DDW-2024-014079

GREEN RIVER
FILTER SKID PILOT
STUDY
Green River Water
Treatment Plant

Green River, Utah

Filter Pilot Skid Operations

January 2024

AE2S Project #: P16019-2023-001 020

P16019-2023-001 Page i
GREEN RIVER FILTER SKID PILOT STUDY
Green River Water Treatment Plant
For
Town of Green River, Emery County, Utah

January 2024
Professional Certification
I hereby certify that this report was prepared by me or under my direct supervision and that I am
a duly Registered Professional Engineer under the laws of the State of Utah.

Name: Marie Owens, PE
Company: Advanced Engineering and Environmental Services, Inc. (AE2S)
Date: Registration Number:

Prepared By:
Advanced Engineering and Environmental Services, Inc. (AE2S)
3400 Ashton Blvd. Suite 105
Lehi, UT 84043

P16019-2023-001 Page ii
Table of Contents
1. INTRODUCTION AND BACKGROUND 3
1.1 Project Background ................................................................................................................................................................. 3
1.2 Alternative Treatment Strategies ....................................................................................................................................... 4
1.3 Data Used in This Study......................................................................................................................................................... 5
1.4 Pilot Study Objectives ............................................................................................................................................................ 6
2. PILOT STUDY PROTOCOL 7
2.1 Equipment and Setup ............................................................................................................................................................. 7
2.1.1 Filter Pilot Skid ...................................................................................................................................................... 7
2.1.2 Filter Media and Filter Arrangement ............................................................................................................ 8
2.1.3 Project Phases ....................................................................................................................................................... 9
2.2 Data Collection .......................................................................................................................................................................... 9
3. WATER QUALITY DATA RESULTS & DISCUSSION 11
3.1 Organic Samples .................................................................................................................................................................... 11
3.1.1 TOC Removal ..................................................................................................................................................... 12
3.1.2 Bioactivity ............................................................................................................................................................ 15
3.2 Disinfection Byproducts (DBPs) ....................................................................................................................................... 17
3.2.2 Immediate DBPs Present ................................................................................................................................ 17
3.2.3 Formation Potential of DBPs ........................................................................................................................ 18
3.2.4 Prechlorinating Treatment Residuals ........................................................................................................ 20
3.3 Turbidity and UV254 ............................................................................................................................................................ 21
3.4 Headloss ................................................................................................................................................................................. 23
4. RECOMMENDATIONS 25
4.1 Media Type and Bed Characteristics ............................................................................................................................. 25
4.2 Media Replacement ............................................................................................................................................................. 26
4.3 Process Optimization ........................................................................................................................................................... 26
4.4 Summmary ............................................................................................................................................................................... 27
5. ACKNOWLEDGEMENTS 29
6. APPENDICES 30
6.1 Appendix A – GRWTP Compiled WQ Data 2021 - 2023 ........................................................................................ 30
6.2 Appendix B – Independent Laboratory Water Sample Data ................................................................................ 33
6.3 Appendix C – Additional Study Figures........................................................................................................................ 35
6.4 Appendix D – Previous Reference Report ................................................................................................................... 37

Green River Filter Skid Pilot Study
Pilot Study Final Report
Introduction and Background
January 2024

P16019-2023-001 Page 3
1. Introduction and Background
1.1 Project Background
The Green River Water Treatment Plant (GRWTP) treats nearly 500,000 gallons per day (gpd)
from the Green River and serves around 900 customers. The water quality of the Green River
varies throughout the year due to natural hydrologic cycles. As shown in Figure 1.1, the facility is
a conventional treatment process consisting of coagulation, flocculation, sedimentation, dual
media filtration, and chlorine disinfection.

Figure 1.1: Schematic of the process flow diagram (PFD) of GRWTP.
The GRWTP has two separate, but identical, treatment trains of 500 gpm capacity each. There is
a common rapid mix where potassium permanganate is added early to address taste and odor
and alum is used as the primary coagulant. Polymer coagulant and flocculant aids are also
utilized. The plant flow is then split into the two trains with three-stage tapered flocculation and
tube settlers. Each filter is tied directly to its preceding pre-treatment basins and consists of a
layer of sand and a layer of anthracite coal. The operations team has found that the filters
produce lower turbidity effluent water if they feed chlorine at the filter influent. There is no
instrumentation to ensure the dose or equalize the chlorine feed between the two filters. An
onsite sodium hypochlorite generator provides the chlorine dose for both the pre-filter as well
as the clearwell location treatments. The required contact time (CT) credit is achieved entirely in
the clearwell, and the necessary secondary disinfection requirement is met by the residual
chlorine leaving the clearwell before the distribution system.
Green River has had reoccurring events where the disinfection by-product (DBP) levels within
the distribution system rise above the regulatory limits of 80 ug/L for trihalomethanes (THMs)
and 60 ug/L of haloacetic acids (HAAs). DBPs are formed over time whenever the precursors of
total organic carbon (TOC) are in the presence of chlorine. The existing process filters remove
enough TOC to reach concentrations around 2.7 mg/L and lower the turbidity to about 0.1 NTU.
The largest concerns of the facility surround the formation of DBPs throughout their distribution
system that tend to rise out of compliance during the spring and autumn seasons due to
Green River Filter Skid Pilot Study
Pilot Study Final Report
Introduction and Background
January 2024

P16019-2023-001 Page 4
increases in the Green River TOC levels entering the GRWTP, increases in chlorine dosing to
achieve secondary disinfection requirements, and increases in travel times in the system due to
low flow seasons. The DBPs leaving the facility’s clearwell on site are typically between 10 to 50
mg/L for HAAs and 10-60 mg/L for THMs.
Green River needs a way to reduce TOC concentrations entering the clearwell to reduce the
formations of DBPs throughout the distribution system and stay in compliance with federal
maximum contaminant levels (MCLs). This pilot study was designed to optimize TOC removal
strategies by various granular activated carbon (GAC) filtration medias, bed depths, contact
times, and loading rates.
1.2 Alternative Treatment Strategies
GAC filtration is the recommended strategy for DBP formation control for the City of Green
River. After discussions with multiple stakeholders, a few different treatment options were
eliminated to render GAC as the strongest candidate. Table 1.1 below highlights takeaways from
the feasibility of other treatment options.
Table 1.1: Alternative treatment strategies discussed with the State of Utah and the City.
Treatment Strategy Conclusion
Ozonation • Green River does not have the personnel to fine-tune and facilitate
the integration of this system
• Similar in cost to GAC filtration and no pilot study to define design
criteria
Ozonation and GAC
filtration in series
• Green River does not have the personnel or budget to manage two
new processes
• Though very effective in process, Green River does not need this kind
of removal to remain in EPA compliance
Air-Stripping for
DBPs
• The current storage tanks in Green River are not strategically placed
to effectively remove DBPs
• New storage tanks would need to be built and a hydraulic analysis
would need to be performed to place them
• Air-stripping would cause problems with chlorine residuals and need
strict monitoring to manage
Chloramination • Green River does not have the personnel to facilitate this transition
• State concerned about the transition due to water chemistry changes
in the distribution system
Green River Filter Skid Pilot Study
Pilot Study Final Report
Introduction and Background
January 2024

P16019-2023-001 Page 5
1.3 Data Used in This Study
GRWTP collects TOC and DBP water samples each calendar quarter. Since the beginning of
2021, there have been 4 samples that fall out of compliance and 5 other samples within 20% of
federal regulations (60 ug/L for HAAs and 80 ug/L for THMs). Times where DBP concentrations
fell out of compliance correlated well with springtime bumps in TOC concentrations from the
runoff season of the Green River. DBP concentrations also spike around lower autumn flows
before water temperatures drop significantly, though TOC was not as inflated this time of year
as in the spring. Therefore, TOC removal is essential to mitigating DBP formations and staying
within compliance regulations. Other factors such as water age also greatly affect the formation
of DBPs but were not included as a focus of this pilot study.

Figure 1.2: Compliance DBP and TOC data for the Green River System. WS001 is the facility inlet,
TP001 is the facility effluent, MR001 for middle of the distribution system, and MD001 for the far
end of the distribution system.
A previous study conducted a DBP Mitigation Study in 2021 (Appendix D) on the success of GAC
medias in the removal of TOC from the city’s water source in comparison to alternative
treatment options such as ozonation and enhanced coagulation. The main points from this
report concluded; GAC is the optimal choice for Green River, an effective bed contact time
(EBCT) of 10 minutes and 20 minutes showed no significant difference in TOC removal,
Green River Filter Skid Pilot Study
Pilot Study Final Report
Introduction and Background
January 2024

P16019-2023-001 Page 6
chlorination before the GAC media did not significantly affect removal, and an estimated time to
media exhaustion was 4 months.
1.4 Pilot Study Objectives
The objective of this 12-Week filter skid pilot study is to evaluate the effectiveness of GAC
filtration to reduce DBP formations as well as the type of GAC media and some key design
parameters (EBCT and a media replacement schedule). These parameters will inform a GAC filter
design that will address Green River’s compliance issues with DBP formations through their
distribution system. The piloting data focuses on water quality goals such as the removal of DBP
precursors (TOC), reduced turbidity, and enhanced taste/odor removal while monitoring other
parameters of pH, bioactivity, and temperature.

Green River Filter Skid Pilot Study
Pilot Study Final Report
Pilot Study Protocol
January 2024

P16019-2023-001 Page 7
2. Pilot Study Protocol
2.1 Equipment and Setup
2.1.1 Filter Pilot Skid
The pilot filter skid used for the 12-Week study was designed by Intuitech. A 3,000-gallon tank
was placed in line behind the existing process filters and before the skid to maximize pilot run
time due to the intermittent operation of the existing facility. Figure 2.1 shows the process flow
within the facility after pilot skid installation. Sampling locations are strategically placed to
identify where chemical transformations occur.

Figure 2.1: Process flow diagram with pilot skid installation and blue tags for sampling locations.
The filter skid consists of four 6-inch columns that can be operated independently or in
customized series (Figure 2.2). Each column measures flow rate, headloss, pH, temperature, and
turbidity every five minutes. The automated air scouring backwash system consists of one
blower pump with customizable backwash flow rates and times. Chemical feed ports are
available, though were not used during this study. Sample ports are located along the length of
each column for convenient sampling at various bed depths and a sample-port box on the side
of the skid for effluent stream sampling.
Green River Filter Skid Pilot Study
Pilot Study Final Report
Pilot Study Protocol
January 2024

P16019-2023-001 Page 8

Figure 2.2: Intuitech filter pilot skid in place at GRWTP.
Online measurements were taken every five minutes during skid operation. Each column was run
independently and in parallel to compare media types and bed depths.
2.1.2 Filter Media and Filter Arrangement
Three media types and two media depths were tested in this study (Table 2.1). Calgon F400 was
tested at two depths to estimate how media depth impacts TOC removal and alters media
breakthrough timelines. Calgon F400 is bituminous-activated and provides a relatively large
micropore structure that fosters TOC removal. Norit HydroDarco 4000 is lignite-activated and
provides a larger micropore structure that may enhance TOC removal. Jacobi Aquasorb CS
media is derived from coconut shells instead of coal and provides a smaller micropore structure
than both the Calgon and Norit medias selected, though may provide a greater number of
channels for more effective TOC removal.
Table 2.1: Outline of media specifications for each column.

Green River Filter Skid Pilot Study
Pilot Study Final Report
Pilot Study Protocol
January 2024

P16019-2023-001 Page 9
2.1.3 Project Phases
Three distinct loading rates were tested to compare the effects of EBCT on TOC removal and
effluent turbidities for each media type. These loading rates correlated to flow rates of 1.00
GPM, 1.75 GPM, and 0.80 GPM, chronologically (Table 2.2).
Table 2.2: Pilot Study testing schedule.
Week of Study Loading Rate 48” Column EBCT 72” Column EBCT
1
5.1 gpm/ft2 6 minutes 9 minutes 2
3
4
5
8.9 gpm/ft2 3 minutes 5 minutes 6
7
8
9
4.1 gpm/ft2 7 minutes 11 minutes 10
11
12

2.2 Data Collection
Daily water samples were collected to monitor general water chemistry characteristics and
leverage weekly laboratory samples for TOC removals and DBP concentrations. Formation
potentials tests to replicate the formation of DBPs in Green River’s distribution system were also
conducted. Locations described in Figure 2.1 were used for sampling to assess water chemistry
throughout the plant and through the filter skid effluent streams. As shown in Table 2.3, daily
samples and online measurements are designed to inform differences and discrepancies in
laboratory sampling across columns and weeks. The GRWTP staff assisted in collecting these
samples and monitoring skid operations daily. Laboratory samples were taken to Chemtech-
Ford Laboratory in Sandy, Utah and were collected by AE2S personnel (Table 2.4).
Measurements for pH, DO, temperature, and HPC inform the presence and productivity of
biological agents in each column. UV254 was collected to determine the appropriateness of it as
a surrogate for TOC and turbidity is both regulated and is a measure of the aesthetic quality of
the water. Free chlorine helps to identify areas where DBPs can form in the facility. TOC, DOC,
HAA and THM are all quantitative water quality parameters with MCL and treatment technique
requirements. The formation potentials of HAA and THM were 7-day resting tests at ambient
Green River Filter Skid Pilot Study
Pilot Study Final Report
Pilot Study Protocol
January 2024

P16019-2023-001 Page 10
temperature in the presence of excess chlorine to explore the potential for DBP formation
through the distribution system. On the final week, a ‘simulated distribution system” (SDS) test
was taken, which provides a perspective on what TOC concentrations can leave the plant to keep
DBP formations below compliance regulations across the water service area. This SDS test
replicated real water conditions leaving the facility with a chlorine residual around 2 mg/L and a
water age at the far end of the distribution system estimated at 5 days.
Table 2.3: Water quality parameters and their corresponding analysis types.

Table 2.4: Weekly and biweekly laboratory sample location collection. Data was not collected in
blacked out regions.

Green River Filter Skid Pilot Study
Pilot Study Final Report
Water Quality Data Results & Discussion
January 2024

P16019-2023-001 Page 11
3. Water Quality Data Results & Discussion
The Calgon media showed the best removal of TOC for the majority of the 12-Week study. The
Norit media also performed strongly at TOC removal. The Jacobi media did not prove to be
effective on the specific TOC and water quality. Calgon media was tested at two depths and a
50% increase in bed depth resulted in about a 35% increase in TOC removal near the beginning
of the media lifetime, though demonstrated a comparable removal at higher exhaustions. The
increased performance was attributed to the presence of additional adherence sites in the GAC
micropores due to more particles being present in a larger bed volume and the extended EBCT
associated with this larger bed volume. Table 3.1 shows the number of bed volumes until two
different exhaustion criteria based on data from Figure 3.2.
Table 3.1: The number of bed volumes treated in relation to corresponding exhaustion criteria.

Both Calgon and Norit medias are very responsive to changes in loading rate due to the related
changes in EBCT. As contact time with the media increases, removal has more time to take place.
The best TOC removal and DBP formation potential reduction was seen with a loading rate of
4.1 gpm/ft2, associated with the longest EBCTs for each column (11 minutes for the 72” column
and 7 minutes for all 48” columns). EBCT was found to be the main variable in controlling TOC
removal and DBP formation potential reduction. Each media reduced DBP formations for both
HAAs and THMs. Appendix A provides laboratory tested compliance values for Green River,
Appendix B provides weekly laboratory water sample testing throughout the study duration,
Appendix C provides additional study figures and data, and Appendix D provides the preliminary
report used as a reference for this study.

3.1 Organic Samples
TOC concentrations are a controllable driver of DBP formation. The pilot study showed a direct
correlation between an increase in TOC removal with the reduction of DBP formation. This is, in
fact, a causal relationship, though not a necessarily a one to one relationship, since other
variables also impact the formation of DBPs from TOC and chlorine. The results confirm that
Green River can and should use the TOC removal data to monitor the exhaustion of the GAC
media and determine when it needs to be replaced or regenerated.
Green River Filter Skid Pilot Study
Pilot Study Final Report
Water Quality Data Results & Discussion
January 2024

P16019-2023-001 Page 12
3.1.1 TOC Removal
DOC was analyzed along with TOC on the raw water to confirm that it is similar enough to
disregard the difference. The comparison was conducted through the entire 12-week study with
consistent similarity. TOC removal was successful in Columns 1, 2, 3, where Calgon and Norit
media were used. Jacobi media was successful in removing some DBPs that formed in the
existing process filters and improving turbidity, though removed little to no TOC and resulted in
high DBP formation potentials as shown in Figure 3.1. Due to this, Jacobi is not recommended as
an acceptable media choice.
Three loading rates were evaluated throughout the course of the study with three
corresponding EBCTs. EBCT is integral to giving the media enough time to remove TOC and
other constituents from the influent water. A previous study from 2021, found in Appendix D,
evaluated EBCTs of 10 and 20 min, rendering comparable TOC removal results. However, this
pilot study evaluated shorter EBCTs to give Green River further flexibility in how they choose to
operate the proposed GAC facility. The 72” Calgon media column experienced EBCTs of 5, 9, and
11 minutes. The other 48” columns with Calgon, Norit, and Jacobi media experienced EBCTs of 3,
6, and 7 minutes. There were significant differences in the TOC removal rates and subsequent
breakthrough to TOC across the GAC filters. Figure 3.1 shows these results for each pilot column
and at each of the tested loading rates (reflecting the EBCT values in Table 2.2). Shorter EBCTs
allow for reduced filter footprints and lower material costs, though can overwhelm the media,
and render lower TOC removals. The combination of high loading rates, low EBCTs, and a
uniquely small micropore structure of Jacobi media explain its poor performance.

Figure 3.1: TOC removal in relation to days of 8-hour average operation at 7 gpm/ft2.
Green River Filter Skid Pilot Study
Pilot Study Final Report
Water Quality Data Results & Discussion
January 2024

P16019-2023-001 Page 13
The data in Figure 3.1 trends well with the previous study data from the performed RSSCTs.
“TOC in” is defined as the influent to the pilot skid (effluent of the existing process filters) and
“TOC out” is defined as the effluent of the pilot skid filters. Table 2.2 can be used for identifying
the corresponding EBCTs to each loading rate tested. Table 3.2 is derived from each EBCT for
each column of TOC data to highlight the importance of EBCT on media exhaustion rates. By
extrapolating each TOC removal trend in Figure 3.1, the number of days to reach a defined
media exhaustion is possible. In reality, columns where the number of days to exhaustion
decreases as exhaustion increases are near horizontal trends (4.1 gpm/ft2 data for both Calgon
and Norit).
Table 3.2: Media exhaustion durations Calgon and Norit medias.

The media lifetimes in Table 3.2 are based on an 8-hour operational day. The 4.1 gpm/ft2 data
for Calgon rests around 80% exhaustion (which continues to be true up to 560 days), whereas
Norit TOC removal during this time rests just under 90% exhaustion (which continues to be true
up to 450 days). The timelines listed in Table 3.2 indicate that the media can last 1-2 years with
EBCTs greater than 7-minutes. Figure 3.2 reflects approximate media lifetimes by averaging out
TOC removal data from Figure 3.1 across loading rates.

Figure 3.2: Bed volume estimates in relation to TOC breakthroughs.
Green River Filter Skid Pilot Study
Pilot Study Final Report
Water Quality Data Results & Discussion
January 2024

P16019-2023-001 Page 14
Figures 3.2, Table 3.1 and Table 3.2 demonstrate the importance of media replacement and
determining an appropriate exhaustion limit. The referenced previous study found that 60% TOC
exhaustion took place around roughly 8,000 bed volumes. This is a longer media lifetime than
the results of this pilot study due to longer EBCTs. The 72” Calgon column at a loading rate of
4.1 gpm/ft2 correlated the closest to a previously conducted report, likely due to having the
longest EBCT tested (11 minutes). Although strong TOC removal (below 60% exhaustion) is
exhausted by roughly 5,000 bed volumes in Figure 3.2. Section 3.2 discusses appropriate
exhaustion criteria for this application and the relationship between TOC removal and DBP
formation potential reduction. Timing media replacements to consider TOC, HAA, and THM
increases in the spring and fall as shown in Figure 1.2 will be essential to maximizing media
lifetime in the facility and minimizing costs to Green River.
In order to remove around 15% of TOC, Green River would need to remove on average, 545
grams TOC per day. By using weekly TOC removal masses shown in Figure 3.3 and comparing to
the amount of TOC removal required, the volume of media required could be determined for
the final design. Both the Calgon and Norit medias required nearly 460 cubic feet of media to
meet this requirement. At higher exhaustions, bed depth does less to enhance removal and
Norit and Calgon medias compete closely.

Figure 3.3: TOC mass removed versus media exhaustion levels.
Figure 3.4 shows the cumulative mass of TOC removed during the study in relation to media
exhaustion levels. Calgon media can remove the greatest mass of TOC with the lowest
exhaustion levels. The 72” and 48” Calgon columns removed a total of 450 grams and 345 grams
of TOC, respectively. The 48” Norit and Jacobi columns removed a total of 225 grams and 60
grams of TOC total.
Green River Filter Skid Pilot Study
Pilot Study Final Report
Water Quality Data Results & Discussion
January 2024

P16019-2023-001 Page 15

Figure 3.4: Cumulative mass of TOC removed in relation to media exhaustion levels.
EBCT is another important factor in filter size and the greater of the two criteria (media volume
required and EBCT) should be used for design. The best performing EBCT for the 48” columns
was 7 minutes, and therefore a bed depth of 7.5 feet total or a volume of 590 cubic feet total is
needed to reach an EBCT of 8 minutes to ensure a balance between adequate TOC removal and
filter footprint. Between the 48” and 72” Calgon columns tested, a 50% increase in EBCT due to
the altered bed depths related to a 30% increase in total TOC removal (Figure 3.4).
3.1.2 Bioactivity
Biological communities in GAC filters are known to be very productive in water treatment. HPC
samples showed bioactivity in the water of each column by the second week of the study. Figure
3.5 shows the correlation between water temperatures and bioactivity in the water during the
12-weeks. Numbers above 1000 CFL/mL are considered to demonstrate the presence of an
active community. Around October 11th, water temperatures dropped beneath 16oC in the
facility and the biological communities experienced about a 90% decrease in concentration over
the next week.
Green River Filter Skid Pilot Study
Pilot Study Final Report
Water Quality Data Results & Discussion
January 2024

P16019-2023-001 Page 16

Figure 3.5: Water HPC and temperature over time.
HPC tests measure the presence of biological communities in water samples rather than media
samples. From this, it cannot be explicitly concluded that the GAC media completely lost the
bioactivity during the study as temperatures dropped, due to potential biofilm formation and a
reduction in biological sluffing from the media that may be taking place instead. Instead, the
HPC data reflects the quick recruitment of microorganisms at the beginning of the study (within
1-2 weeks). There is no evidence suggesting that the presence of the microorganisms negatively
affected TOC removal rates. Additionally, bituminous and lignite activated carbon medias are
marketed as having micropore structures that are large enough to accommodate biological
growth without blocking pore openings from TOC removal.
Dissolved oxygen (DO) concentrations are directly related to flow turbulence and indirectly
related to both bioactivity and temperature. Temperature is a large driver due to the solubility of
gases in liquid systems increasing as temperatures decrease. As bioactivity and water
temperatures decrease, an increase in DO was seen as shown in Figure 3.6. DO concentrations in
the filter skid columns were likely lower than in the plant due to slower moving water and their
containment in columns. These results indicate a biological presence in the column and the lack
thereof in the existing facility. The outlier data points are likely a result of sample recording
taking place after columns were out of service for a significant portion of time, reducing
available oxygen transport in the water/media.
Green River Filter Skid Pilot Study
Pilot Study Final Report
Water Quality Data Results & Discussion
January 2024

P16019-2023-001 Page 17

Figure 3.6: Dissolved oxygen and temperature over time.
3.2 Disinfection Byproducts (DBPs)
The facility’s main concern is the formation of DBPs throughout the distribution system. DBP
measurements were sampled on a biweekly basis to monitor the ability of each media in
removing DBPs, and to verify that concentrations within the plant were not greatly varying
throughout the length of the study. The facility currently prechlorinates at their process filters to
manage filter effluent turbidity. However, this begins the DBP formation within the plant before
reaching the clearwell or storage tanks.
3.2.2 Immediate DBPs Present
The facility currently forms about 2-4 µg/L of HAAs and 5-10 µg/L of THMs across their process
filters with only a 38-minute contact time from the prechlorine feed location to the filter effluent
as shown in Figure 3.7. The compliance DBP data sown in Figure 1.2 confirms that the facility
produces around 20 ug/L of HAAs and 30 µg/L of THMs at the clearwell effluent before the
water enters the distribution system.
Green River Filter Skid Pilot Study
Pilot Study Final Report
Water Quality Data Results & Discussion
January 2024

P16019-2023-001 Page 18

Figure 3.7: DBPs before and after pilot skid GAC filters.
THM removal was on average 70% by the 72” Calgon column, 40% by the 48” Calgon column,
50% by 48” Norit, and 50% by 48” Jacobi. HAA removal was on average 90% by the 72” Calgon
column, 80% by the 48” Calgon column, 100% by 48” Norit, and 30% by 48” Jacobi. This is
despite the extra time (up to 15 hours) associated with the existing filter effluent stream being
fed into a tank and then routed to the skid.
3.2.3 Formation Potential of DBPs
Measuring the formation potential of DBPs was performed by flooding the sample with an
excess of chlorine (TOC limited) and holding samples in an ambient room for a full seven days.
At the end of this period, DBP concentrations were measured. These samples provide
information on the reduction of DBP formation potentials and can be used to estimate potential
DBP formations that could be encountered in the future with the new GAC filters in operation.

Figure 3.8: Relationship between DBP formation potentials and TOC effluent concentrations.
Green River Filter Skid Pilot Study
Pilot Study Final Report
Water Quality Data Results & Discussion
January 2024

P16019-2023-001 Page 19
Figure 3.8 identifies TOC as a driver of DBP formation, highlighting the importance of additional
TOC removal for Green River. Figure 3.9 displays the effectiveness of GAC filtration in lowering
DBP formation potentials. This data validates the ability of GAC to bring future DBP
concentrations at the longest water age in the distribution system below MCLs. From Figure 1.2,
a minimum of 12% reduction of DBP formations in WTP finished water would prevent all
quarterly DBP compliance samples taken by Green River since 2021 from surpassing DBP MCLs.
From this, staying under an 85% TOC exhaustion criteria during times when Green River
struggles with compliance is a strong indicator that the facility will restrict DBP formations to
stay consistently in compliance. Figure 3.9 shows a formation potential exhaustion curve. Each
logarithmic curve follows closely to the TOC reduction curves shown in Figure 3.1. Figures 3.1
and 3.9 are related through Figure 3.8. With this correlation between DBP formation potential
and the TOC removal to set a maximum media exhaustion standard, we can estimate the time
between media replacement/regeneration to extend beyond 500 days of 8-hour long operation
at the longest EBCTs tested shown in Figure 3.2.

Figure 3.9: The reduction of DBP formation potentials.
During Week 12 sampling, a customized simulated distribution system (SDS) test was performed
with a 5-day holding time and a 2 mg/L addition of chlorine. Using Equation 3.1, these values
were compared to the Week 12 maximum formation potential data to determine a percent
reduction in DBP formation between the maximum potential and the SDS potential. Both the
SDS data values and the “SDS : Maximum” Ratios are shown in Table 3.3. The maximum
formation potential values collected throughout the study are shown in Appendix B.2, and are
based on a 7-day holding time and an excess of chlorine to drive DBP formations to the
maximum. This percentage was then used to adjust the maximum formation potential data for
the final phase of the study to produce calculated values that represent what real DBP
measurements could resemble on the far end of the distribution system in Green River and are
shown in Table 3.4.
"𝑅𝐷𝑅∶𝑀𝑎𝑥𝑖𝑚𝑡𝑚" 𝑅𝑎𝑡𝑖𝑜=𝑆𝐷𝑆 𝑙𝑒𝑎𝑟𝑟𝑟𝑒𝑙𝑒𝑙𝑟
𝑙𝑎𝑥𝑖𝑙𝑟𝑙 𝑒𝑙𝑟𝑙𝑎𝑟𝑖𝑙𝑙 𝑙𝑙𝑟𝑒𝑙𝑟𝑖𝑎𝑙 𝑙𝑒𝑎𝑟𝑟𝑟𝑒𝑙𝑒𝑙𝑟 (Eq. 1)
Green River Filter Skid Pilot Study
Pilot Study Final Report
Water Quality Data Results & Discussion
January 2024

P16019-2023-001 Page 20
Table 3.3: Customized simulated distribution system (SDS) test results and their ratio with the
maximum formation potential data from 12/6/2023.

Table 3.4: Adjusted maximum formation potential values based on “SDS : Maximum” Ratio.

The Calgon F400 media and Norit HydroDarco40000 performed exceptionally and all HAA and
THM values remained under the MCLs and the yellow highlighted cell is the only calculated
value that fell within 10% of the corresponding MCL. The calculated SDS values in Table 3.4 were
produced only for the final third of the study because SDS data was only collected at a loading
rate of 4.1 gpm/ft2 to produce information on a “SDS : Maximum” Ratio. These values indicate
that the bed volumes tested, and days of operation tested are within a reasonable lifetime for
the GAC media to HAA and THM values underneath DBP compliance MCLs. The data presented
in in Table 3.2 corresponds to EBCTs of 11 minutes for the 72” column and 7 minutes for the 48”
columns. When compared to Green River’s compliance data taken on November 21st, 2023
(Table A.1), the plant effluent DBP formations are less than the dual filter effluent DBP
formations shown in Table 3.2, indicating that this SDS test considered longer water ages or
larger chlorine residuals than the true values, and that these numbers are an overestimate of
what is realistically forming in the distribution system. In other words, these are a conservative
estimate for projected DBP concentrations that would exist below the facility with GAC filtration
in place because of the SDS values taken on December 6th, 2023 are significantly higher than the
true DBP concentrations measured by Green River on November 21st, 2023.
3.2.4 Prechlorinating Treatment Residuals
The chlorine residuals leaving the existing process filters have the potential to vary greatly due
to the lack of equipment available to monitor the amount of chlorine added upstream. Figure
3.8 shows the inconsistency of chlorine added during pretreatment. This highlights the variability
of potential DBP formations that could be taking place within the WTP, though are likely limited
due to the relatively short residence time that water experiences within the facility. The chlorine
concentration in the process filters is not directly monitored so the operations team is unable to
Green River Filter Skid Pilot Study
Pilot Study Final Report
Water Quality Data Results & Discussion
January 2024

P16019-2023-001 Page 21
confidently know the amount of chlorine being fed at this location. Process optimization for this
step could take one or more of the following forms:
• Complete removal of chlorination at this point
• A change to the addition of a filter aid instead of the use of chlorine, and
• Direct monitoring system for the chlorine at this feed point.
This highlights an area of the facility that could benefit from process optimization, either from
the removal of this step, or a more direct monitoring system to minimize chlorine usage.
Utilization of these strategies can maintain turbidity removal through the process filters and
minimize DBP formations within the facility. The current operation of the facility utilizes this step
to further reduce turbidities off the process filters.
By measuring free chlorine at both the existing process filter influent and effluent, the fraction of
free chlorine remaining can provide information on the amount of chlorine taken up by
biological communities or transformed into DBPs. Figure 3.10 highlights the variance in residuals
leaving the process filters and displays a process optimization opportunity with tighter control
on free chlorine addition.

Figure 3.10: The fraction of free chlorine residuals remaining after the existing process filters in
relation to the influent.
3.3 Turbidity and UV254
Both turbidity and UV254 measurements play a large role in aesthetic water quality standards.
Green River is required to achieve finished water turbidities below 0.3 NTU, though frequently
produces under 0.1 NTU water throughout the year with the help of coagulation, flocculation,
and prechlorination as shown in Figure 3.11. The settled UV254 values after the existing process
filters range from 0.05-0.10 cm-1.
Green River Filter Skid Pilot Study
Pilot Study Final Report
Water Quality Data Results & Discussion
January 2024

P16019-2023-001 Page 22

Figure 3.11: Turbidity measurements from existing filter effluents.
Turbidities at the existing process filter effluents during the study varied greatly, though each
filter skid column was able to stabilize effluent turbidities to an average of less than 0.05 NTU.
Turbidities on this order will provide enhanced aesthetic drinking water characteristics to users
of the municipal distribution system. Figure 3.12 displays the filter skid effluent turbidities, in
which peaks represent daily start up readings from nightly shutdowns.

Figure 3.12: Turbidity measurements from filter skid effluent streams.
Figure 3.13 shows UV254 absorbance readings for all pilot test sample locations. UV254
correlates with TOC concentrations, allowing Figure 3.13 to provide evidence of the reduced
TOC concentrations seen in the filter skid effluent streams in comparison to the effluent of the
Green River Filter Skid Pilot Study
Pilot Study Final Report
Water Quality Data Results & Discussion
January 2024

P16019-2023-001 Page 23
existing process filters. GAC filter effluent absorbances are lower than existing filter absorbances
and finished water absorbances for the duration of the study to significant TOC removal.

Figure 3.13: UV254 removal performance for pilot columns.
3.4 Headloss
The pilot skid monitored headlosses in each column for the study duration, a parameter integral
to sizing GAC filter pumps. The large headloss changes shown in Figure 3.14 coincide with
updating the loading rates to test various EBCTs. Headlosses around 4 feet will be expected for
the GAC facility based on a slightly lower loading rate than tested in the middle of the study and
a similar bed depth. The 72” Calgon media and the 48” Norit media provided the greatest
headloss in comparison to other columns due to bed depth and particle size, respectively. 48”
Calgon media on average experienced the lowest headloss against the other 48” columns, likely
due to the particle size and bituminous activated structure. Intermittent noise in the middle of
the study shown in Figure 3.14 correlates with the inconsistent operation schedule of the pilot
skid due to the inability to run during the night at the higher loading rates tested.
Green River Filter Skid Pilot Study
Pilot Study Final Report
Water Quality Data Results & Discussion
January 2024

P16019-2023-001 Page 24

Figure 3.14: Column headloss for each media.
Table 3.5: Average column headloss in relation to the selected loading rate
Loading Rate 72” Calgon 48” Calgon 48” Norit 48” Jacobi
4.1 gpm/ft2 1.3 0.8 1.4 1.0
5.1 gpm/ft2 2.1 1.5 2.0 1.6
8.9 gpm/ft2 4.0 2.9 4.3 2.9
From the data in Table 3.5, a loading rate of 7 gpm/ft2 and bed depth of four feet would result
in an average headloss around 2.1 feet for Calgon and 3.1 feet for Norit. These headlosses are
associated with a bed volume of 0.8 ft3 in each 48” column. The differences in headloss across
media types, loading rates and bed depths are critical for informing pump sizing for the GAC
facility.

Green River Filter Skid Pilot Study
Pilot Study Final Report
Recommendations
January 2024

P16019-2023-001 Page 25
4. Recommendations
The 12-Week pilot study successfully addressed the feasibility of GAC filtration for Green River
and clarified appropriate design criteria for a full-scale design. Through online pilot skid
measurements, daily grab sampling, and weekly water chemistry analysis, GAC media
performances were characterized to provide recommended media types, filter bed dimensions
and configurations, and appropriate loading rates.
AE2S recommends two (2) GAC filter units run in a lead-lag configuration.
Each filter is to have no less than 8,000 lbs of media and replacement
should take place for each filter every other year, on alternating years for
this media amount. A bituminous coal GAC media with a particle mesh size
of 12x40 is recommended for bid.
Media replacement schedules can be elongated with greater bed volumes. A minimum EBCT of
8 minutes through the GAC filters is recommended for the entire GAC installation (4 minutes in
each of the two filter vessels). A loading rate below 10 gpm/ft2 is ideal for operation.
Backwashing should be performed once every other week during normal operation, and once a
week during periods of non-operation. Under no circumstances should backwashes be
performed less than half of this frequency. AE2S also recommends a blend of process
optimization strategies which include, but are not limited to, discontinuing prechlorination
before the existing process filters, updating equipment to monitor and manage chemical feeds,
the addition of a filter aid to compensate for the removal of chlorine onto the filters, and the
addition of a buffer to manage pH during the mixing of alum to enhance the coagulation step.
4.1 Media Type and Bed Characteristics
Both Calgon F400 and Norit HydroDarco4000 performed very well in the reduction of DBP
formation potentials, TOC removal, turbidity, and other parameters of interest. The poor
performance documented by Jacobi Aquasorb CS media is likely a result of the smaller
micropore structure associated with coconut shell medias that get blocked by larger TOC
molecules. TOC removals were adequate for nearly 500 days of 8-hour operation for both
Calgon F400 and Norit HydroDarco4000, indicating that media replacement schedules around
every two (2) years would be sufficient for the proposed bed volume. Additionally, Norit GAC300
is a bituminous-activated media with a similar micropore structure and particle size to Calgon
F400. From this, it is expected to perform similarly to Calgon F400, and would be acceptable for
this application. The best performing media for this application is bituminous coal based with a
particle mesh of 12x40.
The longest EBCTs tested provided the best TOC removal and longest exhaustion rate for each
media. An EBCT of 8 minutes would stay close within the 10 to 20 minute EBCTs of the Brown &
Caldwell study while minimizing filter sizes, minimizing media costs, and maintaining DBP
compliance. An 8-minute EBCT relates to 600 ft3, or 8,000 lbs of media. To extend media lifetime,
operators can alternate filter replacements so the filter with the fresh media is placed in
Green River Filter Skid Pilot Study
Pilot Study Final Report
Recommendations
January 2024

P16019-2023-001 Page 26
operation downstream (lag) to the filter with the older media (lead). This prevents excessive
removal during the early use of the media, saving pore space for better long-term removal.
Additionally, lead-lag configurations also provide redundancy and integrity to a system. Each
filter can be run independently as well as in series to facilitate maintenance needs and media
replacement schedules.
Biological activity was able to establish within a matter of weeks in each filter, though
temperatures below 16oC were too cold to support microbial communities. Biological presence
was not needed in the filters to have successful TOC removals but will enhance that removal if
present and managed. During seasons where GAC filters are not operated, it is recommended
that filters be run for two-hours once a week to prevent stagnation in filters and overgrowth.
4.2 Media Replacement
Media lifetimes around 500 days provide roughly two-years of use before needing replacement.
By splitting the design into two filters to maximize operational flexibility, it is recommended that
each filter be replaced once every other year, though alternating each year so that the facility is
never run with two fresh media filters or have the potential for both filters to have exhausted
media at the same time. Media replacement should only happen before the spring increase in
raw water TOC or the late summer increase in raw water TOC to ensure that the freshest media
combats the times with highest TOC concentrations. GAC filtration is not required during the
winter months to maintain compliance with DBP MCLs. A more frequent backwash schedule of
once a week is recommended if the operations team chooses to take the GAC filters offline to
extend the media life. This increased backwash schedule will manage any present biologic
growth and prevent biofouling of the GAC filters when not in use.
Each pressure filter is expected to cost $16,000 to replace 8,000 lbs of Calgon F400 media or
Norit GAC 300 based on a $2.00 per pound estimate. The same amount of Norit
HydroDarco4000 media is expected to cost $24,000 due to a $3.00 per pound estimate. By
replacing one filter a year, a budgeted annual media replacement cost of $20,000 would be
reasonable to account for media cost inflations and potential increases in media bed depths if
needed.
As media amounts in each filter increase, media lifetimes increase as well. Updated media
replacement schedules should be reevaluated if the chosen media operational amounts are
greater than 8,000 pounds in each filter.
4.3 Process Optimization
Though GAC filtration is a successful method to treat DBP formations, process optimization will
assist in improving water treatment capabilities of the GRWTP and extending media lifetime.
Process optimization recommendations come in the form of updated equipment and refining
existing process steps.
Prechlorinating before process filters has shown to be an effective technique in reducing
turbidity readings, though results in higher ultimate DBP formation by allowing them to form
Green River Filter Skid Pilot Study
Pilot Study Final Report
Recommendations
January 2024

P16019-2023-001 Page 27
earlier and before a portion of TOC can be removed. Online chlorine monitoring capabilities in
the facility are limited to the finished water tap. This location is used to manage the chlorine
addition before the existing process filters. Using this location to manage a process far upstream
is complicated and has reduced monitoring power. Removing or amending this step would be
integral to ensuring DBP compliance. Amending the prechlorination step can come in the form
of a complete removal of the prechlorination step, substitution with an alternative filter aid,
and/or a reduction and heightened monitoring of prechlorination. The resulting drop in
turbidity removal from the process filters will be compensated for by the GAC filters.
Chemical feed pumps are outdated diaphragm pumps that require excessive maintenance to
minimize variance in chemical addition. The ability to further manage and monitor these
chemical feed systems will enhance facility operations.
Alum, the primary coagulant chemical for the plant, is controlled by one of these outdated
diaphragm pumps. Alum performs the best when used below a pH of 7. The facility currently
experiences raw water with a pH of 8.5 and the treated water retains a pH of 7.5. This provides
an opportunity to enhance coagulation and flocculation of suspended solids by adjusting the pH
at or before the rapid mix. This would help with overall TOC removal in the existing plant.
Lowering facility flows during spring and early fall TOC spikes is another strategy to improve
TOC removal from the existing process filters by increasing the EBCT in the GAC filters. Although
this strategy would increase plant operation hours to maintain the necessary daily plant
production, it may be a useful tool in conjunction with GAC filtration to ensure compliance with
DBPs.
Replacing or chemically cleaning the existing anthracite media is an additional optimization
strategy that may play a role in improving TOC removal and mediating DBP formation
potentials. Anthracite media has a recommended media replacement schedule of every 10 to 20
years. The existing media has been in place since 1999. Chemical cleaning is possible when
anthracite is still within original purchased specifications for particle sizes, though is fouled
through a combination of inadequate backwash rates, uneven backwash flow distribution, or
age.
4.4 Summary
This pilot study has informed many recommendations for GRWTP DBP compliance. Each
recommendation should be evaluated based on cost and impact on heightened water quality.
The recommendations discussed in this section include:
• Installation of a GAC pressure filtration system
o Lead-lag skid system with a total EBCT of 8 minutes at minimum through
both filters together
o Bituminous coal granular activated carbon media in a 12 x 40 mesh size
Green River Filter Skid Pilot Study
Pilot Study Final Report
Recommendations
January 2024

P16019-2023-001 Page 28
o Backwash schedule of every two-weeks during operation and once a week
when filters are offline
• Modification of current prechlorination step
o Removal of this step, reduction of chlorine addition, substitute with an
alternative filter aid, or heightened monitoring
• More accurate chemical additions as a result of updating older diaphragm
chemical pumps
• A pH adjustment to heighten effectiveness of alum
• Lower facility flow rates during periods of increased DBP formations in the spring
and fall
• Replacement of current anthracite media

Green River Filter Skid Pilot Study
Pilot Study Final Report
Acknowledgements
January 2024

P16019-2023-001 Page 29
5. Acknowledgements
Lastly, Advanced Engineering and Environmental Services LLC. (AE2S) would like to thank Green
River’s involvement in this study, though both the recognition of the importance of piloting and
the day-to-day assistance in skid operation and data collection. AE2S would like to thank
Johansen & Tuttle for their integral role as consultant to Green River, as they have worked hard
to coordinate well with us in our desire to provide the best process design. Additionally, AE2S
would like to thank the State of Utah and the Division of Drinking Water for their commitment
to providing clean and safe drinking water to all communities across the state, and their role in
guiding our pilot study objectives for a comprehensive and long-lasting design.

Green River Filter Skid Pilot Study
Pilot Study Final Report
Appendices
January 2024

P16019-2023-001 Page 30
6. Appendices

6.1 Appendix A – GRWTP Compiled WQ Data 2021 - 2023
Table A.1: Compiled TOC and DBP data for GRWTP from 2021 to 2023. For DBP samples, yellow
cells designate values within 20% of the MCL and red cells designate values above the MCL.

Green River Filter Skid Pilot Study
Pilot Study Final Report
Appendices
January 2024

P16019-2023-001 Page 31

Figure A.2: Daily flow rates through facility for January 2021 to January 2024. Total facility flow
increased to about 550 GPM in July 2023.

Figure A.3: UV254 from the dual media influent location (02). This is the settled UV254 at the top of
the existing process filters.
Green River Filter Skid Pilot Study
Pilot Study Final Report
Appendices
January 2024

P16019-2023-001 Page 32

Figure A.4: Raw water and finished water pH measurements. Coagulation and flocculation additives
lower pH about 1 pH unit.

Green River Filter Skid Pilot Study
Pilot Study Final Report
Appendices
January 2024

P16019-2023-001 Page 33
6.2 Appendix B – Independent Laboratory Water Sample
Data
Table B.1: DBP concentrations for filter skid influent and effluent streams.

Table B.2: Maximum DBP formation potentials dosed with an excess of chlorine and a 7-day
holding time for filter skid influent and effluent streams.

Green River Filter Skid Pilot Study
Pilot Study Final Report
Appendices
January 2024

P16019-2023-001 Page 34
Table B.3: Dissolved organic carbon (DOC) concentrations for all sample locations.

Table B.4: Total organic carbon (TOC) concentrations for all sample locations.

Table B.5: Heterotrophic plate count (HPC) values for filter skid influent and effluent streams.

Green River Filter Skid Pilot Study
Pilot Study Final Report
Appendices
January 2024

P16019-2023-001 Page 35
6.3 Appendix C – Additional Study Figures

Figure C.1: TOC Removal in relation to bed volumes of water treated. Trendlines for each column
are split up into three different regions that correlate to loading rates of 5.1 gpm/ft2, 8.9 gpm/ft2,
and 4.1 gpm/ft2, moving left to right.

Figure C.2: UV254 daily samples collected throughout study duration.
Green River Filter Skid Pilot Study
Pilot Study Final Report
Appendices
January 2024

P16019-2023-001 Page 36

Figure C.3: Free chlorine daily samples collected throughout study duration. Most process filter
influent data was collected for each filter train independently (Filter 1 Media Influent and Filter 2
Media Influent), though the first week of data was collected as a composite sample (Dual Media
Influent).

Figure C.4: pH daily samples collected throughout study duration.
Green River Filter Skid Pilot Study
Pilot Study Final Report
Appendices
January 2024

P16019-2023-001 Page 37
6.4 Appendix D – Previous Reference Report

SEE ATTACHED BELOW.

Technical Memorandum

Limitations:
This document was prepared solely for the City of Green River (City) in accordance with professional standards at the time the services were
performed and in accordance with the contract between the City and Brown and Caldwell dated January 29, 2020. This document is governed by the
specific scope of work authorized by the City; it is not intended to be relied upon by any other party except for regulatory authorities contemplated by
the scope of work. We have relied on information or instructions provided by the City and other parties and, unless otherwise expressly indicated,
have made no independent investigation as to the validity, completeness, or accuracy of such information.

6955 Union Park Center
Suite 270
Midvale, UT 84047

Phone: [phone]
Fax: [phone]

Prepared for: City of Green River
Project Title: Water System Analysis
Project No.: 154789
Technical Memorandum (No. 4)
Subject: Summary Report for Bench-scale Pre-ozonation and Granular Activated Carbon (GAC) Rapid
Small-Scale Column Test (RSSCT)
Date: December 15, 2021
To: Bryan Meadows, Public Works Director, City of Green River
From: Adam Jones, P.E., Brown and Caldwell

Prepared by: Rosa Yu, Ph.D., Brown and Caldwell

Reviewed by: Adam Jones, P.E., Brown and Caldwell
Laurie Sullivan, P.E., Brown and Caldwell

Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

ii
Use of contents on this sheet is subject to the limitations specified at the beginning of this document.
TM4 Bench Pre-ozonation and RSSCT Testing FINAL
Table of Contents
List of Figures ............................................................................................................................................................ ii
List of Tables ............................................................................................................................................................. iv
Section 1: Background Introduction ........................................................................................................................ 1
1.1 Green River Water Treatment Plant .................................................................................................................. 1
1.2 Advanced Water Treatment Processes for DBP Control .................................................................................. 2
1.3 Bench-scale Treatability Testing ....................................................................................................................... 3
Section 2: Jar Testing Results.................................................................................................................................. 3
2.1 Jar Test Objectives ............................................................................................................................................. 3
2.2 Jar Test Matrix .................................................................................................................................................... 4
2.2.1 Raw Water Characterization ............................................................................................................. 4
2.2.2 Jar Test of Alternative Coagulants and Coagulant Doses ............................................................... 4
2.3 Jar Test Results and Discussions ..................................................................................................................... 5
2.3.1 Raw Water Characterization ............................................................................................................. 5
2.3.2 Alternative Coagulants and Coagulant Doses ................................................................................. 6
Section 3: Pre-ozonation Testing Results ............................................................................................................. 11
3.1 Ozone Test Objectives ..................................................................................................................................... 11
3.2 Ozone Test Matrix ............................................................................................................................................ 11
3.2.1 Ozone Demand Test ........................................................................................................................ 11
3.2.2 Jar Test of Ozonated Raw Water ..................................................................................................... 11
3.2.3 SDS Test of Ozonated and Coagulated Effluent ............................................................................ 12
3.3 Ozone Test Results and Discussions .............................................................................................................. 13
3.3.1 Raw Water Ozone Demand ............................................................................................................. 13
3.3.2 Jar Test of Ozonated Raw Water ..................................................................................................... 14
3.3.3 Control of DBP Formation by Pre-ozonation ................................................................................... 19
Section 4: GAC RSSCT Results .............................................................................................................................. 22
4.1 GAC Rapid Small-scale Column Test (RSSCT) Objectives .............................................................................. 22
4.2 GAC RSSCT Matrix ............................................................................................................................................ 22
4.2.1 GAC RSSCT ....................................................................................................................................... 22
4.2.2 Free Chlorine Demand and SDS Test of RSSCT Influent and Effluents ....................................... 23
4.3 GAC RSSCT Results and Discussions ............................................................................................................. 24
4.3.1 GAC Adsorption ................................................................................................................................ 24
4.3.2 SDS DBP Formation ......................................................................................................................... 27

List of Figures
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

iii
Use of contents on this sheet is subject to the limitations specified at the beginning of this document.
TM4 Bench Pre-ozonation and RSSCT Testing FINAL
Figure 1-1. Treatment processes at the Green River Water Treatment Plant (GRWTP). ..................................... 1
Figure 1-2. Pre-ozonation and GAC treatment train configurations. ..................................................................... 3
Figure 2-1. Phipps and Bird six-paddle programmable jar tester with 2-liter square B-KER2 gator jars............ 4
Figure 2-2. Settled water percent TOC removal as a function of coagulant dose in molar active ingredient
(i.e., mM as aluminum or iron) .......................................................................................................................... 7
Figure 2-3. Settled water percent DOC removal as a function of coagulant dose in molar active ingredient
(i.e., mM as aluminum or iron). ......................................................................................................................... 8
Figure 2-4. Coagulated water zeta potential and settled water DOC concentrations as a function of coagulant
dose in molar active ingredient (i.e., mM as aluminum or iron). .................................................................... 8
Figure 2-5. Settled water percent UVA 254 nm removal as a function of coagulant dose in molar active
ingredient (i.e., mM as aluminum or iron). ....................................................................................................... 9
Figure 2-6. Settled water turbidity as a function of coagulant dose in molar active ingredient (i.e., mM as
aluminum or iron)............................................................................................................................................... 9
Figure 3-1. Ozone residual as a function of ozone to TOC ratio and ozone contact time. Ozone demand test
was performed in July 2021 with raw water TOC concentration of 4.5 mg/L. ............................................ 14
Figure 3-2. Ozone residual as a function of ozone to TOC ratio and ozone contact time. Ozone demand test
was re-conducted in October 2021 with raw water TOC concentration of 3.1 mg/L. ................................. 14
Figure 3-3. Impact of ozone dose on percent TOC removal by enhanced coagulation with alum. ................... 15
Figure 3-4. Impact of ozone dose on percent DOC removal by enhanced coagulation with alum. .................. 16
Figure 3-5. Impact of ozone dose on percent UVA removal by enhanced coagulation with alum. ................... 16
Figure 3-6. Impact of ozone dose on percent turbidity removal by enhanced coagulation with alum. ............ 17
Figure 3-7. The SDS formation of TTHM as a function of ozone dose, alum dose, and hold time (i.e., reaction
time, or water age). Raw water TOC concentration was 3.5 mg/L. .............................................................. 20
Figure 3-8. The SDS formation of HAA5 as a function of ozone dose, alum dose, and hold time (i.e., reaction
time, or water age). Raw water TOC concentration was 3.5 mg/L. .............................................................. 20
Figure 4-1. Normalized DOC breakthrough (i.e., DOC/DOC0) as a function of throughput in number of bed
volumes for the pre-chlorinated and non-chlorinated RSSCTs. Full-scale GAC post-filter adsorber
EBCT=10 min. DOC0=2.75 mg/L (n=41). ....................................................................................................... 25
Figure 4-2. Normalized DOC breakthrough (i.e., DOC/DOC0) as a function of throughput in number of bed
volumes for the pre-chlorinated and non-chlorinated RSSCTs. Full-scale GAC post-filter adsorber
EBCT=20 min. DOC0=2.75 mg/L (n=41). ....................................................................................................... 25
Figure 4-3. Normalized DOC breakthrough (i.e., DOC/DOC0) as a function of throughput in number of bed
volumes for the pre-chlorinated and non-chlorinated RSSCTs at EBCT=10 and 20 minutes. DOC0=2.75
mg/L (n=41). .................................................................................................................................................... 26
Figure 4-4. DOC breakthrough as a function of throughput in number of bed volumes and EBCT for the pre-
chlorinated RSSCT. DOC0=2.67 mg/L (n=22). ............................................................................................... 26
Figure 4-5. DOC breakthrough as a function of scaled operation time and EBCT for the pre-chlorinated
RSSCT. DOC0=2.67 mg/L (n=22). Green dashed line indicates the finished water DOC target (i.e., 2.2
mg/L) to comply with TTHM MCL of 80 µg/L after 10 days of water residence time. ................................ 27
Figure 4-6. SDS TTHM and HAA5 formation (10 days) as a function of GAC effluent DOC concentration ....... 28
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

iv
Use of contents on this sheet is subject to the limitations specified at the beginning of this document.
TM4 Bench Pre-ozonation and RSSCT Testing FINAL
Figure 4-7. SDS TTHM and HAA5 formation (6 days) as a function of GAC effluent DOC concentration. ........ 29

List of Tables
Table 2-1. Jar Test Chemical Dosing Matrix ............................................................................................................ 5
Table 2-2. Jar Test Mixing Regime ........................................................................................................................... 5
Table 2-3. Raw Water Characterization Results ..................................................................................................... 6
Table 2-4. Jar Test Results ..................................................................................................................................... 10
Table 2-5. Enhanced Coagulation: Required Percent TOC Removals ..................... Error! Bookmark not defined.
Table 3-1. Ozone Demand Test Matrix .................................................................................................................. 11
Table 3-2. Ozone and Jar Test Matrix .................................................................................................................... 12
Table 3-3. SDS Test Matrix for Ozonated and Coagulated Effluents .................................................................. 12
Table 3-4. Jar Test Results of Ozonated and Non-ozonated Raw Water ............................................................ 18
Table 3-5. Pre-ozonation and Enhanced Coagulation SDS DBP Formation ....................................................... 21
Table 4-1. GAC RSSCT Design ............................................................................................................................... 22
Table 4-2. Free Chlorine Demand and SDS Test Matrix for Pre-chlorinated RSSCT Influent and Effluents ..... 24
Table 4-3. Instantaneous DBP Formation from Pre-chlorination ........................................................................ 30
Table 4-4. GAC RSSCT SDS DBP Formation ......................................................................................................... 30
Table 4-5. Chlorine Dose and Residual Data for All SDS Samples ..................................................................... 31

Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

1
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL
Section 1: Introduction
This technical memorandum presents the results of testing conducted to evaluate enhanced coagulation,
ozone, and granular activated carbon (GAC) for removal of total organic carbon (TOC) to control disinfection
byproducts (DBPs). The City of Green River’s highest DBP concentrations were primarily total trihalome-
thanes (TTHMs). In 2019 TTHMs ranged from 60 µg/L to 90 µg/L. Current TOC removal has historically aver-
aged 31 % removal, based on their raw water TOC and alkalinity, they are required (per the Stage 1 D/DBP
Rule) to achieve between 15 to 35% removal. The object of this testing is to identify the most reliable and
cost-effective way to remove total organic carbon to minimize DBPs.
Table 1-1. Enhanced Coagulation: Required Percent TOC Removals
Raw Water TOC Raw Water Alkalinity
mg/L as CaCO3
mg/L <60 60-120 >120
>2-4 35% 25% 10%
>4-8 45% 35% 25%
>8 50% 40% 30%
1.1 Green River Water Treatment Plant
The City of Green River owns and operates the Green River Water Treatment Plant (GRWTP), which treats an
average of 1,000 gallons per minute (gpm) of surface water from the Green River by conventional coagula-
tion, flocculation, sedimentation, and granular media filtration processes. Due to high turbidity of the raw
water from the river (i.e., 50 to 1,000 NTU), a portion of the river water is pumped into the pre-sedimentation
ponds to allow particles to settle before entering the GRWTP.
The GRWTP contains package plant equipment manufactured by Filter Tech Systems and was constructed in
2000. Treatment at the GRWTP consists of coagulation (or rapid mix) using an inline mixer, three-stage ta-
pered flocculation, sedimentation via tube settlers, and dual-media filtration by sand and anthracite. Free
chlorine is dosed post dual-media filtration and disinfection credit is obtained within a clearwell before fin-
ished water is discharged into the distribution system. Figure 1-1 shows the process flow diagram of the
GRWTP and the typical chemical dosing locations at the plant.

Figure 1-1. Treatment processes at the Green River Water Treatment Plant (GRWTP).
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

2
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL
1.2 Advanced Water Treatment Processes for DBP Control
DBPs are a broad groups of compounds that are formed upon drinking water disinfection when dissolved
organic matter (DOM) reacts with disinfectants such as free chlorine or chloramines.
Currently, the United States Environmental Protection Agency (USEPA) regulates two classes of halogenated
organic DBPs, specifically total trihalomethanes (TTHM) and five haloacetic acids (HAA5) under the Stage 1
and Stage 2 Disinfectants and Disinfection Byproducts (D/DBP) Rules. The maximum contaminant levels
(MCLs) for these two groups of DBPs are 80 µg/L for TTHM and 60 µg/L for HAA5. Compliance with the DBP
MCLs is based on a locational running annual average (LRAA). Historically, the City of Green River has seen
TTHM running annual average concentrations at certain sampling locations exceeding the MCL of 80 µg/L.
For this reason, the goal of the bench-scale treatability testing was to identify advanced water treatment pro-
cesses for the control of DBP precursors to assist the GRWTP in complying with the DBP regulations.
Pre-ozonation and adsorption by granular activated carbon (GAC) are two advanced water treatment pro-
cesses that are commonly applied in drinking water treatment for the control of DBP precursors. Ozone can
partially oxidize DBP precursors, particularly humic substances, and therefore control subsequent DBP for-
mation upon chlorination. Previous studies have shown that ozonation at near neutral pH results in a net
decrease in trihalomethane (THM) formation, whereas the impact of ozone on haloacetic acid (HAA) for-
mation depends on the nature of DOM present in the source waters. Similar to THM precursors, trihaloacetic
acid (THAA) precursors could be oxidized by ozone. In contrast, dihaloacetic acid (DHAA) precursors are often
negatively affected by ozone. Due to the greater extent in the reduction of THAA formation, HAA5 concentra-
tion has been shown to likely decrease when pre-ozonation is implemented. In addition to DBP precursor
control, it was reported that ozone dose between 0.4 to 0.8 mg of ozone per mg of dissolved organic carbon
(DOC) could provide additional benefits in enhancing DOM coagulation and thus turbidity removal.
GAC is another advanced treatment technology that can be applied in the form of post-filter contactors to
control DBP formation by adsorbing their organic precursors. Due to the heterogeneous nature of DBP pre-
cursors, their adsorbabilities to GAC differ, so treatment performance and economics (represented by carbon
use rate, or CUR) of the GAC adsorption process are site-specific and depend on the finished water DOC tar-
gets.
The objective of the bench-scale treatability testing was to compare the pre-ozonation treatment train
(ozone/coagulation/flocculation/sedimentation/filtration) with the GAC treatment train (coagulation/floccu-
lation/sedimentation/filtration/GAC) in removing DBP precursors and achieving overall finished water quality
goals. Figure 1-2 shows the pre-ozonation vs GAC treatment train configurations.
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

3
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Figure 1-2. Pre-ozonation and GAC treatment train configurations.
1.3 Bench-scale Treatability Testing
This report summarizes the bench-scale treatability testing results of advanced treatment processes treating
raw water from GRWTP. All bench-scale treatability testing was conducted by Brown and Caldwell (BC) at
BC’s Water Treatability Laboratory in Nashville, Tennessee. The tests consisted of the following three
phases: jar test, pre-ozonation test, and GAC rapid small-scale column test (RSSCT). All non-treated (i.e., ex-
perimental controls) and treated effluents were evaluated for TTHM and HAA5 formation potentials using
simulated distribution system (SDS) tests as a measure of treatment performance.
This summary report is organized as the following:
• Section 2: Jar Testing Results
• Section 3: Pre-ozonation Testing Results
• Section 4: GAC Rapid Small-scale Column Testing Results
• Section 5: Recommendations
Section 2: Enhanced coagulation Results
2.1 Jar Test Objectives
A series of jar tests were conducted to evaluate enhanced coagulation performance using alum (current co-
agulant), ferric chloride and aluminum chlorohydrate. Additionally, based on the results of this jar testing, a
single coagulant will be selected for the additional testing with ozone.
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

4
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Currently, alum is applied at on average 37 mg/L for enhanced coagulation as per the requirements based
on the Stage 1 and Stage 2 D/DBP Rules. A cationic polymer (i.e., SumaClear P30) and a nonionic polymer
(i.e., Magnafloc LT20) are dosed during rapid mix and the end of stage-one flocculation as coagulant aid and
flocculant aid, respectively.
The plant’s typical alum dose meets required TOC percent removals given raw water TOC concentrations and
alkalinity but treated water TOC concentrations are not low enough to consistently meet the DBP MCLs. For
this reason, alternative coagulants (i.e., ferric chloride [FeCl3] and aluminum chlorohydrate [ACH]) and higher
coagulant doses were assessed to determine whether higher TOC percent removals could be achieved by
enhanced coagulation.
2.2 Jar Test Matrix
2.2.1 Raw Water Characterization
Raw water was collected from the raw water pipeline downstream of the pre-sedimentation ponds of the
GRWTP in three 55-gallon barrels on July 1, 2021. Collected sample was freight shipped to BC’s Water Treat-
ability Laboratory in Nashville, Tennessee and was stored at 4 °C until bench-scale testing commenced.
At BC’s Water Treatability Laboratory, the raw water sample was characterized for pH, UV absorbance (UVA)
at 254 nm, TOC, DOC, turbidity, alkalinity, total dissolved solids (TDS), ammonia-nitrogen, nitrate-nitrogen,
and nitrite-nitrogen. In parallel, the same raw water sample was submitted to Eurofins Eaton Analytical (EEA)
for the characterization of bromide, total iron, and total manganese concentrations.
2.2.2 Jar Test of Alternative Coagulants and Coagulant Doses
Ferric chloride (FeCl3) and aluminum chlorohydrate (ACH) were evaluated through jar tests as alternative co-
agulants to alum. Jar tests were conducted using a Phipps and Bird programmable jar tester (Figure 2-1) and
2-liter square plastic beakers (often referred to as gator jars) under ambient raw water pH. Each coagulant
was tested at six different doses, whereas the molar amount of active ingredient (i.e., the molar concentra-
tion of iron or aluminum per liter of sample) was kept the same across the three types of coagulants to allow
for direct performance comparisons. Detailed jar test matrix is listed below in Table 2-1.

Figure 2-1. Phipps and Bird six-paddle programmable jar tester with 2-liter square B-KER2 gator jars.
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

5
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Table 2-1. Jar Test Chemical Dosing Matrix
Parameter Units Jar 1 Jar 2 Jar 3 Jar 4 Jar 5 Jar 6
pH S.U. Ambient
Molar Active Ingredient mM of Al or Fe 0 0.05 0.09 0.14 0.18 0.24
Alum Dose1 mg/L 0 15 30 45 60 80
FeCl3 Dose2 mg/L 0 12 24 37 49 65
ACH Dose3 mg/L 0 5 9 14 19 25
Permanganate Dose mg/L 0 0.7
Coagulant Aid (Cationic Polymer, SumaClear P30) Dose mg/L 0 1.5
Flocculant Aid (Nonionic Polymer, Magnafloc LT20) Dose mg/L 0 0.12
Notes:
1 Alum molecular weight: 666.201, 5 g/100 mL
2FeCl3 molecular weight: 270.3, 5 g/100 mL
3ACH molecular weight: 210.48, 2 g/100 mL

The jar test procedures were tailored to closely match the conditions at the full-scale plant, including the se-
quence of chemical addition and rapid mix, flocculation, and sedimentation mixing regimes. Table 2-2 pre-
sents the mixing energy and mixing time applied during the jar tests.

Table 2-2. Jar Test Mixing Regime
Process Mixing Speed Mixing Time
revolutions per minute (rpm) minutes (min)
Rapid Mix 300 0.5
Stage-1 flocculation 80 5.0
Stage-2 flocculation 68 5.0
Stage-3 flocculation 35 20
Sedimentation 0 18.5
After prescribed sedimentation time (Table 2-2), settled water was withdrawn from the fixed sampling port
located at the 10 centimeter (cm) settling-distance level of each jar to minimize disturbance to the settling
flocs. The first portion of sample taken from the fixed sampling port was discarded. All settled water samples
were then characterized for pH, UVA at 254 nm, TOC, DOC, and turbidity. A small aliquot of sample (i.e., 5
milliliters [mL]) was taken from each jar at the beginning of Stage-3 flocculation for the measurement of zeta
potential, a parameter that reflects the effective charge on the particles and dissolved organic matter and
determines particle stability and organic matter coagulability.
2.3 Jar Test Results and Discussions
2.3.1 Raw Water Characterization
Results of raw water characterization are presented below in Table 2-3. The raw water sample was relatively
high in pH and alkalinity with an initial TOC concentration of 4.2 mg/L (N=3), indicating potential challenges
in achieving required percent TOC removal without low pH coagulation using alum or alternative coagulants.
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

6
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Table 2-3. Raw Water Characterization Results
Bar-
rel
No.
pH UVA at 254
nm TOC Alkalinity Turbidity Total Dis-
solved Solids Bromide Total Iron Total Manganese
S.U. cm-1 mg/L mg/L as
CaCO3 NTU mg/L µg/L mg/L µg/L
1 8.53 0.179 4.34 123.4 32.6 385 18 1.2 32
2 8.53 0.219 4.26 175.0 49.4 422 18 1.2 43
3 8.19 0.176 4.04 165.0 44.9 276 19 1.2 34

2.3.2 Alternative Coagulants and Coagulant Doses
Figures 2-2 to 2-6 present all jar test results, including settled water TOC, DOC, UVA at 254 nm, turbidity, and
their percent removals as a function of coagulant dose. All coagulant doses in Figures 2-2 to 2-6 are shown
in molar concentration of active ingredient (i.e., mM of aluminum or iron) to allow for direct performance
comparisons among the three coagulant types. DOC was measured after settled water filtration using 0.45-
µm disc membrane filters. Based on the results of jar testing, it is recommended that alum continue to be
used for coagulation.
Key observations include:
• At the same molar metal doses, both ferric chloride and ACH outperformed alum in percent TOC re-
moval (Figure 2-2) by an average of 5 percent. However, this improvement in TOC removal was not
significant enough to justify the switch from alum to either ferric chloride or ACH as the primary coag-
ulant. For this reason, alum was still applied as the primary coagulant for the following treatability
tests (i.e., pre-ozonation tests and GAC RSSCT).
• At the same molar metal doses, all three coagulants resulted in equivalent percent DOC removals
(Figure 2-3). It is noteworthy that DOC is a better parameter than TOC in jar tests to evaluate the re-
moval of DOM by coagulation. This is because sampling from jar tests for TOC quantification is diffi-
cult and could lead to errors since TOC concentrations depend on particulate settling, which is scale
dependent, whereas the chemistry of coagulation reactions between coagulant and DOM is inde-
pendent of scale. In general, a water treatment plant with filtration (e.g., granular media filtration)
will remove particulate organic carbon that does not settle in the test jars. As a result, selection of
coagulant dose based on TOC concentrations or percent TOC removals may lead to overdosing,
higher chemical costs, and greater sludge production.
• Percent TOC and percent DOC removals both increased with increasing coagulant dose. However,
there was a point of diminishing returns with dosing more coagulant above 60 mg/L of alum as prod-
uct, or 49 mg/L of ferric chloride or 19 mg/L of ACH as product (Table 2-4) as TOC and DOC remov-
als plateaued.
• To meet enhanced coagulation requirements and achieve required percent TOC removal in Table 2-5
(i.e., 25 percent TOC removal based on raw water TOC of 4.2 mg/L and alkalinity of 154 mg/L as
CaCO3), a minimum alum dose of 60 mg/L as product or 0.18 mM as aluminum is required.
• 30 percent DOC removal was achieved by enhanced coagulation with alum dose of 60 mg/L as prod-
uct at ambient water pH of 8.5.
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

7
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL
• Zeta potential is a good parameter to optimize coagulant dose due to its strong correlation with set-
tled water DOC concentration (Figure 2-4). Zeta potential increased to neutral values with increasing
coagulant doses, resulting in decreased settled water DOC concentrations. Lowest treated water
DOC concentration was achieved when zeta potential was close to neutral (e.g., 0 millivolts).
• At the same molar metal doses, all three coagulants resulted in the same percent UVA reduction (Fig-
ure 2-5).
• At equivalent molar metal doses, alum and ferric chloride outperformed ACH in turbidity removal.
Settled water turbidity was constantly below 2.0 NTU when alum and ferric chloride doses were
above 30 mg/L and 24 mg/L as product, respectively (Figure 2-6).

Figure 2-2. Settled water percent TOC removal as a function of coagulant dose in molar active ingredient (i.e., mM as
aluminum or iron)
0%
5%
10%
15%
20%
25%
30%
35%
0.00 0.05 0.10 0.15 0.20 0.25 0.30
Pe
r
c
e
n
t

T
O
C

R
e
m
o
v
a
l

(
%
)
Coagulant Dose (mM as Al or Fe)
Alum
Ferric Chloride
ACH
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

8
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Figure 2-3. Settled water percent DOC removal as a function of coagulant dose in molar active ingredient (i.e., mM as
aluminum or iron).

Figure 2-4. Coagulated water zeta potential and settled water DOC concentrations as a function of coagulant dose in
molar active ingredient (i.e., mM as aluminum or iron).
0%
5%
10%
15%
20%
25%
30%
35%
40%
0.00 0.05 0.10 0.15 0.20 0.25 0.30
Pe
r
c
e
n
t

D
O
C

R
e
m
o
v
a
l

(
%
)
Coagulant Dose (mM as Al or Fe)
Alum
Ferric Chloride
ACH
0
1
2
3
4
5
-25
-20
-15
-10
-5
0
5
10
15
20
25
0.00 0.05 0.10 0.15 0.20 0.25 0.30
Se
t
t
l
e
d

W
a
t
e
r

D
O
C

(
m
g
/
L
)

(
s
o
l
i
d

l
i
n
e
s
)
Ze
t
a

P
o
t
e
n
t
i
a
l

(
m
V
)

(
d
a
s
h
e
d

l
i
n
e
s
)
Coagulant Dose (mM as Al or Fe)
Alum
Ferric Chloride
ACH
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

9
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Figure 2-5. Settled water percent UVA 254 nm removal as a function of coagulant dose in molar active ingredient (i.e.,
mM as aluminum or iron).

Figure 2-6. Settled water turbidity as a function of coagulant dose in molar active ingredient (i.e., mM as aluminum or
iron).

0%
10%
20%
30%
40%
50%
60%
70%
0.00 0.05 0.10 0.15 0.20 0.25 0.30
UV
A

2
5
4

n
m

R
e
m
o
v
a
l
Coagulant Dose (mM as Al or Fe)
Alum
Ferric Chloride
ACH
0
5
10
15
20
25
30
0.00 0.05 0.10 0.15 0.20 0.25 0.30
Se
t
t
l
e
d

W
a
t
e
r

T
u
r
b
i
d
i
t
y

(
N
T
U
)
Coagulant Dose (mM as Al or Fe)
Alum
Ferric Chloride
ACH
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0.00 0.10 0.20 0.30
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

10
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Table 2-4. Jar Test Results
Coagulant
Coagulant
Dose
KMnO4
Dose
Cationic
Polymer
Dose
Nonionic
Polymer
Dose
Zeta
Potential pH TOC TOC
Removal DOC DOC
Removal
UVA at
254 nm
UVA
Removal Turbidity
mg/L mg/L mg/L mg/L mV SU mg/L % mg/L % cm-1 % NTU
Alum
0 0 0 0 -15.3 8.74 4.51 0% 4.17 0% 0.091 0% 26.5
15 0.7 1.5 0.12 -8.3 8.00 4.34 4% 3.6 14% 0.063 31% 2.74
30 0.7 1.5 0.12 -6.3 7.78 3.81 16% 3.2 23% 0.053 42% 1.19
45 0.7 1.5 0.12 -5.7 7.62 3.70 18% 3.18 24% 0.048 47% 1.10
60 0.7 1.5 0.12 -4.1 7.49 3.38 25% 2.93 30% 0.042 54% 1.10
80 0.7 1.5 0.12 -3.8 7.35 3.22 29% 2.79 33% 0.039 57% 1.27
FeCl3
0 0 0 0 -22.2 8.45 4.29 0% 4.1 0% 0.09 0% 27.6
12 0.7 1.5 0.12 -11.9 8.01 3.94 8% 3.69 10% 0.062 31% 2.72
24 0.7 1.5 0.12 -10.4 7.69 3.60 16% 3.24 21% 0.053 41% 1.68
37 0.7 1.5 0.12 -11.3 7.56 3.28 24% 3.04 26% 0.046 49% 0.733
49 0.7 1.5 0.12 -10.7 7.43 3.01 30% 2.78 32% 0.041 54% 0.832
65 0.7 1.5 0.12 -8.1 7.32 2.87 33% 2.55 38% 0.036 60% 0.94
ACH
0 0 0 0 -19.8 8.46 4.32 0% 4.12 0% 0.091 0% 26.3
5 0.7 1.5 0.12 -10.2 8.39 3.80 12% 3.38 18% 0.058 36% 1.80
9 0.7 1.5 0.12 -6.4 8.31 3.53 18% 3.3 20% 0.053 42% 1.81
14 0.7 1.5 0.12 -3.4 8.26 3.27 24% 3.15 24% 0.048 47% 1.81
19 0.7 1.5 0.12 -3.7 8.17 3.05 29% 2.86 31% 0.042 54% 1.95
25 0.7 2 0.12 -2.1 8.12 3.04 30% 2.78 33% 0.041 55% 2.02
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

11
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Section 3: Pre-ozonation Testing Results
3.1 Ozone Test Objectives
An ozone demand study was first performed to determine the proper ozone dose that would result in com-
plete ozone decay (i.e., zero ozone residual) after 10 minutes of contact time. Pre-ozonated raw water was
then jar tested to evaluate the potential impact of ozonation on coagulant dose and the control of DBP for-
mation.
3.2 Ozone Test Matrix
3.2.1 Ozone Demand Test
Two batches of raw water were ozonated at different ozone to TOC ratios ranging from 0.5: 1, 0.75: 1, to 1: 1
mg of ozone per mg of carbon. Ozone residual was monitored as a function of reaction time from 0.5
minutes up to 10 minutes to determine the appropriate dose for pre-ozonation. Two separate ozone de-
mand tests were performed separated by 2 months. Table 3-1 contains a detailed ozone demand test ma-
trix.

Table 3-1. Ozone Demand Test Matrix
Testing Time Raw Water TOC Ozone: TOC Ozone Dose Reaction Time
mg/L mg of O3/mg of C mg/L minutes
July 2021 4.5
0.5: 1 2.3
0.5, 1.0, 1.5, 2, 3, 5, 8, 10
0.75: 1 3.4
1: 1 4.5
October 2021 3.1 0.5: 1 1.6
0.75: 1 2.3

Ozone was dosed using the bench-scale batch ozonation method, which involved the preparation of an
ozone stock solution in ultrapure ozone-demand-free water and the addition of ozone stock solution into the
sample to be treated at prescribed volumes to achieve target ozone doses. Batch ozonation is the simplest
method in ozonating aqueous samples and allows for high accuracy in ozone dosing.
Ozone stock solution concentration was determined by directly measuring UV absorbance at 260 nm, at
which wavelength, ozone has the highest molar absorptivity of 3,290 M-1 cm-1. Residual ozone concentra-
tions in the treated samples were determined based on the indigo colorimetric method. Ozone bleaches this
blue-colored dye at a stoichiometric ratio, and the loss in UV absorbance at 600 nm is linearly correlated
with aqueous ozone concentration.
3.2.2 Jar Test of Ozonated Raw Water
In this experiment, raw water was first ozonated at two ozone to TOC ratios, 0.5: 1 and 0.75: 1 mg of ozone
per mg of carbon, as determined through the ozone demand test. Ozonated samples were then jar tested
using alum as the primary coagulant at six different doses. One set of control jar test was conducted in
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

12
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL
parallel on raw water that was not ozonated to evaluate the impact of pre-ozonation on settled water quality
in pH, TOC, DOC, UVA at 254 nm, and turbidity. Table 3-2 shows a detailed ozone and jar test matrix.

Table 3-2. Ozone and Jar Test Matrix
Jar O3:TOC
Ozone
Contact
Time Coagulant Coagulant Dose
Cationic
Polymer
Dose
Nonionic
Polymer
Dose
mg of O3/mg of C minutes mg/L as Product mg/L mg/L
1
0 --
Alum
0 0 0
2 20 1.5 0.12
3 40 1.5 0.12
4 60 2.0 0.12
5 80 2.0 0.12
6 100 2.0 0.12
1
0.5:1 10
0 0 0
2 20 1.5 0.12
3 40 1.5 0.12
4 60 2.0 0.12
5 80 2.0 0.12
6 100 2.0 0.12
1
0.75:1 10
0 0 0
2 20 1.5 0.12
3 40 1.5 0.12
4 60 2.0 0.12
5 80 2.0 0.12
6 100 2.0 0.12

3.2.3 SDS Test of Ozonated and Coagulated Effluent
Jar tests were followed by settled water characterization for pH, UVA at 254 nm, TOC, DOC, and turbidity. A
total of eight ozonated and/or coagulated samples (Table 3-3) were assessed for DBP formation via simu-
lated distribution system (SDS) tests to evaluate the impact of ozone on the control of DBP formation.
SDS conditions include incubation temperature of 20 ± 1 °C, ambient water pH, and target chlorine residual
of 0.5 mg/L as Cl2 after 6 days and 0.2 mg/L as Cl2 after 10 days. A chlorine demand study was conducted
prior to the SDS tests to determine required free chlorine dose for each sample to achieve the target chlo-
rine residuals of 0.5 mg/L as Cl2 and 0.2 mg/L as Cl2 after hold time of 6 days and 10 days, respectively.
At hold time of 2, 6, and 10 days, free chlorine residual in each sample was measured using the Hach®
Method 8167. In parallel, DBP samples were collected and analyzed for TTHM and HAA5 formation by EEA
using EPA Method 551.1 for TTHM and 552.2 for HAA5.

Table 3-3. SDS Test Matrix for Ozonated and Coagulated Effluents
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

13
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL
Sample O3: TOC Alum Dose Cl2 : DOC for Chlorine Demand Test SDS Hold Time
mg of O3 per mg of C mg/L as product mg of Cl2 per mg of C days
1 (control) 0 40
1: 1
1.5: 1
2: 1
2 days
6 days
10 days
2 (control) 0 60
3 (control) 0.5: 1 0
4 (control) 0.75 :1 0
5 0.5: 1 40
6 0.5 :1 60
7 0.75: 1 40
8 0.75 :1 60

3.3 Ozone Test Results and Discussions
3.3.1 Raw Water Ozone Demand
Raw water ozone demand test results are presented in Figures 3-1 and 3-2. Ozone residuals are shown as a
function of ozone to TOC ratio and ozone contact time. The two batches of raw water tested had initial TOC
concentrations of 4.5 mg/L and 3.1 mg/L. The second raw water sample was stored for a few additional
months before it was used in the ozone bench test. The decrease in raw water TOC concentration was po-
tentially due to the long sample storage time and biodegradation of dissolved organic matter present in the
water.
Regardless of the initial TOC concentrations, an ozone dose at ozone to TOC ratio of 0.5: 1 mg of ozone per
mg of carbon resulted in complete ozone decay after 8 minutes of ozone contact time. In comparison, a
dose at ozone to TOC ratio of 0.75: 1 mg of ozone per mg of carbon yielded an ozone residual of less than
0.5 mg/L and zero ozone residual after 10 minutes in the two raw water samples tested, respectively. An
ozone dose at ozone to TOC ratio of 1: 1 mg of ozone per mg of carbon yielded an ozone residual of 2 mg/L
after 10 minutes of ozone contact, which would require chemical quenching prior to downstream conven-
tional treatment processes and therefore was considered as too high for this application.
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

14
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Figure 3-1. Ozone residual as a function of ozone to TOC ratio and ozone contact time. Ozone demand test was per-
formed in July 2021 with raw water TOC concentration of 4.5 mg/L.

Figure 3-2. Ozone residual as a function of ozone to TOC ratio and ozone contact time. Ozone demand test was re-
conducted in October 2021 with raw water TOC concentration of 3.1 mg/L.
3.3.2 Jar Test of Ozonated Raw Water
To evaluate the potential impact of pre-ozonation and ozone dose on DBP precursor control, raw water was
ozonated at two doses with respective ozone to TOC ratios of 0.5: 1 and 0.75: 1, followed by jar test using
alum as the primary coagulant at 6 different doses. One set of control jar test was conducted in parallel on
raw water that was not ozonated to evaluate the impact of pre-ozonation and ozone dose on settled water
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
0 1 2 3 4 5 6 7 8 9 10 11
Oz
o
n
e

R
e
s
i
d
u
a
l

(
m
g
/
L
)
Ozone Contact Time (minutes)
Ozone:TOC=1:1
Ozone:TOC=0.75:1
Ozone:TOC=0.5:1
TOC0=4.5 mg/L
0.0
0.5
1.0
1.5
2.0
2.5
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0
Oz
o
n
e

R
e
s
i
d
u
a
l

(
m
g
/
L
)
Ozone Contact Time (minutes)
Ozone:TOC=0.75:1
Ozone:TOC=0.5:1
TOC0=3.1 mg/L
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

15
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL
quality in pH, TOC, DOC, UVA at 254 nm, and turbidity. Results of the jar tests are shown below in Table 3-4
and Figures 3-3 to 3-6.
Key observations include:
• Regardless of ozone dose, pre-ozonation did not have a significant impact on enhanced coagulation
performance in TOC (Figure 3-3), UVA (Figure 3-5), and turbidity (Figure 3-6) removal. At the same
alum doses, settled water percent TOC, UVA, and turbidity removals were comparable between the
non-ozonated and ozonated raw waters and between the ozonated samples at two ozone doses (i.e.,
ozone to TOC ratios of 0.5:1 and 0.75:1 mg of ozone per mg of carbon).
• High alum doses (≥ 80 mg/L as product) combined with pre-ozonation resulted in decreased perfor-
mance and low settled water turbidity removal compared to enhanced coagulation of non-ozonated
raw water (Figure 3-6).
• High pre-ozone doses enhanced DOC removal. A high ozone dose resulted in greater percent DOC
removal by enhanced coagulation at the same alum doses (Figure 3-4).

Figure 3-3. Impact of ozone dose on percent TOC removal by enhanced coagulation with alum.
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
0 20 40 60 80 100 120
Pe
r
c
e
n
t

T
O
C

R
e
m
o
v
a
l

(
%
)
Alum Dose (mg/L as product)
Non-ozonated
Ozone:TOC=0.5:1
Ozone:TOC=0.75:1
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

16
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Figure 3-4. Impact of ozone dose on percent DOC removal by enhanced coagulation with alum.

Figure 3-5. Impact of ozone dose on percent UVA removal by enhanced coagulation with alum.
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
0 20 40 60 80 100 120
Pe
r
c
e
n
t

D
O
C

R
e
m
o
v
a
l

(
%
)
Alum Dose (mg/L as product)
Non-ozonated
Ozone:TOC=0.5:1
Ozone:TOC=0.75:1
0%
10%
20%
30%
40%
50%
60%
70%
0 20 40 60 80 100 120
Pe
r
c
e
n
t

U
V
A

R
e
m
o
v
a
l

(
%
)
Alum Dose (mg/L as product)
Non-ozonated
Ozone:TOC=0.5:1
Ozone:TOC=0.75:1
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

17
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Figure 3-6. Impact of ozone dose on percent turbidity removal by enhanced coagulation with alum.

0%
20%
40%
60%
80%
100%
0 20 40 60 80 100 120
Pe
r
c
e
n
t

T
u
r
b
i
d
i
t
y

R
e
m
o
v
a
l

(
%
)
Alum Dose (mg/L as product)
Non-ozonated
Ozone:TOC=0.5:1
Ozone:TOC=0.75:1
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

18
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Table 3-4. Jar Test Results of Ozonated and Non-ozonated Raw Water
Jar
O3:TOC Alum Dose
Cationic
Polymer
Dose
Nonionic
Polymer
Dose
pH TOC TOC
Removal DOC DOC
Removal
UVA at
254 nm
UVA
Removal Turbidity Turbidity
Removal
mg of O3/mg of C mg/L as
Product mg/L mg/L SU mg/L % mg/L % cm-1 % NTU %
1
0
0 1.5 0.12 8.17 3.555 0% 2.979 0% 0.084 0% 33.6 0%
2 20 1.5 0.12 7.91 2.617 26% 2.456 18% 0.052 38% 3.85 89%
3 40 1.5 0.12 7.76 2.465 31% 2.358 21% 0.045 46% 4.95 85%
4 60 2 0.12 7.58 2.002 44% 1.966 34% 0.038 55% 3.84 89%
5 80 2 0.12 7.44 2.101 41% 2.002 33% 0.035 58% 4.88 85%
6 100 2 0.12 7.33 1.955 45% 1.767 41% 0.03 64% 3.64 89%
1
0.5:1
0 0 0 7.52 3.69 0% 2.92 0% 0.049 0% 33.0 0%
2 20 1.5 0.12 7.45 2.7 27% 2.477 15% 0.028 43% 1.54 95%
3 40 1.5 0.12 7.39 2.361 36% 2.25 23% 0.024 51% 2.94 91%
4 60 2 0.12 7.18 2.362 36% 2.148 26% 0.023 53% 3.48 89%
5 80 2 0.12 7.12 2.314 37% 2.163 26% 0.019 61% 6.64 80%
6 100 2 0.12 6.97 2.3 38% 1.98 32% 0.018 63% 9.32 72%
1
0.75:1
0 0 0 7.42 3.491 0% 3.313 0% 0.044 0% 35.5 0%
2 20 1.5 0.12 7.33 2.444 30% 2.364 29% 0.025 43% 3.38 90%
3 40 1.5 0.12 7.24 2.328 33% 2.226 33% 0.022 50% 4.01 89%
4 60 2 0.12 7.16 2.312 34% 2.115 36% 0.02 55% 5.99 83%
5 80 2 0.12 7.14 2.182 37% 2.097 37% 0.017 61% 8.69 76%
6 100 2 0.12 7.07 2.526 28% 1.983 40% 0.016 64% 11.1 69%
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

19
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL
3.3.3 Control of DBP Formation by Pre-ozonation
A total of eight ozonated and/or coagulated samples (Table 3-3) were assessed for DBP formation via SDS
tests to evaluate the impact of pre-ozonation on the control of DBP formation. Results of the SDS tests are
shown below in Table 3-5 and Figures 3-7 and 3-8.
The following are key observations and conclusions:
• None of the alternatives met the TTHM goal of 80 ug/L with a sufficient factor of safety for the 10-
day test (two of preozone plus EC resulted in a 10 day TTHM of 79 µg/L).
• Compared to ozonation alone or coagulation alone, ozonation prior to enhanced coagulation with
alum resulted in lower TTHM and HAA5 formation over reaction time from 2 days up to 10 days.
• A higher ozone dose (i.e., 0.75 mg of ozone per mg of carbon) in combination with a higher alum
dose (i.e., 60 mg/L as product) resulted in lowest TTHM and HAA5 formation under SDS conditions
at all three prescribed reaction times.
• At an average water age of 6 days, pre-ozonation and enhanced coagulation at all ozone and alum
dose combinations controlled TTHM formation without exceeding the MCL of 80 µg/L. It is notewor-
thy that TOC concentration of the raw water sample was 3.5 mg/L, which is relatively lower than typi-
cal raw water TOC observed at the GRWTP (monthly average TOC concentration ranges from 3 to 6
mg/L). This lower raw water TOC concentration might also have contributed to the compliance with
the TTHM MCL. The reliance of pre-ozonation for TTHM formation control, particularly when raw wa-
ter TOC concentration seasonally spikes, will require further evaluation.
• At a water age of 10 days, the compliance with TTHM MCL required a high alum dose (i.e., 60 mg/L
as product) for enhanced coagulation regardless of the ozone dose. Pre-ozonation and coagulation
at lower ozone and alum dose combinations (i.e., alum dose of 40 mg/L as product) yielded TTHM
formation that exceeded the MCL of 80 µg/L under SDS conditions.
• HAA5 MCL of 60 µg/L was not exceeded under all ozone and alum dose combinations after all three
prescribed hold times. HAA5 SDS chlorination of ozonated and coagulated effluents yielded values
well below the MCL.
• In general, pre-ozonation at doses ranging from 0.5 to 0.75 mg of ozone per mg of carbon had posi-
tive impact on the overall control of TTHM and HAA5 formation. Compared to enhanced coagulation
alone, pre-ozonation followed by coagulation significantly lowered the TTHM and HAA5 formation
even though high alum dose (i.e., 60 mg/L as product) was required to control TTHM formation be-
low the MCL.
• The necessary ozone and alum doses will be driven by raw water TOC concentration as well as TTHM
MCL given that all HAA5 SDS concentrations were well below the MCL of 60 µg/L after pre-ozonation
and enhanced coagulation regardless of ozone and alum doses.
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

20
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Figure 3-7. The SDS formation of TTHM as a function of ozone dose, alum dose, and hold time (i.e., reaction time, or
water age). Raw water TOC concentration was 3.5 mg/L.

Figure 3-8. The SDS formation of HAA5 as a function of ozone dose, alum dose, and hold time (i.e., reaction time, or
water age). Raw water TOC concentration was 3.5 mg/L.

Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

21
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Table 3-5. Pre-ozonation and Enhanced Coagulation SDS DBP Formation
Sample O3:TOC Coagulant Dose DOC Chlorine Dose Reaction Time Chlorine Residual CHCl3 CHCl2Br CHClBr2 CHBr3 TTHM MCAA DCAA DBAA TCAA HAA5
mg O3/mg C mg/L mg/L mg/L as Cl2 days mg/L as Cl2 µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L
1 0 40 2.36 3.2
2 1.32 49 17 4.2 0.51 71 ND 14 ND 20 34
6 0.57 74 24 5.8 ND 104 2.8 22 ND 26 51
10 0.24 88 23 5.5 ND 117 3 25 ND 29 57
2 0 60 1.97 3.0
2 1.2 39 16 4.3 ND 59 ND 11 ND 14 25
6 0.59 59 21 6 ND 86 ND 18 1.8 20 40
10 0.29 77 23 6.4 ND 106 2.6 22 ND 23 48
3 0.5:1 0 2.92 5.1
2 2.06 63 18 6.4 ND 87 ND 20 ND 23 43
6 0.95 78 25 8 ND 111 5.4 34 2 34 75
10 0.45 130 26 7.8 ND 164 4.2 41 1.1 39 85
4 0.75:1 0 3.31 5.2
2 2.08 63 16 5.3 ND 84 3.3 23 ND 24 50
6 0.89 98 22 7.2 ND 127 8.1 34 2.3 32 76
10 0.47 120 23 7.1 ND 150 4.2 42 ND 39 85
5 0.5:1 40 2.25 3.6
2 1.14 32 12 5.2 0.55 50 ND 12 ND 9.5 22
6 0.63 50 17 7 0.62 75 3.6 18 1.9 14 38
10 0.17 65 19 8 0.5 93 3.2 23 1.1 18 45
6 0.5:1 60 2.15 3.4
2 1.32 24 11 4.9 ND 40 ND 9.4 ND 7.2 17
6 0.72 40 16 6.3 0.66 63 3.4 17 2 11 33
10 0.47 53 18 7.5 ND 79 2.7 21 1.2 15 40
7 0.75:1 40 2.23 3.6
2 1.44 27 11 4.7 0.5 43 ND 7.8 ND 6.4 14
6 0.74 43 15 6.1 0.51 65 4.1 17 1.2 11 33
10 0.44 61 18 6.8 ND 86 2.8 23 1.1 16 43
8 0.75:1 60 2.12 3.4
2 1.34 23 10 4.2 ND 37 ND 9.1 ND 6.1 15
6 0.75 37 14 6 0.56 58 ND 15 1.4 9.7 26
10 0.48 55 17 6.8 0.59 79 2.6 21 1.1 14 39
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

22
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL
Section 4: GAC RSSCT Results
4.1 GAC Rapid Small-scale Column Test (RSSCT) Objectives
The primary objective of the RSSCT was to evaluate the performance of a single GAC post-filter adsorber for the
control of DOC and DBP precursors for the GRWTP. Another key objective was to determine the DOC target in
the final treated effluent to meet TTHM and HAA5 MCLs.
4.2 GAC RSSCT Matrix
4.2.1 GAC RSSCT
The detailed methodology used for the design, construction, and operation of the GAC RSSCT is outlined in the
“ICR Manual for Bench- and Pilot-Scale Treatment Studies”, EPA 814-B-96-003, April 1996. The RSSCT was de-
signed using the proportional diffusivity (PD) approach to simulate full-scale post-filter adsorber empty bed con-
tact time (EBCT) of 10 minutes and 20 minutes (i.e., two 10-minute columns in series). The design conditions
and relevant design equations are listed below in Table 4-1. The GAC used in this study was Filtrasorb 400 sup-
plied by Calgon Carbon, which is a bituminous-based GAC with 12×40 mesh size and was ground to a mesh
size of 100×200 for the small-scale columns.
One batch (i.e., 50 gallons) of raw water was first coagulated with alum at a dose of 60 mg/L (as product) and
was filtered through a cartridge filter with 0.45 µm pore size opening and used as the influent for the RSSCT.
Two RSSCTs were run in parallel, one with free chlorine added to the influent and the other without free chlo-
rine addition. The pre-chlorinated influent was tested to simulate the scenario where GAC treatment is down-
stream of disinfection in the existing clearwell. Influents and effluents of both RSSCTs were sampled daily to
monitor DOC and UVA breakthrough.

Table 4-1. GAC RSSCT Design
Parameter Symbol Units Value Design Equation
Carbon
Manufacturer -- -- Calgon -
Product -- -- F400 -
Type -- -- Bituminous -
Dry Bed Density ρb g/cm3 0.56 -
Bed Porosity εb g/cm3 0.37 -
Particle Porosity εp g/cm3 0.5 -
Apparent Particle Density εap g/cm3 0.89 ρb/(1- εb)
Method -- -- PD PD or CD
X -- -- 1 X = 1 (PD) or X = 0 (CD)
Water
Temperature T °C 23 -
Kinematic Viscosity kv m2/s 9.34E-07 -
Density of Water ρw kg/m3 997.6 -
Dynamic Viscosity dv g•cm-1•s-1 0.0093 -
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

23
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL
Table 4-1. GAC RSSCT Design
Parameter Symbol Units Value Design Equation
Large
Column
Upper Sieve Size (Large Scale) -- -- 12 -
Lower Sieve Size (Large Scale) -- -- 40 -
Media Diameter dp LC mm 0.92 -
Empty Bed Contact Time EBCTLC min 10 -
Hydraulic Loading Rate vLC m/hr 4.5 -
Reynolds No. ReLC -- 3.3 -
RSSCT
Column
Upper Sieve Size (Large Scale) -- -- 100 -
Lower Sieve Size (Large Scale) -- -- 200 -
Media Diameter dp SC mm 0.11 -
Column Diameter Dcol mm 4.76 -
Flow Rate Qsc mL/min 1.5 -
Hydraulic Loading Rate vsc m/hr 5.1 vsc = Qsc/A
Ideal HLR videal m/hr 38.3 -
Column Area A cm2 0.18 A = π•(Dcol)2/4
Aspect Ratio AR -- 44 AR = Dcol/dp SC
Scaling Factor SF -- 8.5 SF = dLC/dSC
Empty Bed Contact Time EBCTSC min 1.18 EBCTSC = EBCTLC/SF 2-x
Reynold No. ReSC -- 0.44 -
Bed Volume V mL 1.76 V = A•Lsc
Bed Length Lsc cm 9.9 Lsc = vsc•EBCTSC
Mass GAC Required MGAC g 0.8 MGAC = EBCTsc•Qsc•ρb
No. of BV per Day -- -- 1224 -
Volume of Water per Day -- L 2.2 -

4.2.2 Free Chlorine Demand and SDS Test of RSSCT Influent and Effluents
For the pre-chlorinated RSSCT, a chlorine demand study was first performed to determine the chlorine dose for
the column feed that would result in a free chlorine residual of less than 0.2 mg/L as Cl2 after 30 minutes of
chlorine contact time. A high free chlorine residual should be avoided in the GAC influent to prevent GAC from
being used for free chlorine quenching other than dissolved organic matter (or DBP precursor) adsorption. The
instantaneous TTHM and HAA5 formation as a result of pre-chlorination were quantified as part of the free chlo-
rine demand study. A free chlorine dose of 1.2 mg/L as Cl2 was found to yield the target chlorine residual of
less than 0.2 mg/L as Cl2 after 30 minutes. This dose was used to chlorinate 3 liters of coagulated and filtered
effluent on a daily basis to feed the small-scale columns.
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

24
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL
Since no difference was observed in DOC and UVA breakthrough between the pre-chlorinated and the non-chlo-
rinated RSSCTs, only pre-chlorinated RSSCT influent and effluents were collected at DOC breakthrough of 10,
30, 50, 70 percent and were further chlorinated to determine the free chlorine demand and TTHM and HAA5
formation. Detailed free chlorine demand and SDS test matrix for pre-chlorinated RSSCT influent and effluents
is listed below in Table 4-2.

Table 4-2. Free Chlorine Demand and SDS Test Matrix for Pre-chlorinated RSSCT Influent and Effluents
Sample DOC Percent Breakthrough Cl2 : DOC for Chlorine Demand Test SDS Test Conditions
% mg of Cl2 per mg of C /
1 10%
1: 1
1.5: 1
2: 1
20 ± 1°C
Ambient water pH
6-day chlorine residual=0.5 mg/L as Cl2
10-day chlorine residual=0.2 mg/L as Cl2
2 30%
3 50%
4 70%
5 100% (RSSCT Influent)

4.3 GAC RSSCT Results and Discussions
4.3.1 GAC Adsorption
The non-chlorinated and pre-chlorinated RSSCT results are presented in Figures 4-1 to 4-5.
Figures 4-1 and 4-2 present the normalized DOC breakthrough (i.e., DOC/DOC0) as a function of throughput in
number of bed volumes in both columns (with and without pre-chlorination) at EBCT of 10 and 20 minutes, re-
spectively. The influent DOC to these columns were 2.75 mg/L (n=41). Throughput (i.e., number of bed vol-
umes) was calculated from the RSSCT volume of water treated divided by the GAC bed volume in the small-
scale columns.
DOC in all RSSCT effluents showed an immediate breakthrough of approximately 10% of the influent DOC,
which represents the non-adsorbable fraction of organic carbon in the source water. The typical range of this
fraction is 5 to 15% (Roberts and Summers, 1982). Perhaps most importantly, Figures 4-1 and 4-2 show that
there is no impact of pre-chlorination on GAC performance in terms of DOC adsorption at both EBCTs.
For both non-chlorinated and pre-chlorinated RSSCTs, DOC breakthrough of the 10 minute column (Figure 4-1)
began after approximately 1,000 bed volumes and increased with increasing throughput until 20,000 bed vol-
umes at a DOC concentration approaching that of the influent. DOC breakthrough of the 20 minute column (Fig-
ure 4-2) began around 2,000 bed volumes and continued until about 60 percent DOC breakthrough at 10,000
bed volumes. Expressed on the same throughput basis, Figure 4-3 shows that there is no difference in GAC per-
formance between the two EBCTs.
Taken together, GAC performance (i.e., DOC breakthrough) was not impacted by either pre-chlorination or EBCT
for this source water.
Figures 4-4 and 4-5 present the breakthrough of DOC in the column with prechlorination as a function of
throughput in number of bed volumes and scaled operation time, respectively. Scaled operation time is the pro-
jection of the full-scale operation time for a single adsorber, which is calculated as throughput (bed volumes)
times the full-scale EBCT. The influent DOC to these two columns was 2.67 mg/L (n=22). The DOC in the RSSCT
effluent showed an immediate breakthrough of approximately 0.25 mg/L or 10 percent of the influent DOC re-
gardless of EBCT. DOC breakthrough of the 10 minute column began 10 days after the start of the run (Figure
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

25
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL
4-5) and increased with time until 180 days at a DOC concentration approaching that of the influent. DOC
breakthrough of the 20 minute column also increased with increasing scaled operation time and continued un-
til the end of the test at a DOC of 1.9 mg/L at 170 days. Expressed on a throughput (number of bed volumes)
basis, Figure 4-4 shows that there is no difference in GAC performance at the two EBCTs.

Figure 4-1. Normalized DOC breakthrough (i.e., DOC/DOC0) as a function of throughput in number of bed volumes for the
pre-chlorinated and non-chlorinated RSSCTs. Full-scale GAC post-filter adsorber EBCT=10 min. DOC0=2.75 mg/L (n=41).

Figure 4-2. Normalized DOC breakthrough (i.e., DOC/DOC0) as a function of throughput in number of bed volumes for the
pre-chlorinated and non-chlorinated RSSCTs. Full-scale GAC post-filter adsorber EBCT=20 min. DOC0=2.75 mg/L (n=41).
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 5000 10000 15000 20000 25000
DO
C

/

D
O
C
0
(%
)
Throughput (Number of Bed Volumes)
Pre-chlorinated EBCT=10 min
Non-chlorinated EBCT=10 min
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 2500 5000 7500 10000 12500
DO
C

/

D
O
C
0
(%
)
Throughput (Number of Bed Volumes)
Pre-chlorinated EBCT=20 min
Non-chlorinated EBCT=20 min
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

26
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Figure 4-3. Normalized DOC breakthrough (i.e., DOC/DOC0) as a function of throughput in number of bed volumes for the
pre-chlorinated and non-chlorinated RSSCTs at EBCT=10 and 20 minutes. DOC0=2.75 mg/L (n=41).

Figure 4-4. DOC breakthrough as a function of throughput in number of bed volumes and EBCT for the pre-chlorinated
RSSCT. DOC0=2.67 mg/L (n=22).
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 5000 10000 15000 20000 25000
DO
C

/

D
O
C
0
(%
)
Throughput (Number of Bed Volumes)
Pre-chlorinated EBCT=10 min
Pre-chlorinated EBCT=20 min
Non-chlorinated EBCT=10 min
Non-chlorinated EBCT=20 min
0.0
0.5
1.0
1.5
2.0
2.5
0 5,000 10,000 15,000 20,000 25,000 30,000
DO
C

(
m
g
/
L
)
Throughput (Number of Bed Volumes)
PreCl2 10 min
PreCl2 20 min
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

27
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Figure 4-5. DOC breakthrough as a function of scaled operation time and EBCT for the pre-chlorinated RSSCT. DOC0=2.67
mg/L (n=22). Green dashed line indicates the finished water DOC target (i.e., 2.2 mg/L) to comply with TTHM MCL of 80
µg/L after 10 days of water residence time.
4.3.2 SDS DBP Formation
The concentrations of TTHM and HAA5 formed in the RSSCT influent by pre-chlorination are shown in Table 4-3,
and instantaneous TTHM and HAA5 formation was about 14 percent (TTHM) to 15 percent (HAA5) of the SDS
DBP formation after 10 days of reaction time.
Samples of the RSSCT effluent at 4 different DOC concentrations (0.3, 0.8, 1.4, and 2.0 mg/L) were taken to
evaluate DBP formation and determine DOC treatment target. Sample of the RSSCT influent was also chlorin-
ated to simulate DBP formation without GAC treatment. TTHM and HAA5 formation and chlorine residual as a
function of effluent DOC are shown in Table 4-4 for SDS hold times of 2, 6, and 10 days. Chlorine demand of
each SDS sample was determined based on data shown in Table 4-5.
SDS chlorination of the influent yielded TTHM formation of 120 µg/L that far exceeded the MCL at a hold time
of 10 days, while HAA5 formation was below the MCL at 46 µg/L (Table 4-4). As shown in Figure 4-6, in all
cases, GAC treatment controlled the TTHM and HAA5 formation to values less than the MCLs. A comparison of
TTHM and HAA5 formed after 10 days to that formed after 6 days is shown in Table 4-4. The 10 day formation
was 14 percent higher for TTHM and 10 percent higher for HAA5. These values are as expected as HAA for-
mation kinetics are fast, so less formation is expected over long reaction times, TTHM kinetics are slower so
more of an increase in concentration over time is expected.
The SDS DBP formation results indicate TTHM formation will be more important than HAA5 formation and will
drive the DOC target for the GAC effluent. The SDS formation of TTHM and HAA5 as a function of GAC effluent
DOC concentration for a hold time of 6 days (i.e., average distribution system residence time) is shown in Figure
0.0
0.5
1.0
1.5
2.0
2.5
0 50 100 150 200
DO
C

(
m
g
/
L
)
Scaled Operation Time (days)
PreCl2 10 min
PreCl2 20 min
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

28
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL
4-6. Also shown are the TTHM target of 64 µg/L and HAA5 target of 48 µg/L (i.e., 80 percent of MCL). The TTHM
treatment objective translates to DOC treatment objective of 2.0 mg/L with pre-chlorination (indicated by the
red arrow in Figure 4-6). This means that a single GAC post-filter adsorber run to an effluent DOC concentration
of 2.0 mg/L and all formed TTHM and HAA5 concentrations should be lower than their respective MCLs. In ad-
dition, the expected scaled operation time (SOT) to achieve the DOC treatment target of 2.0 mg/L is shown in
Figure 4-5 under a single adsorber scenario at EBCT of 10 minutes, which is 80 days. Since there is no differ-
ence in GAC performance on the same throughput basis, the expected SOT for an adsorber with 20-minute
EBCT is 160 days.
As indicated above, the RSSCT results suggest little impact of pre-chlorination on DOC breakthrough and DBP
formation control when a single GAC post-filter adsorber is used. This may be caused by a low pre-chlorination
dose: only 14 percent of the SDS TTHM in the influent was formed by pre-chlorination. In addition, the pre-chlo-
rination dose, 1.2 mg/L, was only 30% of the required chlorine dose to carry a free chlorine residual after 10
days.
RSSCT results also indicate there is no significant benefit of longer EBCT. At an EBCT of 10 minutes, the GAC
contactor can run for 2.7 months under a single adsorber scenario, and about 5 months under a blending sce-
nario of 2 parallel contactors (i.e., approximately 50 percent longer SOT). This compares to 5 months for a sin-
gle adsorber at an EBCT of 20 minutes, and 10 months if blending 2 parallel contactors.

Figure 4-6. SDS TTHM and HAA5 formation (10 days) as a function of GAC effluent DOC concentration
y = 11.067x2 + 12.383x
R² = 0.9938
y = 4.7981x2 + 2.9728x
R² = 0.973
0
20
40
60
80
100
120
140
0.0 0.5 1.0 1.5 2.0 2.5 3.0
10

D
a
y

S
D
S

D
B
P

F
o
r
m
a
t
i
o
n

(
µ
g
/
L
)
DOC Concentration (mg/L)
10 Day TTHM
10 Day HAA5
TTHM MCL
HAA5 MCL
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

29
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Figure 4-7. SDS TTHM and HAA5 formation (6 days) as a function of GAC effluent DOC concentration.

y = 14.299x2 + 3.2834x
R² = 0.9903
y = 5.4109x2 + 1.6941x
R² = 0.9615
0
20
40
60
80
100
120
140
0.0 0.5 1.0 1.5 2.0 2.5 3.0
6
D
a
y

S
D
S

D
B
P

F
o
r
m
a
t
i
o
n

(
µ
g
/
L
)
DOC Concentration (mg/L)
6 Day TTHM
6 Day HAA5
TTHM Target
HAA5 Target
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

30
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Table 4-3. Instantaneous DBP Formation from Pre-chlorination
Sample DOC Chlorine Dose
Reac-
tion
Time
Chlorine
Residual CHCl3 CHBrCl2 CHBr2Cl CHBr3 TTHM MCAA DCAA DBAA TCAA HAA5
mg/L mg/L as Cl2 min mg/L as Cl2 μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L
RSSCT Influent 2.7 1.20 30 0.2 12 3.1 1.1 ND 16.2 ND 4.4 ND 2.8 7.2

Table 4-4. GAC RSSCT SDS DBP Formation
Sample DOC DOC Breakthrough Chlorine Dose Reaction Time Residual CHCl3 CHCl2Br CHClBr2 CHBr3 TTHM MCAA DCAA DBAA TCAA HAA5
mg/L % mg/L as Cl2 days mg/L as Cl2 μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L
1 0.3 11% 1.00
2 0.67 ND 0.75 0.89 ND 1.6 ND ND ND ND ND
6 0.44 0.86 1.1 1.4 0.78 4.1 ND ND ND ND ND
10 0.24 1.2 1.8 2.5 1.4 6.9 ND ND ND ND ND
2 0.79 29% 1.20
2 0.60 2.7 3.3 3.2 0.92 10.1 ND 4.4 ND ND 4.4
6 0.37 5.8 5.3 4.8 1.2 17.1 ND 5.6 1.1 1.1 7.8
10 0.05 7.6 6.2 5.4 1.3 20.5 ND 5.7 1.1 1.2 8.0
3 1.35 49% 1.55
2 1.23 8.4 5.9 4.3 0.7 19.3 ND 8.9 ND 2.2 11.1
6 0.49 18 7.9 4.4 0.73 31.0 ND 11 1.0 2.8 14.8
10 0.30 21 10 5.8 0.86 37.7 ND 10 1.0 3.8 14.8
4 1.86 68% 1.86
2 0.98 18 7.4 3.8 ND 29.2 ND 6.0 ND 4.7 10.7
6 0.54 34 9.8 4.4 0.71 48.9 ND 8.5 ND 7.9 16.4
10 0.25 38 12 5.3 0.6 55.9 ND 9.8 ND 8.1 17.9
5 2.75 100% 3.80
2 1.94 61 19 5.6 ND 85.6 2.4 18 ND 17 37.4
6 0.57 92 21 6.1 0.5 119.6 3.2 22 ND 22 47.2
10 0.21 90 23 6.7 ND 119.7 3.7 21 ND 21 45.7
Summary Report for Bench-scale Pre-ozonation and GAC RSSCT

31
Use of contents on this sheet is subject to the limitations specified at the beginning of this document. TM4 Bench Pre-ozonation and RSSCT Testing FINAL

Table 4-5. Chlorine Dose and Residual Data for All SDS Samples
Sample DOC DOC Breakthrough Chlorine Dose 2-day Residual 6-day Residual 10-day Residual
mg/L % Cl2:DOC mg/L as Cl2 mg/L as Cl2 mg/L as Cl2 mg/L as Cl2
1 0.30 11%
1:1 0.30 0.14 0.04 0.00
1.5:1 0.45 0.31 0.09 0.00
2:1 0.60 0.45 0.13 0.03
2 0.79 29%
1:1 0.80 0.64 0.27 0.09
1.5:1 1.20 0.93 0.46 0.3
2:1 1.60 1.18 0.74 0.54
3 1.35 49%
1:1 1.33 0.88 0.35 0.19
1.5:1 2.00 1.35 0.8 0.58
2:1 2.67 1.81 1.3 0.95
4 2.00 68%
1:1 1.87 1.11 0.55 0.22
1.5:1 2.80 1.67 1.01 0.64
2:1 3.74 2.54 1.8 1.36
5 2.75 100%
1:1 2.67 0.95 0.05 0.02
1.5:1 4.00 1.5 0.62 0.22
2:1 5.34 2.46 1.42 1.06
Section 5: Recommendation
Bench scale testing of pre-ozonation and GAC on raw water collected after the pre-sedimentation basis as
the GRWTP shows that GAC performs better for DBP reduction than pre-ozonation. The City of Green River
should perform a lifecycle cost analysis between the two technologies to determine the best overall solution.