Science of the Total Environment 512513 (2015) 637644

Contents lists available at ScienceDirect

Science of the Total Environment

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

Characterization of hydraulic fracturing owback water in Colorado:

Implications for water treatment
Yaal Lester a, Imma Ferrer b, E. Michael Thurman b, Kurban A. Sitterley a, Julie A. Korak a,
George Aiken c, Karl G. Linden a,
a
b
c

Department of Civil, Environmental, and Architectural Engineering, University of Colorado, UCB 428, Boulder, CO 80309, United States
Center for Environmental Mass Spectrometry, University of Colorado, UCB 428, Boulder, CO 80309, United States
US Geological Survey, 3215 Marine Street, Boulder, CO, 80303, United States

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

Flowback from the DJ basin (CO) was

analyzed using a suite of analytical tools.
The owback contained salts, metals
and high levels of dissolved organic
matter.
Organic matter was comprised of fracturing uid additives and degradation
products.
Water analysis and treatability tests
served to indicate suitable treatments.
Example: aeration/precipitation/disinfection deemed suitable for
owback recycling

a r t i c l e

i n f o

Article history:
Received 9 December 2014
Received in revised form 17 January 2015
Accepted 17 January 2015
Available online 3 February 2015
Editor: D. Barcelo
Keywords:
Hydraulic fracturing
Flowback
Fracking
Water quality
Water treatment
Water reuse
DenverJulesburg basin

a b s t r a c t
A suite of analytical tools was applied to thoroughly analyze the chemical composition of an oil/gas well owback
water from the DenverJulesburg (DJ) basin in Colorado, and the water quality data was translated to propose effective treatment solutions tailored to specic reuse goals. Analysis included bulk quality parameters, trace organic
and inorganic constituents, and organic matter characterization. The owback sample contained salts (TDS =
22,500 mg/L), metals (e.g., iron at 81.4 mg/L) and high concentration of dissolved organic matter (DOC = 590
mgC/L). The organic matter comprised fracturing uid additives such as surfactants (e.g., linear alkyl ethoxylates)
and high levels of acetic acid (an additives' degradation product), indicating the anthropogenic impact on this
wastewater. Based on the water quality results and preliminary treatability tests, the removal of suspended solids
and iron by aeration/precipitation (and/or ltration) followed by disinfection was identied as appropriate for
owback recycling in future fracturing operations. In addition to these treatments, a biological treatment (to remove
dissolved organic matter) followed by reverse osmosis desalination was determined to be necessary to attain water
quality standards appropriate for other water reuse options (e.g., crop irrigation). The study provides a framework
for evaluating site-specic hydraulic fracturing wastewaters, proposing a suite of analytical methods for characterization, and a process for guiding the choice of a tailored treatment approach.
2015 Elsevier B.V. All rights reserved.

Corresponding author.
E-mail address: karl.linden@colorado.edu (K.G. Linden).

http://dx.doi.org/10.1016/j.scitotenv.2015.01.043
0048-9697/ 2015 Elsevier B.V. All rights reserved.

638

Y. Lester et al. / Science of the Total Environment 512513 (2015) 637644

1. Introduction
Increased use of hydraulic fracturing for unconventional gas
exploration has resulted in substantial interest in the environmental
impacts of this technology. One of the major concerns is the generation
of large volumes of owback water, and its potential adverse effects on
the environment and on human health (Vengosh et al., 2014). During
hydraulic fracturing, up to 4 million gal of water-based uid is injected
into a single drilled well, to initiate and expand fractures, as well as for
proppant transport (i.e., sand or ceramic materials); of which 1070% is
subsequently recovered as owback (API, 2010). This impaired water
stream contains high concentrations of potentially hazardous organic
and inorganic constituents and requires some degree of treatment
before the water is reused or discharged into the environment.
Until recent years, the most popular treatment for owback was
the disposal of the water into deep injection wells. However, this technique is becoming less viable due to limited access to disposal wells in
many drilling locations, and increased pressure from regulatory authorities and the public for a more sustainable solution (Gregory et al., 2011;
Lutz et al., 2013). These pressures, along with stresses on local water
resources and potential cost savings, have led the industry to search
for alternative treatment solutions, many of which would enable benecial water reuse (Mauter et al., 2014).
The choice of suitable treatment approaches depends largely on the
composition of the owback water, which is typically a mixture of the
original injected additives as well as substances from the subsurface formation that accumulate in the water down-hole. The original additives,
injected to facilitate the fracturing process and prevent problems during
well operation mainly include gelling agents, friction reducing polymers, corrosion and scale inhibitors, and biocides (FracFocus, 2014).
Some of these chemicals are recovered unchanged, while others either
react under the conditions of the fracturing process or are microbially
altered to form degradation and reaction products (Orem et al., 2014).
Olsson et al. (2013), for instance, found high concentrations of acetate
and formate in owback water from two different fracturing sites in
Germany (up to 480 and 190 mg/L, respectively), whereas these substances were present at extremely low concentrations in formation
water from a well unimpacted by hydraulic fracturing. The researchers
concluded that acetate and formate were likely the degradation products of polymers used in the fracturing uid.
Substances in owback water coming from the subsurface may
include salts, metals and soluble organic compounds, depending
on the location and nature of the geological formation. Acharya
et al. (2011) showed large variations in the concentration of total
dissolved solids (TDS) in owback water from different shales,
with average values ranging from 13,000 mg/L (Fayetteville shale)
to 120,000 mg/L (Marcellus shale). Boschee (2014) reported TDS
levels in the range of 20,00065,000 mg/L for owback and produced water from the DenverJulesburg (DJ) basin, and 150,000
300,000 mg/L for water from the Bakken formation.
Clearly, understanding the composition of the owback water is
essential for assessing its reuse potential and for implementing a suitable treatment process to improve water sustainability. The goals of
this research were to (i) demonstrate the use of a suite of analytical
tools for an in depth analysis of hydraulic fracturing owback, using
the DJ basin (CO) as an example, and, (ii) use the analytical data to
propose suitable strategies for the treatment of the owback water for
benecial reuse. In addition, the paper provides information on the
composition of owback water in the DJ basin, which is presently highly
underexplored (relative to other basins).
2. Material and methods
The owback water was a composite sample taken by the well operator and delivered under a non-disclosure agreement, mainly with
regard to sampling period, specic well location, and composition of

the original fracturing uid. The sample was ltered upon receipt with
a 0.45 m cellulose acetate lter (prewashed with 1 L deionized water
and 200 mL owback) and analyzed for general quality characteristics,
inorganic constituents, and trace organic compounds. The selection of
examined parameters was based on a list of potential owback constituents, established by Hayes (2009) following comments received from
regulatory agencies and members of the oil and gas industry.

2.1. General quality parameters

Dissolved organic carbon (DOC) was measured after dilution using a
TOC-VSCH analyzer (Shimadzu Corp., Japan). Chemical oxygen demand
(COD) was measured using HACH (Loveland, CO) COD kits and a
DR5000 spectrophotometer (after water dilution). Biochemical oxygen
demand (BOD5) was analyzed according to standard method SM5210B.
Volatile Fatty Acids (e.g., acetic acid) and inorganic anions (e.g., Cl, Br)
were measured using ion chromatography (IC Dionex 4500I). Total
recoverable phenolics were measured in accordance with EPA method
420.4. Other general quality parameters were analyzed using either
EPA or standard methods (see Table 1). Analyses were conducted at
least in duplicates in most cases, with relative standard deviations typically below 10%. Naturally occurring radioactive materials (NORM),
occasionally detected in owback, were not analyzed in the present
study, because they are not considered owback contaminants of concern in Colorado (STRONGER, 2011).

2.2. Elemental analysis

Selected elements were analyzed at the Laboratory for Environmental and Geological Studies LEGS (Department of Geological Sciences,
University of Colorado at Boulder, CO), using a Perkin Elmer SCIEX inductively coupled plasma mass spectrometer (ICP-MS, Elan DRC-e).
Samples were rst diluted by 10, 100 and 1000, and the average
value is reported (for results above the analytical method detection
limit [MDL]). Concentrations of the major ions (e.g., Ca2+, Mg2+, Na+)
were validated by inductively coupled plasma optical emission spectroscopy (ICP-OES, ARL 3410+, Thermo Scientic).

Table 1
General quality parameters, inorganic ions and select organic compound concentrations
for the owback sample.
Parameter

Analytical method

Units

Bulk parameters
pH
DOC
Alkalinity
TDS
COD
BOD5
Oil & grease
TSS

Value

SM 5310B
HACH 8203
SM 2540C
EPA 5220D
SM 5210B
EPA 1664A
SM 2540D

mgC/L
mg CaCO3/L
mg/L

Inorganic ions
Ammonia as N
NO3 as N
Chloride (Cl)
Cyanide, total
Bromide (Br)
Sulde (S2)
Sulfate (SO42)

EPA 320.1
IC
IC
EPA 335.4
IC
SM 4500 S2 D
IC

mg/L

24.7
5.2
13,600
0.055
87.2
0.31
1.3

Others
Total recoverable phenolics
Acetic acid
n-Butyric acid
Propionic acid

EPA 420.4
VFA-IC
VFA-IC
VFA-IC

mg/L

1.4
1600
19
33

6.8
590
150
22,500
1218
1100
59
360

Y. Lester et al. / Science of the Total Environment 512513 (2015) 637644

2.3. Organic constituents

Volatile and semi-volatile organic compounds (VOCs and SVOCs)
were analyzed using gas chromatography coupled with mass spectrometry (VOCs Hewlett Packard 5890 GC and 5972 MSD; SVOCs Agilent
6890N GC and Hewlett Packard 5973 MSD), in accordance with EPA
methods 8260B (volatiles) and 8270C (semi-volatiles). Blank sample
and external standard calibration were used for analyte quantication. Trace organic compounds (TOrCs) were detected using ultra
high denition quadrupole time-of-ight mass spectrometry (LC
QTOF-MS) Model 6540 Agilent (Agilent Technologies, Santa Clara,
CA, USA). Accurate mass spectra were recorded across the range
501000 m/z at 2 GHz (Thurman et al., 2014). Quality control of
the LCQTOF-MS is detailed in the Supporting information (text S1).
Fractionation and characterization of dissolved organic matter
(DOM) were performed at the U.S. Geological Survey laboratory
(Boulder, CO), using hand-packed XAD resin columns, following a
procedure described by Aiken et al. (1992). In brief, 1 L of diluted
owback sample (DOC ~ 10 mg/L) was acidied (pH 2) and passed
through XAD-8 and XAD-4 columns connected in series. Each column
was subsequently back-eluted using NaOH, and the concentration of
the DOC was measured in each fraction. This procedure enables the
fractionation of DOM into hydrophobic acids and neutrals (HPOA and
HPON), transphilic acids and neutrals (TPIA and TPIN), and low
molecular weight hydrophilic compounds (HPI). The fractionated
samples were additionally characterized by light absorption and
uorescence spectroscopy using a Varian Cary-Bio100 spectrophotometer (Agilent Technologies, Santa Clara, CA) and a spectrouorometer
(Horiba Aqualog, NJ), as described in detail in the Supporting information
(text S2).
2.4. Treatability tests
Several treatability tests were carried out to assess the potential of
biological processes and advanced oxidation for the removal of DOM
from the owback water. A detailed description of the experiments is
given in text S3 in the Supporting information. In brief, activated sludge
tests were conducted at the University of Colorado at Boulder, using a
bench-scale Sequencing Batch Reactor (SBR) and acclimated sludge
from Boulder (CO) 75th Street Wastewater Treatment Facility. UV/
H2O2 treatment was evaluated in a medium pressure (MP) quasicollimated UV apparatus (Calgon Carbon Corp., Pittsburgh, PA), using
H2O2 concentrations up to 100 mg/L (in the upper range of a UV/H2O2
application; Lester et al., 2010) and UV dose up to 3000 mJ/cm2. O3/
H2O2 experiments were carried out in a semi-batch 1 L ozone reactor,
with ozone gas continually sparged into the owback solution, to
achieve an ozone dose of up to 1 mg O3/mg DOC (in the upper range
of wastewater application; Zimmermann et al., 2011). Hydrogen peroxide was added at 350 mg/L, to obtain O3/H2O2 w/w ratio of 1.4 (typical
for O3/H2O2; Mller and Jekel, 2001). Experiments combining solar light
and chlorine were carried out in a 1000 W Xe lamp solar simulator
(Model Sol1A, Oriel Instruments, Stratford, CT), using up to 80 mM
NaOCl (Chan et al., 2012) and irradiation time up to 24 h. The same
solar simulator and exposure times were used during photo-Fenton experiments, by adjusting the pH of the owback water to 3 and adding
H2O2 at 500 mg/L (using indigenous iron (ca. 80 mg/L) and Fe:H2O2 in
a molar ratio of 1:10; Shemer et al., 2006). All experiments were conducted in duplicate at a relatively constant temperature of 23 3 C.
3. Results and discussion
3.1. General quality parameters
The owback water general quality parameters are summarized in
Table 1. The concentration of TDS was measured as 22,500 mg/L,
which is in the typical range for owback and produced water from

639

the DJ basin (20,00065,000 mg/L; Boschee, 2014). Concentration of

Cl corresponds to approximately 60.5 wt.% of the TDS, which is in excellent agreement with the theoretical value of 60.8 wt.% for TDS composed exclusively of NaCl. On the other hand, Na + concentration
corresponds to only ~ 30.9 wt.% of TDS, instead of the expected
39.5% for NaCl (see Table 2). This difference may be explained by
the contribution of other salts such as KCl, CaCl2, and MgCl2 to the
water TDS, as implied by the high concentrations of K+, Ca2+, and
Mg2+ (Table 2). Other key quality parameters i.e., DOC 590 mg/L,
pH 6.8, oil & grease 59 mg/L, and ammonia 24.7 mg/L-N, were all
within typical ranges previously reported for owback waters
(Gregory et al., 2011; Hayes, 2009).
Acetic acid was detected at an extremely high concentration of
1600 mg/L (Table 1). High levels of acetic acid were previously reported
in fracturing owback from the Marcellus shale (Wolford, 2011) and
from different sites in Germany (Olsson et al., 2013). Although this constituent is often used at trace amount as a fracturing uid additive, to
prevent the precipitation of metal oxides (FracFocus, 2014), its high
concentration in the owback sample is more likely the result of anaerobic microbial degradation of the uid's biopolymers down-hole
(Olsson et al., 2013; Wolford, 2011). Alternatively, acetic acid can be
naturally found in formation water from biogenic gas deposits (which
is not the case for the sample reported on here), as an intermediate substrate for acetoclastic methanogenesis (Dahm et al., 2013). The high
concentration of acetic acid can further explain the high level of BOD5
and BOD5/COD ratio (i.e., 0.9), indicative of high biodegradability of
the water's organic matter (Metcalf & Eddy, 2003).
3.2. Elemental analysis
Concentrations of key elements in the owback water sample are
summarized in Table 2 (the full list is given in Table S1 of the
Supporting information). Sodium (Na) is by far the dominant metal
(i.e., 6943.9 mg/L), followed by Ca, Mg and K. Iron (Fe) was also detected
at relatively high concentrations (i.e., 81.4 mg/L), which might be of concern if the water was recycled for fracturing operations (due to potential
scaling). The concentration of other ions frequently found in fracturingrelated wastewater and known for their scaling potential, such as Ba
and Sr, was relatively low in the owback water sample.
Toxic heavy metals of environmental and health concern, often
associated with industrial wastewaters (e.g., Cr, Cu, Ni), were either
not detected in the owback sample or, detected at relatively low

Table 2
Elements identied in the owback water sample.
Metal

Symbol

Concentration
(mg/L)

Aluminum
Arsenic
Boron
Barium
Calcium
Chromium
Cesium
Copper
Iron
Potassium
Lithium
Magnesium
Manganese
Sodium
Nickel
Rubidium
Silicon
Strontium
Titanium
Vanadium
Zinc

Al
As
B
Ba
Ca
Cr
Cs
Cu
Fe
K
Li
Mg
Mn
Na
Ni
Rb
Si
Sr
Ti
V
Zn

0.064
0.067
3.105
8.542
524.1
0.058
0.073
0.288
81.42
101.3
3.519
106.4
1.471
6943.9
0.042
0.230
19.65
60.25
0.028
0.120
0.051

640

Y. Lester et al. / Science of the Total Environment 512513 (2015) 637644

concentrations (b1 mg/L). These levels are in the lower range for
owback water (Hayes, 2009) and within the range typically found in
municipal wastewater treatment plants (Karvelas et al., 2003; Olujimi
et al., 2012).
3.3. Organic constituents
3.3.1. Volatile organic compounds
The owback sample was analyzed for over 70 different VOCs, including compounds typically detected in gas and oil affected water
(e.g., benzene, toluene, ethylbenzene, and xylene BTEX) and other
chemicals. The full list is provided in the Supporting information
(Table S2). Of the analyzed compounds, only (total) xylenes, acetone
and 2-butanone were detected at levels above the MDL (Table 3). Xylenes are aromatic hydrocarbons commonly detected in water associated with oil and gas producing formations. Acetone and 2-butanone are
solvents commonly used in industry for cleaning purposes (Hayes,
2009). In addition, their formation may result from microbial degradation of the fracturing uid's biopolymers downhole. Both cases imply
that these compounds are of anthropogenic origin.
3.3.2. Semivolatile organic compounds
A total of 112 SVOCs were analyzed (Table S3 in Supporting
information). Less than 10% of the compounds were detected in the
owback sample (Table 3), all at trace concentrations of b1 mg/L.
Some of the detected SVOCs are commonly associated with fossil fuel
related water (e.g., methylated phenols; Hayes, 2009), and some are
considered human carcinogens (e.g., 1,4-dioxane; U.S. EPA, 2009).
3.3.3. Trace organic contaminants
Total LCTOF-MS ion chromatogram obtained after the analysis of
the sample in positive ion mode is presented in Fig. 1. Multiple peaks
can be seen in this gure, indicating the presence and detection of
numerous trace organic compounds. One of these compounds was identied as cocamidopropyl dimethylamine, a commercial surfactant
widely used in cosmetics and detergents. This compound was identied
by its accurate mass of 229.2274, corresponding to the protonated molecule and elemental formula of C13H28N2O.
Further conrmation of the presence of this compound was
achieved using MSMS fragmentation (Fig. S1 in Supporting information).
Two major fragments were observed at m/z 184.1696 and 127.1117,
which originate from diagnostic losses of these type of compounds.
Cocamidopropyl dimethylamine is usually used as a mixture of homologues ranging from C6 to C14 alkyl chains. In this case, only the C7
alkyl chain was observed, no other homologues from this family of compounds were seen in this sample. The estimated concentration in this
sample was in the mg/L range. It is believed that this compound is
Table 3
Volatile and semivolatile organic compounds identied in the owback water sample.
Compound

Concentration (g/L)

VOCs
Acetone
2-Butanone
Xylenes, total

16,000
240
30

SVOCs
1,4-Dioxane
2-Methylphenol
3&4 methylphenol
2-Methylnaphthalene
Dimethyl phthalate
Phenanthrene
Pyrene
Butyl benzyl phthalate
Bis(2-ethylhexyl) phthalate
Phenol
2,4-Dimethylphenol

60
150
170
4
15
3
0.9
4.2
29
830
790

added as part of a mixture of surfactants to enhance foaming and lubricating capabilities of the fracking solution (FracFocus, 2014; US House
of Representatives, 2011).
Another group of chromatographic peaks was detected from 22 to
26 min (Fig. 1). These compounds are separated by 44.0262 mass
units and are attributed to the ethoxylate group (\CH2CH2O\). The
accurate mass of each of these compounds ts a linear alkyl ethoxylate
from C-8 to C-10, with ethoxylate chains from EO-4 to EO-8 as the major
components. This class of compounds was previously reported in fracturing owback water by Thurman et al. (2014) but with different
chain lengths of carbon from C-11 to C-14, and at longer retention
times than the ones reported here. The linear alkyl ethoxylates are
known fracturing uid additives typically used to enhance surfactant
properties and movement of sand and oily substances in the fracturing
process (FracFocus, 2014). Although individual standards for the specific compounds are not yet available, the estimated concentrations are in
the mg/L range (based on standards of compounds from the same class
with longer chain lengths; Thurman et al., 2014).
3.3.4. Dissolved organic matter fractionation
The goal herein was to demonstrate an analytical tool for the characterization of owback organic matter and to validate the main organic
constituents detected by DOC analysis, liquid and ion chromatography.
DOM fractionation analysis of the owback water using XAD resins indicated that the DOM is composed primarily of low molecular weight
hydrophilic compounds (HPI N 61%; Fig. 2). This nding agrees well
with the high level of acetic acid noted earlier (Table 1), since this compound is innitely soluble, and is expected to pass completely through
the XAD-8/XAD-4 columns (Aiken et al., 1992). The smallest fraction,
the hydrophobic organic neutrals (HPON ~ 3%), is believed to consist
mostly of the linear alkyl ethoxylates, as noted above.
Further analysis of the DOM fractions using uorescence measurements demonstrated that the hydrophobic organic acid (HPOA) fraction,
typically associated with humic material, has different characteristics
than what is commonly observed in natural waters (e.g., ground water).
Humic material in ground water normally exhibits uorescence signatures that have a distribution of intensity through the Peak A (excitation
240270 nm, emission 380470 nm) and C (Ex 300340 nm, Em 400
480 nm) regions (Dahm et al., 2013; Coble, 1996), as exemplied in
Fig. S2 of the Supporting information for ground water of different sources.
The owback water exhibited the strongest uorescence intensity at
270 nm excitation and 300 nm emission (Peak B region), and this signature is mainly isolated to the HPOA fraction (Fig. 3). The HPOA fraction
also has a less intense uorescence peak with an emission at 350 nm.
Additionally, the specic UV absorbance of the HPOA fraction was
1.1 L/m mgC, which is relatively low for humic substances (Weishaar
et al., 2003). All other fractions (HPI, TPIA and HPON) were weakly uorescent (Fig. S3 in Supporting information).
In natural waters, uorescence in the Peak B region is commonly
associated with proteinaceous material and microbial activity due the
amino acid tyrosine (Coble, 1996). In this case, it is more likely that
the uorescence signature observed stems from phenolic compounds
not associated with proteins (i.e., phenol and 2,4-dimethylphenol;
Table 3). These compounds uoresce in the Peak B region (Aiken,
2014; Monti et al., 1993). If the signal were derived from primarily microbial products, prior work suggests that the Peak B signal should also
be present in other fractions and not strictly conned to the HPOA fraction (Chen et al., 2003; Nguyen et al., 2005). Contrary to what has been
concluded in previous studies, the suite of analytical methods in this
study suggests that the strong uorescence in the Peak B region for
owback water is not due to microbial processes and amino acids, but
phenolic compounds.
Little research has been conducted regarding the uorescence characterization of owback water or the source of the weaker uorescence
peak with a 350 nm emission. Dahm et al. (2013) measured dominant
uorescence peaks at emission wavelengths less than 400 nm for

Y. Lester et al. / Science of the Total Environment 512513 (2015) 637644

641

Alcohol ethoxylates (C8-C10)

Cocamidopropyl dimethylamine

Fig. 1. Total ion chromatogram (by LC/TOF-MS) of the owback sample in positive ion mode. The X and Y axes represent the acquisition time and the relative size of peaks (in counts)
respectively.

produced water from coalbed methane (CBM) formations. These formations are geologically different from the natural gas wells in the DJ Basin
in that aquifers lie adjacent to CBM formations. Investigations into crude
oil in the Gulf of Mexico have also associated uorescence signatures
similar to those in Fig. 3 with the presence of hydrocarbons (Bianchi
et al., 2014). These observations suggest that waters impacted by hydrocarbons (e.g., owback, produced water, and crude oil in ocean) may
have a similar uorescence signature characterized by blue-shifted
intensity (predominantly less than 400 nm).
The above analytical results (particularly the general water quality
parameters and organic constituents) show a clear ngerprint of fracturing uid additives in the recovered owback sample, emphasizing
the anthropogenic inuence on the water and a potential need for suitable treatment before reuse or discharge into the environment. It is
noteworthy that some of the substances identied herein can potentially serve as indicators for the presence of fracturing uid constituents in
owback, such as surfactants and phenolics (easily detected by LCMS
and uorescence, respectively).
3.4. Tailored treatments for owback reuse
The choice of owback treatment(s) mainly depends on the quality
of the raw water and the intended use of the treated efuent. The two
principal pathways available for benecial owback reuse are:
(i) recycling for future fracturing operation and, (ii) reuse outside the
gas industry (e.g., crop irrigation). Each of these options requires a different efuent quality and treatment approach. Herein, optional

treatment trains for the recycle/reuse of the owback are proposed,

based on the water quality data and intended reuse application. Preliminary treatability results are presented to support the applicability of the
proposed treatments.
3.4.1. Flowback recycle for future fracturing operations
Recycling of owback water on-site has the advantage of minimizing water transportation via truck and is very often the least
expensive reuse strategy, since it typically involves a relatively low
level of treatment to achieve an acceptable quality of reusable water.
The key water quality parameters of concern include scaling, plugging
and fouling agents, and the presence of substances that limit the effectiveness of friction reducers such as dissolved solids (Acharya et al.,
2011; Hayes, 2008).
Scale forming constituents (e.g., calcium, barium, sulfate, and carbonates) can potentially form stable carbonate and sulfate precipitates in the wellbore (e.g., CaCO3, BaSO4), reducing gas production
from the well (Hayes, 2008; Kondash et al., 2014; Guerra et al., 2011).
In addition, high levels of iron and manganese can result in the formation of undesirable hydroxide precipitates. Evaluation of the chemical
composition of the DJ owback water sample using the PHREEQC speciation modeling software (PHREEQC, 2011) showed that ferric hydroxide (Fe(OH)3) is the only precipitate of concern (Table 4).
HPOA
450
5

HPI - Low MW hydrophilic compounds

HPOA - Hydrophobic organic acids
TPIA - Transphilic organic acids
HPON - Hydrophobic organic neutrals

60%

40%

20%

Excitation Wavelength, nm

Percent of total DOC

80%
400
4

350

2
300
1

0%
HPI

HPOA

TPIA

HPON

Fig. 2. Distribution of organic matter fractions obtained by the chromatographic separation of the owback DOM using XAD resins.

250
300

350

400
450
500
Emission Wavelength, nm

550

Fig. 3. Fluorescence excitation-emission matrix of the owback water HPOA fraction.

642

Y. Lester et al. / Science of the Total Environment 512513 (2015) 637644

Precipitation of Fe(OH)3 can be avoided by removing the iron from the

recycled owback, using, for example, aeration/precipitation. Benchscale tests, which included 20 min air sparging of a 100 mL owback
sample, followed by 120 min settling time, resulted in more than 95% removal of the iron in the owback water (from 81 mg/L to ~3 mg/L).
Plugging and fouling agents include suspended solids (e.g., clay)
and microorganisms. Suspended solids can be removed by various
separation techniques such as hydrocyclones, ltration, and settling/
sedimentation (also effective for the removal of suspended oil and
grease). Many of these processes were already shown to be effective
in treating oil and gas wastewaters (Sinker, 2007; Veil, 2010; Acharya
et al., 2011). Microorganisms can be inactivated using a variety of disinfection methods such as UV light, chlorine, and ozone (IOGCC and ALL
Consulting, 2006; Kidder et al., 2011).
High levels of salt in recycled owback can reduce the effectiveness
of friction reducers, added to increase the pumping rate of the fracturing
uid and proppant (in slickwater fracturing) (Kamel and Shah, 2009).
The generally accepted upper limit for TDS in fracturing uid is in the
range of 16,00025,000 mg/L (Vengosh et al., 2014), but can reach as
high as 80,000 mg/L in extreme cases (US EPA, 2011). The concentration
of TDS in the DJ water from the well sampled in this study is within this
range (i.e., 22,500 mg/L; Table 1), which indicates that some owback
water in the DJ basin could be used in recycling for fracturing operation
without any additional desalination. Therefore, a viable treatment train
for recycle of the owback water in fracturing could consist principally
of aeration (for iron), followed by sedimentation/ltration (with or
without occulation, for suspended solids and colloids), and disinfection (for microorganisms).
3.4.2. Flowback reuse outside the gas industry
Potential owback water reuse applications outside the gas industry
include for example crop irrigation, livestock watering, and indirect
potable reuse (IOGCC and ALL Consulting, 2006). Some full-scale
examples for reuse application projects are described elsewhere
(ALL Consulting, 2003; Veil et al., 2004). These practices will require
water of better quality than is typical for recycled owback waters,
and therefore a higher level of treatment.
For example, most reuse applications require TDS levels signicantly
lower than the values measured in the DJ owback water. The US EPA's
recommendations for TDS in treated wastewater for long-term irrigation is 5002000 mg/L. Recommendations for livestock watering are
TDS b 10,000 mg/L (ALL Consulting, 2003) and for discharge to surface
water in Pennsylvania TDS b 500 mg/L (FracFocus, 2014). Therefore,
reuse of owback water for most applications will require an additional
desalination step.
Reverse osmosis (RO) separation is the most popular desalination
technique for water at TDS levels b~35,000 mg/L (composed mainly of
sodium and chlorides) (Semiat and Hasson, 2012) and is suitable for

some DJ owback water. The relatively small footprint of the system

and its high mobility make it attractive for mobile on-site treatment of
fracturing wastewater, as demonstrated in many oil/gas exploration
sites throughout the US (Boschee, 2012; Mondal and Wickramasinghe,
2008; Veil, 2010). Moreover, RO treatment is expected to remove
other hazardous constituents, such as heavy metals and toxic organic
compounds (Fu and Wang, 2011).
In addition to salts, DOM may also limit the reuse of owback water.
For example, the limit of BOD5 concentration in treated efuent
discharged to surface water is regulated in Colorado at 45 mg/L
(Colorado Water Quality Control Commission, 2012). Furthermore,
DOM may reduce the efciency of RO desalination, through biofouling
and decreasing of permeate ux (Lester et al., 2014; Ozgun et al.,
2013). Removal of organic matter from wastewater is commonly carried out by biological processes, activated carbon adsorption or advanced oxidation processes (AOPs). In the present case, the owback
DOM is mostly composed of acetic acid, a highly biodegradable compound (as indicated by the high ratio of BOD5/COD ~ 0.9; Table 1).
Therefore, biologically-based treatment will be suitable to reduce the
concentration of DOM. This estimation is supported by preliminary experiments conducted in an activated sludge SBR, resulting in more than
50% DOC removal after only 6 h of treatment (Fig. 4). Activated sludge
and other biological treatments (e.g., membrane bio-reactor-MBR)
were previously demonstrated as an effective solution for treating dissolved organic matter in fracturing wastewaters (Boschee, 2012;
Freire et al., 2001; Doran, 1997).
The potential of various AOPs was also evaluated (i.e., UV/H2O2,
O3/H2O2, Solar light/Chlorine, and photo-Fenton). This technology
is known to degrade organic matter through the highly reactive hydroxyl radical (HO) (Legrini et al., 1993). The examined treatments
only marginally reduced the DOC level of the owback water, as exemplied in Fig. 5 for high doses of light and oxidants (e.g., 10% DOC removal by O3/H2O2, at O3 and H2O2 doses of 500 mg/L and 350 mg/L
respectively).
The efcacy of an AOP treatment is inuenced by the reaction rate of
HO with the target contaminant and the presence of non-target HO
scavengers (in our case mainly chlorides and bromides, as iron was
removed prior to AOP). The reaction rate constants of HO with acetic
acid (1.9 107 M 1 s 1), chloride (7.5 104 M 1 s 1), and bromide
(8.5 108 M1 s1) were calculated in laboratory water, using a lowpressure UV/H2O2 apparatus and a procedure adapted from Lester
et al. (2013) (Text S4 in the Supporting information). These rate constants suggest that most of the generated HO will be scavenged by bromides in the owback (63%), and only ~ 35% will react with DOM
(assuming that it is mostly composed of acetic acid), though not

Phase

SI

log IAP

log K*

Anhydrite
Aragonite
Artinite
Barite
Calcite
Cristobalite
Dolomite
Epsonite
Ferrihydrite(am)
Gypsum
Magnesite(nat)
Mg(SO4)(s)
Witherite

3.70
2.83
10.53
0.52
2.66
0.01
5.71
6.84
6.36
3.54
2.77
17.78
4.93

8.14
11.14
0.72
10.49
11.14
3.15
22.83
8.72
8.90
8.15
11.68
8.68
13.49

4.44
8.31
9.81
9.97
8.48
3.16
17.12
1.88
2.54
4.61
8.91
9.10
z8.56

Ca(SO4)
CaCO3
Mg2(CO3)(OH)2:3H2O
Ba(SO4)
CaCO3
SiO2
CaMg(CO3)2
Mg(SO4):7H2O
Fe(OH)3
CaSO4:2H2O
MgCO3
Mg(SO4)
Ba(CO3)

[DOC]/[DOC]0

0.8
Table 4
Scaling indices (SI), ion-activity products (IAP), and apparent equilibrium constant
(log K*) for the main minerals in the owback sample using the PHREEQC speciation
software.

0.6
0.4
0.2
0
0

Time (h)
Fig. 4. The removal of owback water dissolved organic matter over time in an activated
sludge SBR, using acclimated sludge from Boulder CO 75th Street Wastewater Treatment
Facility. DOC0 represents the original DOC concentration. Experiments were conducted
in duplicate, with average values and standard errors reported.

Y. Lester et al. / Science of the Total Environment 512513 (2015) 637644

[DOC]/[DOC]0

643

Future study will examine the proposed framework, using owback

water samples from different locations in the DJ basin and other areas,
to validate the usefulness of the analytical tools and the efcacy of the
suggested treatment trains.

0.8
0.6

Acknowledgments

0.4
0.2
0
UV/H2O2

O3/H2O2

Solar light/ photo-Fenton

Chlorine

Fig. 5. Removal of dissolved organic carbon by various AOPs. DOC0 refers to the original
DOC concentration. Operational parameters: (i) UV/H2O2 UV dose 3000 mJ/cm2,
[H2O2]0 = 100 mg/L; (ii) O3/H2O2 O3 dose 500 mg/L, H2O2 dose 350 mg/L; (iii) solar
light/chlorine light dose (290400 nm) 5.9 105 mJ/cm2 (24 h), [Cl]0 = 80 mM;
(iv) photo-Fenton (with solar light up to 24 h exposure time) [iron]0 = 80 mg/L,
[H2O2]0 = 500 mg/L. Experiments were conducted in duplicate, with average values and
standard errors reported.

This research was supported by the AirWaterGas Sustainability

Research Network, funded by the National Science Foundation under
Grant No. CBET-1240584. Any opinion, ndings, and conclusions or recommendations expressed in this paper are those of the author(s) and do
not necessarily reect the views of the National Science Foundation. The
authors would like to thank Kenna Butler (USGS) for analytical assistance, Dr. Ori Lahav (Israel Institutes of Technology) for his help with
the PHREEQC speciation modeling software, Dr Tzahi Cath and Daniel
Freedman for their technical support, and Dr. Clara Duca and Dr. James
Rosenblum for their assistance. Any use of trade, rm, or product
names is for descriptive purposes only and does not imply endorsement
by the U.S. Government.
Appendix A. Supplementary data

necessarily to mineralization (Bigda, 1995), which partially explains the

low removal of DOC by AOP.
In addition to the bulk parameters TDS and DOC, specic contaminants such as boron and 1,4-dioxane may limit the reuse of the DJ
owback. The US EPA's recommended level for boron in reclaimed
water for short term irrigation is 2 mg/L, while the Colorado standard
for 1,4-dioxane in groundwater is 0.35 g/L (Colorado Water Quality
Control Commission, 2012). Thus extra care is warranted when considering aquifer recharge as an alternative for owback water reuse. Boron
and 1,4-dioxane are poorly removed by RO (Glueckstern and Priel,
2003; Kegel et al., 2010), thus additional desalination stages or other
specic polishing treatments such as ion exchange (Glueckstern and
Priel, 2003) or UV/H2O2 AOP (Stefan and Bolton, 1998) are required
for their removal.
Based on the above analysis and treatability tests, a combination of
biological treatment, aeration/ltration, and RO separation might present an attractive treatment train for owback reuse in the DJ basin. In
addition, UV/H2O2 advanced oxidation should be considered as a nal
step (after bromides are removed by RO), both for disinfection and for
the elimination of 1,4-dioxane. This proposed treatment train will be
experimentally examined in future research.

4. Conclusions
Treating hydraulic fracturing owback water for benecial reuse is
becoming an attractive alternative to their environmental discharge,
especially in areas with limited access to disposal wells and/or local
water scarcity. Enabling appropriate reuse of owback requires analysis
of the water and selection of suitable treatment approaches, tailored to
the quality of the water and the intended reuse application. This framework was demonstrated on owback water from the DJ basin, where
different analytical methods were shown to provide essential information on the quality of owback water, informing the selection of suitable
treatment trains for water reuse. For example, ion chromatography
(validated by DOM characterization) showed the presence of high concentrations of acetic acid in the owback. These data could be translated
for the selection of biological process as the most attractive treatment
solution for the removal of DOM (which was subsequently conrmed
by treatability tests). Other data such as the concentrations of metals
and trace organic compounds also contributed to the choice of suitable
treatments. The analytical tools presented herein can be used, together
or separately, for the characterization of other fracturing wastewater,
enabling the treatment and reuse of the water, and contributing to
water sustainability in the gas/oil industry.

Complete lists of the analyzed VOCs, SVOCs and metals, MSMS fragmentation of cocamidopropyl dimethylamine by LC/QTOF-MS and a
detailed description of the advanced oxidation experiments.
References
Acharya, H.R., Henderson, C., Matis, H., Kommepalli, H., Moore, B., Wang, H., 2011. Cost
Effective Recovery of Low-TDS Frac Flowback Water for Re-use. U.S. Department of
Energy, Washington, DC (DE-FE0000784).
Aiken, G.R., 2014. Fluorescence and dissolved organic matter: a chemist's perspective. In:
Coble, P.G., Lead, J., Baker, A., Reynolds, D., Spencer, R.G.M. (Eds.), Aquatic Organic
Matter Fluorescence. Cambridge University Press, New York.
Aiken, G.R., Mcknight, D.M., Thorn, K.A., Thurman, E.M., 1992. Isolation of hydrophilic
organic-acids from water using nonionic macroporous resins. Org. Geochem. 18,
567573.
ALL Consulting, 2003. Handbook on coal bed methane produced water: management and
benecial use alternatives, chapter 5 benecial use alternatives. Prepared For,
Ground Water Protection Research Foundation U.S. Department of Energy National
Petroleum Technology Ofce Bureau of Land Management (Accessed from http://
www.all-llc.com/CBM/BU/index.htm).
API American Petroleum Institute, 2010. Water Management Associated with Hydraulic
Fracturing: API Guidance Document HF2. 1st ed. .
Bianchi, T.S., Osburn, C., Shields, M.R., Yvon-Lewis, S., Young, J., Guo, L.D., Zhou, Z.Z., 2014.
Deepwater horizon oil in gulf of mexico waters after 2 years: transformation into the
dissolved organic matter pool. Environ. Sci. Technol. 48, 92889297.
Bigda, R.J., 1995. Consider Fenton's chemistry for wastewater treatment. Chem. Eng. Prog.
91, 6266.
Boschee, P., 2012. Handling Produced Water From Hydraulic Fracturing. Oil and Gas
Facilities.
Boschee, P., 2014. Produced and Flowback Water Recycling and Reuse. Oil and Gas
Facilities.
Chan, P.Y., El-Din, M.G., Bolton, J.R., 2012. A solar-driven UV/chlorine advanced oxidation
process. Water Res. 46, 56725682.
Chen, W., Westerhoff, P., Leenheer, J.A., Booksh, K., 2003. Fluorescence excitation
emission matrix regional integration to quantify spectra for dissolved organic
matter. Environ. Sci. Technol. 37, 57015710.
Coble, P.G., 1996. Characterization of marine and terrestrial dom in seawater using excitation emission matrix spectroscopy. Mar. Chem. 51, 325346.
Colorado Water Quality Control Commission, 2012. Regulation #62 Regulations for
Efuent Limitations (5CCR 1002-62).
Dahm, K.G., van Straaten, C.M., Munakata-Marr, J., Drewes, J.E., 2013. Identifying well
contamination through the use of 3-d uorescence spectroscopy to classify coalbed
methane produced water. Environ. Sci. Technol. 47, 649656.
Doran, G., 1997. Developing a Cost Effective Environmental Solution for Produced Water
and Creating a New Water Resource. Department of Energy (DE-FC22-95MT95008
Report).
Metcalf & Eddy, 2003. Wastewater Engineering: Treatment and Reuse. McGraw-Hill Inc.,
New York.
Fracfocus Chemical Disclosure Registry, 2014. http://fracfocus.org/chemical-use
(Accessed January 2014).
Freire, D.D.C., Cammarota, M.C., Sant'anna, G.L., 2001. Biological treatment of oil eld
wastewater in a sequencing batch reactor. Environ. Technol. 22, 11251135.
Fu, F.L., Wang, Q., 2011. Removal of heavy metal ions from wastewaters: a review.
J. Environ. Manag. 92, 407418.
Glueckstern, P., Priel, M., 2003. Optimization of boron removal in old and new SWRO systems. Desalination 156, 219228.

644

Y. Lester et al. / Science of the Total Environment 512513 (2015) 637644

Gregory, K.B., Vidic, R.D., Dzombak, D.A., 2011. Water management challenges associated
with the production of shale gas by hydraulic fracturing. Elements 7, 181186.
Guerra, K., Dahm, K., Dundorf, S., 2011. Oil and gas produced water management and
benecial use in the western United States. Science and Technology Program Report
No. 157. U.S. Department of the Interior, Bureau of Reclamation, Denver, CO.
Hayes, T., 2008. Proceedings and Minutes of the Hydraulic Fracturing Expert Panel XTO
Facilities, Fort Worth.
Hayes, T., 2009. Sampling and analysis of water streams associated with the development
of Marcellus Shale gas. Final Report for the Marcellus Shale Coalition.
Kamel, A., Shah, S.N., 2009. Effects of salinity and temperature on drag reduction characteristics of polymers in straight circular pipes. J. Pet. Sci. Eng. 67, 2333.
Karvelas, M., Katsoyiannis, A., Samara, C., 2003. Occurrence and fate of heavy metals in the
wastewater treatment process. Chemosphere 53, 12011210.
Kegel, F.S., Rietman, B.M., Verliefde, A.R.D., 2010. Reverse osmosis followed by activated
carbon ltration for efcient removal of organic micropollutants from river bank ltrate. Water Sci. Technol. 61, 26032610.
Kidder, M., Palmgren, T., Ovalle, A., Kapila, M., 2011. Treatment options for reuse of frac
owback and produced water from shale. World Oil, p. 232.
Kondash, A.J., Warner, N.R., Lahav, O., Vengosh, A., 2014. Radium and barium removal
through blending hydraulic fracturing uids with acid mine drainage. Environ. Sci.
Technol. 48, 13341342.
Legrini, O., Oliveros, E., Braun, A.M., 1993. Photochemical processes for water-treatment.
Chem. Rev. 93, 671698.
Lester, Y., Avisar, D., Mamane, H., 2010. Photodegradation of the antibiotic
sulphamethoxazole in water with UV/H2O2 advanced oxidation process. Environ.
Technol. 31, 175183.
Lester, Y., Avisar, D., Mamane, H., 2013. Ozone degradation of cyclophosphamide effect of
alkalinity and key efuent organic matter constituents. Ozone Sci. Eng. 35, 125133.
Lester, Y., Tesfayohanes, Y., Morrissey, I., Linden, K.G., 2014. Can we treat hydraulic
fracturing owback with a conventional biological process? The case of guar gum.
Environ. Sci. Technol. Lett. 1, 133136.
Lutz, B.D., Lewis, A.N., Doyle, M.W., 2013. Generation, transport, and disposal of wastewater
associated with marcellus shale gas development. Water Resour. Res. 49, 647656.
Mauter, M.S., Alvarez, P.J.J., Burton, A., Cafaro, D.C., Chen, W., Gregory, K.B., Jiang, G., Li, Q.,
Pittock, J., Reible, D., Schnoor, J.L., 2014. Regional variation in water-related impacts of
shale gas development and implications for emerging international plays. Environ.
Sci. Technol. 48, 82988306.
Mondal, S., Wickramasinghe, S.R., 2008. Produced water treatment by nanoltration and
reverse osmosis membranes. J. Membr. Sci. 322, 162170.
Monti, S., Kohler, G., Grabner, G., 1993. Photophysics and photochemistry of methylated
phenols in beta-cyclodextrin inclusion complexes. J. Phys. Chem. 97, 1301113016.
Mller, J.P., Jekel, M., 2001. Comparison of advanced oxidation processes in ow-through
pilot plants (part I). Water Sci. Technol. 44, 303309.
Nguyen, M.L., Westerhoff, P., Baker, L., Hu, Q., Esparza-Soto, M., Sommerfeld, M., 2005.
Characteristics and reactivity of algae-produced dissolved organic carbon.
J. Environ. Eng. Asce 131, 15741582.
Olsson, O., Weichgrebe, D., Rosenwinkel, K.H., 2013. Hydraulic fracturing wastewater in
Germany: composition, treatment, concerns. Environ. Earth Sci. 70, 38953906.
Olujimi, O.O., Fatoki, O.S., Odendaal, J., Daso, A.P., Oputu, O., 2012. Preliminary investigation
into occurrence and removal of arsenic, cadmium, mercury, and zinc in wastewater
treatment plants in Cape Town and Stellenbosch. Pol. J. Environ. Stud. 21, 17551765.

Orem, W., Tatu, C., Varonka, M., Lerch, H., Bates, A., Engle, M., Crosby, L., Mclntosh, J., 2014.
Organic substances in produced and formation water from unconventional natural
gas extraction in coal and shale. Int. J. Coal Geol. 126, 2031.
Ozgun, H., Ersahin, M.E., Erdem, S., Atay, B., Kose, B., Kaya, R., Altinbas, M., Sayili, S., Hoshan,
P., Atay, D., Eren, E., Kinaci, C., Koyuncu, I., 2013. Effects of the pre-treatment alternatives on the treatment of oil-gas eld produced water by nanoltration and reverse
osmosis membranes. J. Chem. Technol. Biotechnol. 88, 15761583.
PHREEQC PH Redox Equilibrium, 2011. version 2.18.00 http://pfw.antipodes.nl/
download.html.
Semiat, R., Hasson, D., 2012. Water desalination. Rev. Chem. Eng. 28, 4360.
Shemer, H., Kunukcu, Y.K., Linden, K.G., 2006. Degradation of the pharmaceutical Metronidazole via UV, Fenton and photo-Fenton processes. Chemosphere 63, 269276.
Sinker, A., 2007. Produced water treatment using hydrocyclones: theory and practical
applications. 14th Annual International Petroleum Environmental Conference
(Houston, Texas) (http://ipec.utulsa.edu/Conf2007/Papers/Sinker_114.pdf).
Stefan, M.I., Bolton, J.R., 1998. Mechanism of the degradation of 1,4-dioxane in dilute
aqueous solution using the UV hydrogen peroxide process. Environ. Sci. Technol.
32, 15881595.
STRONGER State Review of Oil&Natural GAs Environmental Regulations, 2011. Colorado Hydraulic Fracturing State Review.
The Interstate Oil and Gas Compact Commission (IOGCC), ALL Consulting, 2006. A Guide
to Practical Management of Produced Water From Onshore Oil and Gas Operations in
the United States (DE-PS26-04NT15460-02).
Thurman, E.M., Ferrer, I., Blotevogel, J., Borch, T., 2014. Analysis of hydraulic fracturing
owback and produced waters using accurate mass: identication of ethoxylated
surfactants. Anal. Chem. 86, 96539661.
US House of Representatives, Committee on Energy and Commerce, 2011. Chemicals
Used in Hydraulic Fracturing.
US EPA, 2009. Emerging contaminant 1,4-dioxane. Fact Sheet.
US EPA, 2011. Proceedings of the Technical Workshops for the Hydraulic Fracturing
Study: Water Resources Management (EPA 600/R-11/048).
Veil, J.A., 2010. Water Management Technologies Used by Marcellus Shale Gas Producers.
United States Department of Energy, National Energy Technology Laboratory.
Veil, J.A., Puder, M.G., Elcock, D., Redweik, R.J., 2004. A White Paper Describing Produced
Water From Production of Crude Oil, Natural Gas, and Coal Bed Methane. Argonne
National Laboratory, Chicago, IL (W-31-109-Eng-38).
Vengosh, A., Jackson, R.B., Warner, N., Darrah, T.H., Kondash, A., 2014. A critical review of
the risks to water resources from unconventional shale gas development and hydraulic fracturing in the United States. Environ. Sci. Technol. 48, 83348348.
Weishaar, J.L., Aiken, G.R., Bergamaschi, B.A., Fram, M.S., Fujii, R., Mopper, K., 2003. Evaluation of specic ultraviolet absorbance as an indicator of the chemical composition
and reactivity of dissolved organic carbon. Environ. Sci. Technol. 37, 47024708.
Wolford, R., 2011. Characterization of organics in the Marcellus Shale owback and produced water. Master of Science. The Pennsylvania State University.
Zimmermann, S.G., Wittenwiler, M., Hollender, J., Krauss, M., Ort, C., Siegrist, H., von
Gunten, U., 2011. Kinetic assessment and modeling of an ozonation step for
full-scale municipal wastewater treatment: micropollutant oxidation, by-product
formation and disinfection. Water Res. 45, 605617.