The US Environmental Protection Agency (EPA) conducted an Arsenic Demonstration Program whereby the Agency purchased, installed, and evaluated the performance and cost of 50 small water arsenic removal treatment systems in the USA. A major goal of the program was to collect high-quality cost data (capital and operational and maintenance (O&M)) from the long-term operation (1–4 years) of these systems. The technologies consisted of adsorptive media (AM), iron removal (IR), coagulation/filtration (C/F), ion exchange (IX), reverse osmosis, and point-of-use devices, which reduced the arsenic to less than the EPA maximum contaminant level of 10 μg/L. This paper presents the capital and O&M cost of 48 treatment systems ranging in size from 10 to 770 gal/min (gpm) (38–2,915 L/min). The capital cost of the systems ranged from $477 to$6,171 per gpm ($126–$1,632 per L/min) of design flow and the O&M cost from $0.07 to$22.88 per 1,000 gal ($0.02–$6.05 per 1,000 L) of treated water. AM had a lower capital, but a higher O&M cost than IR, C/F, and IX. The media replacement cost for the AM systems averaged 80% of the O&M cost and was the main cause of the higher O&M cost of the AM systems.

## INTRODUCTION

In 2001, the US Environmental Protection Agency (EPA) revised the arsenic maximum contaminant level (MCL) in drinking water from 0.050 mg/L (50 μg/L) to 0.010 mg/L (10 μg/L) and required public water systems to comply with the new standard by 23 January 2006 (EPA 2001, 2003a). This change in the MCL was estimated to impact approximately 5,000 small, public water systems serving populations of ≤3,300.

Although many small systems have installed treatment systems to achieve compliance with the revised arsenic standard, a significant number still face the challenge of installing a water treatment system to achieve arsenic compliance. Most of the existing, non-compliant small water systems have very limited resources and, therefore, finding a low cost, affordable technology (capital and operation and maintenance (O&M)) is of major interest to them. Although many small arsenic treatment systems have been operating for the past 6–8 years, very few data on the capital and operating costs of these systems have been collected and reported in the literature (Mohan & Pittman 2007; Hilkert et al. 2010). Treatment capital and O&M cost data from operated systems can be extremely useful to non-compliant, small water systems in selecting an affordable arsenic removal technology.

In 2002, the EPA initiated an Arsenic Treatment Technology Demonstration Program (ADP) with the major objective of providing information on cost-effective, small system treatment technologies. The program was funded by the EPA to reduce the financial risk of the small systems with their limited resources having to select and purchase a new, unproven technology with little or no performance or cost information available to them. A key goal of the program was to collect high-quality performance and cost data from the long-term operation (1–4 years) of full-scale systems under the real world operating conditions. To achieve this goal, the EPA purchased, installed, and evaluated the performance and cost of the treatment systems that were selected through a cooperative decision process involving the water systems, the state drinking water agencies, and the EPA (Wang et al. 2005).

Between 2003 and 2011, the EPA initially partnered with 50 small water systems in 27 different states to conduct performance evaluation studies on over 15 different, commercially available, full-scale treatment technologies operating under a variety of conditions. One utility withdrew from the program and was replaced by a single residential home, point-of-entry (POE) treatment study which reduced state participation to 26. The source waters were all ground water and the arsenic concentration in the source water ranged from 13 to 150 μg/L.

Five basic technologies were selected for study and all were fundamentally proven technologies (Chen et al. 1999; EPA 2003b; Amy et al. 2005; Mohan & Pittman 2007; Giles et al. 2011) for small systems. The treatment systems include: adsorptive media (AM), iron removal (IR), coagulation/filtration (C/F), ion exchange (IX), and reverse osmosis (RO). At four sites, a combination of two technologies, iron removal followed by adsorptive media (IR/AM), was installed. All of the treatment systems reduced and maintained the arsenic concentration in the treated water to below the EPA MCL of 10 μg/L.

A single or combination treatment approach was selected for demonstration at each site except for two. System modification was evaluated at a North Dakota site where an existing IR system was reducing the arsenic from around 130 to 30 μg/L (Condit et al. 2006). At an Oregon site (Oregon Institute of Technology), three AM systems were installed; one at the entry point (POE) of three different campus buildings as an alternative to central wellhead treatment (Chen et al. 2011). A map of the locations of the 50 demonstration sites is shown in Figure 1. The ADP ended on 30 October 2011 and the results from each demonstration project are documented in individual performance evaluation reports available online at www.epa.gov/nrmrl/wswrd/dw/arsenic/. A detailed report of all the cost data (capital and O&M) from these individual project reports is also available at the same website (Wang & Chen 2011).

Figure 1

Location of the 50 arsenic treatment technology demonstration program sites (states (shaded green/gray) with arsenic demonstration program site(s); o = site location). The full color version of this figure is available online at http://www.iwaponline.com/jws/toc.htm.

Figure 1

Location of the 50 arsenic treatment technology demonstration program sites (states (shaded green/gray) with arsenic demonstration program site(s); o = site location). The full color version of this figure is available online at http://www.iwaponline.com/jws/toc.htm.

This paper summarizes the capital and O&M costs of arsenic removal treatment systems of the ADP treating flows of 10–770 gpm (all gpm units can be converted to L/min by using a multiplier of 3.78). The capital and O&M information presented are the actual cost data obtained during the ADP performance evaluation studies. The treatment systems were purchased between 2003 and 2008 and the O&M cost data were collected during a 7 year period from 2004 through 2010. Normalizing all of the capital and O&M cost data collected during the 7 year period to 1 year would be a difficult task because of varying operational periods and the variation in designs of the individual treatment systems. Thus, the data presented are not intended to be used for precise estimates of current day or future costs of similar arsenic removal systems. The operating and capital costs presented are likely lower than current day costs due to inflation, with the current consumer price index at approximately 15% higher than the average index of the 2004–2008 year period. The cost data do, however, provide rough cost estimates and are useful for comparing the costs (capital and operating) of the different technologies and designs.

## TREATMENT TECHNOLOGIES

The most frequently selected technology was the AM process that consisted of 28 systems at 26 sites in 15 different states. The 28 systems included nine different AM products with E33 (iron based media) being the most frequently used media (13 systems).

The second most frequently selected technology was IR (10 systems). Within a short operational time period, two of these systems required additional iron (i.e. ferric chloride) in order to reduce the arsenic to below 10 μg/L. Four different filtration media products were used with Macrolite® being the most commonly used media (seven systems).

Three of the four sites that installed an IR/AM combination system selected it because of the potential of iron fouling of the AM. The IR process not only helps protect the AM system from iron fouling but also removes some arsenic, thereby extending the life of the AM. The fourth dual system was selected for its capability to reduce the arsenic to very low levels. At this site, the IR system lowered the source water arsenic from 23 μg/L to slightly below 10 μg/L and the AM system further reduced the arsenic to around 1–2 μg/L at start-up.

The remaining technologies consisted of C/F (4), IX (2), and RO (1). The C/F and IR processes are very similar with both technologies using iron to remove arsenic. C/F requires the addition of an iron coagulant while IR takes advantage of natural iron in source water. Some, but not all, of the C/F and IR systems included very simple contact tanks (no flocculation). IX was selected by two water systems because their source waters contained a co-contaminant, nitrate. IX is effective in removing both arsenic (arsenate) and nitrate. The central RO system was also selected because of a co-contaminant, antimony.

Excluding RO technology, the other five technologies all use metal/fiberglass tanks to house an adsorptive media, a filtration media or an IX resin depending on the technology (Figure 2). This common design feature is a major cost component of each treatment system. The number, size, and layout of these tanks are a function of the design flow, water quality, and the arsenic removal mechanism. Ancillary equipment, such as pH adjustment, chemical feed (ferric chloride and liquid chlorine) systems, chemical contact tanks, and backwash holding tanks are additional cost items that vary with the technology and the need for additional treatment.

Figure 2

Basic design of five arsenic removal treatment systems.

Figure 2

Basic design of five arsenic removal treatment systems.

Table 1 provides source water quality parameters (As, Fe, and pH) and system design flows for the six treatment technology categories for the 49 systems in the study. All three water quality parameters had some impact on the selection and design of the treatment system for each site. Iron was particularly significant because of its affinity for arsenic and its potential for media fouling. When the iron concentration of the source water was above the iron secondary maximum contaminant level of 0.3 mg/L, an IR or IR/AM system was selected except for one system (AM only) (Figure 3). AM system was selected for source water with 2.3 mg/L of iron because the site had an existing water softener that reduced the iron to below 0.3 mg/L. When the iron concentration of the source water was below 0.3 mg/L, other factors came into play in the technology selection process which led the decision to select AM, C/F, IX, or RO.

Table 1

Summary of treatment system information and source water quality

Treatment systems AM IR IR + AM C/F IX RO
Number of systems 28 10
Design flow rate, L/m 38–2,423 76–2,915 95–946 946–2,083 946–2,040 a
Arsenic, μg/L 13–88 14–39 15–42 18–56 17–44 18
Iron, mg/L <0.025–2.3 2.5–0.43 0.33–1.6 <0.025–0.16 <0.025 0.025
pH, units 6.9–9.6 6.8–7.9 7.2–7.9 7.5–8.3 7.4–8.0 7.9
Treatment systems AM IR IR + AM C/F IX RO
Number of systems 28 10
Design flow rate, L/m 38–2,423 76–2,915 95–946 946–2,083 946–2,040 a
Arsenic, μg/L 13–88 14–39 15–42 18–56 17–44 18
Iron, mg/L <0.025–2.3 2.5–0.43 0.33–1.6 <0.025–0.16 <0.025 0.025
pH, units 6.9–9.6 6.8–7.9 7.2–7.9 7.5–8.3 7.4–8.0 7.9

aRO feed flow rate of 3,000 gpd (11,356 L/d) with permeate production rate of 1,500 gpd (5,678 L/d)).

Figure 3

Relationship of treatment system type to iron concentration of source water.

Figure 3

Relationship of treatment system type to iron concentration of source water.

## COST DATA METHODOLOGY

### Capital costs

A request for proposal (RFP) was solicited for each demonstration project that required the proposer to describe the treatment system and provide an estimate of capital cost and operating cost of the system. The number of proposals received for each project ranged from one to eight. It was not uncommon for the highest capital cost estimate to be as much as 400% more than the lowest cost estimate. Although the estimated cost was a significant factor considered in the proposal selection process, it was not the only factor. Other factors considered were the estimated performance, residuals, ease and simplicity of operation, footprint, pre- and post-treatment requirements, and adaptability to site conditions. Because the ADP was limited to small system participation, the majority of the technical proposals were received from equipment vendors proposing pre-built, skid-mounted systems.

After a project proposal was selected, the equipment vendor was required to submit a more detailed cost proposal using an ADP template. This template required a breakdown of the total cost into three major categories: equipment, engineering, and installation. These three categories are the costs that make up the capital cost estimates used for comparison. In most cases, the vendors provided even greater breakdowns of the three categories. Included under the equipment cost category were items such as the skid-mounted system (tanks, piping, valves, etc.), adsorptive/filtration/resin media, miscellaneous materials and supplies and freight, user's manual, and vendor's labor. Also included were costs for chemical feed systems, such as pH control, pre-chlorination, and/or iron addition. The cost of backwash recycle equipment was provided when backwash recycle was included in system design.

The engineering category included the cost for the design, layout and footprint of the system within the building, design of the piping connections up to the distribution tie-in points in the building, and the design of the electrical connection and conduit plan. The engineering cost also included the cost for the submission of the engineering plans to the state regulators for permit review and approval. In most cases, the vendors contacted and hired (subcontractor) a local engineering firm recommended by the water system who had experience with the state permit process. The installation cost included the costs for the equipment and labor to unload and install the system, the piping tie-ins and electrical work, the loading and backwashing of the media, system shakedown and start-up, and operator's training.

The cost of any land acquisition, building construction, site preparation, storage facilities, and residual handling (if required) was outside of the scope of this program and funded separately by the host sites; thus, these items are not included in capital cost.

### O&M cost

O&M costs were categorized into media replacement (AM systems only), chemicals, electricity, and labor. These data were collected by the operator using standardized data collection forms designed for the ADP. The cost data were collected primarily from the first year of system operation except for the media replacement costs, most of which were not incurred until 14–45 months into operation (depending on the media life).

Of the 28 AM systems, only 15 had to replace their exhausted media during the study period (paid for by the ADP); thus, complete O&M cost was obtainable for only these 15 systems. Two of the systems experienced multiple media change out with different types of media. For AM systems, the media replacement cost was broken down into media, labor, and other related costs, such as media analysis (toxicity characteristic leaching procedure), media disposal, and freight. The labor cost for AM replacement was the cost incurred by a contractor who provided the media replacement services.

The cost of all chemicals associated with the new treatment system, such as pH adjustment and pre-oxidation chemicals, ferric chloride, and IX brine, was paid for by the EPA under the ADP agreement with the water system. Any chemicals associated with any pre-existing system operation, such as chlorination/disinfection, softeners, fluoridation, residual treatment, etc., were paid by the host sites and not included in the O&M cost data.

The electricity cost was tracked by comparing monthly electrical bills (paid by the host sites) before and after 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 compressor, pump, etc.), the average operational hours, and the local electricity unit price. The local electricity unit price ranged from $0.08 to$0.14 per kWh.

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 includes activities such as filling field logs, performing system inspection, ordering inventory, and others as recommended by the vendors. EPA study-related activities such as performing field measurements, collecting and shipping samples, and communicating with the ADP Project Study Lead, were not used in the cost analysis. The labor rate was provided by the operators.

## RESULTS AND DISCUSSION

### Capital cost

#### Total capital cost

The total capital cost and some design information of 48 treatment systems of the ADP are provided in Table 2 (AM systems) and Table 3 (all other systems except for RO). The RO system (school) information is excluded because of the unique nature of this project whereby a separate water distribution line was installed in the school to provide RO treated water to only the drinking water outlets and kitchen. The total capital cost of the systems ranged from $14,000 for a small 22 gpm AM system to$427,407 for a 770 gpm IR system. The total capital cost of each system plotted as a function of its design flow rate is shown in Figure 4. The linear regression of the data set yielded an R2 value of 0.74. A visual observation of these data indicates significantly more scatter of the total costs of the systems treating flows of ≤100 gpm than those treating ≥100 gpm, particularly with the AM systems. The difference can be more clearly seen when the data are divided into two design flow categories: ≤100 and ≥100 gpm (Figure 5(a) and 5(b)).

Table 2

Media Tanks      Equipment Engineering I & SSUa
No. ID Year Design flow gpm (L/min) EBCT Design No. Tanks material EBCT min pH system Total capital cost ($)$/gpm design flow Cost ($) Cost ($) Cost ($) BL 2005 10 (38) ARM 200 Series FRP 3.4 No 27,255 2,726 10,435 38 11,000 40 5,820 21 GF 2004 10 (38) E33 Series FRP 3.7 No 34,201 3,420 22,431 66 4,860 14 6,910 20 SU 2005 12 (45) A/I Complex Series Polyglass 1.0 No 16,930 1,411 8,640 51 3,400 20 4,890 29 WA 2004 14 (53) A/I Complex Series (2) Polyglass 1.6 No 16,475 1,177 10,790 65 1,800 11 3,885 24 PF 2008 15 (57) ArsenXnp Series FRP 1.1 No 17,255 1,150 11,345 66 b 5,910 34 WS 2008 20 (76) Adsorbsia Series FRP 2.8 No 51,895 2,595 30,215 58 10,110 19 11,570 22 DM 2004 22 (83) A/I Complex Series (2) Polyglass No 14,000 636 8,990 64 2,400 17 2,610 19 KF(1) 2005 30 (114) ArsenXnp Series (2) FRP 2.5 No 55,847 1,862 39,108 70 9,941 18 6,798 12 VV 2004 37 (140) AAFS50 Series FRP 3.5 Yes 228,309 6,171 122,544 54 50,659 22 55,106 24 10 BR 2005 40 (151) E33 Series CS 4.1 Yes 138,642 3,466 94,662 68 24,300 18 19,680 14 11 BW 2004 40 (151) G2 Series SS 16 Yes 166,050 4,151 105,350 63 17,200 10 43,500 26 12 RR 2004 45 (170) E33 Series FRP 3.7 No 88,307 1,962 63,785 72 11,372 12 13,150 15 13 LI 2005 50 (189) ArsenXnp Parallel FRP 4.0 No 114,070 2,281 82,470 72 12,800 11 18,800 16 14 KF(2) 2005 60 (227) ARM 200 Series FRP 2.5 No 59,516 992 41,689 70 10,587 18 7,240 12 15 KF(3) 2005 60 (227) Adsorbsia Series FRP 2.5 No 73,258 1,221 51,314 70 13,032 18 8,912 12 16 TN 2007 63 (238) E33 Parallel FRP 4.5 Yes 115,306 1,830 86,018 75 12,897 11 16,391 14 17 LD 2007 75 (283) ArsenXnp Series FRP 2.8 No 87,892 1,172 60,678 69 14,214 16 13,000 15 18 WM 2006 100 (379) E33 Parallel CS 5.7 Yes 149,221 1,492 103,897 70 25,310 17 20,014 13 19 RF 2003 120 (454) E33 Parallel FRP 3.7 Yes 131,692 1,097 105,805 80 4,672 21,215 16 20 TE 2005 150 (568) Isolux Parallel CS 0.6 No 76,840 512 58,500 76 8,500 11 9,840 13 21 AL 2005 150 (568) E33 Series FRP 3.1 No 179,750 1,198 124,103 69 14,000 41,647 23 22 NP 2006 160 (605) E33 Parallel FRP 3.3 Yes 143,113 894 116,645 82 11,683 14,830 10 23 GE 2007 200 (757) E33 Parallel CS 3.7 No 139,149 696 101,290 73 19,545 14 18,314 13 24 SV 2004 3,000 (1,136) E33 Parallel FRP No 211,000 703 129,500 61 36,700 17 44,800 21 25 AN 2003 320 (1,211) E33 Parallel FRP 3.6 No 153,000 478 112,000 73 23,000 15 18,000 12 26 RN 2005 350 (1,325) GFH Parallel CS 5.2 No 232,147 663 157,647 68 16,000 58,500 25 27 TA 2005 450 (1,703) E33 Parallel FRP 3.6 No 296,644 659 202,685 68 32,750 11 61,209 21 28 BC 2003 640 (2,423) E33 Parallel FRP 3.7 No 305,000 477 218,000 71 35,500 12 51,500 17 Media Tanks Equipment Engineering I & SSUa No. ID Year Design flow gpm (L/min) EBCT Design No. Tanks material EBCT min pH system Total capital cost ($) $/gpm design flow Cost ($) Cost ($) Cost ($)
BL 2005 10 (38) ARM 200 Series FRP 3.4 No 27,255 2,726 10,435 38 11,000 40 5,820 21
GF 2004 10 (38) E33 Series FRP 3.7 No 34,201 3,420 22,431 66 4,860 14 6,910 20
SU 2005 12 (45) A/I Complex Series Polyglass 1.0 No 16,930 1,411 8,640 51 3,400 20 4,890 29
WA 2004 14 (53) A/I Complex Series (2) Polyglass 1.6 No 16,475 1,177 10,790 65 1,800 11 3,885 24
PF 2008 15 (57) ArsenXnp Series FRP 1.1 No 17,255 1,150 11,345 66 b  5,910 34
WS 2008 20 (76) Adsorbsia Series FRP 2.8 No 51,895 2,595 30,215 58 10,110 19 11,570 22
DM 2004 22 (83) A/I Complex Series (2) Polyglass No 14,000 636 8,990 64 2,400 17 2,610 19
KF(1) 2005 30 (114) ArsenXnp Series (2) FRP 2.5 No 55,847 1,862 39,108 70 9,941 18 6,798 12
VV 2004 37 (140) AAFS50 Series FRP 3.5 Yes 228,309 6,171 122,544 54 50,659 22 55,106 24
10 BR 2005 40 (151) E33 Series CS 4.1 Yes 138,642 3,466 94,662 68 24,300 18 19,680 14
11 BW 2004 40 (151) G2 Series SS 16 Yes 166,050 4,151 105,350 63 17,200 10 43,500 26
12 RR 2004 45 (170) E33 Series FRP 3.7 No 88,307 1,962 63,785 72 11,372 12 13,150 15
13 LI 2005 50 (189) ArsenXnp Parallel FRP 4.0 No 114,070 2,281 82,470 72 12,800 11 18,800 16
14 KF(2) 2005 60 (227) ARM 200 Series FRP 2.5 No 59,516 992 41,689 70 10,587 18 7,240 12
15 KF(3) 2005 60 (227) Adsorbsia Series FRP 2.5 No 73,258 1,221 51,314 70 13,032 18 8,912 12
16 TN 2007 63 (238) E33 Parallel FRP 4.5 Yes 115,306 1,830 86,018 75 12,897 11 16,391 14
17 LD 2007 75 (283) ArsenXnp Series FRP 2.8 No 87,892 1,172 60,678 69 14,214 16 13,000 15
18 WM 2006 100 (379) E33 Parallel CS 5.7 Yes 149,221 1,492 103,897 70 25,310 17 20,014 13
19 RF 2003 120 (454) E33 Parallel FRP 3.7 Yes 131,692 1,097 105,805 80 4,672 21,215 16
20 TE 2005 150 (568) Isolux Parallel CS 0.6 No 76,840 512 58,500 76 8,500 11 9,840 13
21 AL 2005 150 (568) E33 Series FRP 3.1 No 179,750 1,198 124,103 69 14,000 41,647 23
22 NP 2006 160 (605) E33 Parallel FRP 3.3 Yes 143,113 894 116,645 82 11,683 14,830 10
23 GE 2007 200 (757) E33 Parallel CS 3.7 No 139,149 696 101,290 73 19,545 14 18,314 13
24 SV 2004 3,000 (1,136) E33 Parallel FRP No 211,000 703 129,500 61 36,700 17 44,800 21
25 AN 2003 320 (1,211) E33 Parallel FRP 3.6 No 153,000 478 112,000 73 23,000 15 18,000 12
26 RN 2005 350 (1,325) GFH Parallel CS 5.2 No 232,147 663 157,647 68 16,000 58,500 25
27 TA 2005 450 (1,703) E33 Parallel FRP 3.6 No 296,644 659 202,685 68 32,750 11 61,209 21
28 BC 2003 640 (2,423) E33 Parallel FRP 3.7 No 305,000 477 218,000 71 35,500 12 51,500 17

$/gpm can be converted to$/L min using a multiplier factor of 0.264.

aInstallation and system start-up.

bIncluded in equipment cost.

Table 3

Capital costs and design information of iron removal, iron removal plus adsorptive media, coagulation/filtration, and ion exchange systems

Filters    Equipment Engineering I & SSUc
No. Site ID Year Design flow rate gpm (L/min) Contact tanks No. Material Rate gpm/ft2a Filtration media Total capital cost ($)$/gpm design flowb Cost ($) Cost ($) Cost ($) IR SC 2005 20 (76) FRP 5.4 Macrolite 63,547 3,177 20,227 35 20,227 32 12,410 22 DV 2005 45 (170) FRP 9.4 Macrolite® 60,500 1,344 19,790 33 20,580 34 20,130 33 FC 2008 60 (227) None FRP 2.1 G2® 128,118 2,135 103,118 80 7,500 17,500 14 WV 2009 96 (363) None CS 3.4 GreensandPlus™ 161,560 1,683 90,750 56 22,460 14 48,350 30 CM 2004 140 (530) FRP 10 Macrolite® 270,530 1,932 159,419 59 39,344 15 71,767 27 SA 2006 250 (946) FRP 10 Macrolite® 287,159 1,149 160,875 56 49,164 17 77,120 27 SD 2006 340 (1,287) Yes 3 cellsd Al 2.5 Silica Sand 364,916 1,073 205,800 56 27,077 132,039 36 GV 2009 375 (1,420) None FRP 10 Macrolite® 332,584 887 196,542 59 48,057 14 87,985 26 PW 2005 400 (1,514) CS 10 Macrolite® 334,573 836 224,994 67 30,929 78,650 24 10 AR 2006 770 (2,915) CS 10 Macrolite® 427,407 555 281,048 66 50,770 12 95,589 22 IR/AM WL 2008 30 (114) None FRP 4.8 Birm®/Filox™ + GTO 66,362 2,212 46,267 70 3,850 16,245 24 GS 2008 25 (95) None FRP AD26 + E33 55,423 2,217 31,735 57 11,278 20 12,410 22 SF 2005 250 (946) None CS 6.1 AD26 + E33 292,252 1,169 212,826 73 27,527 51,899 18 ST 2006 250 (946) Yes 4 cellsd CS/RFP 2.6 Silica Sand + E33 367,838 1,471 273,873 74 16,520 77,445 21 C/F CL 2009 250 (946) None CS 5.2 AD GS+ 216,876 868 161,650 75 21,726 10 33,500 15 TF 2006 250 (946) FRP 10 Macrolite® 305,447 1,222 168,142 55 53,534 17 83,870 27 FE 2006 375 (1,420) FRP 10 Macrolite® 334,297 891 201,292 60 44,520 13 78,650 24 OK 2008 550 (2,082) CS 10 Electromedia®424,817 772 296,430 70 48,332 11 80,055 19 IE FL 2005 250 (946) NA FRP 12.5 Purolite Arsenex II 286,388 1,146 173,195 61 35,616 12 77,574 27 VA 2006 540 (2,044) NA FRP 10 Purolite A300E 395,434 732 260,194 66 49,840 13 85,400 22 Filters Equipment Engineering I & SSUc No. Site ID Year Design flow rate gpm (L/min) Contact tanks No. Material Rate gpm/ft2a Filtration media Total capital cost ($) $/gpm design flowb Cost ($) Cost ($) Cost ($)
IR
SC 2005 20 (76) FRP 5.4 Macrolite 63,547 3,177 20,227 35 20,227 32 12,410 22
DV 2005 45 (170) FRP 9.4 Macrolite® 60,500 1,344 19,790 33 20,580 34 20,130 33
FC 2008 60 (227) None FRP 2.1 G2® 128,118 2,135 103,118 80 7,500 17,500 14
WV 2009 96 (363) None CS 3.4 GreensandPlus™ 161,560 1,683 90,750 56 22,460 14 48,350 30
CM 2004 140 (530) FRP 10 Macrolite® 270,530 1,932 159,419 59 39,344 15 71,767 27
SA 2006 250 (946) FRP 10 Macrolite® 287,159 1,149 160,875 56 49,164 17 77,120 27
SD 2006 340 (1,287) Yes 3 cellsd Al 2.5 Silica Sand 364,916 1,073 205,800 56 27,077 132,039 36
GV 2009 375 (1,420) None FRP 10 Macrolite® 332,584 887 196,542 59 48,057 14 87,985 26
PW 2005 400 (1,514) CS 10 Macrolite® 334,573 836 224,994 67 30,929 78,650 24
10 AR 2006 770 (2,915) CS 10 Macrolite® 427,407 555 281,048 66 50,770 12 95,589 22
IR/AM
WL 2008 30 (114) None FRP 4.8 Birm®/Filox™ + GTO 66,362 2,212 46,267 70 3,850 16,245 24
GS 2008 25 (95) None FRP AD26 + E33 55,423 2,217 31,735 57 11,278 20 12,410 22
SF 2005 250 (946) None CS 6.1 AD26 + E33 292,252 1,169 212,826 73 27,527 51,899 18
ST 2006 250 (946) Yes 4 cellsd CS/RFP 2.6 Silica Sand + E33 367,838 1,471 273,873 74 16,520 77,445 21
C/F
CL 2009 250 (946) None CS 5.2 AD GS+ 216,876 868 161,650 75 21,726 10 33,500 15
TF 2006 250 (946) FRP 10 Macrolite® 305,447 1,222 168,142 55 53,534 17 83,870 27
FE 2006 375 (1,420) FRP 10 Macrolite® 334,297 891 201,292 60 44,520 13 78,650 24
OK 2008 550 (2,082) CS 10 Electromedia®424,817 772 296,430 70 48,332 11 80,055 19
IE
FL 2005 250 (946) NA FRP 12.5 Purolite Arsenex II 286,388 1,146 173,195 61 35,616 12 77,574 27
VA 2006 540 (2,044) NA FRP 10 Purolite A300E 395,434 732 260,194 66 49,840 13 85,400 22

agpm/ft2 can be converted to L/min/m2 by using a multiplier factor of 40.7.

b$/gpm can be converted to$/L/min by using a multiplier factor of 0.264.

cInstallation and system start-up.

dAeralater system.

Figure 4

Relationship between total capital cost of all treatment systems and design flow rate (gpm units can be converted to L/min by using a multiplier factor of 3.78).

Figure 4

Relationship between total capital cost of all treatment systems and design flow rate (gpm units can be converted to L/min by using a multiplier factor of 3.78).

Figure 5

Relationship of total capital cost of water systems and design flow rate by system category (gpm units can be converted to L/min by using a multiplier factor of 3.78).

Figure 5

Relationship of total capital cost of water systems and design flow rate by system category (gpm units can be converted to L/min by using a multiplier factor of 3.78).

The wide variation of the total capital costs for the systems treating ≤100 gpm (Figure 5(b)) can be attributed to the extensive variations of the design features of these very small systems, particularly for the AM systems. The major AM design features that impact capital cost include the empty bed contact time (EBCT), tank configuration (series vs. parallel), tank material, and the amount of instrumentation and controls. For these small AM systems, the cost of media becomes a major component of the total cost and in the ADP nine different media products were used that had costs ranging from $40 to over$500 per cubic foot. The EBCT, which determines the tank(s) size, varied from 1.5 to 16 min (all $/cu ft units can be converted to$/L by using a multiplier of 0.035). Another factor impacting the tank(s) cost is the material: carbon steel, stainless steel (SS), FRP, and polyglass. SS tanks cost more than the carbon steel tanks which cost more than FRP and polyglass tanks. Five AM systems had equipment for pH adjustment and four had backwash storage/recycle tanks, all of which increase the total capital cost. Although some water quality parameters, such as the arsenic concentration of the source water, have an impact on the O&M cost (bed life), they are generally minor factors in system design and equipment. One exception was pH, whereas all AM vendors recommended pH adjustment (equipment) when the pH exceeded 9.

For the very small AM systems, the highest cost system ($228,309), with a design flow rate of only 40 gpm, had pH adjustment, backwash recycle, and extensive instrumentation accounting for the higher than expected total capital costs. The second highest cost AM system ($166,050) had a very high EBCT (16 min), SS tanks, and two pH control systems for raw and treated water. For the most part, the low-cost systems had small EBCTs (1.5–3 min), used small, inexpensive polyglass tanks and had no backwash capability, automatic controls, or pH adjustment. The wide variation in costs of the very small systems is consistent with the cost data reported by Hilkert et al. (2010) on AM systems in California.

The total capital costs data shown in Figure 5(a) (flows ≥ 100 gpm) show less variation for the AM systems that is attributed to the majority of these systems having similar design features. The AM systems were mainly E33 media systems of parallel design with (FRP) tank systems having EBCTs around 5 min. The designs of the other systems (IR, C/F, and IX) were somewhat similar to the AM systems consisting of parallel design (FRP) pressure filtration tank systems with comparable filtration rates.

As shown in Figure 5(a), the total capital costs of the AM systems treating ≥100 gpm were lower than those of other systems (shown as a group). The average cost on a $/gpm basis for the AM systems was 21% lower than the IR/CF systems cost;$806 versus $1,019. The slightly higher capital cost of the IR, C/F, IR/AM, and IX systems is likely due to the additional ancillary equipment that include contact tanks (C/F systems) and the instrumentation and controls of these systems. This difference in cost is also consistent with the California systems cost data reported by Hilkert et al. (2010) where the IR and C/F systems had higher capital costs than the AM systems. #### Equipment, engineering, and installation and start-up cost A breakdown of the total capital costs into the three cost categories of equipment, engineering, and installation and start-up is provided in Tables 2 and 3. Except for two systems, the highest cost category for all of the systems was equipment. When broken down by types of system into two groups, AM and all other systems (Figure 6(a) and 6(b)), the average percent for the equipment costs for the two groups is 67 and 61%, respectively. Considering all of the systems as a whole, equipment cost amounted to around two-thirds of the total capital cost. Figure 6 Percentage of three cost categories of the total capital costs of adsorptive media systems and other systems (gpm units can be converted to L/min by using a multiplier factor of 3.78). Figure 6 Percentage of three cost categories of the total capital costs of adsorptive media systems and other systems (gpm units can be converted to L/min by using a multiplier factor of 3.78). The second highest cost category was installation and start-up with the average percent for the AM system at 18% and the other systems group slightly higher at 24%. The engineering category averaged 15 and 14% for the AM and other systems, respectively. Equipment cost proposals were solicited for each project in the ADP. Assuming that most small water systems would likely follow the same process of seeking cost proposals from equipment vendors for an arsenic treatment system, as a rule of thumb, the water system should expect the total capital cost (that included equipment, engineering, and installations and start-up) to be as follows: #### Operational and maintenance cost The O&M costs for 35 arsenic removal systems are provided in Table 4 (15 AM systems) and Table 5 (20 other systems). During the ADP evaluation studies, only 15 systems changed out the adsorptive media and thus the O&M cost data in Table 4 are limited to these 15 AM systems. Table 4 Operational and maintenance cost of adsorptive media systems Mediab Chemicals Labor No.a Site ID Design flow rate L/min Media BVs to break through × 1000 Total O/M cost$/Kgal Cost $/Kgal % of total O/M Electricity$/Kgal Type Cost $/Kgal Avg weekly hours Labor rates$/hr Cost $/Kgal GF 10 (38) E33 27,874 2.34 2.01 86 0.00 None 0.00 0.5 21.00 0.33 SU 12 (45) A/I Complex 7,660 12.06 8.96 74 0.00 None 0.00 0.33 30.00 3.10 WA 14 (53) A/I Complex 5,100 22.88 22.05 96 0.00 None 0.00 0.75 20.00 0.83 WA 14 (53) GFH 11,600 10.44 9.44 90 0.00 None 0.00 0.75 20.00 1.00 WA 14 (53) CFH 15,300 5.52 4.76 86 0.00 None 0.00 0.75 20.00 0.76 PF 15 (57) LayneRT 15,000 7.67 5.31 69 0.00 None 0.00 1.6 20.00 2.36 DM 22 (83) A/I Complex 5,814 10.86 9.99 92 0.00 None 0.00 0.5 20.00 0.87 VV 37 (140) AAFS50 10,364 2.74 2.56 93 0.16 None 0.00 0.4 21.00 0.03 VV 37 (140) AAFS50 23,031 2.06 1.15 56 0.16 Acid 0.61 2.4 21.00 0.14 VV 37 (140) ARM 200 25,717 3.40 3.32 98 0.16 None 0.00 0.4 21.00 0.03 11 BW 40 (151) G2® 3,064 5.11 4.30 84 0.00 Acid/base 0.11/0.36 2.33 20.00 0.34 12 RR 45 (170) E33 52,151 0.86 0.64 74 0.01 None 0.00 1.67 21.00 0.22 14 KF 60 (227) ARM 200 13,940 5.82 5.37 92 0.00 None 0.00 2.5 21.00 0.45 17 LD 75 (283) ArsenXnp 66,794 0.98 0.58 59 0.00 None 0.00 21.00 0.40 20 TE 150 (568) Isolux™ 80,000 1.16 1.02 88 0.00 None 0.00 2.5 37.50 0.14 21 AL 150 (568) E33 38,140 0.61 0.36 59 0.00 None 0.00 4.67 19.50 0.25 24 SV 300 (1,136) E33 78,393 0.61 0.30 49 0.05 Replacement parts 0.03 1.75 21.80 0.23 25 AN 320 (1,211) E33 50,191 0.75 0.66 89 0.00 Replacement parts 0.03 1.75 18.20 0.05 26 RN 350 (1,325) GFH® 7,200 5.69 5.51 97 0.00 None 0.00 2.5 35.00 0.18 Mediab Chemicals Labor No.a Site ID Design flow rate L/min Media BVs to break through × 1000 Total O/M cost$/Kgal Cost $/Kgal % of total O/M Electricity$/Kgal Type Cost $/Kgal Avg weekly hours Labor rates$/hr Cost $/Kgal GF 10 (38) E33 27,874 2.34 2.01 86 0.00 None 0.00 0.5 21.00 0.33 SU 12 (45) A/I Complex 7,660 12.06 8.96 74 0.00 None 0.00 0.33 30.00 3.10 WA 14 (53) A/I Complex 5,100 22.88 22.05 96 0.00 None 0.00 0.75 20.00 0.83 WA 14 (53) GFH 11,600 10.44 9.44 90 0.00 None 0.00 0.75 20.00 1.00 WA 14 (53) CFH 15,300 5.52 4.76 86 0.00 None 0.00 0.75 20.00 0.76 PF 15 (57) LayneRT 15,000 7.67 5.31 69 0.00 None 0.00 1.6 20.00 2.36 DM 22 (83) A/I Complex 5,814 10.86 9.99 92 0.00 None 0.00 0.5 20.00 0.87 VV 37 (140) AAFS50 10,364 2.74 2.56 93 0.16 None 0.00 0.4 21.00 0.03 VV 37 (140) AAFS50 23,031 2.06 1.15 56 0.16 Acid 0.61 2.4 21.00 0.14 VV 37 (140) ARM 200 25,717 3.40 3.32 98 0.16 None 0.00 0.4 21.00 0.03 11 BW 40 (151) G2® 3,064 5.11 4.30 84 0.00 Acid/base 0.11/0.36 2.33 20.00 0.34 12 RR 45 (170) E33 52,151 0.86 0.64 74 0.01 None 0.00 1.67 21.00 0.22 14 KF 60 (227) ARM 200 13,940 5.82 5.37 92 0.00 None 0.00 2.5 21.00 0.45 17 LD 75 (283) ArsenXnp 66,794 0.98 0.58 59 0.00 None 0.00 21.00 0.40 20 TE 150 (568) Isolux™ 80,000 1.16 1.02 88 0.00 None 0.00 2.5 37.50 0.14 21 AL 150 (568) E33 38,140 0.61 0.36 59 0.00 None 0.00 4.67 19.50 0.25 24 SV 300 (1,136) E33 78,393 0.61 0.30 49 0.05 Replacement parts 0.03 1.75 21.80 0.23 25 AN 320 (1,211) E33 50,191 0.75 0.66 89 0.00 Replacement parts 0.03 1.75 18.20 0.05 26 RN 350 (1,325) GFH® 7,200 5.69 5.51 97 0.00 None 0.00 2.5 35.00 0.18 aReference number; see Table 2. bMedia changed out of all tanks of both series and parallel design systems. Table 5 Operational and maintenance cost of iron removal, iron removal plus adsorptive media, coagulation/filtration, and ion exchange systems Chemicals Electricity Labor No. Site ID Design flow rate gpm (L/min) Total O&M costs ($/Kgal) Type Cost ($/Kgal)a % of total O&M Cost ($/Kgal)a % of total O&M Average weekly hours Labor rate ($/hr) Cost ($/Kgal)a % of total O&M
IR
SC 20 (76) 0.36 KMnO4 $0.07 19 0.01 0.42 21.00 0.28 78 DV 45 (170) 0.26 NaClO$0.09 34 0.06 24 0.42 10.80 0.11 42
FC 60 (227) 2.26 NaClO $0.00 0.00 1.67 22.00 2.26 100 WV 96 (363) 0.65 NaMnO4$0.37 57 0.16 25 1.75 15.00 0.12 18
CM 140 (530) 0.29 FeCl3b $0.03 10 0.04 14 2.5 21.00 0.22 76 SA 250 (946) 0.43 NaClO$0.05 12 0.01 1.75 10.00 0.37 86
SD 340 (1,287) 0.27 NaClO $0.04 15 0.16 59 4.5 18.00 0.07 26 GV 375 (1,420) 0.55 NaClO$0.00 0.03 10 24.00 0.52 95
PW 400 (1,514) 0.17 FeCl3b $0.01 0.05 29 2.5 30.00 0.11 64 10 AR 770 (2,915) 0.07 KMnO4$0.03 43 0.00 2.5 $30.00 0.04 57 IR + AM WL 30 (114) 1.93b None$0.00 0.39 20 30.0 1.54 80
GS 25 (95) 2.90b NaClO $0.33 11 0.00 1.6 16.0 2.57 89 SF 250 (946) 0.33b NaClO$0.17 51 0.00 2.33 21.0 0.16 48
ST 250 (946) 0.16b NaClO $0.00 0.08 50 1.7 16.3 0.08 50 C/F CL 250 (946) 0.46 FeCl3$0.07 15 0.06 13 22.0 0.33 72
TF 250 (946) 0.18 FeCl3 $0.02 0.01 4.7 19.6 0.16 88 FE 375 (1,420) 0.31 FeCl3$0.05 16 0.05 15 5.25 30.0 0.21 69
OK 550 (2,082) 0.18 FeCl3, NaClO $0.03,$0.01 17 0.08 44 5.25 30.0 0.06 33
IE
FL 250 (946) 0.62 Salt $0.49 79 0.08 13 2.5 21.00 0.05 VA 540 (2,044) 0.35 Salt$0.29 83 0.03 3.3 21.00 0.03 10
Chemicals Electricity Labor
No. Site ID Design flow rate gpm (L/min) Total O&M costs ($/Kgal) Type Cost ($/Kgal)a % of total O&M Cost ($/Kgal)a % of total O&M Average weekly hours Labor rate ($/hr) Cost ($/Kgal)a % of total O&M IR SC 20 (76) 0.36 KMnO4$0.07 19 0.01 0.42 21.00 0.28 78
DV 45 (170) 0.26 NaClO $0.09 34 0.06 24 0.42 10.80 0.11 42 FC 60 (227) 2.26 NaClO$0.00 0.00 1.67 22.00 2.26 100
WV 96 (363) 0.65 NaMnO4 $0.37 57 0.16 25 1.75 15.00 0.12 18 CM 140 (530) 0.29 FeCl3b$0.03 10 0.04 14 2.5 21.00 0.22 76
SA 250 (946) 0.43 NaClO $0.05 12 0.01 1.75 10.00 0.37 86 SD 340 (1,287) 0.27 NaClO$0.04 15 0.16 59 4.5 18.00 0.07 26
GV 375 (1,420) 0.55 NaClO $0.00 0.03 10 24.00 0.52 95 PW 400 (1,514) 0.17 FeCl3b$0.01 0.05 29 2.5 30.00 0.11 64
10 AR 770 (2,915) 0.07 KMnO4 $0.03 43 0.00 2.5$30.00 0.04 57
IR + AM
WL 30 (114) 1.93b None $0.00 0.39 20 30.0 1.54 80 GS 25 (95) 2.90b NaClO$0.33 11 0.00 1.6 16.0 2.57 89
SF 250 (946) 0.33b NaClO $0.17 51 0.00 2.33 21.0 0.16 48 ST 250 (946) 0.16b NaClO$0.00 0.08 50 1.7 16.3 0.08 50
C/F
CL 250 (946) 0.46 FeCl3 $0.07 15 0.06 13 22.0 0.33 72 TF 250 (946) 0.18 FeCl3$0.02 0.01 4.7 19.6 0.16 88
FE 375 (1,420) 0.31 FeCl3 $0.05 16 0.05 15 5.25 30.0 0.21 69 OK 550 (2,082) 0.18 FeCl3, NaClO$0.03, $0.01 17 0.08 44 5.25 30.0 0.06 33 IE FL 250 (946) 0.62 Salt$0.49 79 0.08 13 2.5 21.00 0.05
VA 540 (2,044) 0.35 Salt $0.29 83 0.03 3.3 21.00 0.03 10 a$/Kgal can be converted to $/KL by using a multiplier factor of 3.78. bIron addition to increase arsenic removal. Media replacement cost not incurred during the study period; thus, not included in the total O&M cost. Figure 7 presents the O&M costs on a 1,000 gal (Kgal) (all Kgal and$/Kgal units can be converted to LK and $/KL by using a multiplier of 0.035) of treated water basis for all treatment systems by their design flow rate. As reported in Tables 4 and 5 and shown in Figure 7, the total O&M costs for the AM systems had a very wide range from$0.61 to $22.88 compared to a range of only$0.07 to $2.90 for the other systems. Moreover, if the three very high O&M costs ($2.90, $2.26, and$1.93) of these non-AM systems are excluded, the cost range of the other 17 systems is only $0.07–0.65 and the average is$0.34/Kgal.

Unquestionably, the average O&M cost for the AM systems is substantially higher than the other types of systems, particularly for the very small systems (flows ≤100 gpm), as more clearly shown in Table 6. Furthermore, as shown in Table 6, the average unit O&M costs for all of the systems are significantly higher for the smaller size (≤100 gpm) systems than the larger systems (≥100 gpm).

Figure 7

Total O/M cost verses system design flow of all systems (gpm units can be converted to L/min by using a multiplier factor of 3.78; $/1,000 gal can be converted to$/1,000 L by using a multiplier factor of 3.78).

Figure 7

Total O/M cost verses system design flow of all systems (gpm units can be converted to L/min by using a multiplier factor of 3.78; $/1,000 gal can be converted to$/1,000 L by using a multiplier factor of 3.78).

Table 6

Summary of the operational and maintenance cost of arsenic removal systems

Technology Total O&M (Avg) ($/Kgal)a Media replacement cost ($/Kgal)a Chemical cost ($/Kgal)a Electricity cost ($/Kgal)a Labor cost ($/Kgal)a Systems <100 gpm (379 L/min) AM 6.47 5.58 0.08 0.03 0.78 IR and CF 1.39 NA 0.14 0.10 1.15 Systems > =100 gpm (379 L/min) AM 1.76 1.57 0.01 0.01 0.17 IR and CF 0.28 NA 0.04 0.05 0.19 IX 0.49 NA 0.39 0.06 0.04 Technology Total O&M (Avg) ($/Kgal)a Media replacement cost ($/Kgal)a Chemical cost ($/Kgal)a Electricity cost ($/Kgal)a Labor cost ($/Kgal)a
Systems <100 gpm (379 L/min)
AM 6.47 5.58 0.08 0.03 0.78
IR and CF 1.39 NA 0.14 0.10 1.15
Systems > =100 gpm (379 L/min)
AM 1.76 1.57 0.01 0.01 0.17
IR and CF 0.28 NA 0.04 0.05 0.19
IX 0.49 NA 0.39 0.06 0.04

a$/Kgal can be converted to$/KL by using a multiplier factor of 3.78.

The higher O&M costs of the AM systems are attributed to media replacement whose costs ranged from 49 to 97% of the total AM O&M costs with the average of all 15 systems at 81% (Tables 4 and 6). Media replacement cost is a function of the media cost and bed life (the number of bed volumes (BVs) of treated water before change out). Media products in the ADP ranged from around $80–595 per cubic feet and the number of BVs of water treated before change out ranged from a low of 3,064 to a high of 80,000 BVs. To clearly illustrate the impact of media life, the total O&M cost presented on BV of treated water is shown in Figure 8. The three highest O&M costs were for systems using high cost media products that had a very short bed life. Several systems achieved over 100,000 BVs of treated water, but because the media was not changed out during the ADP studies, their O&M costs are not included in the data set. The O&M cost for these systems would likely be less than the costs of systems treating 80,000 BVs. Figure 8 Relationship between AM O/M cost and BVs of treated water ($/1,000 gal can be converted to $/1,000 L by using a multiplier factor of 3.78). Figure 8 Relationship between AM O/M cost and BVs of treated water ($/1,000 gal can be converted to $/1,000 L by using a multiplier factor of 3.78). At two sites (WA and VV sites), the adsorptive media were changed out three times with the objective to lower the O&M cost. At the WA site, each change resulted in a lower O/M cost mainly the result of increased bed life (Figure 9). At the VV site, pH adjustment was found to lower the media cost by extending the bed life of the original AA media by about 12%. However, the change to a higher performing media product and without pH adjustment did not reduce the O&M cost because of the significantly higher media cost (Figure 9). Figure 9 Impact of media and pH adjustment on total O/M cost ($/1,000 gal can be converted to $/1,000 L by using a multiplier factor of 3.78;$/cf can be converted to $/L by using a multiplier factor of 0.035). Figure 9 Impact of media and pH adjustment on total O/M cost ($/1,000 gal can be converted to $/1,000 L by using a multiplier factor of 3.78;$/cf can be converted to $/L by using a multiplier factor of 0.035). Labor was the highest cost category for the IR (65%) and C/F (66%) systems and the second highest cost category for the AM (16%) and the IX systems (9%). Labor cost reported on a unit basis of$/Kgal of treated water is a function of the operator's time and salary and the amount of water produced by the water treatment system. A review of the operator time records found that the average time, hours per week, spent by the operator average 2.1 hours for the AM systems and slightly higher at 3.2 hours for the other systems.

Labor rates ranged from $10 to$37.50/hour with the average being $22/hour. Thus, with weekly labor cost (operator hours × operator rate) being a fixed cost, the unit O/M labor cost is dependent on the amount of treated water produced during a week. Surprisingly, the systems utilization rates (percentage of time system treating water) were very low, less than 50%, except for one system (Figure 10). By increasing the production time of the systems, the unit labor cost could be lowered. However, the amount of treated water produced is dependent on water demand. If and when a system's water demand is increased, the labor cost on a unit basis will decrease. Because the labor cost is only around 16% of the total AM systems O&M costs, increasing water production for the AM systems would have only a small impact on the total O&M cost. Figure 10 Relationship of system utilization rate to design flow rate of all systems (gpm units can be converted to L/min by using a multiplier factor of 3.78). Figure 10 Relationship of system utilization rate to design flow rate of all systems (gpm units can be converted to L/min by using a multiplier factor of 3.78). The only treatment systems that had significant chemical costs were the two IX systems (salt for regeneration) and three AM system that required pH adjustment (Tables 4 and 5). Electrical costs were also very low with only five systems reporting power costs greater than$0.08/Kgal.

The lower O&M cost is a significant advantage of IR/CF over AM as long as the facility is able to handle the residual disposal at a low cost. All of the IR and CF systems in the ADP had access to sewers and, therefore, residual disposal was not a significant cost issue. Although the cost of residual disposal for the ADP systems was not included in the O&M cost and was not an issue for these systems, residual disposal could be a significant cost issue depending on local conditions and State regulations and must be considered in the technology selection process.

## CONCLUSIONS

The review of the cost capital and O&M data compiled from the EPA Arsenic Demonstration Program has led to the following conclusions.

### Capital cost

• (1) Many factors affect the total cost of an arsenic treatment system that include system flow rate, vessel design and material, media type and quantity, pre- and/or post-treatment requirements, level of instrumentation and controls, and site-specific conditions.

• (2) Because of the wide variation in designs of the smaller (≤100 gpm) AM systems, the equipment costs of these systems can vary significantly for the same size system.

• (3) The designs of the larger (≥100 gpm) AM systems were found to be somewhat similar and, therefore, the equipment costs of the systems had less variation for the same type and size of systems.

• (4) For the larger (≥100 gpm) systems, the average cost (\$/gpm) for the AM systems was found to be 21% lower than those for the IR, C/F, and IX systems. The reason for the slightly more expensive IR, C/F, and IX systems was due to the extra ancillary equipment and controls, such as contact tanks, iron addition system, and brine tanks associated with these systems.

• (5) A breakdown of the total capital cost into three categories (equipment, engineering, and installations and start-up cost) found the equipment cost to be the major component of all treatment systems and amounts to approximately two-thirds of the total capital cost.

• (6) A rough estimate of the total capital cost using an equipment vendor's estimated system cost can be made by increasing the estimated equipment cost by 50%.

### O&M cost

• (1) AM systems have significantly higher O&M costs than all other treatment systems because of the need to replace the adsorptive media when exhausted for arsenic removal.

• (2) For AM systems, media replacement amounts to around 80% of the total O&M cost.

• (3) AM systems can reduce the O&M cost by replacing the existing media with a lower cost media of equal performance or a media of the same cost, but having a higher arsenic removal capacity. Replacing the existing media project that has the combination of low unit cost and longer bed life provides the lowest O&M cost reduction.

• (4) AM systems, except for those designed with very short EBCTs, can reduce their O&M cost by easily converting the system over to a C/F system providing that the backwash water can be easily disposed of. The conversion process will require some capital cost investment in a chemical feed system and a filtration media.

## ACKNOWLEDGEMENTS

The authors wish to thank the Battelle project team for coordinating the performance evaluation studies, data handling, and analyses efforts. The authors also acknowledge the system operators for their data collection efforts. The US Environmental Protection Agency, through its Office of Research and Development, funded, managed, and collaborated in the research described herein. This paper has been subjected to the Agency's administrative review and has been approved for external publication. Any opinions expressed in this paper are those of the author(s) and do not necessarily reflect the views of the Agency, therefore, no official endorsement should be inferred. Any mention of trade names or commercial products does not constitute endorsement or recommendation for use.

## REFERENCES

REFERENCES
Amy
G.
Chen
H.
Drizo
A.
Von Guten
U.
Brandhuber
P.
Hund
R.
Chowdhury
Z.
Kommineni
S.
Sinha
S.
Jekel
M.
Banerjee
K.
2005
Adsorbent Treatment Technologies for Arsenic Removal
.
AWWA Research Foundation
,
Denver, CO
.
Chen
H.
Frey
M. M.
Clifford
D.
McNeill
L. S.
Edwards
M.
1999
Arsenic treatment considerations
.
J. Am. Water Works Assoc.
91
(
3
),
74
85
.
Chen
A. S. C.
Lal
V.
Wang
L.
2011
Arsenic Removal from Drinking Water by Point of Entry/Point of Use Adsorptive Media.
US EPA Demonstration Project at Oregon Institute of Technology at Klamath Falls
,
OR
,
Final Performance Evaluation Report, EPA/600/R-11/035, EPA, Washington, DC
.
Condit
W. E.
Chen
A. S. C.
Wang
L.
2006
Arsenic Removal from Drinking Water by Process Modifications to Coagulation/Filtration
.
US EPA Demonstration Project at Lidgerwood
,
ND
,
Final Evaluation Report. EPA/600/R-06/159, EPA, Washington, DC
.
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, Washington, DC
.
EPA
2003a
Minor Clarification of the National Primary Drinking Water Regulation for Arsenic
.
Federal Register 40 CFR Part 141, Washington, DC
.
EPA
2003b
Arsenic Treatment Technology Evaluation Handbook for Small Systems
.
EPA/816-R-03-014
,
Washington, DC
.
Giles
D. E.
Mohapatra
M.
Issa
T. B.
Anand
A.
Singh
P.
2011
.
J. Environ. Manage.
92
(
8
),
3011
3022
.
Hilkert
E. J.
Young
T. M.
Green
P. G.
Darby
J. L.
2010
.
J. Am. Water Resour. Assoc.
46
(
60
),
1238
1252
.
Mohan
D.
Pittman
C. U.
Jr.
2007
.
J. Hazard. Mater.
142
(
1–2
),
1
53
.
Wang
L.
Chen
A. S. C.
2011
Costs of Arsenic Removal Technologies for Small Systems
.
US Environmental Protection Agency Technology Demonstration program
.
EPA/600/R-11/090, EPA
,
Cincinnati, OH
.
Wang
L.
Condit
W. E.
Chen
A. S. C.
2005
Technology Selection and System Design
.
US Environmental Protection Agency Technology Demonstration program
.
EPA/600/R-05/001, EPA
,
Cincinnati, OH
.