-effective
technologies to help small community water systems (< 10,000 customers) meet the new arsenic standard,
and to provide technical assistance to operators of small systems to reduce compliance costs. As part of
this Arsenic Rule Implementation Research Program, EPA's Office of Research and Development (ORD)
proposed a program to conduct a series of full-scale, onsite demonstrations of arsenic removal technology
projects, process modifications, and engineering approaches applicable to small systems.
With EPA program funds and additional funding from Congress during fiscal years 2005, 2006 and 2007,
EPA conducted three rounds of demonstration projects: Round 1 (12 projects), Round 2 (28 projects) and
Round 2a (10 projects). The selections of the treatment technologies were made from solicited proposal
through a joint effort of EPA, respective state regulators, and host sites. Figure 1-1 is a map showing the
locations of the 50 demonstration projects.
Technologies selected for the 50 projects included adsorptive media (AM), iron removal (IR),
coagulation/filtration (CF), ion exchange (IX), reverse osmosis (RO), point-of-use (POU), and
system/process modification. Table 1-1 summarizes the locations, technologies, vendors, system
flowrates, and key source water quality parameters (including As, iron [Fe], and pH). The table is
-------
Treatment
Technology
No. of Systems
AM
28
IR
10
IR + AM
4
C/F
4
IX
2
RO
1
POU
2
System
Mod
1
Figure 1-1. Locations of 50 Arsenic Demonstration Projects
organized by four sections of the country: Northeast/Ohio, Great Lakes/Interior Plains, Midwest/
Southwest, and Far West. Each demonstration location was assigned to a two-letter identification (ID)
code, which was used throughout this report for system identification. Table 1-2 presents the number of
systems for each type of technologies and the section of this report where the cost information is
presented.
This report consists of six sections. Section 1 is a brief introduction. Section 2 presents the cost
information of 28 AM systems demonstrated at 26 sites (one site had three AM systems). Section 3
presents the cost information of 18 IR/CF systems demonstrated at 18 sites, including 10 IR systems
(including two requiring supplemental iron addition), four IR/AM systems, and four CF systems. Section
4 presents the cost information of other technologies each demonstrated at one or two sites using IX, RO,
POU, or system/process modification. Section 5 summarizes and compares the costs for AM, IR/CF, and
IX systems. Section 6 contains a list of references cited in this report.
-------
Table 1-1. Summary of 50 Arsenic Removal Demonstration Locations, Technologies, and Source Water Quality
State
Demonstration
Location
(Two-Letter ID)
Site Name
Technology (Media)
Vendor
Design
Flow rate
(gpm)
Source Water Quality
As
(Mg'L)
Fe
(Mg/L)
PH
(S.U.)
Northeast/Ohio
ME
ME
NH
NH
NH
VT
NY
CT
CT
DE
MD
PA
OH
OH
Carmel (CE)
Wales (WA)
Bow (BW)
Goffstown (GF)
Rollinsford (RF)
Dummerston (DM)
Houghton (HT)(C)
Woodstock (WS)
Pomfret (PF)
Felton (FE)
Stevensville (ST)
Conneaut Lake (CL)
Buckeye Lake (BL)
Springfield (SF)
Carmel Elementary School
Springbrook Mobile Home Park
White Rock Water Company
Orchard Highlands Subdivision
Rollinsford Water/Sewer District
Charette Mobile Home Park
Town of Caneadea
Woodstock Middle School
Seely-Brown Village
Town of Felton
Queen Anne's County
Conneaut Lake Park
Buckeye Lake Head Start Building
Chateau Estates Mobile Home Park
RO
AM (A/I Complex)
AM(G2)
AM(E33)
AM(E33)
AM (A/I Complex)
IR (Macrolite®)
AM (Adsorbsia'™)
AM (ArsenXnp)
CF (Macrolite®)
AM(E33)
CF (AD GS+)
AM (ARM 200)
IR & AM (E33)
Norlen's Water
ATS
ADI
AdEdge
AdEdge
ATS
Kinetico
Siemens
SolmeteX
Kinetico
STS
AdEdge
Kinetico
AdEdge
l,200gpd
14
40w
10
100
22
550
17
15
375
300
250
10
250(eJ
18.2
39.1W
46.4
29.7
37.7
42.2
27W
24.7
25
34.4W
20.1W
29W
15.4(a)
22.7W
<25
<25
<25
<25
297
<25
l,806(d)
27
<25
26
269(d)
188(d)
2,290(d)
l,102(d)
7.9
8.5
7.3
7.1
7.7
7.7
7.6
7.1
7.3
8.3
7.8
7.8
7.4
7.2
Great Lakes/Interior Plains
MI
MI
MI
WI
IN
IN
IL
IL
WI
MN
MN
MN
MN
ND
SD
Brown City (BC)
Pentwater (PW)
Sandusky (SD)
Delavan (DV)
Goshen (GS)
Fountain City (FC)
Waynesville (WV)
Geneseo Hills (GE)
Greenville (GV)
Climax (CM)
Sabin (SA)
Sauk Centre (SC)
Stewart (ST)
Lidgerwood (LW)
Lead (LD)
City of Brown City
Village of Pentwater
City of Sandusky
Vintage on the Ponds
Clinton Christian School
Northeaster Elementary School
Village of Waynesville
Geneseo Hills Subdivision
Town of Greenville
City of Climax
City of Sabin
Big Sauk Lake Mobile Home Park
City of Stewart
City of Lidgerwood
Terry Trojan Water District
AM(E33)
IR/IA (Macrolite®)
IR (Aeralater*1)
IR (Macrolite®)
IR & AM (E33)
IR (G2)
IR (GreensandPlus'™)
AM(E33)
IR (Macrolite®)
IR/IA (Macrolite®)
IR (Macrolite®)
IR (Macrolite®)
IR &AM (E33)
Process Modification
AM (ArsenXnp)
STS
Kinetico
Siemens
Kinetico
AdEdge
US Water
Peerless
AdEdge
Kinetico
Kinetico
Kinetico
Kinetico
AdEdge
Kinetico
SolmeteX
640
400
340(e)
45
25
60
96
200
375
140
250
20
250
250
75
15.3W
17.7W
11. 4W
18.9(a)
28.6W
29.4W
32W
19.6W
5.6W
36.5W
41.8
27.5W
44.8W
146W
22.2
177(d)
426CT
896(d)
l,392(d)
741(d)
1,865W
2,543(d)
554W
2,068(d)
540(d)
l,350(d)
25385(d)
U88(d)
1,325W
<25
7.9
7.9
7.2
7.5
7.3
7.6
7.1
7.2
7.3
7.5
7.3
7.3
7.9
7.2
7.2
Midwest/Southwest
UT
LA
Willard (WL)
Arnaudville (AR)
Hot Springs Mobile Home Park
United Water Systems
IR & AM (Adsorbsia'™)
IR (Macrolite®)
Filter Tech
Kinetico
30
770(e)
13.2
32.7W
276(d)
2,059(d)
7.6
6.8
-------
Table 1-1. Summary of 50 Arsenic Removal Demonstration Locations, Technologies, and Source Water Quality (Continued)
State
TX
TX
TX
NM
NM
NM
AZ
AZ
AZ
Demonstration
Location
(Two-Letter ID)
Alvin (AV)
Bruni (BR)
Wellman (WM)
Anthony (AN)
Nambe Pueblo (NP)
Taos (TA)
Rimrock (RR)
Tohono O'odham
Nation (TN)
Valley Vista (VV)
Site Name
Oak Manor Municipal Utility
District
Webb Consolidated Independent
School District
City of Wellman
Desert Sands Mutual Domestic
Water Consumers Association
Nambe Pueblo Tribe
Town of Taos
Arizona Water Company
Tohono O'odham Utility Authority
Arizona Water Company
Technology (Media)
AM(E33)
AM(E33)
AM(E33)
AM(E33)
AM(E33)
AM(E33)
AM(E33)
AM(E33)
AM (AAFS50/ARM 200)
Vendor
STS
AdEdge
AdEdge
STS
AdEdge
STS
AdEdge
AdEdge
Kinetico
Design
Flow rate
(gpm)
150
40
100
320
145
450
45W
50
37
Source Water Quality
As
fag/L)
40.2(a)
57.6W
36
23.5W
32.2
16.9
59.7
34.9
39.4
Fe
(Mg/L)
63
32
<25
80
<25
31
<25
<25
<25
PH
(S.U.)
7.8
8.2
7.8
7.8
9.0
9.6
6.9
8.0
7.7
Far West
MT
ID
ID
WA
OR
OR
NV
CA
CA
CA
Three Forks (TF)
Fruitland (FL)
Homedale (HD)
Okanogan (OK)
Klamath Falls (KF)
Vale (VA)
Reno (RN)
Susanville (SU)
Lake Isabella (LI)
Tehachapi (TE)
City of Three Forks
City of Fruitland
Sunset Ranch Development
City of Okanogan
Oregon Institute of Technology (OIT)
City of Vale
South Truckee Meadows General
Improvement District
Richmond School District
Upper Bodfish Well CH2-A
Golden Hills Community Service
District
CF (Macrolite®)
IX (A300E)
POU R0(1)
CF (Electromedia-I®)
AM(Adsorbsia /
ARM 200/ArsenXnp)
and POU AM (ARM 200)(g)
IX (Arsenex II)
AM (GFH)
AM (A/I Complex)
AM (ArsenXnp)
AM (Isolux)
Kinetico
Kinetico
Kinetico
Filtronics
Kinetico
Kinetico
Siemens
ATS
VEETech
MEI
250
250
9 unit
550
60/60/30
540
350
12
50
150
84
42.5
57.8
17.9
29.8
22.6
67.2
31.7
41.7
12.7
<25
<25
112
78(d)
<25
<25
<25
37
<25
<25
7.5
7.6
7.3
7.6
8.0
7.4
7.1
8.4
6.9
7.6
AM = adsorptive media process; CF = coagulation/filtration; IR = iron removal; IR/IA = iron removal with iron addition; IX = ion exchange process; RO = reverse osmosis
ATS = Aquatic Treatment Systems; MEI = Magnesium Elektron, Inc.; STS = Severn Trent Services
(a) Arsenic existing mostly as As(III).
(b) Design flowrate reduced by 50% due to system reconfiguration from parallel to series operation.
(c) Selected originally to replace Village of Lyman, NE site, which withdrew in June 2006; withdrew in 2007 and later replaced by residential systems in Lewisburg, OH.
(d) Iron existing mostly as Fe(II).
(e) Facilities upgraded systems in Springfield, OH from 150 to 250 gpm, Sandusky, MI from 210 to 340 gpm, and Amaudville, LA from 385 to 770 gpm.
(f) Including nine under-the-sink units.
(g) Including eight under-the-sink or inside-a-drinking-fountain cartridges.
-------
Table 1-2. Number of Demonstration Systems for Each Type of
Arsenic Removal Technology
Technology Type
Adsorptive Media
Iron Removal (Oxidation/Filtration)
Iron Removal and Adsorptive Media Combined
Coagulation/Filtration
Ion Exchange
Reverse Osmosis
Point-of-Use Reverse Osmosis
Point-of-Use Adsorptive Media
System/Process Modifications
Number of
Systems
28W
10(b)
4
4
2
1
l(c>
j(d)
1
Report
Section
2
3
4
Not included
(a) 28 AM systems demonstrated at 26 sites with one having three AM
systems.
(b) Two IR systems used supplemental iron addition.
(c) Including nine under-the-sink units.
(d) Including eight under-the-sink or inside-a-drinking-fountain cartridges.
-------
2.0 ADSORPTIVE MEDIA SYSTEMS
AM systems were selected at 26 of the 50 demonstration project locations as the main treatment process
for arsenic removal. The 26 water systems consisted of five non-transient non-community water systems
(NTNCWS) and 21 community water systems (CWS). Table 2-1 lists AM demonstration locations,
technologies and vendors, and study durations in the order of design flowrates. Because the Klamath
Falls (KF) site had three POE AM systems, labeled as 4a, 4b, and 4c in Table 2-1, a total of 28 AM
systems were demonstrated at the 26 sites. Performance of each system was evaluated for 14 to 45
months with more extensive sampling and analysis conducted in the first 12 to 18 months and less
thereafter. Detailed information about the performance and capital and O&M costs on each system can
be found in individual performance evaluation reports provided on the EPA Arsenic Demonstration
Program Web site at http://www.epa.gov/ORD/NRMRL/wswrd/dw/arsenic/index.html.
Table 2-1. Summary of AM Demonstration Locations, Technologies, and Study Durations
No.
Site
ID
Demonstration
Location
Technology
(Media)
Vendor
Design
Flowrate
(gpm)
Study
Duration
Length
of Study
(mon)
Non-Transient Non-Community Water Systems
1
2
o
5
4a
4b
4c
5
BL
SU
ws
KF
BR
Buckeye Lake, OH
Susanville, CA
Woodstock, CT
Klamath Falls, OR
Bruni, TX
ARM 200
A/I Complex 2000
Adsorbsia™ GTO™
ArsenXnp
ARM 200
Adsorbsia™ GTO™
E33
Kinetico
ATS
Siemens
Kinetico
AdEdge
10
12
17
30
60
60
40
06/06-02/10
09/05-06/07
02/09-09/10
12/05-08/09
12/05-08/09
02/06-08/09
12/05-05/08
44
21
19
45
45
43
30
Community Water Systems
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
GF
WA
PF
DM
W
BW
RR
LI
TN
LD
WM
RF
TE
AL
NP
GE
SV
AN
RN
TA
BC
Goffstown, NH
Wales, ME
Pomfret, CT
Dummerston, VT
Valley Vista, AZ
Bow, NH
Rimrock, AZ
Lake Isabella, CA
Tohono O'odham
Nation, AZ
Lead, SD
Wellman, TX
Rollinsford, NH
Tehachapi, CA
Alvin, TX
Nambe Pueblo, NM
Geneseo Hills, IL
Stevensville, MD
Anthony, NM
Reno, NV
Taos, NM
Brown City, MI
E33
A/I Complex 2000
ArsenXnp
A/I Complex 2000
AAFS50
G2®
E33
ArsenXnp
E33
ArsenXnp
E33
E33
Isolux™
E33
E33
E33
E33
E33
GFH®
E33
E33
AdEdge
ATS
SolmeteX
ATS
Kinetico
ADI
AdEdge
VEETech
AdEdge
SolmeteX
AdEdge
AdEdge
MEI
STS
AdEdge
AdEdge
STS
STS
Siemens
STS
STS
10
14
15
22
37
40W
45(a)
50
50
75
100
100
150
150
145
200
300
320
350
450
640
04/05-08/07
03/05-08/07
02/09-09/10
06/05-10/06
06/04-08/06
10/04-09/06
06/04-03/07
10/05-03/07
02/08-03/10
04/08-05/10
08/06-04/08
02/04-05/06
10/05-03/07
04/06-04/08
05/07-09/09
05/08-07/10
06/04-04/07
01/04-08/06
09/05-07/07
02/06-10/07
05/04-05/07
28
29
20
16
14
23
33
17
25
25
20
27
17
24
28
26
34
31
22
20
36
ATS = Aquatic Treatment Systems; MEI = Magnesium Elektron, Inc.; STS = Severn Trent Services
(a) Design flowrate reduced by 50% due to system reconfiguration from parallel to series operation.
-------
2.1 Overview of AM Demonstration Sites
Table 2-2 summarizes the AM demonstration site information. All five NTNCWS were schools,
including one university having three point-of-entry (POE) systems loaded with different types of media.
Most of these facilities were classified as very small (serving 25 to 500 of people) and small (serving 501
to 3,300 of people) water systems. The wells supplying the demonstration systems were operated less
than 10 hr/day at most of the sites. Five systems were operated on demand, with varying flowrates
corresponding to momentary water demands in the distribution systems. Average daily demands varied
from 450 to 17,562 gal for NTNCWS and from 1,565 to 152,280 gal for CWS. Annual productions
ranged from 0.1 to 6 million gallons (MG) for NTNCWS and from 0.6 to 51 MG for CWS. The ratio of
the annual production to the system maximum capacity represents a hydraulic utilization rate, varying
from 2 to 19% for NTNCWS and 5 to 96% for CWS.
Source water quality plays an important role in technology selection and design and operation of a
treatment system because it can affect the performance of a technology and treatment cost. Table 2-3
provides average values of several key water quality parameters of source waters treated by the AM
systems. Arsenic concentrations in source waters ranged from 12.7 to 67.2 (ig/L across all 26
demonstration locations. At nine of 26 sites, soluble As(III) was the most prevalent form of arsenic in the
source waters. Among these nine sites, two sites (BL and GE) had total iron levels (primarily as soluble
Fe[II]) above its secondary MCL (SMCL) of 300 ng/L and three sites (BL, RF, and AL) had total
manganese levels above its SMCL of 50 |o,g/L. The BL site had a pre-existing softener that removed iron
and manganese from source water before adsorption. In general, if a source water contains Fe(II) and/or
Mn(II) above the respective MCL, an iron removal (IR) or an IR/AM process mostly likely would be
selected for arsenic and iron removal.
The arsenic removal capacity of an AM is strongly dependent on solution pH. Most AMs adsorb arsenic
more effectively at a pH value of 5.5 to 7.5, and their adsorptive capacities increase with decreasing pH.
Adjusting the pH of raw water can increase the media capacity and lower the operating cost; however, the
pH control equipment increases the system cost and the complexity of operation. Source water pH values
ranged from 6.9 to 9.6 across all 26 demonstration locations. At 17 locations, source water pH values
were higher than 7.5, which led to the use of pH adjustment to lower the pH at seven of these 17 locations
(see Section 2.3.4).
2.2 Overview of AM Demonstration Technologies
Nine different types of media were evaluated, including three iron-based media, either granular ferric
oxide (ARM 200 and E33) or granular ferric hydroxide (GFH®); four iron-modified media, either
alumina-based (A/I Complex 2000 and AAFS50), silica-based (G2), or resin-based (ArsenXnp); one
titanium oxide media (Adsorbsia™ GTO™); and one zirconium oxide media (Isolux™). All of these
media have NSF Standard 61 certification for use in drinking water applications. Over the course of the
study, some newer versions of the media were developed with slight modifications to the older versions.
For example, ARM 300 is a newer version of ARM 200 with a slightly different mesh size and density.
E33-P is a pelletized media, which is 25% denser than its granular counterpart, E33-G (thus, its cost per
cubic foot is higher than E33-G). Both media have a similar arsenic adsorptive capacity on a weight
basis. E33-P was designed for more robust applications such as frequent backwashes, but because of lack
of apparent benefits, the manufacturer had stopped recommending the use of this type of media for
arsenic removal in 2010. LayneRT, a newer version of hybrid adsorbent manufactured by Dow
Chemical, was used to replace the original ArsenXnp during the media change-out at two demonstration
sites. Table 2-4 summarizes the major characteristics of these nine media.
-------
Table 2-2. Summary of AM Demonstration Sites
No.
Site
ID
Design
Flow
rate
(gpm)
Average
Flow
rate
(gpm)
Daily Op
Time
(hr/day)
Average
Daily
Demand
(gpd)
Annual
Production
(kgal)
Utilization
Rate(a)
(%)
Pre-existing
Treatment
Non-Transient Non-Community Water Systems
1
2
3
4a
4b
4c
5
BL
SU
ws
KF
KF
KF
BR
10
12
20
30
60
60
40
On demand
9.3
16.4
1.1
1.0
On demand
On demand
On demand
40
4.2
450
730
984
1,341
17,562
4,580
10,080
83
181
349
489
6,022
1,672
3,679
2
3
3
3
19
5
17
Softener, C12
None
Softener
Gas C12
C12
Community Water Systems
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
GF
WA
PF
DM
W
BW
RR
LI
TN
LD
WM
RF
TE
AL
NP
GE
SV
AN
RN
TA
BC
10
14
15
22
37
40
45
50
63
75
100
120
150
150
160
200
300
320
350
450
640
13
10.4
9.6
6.1
36
41
31
23
60.1
71.5
91
82
79.3
129
114
32.0(c)
207
260
275
503
564
5.4
3.7
3.6
7.6
24(b)
9.5
12or24(b)
18.5
4.4
12
5.9
9.7
19.6
6.7
12.3
2.6
6.2
7.0
3.8
3.9
4.5
4,212
2,618
2,074
1,565
51,840
23,370
NA
25,783
15,276
46,866
32,214
48,977
93,257
51,393
84,132
NA
77,004
109,200
62,700
117,702
152,280
1,509
955
706
571
18,750
8,530
8,508
9,318
5,755
18,790
11,758
21,243
34,039
18,928
30,709
14,868
28,106
40,395
22,885
42,961
51,334
29
13
9
5
96w
41
36
35
17
48
22
34
43
24
37
14
18
24
12
18
15
Aeration for radon
None
Birm18
C12
C12
C12, AA, caustic
C12
Air, C12, poly-PO4
C12
C12
C12
C12
C12
Gas C12, poly-PO4
C12
C12,F
Gas C12, poly-PO4
C12
C12
C12
C12
(a) Ratio of a system's average annual production to its maximum capacity at design flowrate.
(b) Wells at W and RR operated for 12 or 24 hr daily for study purposes.
(c) On demand.
AA = activated alumina; Air = aeration; NA = not available
2.3
AM System Design and Configuration
Because of varying site conditions and source water quality, the design and basic components of the AM
systems varied among the demonstration sites. Table 2-5 summarizes the design and basic components of
the 28 AM systems. The system flowrate, media vessel design, media type and quantity, and any pre-
and/or post-treatment requirement affected the system performance and cost. In addition, the system
instrumentation and controls also affected the system cost. These parameters and cost factors are
discussed as follows.
-------
Table 2-3. Summary of AM Site Source Water Quality
No.
Site
ID
Total As
(HS/L)
As (III)
(HS/L)
Total Fe
(HS/L)
Total Mn
(HS/L)
Total P
(HS/L)
Silica(a)
(mg/L)
TOC
(mg/L)
pH
(S.U.)
Non-Transient Non-Community Water Systems
1
2
o
J
4
5
BL
SU
ws
KF
BR
15.4
31.7
24.7
29.8
57.6
11.3
12.1
5.8
0.3
37.5
2,290
37
27
<25
32
85.7
5.4
17.5
0.4
5.1
<10
<10
<10
<10
<10
15.3
14.1
15.8
30.0
41.8
2.0
1.0
1.0
<0.7
0.9
7.4
8.4
7.1
8.0
8.2
Community Water Systems
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
GF
WA
PF
DM
W
BW
RR
U(b>
TO
LD
WM
RF
TE
AL
NP^
GE
SV
AN
RNW
TA
BC
29.7
39.1
25.2
42.2
39.4
46.4
59.7
41.7
34.9
22.2
36.0
37.7
12.7
40.2
32.2
19.6
20.1
23.5
67.2
16.9
15.3
0.5
38.7
3.2
1.8
0.6
0.5
2.2
0.4
0.5
0.4
1.3
16.8
2.5
31.5
0.7
14.3
19.1
21.7
0.3
0.3
13.1
<25
<25
97
<25
<25
<25
<25
<25
<25
<25
<25
297
<25
63
<25
554
269
80
<25
31
177
o o
5.5
21.9
56.8
9.0
1.0
2.3
0.3
0.2
0.7
0.6
0.6
106.0
4.0
55.1
0.8
8.0
2.9
9.6
0.1
1.3
16.2
71
33
180
<10
11
<10
10
<10
<10
6
<10
81.5
<10
40.7
<10
49.8
17.3
<10
115
<10
<10
25.4
10.5
15.1
12.6
19.0
19.7
25.6
43.4
26.2
16.4
46.8
15.3
27.7
15.3
14.1
23.3
14.6
38.0
72.6
32.8
9.0
<0.7
0.7
<1.0
0.7
0.5
0.7
NA
0.7
0.7
<1
1.3
<1.0
0.7
0.7
<1.0
1.9
0.5
1.6
<1.0
0.7
0.5
7.1
8.5
7.9
7.7
7.7
7.3
6.9
6.9
8.0
7.2
7.8
7.7
7.6
7.8
9.0
7.2
7.8
7.8
7.1
9.6
7.9
(a) as SiO2.
(b) Source water also contained
(c) Source water also contained
elevated uranium.
elevated antimony.
2.3.1 System Flowrate. As shown in Table 2-5, system design flowrates varied from 10 to 60 gpm
for NTNCWS and from 10 to 640 gpm for CWS. The design flowrate of an AM system was determined
by the well capacity or peak flowrate. It was used to size the treatment system, thus affecting the system
capital investment cost (Section 2.4). Average system flowrates as measured during the performance
evaluation studies often were lower than the respective design flowrates. The average flowrate of an AM
system affected media performance and O&M cost, as discussed in Section 2.5.
2.3.2 Tank Design. Most of the AM systems evaluated used two or more media tanks arranged
either in series or in parallel. Since a lead/lag system requires extra media and media tanks than a parallel
system, it often costs more than the parallel system treating the same flow. Smaller systems tend to use a
lead/lag configuration. For example, all seven NTNCWS and eight out of 10 CWS with flowrates below
100 gpm were configured in series; whereas 10 out of 11 CWS equal to or greater than 100 gpm were
configured in parallel. Systems in lead/lag configuration often had one or two treatment trains, each with
a pair of tanks. Exceptions were the ATS systems demonstrated at SU, WA, and DM where one
treatment train consisted of three adsorption vessels in series. Systems in parallel configuration had at
least two treatment trains with one tank in each train. The RN and TA sites each had three vessels in
-------
Table 2-4. Properties of AM Used for EPA Demonstration Projects
Parameter
Matrix/ Active Ingredient
Physical Form
Color
Bulk Density (g/cm3 [lb/ft3])
Moisture Content (%)
BET Area (m2/g)
Particle Size Distribution/
Effective Size
Manufacturer
No. of EPA Demo Sites
Parameter
Matrix/ Active Ingredient
Physical Form
Color
Bulk Density (g/cm3 [lb/ft3])
Moisture Content (%)
BET Area (m2/g)
Particle Size Distribution/
Effective Size
Manufacturer
No. of EPA Demo Sites
Parameter
Matrix/ Active Ingredient
Physical Form
Color
Bulk Density (g/cm3 [lb/ft3])
Moisture Content (%)
BET Area (m2/g)
Particle Size Distribution/
Effective Size
Manufacturer
No. of EPA Demo Sites
A/I Complex 2000
-------
parallel and the BC site had four vessels in parallel. Figures 2-1A through 2-1F show photographs of
selected AM systems with different tank designs and configurations.
Lead/lag and parallel systems can be interchangeable with minor modifications. For example, the BW
and RR systems were originally designed for parallel operation, but were re-configured to lead/lag to treat
about half of the flow or less. The GE system was originally designed as a lead/lag system, but changed
to parallel to treat twice the flow. In theory, when a parallel system is changed to lead/lag, the flow-
normalized cost would double due to a 50% reduction in flowrate.
Tank size and material also affected the system cost. An adsorption tank was sized to hold an appropriate
amount of media required for treatment. Tank sizes varied from 10-in x 54-in (smallest) to 72-in x 72-in
(largest) with a diameter of 10, 12, 18, 20, 24, 36, 42, 48, 54, 63, 66, or 72 in and a height of 48, 52, 54,
60, 65, 72, 80, or 86 in. Adsorption tanks were constructed of fiberglass reinforced plastic (FRP),
polyglass, carbon steel (CS), or stainless steel (SS). The steel tanks were American Society of
Mechanical Engineers (ASME)-coded for a pressure rating of at least 100 psi. The FRP tanks were rated
for 100 to 150 psi. 17 out of 26 sites used FRP tanks and five used CS tanks. The three ATS sites used
small polyglass tanks. Only one site used 72-in x72-in SS tanks, the largest tanks used for the
demonstration program. Both FRP and CS tanks could be used for treatment, but the cost of smaller FRP
tanks often was lower than that of smaller CS tanks. The cost of larger FRP tanks converged with that of
larger CS tanks.
Tank openings and internal arrangements such as upper and bottom distributors and laterals varied among
different types of tanks. For example, smaller tanks often have only one opening on the top with a riser
tube. Larger tanks had top and bottom openings; some even had side openings for viewing and/or media
loading. The internal distributors and laterals were constructed mostly of polyvinyl chloride (PVC) or SS.
2.3.3 Media Type and Volume. The media volume was determined by the design flowrate and
empty bed contact time (EBCT) required. Table 2-6 presents design and average EBCTs for the 28
systems sorted by the media type and tank configuration.
Of the nine media, Isolux™ had the shortest design EBCT of 0.6 min because it is a powder material with
much finer particle sizes (<50 urn) and, therefore, much faster adsorption kinetics than those of granular
media. Isolux™ was filled into cartridges, each with an annular space sandwiched between two thin
layers of tubular membrane made of porous polyethylene (PE) material. The cartridges were then loaded
into adsorption modules and operated in cross-flow, unlike the downflow used by granular media. A/I
Complex 2000 had a short design EBCT, i.e., 0.9 to 1.6 min per tank. But the EBCT for the entire system
was tripled due to the use of three vessels in series. G2® had the longest EBCT of 15.9 min (per tank).
The G2® system was originally designed for a different site to treat 75 gpm of flow using two tanks in
parallel at an EBCT of 17 min. Because the site withdrew from the demonstration program and was
replaced by the BW site to treat a smaller flowrate of 40 gpm, the two G2® tanks were reconfigured to
lead/lag. For E33 media, the design EBCT ranged from 3.3 to 5.7 min for the parallel systems, slightly
longer than that for the lead/lag systems, i.e., 3.1 to 4.1 min. For ArsenXnp, the design EBCT was 4.0 min
for a parallel system and 1.1 to 2.8 min for lead/lag systems.
11
-------
Table 2-5. Summary of AM System Design and Components
No.
Site
ID
Floy
(21
D
vrate
3m)
A
Tank Design
Configu-
ration
No. of
Trains
Tanks
per
Train
Total
Tanks
Tank
Size
(in)
Tank
Materials
Adsorptive Media
Media Type
Volume
per
Vessel
(ft3)
Total
Volume
(ft3)
EBCT(a)
(D/A)
(min)
Pre-
treatment
Post-
treatment
Non-Transient Non-Community Water Systems
1
2
3
4a
4b
4c
5
BL
SU
ws
KF
KF
KF
BR
10
12
20
30
60
60
40
Vary
9.3
16.4
Vary
Vary
Vary
40
Series
Series
Series
Series
Series
Series
Series
1
1
1
2
1
1
1
2
3
2
2
2
2
2
2
3
2
4
2
2
2
18x65
10 x 54
24 x72
18 x65
36x72
36x72
42x72
FRP
Polyglass
FRP
FRP
FRP
FRP
CS
ARM 200
A/I Complex
2000
Adsorbsia™
GTO™
ArsenXnp
ARM 200
Adsorbsia™
GTO™
E33
4.5
1.5
7.5
5
20
20
22
9
4.5
15
20
40
40
44
3.4 (D)
Varying (A)
1.0 (D)
1.2 (A)
2.8 (D)
3.4 (A)
2.5 (D)
Vary (A)
2.5 (D)
Vary (A)
2.5 (D)
Vary (A)
4.1(D)
4.1 (A)
NaOCl,
softening
Oxidation
Columns
None
C12
C12
C12
NaOCl,
PH (C02)
None
None
None
None
None
None
None
Community Water Systems
6
1
8
9
10
11
12
13
14
GF
WA
PF
DM
VV
BW
RR
LI
TN
10
14
15
22
37
40W
45ibj
50
63
13
10.4
9.6
6.1
36
41
31
23
60.1
Series
Series
Series
Series
Series
Series
Series
Parallel
Parallel
1
2
1
2
1
1
1
1
2
2
3
2
3
2
2
2
1
1
2
6
2
6
2
2
2
2w
2
18x65
10 x 54
12x52
10 x 54
36 x72
72 x72
36x72
42x60
36x72
FRP
Polyglass
FRP
Polyglass
FRP
SS
FRP
FRP
FRP
E33
A/I Complex
2000
ArsenXnp
A/I Complex
2000
AAFS50
ARM 200
G281
E33
ArsenXnp
E33
5
1.5
2.3
1.5
16.7,
22
85
22
27
19
10
9
4.6
9
33.4,
44
170
44
54
38
3.7 (D)
2.9 (A)
1.6(D)
2.2 (A)
1.1 (D)
1.8 (A)
1.0 (D)
3.7 (A)
3.5 (A)
4.6 (A)
16 (D)
16 (A)
3.7 (D)
5.3 (A)
4.0 (D)
8.8 (A)
4.5 (D)
4.7 (A)
None
Oxidation
Columns
None
NaOCl
NaOCl,
pH (acid)
NaOCl,
pH (acid)
NaOCl
None
NaOCl,
PH (C02)
Aeration to
remove
Radon
None
Birm®
(old)
None
None
PH
(NaOH)
None
NaOCl,
Poly-P04,
Aeration
None
-------
Table 2-5. Summary of AM System Design and Components (Continued)
No.
15
16
17
18
19
20
21
22
23
24
25
26
Site
ID
LD
WM
RF
TE
AL
NP
GE
SV
AN
RN
TA
BC
Flo\
(SI
D
75
100
120
150
150
160
200
300
320
350
450
640
vrate
3m)
A
71.5
90
82
79.3
129
114
32
207
260
275
503
564
Tank Design
Configu-
ration
Series
Parallel
Parallel
Parallel
Series
Parallel
Parallel
Parallel
Parallel
Parallel
Parallel
Parallel
No. of
Trains
1
2
2
4
1
2
2
2
2
3
3
4
Tanks
per
Train
2
1
1
1
2
1
1
1
1
1
1
1
Total
Tanks
2
2
2
4
2
2
2
2
2
3
3
4
Tank
Size
(in)
42x72
48x72
48x72
20 x48
63 x 86
48 x72
54 x60
63 x86
63 x80
66x72
63 x 86
63 x 80
Tank
Materials
FRP
CS
FRP
CS
FRP
FRP
CS
FRP
FRP
CS
FRP
FRP
Adsorptive Media
Media Type
ArsenXnp
E33
E33
Isolux™
E33
E33
E33
E33
E33
GFFf3
E33
E33
Volume
per
Vessel
(ft3)
28
38
30
2.9
53.6,
70.3
35.6
49
80
76
80
71-73
80
Total
Volume
(ft3)
56
76
60
11.6
124
71.2
98
160
152
240
215
320
EBCT(a)
(D/A)
(min)
2.8 (D)
2.9 (A)
5.7 (D)
6.3 (A)
3.7 (D)
5.5 (A)
0.6 (D)
1.1 (A)
3.1(D)
3.6 (A)
3.3 (D)
4.7 (A)
3.7 (D)
22.9 (A)
4.0 (D)
5.8 (A)
3.6 (D)
4.4 (A)
5.2 (D)
6.5 (A)
3.6 (D)
3.2 (A)
3.7 (D)
4.2 (A)
Pre-
treatment
None
NaOCl,
pH (acid)
NaOCl,
PH (C02)
NaOCl
Gas C12
NaOCl,
pH (CO2)
NaOCl
NaOCl
NaOCl
NaOCl
PH (C02)
NaOCl
Post-
treatment
NaOCl
None
None
None
None
None
None
Poly-PO4
None
NaOCl
C12, HOC1
(MOX)
NaOCl
(a) EBCT for one vessel only.
(b) System flowrate reduced to 50% after being reconfigured to lead/lag.
(c) One vessel in service and one in standby.
A = average; CS = carbon steel; D = design; EBCT = empty bed contact time; FRP = fiberglass reinforced plastic; SS = stainless steel
-------
Figure 2-1A. 20-gpm Adsorbsia™ GTO™ Media System by Siemens
(Two FRP Vessels in Series)
Figure 2-1B. 14-gpm As/I Complex 2000 Media System by ATS
(Two Trains of Three Polyglass Vessels in Series)
14
-------
Figure 2-1C. 40-gpm G2® Media Arsenic Adsorption System by ADI
(Two Stainless Steel Vessels in Series)
Figure 2-1D. 150-gpm Isolux™-302M Media Arsenic Adsorption System by MEI
(Nine Replaceable Media Cartridges in Each Carbon Steel Vessel)
15
-------
Figure 2-1E. 160-gpm E33 Media Arsenic Adsorption System by AdEdge
(Two FRP Vessels in Parallel)
Figure 2-1F. 450-gpm E33 Media Arsenic Adsorption System by Severn Trent Services
(Three FRP Vessels in Parallel)
16
-------
Table 2-6. EBCT vs. Media Type and Tank Configuration
Media Type
A/I Complex 2000
AAFS50
Adsorbsia™ GTO™
ARM 200
ArsenXnp
E33
G2S
GFH®
Isolux™
Design EBCT
Lead/Lag00
0.9-1.6 (3)
4.4 (1)
2.5, 2.8 (2)
2.5, 3.4 (2)
1.1-2.8(3)
3.1-4.1 (4)
15.9(1)
NA
NA
Parallel
NA
NA
NA
NA
4.0(1)
3.3-5.7 (9)
NA
5.1(1)
0.6(1)
Average EBCT
Lead/Lag(a)
1.2-3.7 (3)
3.5(1)
3.4(1)
Varying
2.9(1)
2.9-5.3 (4)
15.5(1)
NA
NA
Parallel
NA
NA
NA
NA
8.8(1)
3.4-6.3 (9)
NA
6.5(1)
1.9(1)
(a) EBCT calculated for one tank.
Numbers in parentheses indicate number of systems.
EBCT = empty bed contact time
2.3.4 Pre- and Post-Treatment. The most common pre-treatments for AM systems are pH
adjustment and pre-oxidation. Any new pre- and/or post-treatment for AM systems will have an impact
on the total capital investment cost and must be taken into consideration when attempting to compare the
costs of different systems.
Because the adsorptive capacity of a media increases with decreasing pH, lowering the water pH can
extend media life and improve media performance. As shown in Table 2-5, eight out of 28 AM systems
were equipped with pH adjustment/control systems, although one site decided not to use it after its
installation. Among these seven systems, five used CO2 gas and two used mineral acid to lower raw
water pH. Figure 2-2 shows a composite of photographs of a CO2pH adjustment/control system, which
consisted of a liquid CO2 supply assembly, an automatic pH control panel, a CO2 membrane assembly,
and a pH probe located downstream of the membrane module. Only one site used NaOH to bring the
effluent pH back to near neutral.
When source water contained soluble As(III), a pre-oxidation step was included to oxidize it to As(V). If
a site already disinfected water with NaCIO or gas C12, the chlorination point was moved to ahead of the
AM system to oxide As(III). Out of the 26 sites, 18 sites used pre-chlorination, two used oxidation
columns, and the remaining six did not use any pre-oxidation. However, not all 18 sites using pre-
chlorination had soluble As(III) in raw water. For example, raw water at the VV site did not have soluble
As(III), but was pre-chlorinated to prevent algae growth in the adsorption tanks. If raw water contained
high concentrations of Fe(II) and/or Mn(II), then a more elaborate pre-treatment, such as iron removal,
would be used to protect AM from being clogged and/or fouled by iron and manganese coatings.
Other pre-existing treatment processes, such as softening, aeration, Birm®, and phosphate addition,
remained on site as long as they did not interfere with the arsenic treatment.
2.3.5 Instrumentation and Controls. System instrumentation and controls varied among
different systems in terms of quality, material, level of complexity/automation, and functionality. Such
variations had an impact on the total capital investment cost and must be taken into consideration when
attempting to compare costs of different systems.
17
-------
Figure 2-2. Carbon Dioxide Gas Flow Control System for pH Adjustment
(Clockwise from Top Left: Liquid CO2 Supply Assembly;
Automatic pH Control Panel; CO 2 Membrane Module; Port for pH Probe)
A fully automatic instrumentation and control system included a programmable logic controller (PLC)
and operator's interface panel (OIP), software, automatic instrumentation (sensors, transmitters,
controllers, alarms, electrical conductors, pneumatic tubing, etc.), and automatically controlled equipment
(valves, pumps, chemical feed pumps, air compressors, etc.). The instrument could monitor pH, flow,
level, pressure, and temperature. Some even had a remote dial-in capability for troubleshooting.
Automatic operations reduced operator's efforts, but increased the cost for instrumentation and control
equipment as well as the skill level required of the operator to maintain more sophisticated equipment.
Some systems only had a controller box on top of a media tank. The AM systems were suitable for semi-
automatic or manual operation because there were not many "moving parts". The three AM systems at
KF were designed for complete manual operations. There was no electrical connection for each of the
three systems; all flow meters and pressure gauges were mechanical and all valves were manual. Pressure
was the driving force to push water through the treatment systems. During system backwash, manual
valves were physically opened and closed to change flow paths and adjust flowrates.
18
-------
2.4 AM System Capital Investment Costs
This section begins with a review of total capital investment costs, and then breaks down the discussion
into three cost categories: equipment, site engineering, and installation.
2.4.1 Total Capital Investment Costs. Capital investment costs for the 28 AM demonstration
systems are categorized into three groups: NTNCWS, small CWS (<100 gpm), and large CWS (>100
gpm), as shown in Table 2-7. The KF site had three separate POE systems, which were counted as three
NTNCWS. One system located in the Resident Hall (Site 4b) supplied water to students living in the
dorms year around, including breaks. Therefore, it was not a typical NTNCWS.
Total capital investment costs ranged from $14,000 for the 22-gpm DM system to $305,000 for the 640-
gpm BC system. Figures 2-3 and 2-4 present the total capital investment costs as a function of design
flowrates for smaller (<100 gpm) and larger systems (>100 gpm), respectively. Because tank
configuration could affect system costs, lead/lag and parallel systems were plotted separately in each
figure. All seven NTNCWS and eight out of 10 small CWS were lead/lag systems, whereas all but one
large CWS were parallel systems. Thus, the effect of tank configuration on costs could not be separated
from that of system flowrates. Even though there were insufficient data to compare costs of systems with
similar sizes but different configurations, lead/lag systems are generally more expensive than their
parallel counterparts.
Among the seven NTNCWS, the BR system had the highest total capital investment cost of $138,642 due
largely to three contributing factors: a CO2 pH control system, two large CS vessels, and a more advanced
system control. Among the smaller CWS (<100 gpm), the VV system had the highest total capital
investment cost at $228,309, partly because it was equipped with a mineral acid pH control system, a
backwash recycle system, and extra monitoring and control devices (see Figures 2-5 and 2-6). The BW
system cost ranked the second highest at $166,050, due mainly to the use of two large (72-in x 72-in) SS
tanks and two pH control systems for raw and treated water (see Figure 2-1C). The three A/I Complex
2000 systems at SU, WA, and DM had the lowest costs because they used small, inexpensive polyglass
tanks (10-in x 54-in) without the backwash capability or automatic controls (see Figure 2-1B).
The data for the larger CWS systems, as shown in Figure 2-4, indicate a stronger correlation between
capital investment costs and system design flowrates. Curve fittings were performed on the data set for
12 parallel systems, yielding an R2 of 0.817 for linear regression. This result might be attributed to the
fact that most of these systems used E33 and similar iron-based media for arsenic removal.
To further compare system capital investment costs, the capital cost of each system was divided by its
design capacity in gpm and gpd and the results are presented in Table 2-7 and plotted against system
design flowrates in Figures 2-7 and 2-8. Normalize costs for NTNCWS ranged from $992 to $3,466/gpm
(or $0.69 to $2.41/gpd) and averaged $2,039/gpm (or $1.42/gpd). Normalized costs for smaller CWS
(<100 gpm) ranged from $636 to $6,171/gpm (or $0.44 to $4.29/gpd) and averaged $2,395(or $1.66/gpd).
These normalized costs scattered widely and did not show a clear trend. Normalized costs for larger
CWS (>100 gpm) ranged from $477 to $l,492/gpm (or $0.33 to 1.04/gpd) and averaged $806 (or
$0.56/gpd). As shown in Figure 2-8, these normalized costs clearly showed a decreasing trend with
system flowrates due to the economy of scale.
Unit costs of the 28 AM systems expressed as 1,000 gal of water treated are also shown in Table 2-7. To
calculate the unit cost, the capital investment cost of an AM system was first converted to an annualized
cost using a capital recovery factor (CRF) of 0.09439 based on a 7% interest rate and a 20-year return
period and then divided by the design or average annual water production rate. The design annual
production is the maximum amount of water that can be produced by a system assuming that it is operated
19
-------
Table 2-7. Total Capital Investment Costs for AM Systems
No.
Site
ID
Media
Type
Design
Flow
Rate
(gpm)
Total
Capital
Cost
($)
Normalized
Capital
Cost
($/gpm)
Normalized
Capital
Cost
($/gpd)
Annualized
Cost
($/yr)
Unit Cost
(/kgal of water)
D(a)
A
Utilization
Rate*'
(%)
Non-Transient Non- Community Water Systems
1
2
3
4a
4b
4c
5
BL
SU
ws
KF
KF
KF
BR
ARM 200
A/I Complex
Adsorbsia'™
ArsenXnp
ARM 200
Adsorbsia'™
E33
Minimum
Maximum
Average
10 (S)
12 (S)
20 (S)
30 (S)
60 (S)
60 (S)
40 (S)
10
60
$27,255
$16,930
$51,895
$55,847
$59,516
$73,258
$138,642
$16,930
$138,642
$2,726
$1,411
$2,595
$1,862
$992
$1,221
$3,466
$992
$3,466
$2,039
$1.89
$0.98
$1.80
$1.29
$0.69
$0.85
$2.41
$0.69
$2.41
$1.42
$2,573
$1,598
$4,898
$5,271
$5,618
$6,915
$13,086
$1,598
$13,086
$0.49
$0.25
$0.47
$0.33
$0.18
$0.22
$0.62
$0.18
$0.62
$0.37
$31.36
$8.90
$14.03
$10.77
$0.93
$4.14
$3.56
$0.93
$31.36
$10.53
2
3
3
3
19
5
17
2
19
8
Community Water Systems (<100 gpm)
6
7
8
9
10
11
12
13
14
15
GF
WA
PF
DM
W
BW
RR
LI
TN
LD
E33
A/I Complex
ArsenXnp
A/I Complex
AAFS50
G2S
E33
ArsenXnp
E33
ArsenXnp
Minimum
Maximum
Average
10 (S)
14 (S)
15 (S)
22 (S)
37 (S)
40 (S)
45 (S)
50 (P)
63 (P)
75 (S)
10
75
$34,201
$16,475
$17,255
$14,000
$228,309
$166,050
$88,307
$114,070
$115,306
$87,892
$14,000
$228,309
$3,420
$1,177
$1,150
$636
$6,171
$4,151
$1,962
$2,281
$1,830
$1,172
$636
$6,171
$2,395
$2.38
$0.82
$0.80
$0.44
$4.29
$2.88
$1.36
$1.58
$1.27
$0.81
$0.44
$4.29
$1.66
$3,228
$1,555
$1,629
$1,321
$21,550
$15,673
$8,335
$10,767
$10,884
$8,296
$1,321
$21,550
$0.61
$0.21
$0.21
$0.11
$1.11
$0.75
$0.35
$0.41
$0.33
$0.21
$0.11
$1.11
$0.43
$2.13
$1.63
$2.31
$2.31
$1.15
$1.84
$0.98
$1.16
$1.89
$0.44
$0.44
$3.56
$1.58
29
13
9
5
96W
41
36
35
17
48
5
48
26
Community Water Systems (>100 gpm)
16
17
18
19
20
21
22
23
24
25
26
WM
RF
TE
AL
NP
GE
SV
AN
RN
TA
BC
E33
E33
Isolux™
E33
E33
E33
E33
E33
G¥H®
E33
E33
Minimum
Maximum
Average
100 (P)
120 (P)
150 (P)
150 (S)
160 (P)
200 (P)
300 (P)
320 (P)
350 (P)
450 (P)
640 (P)
100
640
$149,221
$131,692
$76,840
$179,750
$143,113
$139,149
$211,000
$153,000
$232,147
$296,644
$305,000
$76,840
$305,000
$1,492
$1,097
$512
$1,198
$894
$696
$703
$478
$663
$659
$477
$477
$1,492
$806
$1.04
$0.76
$0.36
$0.83
$0.62
$0.48
$0.49
$0.33
$0.46
$0.46
$0.33
$0.33
$1.04
$0.56
$14,085
$12,430
$7,253
$16,967
$13,508
$13,134
$19,916
$14,442
$21,912
$28,000
$28,789
$7,253
$28, 789
$0.27
$0.20
$0.09
$0.22
$0.16
$0.12
$0.13
$0.09
$0.12
$0.12
$0.09
$0.09
$0.27
$0.14
$1.20
$0.59
$0.21
$0.90
$0.44
$0.88
$0.70
$0.37
$0.96
$0.65
$0.56
$0.21
$1.20
$0.68
22
34
43
24
37
14
18
24
12
18
15
12
43
24
(a) System's maximum capacity at design flowrate, operating 24 hr a day, 365 days a year.
(b) Ratio of a system's average annual production rate to its maximum capacity at design flowrate.
(c) W system operated full time for testing purposes. Data not included in statistics.
A = Average; D = Design; P = parallel configuration; S = series configuration
20
-------
$250,000
$200,000 -
ANTNCWS
»CWS - Lead/Lag
• CWS-Parallel
o $150,000
O
3
'a.
ra
- $100,000 •
S
o
P-
$50,000
«n -
A
• •
* *
A
A A
20 40 60 80
Design Flow/rate (gpm)
100
Figure 2-3. Total Capital Investment Costs of Smaller AM Systems (<100 gpm)
$350,000
$300,000
_. $250,000
to
0 $200,000
s
Q.
(3 $150,000 1
3
"~ $100,000
$50,000
1C
/'''''•
y = 372.62x +82123
R2 = 0.817
• ,,-'' (CS - Parallel)
• ,,-'
* ,--'
^^-'' •
,"-''
* CWS - Lead/Lag
• CWS - Parallel
Linear (CWS - Parallel)
)0 200 300 400 500 600 700
Design Flowrate (gpm)
Figure 2-4. Total Capital Investment Costs of Larger AM Systems (>100 gpm)
21
-------
Figure 2-5. AM Treatment System Components at VV by Kinetico
(Clockwise from Top: POE Well No. 2 and Bypass Piping; Acid Addition Setup;
In-Line pH Transmitter; Adsorption Tanks and Lower Distributor; and Main Control Panel)
Figure 2-6. Backwash Recycling System at VV
(Clockwise from Left: 1,800-gal Holding Tank; Recycle Pump and Bag Filter;
and Backwash Flowrate Indicator and Pump Box)
22
-------
at the design flowrate, 24 hours a day, 365 days a year. In reality, most systems, particularly small ones,
do not operate at the design flowrate or 24 hours a day, 365 days a year. Therefore, the unit cost based on
the average production rate is always higher than that based on the maximum possible production
capacity.
The ratio of the average production to the maximum capacity, or utilization rate, affected the unit capital
cost. In general, the lower the utilization rate, the higher the unit cost. Figure 2-9 presents average unit
costs verses utilization rates for three groups: NTNCWS, small CWS (<100 gpm), and large CWS
(>100 gpm).
$5.00 -
$4.00 -
T3
Q.
O)
$3.00
$2.00
0.
TO
O
A NTNCWS
* CWS - Lead/Lag
• CWS - Parallel
$1.00
$0.00
20 40 60
Design Flowrate (gpm)
80
100
Figure 2-7. Smaller AM System Capital Investment Costs per gpd of Design Capacity (<100 gpm)
Comparison of the data in the three groups revealed some interesting observations. For example, the
systems in the NTNCWS and small CWS groups had comparable flow ranges. However, because the
systems in the NTNCWS group had a significantly lower utilization rate than those in the small CWS
group, i.e., 8% vs. 26% (on average), their unit costs per 1,000 gal of water treated were significantly
higher than those for the systems in the small CWS group, i.e., $10.53 vs. $1.58 (on average). On the
other hand, the systems in the small and large CWS groups had comparable utilization rates, i.e., 26% vs.
24% (on average), and the system unit costs of the small CWS group were more than twice the costs for
the large CWS group, i.e., $1.58 vs. $0.68 (on average). Therefore, the systems in the NTNCWS group
had the highest unit costs due to small sizes and low utilization rates. An NTNCWS could consider using
a smaller system with a larger storage capacity to achieve a higher utilization rate, thus a lower unit cost.
23
-------
$1.20
$1.00 '
•o
§> $0.80
0)
Q.
8 $0.60
O
$0.40
$0.20
$0.00
* CWS - Lead/Lag
• CWS - Parallel
Log. (CWS-Parallel)
100 200 300 400 500 600 700
Design Flowrate (gpm)
Figure 2-8. Larger AM System Capital Investment Costs per gpd of Design Capacity (>100 gpm)
$100.00
"TO
0)
| $10.00
to
o
o
3 $1.00 -
'o.
TO
O
ANTNCWS
*CWS<100 gpm
• CWS>100 gpm
A A
****.*
•• A "• *
.•
. •
•
IpU. 1U
0% 10% 20% 30% 40% 50% 60%
Utilization Rate
Figure 2-9. AM System Unit Costs per 1,000 gal of Water Treated as
a Function of Utilization Rates
24
-------
2.4.2 Equipment Cost. Treatment equipment including filtration vessels, piping and valves, and
instrument and controls was mostly skid-mounted on a steel frame. The equipment cost for an AM
system included the cost for the skid-mounted system, AM and under-bedding media, miscellaneous
materials and supplies, freight, user's manual, and vendor's labor. It also included the cost for pH
adjustment and/or pre-oxidation equipment. In one or two cases, the cost of backwash recycle equipment,
such as backwash storage tank(s) and recycle pump, was also included in the equipment cost if it was part
of the original proposal selected for the demonstration study.
Equipment costs for the AM systems ranged from $8,640 for the 12-gpm SU system to $218,000 for the
640-gpm BC system, as shown in Table 2-8. On average, the equipment costs accounted for 61%, 67%,
and 72% of the total capital investment costs for NTNCWS, smaller CWS (<100 gpm), and larger CWS
(>100 gpm), respectively. Equipment cost data were plotted as a function of flowrates in Figure 2-10 for
smaller systems (<100 gpm) and in Figure 2-11 for larger systems (>100 gpm). Because the equipment
costs made up the highest percentage of the total capital investment costs, equipment cost curves were
similar, as expected, to the total capital investment cost curves shown in Figures 2-3 and 2-4. Factors
contributing to the highest or the lowest equipment cost for the BR, VV, BW, and three A/I Complex
2000 systems were discussed in Section 2.4.1. Curve fittings were performed on the data set for 12
parallel systems (>100 gpm), yielding an R2 of 0.8002 for linear regression.
2.4.3 Site Engineering Cost. The site engineering cost for an AM system included the cost for the
development of a system layout within the treatment building, design of piping connections to the inlet
and distribution tie-in points in the building, and design of electrical connections. The site engineering
cost also included the cost for the submission of engineering plans to relevant state agencies for permit
review and approval.
Engineering costs for the AM treatment systems ranged from $1,800 for the 14-gpm WA system to
$50,659 for the 37-gpm VV system. These costs represent, on average, 20%, 14%, and 12% of the total
capital investment costs for NTNCWS, smaller CWS (<100 gpm), and larger CWS (>100 gpm),
respectively (see Table 2-8). As expected, the percentage decreased as the size of the system increased.
2.4.4 Installation Cost. The installation cost for an AM system included equipment and labor to
unload and install the system, perform piping tie-ins and electrical connections, load and backwash AM,
perform system shakedown and startup, and conduct operator's training. Piping tie-ins were completed
using ductile iron or polyvinyl chloride (PVC) pipe, valves, and fittings. Figure 2-12 is a photograph
showing media loading at the VV site. Installation costs for the treatment systems ranged from $2,610 for
the 22-gpm DM system to $61,209 for the 450-gpm TA system. These installation costs represented
20%, 19%, and 16% of the total capital investment costs for NTNCWS, smaller CWS (<100 gpm), and
larger CWS (>100 gpm), respectively (see Table 2-8). Again, the percentage decreased as the size of the
system increased, as expected.
2.5 AM System O&M Costs
O&M costs evaluated included the cost for media replacement and disposal, chemical supply, electricity
consumption, and labor to operate the treatment systems. Of the 28 AM systems, 15 systems had spent
media replaced during the study period and therefore more complete O&M costs were available. Table 2-
9 summarizes the O&M costs with cost breakdowns for the 15 systems with media replacement. Two of
the systems, i.e., WA and VV, experienced multiple change-outs with different media types. For the 13
systems without media replacement, estimated replacement costs were provided in individual final
performance evaluation reports. Because costs were not actually incurred, the estimates were not used in
the cost analysis herein. Each cost component is discussed below.
25
-------
Table 2-8. Summary of Equipment, Site Engineering, and Installation Costs of AM Systems
No.
Site
ID
Media
Type
Design
Flow
Rate
(gpm)
Total
Capital
Cost
($)
Equipment
Cost
%of
Total
Site
Engineering
Cost
%of
Total
Installation
&Startup
Cost
%of
Total
Non-Transient Non-Community Water Systems
1
2
3
4a
4b
4c
5
BL
SU
ws
KF
KF
KF
BR
ARM 200
A/I Complex
Adsorbsia™
ArsenXnp
ARM 200
Adsorbsia™
E33
Minimum
Maximum
Average
10
12
20
30
60
60
40
10
60
$27,255
$16,930
$51,895
$55,847
$59,516
$73,258
$138,642
$16,930
$73,258
$10,435
$8,640
$30,215
$39,108
$41,689
$51,314
$94,662
$8,640
$51,314
Community Water Systems
6
7
8
9
10
11
12
13
14
15
GF
WA
PF
DM
W
BW
RR
LI
TN
LD
E33
A/I Complex
ArsenXnp
A/I Complex
AAFS50
G2S
E33
ArsenXnp
E33
ArsenXnp
Minimum
Maximum
Average
10
14
15
22
37
40
45
50
63
75
10
75
$34,201
$16,475
$17,255
$14,000
$228,309
$166,050
$88,307
$114,070
$115,306
$87,892
$14,000
$228,309
$22,431
$10,790
$11,345
$8,990
$122,544
$105,350
$63,785
$82,470
$86,018
$60,678
$8,990
$122,544
Community Water Systems
16
17
18
19
20
21
22
23
24
25
26
WM
RF
TE
AL
NP
GE
SV
AN
RN
TA
BC
E33
E33
Isolux™
E33
E33
E33
E33
E33
GFH«
E33
E33
Minimum
Maximum
Average
100
120
150
150
160
200
300
320
350
450
640
100
640
$149,221
$131,692
$76,840
$179,750
$143,113
$139,149
$211,000
$153,000
$232,147
$296,644
$305,000
$76,840
$305,000
$103,897
$105,805
$58,500
$124,103
$116,645
$101,290
$129,500
$112,000
$157,647
$202,685
$218,000
$58,500
$218,000
38
51
58
70
70
70
68
38
70
61
$11,000
$3,400
$10,110
$9,941
$10,587
$13,032
$24,300
$3,400
$13,032
40
20
19
18
18
18
18
18
40
20
$5,820
$4,890
$11,570
$6,798
$7,240
$8,912
$19,680
$4,890
$11,570
21
29
22
12
12
12
14
12
34
20
f<100gpm)
66
65
66
64
54
63
72
72
75
69
54
75
67
$4,860
$1,800
.(a)
$2,400
$50,659
$17,200
$11,372
$12,800
$12,897
$14,214
$1,800
$50,659
14
11
_(a)
17
22
10
13
11
11
16
10
22
14
$6,910
$3,885
$5,910
$2,610
$55,106
$43,500
$13,150
$18,800
$16,391
$13,000
$2,610
$55,106
20
24
34
19
24
26
15
16
14
15
14
26
19
f>100gpm)
70
80
76
69
82
73
61
73
68
68
71
61
82
72
$25,310
$4,672
$8,500
$14,000
$11,638
$19,545
$36,700
$23,000
$16,000
$32,750
$35,500
$4,672
$35,500
17
4
11
8
8
14
17
15
7
11
12
4
17
12
$20,014
$21,215
$9,840
$41,647
$14,830
$18,314
$44,800
$18,000
$58,500
$61,209
$51,500
$9,840
$61,209
13
16
13
23
10
13
21
12
25
21
17
13
25
16
(a) Included in equipment cost.
26
-------
$140,000
$120,000
_, $100,000
,3 $80,000
100 gpm)
27
-------
Figure 2-12. E33 Media Loading
2.5.1 Media Replacement Cost. As shown in Table 2-9, media replacement costs represented the
majority of O&M costs, accounting for 39% to 97% of O&M costs (averaging 79%). The media
replacement cost included the cost for replacement media, labor (for replacement services), spent media
analysis (i.e., Toxicity Characteristic Leaching Procedure [TCLP]), spent media disposal, and freight. All
spent media passed the TCLP test and were disposed off as non-hazardous wastes (the exact disposal
facilities were not tracked by the study). Table 2-10 presents breakdowns of actual media replacement
costs for the 15 systems, including multiple replacements for the WA and VV systems. To help
understand the costs, the table also summarizes data that affected media replacement, including
replacement media type, media life (at the time of replacement), volume throughput (in gallons and bed
volumes [BV]), and quantity replaced.
The cost analysis also included unit media replacement costs (in $/ft3 or $/1,000 gal of water treated)
obtained by dividing lump-sum media replacement costs by either respective media quantities or volume
throughputs (gallons of water treated to reach 10-ng/L arsenic in system effluent). The results of these
calculations are also shown in Table 2-10 for comparisons among different media across different sites.
Table 2-11 summarizes media replacement costs of different media types occurred at one or multiple
demonstration sites, i.e., five for E33, three for A/I complex 2000, two each for ARM 200, LayneRT, and
GFH®, and one each for AAFS50, G2®, and Isolux™. Adsorbsia™ GTO™ was not replaced at either of
the two sites during the study period; therefore, the estimated cost was presented instead.
28
-------
Table 2-9. O&M Costs for AM Systems with Media Replacement
No.
Site
ID
Design
Flow rate
(gpm)
Total
O&M
Costs
(S/kgal)
Media Replacement
Replacement
Media
Type
Cost
(S/kgal)
%of
Total
O&M
Electricity
Cost
(S/kgal)
Chemicals
Type
Cost
(S/kgal)
Labor
Average
Weekly
Hours
Labor
Rate
(S/hr)
Cost
(S/kgal)
Non-Transient Non-Community Water Systems
1
4b
SU
KF
12
60
$12.06
$5.82
A/I Complex
ARM 200
$8.96
$5.37
74
92
$0.000
$0.000
No
No
$0.00
$0.00
0.33
2.5
$30.0
$21.0
$3.10
$0.45
Community Water Systems
6
1
8
9
10
11
12
15
18
19
22
23
24
GF
WA
PF
DM
VV
BW
RR
LD
TE
AL
SV
AN
RN
Minimum
Maximum
Average
10
14
15
22
37
40
45
75
150
150
300
320
350
$2.34
$22.88
$10.44
$5.52
$7.67
$10.86
$2.74
$1.48
$1.79
$5.11
$0.86
$0.98
$1.16
$0.61
$0.61
$0.75
$5.69
$0.61
$22.88
$4.61
E33
A/I Complex
GFH
CFH
LayneRT
A/I Complex
AAFS50
AAFS50
ARM 200
G2®
E33
ArsenXnp
Isolux™
E33
E33
E33
GFH®
$2.01
$22.05
$9.44
$4.76
$5.31
$9.99
$2.56
$0.58
$1.61
$4.30
$0.64
$0.58
$1.02
$0.36
$0.30
$0.66
$5.51
$0.30
$22.05
$4.15
86
96
90
86
69
92
93
39
90
84
74
59
88
59
49
89
97
39
97
79
$0.000
$0.000
$0.000
$0.000
$0.000
$0.000
$0.157
$0.157
$0.157
$0.001
$0.008
$0.000
$0.001
$0.000
$0.050
$0.001
$0.001
$0.00
$0.16
$0.03
No
No
No
No
No
No
No
Acid
No
Acid/Base
No
No
No
No
Replacement parts
Replacement parts
No
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
$0.61
$0.00
$0.11/0.36
$0.00
$0.00
$0.00
$0.00
$0.03
$0.03
$0.00
$0.00
$0.61
$0.07
0.5
0.75
0.75
0.75
1.6
0.5
0.4
2.4
0.4
2.33
1.67
7.0
2.5
4.67
1.75
1.75
2.5
0.4
7.0
1.9
$21.0
$20.0
$20.0
$20.0
$20.0
$20.0
$21.0
$21.0
$21.0
$20.0
$21.0
$21.0
$37.5
$19.5
$21.8
$18.2
$35.0
$18.2
$37.5
$22.4
$0.33
$0.83
$1.00
$0.76
$2.36
$0.87
$0.03
$0.14
$0.03
$0.34
$0.22
$0.40
$0.14
$0.25
$0.23
$0.05
$0.18
$0.03
$2.36
$0.36
-------
Table 2-10. Breakdowns of Media Replacement Costs
No.
7
9
2
10
10
10
4b
6
19
23
12
22
11
24
7
7
18
8
15
Site
ID
WAI
DM
SU
Wl
W2
W3
KF
GF
AL
AN
RR
SV
BW
RN
WA2
WAS
TE
PF
LD
Design
Flow
Rate
(gpm)
14 (S)
22 (S)
12 (S)
37 (S)
37 (S)
37 (S)
60 (S)
10 (S)
150 (S)
320 (P)
45 (S)
300 (P)
40 (S)
350 (P)
14 (S)
14 (S)
150 (P)
15 (S)
75 (S)
Media Type, Run Length, and Quantity Replaced
Replace-
ment Media
Type
A/P & A/I(d)
A/I Complex
A/I Complex
AAFS50
AAFS50
ARM 200
ARM 200
E33-G
E33-P
E33-P
E33-G
E33-G
G2KJ
GFH®
Filox™/GFH®
Filox™/CFH(e)
Isolux™
LayneRT
LayneRT
Media
Life
(mon)
6
8
18
2
5
5.5
13.5
17
24
18
25
-42
13
7
12
12
~4
10.5
20
Volume
of Water
Treated'3'
(gal)
342,000
391,400
257,832
3,411,000
7,580,000
8,464,000
2,085,424
2,085,000
35,375,613
46,553,000
17,164,000
93,820,742
3,896,000
12,925,440
391,000
516,000
6,941,440
516,120
27,978,780
Volume
of Water
Treated'1''
(BV)
5,100
5,814
7,660
10,364
23,031
25,717
13,940
27,874
38,140
50,191
52,151
78,393
3,064
7,200
11,600
15,300
80,000
15,000
66,794
Media
Volume
(ft3)
3/9
6
3
44
22
22
20
5
48
124
22
160
170
240
1.5/4.5
1.5/4.5
11.6
2.3
28
Media Replacement Costs
Media
Unit
Cost
($/ft3)
$517
$517
$450
$99
$99
$500
$385
$300
$165
$202
$265
$156
$40
$240
$595
$320
$559
$852(f)
$480
Total
Media
Cost
($)
$6,204
$3,102
$1,350
$4,350
$2,175
$11,000
$7,700
$1,500
$7,920
$25,048
$5,830
$24,928
$6,800
$57,600
$2,993
$1,755
$6,484
$1,960
$13,440
Labor
Cost
($)
$520
$260
$0
Other
Costs'0'
($)
$845
$548
$960
$4,375
$2,188
$2,610
$3,500
$1,850
$1,000
$4,130
$4,240
$2,120
$8,272
$12,950
$500
$500
Facility (g)
$360
Facility (g)
$849
$3,760
$1,722
$838
$680
$1,680
$608
$201
$200
$596
$420
$2,693
Total
MR
Cost
($)
$7,569
$3,910
$2,310
$8,725
$4,363
$13,610
$11,200
$4,199
$12,680
$30,900
$10,908
$27,728
$16,752
$71,158
$3,693
$2,455
$7,080
$2,740
$16,133
Unit
MR
Cost
($/ft3)
$631
$652
$770
$198
$198
$619
$560
$840
$264
$249
$496
$173
$99
$296
$616
$409
$610
$1,191
$576
Unit
MR
Cost
($/kgal)
$22.05
$9.99
$8.96
$2.56
$0.58
$1.61
$5.37
$2.01
$0.36
$0.66
$0.64
$0.30
$4.30
$5.51
$9.44
$4.76
$1.02
$5.31
$0.58
(a) System throughput at time of reaching 10-ug/L arsenic in system effluent.
(b) For lead/lag system, B V calculated based on media in both lead and lag vessels.
(c) Other costs including spent media analysis, spent media disposal, and freight.
(d) A/P Complex 2002 oxidizing media and A/I Complex 200 adsorptive media manufactured by ATS.
(e) CFH-12 adsorptive media manufactured by Kemira Water Solutions.
(f) Including cost of media vessel.
(g) Provided by facility.
B V = bed volumes; G = granular; MR = media replacement; P = parallel or pelletized; S = series
-------
Table 2-11. Replacement Costs of Various Types of AM
Media
Type
A/I Complex 2000
AAFS50
Adsorbsia™ GTO™
ARM 200
ArsenXnp/LayneRT
E33
G2KJ
GFH®
Isolux™
No.
of
Systems
3
1
2
2
2
5
1
2
1
Media
Cost
Only
($/ft3)
450-517
99
449(b), 678(b)
385; 500
480; 852(d)
165-300
40
240; 595
559
Media
Replacement
Unit Cost
($/ft3)
631-770
198
774(b- d)
560; 619
576; l,191(d)
173-840
99
296; 616
610
Media
Run
Length
(BV)
5,100-7,700
23,000(a), 10,400
>5,240(c);>21,900(c)
13,900; 25,700
15,000; 66,800
27,900-78,400
3,100ta)
7,200; 11,600
80,000
Normalized
Replacement
Cost
($/kgal of Water)
8.96-22.05
0.58(a); 2.56
<10.66(c); <2.30(c)
1.61; 5.37
0.58; 5.3 l(d)
0.30-2.01
4.30
5.51; 9.44
1.02
(a) With pH adjustment.
(b) Estimates provided by vendor.
(c) Based on data at end of study when arsenic had not reached 10 ug/L breakthrough in system effluent.
(d) Including cost of media vessel.
Figure 2-13 plots media replacement costs against media run lengths for eight different media. As shown
in Table 2-11 and Figure 2-13, media performance and costs varied from site to site, even for the same
media type. Different water quality, such as concentrations of arsenic, phosphate, and silica and water
pH, and different system designs in terms of EBCT and series/parallel configuration, could affect media
performance. For example, ArsenXnp achieved 66,800 BV at the LD site but only 15,000 BV at the PF
site. The PF source water had a higher pH (7.9 vs. 7.2) and contained more phosphorus (180 vs. <10
|jg/L as total P) than the LD source water. The PF system also had a shorter EBCT than the LD system
(1.8 vs. 2.9 min per vessel). There are 13 systems using E33 with five having media replacement. Run
lengths of E33 media ranged from 27,900 to 78,400 BV. The shortest run length of 27,900 BV occurred
at the GF site where source water contained 71 (ig/L (on average) of total phosphorus. In general, ferric
oxide or hydroxide media outperformed the iron-modified, alumina- or silica-based media. The poor
performance of GFH® observed at the RN site was caused by high phosphorus (115 (ig/L as total P) and
very high silica (i.e., 72.6 mg/L as SiO2) in source water.
Figure 2-14 plots media replacement unit costs (including replacement media, labor, and spent media
disposal costs) of 13 E33 systems against system design flowrates. Estimated costs were used in the plot
for the systems without media replacement. The data clearly showed that unit media replacement costs
decreased as system sizes increased, due primarily to the scale of economy.
The media replacement cost per 1,000 gal of water treated is a function of the unit media replacement cost
per ft3 and the media run length, as shown by the following equation:
Replacement Cost ($/1,000 gal) = Media Replacement Unit Cost ($/ft3)/(Run Length [BV] x 7.48/1,000)
31
-------
$100.00 v
A/I Complex
• AAFS50
=• ARM 200
CU3
§
<=
^. $10.00 ". GFH
vv
V, O G2
8 »
*- Clsolux
£ •
E • » LayneRT
g- $1.00
-------
Figure 2-15 presents a series of hypothetic cost curves with each representing one media with a certain
unit media replacement cost. The cost curves clearly show that the longer the run lengths are, the lower
the replacement costs (per 1,000 gal of water treated) would be. These cost curves can be used as a
general guideline to compare different media and help select the most cost-effective media. An example
is given below to show how to use these cost curves step by step.
Assumptions:
• Media A costs $200/ft3 and is replaced at 25,000 BV
• Media B costs $400/ft3 and is replaced at 60,000 BV
Solutions:
• Step 1: Find the curve representing Media A with a unit cost of $200/ft3.
• Step 2: On the x-axis, draw a vertical line across 25,000 BV and intercept the $200/ft3 curve
at Point A, find the y value of Point A, which is approximately $1.1/1,000 gal.
• Step 3: Find the curve representing Media B with a unit cost of $400/ft3.
• Step 4: On the x-axis, draw a vertical line across 60,000 BV and intercept the $4007 ft3 curve
at Point B, find the y value of Point B, which is approximately $0.90/1,000 gal.
In this example, Media B's cost is twice as much as Media A's, but its life is more than twice as long as
Media A's. Assuming all other costs, i.e., labor and media disposal, are equal, Media B has a lower
replacement cost (per 1,000 gal of water treated).
$5.00
$4.50
$4.00
$3.50
_ $3.00
"ro
O)
O
o
°- $2.50
5»
In
0 $2.00
$1.50
$1.00
$0.50
$0.00
\\
\
Point A
Point B
10 20 30 40 50 60 70
Media Life (x1,000 Bed Volumes)
90
100
Figure 2-15. Hypothetic Media Replacement Cost Curves
33
-------
2.5.2 Chemical Cost. Chemicals used during AM system operations included CO2 and H2SO4/NaOH
for pH adjustments and sodium hypochlorite (NaOCl) and gas chlorine for pre-oxidation and disinfection.
Table 2-12 presents chemical costs for the pH control systems used at seven sites (note: the pH control
system installed at the WM site was not used).
Table 2-12. Costs of pH Controls for AM Systems
Site
ID
W
BW
BR
TN
RFlb)
NP
TA
Flow
Rate
(gpm)
37
40
40
63
120
160
450
Media
Type
AAFS50
G2®
E33
E33
E33
E33
E33
Chemical(s)
H2SO4
H2SO4,
NaOH
C02
C02
C02
C02
C02
Raw
Water
pH
7.7
7.3(a)
8.2
8.0
7.7
9.0
9.6
Target
pH
6.8
6.5
7.0
7.0
7.4
7.0
7.2
Usage
(Ib/kgal of
Water)
0.58
0.27,
0.57
0.65
0.39
0.12
0.30
0.36
Cost
($/kgal of
Water)
0.61
0.11,
0.36
0.41
0.30
0.11
0.20
0.29
(a) Lower than historical value of 7.7.
(b) CO2 pH control system installed at RF site used for Phase 1, but not for Phase 2.
H2SO4 was available in a 37%, 50%, or 93% solution in 15- or 55-gal drums. NaOH was available in a
25% solution in 15-gal drums and used only at one site to raise pH after treatment. CO2 was supplied
with 50-lb gas cylinders for smaller systems and 380-lb dewars for larger systems. CO2 supply costs
ranged from $0.11 to $0.41 per 1,000 gal of water treated.
Some facilities had pre-existing chlorination for disinfection, which was switched to pre-chlorination if
these facilities required pre-oxidation for soluble As(III) conversion. Because oxidation of soluble
As(III) did not consume a significant amount of chlorine and the chlorine usage did not show any
noticeable increase, the incremental chemical cost was negligible.
2.5.3 Electricity Cost. The electricity cost was tracked by comparing monthly electrical bills
before and after installation of an AM treatment system. If the site did not have a separate meter for the
arsenic treatment system, then the cost was estimated based on power requirements of the major
equipment such as compressor, pump, etc., average operational hours, and local electricity unit price.
Local electricity unit prices ranged from $0.08 to $0.14/kwh provided by the facilities.
The incremental electrical consumption was negligible for most of the sites because the AM systems have
very few "moving" parts and operate mostly intermittently. Electricity costs per 1,000 gal of water
treated ranged from zero to $0.16 and averaged $0.03, as shown in Table 2-9. The highest electricity cost
incurred at the VV site because the VV system was equipped with a number of energy-consuming
components such as a compressor (to supply air to pneumatic valves), an acid metering pump, a backwash
recycling pump, and a heat lamp (during winter time), and operated around the clock for the
demonstration study.
2.5.4 Labor Cost. Each demonstration site was provided with an Operator Labor Log Sheet to
track labor hours used for routine O&M, EPA demonstration study-related activities, repairs, and
miscellaneous activities. The routine O&M included activities such as filling field logs, performing
system inspection, ordering inventory, and others as recommended by vendors. EPA study-related
34
-------
activities such as performing field measurements, collecting and shipping samples, and communicating
with the Battelle Study Lead, were tracked, but not used for cost analysis.
The routine, non-demonstration related labor activities consumed only 10 to 30 min a day, one or several
days a week at most of the AM sites. Average weekly hours ranged from 20 min to 7 hr, averaging 1.9
hr. As shown in Table 2-9, labor rates ranged from $18.2 to $37.5/hr and averaged $22.4/hr (note: these
labor rates might be lower than those in certain regions of the country, such as California, but were actual
numbers provided by the operators). Labor costs per 1,000 gal of water treated varied significantly from
$0.45 to $3.10 forNTNCWS and from $0.03 to $2.39 for CWS due to varying annual water production
rates among the AM demonstration sites. NTNCWS often had a lower demand and a lower utilization
rate than CWS. Therefore, the labor cost (per 1,000 gal of water treated) of a small NTNCWS was higher
than that of a large CWS.
35
-------
3.0 IRON REMOVAL/COAGULATION/FILTRATION SYSTEMS
Of the 50 demonstration sites, 18 sites used IR or CF as the main treatment process, including two
NTNCWS and 16 CWS. Among the 18 systems, four systems had IR followed by AM to remove iron
and arsenic. At these four sites, the main purpose of the IR was to provide protection to the AM systems
against iron fouling although at one site (SF), the AM system was actually used to polish the IR system
effluent because the IR system had already reduced arsenic concentrations to below the MCL.
Table 3-1 lists IR/CF demonstration locations, technologies, and study durations in order of system
design flowrates. The performance evaluation studies for the IR and CF systems were conducted for a
period of 12 to 15 months, except for two systems for which more extensive studies were performed.
Detailed information on system performance and costs can be found in individual final performance
evaluation reports provided on the EPA Arsenic Demonstration Program Web site.
Table 3-1. Summary of IR/CF Demonstration Locations,
Technologies, and Study Durations
No.
Site
ID
Demonstration
Location
Technology
Vendor
Design
Flow rate
(gpm)
Study
Duration
Length
of Study
(mon)
Non-Transient Non-Community Water Systems
1
2
GS
FC
Goshen, IN
Fountain City, IN
IR (AD26)^AM (E33)
IR (G2®)
AdEdge
US Water
25
60
06/08-06/09
09/08-10/09
12
13
Community Water Systems
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
SC
WL
DV
wv
CM
CL
TF
SA
SF
ST
SD
GV
FE
PW
OK
AR
Sauk Centre, MN
Willard, UT
Delavan, WI
Waynesville, IL
Climax, MN
Conneaut Lake, PA
Three Forks, MT
Sabin, MN
Springfield, OH
Stewart, MN
Sandusky, MI
Greenville, WI
Felton, DE
Pentwater, MI
Okanogan, WA
Arnaudville, LA
IR (Macrolite®)
IR (Birm®/Filox™) +
AM (Adsorbsia™ GTO™)
IR (Macrolite®)
IR (GreensandPlus™)
IR/IA (Macrolite®)
CF (AD GS+)
CF (Macrolite®)
IR (Macrolite®)
IR(AD26)+AM(E33)
IR (AERALATERVAM (E33)
IR (AERALATER®)
IR (Macrolite®)
CF (Macrolite®)
IR/IA (Macrolite®)
CF (Electromedia® I)
IR (Macrolite®)
Kinetico
Filter Tech
Kinetico
Peerless
Kinetico
AdEdge
Kinetico
Kinetico
AdEdge
AdEdge
Siemens
Kinetico
Kinetico
Kinetico
Filtronics
Kinetico
20
30
45
96
140
250
250
250
250
250
340
375
375
400
550
770
07/05-10/06
12/08-10/10
07/05-09/06
07/09-09/10
08/04-08/05
12/09-12/10
11/06-02/08
01/06-04/07
09/05-09/06
02/06-02/07
06/06-06/07
08/07-12/07;
05/09-04/10
09/06-11/07
11/05-12/06
08/08-08/09
06/06-09/10
15
22
14
14
12
12
15
15
12
12
12
4;
11
14
13
12
51
AM = adsorptive media; CF = coagulation/filtration; IA = supplemental iron addition; IR = iron removal
3.1
Overview of IR/CF Demonstration Sites
Table 3-2 summarizes the IR/CF demonstration site information. Most of the facilities evaluated were
classified as very small (serving 25 to 500 of people) and small (serving 501 to 3,300 of people) water
systems. The two NTNCWS systems, both schools, were operated fewer than 2 hr/day, whereas most
CWS were operated less than 10 hr/day. Average daily demand was less than 4,000 gal for NTNCWS
36
-------
and varied from 4,500 to 414,000 gal for CWS. Annual productions were less than 1 MG for NTNCWS
and ranged from 1.6 to 139 MG for CWS. Utilization rates were 3 or 4% for NTNCWS and 9 to 48% for
CWS.
Table 3-2. Summary of IR/CF Demonstration Sites
No.
Site
ID
Design
Flow
rate
(gpm)
Average
Flow
rate
(gpm)
Daily
Op Time
(hr/day)
Average
Daily
Demand
(gpd)
Annual
Production
(kgal)
Utilization
Rate00
(%)
Pre-existing
Treatment
Non-Transient Non-Community Water Systems
1
2
GS
FC
25
60
15.2
47
1.9
1.4
1,733
3,956
517
845
4
3
None
C12, softener
Community Water Systems
o
J
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
SC
WL
DV
WV
CM
CL
TF
SA
SF
ST
SD
GV
FE
PW
OK
AR
20
30
45
96
140
250
250
250
250
250
340
375
375
400
550
770
4.0
9.3
20 (max)
84
132
153
206
231
89
190
163
285
263
350
538
335
4.6
23.4
2.6
11.8/5.8
5.6
11.9/4.3
8.9
3.1
9.5
4.7
NA
3.8
6.5
5.1
13.6
14
4,523
8,354
5,981
29,400
38,560
109,242
107,400
32,858
45,700
52,418
166,000
66,037
107,300
102,800
414,000
277,128
1,650
3,049
2,200
10,731
13,800
20,114
27,200
12,200
16,700
19,133
60,300
24,051
38,200
38,300
139,400
101,152
16
19
9
21
19
15
21
9
13
15
34
12
19
18
48
25
None
None
Softener
C12, poly-P04
Gas C12
Gas C12, poly-PO4
C12
Aeration, gravity
filtration, C12
C12, poly-P04
Gas C12, poly-PO4
C12, poly-P04
Gas C12
C12
C12, poly-P04
None
Aeralator, C12,
softener
(a) Ratio of a system's average annual production to its maximum capacity at the design flowrate.
NA = not available
Table 3-3 presents source water quality of the 18 IR/CF sites using average values measured during the
performance evaluation studies. Arsenic concentrations in source waters varied from 11.4 to 84.0 |Jg/L
(excluding the GV water, which contained only 5.6 |Jg/L of total arsenic). Soluble As(III) was the
predominating arsenic species at all but three sites (i.e., WL, TF, and SA). Iron, existing predominantly
as soluble Fe(II), exceeded its SMCL of 300 |o,g/L at 13 sites, with the highest concentration measured at
2,385 |o,g/L at the SC site. Half of the sites had manganese levels above its SMCL of 50 |og/L. Four of
the five low-iron sites, i.e., CL, TF, FE, and OK, added an iron salt to source waters as a coagulant to
remove arsenic. At these sites, the treatment system was considered a CF process. The fifth site, WL,
used dual Birm®/Filox™ media as a pretreatment to AM. The CM and PW sites contained moderate
levels of iron in raw waters, which were insufficient to remove arsenic to below 10 |Jg/L in treated water.
Therefore, supplemental iron was added to the waters at both sites to improve the arsenic removal rates.
During the studies, high phosphate and silica levels were found to affect system performance and reduce
treatment efficiencies. At four sites (i.e., SC, WL, ST, and AR), total phosphate concentrations were over
100 |og/L. Significantly elevated silica concentrations were measured at the TF and AR sites at 48.5 and
42.5 mg/L, respectively. The presence of high total organic carbon (TOC) and ammonia had some effects
37
-------
Table 3-3. Summary of IR/CF Site Source Water Quality
No.
Site
ID
Total As
(Hg/L)
As (III)
(Hg/L)
%
As(III)
Total Fe
(HS/L)
Total Mn
(HS/L)
Total P
(Hg/L)
Silica00
(mg/L)
TOC
(mg/L)
pH
(S.U.)
NH3(b)
(mg/L)
Non-Transient Non-Community Water Systems
1
2
GS
FC
28.6
29.4
20.2
17.7
71
60
741
1,865
82
51
11
11
20.1
15.2
<1.0
1.8
7.3
7.6
0.1
1.0
Community Water Systems
o
J
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
SC
WL
DV
WV
CM
CL
TF
SA
SF
ST
SD
GV^
FE
PW
OK
AR
27.5
13.2
18.9
33.1
36.5
29.0
84.0
41.8
22.7
44.8
11.4
5.6
34.4
17.7
17.9
32.7
21.9
6.0
16.3
24.1
35.8
26.2
0.7
11.6
16.9
35.3
8.7
4.1
29.1
14.9
13.4
24.4
80
45
86
73
98
90
1
28
74
79
76
73
85
84
75
75
2,385
276
1,392
2,298
540
188
<25
1,350
1,102
1,188
896
2,068
26
426
78
2,059
130
116
19
33
136
64
<0.1
341
36
24
25
31
1
27
63
133
135
112
70
91
<30
<10
33
30
<10
301
<10
33
45
57
51
648
24.2
15.4
14.5
22.1
28.7
14.1
48.5
29.9
18.4
25.1
12.0
13.0
9.5
11.2
25.9
42.5
3.3
<1.0
1.8
7.9
<1.0
<1.0
1.7
1.7
<1.0
6.4
<1.0
NA
0.8
2.0
<0.7
1.3
7.3
7.6
7.5
7.5
7.5
7.8
7.5
7.3
7.2
7.9
7.2
7.3
8.3
7.9
7.6
6.8
1.2
0.1
2.9
3.8
0.7
0.1
<0.05
0.2
0.2
1.6
0.3
NA
0.3
0.3
0.1
1.9
(a) as SiO2.
(b) as N.
(c) Source water contained elevated radium.
NA = not analyzed
on the choice of oxidants because of concerns over the trihalomethanes (THMs) formation. For example,
at the SC, WV, ST, and AR sites, KMnO4 was used instead of chlorine to oxidize waters due to elevated
levels of TOC and ammonia. Source water pH values ranged from 6.8 to 8.3. Similar to the AM
processes, the pH had some impact on the performance of the IR/CF processes.
3.2
Overview of IR/CF Demonstration Technologies
Most IR/CF technologies involved a two-step process: (1) oxidation of soluble iron and manganese to
form iron and manganese solids (oxidation of soluble manganese with chlorine had slow reaction
kinetics) and (2) filtration of the solids formed. Arsenic in source waters can be removed by taking
advantage of adsorptive capacities of natural iron particles. The ability of a given IR process to remove
arsenic to meet the arsenic MCL depends largely on the amount of arsenic and natural iron in source
waters (Sorg and Logsdon, 1978; Sorg, 1993; Hering et al, 1996; Gulledge and O'Conner, 1973). As a
rule of thumb, source waters having a soluble iron to soluble arsenic mass ratio of 20:1 or greater can
achieve removal to below the arsenic MCL (Sorg, 2002). If source water has an insufficient amount of
natural iron, arsenic removal can be enhanced with supplemental iron addition.
Some IR/CF system designs had a contact tank following chemical addition(s) but prior to pressure
filtration. The extended contact time may result in an increase in arsenic adsorption/removal. A contact
tank can also help reduce the filter loading rate, thereby increasing filter performance and run time.
However, adding a contact tank would increase the system cost and require additional space.
38
-------
After the oxidation step (with or without a contact tank), water was filtered through a filtration media in
either a pressure or a gravity filter to remove arsenic-laden particles. Filter media included silica
sand/anthracite, GreensandPlus™, and proprietary products, such as Macrolite® by Kinetico (currently
marketed by Fairmont Minerals in Chardon, OH), AD26 by AdEdge (Buford, GA), and Electromedia® I
by Filtronics (Anaheim, CA). An anthracite cap of 12 to 18 in was used to prevent excessive head loss
buildup, thus reducing backwash frequency. Effective removal of iron particles was critical to good
arsenic removal because any iron particles present in filter effluent would likely contain (adsorbed)
arsenic.
Table 3-4 summarizes characteristics of different filtration media used in the IR/CF demonstration
systems. Macrolite® is a low-density, spherical, chemically inert ceramic media, designed for higher
filtration rates (i.e., up to 10 gpm/ft2) than those commonly used for conventional filtration processes.
AD26 is a manganese dioxide-based (MnCh) granular media with physical and chemical properties
similar to Pyrolusite (also known as Pyrolox™) and Filox™. Electromedia® I is processed from naturally
occurring minerals and can also handle a high filtration rate of up to 10 gpm/ft2. GreensandPlus™,
branded as AD GS+ by AdEdge, consists of a silica sand core with a thermally bonded MnCh coating,
designed to withstand greater pressure drops and is less prone to stripping of the coating than standard
manganese greensand. Birm® and Filox™ are MnCh-based media commonly used for iron and
manganese removal. An innovative approach using dual Birm®/Filox™media as an alternative to
chemical oxidation was demonstrated at the WL site as a pre-treatment to AM. Silica sand and anthracite
were used in gravity filters at the ST and SD sites as part of the AERALATER® systems. All of the
media have NSF Standard 61 certification for use in drinking water applications.
3.3 IR/CF System Design and Configuration
Because of varying site conditions and source water qualities, the design and basic components of the
IR/CF systems varied among the demonstration sites. Table 3-5 summarizes the design and basic
components of the 18 IR/CF systems demonstrated. Figures 3-1A through 3-1F show photographs of
different types of IR/CF systems and Figure 3-2 shows photographs of chemical feed systems. System
flowrate, use of contact tank(s), filter vessel design, and level of system instrumentation and controls
affected the system performance and cost, and are discussed in the following subsections.
3.3.1 System Flowrate. As shown in Table 3-5, IR/CF system design flowrates were 25 and 60
gpm for the two NTNCWS systems and ranged from 20 to 770 gpm for CWS. The design flowrate of a
system was determined by the capacity of supply well(s) or the peak flow rate. The design flowrate was
used to size the treatment system, thus affecting the system capital cost (Section 3.4). Average flowrates
measured during the performance evaluation studies often were lower than the corresponding design
flowrates. The average flowrates affected the media performance and operational costs, as discussed in
Section 3.5.
3.3.2 Contact/Detention Tank. As shown in Table 3-5, 12 of the 18 systems were equipped with
one or two contact tanks. The AERALATER® systems at the ST and SD sites consisted of an 11- and 12-
ft-diameter aluminum detention tank, providing 34 and 40 min of residence time, respectively. The
detention tank was equipped with an air diffuser grid to further oxidize and mix the chlorinated water.
For the other 10 pressure filtration systems, contact tank sizes varied from 12-in x 62-in to 96-in x 96-in,
providing a contact time of 1.8 to 20 min. These contact tanks were constructed of FRP or CS with a
pressure rating of at least 100 psi.
39
-------
Table 3-4. Characteristics of Filtration Media Used in EPA Demonstration Projects
Parameter
Matrix/ Active Ingredient
Physical Form
Color
Bulk Density (g/cm3[lb/ft3])
Specific Gravity
Mesh Size (U.S. Standard)
Effective Size (mm)
Uniformity Coefficient
pH Range
Filter Rate (gpm/ft2)
Backwash Rate (gpm/ft2)
Manufacturer
No. of EPA Demo Sites
Parameter
Matrix/ Active Ingredient
Physical Form
Color
Bulk Density (g/cm3[lb/ft3])
Specific Gravity
Mesh Size (U.S. Standard)
Effective Size
Uniformity Coefficient
pH Range
Filter Rate (gpm/ft2)
Backwash Rate (gpm/ft2)
Manufacturer
No. of EPA Demo Sites
Parameter
Matrix/ Active Ingredient
Physical Form
Color
Bulk Density (g/cm3[lb/ft3])
Specific Gravity
Mesh Size (U.S. Standard)
Effective Size (mm)
Uniformity Coefficient
pH Range
Filter Rate (gpm/ft2)
Backwash Rate (gpm/ft2)
Manufacturer
No. of EPA Demo Sites
Macrolite
Ceramic,
chemically inert
Dry nodular granules
Taupe, Brown to Grey
0.86 (54)
2.1
40x60
0.25-0.35
1.1-1.2
Inert
8-10
8-10
Kinetico
9
Birm®
<0.01%MnO2
Dry nodular granules
Black
0.64-0.72 (40-45)
2.0
10x40
0.48
2.7
6.8-9.0
3.5-5
10-12
Clack Corporation
AD26(a)
MnO2 (>80%)
Dry nodular granules
Black
2.0 (125)
3.8
20x40
0.40
1.54
6.5-9.0
8-12
18-20
Unknown
2
Filox™
75-85% MnO2
Dry nodular granules
Black
1.83(114)
3.8-4.0
20x40
0.51
1.45
6.5-9.0
5
25-30
Matt-Son, Inc.
1
Anthracite #1
Coal
Dry, crushed
Black
0.8 (50)
1.6
14x30
0.6-0.8
<1.7
Inert
5
12-18
Clack Corporation
Silica Sand
Silica
Dry
Light brown to light red
1.6-1.92 (100-120)
2.6
16 x50
0.45-0.55
<1.6
Inert
3-5
10-20
Many
2
AD GS+(a)
Silica sand core coated
with MnO2
Dry nodular granules
Black
1.4 (85)
2.4
18x60
0.30-0.35
<1.6
6.2-8.5
2-12
10-12
Unknown
1
GreensandPlus™
Silica sand core coated
with MnO2
Dry nodular granules
Black
1.4 (85)
2.4
18x60
0.30-0.35
<1.6
6.2-8.5
3-5
10-12
Inversand
1
Electromedia®I(b)
Unknown
Dry nodular granules
White
NA
NA
NA
NA
NA
NA
Up to 10
NA
Filtronics
1
(a) Marketed and supplied by AdEdge.
(b) Not disclosed by vendor.
NA = not available
Note: Characteristics of G2 media for FC site shown in Table 2-4.
40
-------
Table 3-5. Summary of IR/CF System Design and Components
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Site
ID
GS
FC
SC
WL
DV
wv
CM
CL
TF
SA
SF
ST
SD
GV
FE
PW
OK
AR
Flow rate
D
(gpm)
25
60
20
30
45
96
140
250
250
250
250
250
340
375
375
400
550
770
A
(gpm)
15.2
47.1
1-15
9.3
20 (max)
84
132
153
206
231
89
188
163
285
263
350
538
335
Chemical
Addition
Oxidant
NaCIO
NaCIO
KMn04
None
NaCIO
NaMnO4
NaCIO
NaCIO
NaCIO
NaCIO
NaCIO
NaCIO
NaCIO
NaCIO
NaCIO
NaCIO
NaCIO
KMn04
Iron
Dose
(mg/L
asFe)
No
No
No
No
No
No
0.5
1.8
2.1
No
No
No
No
No
2.2
0.5
0.9
No
Contact
No.
of
Tanks
None
None
2
None
1
None
2
None
2
2
None
1
1
2
2
1
2
1
Tank
Size
(in)
-
-
36x57
-
12x62
-
42x72
-
63 x86
63 x86
-
132x138
144 x 130
63 x86
48 x72
96 x96
48 x96
132 x 84
Contact
Time
(min)
D
-
-
20
-
1.8
-
5
-
5
6.8
-
34
40
4.5
3
6
2
6.5
A
-
-
103
-
4.1
-
5.5
-
6.2
7.4
-
46
69
5.9
4.3
6.8
2.8
14.9
Filtration
No.
of
Filters
3
4
4
2
2
4
2
3
2
2
3
4 cells
3 cells
3
3
2
1
2
Filter
Size
(in)
13 x54
36x72
13x54
24x72
21 x62
36x72
36x72
54x60
48x72
48x72
36x60
132dia
144 dia
48x72
48x72
60x96
84 x 1 12
84x96
Filter
Media
AD26
G281
Macrolite "
Birm*/Filox™
Macrolite *
GreensandPlus™
Macrolite *
ADGS+
Macrolite *
Macrolite
AD26
anthracite/
silica sand
silica sand
Macrolite
Macrolite
Macrolite
Electromedia I
Macrolite
Media
Volume
(ft3)
Per
Filter
2.3
17.7
1.5
5/5
2.4
14.1
14
40
25
25
19
24/24
75.3
25
25
40
174
75
Total
6.9
70.8
6
10/10
4.8
56.4
28
120
50
50
57
95/95
226
75
75
80
174
150
Filtration
Rate
(gpm/ft2)
D
9
2.1
5.4
4.8
9.4
3.4
10
5.2
10
10
6.1
2.6
2.5
10
10
10
10
10
A
5.6
1.7
1.1
1.4
4.2
3.0
9.1
3.2
8.0
9.2
4.2
2.0
1.4
7.6
7.0
8.9
7.0
4.4
A = average; D = design
-------
Figure 3-1A. 20-gpm Macrolite® Pressure Filtration System by Kinetico
(1. Duplex Units, 2. Contact Tanks, 3. Pressure Filters,
4. Chemical Day Tank, and 5. Totalizer on Raw Water Line)
Figure 3-1B. 35-gpm Birm®/Filox™ and Adsorbsia™ GTO™ System by Filter Tech
42
-------
Figure 3-1C. 140-gpm Macrolite Pressure Filtration System by Kinetico
(Clockwise from Left: Control Panel, Macrolite® Filters, and Contact Tanks)
Figure 3-1D. 250-gpm AD26/E33 Filtration System by AdEdge
43
-------
Figure 3-1E. 340-gpm AERALATER® Filtration System by Siemens
(Clockwise from Left: Inlet Piping from Wells; Air Diffiiser Grid within Detention Tank; Prechlorination
Equipment; AERALATER® Unit with Detention Tank and Gravity Cell Influent; and Discharge Piping)
Figure 3-1F. 550-gpm Electromedia® I Filtration System by Filtronics
44
-------
Figure 3-2. Chlorine and Iron Addition Systems
45
-------
3.3.3 Filter Design. As shown in Table 3-5, the pressure filtration systems demonstrated used two
or more filter tanks in parallel for treatment, except for the Electromedia-I® system at the OK site which
used a single horizontal filter tank. The AERALATER® systems consisted of three- or four-cell gravity
filters. The filter cross-sectional area was determined by the design flowrate and the hydraulic loading
rate. Table 3-6 summarizes design and average filtration rates used by different filter media.
The filter size and material affected the system cost. Pressure filter sizes varied from 13-in x 54-in
(smallest) to 84-in x 112-in (largest) with various diameters and heights. The pressure filters were
constructed of FRP, CS, or SS, whereas the AERALATER® chamber was constructed of either aluminum
or CS. The CS or SS filter tanks were ASME-coded for a pressure rating of at least 100 psi. The FRP
tanks were rated for 100 to 150 psi. The costs of FRP tanks were often lower than those of CS tanks for
smaller tanks, but the costs of the two vessel types converged for larger tanks.
Table 3-6. Filtration Rates of Different Filter Media
Filter Media
Macrolite®
Electromedia® I
AD26
GreensandPlus™
ADGS+
Birmw/Filox™
G2S
Anthracite/Silica sand
No. of
Systems
9
1
2
1
1
1
1
2
Design
Filtration Rate
(gpm/ft2)
5.4-10
10.0
6.1,9.0
3.4
5.2
4.8
2.1
2.5,2.6
Average
Filtration Rate
(gpm/ft2)
1.1-9.2
7.0
4.2, 5.6
3.0
3.2
1.4
1.7
1.4,2.0
3.3.4 Instrumentation and Controls. System instrumentation and controls varied among
different IR/CF systems in terms of material, quality, level of complexity/automation, and functionality.
Such variations had an impact on the total capital investment cost and must be taken into consideration
when attempting to compare the costs of different systems. For example, each Kinetico Macrolite*
system was equipped with a turbidimeter to control the backwash operation, which added cost to the
overall system.
3.4
IR/CF System Capital Investment Costs
This section begins with a review of the total capital investment cost, and then follows with a discussion
of three cost categories: equipment, engineering, and installation.
3.4.1 Total Capital Investment Costs. Capital investment costs for all 18 IR/CF demonstration
systems are presented in Table 3-7 in three categories: NTNCWS, small CWS (<100 gpm), and large
CWS (>100 gpm). Capital investment costs ranged from $55,423 for the 25-gpm GS system to $427,407
for the 770-gpm AR system. Figure 3-3 presents capital investment costs of six smaller IR and IR/AM
systems (<100 gpm) (including two NTNCWS and four small CWS systems) as a function of design
flowrates. Figure 3-4 presents similar data for the larger CWS systems (>100 gpm). The IR, IR/AM,
and/or CF systems were plotted using different legends for easy identification. The data for the IR
systems indicated a stronger correlation between the costs and flowrates on both figures. Curve fitting
using linear regression was performed on the data set for the IR systems, yielding an R2 of 0.8342 and
0.8808 for smaller and larger systems, respectively. Curve fitting was not performed on IR/AM or CF
data due to insufficient data points.
46
-------
Table 3-7. Capital Investment Costs for IR/CF Systems
No.
Site
ID
Technology (Media)
Design
Flow
Rate
(gpm)
Total
Capital
Cost
($)
Normalized
Capital
($/gpm)
Normalized
Capital
($/gpd)
Annualized
Cost
($/yr)
Unit Cost
($/kgal of water)
Design(a)
Average
Utilization
Rate(b)
(%)
Non-Transient Non-Community Water Systems
1
2
GS
FC
IR (AD26)+AM (E33)
ffi. (G2*)
Average
25
60
$55,423
$128,118
$2,217
$2,135
$2,176
$1.54
$1.48
$1.51
$5,231
$12,093
$0.40
$0.38
$0.39
$10.12
$14.32
$12.22
4
3
3.5
Community Water Systems (<100 gpm)
3
4
5
6
SC
WL
DV
WV
IR (Macrolite®)
IR (Birm®/Filox™) +
AM (Adsorbsia™ GTO™)
IR (Macrolite®)
IR (GreensandPlus™)
MZHHWMTW
MaxwwM/w
Average
20
30
45
96
20
96
$63,547
$66,362
$60,500
$161,560
$55,423
$161,560
$3,177
$2,212
$1,344
$1,683
$1,344
$3,177
$2,104
$2.21
$1.54
$0.93
$1.17
$0.93
$2.21
$1.46
$5,998
$6,264
$5,711
$15,250
$5,711
$15,250
$0.57
$0.40
$0.24
$0.30
$0.24
$0.57
$0.38
$3.75
$2.05
$2.61
$1.33
$1.33
$3.75
$2.44
15
19
9
23
9
23
17
Community Water Systems (> 100 gpm)
7
8
9
10
11
12
13
14
15
16
17
18
CM
CL
TF
SA
SF
ST
SD
GV
FE
PW
OK
AR
IMA (Macrolite®)
CF (AD GS+)
CF (Macrolite*)
IR (Macrolite®)
IR (AD26) + AM (E33)
IR (AERALATER®) + AM (E33)
IR (AERALATER*)
IR (Macrolite®)
CF (Macrolite*)
IMA (Macrolite®)
CF (Electromedia® I)
IR (Macrolite®)
Minimum
Maximum
Average
140
250
250
250
250
250
340
375
375
400
550
770
140
770
$270,530
$216,876
$305,447
$287,159
$292,252
$367,838
$364,916
$332,584
$334,297
$334,573
$424,817
$427,407
$216,876
$427,407
$1,932
$868
$1,222
$1,149
$1,169
$1,471
$1,073
$887
$891
$836
$772
$555
$555
$1,932
$1,069
$1.34
$0.60
$0.85
$0.80
$0.81
$1.02
$0.75
$0.62
$0.62
$0.58
$0.54
$0.39
$0.39
$1.34
$0.74
$25,535
$20,471
$28,831
$27,105
$27,586
$34,720
$34,444
$31,393
$31,554
$31,580
$40,098
$40,343
$20,471
$40,343
$0.35
$0.16
$0.22
$0.21
$0.21
$0.26
$0.19
$0.16
$0.16
$0.15
$0.14
$0.10
$0.10
$0.35
$0.19
$1.85
$1.02
$1.06
$2.22
$1.64
$1.80
$0.57
$1.31
$0.83
$0.82
$0.29
$0.40
$0.29
$2.22
$1.15
19
15
21
9
13
15
34
12
19
18
48
25
9
48
21
(a) System's maximum capacity at design flowrate, operating 24 hr a day, 365 days a year.
(b) Ratio of a system's average annual production to its maximum capacity at design flowrate.
AM = adsorptive media; CF = coagulation/filtration; IA = supplemental iron addition; IR = iron removal
-------
$175,000
$150,000
$125,000
O $100,000 -
&
O $75,000 -
I
$50,000
$25,000 -
$0
y= 1429.3X +24460
R2 = 0.8342
IR + AM
IR
Linear (IR)
0
20 40 60
Design Flowrate (gpm)
80
100
Figure 3-3. Total Capital Investment Costs of Smaller IR/CF Systems (<100 gpm)
$450,000
$400,000
8 $350,000 -
Q.
8
-= $300,000 -
o
$250,000
y = 247.33x +242417
R2 = 0.8808
$200,000
* IR
• IR + AM
A C/F
Linear (IR)
100
300 500 700
Design Flowrate (gpm)
900
Figure 3-4. Total Capital Investment Costs of Larger IR/CF Systems (>100 gpm)
48
-------
Similar to the AM systems, the capital investment cost of each IR/CF system was divided by its design
capacity in gpm and gpd and the results are shown in Table 3-7 and Figures 3-5 and 3-6. Normalized
costs for smaller CWS systems (<100 gpm) ranged from $1,344 to $3,177/gpm (or $0.93 to $2.21/gpd)
and averaged $2,104/gpm (or $1.46/gpd). Normalized costs for the larger CWS ranged from $555 to
$l,932/gpm (or $0.39 to $1.34/gpd) and averaged $l,069/gpm (or $0.74/gpd). As expected, the larger
systems had lower average costs per gpm (or gpd) of the design capacity than the smaller ones. Both
Figures 3-5 and 3-6 clearly show a decreasing trend with increasing flowrates, reflecting the economy of
scale.
As stated in Section 3.3, in addition to flowrate, several other design parameters also affected system
costs. A good way of demonstrating the effects of these parameters is to compare the costs and design
features of the five 250-gpm systems, including two CF (at CL and TF), one IR (at SA), and two IR/AM
systems (at SF and ST). Total capital investment costs of these five systems ranged from $216,876 for
the AD GS+ system at CL to $367,838 for the AERALATER®/E33 system at ST (or $868 to $l,471/gpm
or $0.60 to $1.02/gpd). Comparing the two 250-gpm CF systems, the TF system cost was 40% higher
than that of the CL system. The difference could be attributed to at least three factors, i.e., filter media,
contact tank, and instrumentation and control. The TF system used Macrolite*, a more expensive media
than AD GS+ used by the CL system. The TF system included two 63-in x 86-in contact tanks while the
CL system did not use any contact tank. Also, the TF system had more advanced and sophisticated
instrumentation than the CL system. Because Macrolite* had a higher design filtration rate than AD GS+
(8.0 vs. 3.2 gpm/ft2), the TF system used fewer and smaller filter vessels (i.e., two 48-in x 72-in FRP
tanks) than the CL system (i.e., three 54-in x 60-in CS tanks). However, the higher filtration rate did not
result in a lower total system cost because of the other design features as discussed.
Other factors were iron addition and AM systems included in the system design. For example, the TF and
SA sites had identical Macrolite® systems, but the TF system was equipped with iron addition while the
SA system was not. The cost of the TF system (with iron addition) was $18,288, or 6.6% higher than that
of the SA system (without iron addition). Using an AM system for post-treatment also increased the
system cost. The IR/AM systems at the SF and SD sites cost 8 to 36% more than the average of the other
cost of three IR and CF systems without AM.
Unit costs (total capital investment) of the 18 systems expressed as 1,000 gal of water treated are also
shown in Table 3-7. These unit costs were calculated based on the average and maximum annual
production rates similar to those for the AM systems (Section 2.4.1). The ratio of a system's average
annual production to its maximum capacity at the design flowrate is the utilization rate, which affected
the unit capital investment cost. In Figure 3-7, unit costs are plotted against utilization rates for three
groups of systems: NTNCWS, smaller CWS (<100 gpm), and larger CWS (>100 gpm). The systems in
the NTNCWS and smaller CWS groups had comparable flow ranges. However, because the NTNCWS
systems had significantly lower utilization rates than those in the smaller CWS group, i.e., 3.5% vs. 17%
(on average), their unit costs per 1,000 gal were significantly higher than those for the smaller CWS
group (i.e., $12.22 vs. $2.44 on average). On the other hand, because the systems in the smaller and
larger CWS groups had rather comparable utilization rates, i.e., 17% vs. 21% (on average), unit costs of
the systems in the smaller CWS group were about twice of those in the larger CWS group, i.e., $2.44 vs.
$1.15 (on average). Therefore, the NTNCWS systems had the highest unit costs due to small sizes and
low utilization rates.
49
-------
$2.50
$2.00
•c
Q.
O)
„ $1.50 -
o
O
I
ra $1.00
O
S
o
$0.50
$0.00
A A
20 40 60 80 100
Design Flow/rate (gpm)
Figure 3-5. Smaller IR/CF System Capital Investment Costs per gpd of
Design Capacity (<100 gpm)
$1.60
$1.40
•D $1.20
0>
g. $1.00 -
in
o
" $0.80
3
'o.
O $0.60
$
>- $0.40
$0.20
$0.00
* IR
• IR + AM
A C/F
Power (IR)
100
300 500 700
Design Flowrate (gpm)
900
Figure 3-6. Larger IR/CF System Capital Investment Costs per gpd of
Design Capacity (>100 gpm)
50
-------
$100.00 -i
as
O)
o
o
o
100 gpm
$1.00
$0.10
0% 10% 20% 30% 40%
Utilization Rate
50%
60%
Figure 3-7. IR/CF System Unit Capital Investment Costs
as a Function of Utilization Rates
3.4.2 Equipment Cost. Except for the GreensandPlus™ system at WV and the two
AERALATER® package units at ST and SD, all other IR/CF treatment systems were skid-mounted with
filtration vessels, piping and valves, and instrument and controls all mounted on individual steel frames.
The equipment cost of a system generally included the cost for the skid-mounted system, filter media,
miscellaneous materials and supplies, freight, user's manual, and vendor's labor. It also included the cost
for a chemical feed system, if any. In some cases (like at WL, CL, and TF), the cost of backwash recycle
equipment, such as backwash storage tank(s) and recycle pump, was also included in the equipment cost.
Equipment costs for the treatment system ranged from $19,790 for the 45-gpm DV system to $296,430
for the 550-gpm OK system, as shown in Table 3-8. On average, equipment costs accounted for 48% and
64% of total capital investment costs for the smaller CWS (<100 gpm) and larger CWS (MOO gpm),
respectively. Figures 3-8 and 3-9 plot equipment costs against flowrates for the smaller (<100 gpm) and
larger systems (>100 gpm). Because equipment costs made up the highest percentage of the total capital
investment costs, equipment cost curves generally were similar to total capital investment cost curves
shown in Figures 3-3 and 3-4. Curve fittings were performed on the data for the IR systems, yielding an
R2 of 0.5776 and 0.9297 for the smaller and larger systems, respectively.
3.4.3 Site Engineering Cost. Site engineering costs for the IR/CF systems ranged from $3,850 for
the 30-gpm WL system to $53,435 for the 250-gpm TF system. These costs represented, on average,
21% and 12% of total capital investment costs for the smaller (<100 gpm) and larger CWS (>100 gpm),
respectively (see Table 3-8). The percentage decreased as the size of the system increased, as expected.
51
-------
Table 3-8. Summary of Equipment, Site Engineering, and Installation Costs of IR/CF Systems
to
No.
Site
Technology
Design
Flow
Rate
(gpm)
Total
Capital
Cost
($)
Equipment
Cost
%of
Total
Site
Engineering
Cost
%of
Total
Installation
&Startup
Cost
%of
Total
Non-Transient Non-Community Water Systems
1
2
GS
FC
IR (AD26) + AM (E33)
IR (G2®)
Average
25
60
$55,423
$128,118
$91,771
$31,735
$103,118
67,426
57
80
69
$11,278
$7,500
$9,389
20
6
13
$12,410
$17,500
$14,955
22
14
18
Community Water Systems (<100 gpm)
3
4
5
6
SC
WL
DV
WV
IR (Macrolite®)
IR (Birm*/Filox™) +
AM (Adsorbsia™ GTO™)
IR (Macrolite®)
IR (GreensandPlus™)
Minimum
Maximum
Average
20
30
45
96
20
96
$63,547
$66,362
$60,500
$161,560
$60,500
$161,560
$87,992
$22,422
$46,267
$19,790
$90,750
$19,790
$90,750
$44,807
35
70
33
56
33
70
48
$20,227
$3,850
$20,580
$22,460
$3, 850
$22,460
$16,779
32
6
34
14
6
34
21
$20,898
$16,245
$20,130
$48,350
$16,245
$48,350
$26,406
33
24
33
30
24
33
30
Community Water Systems (> 100 gpm)
7
8
9
10
11
12
13
14
15
16
17
18
CM
CL
TF
SA
SF
ST
SD
GV
FE
PW
OK
AR
IMA (Macrolite®)
CF (AD GS+)
CF (Macrolite®)
IR (Macrolite®)
IR (AD26) + AM (E33)
IR (AERALATER®) +AM (E33)
IR (AERALATER®)
IR (Macrolite®)
CF (Macrolite®)
IMA (Macrolite®)
CF (Electromedia-I®)
IR (Macrolite®)
Minimum
Maximum
Average
140
250
250
250
250
250
340
375
375
400
550
770
140
770
$270,530
$216,876
$305,447
$287,159
$292,252
$367,838
$364,916
$332,584
$334,297
$334,573
$424,817
$427,407
$216,876
$427,407
$329,891
$159,419
$161,650
$168,142
$160,875
$212,826
$273,873
$205,800
$196,542
$201,292
$224,994
$296,430
281,048
$159,419
$296,430
$211,908
59
75
55
56
73
74
56
59
60
67
70
66
55
75
64
$39,344
$21,726
$53,435
$49,164
$27,527
$16,520
$27,077
$48,057
$44,520
$30,929
$48,332
$50,770
$16,520
$53,435
$38,117
15
10
17
17
9
4
7
14
13
9
11
12
4
17
12
$71,767
$33,500
$83,870
$77,120
$51,899
$77,445
$132,039
$87,985
$88,485
$78,650
$80,055
$95,589
$33,500
$132,039
$79,867
27
15
27
27
18
21
36
26
26
24
19
22
15
36
24
-------
$125,000
$100,000
« $75,000 -
o
o
4-1
ii
•5 $50,000 -
$25,000 -
$0
y = 1054.1x +782.13
R2 = 0.5776
20 40 60 80
Design Flowrate (gpm)
100
Figure 3-8. Equipment Costs of Smaller IR/CF Systems (<100 gpm)
$350,000
$300,000
to $250,000
o
o
100 gpm)
53
-------
3.4.4 Installation Cost. Installation costs for the IR/CF systems ranged from $16,245 for the 30-
gpm WL system to $132,039 for the 340-gpm SD system. The installation cost of the 12-ft diameter
AERALATER® at the SD site was 70% higher than that of the 11-ft diameter AERALATER® and E33
system at the ST site. These installation costs represented 30% and 24% of total capital investment costs
for the smaller (<100 gpm) and larger CWS (MOO gpm), respectively (see Table 3-8). The percentage
decreased as the size of the system increased, as expected.
3.5 IR/CF System O&M Cost
O&M costs for the IR/CF systems included the cost of chemical supplies, electricity consumption, and
labor to operate the arsenic treatment system. The backwash residual disposal cost was not included.
Table 3-9 is a summary of O&M cost breakdowns for the 18 systems. Total O&M costs ranged from
$0.07 to $1.93 per 1,000 gal of water treated. These costs were obtained from the first year system
operations, when the systems were under warranty and required few repairs. Each cost component is
discussed below.
3.5.1 Chemical Cost. Chemicals used for IR/CF system operations included NaCIO, gas C12,
KMnO4, and/or NaMnO4 for oxidation/disinfection and/or an iron salt for coagulation. Where
chlorination already existed at the facility for disinfection purposes, it was switched to pre-chlorination to
oxidize soluble As(III), Fe(II), and/or Mn(II) before treatment. At sites where source water contained
elevated TOC and ammonia, KMnO4 or NaMnO4 was used instead of chlorine. Incremental costs for
chlorination/oxidation were negligible at three sites (e.g., FC, ST, and GV) and ranged from $0.01 to
$0.37 per 1,000 gal of water treated for the other nine sites.
Iron addition was implemented at six sites, including four CF sites where iron was used as a coagulant
and two IR sites where iron was added to supplement natural iron for better arsenic removal. Table 3-10
presents chemical costs for iron addition at these six sites. A 40% FeCl3 solution in 15- or 55-gal drums
was used at all sites. Iron dose rates ranged from 0.5 to 2.2 mg/L (as Fe). The costs of iron addition
ranged from $0.01 to $0.07 per 1,000 gal of water treated.
Total chemical costs ranged from zero to $0.37 per 1,000 gal of water treated, accounting for zero to 57%
(19% on average) of the total O&M costs.
3.5.2 Electricity Cost. The electricity cost was tracked by comparing the monthly electrical bills
before and after the installation of the arsenic treatment system. If the site did not have a separate meter
for the arsenic treatment system, then the cost was estimated based on the power requirements of the
major equipment such as compressors, pumps, control panels, etc., the average operational hours, and the
local electricity unit price. Local electricity unit prices ranged from $0.06 to $0.14 per kwh provided by
the facilities.
The incremental electrical consumption was negligible for most of the systems. Electricity costs per
1,000 gal of water treated ranged from zero to $0.39 averaged $0.07, as shown in Table 3-9. It accounted
for zero to 59% (19% on average) of the total O&M costs. The highest cost was incurred at the WL site
because the well(s) ran almost around the clock.
3.5.3 Labor Cost. Labor costs accounted for 18 to 95% (61% on average) of the total O&M costs.
Routine, non-demonstration related labor activities consumed only 10 to 30 min a day, one or several
days a week at most of the sites. Average weekly hours ranged from 25 min to 10 hr and averaged 3.4 hr.
As shown in Table 3-9, labor rates ranged from $10.8 to $30/hr and averaged $22.6/hr; these rates might
be lower than those in certain regions of the country, such as California, but were actual numbers
provided by the operators. Labor cost per 1,000 gal of water treated averaged $2.41 for the two
54
-------
Table 3-9. O&M Costs for IR/CF Systems
No.
Site
ID
Technology
Desig
n
Flow
Rate
(gpm)
Total
O&M
Costs
($/kgal)
Chemicals
Type
Cost
($/kgal)
%of
Total
O&M
Electricity
Cost
($/kgal)
%of
Total
O&M
Labor
Average
Weekly
Hours
(hr)
Labor
Rate
($/hr)
Cost
($/kgal)
%of
Total
O&M
Non-Transient Non-Community Water Systems
1
2
GS
FC
IR (AD26)+AM (E33)
IR (G2®)
25
60
$2.90W
$2.26
NaCIO
NaCIO
$0.33
$0.00
11
0
$0.00
$0.00
0
0
1.6
1.67
$16.0
$22.0
$2.57
$2.26
89
100
Community Water Systems
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
SC
WL
DV
WV
CM
CL
TF
SA
SF
ST
SD
GV
FE
PW
OK
AR
IR (Macrolite®)
IR (Birm®/Filox™) +
AM Adsorbsia™ GTO™)
IR (Macrolite®)
IR (GreensandPlus™)
IMA (Macrolite®)
CF (AD GS+)
CF (Macrolite®)
IR (Macrolite®)
IR (AD26)+AM (E33)
IR (AERALATER®) +
AM (E33)
IR (AERALATER®)
IR (Macrolite®)
CF (Macrolite®)
IMA (Macrolite®)
CF (Electromedia® I)
IR (Macrolite®)
M/WOTMOT
Mactzmttw
Average
20
30
45
96
140
250
250
250
250
250
340
375
375
400
550
770
20
770
$0.36
$1.93(a)
$0.26
$0.65
$0.29
$0.46
$0.18
$0.43
$0.33(a)
$0.16(a)
$0.27
$0.55
$0.31
$0.17
$0.18
$0.07
$0.07
$1.93
$0.40
KMnO4
None
NaCIO
NaMnO4
FeCl3
FeCl3
FeCl3
NaCIO
NaCIO
NaCIO
NaCIO
NaCIO
FeCl3
FeCl3
FeCl3
NaCIO
KMnO4
$0.07
$0.00
$0.09
$0.37
$0.03
$0.07
$0.02
$0.05
$0.17
$0.00
$0.04
$0.00
$0.05
$0.01
$0.03,
$0.01
$0.03
0
$0.37
$0.06
19
0
34
57
10
15
9
12
51
0
15
0
16
8
17
43
0
57
19
$0.01
$0.39
$0.06
$0.16
$0.04
$0.06
$0.01
$0.01
$0.00
$0.08
$0.16
$0.03
$0.05
$0.05
$0.08
$0.00
0
$0.39
$0.07
3
20
24
25
14
13
3
2
0
50
59
5
15
29
44
0
0
59
19
0.42
3
0.42
1.75
2.5
6
4.7
1.75
2.33
1.7
4.5
10
5.25
2.5
5.25
2.5
0.4
10.0
3.4
$21.0
$30.0
$10.8
$15.0
$21.0
$22.0
$19.6
$10.0
$21.0
$16.3
$18.0
$24.0
$30.0
$30.0
$30.0
$30.0
10.8
30.0
22.6
$0.28
$1.54
$0.11
$0.12
$0.22
$0.33
$0.16
$0.37
$0.16
$0.08
$0.07
$0.52
$0.21
$0.11
$0.06
$0.04
$0.04
$1.54
$0.27
78
80
42
18
76
72
88
86
48
50
26
95
69
64
33
57
18
95
61
(a) Media replacement cost not incurred during the study period; thus, not included in the total O&M cost.
-------
Table 3-10. Cost of Iron Addition for IR/CF Systems
Site
ID
CM
CL
TF
FE
PW
OK
Technology
IMA (Macrolite®)
CF (AD GS+)
CF (Macrolite®)
CF (Macrolite®)
IMA (Macrolite®)
CF (Electromedia® I)
Flow
rate
(gpm)
140
250
250
375
400
550
Raw
Water
As
Levels
MS/L)
36.5
29.0
84.0
34.4
17.7
17.9
Raw
Water
Fe
Levels
(MS/L)
540
188
<25
26
426
78
Raw
Water
Fe/As
Ratio
15
6
<1
<1
24
4
Fe
Dosage
(mg/L
as Fe)
0.5
1.8
2.1
2.2
0.5
0.9
Cost
($/kgal
of
water)
$0.03
$0.07
$0.02
$0.05
$0.01
$0.03
(a) All sites used a 40% FeCl3 solution.
and varied from $0.04 to $1.54 for the 16 CWS because annual water production rates of the treatment
systems varied significantly. A NTNCWS often had a lower demand and a lower utilization rate than a
CWS. Therefore, the labor cost (per 1,000 gal of water treated) of a smaller NTNCWS tended to be
higher than that of a larger CWS.
56
-------
4.0 OTHER ARSENIC TREATMENT TECHNOLOGIES
This section presents the cost information on two IX, one RO, and two POU arsenic demonstration
systems. Table 4-1 presents demonstration locations, technologies, and study durations. The
performance evaluation study on each IX system lasted much longer than 12 months to address issues of
resin fouling which occurred at both sites. The demonstration of the RO system was conducted for 10
months because RO is a relatively mature technology and because a four-month pilot system had been
previously conducted by EPA at the CE site. Capital investment and O&M cost data collected from these
systems are presented in this section. An overview of the demonstration sites, system design and
configurations is also provided to support the cost data. Detailed information on the performance and
capital investment and O&M costs on the systems can be found in individual performance evaluation
study reports provided on the EPA Arsenic Demonstration Program Web site.
Table 4-1. Summary of IX, RO, and POU Demonstration Locations,
Technologies, and Study Durations
No.
Site
ID
Demonstration
Location
Technology
Vendor
Design
Flowrate
(gpm)
Study
Duration
Length
of Study
(mon)
Non-Transient Non-Community Water Systems
1
2
CE
KF-
POU
Carmel, ME
Klamath Falls, OR
RO (Dual Plumbing
Distribution)
POU ARM 200
Norlen's Water
Kinetico
1,200 gpd
8 units
02/09-12/09
12/05-11/06
10
11
Community Water Systems
o
J
4
5
HD
FL
VA
Homedale, ID
Fruitland, ID
Vale, OR
POURO
IX (A300E)
IX (Arsenex III
PFA300E)
Kinetico
Kinetico
Kinetico
9 units
250
540
07/05-06/06
06/05-02/08
09/06-03/10
12
32
42
AM = adsorptive media; IX = ion exchange; POU = point of use; RO = reverse osmosis
4.1
Overview of Demonstration Sites
Table 4-2 summarizes the IX, RO, and POU demonstration site information, including two NTNCWS and
three CWS. At the CE site, an innovative approach using a POE RO unit coupled with dual plumbing in
the distribution system was demonstrated as a low cost alternative to achieve compliance with arsenic and
antimony MCLs, compared to conventional RO treatment. At the KF site, eight POU ARM 200
cartridges were installed either under a sink or inside a drinking water fountain in eight college buildings.
The HD site consisted of nine residences where a POU RO unit was installed at each residence. FL and
VA are municipal facilities where IX was used to remove both arsenic and nitrate.
Table 4-3 presents average values of several source water quality parameters measured at the five sites
during the performance evaluation studies. Arsenic concentrations in source waters varied from 18.2 to
57.8 |jg/L with soluble As(V) being the predominant arsenic species at all five sites. The source waters
also contained several co-contaminants, including antimony (Sb) at the CE site, nitrate (NO3) at the HD,
FL, and VA sites, and uranium (U) at the HD sites. The presence of these co-contaminants in source
waters was the main reason for selecting RO as the treatment technology at the CE and HD sites and IX at
the FL and VA sites.
57
-------
Table 4-2. Summary of IX, RO, and POU Demonstration Sites
No.
Site
ID
Design
Flow
Rate
(gpm)
Average
Flow
Rate
(gpm)
Daily
Op
Time
(hr/day)
Average
Daily
Demand
(gpd)
Annual
Production
(Kgal)
Utilization
Rate
(%)
Pre-existing
Treatment
Non-Transient Non-Community Water Systems
1
2
CE
KF-
POU
1,200 gpd
NA
0.8 (permeate);
1.2 (reject)
NA
11.7
NA
1,486W
NA
108,912
NA
25%
NA
C12
C12
Community Water Systems
o
3
4
5
HD
FL
VA
NA
250
540
NA
157
534
NA
17.4
9.5
NA
166,895
274,473
NA
65,400
111,100
NA
51%
39%
None except
for softeners
at 3 homes
None
C12
(a) Including 562 gpd potable and 924 gpd non-potable demand.
NA = not applicable
Table 4-3. Summary of IX, RO and POU Site Source Water Quality
Site ID
Parameter
Total As
As(III)
NO3 (as N)
Total Sb
Total U
Total V
Total Fe
Total Mn
Total P
S04
TDS
TOC
Silica
Total Hardness
Total Alkalinity
pH
Unit
ug/L
ug/L
mg/L
ug/L
ug/L
ug/L
ug/L
ug/L
ug/L
mg/L
mg/L
mg/L
mg/L
mg/L(a)
mg/L(a)
S.U.
CE
KF-
POU
HD
FL
VA
Average Values
18.2
0.2
0.2
10.8
NA
0.5
<25
2.2
<10
9.8
255
NA
11.2
217
206
7.9
29.8
0.3
0.7
NA
0.3
35.0
<25
0.4
<10
24
200
<0.7
30
83
116
8.0
57.8
1.5
10.2
NA
27.4
32.4
112
0.6
<10
167
685
1.8
66.5
238
295
7.3
42.5
1.2
10.0
<0.1
19.4
39.3
<25
22.1
320
59
580
1.6
57
249
387
7.6
22.6
1.0
5.4
NA
6.1
54.1
<25
0.4
278
82
514
2.0
55.6
165
329
7.4
(a) as CaCO3.
NA = not available; TDS = total dissolved solids;
TOC = total organic carbon
The presence of total dissolved solids (TDS) and sulfate in source waters could affect the IX system
performance and therefore the treatment cost, but the levels measured at the FL and VA sites were not
high enough to cause adverse effects. However, the presence of TOC and silica in source waters was
found to cause resin fouling at both the FL and VA sites. Water pH values ranged from 7.4 to 8.0. Water
pH does not impact the IX or RO process as it would to the AM process.
58
-------
4.2
IX Demonstration Systems
Four strong based anionic (SBA) IX resins manufactured by Purolite® were evaluated at the FL and VA
sites. At FL, A300E was used to remove arsenic and nitrate. At VA where two studies were conducted,
Arsenex II was used initially in Study Period I. Because of organic fouling, Arsenex II was replaced
during Study Period II with PFA300E top-dressed with A850END. PFA300E was very similar to the
A300E used at FL. All of these resins have NSF Standard 61 certification for use in drinking water
applications. Their physical and chemical properties are presented in Table 4-4.
Table 4-4. Properties of IX Resins Used for EPA Demonstration Projects
Parameters
Polymer Structure
Functional Group
Physical Form and
Appearance
Whole Bead Count
Resin Type
Ionic Form, as Shipped
Shipping Weight (g/L or
[lb/ft3])
Specific Gravity (g/mL)
Mesh Size(b) (Wet)
Bead Size Range (mm)
Uniformity Coefficient
Moisture Retention (%)
Reversible Swelling
Total Exchange Capacity,
Cl" Form (eq/L) (wet,
volumetric)
pH Range
Maximum Temperature
Limit (°C/°F)
Arsenex II
Gel polystyrene
crosslinked with
DVB
Dimethyl ethanol
amine
Opaque spherical
beads
95% minimum
SBA Type II
cr
0.69 (43)
-
16 x50
0.3-1.2
-
42-54
Cr to SO/7NO3~
Negligible
1.0
0-14
100/212
A850END(a)
Gel polyacrylic
crosslinked with
DVB
Trimethylamine
Clear spherical
beads
-
SBA Type I
cr
0.68-0.73
(42.5-45.6)
1.09
-
0.60-0.85
1.70
57-62
cr to OH-
15% (max)
1.25
1-10
85/185
PFA300
Gel polystyrene
crosslinked with
DVB
Dimethyl ethanol
amine
Amber spherical
beads
95% minimum
SBA Type II
cr
0.69 (43)
1.10
25 x40
+0.7 10mm
-------
Table 4-5. Summary of IX System Design and Components
Site ID
Design Flowrate (gpm)
Average Flowrate (gpm)
No. of Tanks
Tank Size (in)
Resin Type
Resin Volume/Tank (ft3)
Total Resin Volume (ft3)
Average Hydraulic Loading
(gpm/ft2)
Design EBCT (min)
Average EBCT (min)
Design Salt Loading (lb/ft3)
Average Salt Loading (lb/ft3)
Salt Saturator (in)
Brine Day Tank (in)
Pre-treatment
FL
250
157
2
48 D x 72 H
A300E
50
100
6.2
3.0
4.8
10
9.5
One, 96 D x 148 H (15-
ton capacity)
One, 610x64 H (685
gal)
Five 20-|am bag filters in
parallel
VA
540
536
2
63 D x 86 H
Arsenex II
93
186
12.3
3.0
2.6
12
12.8
A850END/PFA300E
16.7/81.7
33.4/163.4
12.4
3.0
2.8
10
9.3
Two, 96 D x 120 H (1 1-ton capacity)
Two, 610x97 H (1,050 gal)
Two banks of five 5- or 20-|am bag
filters
Figure 4-1. Photograph of IX-248-As/N System at Fruitland, ID
60
-------
The IX systems were regenerated in a downflow, co-current mode using brine. Triggered automatically
by a throughput setpoint in a PLC, the two IX tanks were regenerated sequentially, each cycling through
the steps of brine draw, slow rinse, and fast rinse before returning to service. The regeneration waste
stream was discharged to the sewer at FL and an evaporation pond outside of the plant at VA.
The IX systems were fully automatic and controlled by the PLC in the central control panel. The control
panel also contained a touch screen OIP that allowed the operator to monitor system flowrate and
throughput since last regeneration. The OIP also allowed the operator to change system setpoints, as
needed, and check status of alarms. Setpoint screens were password-protected so that changes could only
be made by authorized personnel. Typical alarms were for no flow, storage tank high/low, and
regeneration failure.
4.2.2 IX System Capital Investment Costs. Table 4-6 presents total capital investment costs for
the two IX systems. The total capital investment costs included the cost for equipment, site engineering,
and installation as shown in Table 4-7. The cost associated with the new building, sanitary sewer
connection (at FL), construction of an evaporation pond and ancillary equipment (at VA), and other
infrastructure improvement was not included in the capital investment costs.
Table 4-6. Total Capital Investment Costs for IX Systems
Site
FL
VA
Design
Flow
rate
(gpm)
250
540
Total
Capital
Cost
($)
$286,388
$395,434
Normalized
Capital
Cost
($/gpm)
$1,146
$732
Normalized
Capital
Cost
($/Spd)
$0.80
$0.51
Annualized
Capital
Cost(a)
($/yr)
$27,032
$37,325
Unit Cost
(/Skgal of water)
Design(b)
$0.21
$0.13
Average
$0.47
0.34
Utilization
Rate(c)
(%)
44
39
(a) Obtained by applying a CRF of 0.09439 (based on a 7% interest rate and a 20-year return period) to
total capital cost.
(b) System's maximum capacity at design flowrate, operating 24 hr a day, 365 days a year.
(c) Ratio of a system's average annual production to its maximum capacity at design flowrate.
Table 4-7. Summary of Equipment, Site Engineering, and Installation Costs of IX Systems
Site
ID
FL
VA
Design
Flow
Rate
(gpm)
250
540
Total
Capital
Cost
($)
$286,388
$395,434
Equipment
Cost
$173,195
$260,194
%of
Total
61
66
Site
Engineering
Cost
$35,619
$49,840
%of
Total
12
13
Installation
&Startup
Cost
$77,574
$85,400
%of
Total
27
22
The total capital investment cost of the VA system was 38% higher than that of the FL system, but its
capacity was more than double the FL system. Therefore, in terms of the capital cost per gpm or gpd of
the design capacity, the VA system is 36% lower than the FL system. Annualized and unit capital costs
per 1,000 gal of water treated are also presented in Table 4-6. As expected, the unit cost based on the
average production was higher than that based on the maximum capacity. The ratio of the average
production to the maximum capacity, expressed as utilization rate, was comparable for both IX systems,
i.e., 44% for FL and 39% for VA.
61
-------
Equipment Cost. Both IX treatment systems were skid-mounted on a steel frame. Similar to an AM and
a IR/CF system, the equipment cost of an IX system included the cost for the skid-mounted system, resin
media, miscellaneous materials and supplies, freight, user's manual, and vendor's labor. It also included
the cost for the salt delivery system, which consisted of one or two salt saturators, brine day tanks, and
brine pumps. The equipment cost of the VA system was about 50% more than that of the FL system.
The equipment cost accounted for 61% and 66% of the respective total capital investment costs for the FL
and VA systems, making up the highest percentage of the total capital investment costs.
Site Engineering Cost. Site engineering costs included the cost for the necessary design work and
engineering plans preparation. The equipment cost of the VA system was 40% more than that of the FL
system. The engineering cost represented 12 or 13% of the total capital investment costs for both
systems.
Installation Cost. The installation cost of the VA system was about 10% more than that of the FL
system. The equipment cost accounted for 27% and 22% of the total capital cost for the FL and VA
systems, respectively.
4.2.3 IX System O&M Costs. The O&M cost evaluated for the IX systems included the
incremental cost associated with the salt supply, electricity consumption, and labor. The disposal cost of
regeneration residual was not included. Table 4-8 is a summary of the cost breakdowns of the O&M
costs for the two IX systems. The total O&M cost was $0.62 and $0.35 per 1,000 gal of water treated for
the FL and VA systems, respectively. These costs were obtained from the first year system operations,
when any system repairs were covered by the warranties. Each cost component is discussed below.
Table 4-8. O&M Costs for IX Systems
Site
ID
FL
VAW
Design
Flow
Rate
(gpm)
250
540
Total
O&M
Costs
($/kgal)
$0.62
$0.35
Salt Supply
Type
Salt
Salt,
caustic
Cost
($/kgal)
$0.49
$0.29
%of
Total
O&M
79%
83%
Electricity
Cost
($/kgal)
$0.08
$0.03
%of
Total
O&M
13%
8%
Labor
Average
Wkly
Hours
2.5
3.3
Labor
Rate
($/hr)
$21.0
$21.0
Cost
($/kgal)
$0.05
$0.03
%of
Total
O&M
8%
10%
(a) Resin replacement cost not included in total O&M cost.
Salt Supply Cost. The IX system used salt for resin regeneration. Caustic soda was mixed with brine to
help remove organic foulants from the resin periodically. The average salt use rate per 1,000 gal of water
treated was 3.6 Ib at VA and 4.4 Ib at FL. The unit salt price was cheaper at VA ($0.076 verse $0.11/lb)
because VA purchased salt in bulk quantities (i.e., half truck load). The salt costs per 1,000 gal of water
treated were $0.29 at VA and $0.49 at FL, accounting for 83% and 79% of the total O&M costs,
respectively. Optimizing the salt loading during resin regeneration and providing more salt storage
capacities to allow delivery of full truck loads can significantly reduce the overall salt cost.
Electricity Cost. The electricity cost was tracked by comparing the monthly electrical bills before and
after IX system installation. For example, electricity bills at VA were approximately $850/month in 2006
and increased by 29% to $l,100/month in 2007. Thus, the annual increase was $3,000, or $0.028/1,000
gal. The electricity cost per 1,000 gal of water treated was $0.08 at FL. Electricity costs represented 13%
and 8% of the total O&M costs for the FL and VA systems, respectively.
62
-------
Labor Cost. The routine, non-demonstration related labor activities consumed only 10 to 30 min a day,
five days a week. The average weekly hours were 2.5 hr at FL and 3.3 hr at VA. The labor rate was
$21/hr for both sites. Labor costs per 1,000 gal of water treated were $0.03 and $0.05, accounting for 8 to
10% of the total O&M costs.
4.3
RO Demonstration System
A POE RO unit coupled with dual plumbing in the distribution system was demonstrated at the CE site.
This approach involved installing a parallel plumbing system dedicated to the potable water distribution
only. Because most water consumed at the school was for non-potable use (i.e., lavatory), only a portion
of raw water would need to be treated for potable use (i.e., kitchen sinks, drinking fountains, etc). As a
result, a smaller RO system with a separate distribution system was installed to meet the potable water
demand, thus reducing the capital investment and O&M costs.
4.3.1 RO System Design and Configuration. The RO system selected was a Crane
Environmental EPRO-1,200 system consisting of an RO unit, a calcite filter for pH adjustment, two 300-
gal atmospheric storage tanks, a re-pressurization system, and a post-chlorination system. Major
components of the RO unit included a 5-(im sediment filter, a !/2-horsepower (hp) booster pump, and two
2.5-in x 40-in thin-film composite RO membrane modules, as shown on Figure 4-2. The RO permeate
passed through the calcite filter to raise its pH levels to near neutral and then was stored in two 300-gal
1) Pressure gauges
2) RO membrane
3) Flow meters
4) Totalizer
5) TDS monitor
Figure 4-2. EPRO-1,200 RO Unit
63
-------
atmospheric storage tanks. The water from the storage tanks was re-pressurized by a 1-hp booster pump
before entering the potable distribution line. All major functions of the EPRO-1,200 RO unit were
automated and required only minimal operator oversight and intervention. Table 4-9 summarizes key
system design parameters of the treatment system.
Table 4-9. Design Specifications of EPRO-1,200 RO System
Parameter
Value
System Components
No. of Pre-filters
Pre-filter Nominal Pore Size ((im)
No. of RO Membrane Elements
RO Membrane Construction
Size of Membrane Elements
1
5
2
Thin film composite
2.5-inD x40-inH
Operating Specifications
Feed Flowrate (gpd)
Daily Permeate Production Rate (gpd)
Recovery (%)
Min. Rejection (%)
3,000
1,200
40
98
The RO system was rated for 1,200 gpd of permeate production with a 40% recovery (or 2.5:1, that is, for
every 2.5 gal of feed water, 1 gal of permeate water and 1.5 gal of reject water were produced). The
reject water was discharged into the existing septic system. Both permeate and reject water lines were
equipped with flow meters and totalizers, pressure gauges, and sample taps for monitoring purposes.
4.3.2 RO System Capital Investment Cost. The capital investment cost for the RO system was
$20,542, including $8,600 for the dual plumbing and $11,942 forthe EPRO-1,200 RO unit. The dual
plumbing installation cost included $2,650 for plumbing materials and $5,950 for the labor to convert the
existing plumbing into a duplex distribution system. The cost of the EPRO-1,200 RO unit included
$8,471 for equipment and parts, $300 for shipping, and $3,171 for installation.
The capital investment cost of $20,542 was normalized to the system's rated capacity of 1,200 gpd of
permeate, which results in $17.12/gpd of design capacity (see Table 4-10). The unit capital cost based on
the average production rate was higher than that based on the maximum capacity. The ratio of the
average production to the maximum capacity, expressed as utilization rate, was 25%.
Table 4-10. RO System Capital Investment Cost
Site
ID
CE
Design
Flow
rate
(gpd)
1,200
Total
Capital
Costs
($)
Normalized
Capital
Cost
($/gpd)
$20,542 $17.12
Annualized
Capital
Cost(a)
($/yr)
$1,939
Unit Cost
($1,000 gal
of water)
Design00
$4.43
Average
17.80
Utilization
Rate(c)
(%)
25%
(a) Obtained by applying a CRF of 0.09439 (based on 7% interest rate and 20-year return
period) to total capital cost.
(b) System's maximum capacity at design flowrate, operating 24 hr a day, 365 days a year.
(c) Ratio of system's average annual production rate to its maximum capacity at design
flowrate.
64
-------
4.3.3 RO System O&M Cost. The O&M cost included system repairs, electricity consumption,
and labor to operate the system. Regularly scheduled maintenance activities involved replacing sediment
filters on a monthly basis or when the differential pressure was greater than 10% and replenishing calcite
in the calcite filter as it became depleted. Neither was required during the performance evaluation study.
The cost to diagnose and install a faulty RO motor and pump assembly was $351. Annual electricity
consumption was estimated to be 5,078 kwh and cost $376. Routine labor activities consumed 10 min per
day to visually inspect the system and record operational parameters, which translated into $666/yr. The
total annual O&M cost was estimated to be $1,404, or $12.89/1,000 gal of permeate water produced.
4.4 POU RO Demonstration Units
4.4.1 POU RO Unit Design and Configuration. One POU RO unit was demonstrated at each of
nine participating residences for arsenic, nitrate, and uranium removal from source water. Softening of
source water was performed as pretreatment to meet feed water quality requirements for the RO units.
Six POE softeners (three homes had existing softeners) and nine POU RO units were provided by
Kinetico. Each POU RO unit consisted of a 20-(im pre-filter, an RO module with a 1.7-in x 11-in thin
film composite, semi-permeable membrane element, a 3-gal storage tank, and a MACguard post-filter.
The RO units were capable of producing up to 35.5 gpd of permeate water and had a feed water to
permeate water ratio of 2.7 to 1, a 37% recovery rating. The RO units automatically shut down
production after 500 gal of permeate water had been processed and resumed operation only after
replacement of spent pre- and post-filters.
Each system was equipped with a PureMometer Filter Life Indicator to alert users for the remaining
capacity of the filter cartridge. Further, a TDS monitor installed at the kitchen tap measured TDS levels
in treated water. A green light on the monitor indicated that a proper amount of permeate water was
generated and a yellow light indicated that it was not. The RO Plus Deluxe unit has been tested and listed
under NSF Standard 58. Table 4-11 summarizes key performance specifications for the RO Plus Deluxe
unit. Figure 4-3 shows a photograph of the under-the-sink RO unit.
4.4.2 POU RO Costs. The capital investment cost for purchasing and installing six water softeners
and nine RO units was $31,877.50. The equipment cost was $21,732.50 (or 68% of the total capital
investment costs), which included the cost for nine RO units, six water softeners, initial salt fill,
additional sample tap and a water meter, and freight. The installation cost was $10,145 (or 32% of the
total capital investment costs). The lump-sum cost was broken down for individual units. Each water
softener cost $2,395, including $1,585 for equipment and $810 for installation. Each RO unit cost
$1,220, including $1,025 for equipment and $195 for installation.
The O&M cost consisted of salt usage, pre- and post-filter replacement, RO element replacement, and
maintenance. The yearly service contract with the vendor for salt supply was $115 per year. Pre- and
post-cartridge filter replacement at 500 gal of treated water was $86.50. Five out of the nine residences
used 500 gal of treated water during the performance evaluation period. For these five residences, the
one-year O&M cost included $115 for salt supply and $86.50 for filter replacement, totaling $201.50 or
$17 per month. The systems were under warranty for one year; therefore, no maintenance cost was
incurred during the study period. Neither electricity nor labor cost was incurred because the water
softener and the RO unit did not consume electricity and did not require a certified operator.
65
-------
Table 4-11. Kinetico RO Plus Deluxe Unit Performance Specifications
Parameter
Value
System Components
Pre-treatment
No. of RO Membrane Elements
RO Membrane Construction
Membrane Element Size (in)
No. of Post-filters
Permeate Flush
Element Configuration
System Shutoff Control
System Shutdown Volume (gal)
System Controller
Storage Tank
One, 20-um pre-filter
1
Thin film composite
1.7-inDxll-inH
1
Internal Permeate Reservoir
Single
Hydraulic
500
Hydraulic
One, 8-inD x 17-inH (3 gal)
Operating Specifications
Maximum Daily Production (gpd)
Daily Production (gpd)
Discharge Water (or Feed Water)/
Product Water Ratio
Normal Operating Pressure (psi)
75
35.5
2.7 to 1
60
Source: Kinetico.
Figure 4-3. Under-the-Sink RO Plus Deluxe Unit
66
-------
4.5
POU AM Demonstration Units
4.5.1 POU AM Cartridge Design and Configuration. Eight Kinetico POU AM units were
installed either under a sink or inside a drinking water fountain in eight different school buildings at the
KF site, but only three were monitored for their performance. Each POU unit used a single cartridge to
house 600 mL of ARM 200 media for arsenic removal. A shut-off assembly and an indicator on the
outside of the filter head were used to measure and show the relative remaining cartridge capacity, based
on a maximum capacity of 500 gal. When 500 gal of water was processed, the shut-off assembly was
completely closed, preventing any more water from passing through the cartridge. About 11 months into
the performance evaluation study, the school began to install 40 new AdEdge E33 POU units and to
replace the eight Kinetico units with AdEdge units. Each AdEdge POU unit consisted of E33 media in a
polypropylene housing. The approximate flowrate with a system inlet pressure of 60 psi was 1 gpm. The
working pressure ranged from 20 to 125 psi. The unit had a height of 13 in and a diameter of 6.75 in.
Table 4-12 presents the design specifications of Kinetico and AdEdge POU units. Figure 4-4 shows
photographs of the POU units installed under a sink and inside a drinking fountain.
Table 4-12. Design Specifications of Kinetico and AdEdge POU AM Cartridges
Parameter
Housing Material
Cartridge Dimensions (mm)
Housing Dimensions
Height
Width
Diameter
Unit Weight (Ib)
Media Type
Media Volume (mL)
Inlet Connection
Outlet Connection
Paniculate Retention (^m)
Water Pressure (psi)
Flowrate (gpm)
Treatment Capacity (gal)
Kinetico POU Unit
Polypropylene
54 x 265
(Slightly tapered)
-
425 mm
150mm
100mm
11
ARM 200
600
!/4-in Female NPT
!/4-in Female NPT
5.0
20-120
0.7-1.0
490
AdEdge POU Unit
Polypropylene
—
-
13 in
-
6.75 in
4
E33
-
3/8in
!/4in
0.5
30-125
1.0@60psi
-
4.5.2 POU AM Cartridge Costs. The cost of purchasing eight Kinetico POU ARM 200
cartridges was $1,216, or $152 per unit. The cost of purchasing 48 AdEdge POU E33 cartridges was
$9,120, or $215 per unit (these replacement cartridges were purchased by the school). Although the E33
cartridge is 40% higher than the ARM 200 cartridge, the E33 media life was almost three times as long as
ARM 200. For example, one E33 cartridge treated up to 3,000 gal of water to reach 8 |Jg/L of arsenic in
the effluent while the ARM 200 cartridge treated up to 1,000 gal of water to reach 6 |Jg/L of arsenic in the
effluent.
The O&M cost of the POU AM unit consisted of replacing pre- and post-filter as well as AM media.
Neither electricity nor labor cost was incurred because the cartridge did not consume electricity and did
not require a certified operator.
67
-------
Figure 4-4. POU AM Units Installed Under a Sink (top) and Inside a Drinking
Water Fountain (bottom)
68
-------
5.0 COST SUMMARY
This section summarizes capital investment and O&M costs of the AM, IR/CF, and IX systems. The cost
data were divided into two groups with one for systems having design flowrates smaller than 100 gpm
(including both NTNCWS and CWS) and the other for systems equal to or larger than 100 gpm. The
group of smaller systems (<100 gpm) comprised 17 AM and six IR/CF (including two IR/AM) systems.
The group of larger systems (>100 gpm) comprised 11 AM, 12 IR/CF (including two IR/AM), and two
IX systems. The range and average of cost data for the same technology in each group were calculated to
allow for comparison of those within and between the groups. Because many factors can affect the costs
of technologies and the number of systems in each group varies, the results of this cost analysis are valid
only for the specific cost data collected from this study; any conclusions drawn from the cost comparisons
should only be used as a reference.
5.1 Total Capital Investment Costs of Treatment Technologies
Capital investment costs of the full-scale arsenic removal systems/POU units demonstrated under EPA
Rounds 1, 2, and 2a demonstration projects totaled $8,552,428. Table 5-1 summarizes total capital
investment costs for the AM, IR/CF, and IX systems demonstrated. The cost data are plotted in Figures
5-1 and 5-2 for smaller systems (<100 gpm) and in Figures 5-3 and 5-4 for larger (>100 gpm) systems.
The four IR/AM systems were plotted separately on these figures, but were considered as IR systems in
the cost analysis in Table 5-1.
Total capital investment costs of the 17 smaller AM systems scattered widely, ranging from $14,000 to
$228,300. The variations observed were caused by the factors discussed in Section 2. The costs of the
six smaller IR/CF systems also varied, but to a lesser extent, from $55,423 to $161,560. Normalized
costs ranged from $636 to $6,171 per gpm (or $0.44 to $4.29 per gpd) for the smaller AM systems and
$1,344 to $3,177 per gpm (or $0.93 to $2.21 per gpd) for the smaller IR/CF systems. Unit capital costs
per 1,000 gal of water treated ranged from $0.11 to $1.11 for the smaller AM systems and $0.24 to $0.57
for the smaller IR/CF systems. Average values of the normalized and unit costs for the AM systems were
6% and 8%, respectively, higher than those for the IR/CF systems. However, individual data points in
Figures 5-1 and 5-3 do not exhibit any clear trend whether AM or IR/CF is more expensive. If the highest
cost associated with the 37-gpm AM system (that was equipped with a pH control system, a backwash
wastewater recycling system, and excessive instrumentation and controls) was removed from the data set,
average values of the normalized and unit costs for the AM technology would be lower than those of the
IR/CF technology. Therefore, the capital investment costs of the smaller AM and IR/CF systems did not
differ significantly from each other.
For larger treatment systems (>100 gpm), total capital investment costs ranged from $74,840 to $305,000
forthe 11 AM systems, $216,876 to 427,407 forthe 12 IR/CF systems, and $286,388 to $395,434 for the
two IX systems. Normalized costs ranged from $477 to $1,492 per gpm (or $0.33 to $1.04 per gpd) for
the AM systems, $555 to $1,932 per gpm (or $0.39 to $1.34 per gpd) forthe IR/CF systems, and $732 to
$1,146 per gpm (or $0.51 to $0.80 per gpd) forthe IX systems. Unit capital costs per 1,000 gal of water
treated ranged from $0.09 to $0.27 forthe AM systems, $0.10 to $0.35 forthe IR/CF systems, and $0.13
to $0.21 for the IX systems. As shown in Figure 5-4, capital investment costs per gpd generally
decreased with increasing system sizes for all technology types. Average values of the normalized and
unit costs for the AM systems were 25% and 26%, respectively, lower than those for the IR/CF systems.
The trendlines in Figures 5-2 and 5-4 also clearly indicate that the cost of IR/CF is higher than that of
AM. The costs of the two IX systems appear to fit well with those for IR/CF. Therefore, IR/CF and IX
are generally more expensive than AM for systems larger than 100 gpm. Because seven out of the 12
IR/CF systems and both IX systems were supplied by one vendor, it is possible that the cost data were
skewed by this vendor's pricing structure. The larger systems have lower normalized and unit costs than
the smaller systems, reflecting the scale of economy.
69
-------
Table 5-1. Summary of Total Capital Investment Costs
Treatment
Technology
No. of
Systems
Range/
Average
Design
Flow
rate
(gpm)
Total
Capital
Cost
($)
Normalized
Capital
Cost
($/gpm)
Normalized
Capital
Cost
($/gpd)
Unit
Cost
($/kgal)
Equipment
Site
Engineering
Installation
(% of Total
Capital Invest Costs)
Systems <100 gpm
AM
IR/CF
17
6W
Range
Average
Range
Average
10-
75
20-
96
14,000-
228,309
55,423-
161,560
636-
6,171
2,248
1,344-
3,177
2,128
0.44-
4.29
1.56
0.93-
2.21
1.48
0.11-
1.11
0.41
0.24-
0.57
0.38
38-75
65
33-80
55
10-40
16
6-34
18
12-34
19
14-33
26
Systems > 100 gpm
AM
IR/CF
IX
11
12W
2
Range
Average
Range
Average
Range
Average
100-
640
140-
770
250-
540
74,840-
305,000
216,876-
427,407
286,388-
395,434
477-
1,492
806
555-
1,932
1,069
732-
1,146
939
0.33-
1.04
0.56
0.39-
1.34
0.74
0.51-
0.80
0.66
0.09-
0.27
0.14
0.10-
0.35
0.19
0.13-
0.21
0.17
61-82
72
55-75
64
61-66
63
4-17
12
4-17
12
12-13
12
13-25
16
15-36
24
22-27
24
(a) Including two AM systems with IR pretreatment.
-------
$250,000
$200,000
~ «
o $150,000
| « «
^ $100,000
S mm
° m «AM
$50,000 « " IR/CF
•
•P B IR+AM
$0 ,
0 20 40 60 80 100
Design Flowrate (gpm)
Figure 5-1. Total Capital Investment Costs of Smaller AM and IR/CF Systems (<100 gpm)
$500,000
$400,000 a
o $300,000
. _,,-
0 $200,000 • „•--''' B AM
TO • ,.'' " MIVI
$100,000 IR+AM
•
D IX
$0 •
100 200 300 400 500 600 700
Design Flowrate (gpm)
Figure 5-2. Total Capital Investment Costs of Larger AM, IR/CF, and IX Systems (>100 gpm)
71
-------
$4.50 -
$4.00 -
S $3.50 -
T3
Q.
ro $3.00
I
tJ $2.50 -
o
o
g $2.00 -
'a.
0 $1.50 -
S
£ $1.00 -
$0.50 -
$0.00 -
• "AN
*IR/
AIR-
^
•
A\ A •
• •
* : •
i i i i
0 20 40 60 80
Design Flowrate (gpm)
/i
CF
hAM
^
1
100
Figure 5-3. Total Capital Investment Costs per gpd of Design Capacity (<100 gpm)
$1.60
$0.20
$0.00
100 200 300 400 500
Design Flowrate (gpm)
600
700
Figure 5-4. Total Capital Investment Cost per gpd of Design Capacity (>100 gpm)
72
-------
Equipment, site engineering, and installation and startup costs are plotted as a percentage of the respective
total capital investment cost in Figure 5-5 through 5-7. In general, equipment costs accounted for higher
percentages of total capital investment costs for larger systems than for smaller systems. For example,
larger AM and IR/CF system equipment costs accounted for 72% and 64% (on average) of respective
total capital investment costs, whereas smaller system equipment costs accounted for 65% and 55% of
respective total capital investment costs. Regardless of system sizes, AM system equipment costs
accounted for higher percentages of total cost than IR/CF system equipment costs.
Site engineering and installation/startup costs were primarily labor costs. Smaller system site engineering
costs accounted for, on average, 16% and 18% of total capital investment costs for AM and IR/CF,
respectively. These percentage points were higher than the 12% found for larger systems for all three
technology types. Installation and startup costs of IR/CF and IX accounted for higher percentage points
than those of AM, regardless of system sizes. For example, IR/CF system installation/startup costs
accounted for 26% (for smaller systems) and 24% (for larger systems) of total capital investment costs,
whereas AM system installation/startup costs accounted for only 19% and 16% for smaller and larger AM
systems, respectively. The data suggest that the AM systems took less time and were easier to install than
the IR/CF systems. The IR/CF systems frequently include contact tanks, iron addition systems, and
ancillary equipment and controls that require more efforts to install and be field-tested and adjusted. The
same vendor who provided seven of the 12 larger IR/CF systems also might be a factor for the higher
costs observed. Because the larger IR/CF systems had higher total capital investment costs than the AM
systems, the higher percentages of the installation/startup costs also indicated higher costs.
100%
90%
Q 80%
X 70%
*j _
O o 60% -
*- """
S° 50% -
E a)
Q. O)
'5 8 40%
UJ 8
oi 30%
Q.
« 20%
10%
0%
0 100 200 300 400 500
Design Flowrate (gpm)
600
700
Figure 5-5. Equipment Costs as a Percentage of Total Capital Investment Costs
73
-------
50%
I 40% ^
'a.
TO
to _
o TO
O -g 30%
0)1-
~o
0)
D.
20%
10%
0%
# •
0 100 200 300 400 500 600 700
Design Flowrate (gpm)
Figure 5-6. Engineering Costs as a Percentage of Total Capital Investment Costs
50%
ra 40%
=" o 30%
"c 0)
If 20%
jS o
w 5
£ a.
<§ 10%
0%
c
• AM
* IR/CF
AIR+AM
^+ DIX
•
Jf' * ..'%. -
*•":•'-
) 100 200 300 400 500 600 700
Design Flowrate (gpm)
Figure 5-7. Installation/Startup Costs as a Percentage of Total Capital Investment Costs
74
-------
5.2
O&M Cost of Treatment Technologies
Table 5-2 summarizes the O&M costs associated with AM, IR/CF, and IX along with cost breakdowns.
The cost data also are plotted in Figures 5-8 and 5-9 for smaller (<100 gpm) and larger (>100 gpm)
systems, respectively. The four IR/AM systems were plotted separately on these figures, but were
considered as IR systems in the cost analysis in Table 5-2 because media replacement did not occur
during the study period.
Table 5-2. Summary of O&M Costs
Treatment
Technology
No. of
Systems
Range/
Average
Design
Flow
rate
(gpm)
Total
O&M
Costs
Media
Replacement
Cost
Chemical
Cost
Electricity
Cost
Labor
Cost
($/l,000 gal of Water Treated)
Systems with <100 gpm Design Flowrates
AM
IR/CF
14W
6(b)
Range
Average
Range
Average
10-75
20-96
0.86-22.88
6.47
0.26-2.90
1.39
0.58-22.05
5.58
NA
NA
0.00-0.61
0.08
0.00-0.37
0.14
0.00-0.16
0.03
0.00-0.39
0.10
0.03-3.1
0.78
0.11-2.57
1.15
Systems with > 100 gpm Design Flowrates
AM
IR/CF
IX
5
12(b)
2
Range
Average
Range
Average
Range
Average
150-350
140-770
250-540
0.61-5.69
1.76
0.07-0.55
0.28
0.35-0.62
0.49
0.3-5.51
1.57
NA
NA
NA
NA
0.00-0.03
0.01
0.00-0.17
0.04
0.29-0.49
0.39
0.00-0.05
0.01
0.00-0.16
0.05
0.03-0.08
0.06
0.05-0.25
0.17
0.04-0.52
0.19
0.03-0.05
0.04
(a) Two systems experienced multiple media change-outs.
(b) Including two AM systems with IR pretreatment.
NA = not applicable
The data in Table 5-2 and Figures 5-8 and 5-9 indicate that the AM systems had higher O&M costs than
the IR/CF and IX systems, regardless of system sizes. The higher costs observed were attributed
primarily to media replacement costs, which accounted for 86% and 89% of total O&M costs for the
smaller and larger systems, respectively, based on the average values presented in Table 5-2. Media
replacement costs were affected by the media performance and media unit prices as discussed in Section
2.5.1. For the four E33 systems achieving a media life of 38,000 BV and higher, media replacement costs
ranged from $0.30 to $0.66 per 1,000 gal of water treated and the total O&M costs ranged from $0.61 to
$0.86 per 1,000 gal of water treated. Methods to extend the media life through caustic regeneration have
shown promises to reduce the O&M cost of E33 systems (Chen and Wang, 2008; 2009; Sorg et al., 2010).
The O&M costs for the IR/CF and IX systems reported in this study did not include treatment and/or
disposal costs of residuals generated such as backwash wastewater and spent brine/rinse water. Residual
disposal costs could be a significant part of the O&M costs and play an important role in the technology
selection.
Chemical cost was a major O&M cost for the IX process that used salt for resin regeneration. Chemical
costs associated with pH control for AM, iron salts for IR/CF, and/or pre-oxidation of raw water for AM
and IR/CF was insignificant.
75
-------
$25.00
$20.00 -
ra
O)
§ $15.00
^
*J
,§ $10.00
I
$5.00
$0.00
20 40 60
Design Flowrate (gpm)
80
100
Figure 5-8. Smaller System (<100 gpm) Total O&M Costs per 1,000 gal of Water Treated
$6.00 -
$5.00 -
2-
™ $4.00 -
o
^ $3.00 -
to
o
O
S $2.00 -
O
$1.00 -
$0.00 -
1C
B BAI\
*IR/
AIR-
niX
B 1 •' •
)0 200 300 400 500 600
Design Flowrate (gpm)
1
CF
^AM
700
Figure 5-9. Larger System (>100 gpm) Total O&M Costs per 1,000 gal of Water Treated
76
-------
Incremental electricity cost was insignificant for AM, IR/CF, and IX technologies because these
technologies did not require electricity to push water through treatment systems like membrane
technologies. Electricity was consumed to overcome any headless across treatment vessels and to power
system controls and/or chemical feed pumps.
It was difficult to quantify and compare labor cost among different technologies because labor rates
varied geographically and labor hours were subject to specific circumstances at different sites. Average
labor rates were similar for all three technologies, i.e., $22.4/hr for AM (Section 2.4.4), $22.6/hr for
IR/CF (Section 3.5.3), and $21/hr for IX (Section 4.2.3). These labor rates might be lower than those in
certain regions of the country, such as California. Average weekly labor hours required to operate and
maintain the treatment systems were 1.8 hr for AM (Section 2.4.4), 3.4 hr for IR/CF (Section 3.5.3), and
2.5 hr for IX (Section 4.2.3). The data supported the general notion that an AM system was easier to
operate and maintain compared to an IR/CF and an IX system. As shown in Table 5-2, average labor
costs per 1,000 gal of water treated were $0.78 and $1.15 for smaller AM and IR/CF systems,
respectively, and $0.17, $0.19, and $0.04 for larger AM, IR/CF, and IX systems respectively. The higher
labor costs for smaller systems were attributed to the lower water production rates associated with smaller
systems.
77
-------
6.0 REFERENCES
Chen, A.S.C. and L. Wang. 2008. "Regeneration of Arsenic Removal Adsorptive Media." 5th Annual
EPA Drinking Water Workshop: Treatment and Distribution System Compliance Challenges.
Cincinnati, OH, August 5-7.
Chen, A.S.C. and L. Wang. 2009. "Regeneration of a Full-Scale Adsorptive Media Arsenic Treatment
System." 6th Annual EPA Drinking Water Workshop, Cincinnati, OH, August 4-6.
Cornwell, D.A. and O.K. Roth. 2011. Water Treatment Plant Residuals Management. Chapter 22 of
Water Quality & Treatment: A Handbook on Drinking Water, sixth edition, J.K. Edzwald, ed.,
American Water Works Association, Denver, CO. McGraw Hill, New York.
EPA. 2000. Regulations on the Disposal of Arsenic Residuals from Drinking Water Treatment Plants.
EPA/600/R-00/025. U.S. Environmental Protection Agency, National Risk Management
Research Laboratory, Cincinnati, OH.
EPA. 2001. National Primary Drinking Water Regulations: Arsenic and Clarifications to Compliance
and New Source Contaminants Monitoring. Federal Register, 40 CFR Parts 9, 141, and 142.
EPA. 2003. Minor Clarification of the National Primary Drinking Water Regulation for Arsenic.
Federal Register, 40 CFR Part 141.
Gulledge, J.H. and J.T. O'Conner. 1973. "Removal of Arsenic (V) from Water by Adsorption on
Aluminum and Ferric Hydroxides," J. AWWA, 65:8:548.
Hering, J.G., P-Y, Chen, J.A. Wilkie, M. Elimelech, and S. Liang, 1996. "Arsenic Removal by Ferric
Chloride;'J. AWWA, 88:155.
Sorg, T.J. and G.S. Logsdon. 1978. "Treatment Technology to Meet the Interim Primary Drinking Water
Regulations for Inorganics: Part 2," J. AWWA, 70:7.
Sorg, T.J. 1993. "Removal of Arsenic From Drinking Water by Conventional Treatment Methods,"
Proceedings of the 1993 AWWA Water Quality Technology Conference.
Sorg, T.J. 2002. "Iron Treatment for Arsenic Removal Neglected," Opflow, AWWA, 28:11:15.
Sorg, T.J., A.S.C. Chen, L. Wang, and M. Wright. 2010. "Regeneration of Exhausted Arsenic Adsorptive
Media of a Full Scale Treatment System," Inorganic Contaminants Workshop, Denver, CO.
February 28.
78
------- |