EPA/600/R-11/087
                                                                                   August 2011

Arsenic/Radium Removal from Drinking Water by the HMO Process
U.S. EPA Demonstration Project at Greenville, WI

PROJECT SUMMARY

Abraham S.C. Chen, Ryan Stowe, andLili Wang

In 2003, the town of Greenville, Wisconsin was selected for the Round 2 U.S. Environmental Protection
Agency (EPA) Arsenic Demonstration Program and a Kinetico Macrolite® pressure filtration system was
selected for its ability to remove arsenic. Well reconstruction prior to the planned arsenic study, however,
resulted in changes in water quality that resulted in a decrease in arsenic to below the 10-|o,g/L maximum
contaminant level (MCL) and an increase in radium to above the combined MCL of 5 pCi/L (at around 9
pCi/L). With the changes in water quality (high radium), the treatment process was modified to  include
the addition of hydrous manganese oxide (HMO) for radium removal. Between August 3, 2007  and April
19, 2010, system performance was evaluated in two phases. The first study conducted between August 3,
2007, and January 14, 2008, collected system operational, performance, and cost data on arsenic removal.
Operational data included system pressures, flowrates, filter run lengths, and filter backwashing.
Operation and maintenance (O&M) cost was tracked and analyzed per 1,000 gal of water treated. A
special study that focused entirely on the system's ability to remove radium followed a year later between
May 29, 2009, and April 19, 2010.

Introduction

Starting in 2003, U.S. EPA conducted an Arsenic Demonstration Program that consisted of a series of
full-scale, on-site demonstration projects to remove arsenic from drinking water at 50 locations in 26
states. The goal of this program was to evaluate the performance and determine the cost (capital and
operating) of the technologies demonstrated to help small drinking water systems meet the new MCL.
The specific objectives of these projects were to: (1) evaluate the performance of arsenic removal
technologies for use on small systems, (2) determine the  required system O&M and operator skill levels,
(3) characterize process residuals produced by the technologies, and (4) determine the capital and O&M
cost of the technologies.  The program consisted of three rounds of projects: 12 projects in Round 1, 28
projects in Round 2, and 10 projects in Round 2a.

The town of Greenville, Wisconsin was selected for the demonstration program in the second round of the
program. The town had approximately 5,000 residents and their drinking water was supplied by three
wells designated as Wells No. 2,  3, and 4.  Well No. 2 (500 ft deep) was not used because of elevated
arsenic and iron concentrations and, therefore, it was selected for the arsenic demonstration project. Well
No. 2 water was first sampled on September 21, 2004, during the initial site visit by an EPA contractor,
Battelle, and results of the water analyses are shown in Table 1. As shown in the table, the arsenic
concentration was 34 (ig/L, with the soluble fraction (12.2  (ig/L) consisting of 85% As(III) and 15%
As(V). The iron level was extremely high at 14.5  mg/L.

Two vendor proposals were  submitted for the demonstration project from which a Kinetico proposal was
selected.  The Kinetico treatment system proposed for demonstration was an iron removal filtration
process using Macrolite® as  the filtration media. Macrolite® is a low-density,  spherical, and chemically
inert ceramic media designed for a high-rate filtration up to 10 gpm/ft2. The media, originally

-------
manufactured by Kinetico, but currently marketed by Fairmont Minerals, is approved for use in drinking
water applications under NSF Standard 61.
After isolating the upper section (305/308 ft) of Well No. 2, the utility conducted a 72-hr pumping test on
February 25, 2005. The test results on the water samples collected at the end of the pumping test found
the total arsenic concentration at 16.9 (ig/L and the iron concentration at 7.8 mg/L; both values were
approximately half of the levels found during the initial site visit five months earlier (September 21,
2004).

                         Table 1. Analytical  Results of Well No. 2 Water
1
Parameter
Unit
Sampling Date
(Time)
pH
Temperature
DO
ORP
Total Alkalinity (as CaCO3)
Hardness (as CaCO3)
Turbidity
TDS
TOC
Nitrate (as N)
Nitrite (as N)
Ammonia
Chloride
Fluoride
Sulfate
Silica (as SiO2)
Orthophosphate
Phosphorus (as P)
As (total)
As (soluble)
As (paniculate)
As (III)
As(V)
Fe (total)
Fe (soluble)
Mn (total)
Mn (soluble)
U (total)
U (soluble)
V (total)
V (soluble)
Na (total)
Ca (total)
Mg (total)
Ra-226
Ra-228
S.U.
°c
mg/L
mV
mg/L
mg/L
NTU
mg/L
mg/L
mg/L
mg/L
mg/L
mg/L
mg/L
mg/L
mg/L
mg/L
Hg/L
Hg/L
Hg/L
Hg/L
Hg/L
Hg/L
mg/L
mg/L
Hg/L
Hg/L
Hg/L
Hg/L
Hg/L
Hg/L
mg/L
mg/L
mg/L
pCi/L
pCi/L
Test Results
During
Initial Visit
09/21/04
7.3
11
0.68
-113
235
300
89
308
1.0
0.04
0.01
0.05
3.6
0.1
40
28.3
0.06
NA
34.0
12.2
22.1
10.4
1.8
14.5
7.8
46
54
2.9
NA
7.4
NA
12.0
56
39
4.1
5.0
During 72-hr
Pump Test00
02/25/05
(at71hr)
NA
NA
NA
NA
NA
240
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
16.9
10.8
6.1
8.8
2.0
7.8
7.4
51.1
49.8
1.2
1.0
1.2
0.5
9.8
49.1
28.5
1.6
3.0
After Well
Reconstruction
04/06/06
7.5
NA
NA
NA
240
258
NA
320
NA
0.03
0.01
NA
4.5
0.17
32
NA
NA
NA
9.7
NA
NA
NA
NA
6.2
NA
59.0
NA
NA
NA
NA
NA
9.8
54.0
30.0
4.1
4.0
           (a)  Well puckered between 305 and 308 ft; samples taken by end of pump test at 71 hr.

-------
During December 2005, the well was reconstructed to seal off an upper aquifer (St. Peter sandstone from
110 to 220 below ground surface [bgs]) to try to improve the water quality. The well was sampled on
April 6, 2006, and the arsenic level was found to be 9.7 (ig/L (below the arsenic MCL) and the iron level
was slightly lower at 6.2 mg/L. Unfortunately, the combined radium (226 plus 228) concentration
increased to 8.1 pCi/L (above the radium MCL).

Six months later in October 2006, an AquaStream sand removal device was installed because of a
sediment problem.  A new submersible pump was installed at the same time at 175 ft bgs. Although the
arsenic concentration of Well No. 2 was now slightly below the MCL after well reconstruction, the
demonstration project continued because of the high level of total radium. With the major focus now on
radium removal and a minor focus on arsenic, the Macrolite® filtration system was modified slightly by
including hydrous manganese oxide (HMO) addition, which is effective for radium removal.  The
removal of natural iron  also was expected to reduce arsenic to even lower levels in the treated water.

Although treatment system installation, startup and shakedown was complete by March 9, 2007, the
performance evaluation study did not begin until August 2007 because of a pipe break at the outlet end of
the contact tanks caused by high system inlet pressure (-200 psi) and the failure of two safety features,
i.e., a pressure relief valve and a high pressure switch. To ensure proper system operation, a variable
frequency drive (VFD)  was installed and the pressure relief valve was re-oriented to prevent sediment
from entering the valve.

Because arsenic concentration in the well water was below the  MCL, the performance evaluation study
was reduced from the normal one-year demonstration period to only six months from August 2007 to
early January 2008 with the focus on arsenic  removal and system operation. During the six-month
arsenic removal study, radium removal was monitored by the facility. A follow-up special study on
radium removal by the treatment system was  conducted by Battelle from May 2009 through April 2010.

System performance (arsenic) data were collected similar to the other arsenic demonstration projects.
The types of data collected included system operation, water quality (both across the treatment train and
in the distribution system), residuals, and capital and O&M cost. This report summarizes the activities
performed during and the  results obtained from the two study periods.

Materials and Methods

Arsenic Treatment System

The treatment train included prechlorination, HMO addition, adsorption/co-precipitation, and Macrolite®
pressure filtration.  Figure 1 shows a photograph of the treatment system and Table 2 summarizes relevant
system design specifications. The major process steps and system  components include:

       •   Prechlorination - The existing chlorine gas feed system was used to oxidize soluble As(III)
           and Fe(II) and maintain a target total chlorine residual level of 0.5 mg/L (as C12). The system
           consisted of two 150-lb cylinders with a chlorinator unit on each cylinder, an ejector, an
           automatic  switchover system,  and a scale (to track chlorine consumption).

       •   HMO Addition - Following prechlorination, a 3% HMO (or MnO2) pre-formed with
           manganese  sulfate (MnSO4) and  potassium permanganate (KMnO4) was added for radium
           removal. The HMO addition system consisted of a chemical feed pump, a 325-gal
           polyethylene chemical day tank,  an overhead tank mixer, a day tank scale, and a static mixer.
           The target HMO level was 1.75 mg/L (as MnO2).

-------
Figure 1. Three Macrolite® Filters (forground) and One Contact Tank (background)
 Table 2. Design Specifications for Greenville Macrolite® Pressure Filtration System
Parameter
Value
Remarks
Pretreatment
Prechlorination Dosage (mg/L [as C12])
HMO Dosage (mg/L [as MnO2])
5.5
1.75
For a target free chlorine residual of 0.5 mg/L (as C12)
Based on pilot test results from Well No. 4
Contact
No. of Tanks
Tank Size (in)
Contact Time (min)
2
63 D x 86 H
4.5
Arranged in parallel
850 gal each tank
Based on 375-gpm design flowrate
Filtration
No. of Vessels
Vessel Size (in)
Media Volume (ft3/vessel)
Hydraulic Loading Rate (gpm/ft2)
Differential Pressure Across Vessel (psi)
Hydraulic Utilization (%)
o
J
48 D x 72 H
25
10
15
33
In parallel
-
24-in bed depth of 40/60 US standard mesh Macrolite®
Based on 125-gpm/vessel flowrate
Across a clean bed
Typical operating time 8 hr/day
Backwash
Initiating Pressure (psi)
Initiating Standby Time (hr)
Initiating Service Time (hr)
Backwash Rate (gpm/ft2)
Minimum Backwash Time (min)
Maximum Backwash Time (min)
Turbidity Set Point (NTU)
25
48
24
8-10
5
18
20
Across bed at end of filter run
-
-
100 to 125 gpm
-
-
To terminate backwash

-------
       •   Adsorption/Co-precipitation - The pretreated water passed through two 850-gal fiberglass
           reinforced plastic (FRP) tanks configured in parallel for 4.5 min of contact time based on a
           peak flowrate of 375 gpm.

       •   Pressure Filtration - Pressure filtration involved downflow filtration through three 150 psi-
           rated FRP vessels arranged in parallel.  The vessels were floor mounted and piped to a valve
           rack on a stainless steel frame.  Each filtration vessel was filled with 24 in (or 25 ft3) of 40/60
           U.S. standard mesh Macrolite® media underlain by 30/40 U.S. standard mesh garnet fill. The
           flow through each vessel was regulated to 125 gpm, yielding a filtration rate of 10-gpm/ft2.

           Filter backwashing was triggered by a 25-psi pressure drop, a 24-hr service time, or a 48-hr
           standby time. The filters were backwashed one at a time.  To begin a backwash, water in a
           filter was drained and the filter was air sparged for 4 min.  After settling for 5 min, the filter
           was backwashed with treated water at 100 to 125 gpm (8 to 10 gpm/ft2) until the turbidity of
           backwash wastewater had reached a target threshold level  of 20 nephelometric turbidity units
           (NTU). The filter then underwent a filter-to-waste rinse for 3 min before returning to service.
           Wastewater was sent to a 3,000 gal equalization tank and then the sanitary sewer.

       •   Post-treatment -Sodium silicate was applied for corrosion control at a target dosage of 2.6
           mg/L (as NaSi) and fluoride was added at a target dosage of 0.88 mg/L (as F).


The system engineering package, prepared by the town's engineer, Earth Tech, was submitted to the
Wisconsin Department of Natural Resources (WDNR) on November 7, 2005, and a system construction
permit was issued by WDNR on March 8, 2006.  A supplemental submittal on HMO addition was filed
on November 8, 2006, and the approval was granted on November 20, 2006.  Installation of the system by
Kinetico began on October 23, 2006, after successful AquaStream installation and well development. All
system startup and  shakedown activities were complete by March 9, 2007.

System O&M and Cost Data Collection

The plant operator performed daily, weekly, and monthly system O&M and data collection. On a regular
basis, the plant operator recorded system operational data such as pressure, flowrate, totalizer, and hour
meter readings and conducted visual inspections to ensure normal system operations. In case of
problems, the plant operator contacted Battelle and/or the vendor for troubleshooting and corrective
actions. The capital cost for the system consisted of the cost for equipment, site engineering, and
installation. The O&M cost consisted of the cost of chemical supplies, electricity, and labor. The
operator tracked labor for routine system O&M, troubleshooting, and repairs.  The demonstration-related
work was recorded, but not used for cost analysis.

Sample Collection  and Analyses

During the six-month arsenic removal study from August 13, 2007, through January 14, 2008, eight sets
of water samples were collected. Each set consisted of samples from three locations of the treatment
plant (see Figure 2): (1) influent source water (IN), (2) after contact tanks  (AC) and, (3) the combined
effluent from the three filters (TT). Water samples were analyzed onsite for pH, temperature, and
dissolved oxygen. Water samples sent to the Battelle laboratory were analyzed for total alkalinity,
hardness, turbidity, nitrate, fluoride, sulfate, silica, phosphorus, arsenic, iron,  and manganese.  In addition,
four sets of the samples were speciated for arsenic (particulate, dissolved,  As[III] and As[V]) and for total
and soluble iron and manganese.

Each month for four months, a filter backwash wastewater sample was collected from a side stream of
each filter's wastewater discharge line over the duration of a backwash. These backwash water samples

-------
             Monthly
   (3rd week of a 4-week cycle)
   pH(a), temperature^), DC*),
  As speciation, Fe (total and soluble),
              Mn (total and soluble),
     Ca, Mg, F, NO3, SO4, SiO2, PO4,
           turbidity, and/or alkalinity
   pH®, temperature^, DC*),
               C12 (free and total )
-------
(12) were analyzed for pH, total suspended solids (TSS), total dissolved solids (TDS), total arsenic, iron
and manganese, and soluble arsenic, iron and manganese. All backwash wastewater was discharged to
the town's sewer system.

Four sets of weekly baseline distribution system water samples were collected before system startup from
two residences within the town's historic Lead and Copper Rule (LCR) sampling network and one non-
LCR residence. Following system startup, four monthly samples were collected from the same three
locations to determine the impact of the arsenic treatment system on the water quality in the distribution
system. Sampling followed an instruction sheet developed in accordance with the LCR Monitoring and
Reporting Guidance for Public Water Systems (EPA, 2002).  These distribution water samples were
analyzed for pH, alkalinity, arsenic, iron, manages, lead and copper.

During the one-year special radium removal study from May 2009 through April 2010, only monthly
radium samples were collected and analyzed by Battelle.

All relevant sampling logistics for preparation of arsenic speciation kits, sample containers, sample
coolers, and sample shipping and handling, as well as specific requirements for analytical methods,
sample volumes, containers, preservation, and holding times were addressed in an EPA-endorsed Quality
Assurance Project Plan (QAPP) (Battelle, 2004).

Results

System Operation

From August 3, 2007, through January 14, 2008, the  system operated for a total of 625 hr during the 162
days (less 3 days when the system was taken offline due to a leak on the backwash discharge line),
treating approximately 10,830,000 gal of water. During this time period, daily run time averaged 3.9
hr/day and the average daily water demand was 66,852 gal/day (gpd).  System instantaneous flowrates
ranged from 197 to 347 gpm and averaged 285 gpm,  compared to the 289-gpm average flowrate
calculated based on the total water production and total system run time. The system design flowrate was
375 gpm. Average chlorine and HMO dose rates  were  3.7 mg/L (as C12) and 1.7 mg/L (as MnO2),
respectively, compared to the target dose rates of 5.5  mg/L (as C12) and 1.75 mg/L (as MnO2),
respectively. The target chlorine dose rate was calculated based on 6.2 mg/L of soluble Fe(II), which was
much higher than the 2.0 mg/L (on average) measured during the performance evaluation study. The
system was backwashed for a total of 112 times, equivalent to 0.7 time/day  (or 0 to 2 time/day). Because
the average daily run time was 3.9 hr/day, the average filter run length was 5.6 hr.

The system was fully automated with an operator interface, programmable logic controller (PLC), and
modem housed in a central control panel.  The control panel was connected to various instruments used to
track system performance, including inlet and outlet pressure for each filter, system flowrate, backwash
flowrate, and backwash turbidity. All major functions of the treatment system were automated and
required only minimal operator oversight and intervention if all functions were operating as intended.
Under normal operating conditions, the skill set required to operate the system was limited to observation
of the process equipment integrity and operating parameters such as pressure, flow, and system alarms.
The daily demand on the operator was approximately 2 hr to visually inspect the system, record the
operating parameters on the log sheets, and collect and  analyze water samples.  The operator also
performed O&M activities such as monitoring backwash operational issues  and working with the vendor
to trouble shoot and perform minor on-site repairs.

-------
System Performance

Arsenic, Iron, and Manganese Removal.  Figures 3 and 4 present arsenic and iron test results,
respectively, conducted on the sets of water samples collected across the treatment train during the arsenic
removal study. The Well No. 2 source water contained 4.8 to 7.0 ug/L of total arsenic, existing mostly as
soluble As(III). The well water also had 1.8 to 2.3 mg/L of total iron and 29.4 to 32.6 ug/L of total
manganese, existing almost entirely in soluble form.  With the addition of chlorine at a pH value of 7.3
(on average), arsenic was converted to mostly particulate arsenic, which was removed by the Macrolite®
filters to below 1.4 ug/L. Total iron concentrations also were reduced to just above the MCL of 25 ug/L
by the  filters.  The highest iron concentration measured in the filter effluent was 100 ug/L, indicating
some iron breakthrough from the filters. Table 3 provides ranges and averages of all test results
conducted on the water samples.
                                 Total Arsenic Concentrations Across Treatment Train
                                                                    At Wellhead (IN)
                                                                    After Contact Tanks (AC)
                                                                    After Total Combined Effluent (TT)
                                                  Date
                 Figure 3. Total Arsenic Concentrations Across Treatment Train
After prechlorination, HMO addition, and the contact tanks, average total manganese concentrations
increased from 30.6 to 399 ug/L (excluding two outliers on September 11 and 25, 2007, with unusually
low manganese levels after HMO addition) due to HMO addition. Total manganese concentrations were
reduced to 7.7 ug/L after Macrolite® filtration, indicating little manganese breakthrough from the filters
even with the combined loading of iron and HMO solids.

Radium Removal.  After Well No.  2 reconstruction, radium levels in source water increased up to 9.0
pCi/L. With HMO addition to the filtration process, the total radium level decreased to 3.2 pCi/L.  After
this initial confirmation by Battelle, the facility assumed the responsibility for radium sampling and
analysis up until the special study conducted about two years later.  The data collected by the facility from
July 2007 through January 2008 ranged 2.1 to 3.4 pCi/L and averaged 2.9 pCi/L.

-------
                             Total Iron Concentrations Across Treatment Train
   ra
   a  1,500
                                                             -At Wellhead (IN)
                                                             -After Contact Tanks (AC)
                                                             -After Total Combined Effluent (TT)
             Figure 4. Total Iron Concentrations Across Treatment Train

Table 3. Summary of Test Results on Samples Collected During Arsenic Removal Study
                    from August 3, 2007, Through January 14, 2008
Parameter
pH
Temperature
DO
Total Alkalinity(a)
Hardness(a)
Turbidity
Nitrate (as N)
Fluoride
Sulfate
Silica (as SiO2)
Phosphorus (as P)
As (total)
As (soluble)
As (paniculate)
As (III)
As(V)
Fe (total)
Fe (soluble)
Mn (total)
Mn (soluble)
Ca (total)
Mg (total)
Unit
S.U.
°c
mg/L
mg/L
mg/L
NTU
mg/L
mg/L
mg/L
mg/L
ug/L
ug/L
ug/L
ug/L
ug/L
ug/L
mg/L
mg/L
ug/L
ug/L
mg/L
mg/L
Source Water
Range (Average)
7.2-7.4 (7.3)
10.8-18.5 (14.5)
1.4-2.6 (1.9)
271-297 (281)
295-366(314)
26.0-36.0 (30.2)
0.05-0.8 (0.2)
0.2-0.3 (0.3)
16.0-17.0 (17.0)
12.3-13.6(13.1)
18.5-19.5 (33.4)
4.8-7.0 (5.7)
4.0-5.2 (4.7)
0.4-1.1 (0.8)
3.0-4.8(4.1)
0.3-1.0 (0.6)
1.8-2.3 (2.1)
1.8-2.2 (2.0)
29.4-32.6 (30.6)
27.7-29.5 (28.5)
53.6-62.4 (57.2)
33.3-51.2(41.5)
After Contact Tanks
Range (Average)
7.1-7.3 (7.2)
11.0-18.6(14.1)
1.3-2.5 (1.7)
265-294 (277)
273-389(317)
1.8-13.0 (6.2)
O.05 (O.05)
0.2-0.5 (0.3)
17.0-22.0 (19.3)
12.7-13.6(13.0)
19.8^6.2(31.6)
1.9-5.9 (4.4)
0.2-0.4 (0.3)
1.5-5.7 (3.8)
0.1-0.3 (0.2)
<0. 1-0.2 (0.1)
1.8-2.3 (2.0)
0.025-0.064 (0.025)
28.9-798(318)
2.1-7.6(5.3)
54.4-64.8 (58.0)
33.3-55.1 (41.7)
After Filters
Range (Average)
7.1-7.3(7.2)
10.9-18.6(14.5)
1.2-2.1(1.7)
266-302 (276)
267-386(315)
0.4-1.6 (0.8)
0.05-0.4(0.1)
0.2-0.6 (0.3)
17.0-23.0 (19.8)
12.0-29.0 (15.3)
<10 (<10)
0.2-1.4(0.5)
0.1-0.2(0.2)
0.1-0.5(0.2)
0.1-0.3(0.1)
0.1(0.1)
0.025-0.1 (0.028)
0.025 (0.025)
0.8-22.3 (7.7)
0.3-2.3 (1.4)
50.8-65.6 (56.8)
32.3-53.9 (42.0)
  (a) asCaCO3

-------
A special follow-up radium removal study was conducted on the treatment system by Battelle that
consisted of colleting 11 effluent sample from May 2009 through April 2010.  Table 4 presents analytical
results of this special study.

                     Table 4. Radium Test Results, May, 2009 - April, 2010
Date
Location
05/29/09
06/30/09
07/29/09
08/25/09
09/24/09
10/26/09
11/24/09
01/05/10
02/10/10
03/17/10
04/19/10
Average
Ra-226 (pCi/L)
IN
6.1
6.2
5.0
7.5
6.3
8.6
6.8
6.2
9.5
5.9
7.3
6.8
AC
5.2
5.2
5.0
5.8
6.1
8.0
7.1
5.1
7.5
2.9
7.8
5.8
TT
3.0
4.2
3.2
0.3
3.2
5.1
6.7
1.8
1.4
2.3
0.8
3.1
Ra-228 (pCi/L)
IN
4.1
3.5
3.9
4.2
4.2
3.6
3.1
2.1
3.4
3.2
2.9
3.5
AC
3.4
4.2
4.3
4.0
4.0
3.8
4.5
2.0
3.4
1.9
2.2
3.4
TT
1.3
1.8
2.6
3.0
3.0
2.3
2.5
0.5
1.1
1.7
0.4
2.0
Total Radium^
a*IT(pCi/L)
4.3
6.0
5.7
3.3
6.2
7.4
9.1
2.3
2.5
4.0
1.3
4.7
                    (a)  Combined Ra-226 and Ra-228 in pCi/L.

Ra-226 levels in raw water ranged from 5.0 to 9.5 pCi/L and averaged 6.8 pCi/L while Ra-228 levels
were slightly lower, ranging from 2.1 to 4.2 pCi/L and averaging 3.5 pCi/L. Both Ra-226 and 228 levels
remained relatively unchanged after HMO additions at the AC sample location. After filtration, levels in
the treated water ranged from 0.3 to 6.7 pCi/L for Ra-226 and from 0.4 to 3.0 pCi/L for Ra-228. Average
radium levels in the treated water were 3.1 and 2.0 pCi/L for Ra-226 and Ra-228, respectively.

The HMO target dosage was 1.75 mg/L (as MnO2).  In order for the dosing pumps to work properly, the
HMO solution had to be diluted, which resulted in a lower feed rate. The lower feed rate led to a decrease
in radium removal and an increase in total radium levels to above 5 pCi/L on five of the first seven
sampling occasions during the study period. In December 2009, the town hired a local engineering firm
to check the HMO feed rate and dosage.  The firm discovered that the actual dosage (i.e., 0.5 to 0.7 mg/L
[as MnO2]) was lower than the target dosage to achieve the required total radium removal and increased
the dosage to approximately 1.0 mg/L (as MnO2) in January 2010. After the increase in the HMO dosage,
total radium levels in the filter effluent were reduced to 1.3 to 4.0 pCi/L. However,  addition of HMO  at
1.0 mg/L (as MnO2) added extra loading to the filters, causing solids to breakthrough from the filters
prematurely.

Back-washing. Table 5 summarizes the data collected on filter backwash times and backwash water
quantity produced under the field settings. Backwash flowrates averaged 83 gpm or 6.6 gpm/ft2, which is
lower than the 8 to 10 gpm/ft2 design value.
                          Table 5. Summary of Backwash Parameters
Backwash Parameters
Backwash Duration (min/vessel)
Backwash Flowrate (gpm/vessel)
Water Quantity Generated (gal/vessel)
Hydraulic Loading (gal/ft2)
Minimum
8
78
749
6.2
Maximum
18
87
1260
6.9
Average
11
83
899
6.6
                                              10

-------
Table 6 summarizes analytical results of backwash sampling.  As expected, total arsenic, iron, and
manganese existed mainly in the particulate form. Assuming a TSS level of 278 mg/L and a total
wastewater production volume of 2,700 gal per backwash event, 2,840 g of solids, consisting of 0.7 g,
742g, 104 g of arsenic, iron, and manganese would have been produced and discharged to the sewer.
                           Table 6.  Backwash Water Sampling Results
Parameter
pH
Total Suspended Solids (TSS)
Total Dissolved Solids (TDS)
As (total)
As (soluble)
Fe (total)
Fe (soluble)
Mn (total)
Mn (soluble)
Unit
S.U.
mg/L
mg/L
Hg/L
Hg/L
mg/L
mg/L
Hg/L
Hg/L
Range (Average)
7.4-7.5 (7.4)
196-384 (272)
268-296 (278)
24.5-149 (70.4)
0.1-0.6(0.3)
38.8-118(72.6)
<25-113(32.5)
5,912-14,820 (10,175)
0.5-26.3 (8.8)
                    Note: One-half of detection limit used for non-detect samples
                    for calculation.
Distribution Water. Table 7 presents distribution system water sampling results.  Arsenic concentrations
in the distribution system baseline samples collected prior to startup of the treatment system averaged 1.3
ug/L. Because the distribution samples were collected prior to the completion of Well No. 2
reconstruction, the water samples represented the water from Wells No. 3 and/or 4 that contained little
arsenic. After system startup, arsenic levels in the distribution system averaged 1.3 ug/L, which is
slightly higher than the 0.5 ug/L level in the filter effluent. Iron levels in the distribution after system
startup were mostly below the MCL of 25 ug/L with one outlier as high as 138 ug/L.  Manganese levels
in the distribution system samples decreased from an average baseline level of 1.9 ug/L to 1.5 ug/L after
system startup. The use of HMO apparently did not impart manganese to the distribution water. System
operation appeared to slightly elevate copper and lead concentrations from the baseline levels.  Lead
concentration increased from 3.6 ug/L to 7.2 ug/L after startup, while copper levels increased from 144 to
163 ug/L.

System Cost

Capital Cost.  Capital investment for the  Greenville system was $332,584, including $196,542 for
equipment, $48,057 for site engineering,  and $87,985 for system installation, shakedown, and startup.
The total capital cost, not including the HMO addition system, was normalized to the system's rated
capacity of 375 gpm (540,000 gpd), which resulted in $886.89 per gpm  ($0.62 per gpd). The total capital
cost was converted to an annualized cost of $31,392.60/yr using a capital recovery factor of 0.09439
based on a 7 percent interest rate and a 20-year return period. Assuming that the system operates 24 hr a
day, 7 days a week, at the system design flowrate of 375 gpm, the unit capital cost would be $0.16/1,000
gal of water treated. The system operated at 289 gpm for 3.9 hr/day (on average) and would have
produced an estimated 24,683,500 gal  of water in a year.  At this reduced usage rate, the unit capital cost
was increased to $1.27/1,000 gal.

Operation and Maintenance Cost. The O&M cost included expenditures associated with electricity and
labor. Because the system was under warranty during the five-month study period, no cost was incurred
                                               11

-------
                         Table 7. Distribution System Sampling Results
Sampling
No.
Date
a
_o
03
a m
M c
03 a
55 H
hr
W
Q.
s.u.
Alkalinity
mg/L
(K
<
Mg/L
<0
tu
Mg/L
I
Mg/L
.Q
a.
Mg/L
,
Mg/L
DS1 (Non-LCR Residence)
BL1
BL2
BL3
BL4
02/21/06
02/28/06
03/07/06
03/15/06
Average
1
2
3
4
09/25/07
10/09/07
11/06/07
12/04/07
Average
8.5
8.2
8.7
8.2
8.4
9.3
9.3
9.0
9.5
9.3
7.4
7.5
7.6
7.5
7.5
7.4
7.5
7.5
7.6
7.5
311
321
319
310
315
314
299
306
308
307
0.5
0.5
0.5
0.2
0.4
0.3
0.4
0.2
0.4
0.3
<25
<25
<25
<25
<25
<25
<25
<25
<25
<25
11.3
2.8
2.2
1.3
4.4
2.0
1.5
0.8
0.6
1.2
2.9
1.2
1.6
1.7
1.9
10.0
11.7
3.6
1.8
6.8
56.4
27.7
54.1
95.3
58.4
206
214
183
111
178
DS2 (LCR Residence)
BL1
BL2
BL3
BL4
02/21/06
02/28/06
03/07/06
03/15/06
Average
1
2
3
4
09/25/07
10/09/07
11/06/07
12/04/07
Average
14.0
14.5
15.3
13.8
14.4
14.0
14.5
14.0
14.0
14.1
7.7
7.6
7.7
7.8
7.7
7.4
7.5
7.6
7.9
7.6
323
329
328
338
330
352
325
327
324
332
1.8
2.1
2.0
1.1
1.7
2.0
2.1
1.7
2.0
1.9
<25
<25
<25
<25
<25
<25
<25
<25
28.1
16.4
0.8
1.7
2.3
0.4
1.3
0.9
1.6
1.8
2.6
1.7
5.2
1.5
1.5
20.7
7.2
6.7
6.7
5.7
5.6
6.2
287
40.3
38.3
487
214
95.4
86.3
125
89.5
99.0
/)£? (LC« Residence)
BL1
BL2
BL3
BL4
02/21/06
02/28/06
03/07/06
03/15/06
Average
1
2
3
4
09/25/07
10/09/07
11/06/07
12/04/07
Average
Overall Baseline Ave.
Overall Sampling Ave.
7.5
7.6
8.0
8.5
7.9
7.8
8.0
8.0
8.0
7.9
10.2
10.4
7.7
7.7
7.7
7.7
7.7
7.5
7.5
7.7
7.7
7.6
7.6
7.6
328
333
328
338
332
348
305
329
325
327
326
322
1.9
1.9
1.8
1.1
1.7
1.7
2.1
1.9
1.7
1.8
1.3
1.3
<25
<25
<25
<25
<25
<25
138
34.4
27.4
70.8
<25
33.2
0.1
<0.1
<0.1
0.1
O.I
0.1
5.0
1.2
0.5
1.7
1.9
1.5
1.5
1.1
1.1
3.3
1.7
8.7
16.8
4.0
4.9
8.6
3.6
7.2
141
132
154
210
159
222
231
204
186
211
144
163
for repairs. Chlorination was performed prior to the demonstration study so the incremental cost for the
gas chlorination system was negligible. Pre-formed HMO was funded, in part, by the facility and since
HMO addition was not part of the original system design, the associated cost was not included for the cost
analysis. The incremental power cost was estimated based on the change in electric utility bills for a five-
month period before and after system startup and did not include natural gas cost to heat the building.
The power cost was estimated to be $0.03/1,000 gal of water treated (assuming an annual production of
24,401,000 gal).  Under normal operating conditions, the skill requirements to operate the system were
minimal, with a typical daily demand on the operator of 2 hr over a 5-day workweek. Based on this time
commitment and a labor rate of $24/hr, the labor cost was $0.51/1,000 gal of water treated. Therefore,
the total O&M cost was $0.54/1,000 gal.
                                               12

-------
Conclusions

The Greenville, Wisconsin demonstration project showed that the Macrolite® pressure filtration system
was capable of reducing arsenic concentrations in source water from around 6 ug/L to less than 0.5 ug/L
with iron levels in the source water of around 2 mg/L. With the addition of HMO, the filtration system
also was found effective in reducing total radium to below 5 pCi/L using a proper HMO dosage. If the
proper dosage  is not maintained, radium levels in the system effluent can increase above the radium MCL
of 5 pCi/L. Added loadings from HMO addition can cause solids to breakthrough from the filters
prematurely, making it difficult to maintain sustainable system operation. The treatment system did not
appear to have any significant detrimental impact on the levels of copper or lead in distribution water.

References

Battelle. 2004. Revised Quality Assurance Project Plan for Evaluation of Arsenic Removal Technology.
Prepared under Contract No. 68-C-00-185, Task Order No. 0029, for U.S. Environmental Protection
Agency, National Risk Management Research Laboratory, Cincinnati, OH.

EPA. 2002. Lead and Copper Monitoring and Reporting Guidance for Public Water Systems.
EPA/816/R-02/009. U.S. Environmental Protection Agency, Office of Water, Washington, D.C.
                                              13

-------