Science of the Total Environment 544 (2016) 10–17

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Conventional drinking water treatment and direct biofiltration for the

removal of pharmaceuticals and artificial sweeteners: A
pilot-scale approach
Michael J. McKie ⁎, Susan A. Andrews, Robert C. Andrews
University of Toronto, Drinking Water Research Group, 35 St. George St., Toronto, Ontario M5S 1A4, Canada

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Pilot plants treating Otonabee River and

Lake Ontario, Canada water were utilized.
• 9 pharmaceuticals and 2 artificial sweet-
eners were spiked into the pilot systems.
• Conventional treatment and direct
biofiltration were examined for com-
pound removal.
• Coagulation and/or biofiltration removed
at least 7 of 9 pharmaceuticals by N 50%.
• Sweetener removal increased with in-
creasing concentrations of in-line PACl.

a r t i c l e i n f o a b s t r a c t

Article history: The presence of endocrine disrupting compounds (EDCs), pharmaceutically active compounds (PhACs) and arti-
Received 25 August 2015 ficial sweeteners are of concern to water providers because they may be incompletely removed by wastewater
Received in revised form 26 November 2015 treatment processes and they pose an unknown risk to consumers due to long-term consumption of low concen-
Accepted 26 November 2015
trations of these compounds. This study utilized pilot-scale conventional and biological drinking water treatment
Available online xxxx
processes to assess the removal of nine PhACs and EDCs, and two artificial sweeteners. Conventional treatment
Editor: D. Barcelo (coagulation, flocculation, settling, non-biological dual-media filtration) was compared to biofilters with or with-
out the addition of in-line coagulant (0.2–0.8 mg Al3+/L; alum or PACl). A combination of biofiltration, with or
Keywords: without in-line alum, and conventional filtration was able to reduce 7 of the 9 PhACs and EDCs by more than
Pharmaceuticals 50% from river water while artificial sweeteners were inconsistently removed by conventional treatment or
Endocrine disruptors biofiltration. Increasing doses of PACl from 0 to 0.8 mg/L resulted in average removals of PhACs, EDCs increasing
Artificial sweeteners from 39 to 70% and artificial sweeteners removal increasing from ~ 15% to ~ 35% in lake water. These results
Biofiltration suggest that a combination of biological, chemical and physical treatment can be applied to effectively reduce
Drinking water treatment
the concentration of EDCs, PhACs, and artificial sweeteners.
Conventional treatment
© 2015 Elsevier B.V. All rights reserved.

⁎ Corresponding author.
E-mail addresses: m.mckie@mail.utoronto.ca (M.J. McKie), sandrews@civ.utoronto.ca (S.A. Andrews), andrews@ecf.utoronto.ca (R.C. Andrews).

http://dx.doi.org/10.1016/j.scitotenv.2015.11.145
0048-9697/© 2015 Elsevier B.V. All rights reserved.
 M.J. McKie et al. / Science of the Total Environment 544 (2016) 10–17
1. Introduction
Anthropogenic contaminants of interest in drinking water include
pharmaceutically active compounds (PhACs), endocrine disrupting
compounds (EDCs) and artificial sweeteners that are poorly removed
by wastewater treatment facilities (Ternes et al., 2004). EDCs interfere
by mimicking or blocking natural hormones (USEPA, 2001) and PhACs
are used to diagnose, treat, alter, or prevent illness (USEPA, 2012).
These compounds are of interest because of their potential impacts to
the natural environment, the unknown risk they pose to human
consumers when consumed at trace levels for long periods of time
(Safe, 2004; Schwab et al., 2005), and their ability to indicate wastewa-
ter contamination.
Artificial sweeteners are used to sweeten foods and beverages
(Scheurer et al., 2010), and though they have been approved by govern-
ment health organizations worldwide, concern exists because long-
term health impacts are presently unknown (Mawhinney et al., 2011).
Artificial sweeteners have been proposed as an indicator for wastewater
in drinking water sources due to the fact that they are poorly removed
by traditional waste and drinking water treatment processes and are
only present in water sources impacted by anthropogenic activities
(Torres et al., 2011; Oppenheimer et al., 2011).
EDCs, PhACs, and artificial sweeteners have been detected in source
waters around the world and many studies have examined their remov-
al using a range of water treatment processes. Ozonation has been dem-
onstrated to be an effective method of reducing the concentration of
EDCs and PhACs, but may create unknown degradation products
(Westerhoff et al., 2005; Dodd et al., 2009). Coagulation has been exam-
ined for the removal of PhACs and EDCs; however removals are typically
low (b30%) (Diemert and Andrews, 2013). Biologically active filtration
is utilized in drinking water treatment facilities because of its ability to
degrade organic compounds while providing effective physical removal
(LeChevallier et al., 1992). Biofiltration is most commonly used in drink-
ing water treatment to remove organic carbon and disinfection by-
product precursors (Onstad et al., 2008; Chu et al., 2012). Recently,
biofiltration has been shown to remove EDCs, PhACs (Reungoat et al.,
2011), and artificial sweeteners (Mawhinney et al., 2011) from munic-
ipal wastewater and drinking water (Zearly and Summers, 2012).
Biofilters may also be enhanced with low doses of in-line coagulant to
combine physical, chemical and biological processes to improve remov-
al of large organic compounds without adversely impacting headloss
(Azzeh et al., 2015).
This study examined the removal of 9 EDCs and PhACs, as well as the
artificial sweeteners sucralose and acesulfame potassium (acesulfame-
K), through pilot-scale conventional treatment (coagulation, floccula-
tion, settling, non-biological filtration) and biofiltration (with or with-
out coagulant enhancement). The objective of the study was to
quantify the removal of these compounds from drinking water, and to
determine which treatment processes are most effective.
2. Materials and methods
2.1. Compounds of interest
Compounds monitored in this study were selected due to their
occurrence in the natural environment and to examine a range of
physical and chemical properties including: hydrophobicity, solubility,
molecular weight, and acidity (Supplemental information Table S1).
Generally, smaller and more hydrophilic compounds (low molecular
weight, low log Kow and high solubility) have been reported to be
more biodegradable (Kickham et al., 2012). Notably absent from this
group of analytes are antibiotics, such as sulfamethazine and sulfameth-
oxazole, due to the negative impact they may pose on growth within bi-
ological filters. Antibiotics are designed to prevent the growth of
bacteria, and exposing drinking water biofilters to antibiotics may im-
pair their ability to degrade other organics. If the biofilm is too thin it
11
may be unable to prevent the migration of antibiotics to active bacteria
(Cochran et al., 2000), or increased growth rates within the biofilm due
to warmer water temperatures may make the bacteria more susceptible
to the impact of antibiotic exposure (Brown et al., 1988). Based on these
factors, antibiotics were omitted from the list of assessed compounds to
ensure that the biofilm was not negatively impacted for other studies
being completed at these locations.
2.2. Source waters
Pilot-scale studies were completed at the Peterborough Water Treat-
ment Plant (Otonabee River), Peterborough, Ontario, and the R. C. Harris
Water Treatment Plant (Lake Ontario), Toronto, Ontario. Pilot plant in-
fluent water quality is shown in Supplemental information Table S2.
These two waters can be differentiated by higher temperatures and
organic concentrations in the Otonabee River, while both locations have
similar pH and raw water turbidity.
2.3. Pilot plant configurations
The pilot using Otonabee River water as its source was configured to
examine biofiltration with or without in-line alum addition
(0.2 mg Al3 +/L), as well as non-biological, conventional filtration
(Fig. 1a). Two biofilters were operated with an empty bed contact
time (EBCT) of 10 min to meet the requirements of a concurrent
study, while the conventional filter was operated at 15 min to match
full-scale operation. All filters contained 50 cm of anthracite (effective
size d10 = 0.85 mm, uniformity coefficient UC = 1.8) over 50 cm of
sand (d10 = 0.5 mm, UC = 1.8). The conventional treatment train
consisted of alum coagulation (3.0–5.0 mg Al3+/L to match full-scale),
three-stage tapered flocculation, parallel plate settling, and non-
biological filtration. One biofilter was operated without chemical addi-
tion and is referred to as a control. The second biofilter was pre-
treated with 0.2 mg Al3+/L inline alum. The biofilters were backwashed
with their own unchlorinated effluent while the conventional filter was
backwashed with chlorinated water (~ 1 mg/L) from the full-scale
clearwell. All filters were backwashed three times per week. Analytes
of interest were spiked directly into an influent constant head tank
and completely mixed prior to treatment.
A second pilot plant, illustrated in Fig. 1b, consisted of three filters
utilizing ozonated Lake Ontario water (dose = 1 mg/L, contact
time = 8 min, O3:DOC = 1:2), and was used to examine the impact of
polyaluminum hydroxychloride (PACl) dose on biofiltration. One of
the filters acted as a control, with no chemical addition, while the
other two filters received 0.2 and 0.8 mg Al3 +/L PACl, respectively.
The biofilters consisted of 50–150 cm GAC over 15–50 cm of sand and
were operated with a 16 min EBCT to match the typical full-scale flow
rate. All media had been in operation for at least 4 years prior to sam-
pling and was considered to be exhausted in terms of adsorption capac-
ity. All filters were backwashed with their own unchlorinated effluent,
at most once per week. EDCs and PhACs were dosed immediately
following ozone quenching in order to isolate their removal within the
filters and eliminate the impact of oxidation. At the Lake Ontario pilot,
analytes were spiked into a constant head tank immediately prior to
the filters.
To determine if any analyte losses were occurring as a result of
contact with plumbing materials, a filter column was operated without
media. Results showed that losses within the pilot plant were always
b12%.
2.4. Analyte spiking procedure
EDCs and PhACs were dissolved in acetonitrile as per Diemert and
Andrews (2013), and spiked into biofilter influent water to achieve a
nominal concentration of 500 ng/L. Artificial sweeteners were dissolved
in Milli-Q® water (1000 ng/L). Spiking began 7 days prior to sample
12 M.J. McKie et al. / Science of the Total Environment 544 (2016) 10–17

Fig. 1. Schematic of the Otonabee River (a) and Lake Ontario pilot plant (b).

collection to ensure consistent operation and provide biofilm organisms
time to acclimate. A Masterflex C/L tubing peristaltic pump was used to
dose the two sets of samples at a flow rate of 1 mL/h. Flow rates were
confirmed at the pilot plants for a minimum of four hours before spiking
the influent water. Samples were collected in 500 mL amber glass bot-
tles and stored at 4 °C for a maximum of 7 days prior to extraction. Filter
influent concentrations for the Otonabee River Pilot Plant are provided
in Fig. 2, and the Lake Ontario pilot plant are given in Fig. 3. The concen-
tration of these compounds was lower in the Otonabee River water,
likely due to interactions with the higher amounts of organic matter
in river water compared to the lake water. Variations in influent concen-
trations were accounted for in the calculations of percent removals of
the compounds at both locations.
3. Analytical methods
EDCs and PhACs were analyzed using LC/MS/MS based on a method
derived from the Ontario Ministry of the Environment (MOE)
EOP-E3454, version 2.0 (MOE, 2008). Surrogate compounds were
added to samples (400 mL), extracted with 6 cm3 Waters Oasis

Fig. 2. Influent water spiking results — Otonabee River (lower and upper whiskers represent the minimum and maximum values, respectively. Bottom, middle and top of box indicate 25th,
50th and 75th percentiles, respectively).
 M.J. McKie et al. / Science of the Total Environment 544 (2016) 10–17
Fig. 3. Influent water spiking results — Lake Ontario (lower and upper whiskers represent the minimum and maximum values, respectively. Bottom, middle and top of box indicate 25th,
50th and 75th percentiles, respectively).
Hydrophilic-Lipophilic Balance (HLB) solid phase extraction (SPE)
cartridges (Waters, Mississauga, ON) and eluted with methanol.
Methanol extracts were subsampled (1 mL), blown to dryness with
nitrogen, and reconstituted with 200 μL of internal standard solution
prior to analysis by LC–MS–MS. Concentrations of analytes were
determined by correlation with an eight point calibration curve. Calibra-
tion standards were prepared in matrix-matched raw water with each
set of samples and extracted in the same manner. Check standards
were prepared and analyzed after every 10th sample as a quality control
measure.
Analysis was completed at the University of Toronto using an
Agilent 1200 series pump, Agilent Poroshell EC-C18 column
(5 cm × 2.1 cm × 2.7 μm particle size), and Agilent 6460 triple
quadrupole. Neat standards were purchased from Sigma-Aldrich Inc.
(Oakville, ON) while surrogates (d4·acetaminophen, d10·carbamazepine,
d3·naproxen, d4·diclofenac, d6·gemfibrozil, and d4·clofibric acid) and in-
ternal standards (13C6·sulfamethazine phenyl for positive mode and
d14·bisphenol A for negative mode) were purchased from CDN Isotopes
(Pointe-Claire, QC).
Artificial sweeteners were analyzed using the same LC/MS/MS in-
strument used for pharmaceutical analysis and a method developed
by Hoque et al. (2014). Surrogate compounds were added to samples
(200 mL), subsequently extracted using 6 cm3 Oasis MCX solid phase
extraction (SPE) cartridges (Waters, Mississauga, ON) and eluted with
a 5% solution of ammonium hydroxide in methanol. Ammonium hy-
droxide solutions were evaporated until near dryness with nitrogen
and then reconstituted with a solution of 0.1% acetic acid in HPLC
grade water. Concentrations of analytes were determined by correlation
with an eight point calibration curve. Calibration standards were
prepared in matrix-matched raw water with each set of samples and
extracted using the same procedure. Check standards were prepared
in and analyzed after every 10th sample as a quality control measure.
Neat standards (sucralose and acesulfame-K) were purchased from
Sigma-Aldrich Inc. (Oakville, ON); internal standards (d6·sucralose
and d4·acesulfame-K) were purchased from Toronto Research
Chemicals (Toronto, ON).
Adenosine triphosphate (ATP) was quantified via LuminUltra
Deposit Surface Analysis (DSA-100C, Fredericton, NB) following the
manufacturer's instructions. DOC was measured using a persulfate wet
oxidation method as described in Standard Method 5310 D (APHA
et al., 2012) with an O-I Corporation Model 1010 TOC Analyzer (College
13
Station, Texas, USA). All other parameters were measured as per
Standard Methods (APHA et al., 2012).
Four sets of samples were collected from the Otonabee River pilot
plant and three sets of samples were collected from the Lake Ontario
pilot plant. EDC, PhAC and artificial sweetener samples were collected
and prepared in duplicate. As such, each sampling location at the
Otonabee River pilot resulted in 8 samples being analyzed; the Lake
Ontario pilot had a total of 6 samples.
4. Results and discussion
4.1. Characterization of the pilot-scale filters
To characterize filter performance, microbial activity on the biofilter
media was quantified by measuring adenosine triphosphate (ATP) and
the removal of dissolved organic carbon (DOC) was monitored through
the filters. ATP concentrations were observed to be approximately ten
times higher for the Otonabee River when compared to Lake Ontario
pilot, consistent with the differences in the nutrient levels and in their
reported biodegradable organic carbon concentrations (Pharand et al.,
2014).
Filters operated biologically at the Otonabee River pilot had higher
levels of ATP than the conventional filter backwashed with chlorinated
water. The control biofilter had an average ATP of 1800 ng/g media,
while the biofilter receiving 0.2 mg Al3 +/L alum had an average of
1080 ng/g media. These were both much higher than the conventional,
non-biological filter (average of 53 ng/g media). Similar trends were
observed for the Lake Ontario pilot but with lower observed ATP values
associated with lower organic carbon (2 mg/L vs 5 mg/L) and nutrient
levels (b0.1 mg/L vs 0.3 mg/L). The control biofilter at the Lake Ontario
pilot plant had an average ATP of 190 ng/g media, whereas the filter
receiving 0.2 mg Al3+/L PACl measured 117 ng ATP/g media, and the
filter receiving 0.8 mg Al3+/L PACl had an average ATP concentration
of 93 ng/g media. Lower ATP concentrations associated with filters
receiving more PACl is likely a result of nutrient sequestration in the
flocs (Aguilar et al., 2002); however these concentrations were all
higher than the 53 ng/g ATP measured for the Otonabee River conven-
tional filter.
While DOC removal was low (b13%) through all of the filters, there
was approximately 5% more removal through the biofilters treating
Otonabee River water. The control biofilter reduced DOC by 6.5%,
14 M.J. McKie et al. / Science of the Total Environment 544 (2016) 10–17
Fig. 4. PhAC and EDC removal — Otonabee River (vertical bars represent mean ± 1 standard deviation).
whereas the filter dosed with 0.2 mg Al3 +/L alum decreased DOC by
6.7%. Similar removals in the absence of coagulant indicated that biolog-
ical degradation was likely responsible. Filters treating Lake Ontario
water and receiving 0, 0.2, and 0.8 mg Al3+/L PACl removed 0, 2, and
13% of the DOC over the course of the study, respectively. These results
indicate that chemical processes (coagulation) were likely responsible
for DOC removal from Lake Ontario water due to the organics in the
water being less biodegradable. Differences observed between the
dominant removal mechanisms for each location may be a function of
biological activity since Otonabee River filters had ATP concentrations
ten times higher than ones treating Lake Ontario water. Higher concen-
trations of DOC in the Otonabee River suggest more biodegradable
organics were available.
4.2. Removal of pharmaceutical compounds by biological processes
Coagulation and biological treatment of Otonabee River water, ei-
ther alone or in combination, reduced all PhAC and EDC concentrations
by at least 37% (Fig. 4). Coagulation (settled water) removed an average
of 31% (range 18–56%), and subsequent non-biological filtration re-
moved an additional 8% (range 0–28%). Biofiltration with or without
0.2 mg Al3+/L alum removed an average of 44% (range 0–78%) of the
analytes. Five of the nine EDCs and PhACs (acetaminophen, estrone,
gemfibrozil, ketoprofen, and pentoxifylline) were reduced by N50% via
biofiltration and biofiltration with 0.2 mg Al3+/L alum; two of the re-
maining compounds were removed by N 50% using coagulation and
non-biological filtration (clofibric acid and naproxen). These
compounds have carboxylic acid functional groups in their structure
that can form complexes with alum (Motekaitis and Martell, 1984).
Diclofenac and carbamazepine were removed by up to 37% by
Fig. 5. Analyte removal — Lake Ontario Water.
conventional treatment; no removal was observed in the control
biofilter, likely because they are among the least soluble compounds
(Supplemental information Table S1). This makes it more difficult for
these compounds to be taken into cells for biodegradation (USGS,
2014). Biofilter enhancement with 0.2 mg Al3+/L alum improved re-
moval of diclofenac and carbamazepine, averaging 22% and 8%,
respectively.
Trials at the Lake Ontario pilot plant focused on the impact of coag-
ulant dose prior to biofiltration. Three polyaluminum hydroxychloride
(PACl) dosages were examined (0 mg Al3 +/L, 0.2 mg Al3 +/L, and
0.8 mg Al3+/L) to serve as a control, replicate experimental conditions
at the Otonabee River, and mimic the dose utilized at full-scale, respec-
tively. Although results were variable for small (0.2 mg Al3+/L) addi-
tions of PACl, increasing PACl dosage typically improved the removal
of most of the PhACs and EDCs (Fig. 5). The control biofilter, with no
PACl added, removed two of the nine compounds by more than 50%
(average removal was 39%, range 7–86%). Following inline addition of
0.2 mg Al3+/L PACl, the biofilter reduced EDCs and PhACs by 12–87%
with an average removal of 45%. Increasing PACl to 0.8 mg Al3 +/L
improved average reductions to 70%, with eight of the nine compounds
reduced by 50% or more (39–91%).
Based on these results, no strong predictors of contaminant removal
efficiency due to physical properties such as molecular weight, hydro-
phobicity/hydrophilicity, pKa or solubility were identified through
linear or multiple linear regressions. Possible complications may include
complexation with natural organic matter; changes in the biological
community throughout the study; or adsorption of the compounds to
activated carbon or the biofilm. Notably, the removal of naproxen and
carbamazepine decreased following the addition of 0.2 mg Al3 +/L
PACl to Lake Ontario water. Although the cause of this response is not
 M.J. McKie et al. / Science of the Total Environment 544 (2016) 10–17 15

Table 1
Pharmaceutical removals compared to existing literature.

Otonabee River Lake Ontario Literature results

Compound Conventional filtration Control biofiltration No PACl 0.2 mg/L PACl 0.8 mg/L PACl Biological filtration processes Coagulation

Acetaminophen 46% 68% 82% 86% 91% 59 ± 11%1 b20%4

Carbamazepine 35% 5.2% 39% 25% 80% 0.5 ± 1.1%1 32%5
clofibric acid 52% 36% 14% NA 39% 35 ± 6%1 b20%6
Diclofenac 36% 8.7% 12% 27% 74% 21 ± 1%1 b20%4,6
Estrone 33% 60% 86% 87% 77% b1%2 b20%4
Gemfibrozil 29% 73% 6.8% NA 58% 70 ± 7%1 31–37%5
Ketoprofen 38% 78% 29% 28% 63% NA b20%6
Naproxen 52% 29% 33% 12% 72% 75%3 b20%4,6
Pentoxifylline 24% 74% 45% 60% 77% 13%2 b20%4

NA = not available.
1
Zearly and Summers (2012).
2
Snyder et al. (2007).
3
Hallé (2009).
4
Westerhoff et al. (2005).
5
Diemert and Andrews (2013).
6
Simazaki et al. (2008).

known for certain, it is possible that the sub-optimal dose of PACl
altered the surface charge or molecular size of the complex to
prevent the diffusion of the molecule across the membrane layer,
and prevented the biodegradation or adsorption of these complexes as
a result.
Despite significant differences between Otonabee River and Lake
Ontario water quality and biological activity on the filter media, the
calculated percent removals were similar. The control filter at the
Otonabee River pilot plant achieved an average removal of 44%
(0–78%) while the control filter at the Lake Ontario pilot plant averaged
39% (7–85%). These results suggest that these compounds are more
easily biodegraded than bulk DOC, and biological activity may not be
proportional to the expected removals. A comparison to results pub-
lished by others is shown in Table 1.
Zearly and Summers (2012) examined the removal of 34 organic
micropollutants through biologically active sand filters operated at 7.5
and 15 min. Using sand media, as opposed to activated carbon, lessens
the possibility of chemical adsorption and allowed easier quantification
of the impact of biological degradation. The same authors reported re-
sults similar to those obtained in this study. Based on the design of
their system they were able to calculate first-order degradation rates
for the biodegradable components, implying that the removal of these
micropollutants at each location could be improved with longer
EBCTs. However, these columns were lab-scale with raw water collected
approximately every 2 days which may impact the results when scaling
up to pilot or full-scale operation. Snyder et al. (2007) summarized the
results of biofilter pilot testing utilizing a 2 min EBCT. This is much lower
than the EBCT used in this study, so it is not surprising that results of this
study were more effective. Hallé (2009) examined anthracite biofilters
operated with a 14 min EBCT. However, Hallé (2009) allowed the filters
to acclimate with elevated concentrations of micropollutants for a
period of ten months, as opposed to the week provided in this study.
As such, increased removals of naproxen may be expected if biological
organisms are allowed extended acclimation time.
Westerhoff et al. (2005) examined the removal of spiked PhACs and
EDC from river water. Tests were completed at ambient pH (6.8) with
alum doses of 6.3 mg Al3+/L. However, their study only examined floc-
culation and settling, without subsequent filtration processes. As such,
observed removals through conventional treatment were likely higher
due to filtration of the settled water. Similarly, the studies completed
by Diemert and Andrews (2013) and Simazaki et al. (2008) were com-
pleted using bench-scale jar tests to evaluate the impact of coagulation,
but were unable to determine the combined benefits of coagulation and
filtration.
Results from the Otonabee River control biofilter closely matched
results from existing literature except for estrone and pentoxifylline.
Additional research may be required to determine best estimates for
their removal with biofiltration and/or the metabolic pathways leading
to their improved removal. Many of the compounds were removed
more effectively from Lake Ontario when compared to the literature.
Higher removals were likely a function of adsorption to the GAC
media (which was previously assumed to be exhausted as it was in
use for more than four years prior to testing), or to the biomass in the
filters.

Fig. 6. Removal of acesulfame-K — Otonabee River (NR = no removal).

16 M.J. McKie et al. / Science of the Total Environment 544 (2016) 10–17
Fig. 7. Sweetener removal as a function of PACl dose — Lake Ontario.
4.3. Impact of coagulation on removal of artificial sweeteners
Removal of two artificial sweeteners, acesulfame-K and sucralose,
was compared using coagulation and biofiltration. In general, the ob-
served removals of approximately 25% were lower than the removals
of PhACs and EDCs at the same location.
Acesulfame-K was best removed from the Otonabee River by treat-
ment that included the addition of alum, likely because it is generally
negatively-charged (pKa = 2.0) and may form a complex (Froloff
et al., 1998). Coagulation alone removed 19% (70 ng/L) in contrast to
the 8.1% (30 ng/L) removed by biofiltration including 0.2 mg Al3 +/L
inline alum (Fig. 6). Both conventional non-biological filtration, and
the control biofilter reduced acesulfame-K concentrations by only 3%
(9 ng/L). Only conventional coagulation showed consistent removals
of acesulfame-K throughout the study.
Sampling was completed over a period of 5 months in an effort to
examine the impact of seasonal changes on the removal of
acesulfame-K. No clear trend was observed through the four sampling
events. Scheurer et al. (2010) reported biological treatment and coagu-
lation to be poor methods for the removal of acesulfame-K, and that
these processes can be impacted by changes in water temperature.
Based on their inherently poor removal capabilities, it is not unexpected
that these processes were unable to remove acesulfame-K efficiently.
Sucralose was poorly removed from Otonabee River water by any of
the treatment processes examined. None of the treatments provided
consistent results (Supplemental information Fig. S1), similar to results
presented by others (Scheurer et al., 2010). Scheurer et al. (2010) exam-
ined the removal of sucralose by biological treatment (bank filtration),
coagulation and flocculation, ozonation, GAC adsorption and chlorina-
tion. Of these five processes they reported that only GAC adsorption
provided moderate to good removal of sucralose. Other chemical and
biological processes were ineffective removal strategies, as further con-
firmed by the results of this study. Furthermore, there did not appear to
be a seasonal impact on the removal of sucralose, which was expected
since processes that may be impacted by seasonal changes (chemical
and biological) were unable to remove this compound.
Removal of both sweeteners via coagulation and biofiltration for
Lake Ontario water were dependent on PACl dose (Fig. 7). The control
biofilter removed an average of 15% acesulfame-K (range 0–38%) as
well as 18% sucralose (range 0–36%), whereas the biofilter receiving
0.8 mg/L PACl removed 33% acesulfame-K (range 14–50%) and 39%
sucralose (range 26–58%). Only one sample was obtained for the filter
receiving 0.2 mg/L PACl (14% and 23% removals of acesulfame-K and
sucralose, respectively). A strong relationship was evident between
the removal of these artificial sweeteners and the in-line PACl dose
(Fig. 7). These results indicate that the combination of biological, chem-
ical and physical processes may be effective for the reduction of artificial
sweeteners. Further studies should be completed to determine the
impact of PACl doses above 0.8 mg Al3 +/L, and to identify possible
degradation pathways.
5. Conclusions
This study examined removal of nine EDCs and PhACs, and two
artificial sweeteners at pilot-scale. Generally, biofiltration, with or with-
out the addition of in-line coagulant was able to lower concentrations of
PhACs and EDCs in treated water. When considering Otonabee River
water a combination of conventional and biological treatment was
able to reduce seven of the nine EDCs and PhACs by more than 50%.
For Lake Ontario water, biofiltration alone reduced nine of the
compounds by an average of 39%; addition of PACl (up to 0.8 mg/L)
increased the average removals to 70%.
Acesulfame-K and sucralose removal averaged 19% over the study.
Removals in Lake Ontario water were proportional to PACl dose with
higher doses performing better (up to 32% and 39% for acesulfame-K
and sucralose, respectively).
Acknowledgments
This work was funded by the Natural Sciences and Engineering
Research Council of Canada (NSERC) Chair in Drinking Water Research
at the University of Toronto. The authors are grateful to Jules Carlson
and Isabelle Netto (University of Toronto) for their assistance in
conducting the PhAC, EDC and artificial sweetener analyses; John
Armour and colleagues at the Peterborough Utilities Commission for
their pilot plant assistance; and Dave Scott and colleagues at Toronto
Water for their pilot plant efforts.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.scitotenv.2015.11.145.
References
Aguilar, M.I., Llorens, J.S.M., Soler, A., Ortuno, J.F., 2002. Nutrient removal and sludge
production in the coagulation–flocculation process. Water Res. 36, 2910–2919.
American Public Health Association (APHA), American Water Works Association
(AWWA), Water Environment Foundation (WEF), 2012. In: Eaton, A.D., Clesceri,
L.S., Rice, E.W., Greenberg, A.E. (Eds.), Standard Methods for the examination of
water and wastewater, 20th ed. (Washington, USA).
Azzeh, J., Taylor-Edmond, L., Andrews, R.C., 2015. Engineered biofiltration for ultrafiltra-
tion fouling mitigation and disinfection by-product precursor control. Water Sci.
Technol. 15 (1), 124–133.
Brown, M.R.W., Allison, D.G., Gilbert, P., 1988. Resistance of bacterial biofilms to
antibiotics: a growth-rate related effect? J. Antimicrob. Chemother. 22 (6), 777–780.
Chu, W., Gao, N., Yin, D., Deng, Y., Templeton, M.R., 2012. Ozone-biological activated
carbon integrated treatment from removal of precursors of halogenated nitrogenous
disinfection by-products. Chemosphere 86 (11), 1087–1091.
 M.J. McKie et al. / Science of the Total Environment 544 (2016) 10–17
Cochran, W.L., McFeters, G.A., Stewart, P.S., 2000. Reduced susceptibility of thin
Pseudomonas aeruginosa biofilms to hydrogen peroxide and monochloramine.
J. Appl. Microbiol. 88 (1), 22–30.
Diemert, S., Andrews, R.C., 2013. The impact of alum coagulation on pharmaceutically
active compounds, endocrine disrupting compounds and natural organic matter.
Water Sci. Technol. Water Supply 13 (5), 1348–1357.
Dodd, M.C., Kohler, H.E., Von Gunten, U., 2009. Oxidation of antibacterial compounds by
ozone and hydroxyl radical: elimination of biological activity during aqueous ozona-
tion processes. Environ. Sci. Technol. 43, 2498–2504.
Froloff, N., Lloret, E., Martinez, J.-M., Faurion, A., 1998. Cross-adaption and molecular
modeling study of receptor mechanisms common to four taste stimuli in humans.
Chem. Sens. 23 (2), 197–206.
Hallé, C., 2009. Biofiltration in drinking water treatment: reduction of membrane fouling
and biodegradation of organic trace contaminants (PhD Thesis) University of Water-
loo, Waterloo, ON.
Hoque, M.E., Cloutier, F., Arcieri, C., McInnes, M., Sultana, T., Murray, C., Vanrolleghem,
P.A., Metcalfe, C.D., 2014. Removal of selected pharmaceuticals, personal care prod-
ucts, and artificial sweetener in an aerated sewage lagoon. Sci. Total Environ. 487,
801–812.
Kickham, P., Otton, S.V., Moore, M.M., Ikonomou, M.G., Gobas, F.A.P.C., 2012. Relationship
between biodegradation and sorption of phthalate esters and their metabolites in
natural sediments. Environ. Toxicol. Chem. 31 (8), 1730–1737.
LeChevallier, M.W., Becker, W.C., Schorr, P., Lee, R.G., 1992. Evaluating the performance of
biologically active rapid filters. J. - Am. Water Works Assoc. 84 (4), 136–146.
Mawhinney, D.B., Young, R.B., Vanderford, B.J., Borch, T., Snyder, S.A., 2011. Artificial
sweetener sucralose in U.S. drinking water systems. Environ. Sci. Technol. 45,
8716–8722.
Motekaitis, R.J., Martell, A.E., 1984. Complexes of aluminum(III) with hydroxyl carboxylic
acids. J. Am. Chem. Soc. 23 (1), 18–23.
Onstad, G.D., Weinberg, H.S., Krasner, S.W., 2008. Occurrence of halogenated furanones in
U.S. drinking waters. Environ. Sci. Technol. 42 (9), 3341–3348.
Ontario Ministry of the Environment (MOE), 2008. The determination of emerging
organic pollutants in environmental matrices by LC/MS/MS. Laboratory services
branch method EOP-E3454, Etobicoke, ON, Canada.
Oppenheimer, J., Eaton, A., Badruzzaman, M., Haghani, A.W., Jacangelo, J.G., 2011. Occur-
rence and suitability of sucralose as an indicator compound of wastewater loading to
surface waters in urbanized regions. Water Res. 45, 4019–4027.
17
Pharand, L., Van Dyke, M.I., Anderson, W.B., Huck, P.M., 2014. Assessment of biomass in
drinking water biofilters by adenosine triphosphate. J. Am. Water Works Assoc. 106
(10), E433–E444.
Reungoat, J., Escher, B.I., Macova, M., Keller, J., 2011. Biofiltration of wastewater treatment
plant effluent: effective removal of pharmaceuticals and personal care products and
reduction of toxicity. Water Res. 45, 2751–2762.
Safe, S., 2004. Endocrine disruptors and human health: Is there a problem? Toxicology
205, 3–10.
Scheurer, M., Storck, F.R., Brauch, H., Lange, F.T., 2010. Performance of conventional multi-
barrier drinking water treatment plants for the removal of four artificial sweeteners.
Water Res. 44, 3573–3584.
Schwab, B.W., Hayes, E.P., Fiori, J.M., Mastrocco, F.J., Roden, N.M., Cragin, D., Meyerhoff,
R.D., D'Aco, V.J., Anderson, P.D., 2005. Human pharmaceuticals in US surface waters:
a human health risk assessment. Regul. Toxicol. Pharmacol. 42, 296–312.
Simazaki, D., Fujiwara, J., Manabe, S., Matsuda, M., Asami, M., Kunikane, S., 2008. Removal
of selected pharmaceuticals by chlorination, coagulation-sedimentation and
powdered activated carbon treatment. Water Sci. Technol. 58 (5), 1129–1135.
Snyder, S.A., Wert, E.C., Lei, H.D., Westerhoff, P., Yoon, Y., 2007. Removal of EDCs and
pharmaceuticals in drinking and reuse treatment processes. Water Research Foundation,
Denver, CO.
Ternes, T.A., Joss, A., Siegrist, H., 2004. Scrutinizing pharmaceuticals and personal care
Products in wastewater treatment. Environ. Sci. Technol. 38 (20), 392A–399A.
Torres, C.I., Ramakrishna, S., Chiu, C.-A., Nelson, K.G., Westerhoff, P., Krajmalnik-Brown, R.,
2011. Fate of sucralose during wastewater treatment. Environ. Eng. Sci. 28 (5),
325–331.
USEPA, 2001. Removal of endocrine disruptor chemicals using drinking water treatment
processes. Office of Research and Development, Washington, DC (EPA/625/R-00/
015).
USEPA, 2012. Pharmaceuticals and personal products Retrieved on July 19, 2013 from
USEPA Website http://www.epa.gov/ppcp/.
USGS, 2014. Biodegradation. Retrieved on May 20, 2015 from United States Geological
Survey Website. http://toxics.usgs.gov/definitions/biodegradation.html.
Westerhoff, P., Yoon, Y., Snyder, S., Wert, E., 2005. Fate of endocrine-disruptor, pharma-
ceutical, and personal care product chemicals during simulated drinking water treat-
ment processes. Environ. Sci. Technol. 39, 6649–6663.
Zearly, T.L., Summers, R.S., 2012. Removal of trace organic micropollutants by drinking
water biological filters. Environ. Sci. Technol. 46, 9412–9419.