w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 1 4 5 e1 1 5 4
Available online at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/watres
A comparative life cycle assessment of hybrid osmotic
dilution desalination and established seawater desalination
and wastewater reclamation processes
Nathan T. Hancock, Nathan D. Black, Tzahi Y. Cath*
Colorado School of Mines, Division of Environmental Science and Engineering, 1500 Illinois St., Golden, CO 80401, USA
article info
abstract
Article history:
The purpose of this study was to determine the comparative environmental impacts of
Received 19 August 2011
coupled seawater desalination and water reclamation using a novel hybrid system that
Received in revised form
consist of an osmotically driven membrane process and established membrane desalina-
4 December 2011
tion technologies. A comparative life cycle assessment methodology was used to differ-
Accepted 5 December 2011
entiate between a novel hybrid process consisting of forward osmosis (FO) operated in
Available online 13 December 2011
osmotic dilution (ODN) mode and seawater reverse osmosis (SWRO), and two other
processes: a stand alone conventional SWRO desalination system, and a combined SWRO
Keywords:
and dual barrier impaired water purification system consisting of nanofiltration followed
Life cycle assessment
by reverse osmosis. Each process was evaluated using ten baseline impact categories. It
Forward osmosis
was demonstrated that from a life cycle perspective two hurdles exist to further devel-
Osmotic dilution
opment of the ODN-SWRO process: module design of FO membranes and cleaning inten-
Reverse osmosis
sity of the FO membranes. System optimization analysis revealed that doubling FO
Potable water reuse
membrane packing density, tripling FO membrane permeability, and optimizing system
Wastewater treatment
operation, all of which are technically feasible at the time of this publication, could reduce
Seawater desalination
the environmental impact of the hybrid ODN-SWRO process compared to SWRO by more
Energy recovery
than 25%; yet, novel hybrid nanofiltration-RO treatment of seawater and wastewater can
achieve almost similar levels of environmental impact.
ª 2011 Elsevier Ltd. All rights reserved.
1.
Introduction
Seawater and wastewater effluents are two sources of
impaired water that are ubiquitous in coastal areas; however,
both sources require substantial treatment to comply with
EPA standards for potable and non-potable uses (Asano et al.,
2007). The development of modern membrane separation
processes such as reverse osmosis (RO) and nanofiltration
enabled widespread utilization of these sources.
Reclaimed water usually contain much lower concentration of total dissolved solids than seawater, and rarely
requires desalination to achieve reuse standards (WHO,
2003). Yet, reclaimed water may still require treatment
with RO or nanofiltration to remove dissolved contaminants
and enable potable reuse (Asano et al., 2007; Letterman,
1999). While desalination technologies are well developed,
they are not without drawbacks. Energy demand associated
with removing salts from seawater and dissolved contaminants from reclaimed water is far greater than treatment of
fresh water by conventional water treatment processes
(Asano et al., 2007). Desalination technologies are also
subject to negative environmental impacts and poor public
* Corresponding author. Tel.: þ1 303 273 3402; fax: þ1 303 273 3413.
E-mail address: tcath@mines.edu (T.Y. Cath).
0043-1354/$ e see front matter ª 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.watres.2011.12.004
1146
w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 1 4 5 e1 1 5 4
perception related to marine life impingement and entrainment and the discharge of chemicals and concentrated brine
into receiving environments (Lattemann and Höpner, 2008).
1.1.
Water desalination and direct potable reuse e
energy implication
The energy required for desalination by RO and nanofiltration
(NF) is a function of the osmotic pressure (p) differential across
the membrane. The osmotic pressure of a solution must be
overcome with hydraulic pressure to produce water through
a semipermeable RO or NF membrane, and water flux (Jw) may
be approximated by Jw ¼ Aw(DpDp), where Aw is the
membrane water permeability coefficient and Dp is the
hydraulic pressure differential across the membrane.
Typical seawater RO (SWRO) plants operate at approximately 50% recovery; thus, the osmotic pressure generated by
concentrating seawater into brine requires operation at
pressures close to 70 MPa (w1000 psig) (Wilf, 2007). At these
pressures, SWRO requires a substantial input of energy to
power high-pressure pumps. Energy recovery devices are
commonly employed in conjunction with SWRO to transfer
energy (i.e., pressure) from the brine stream into the incoming
feed stream; more than 60% energy saving can be achieved
with the assistance of modern energy recovery devices
(Stover, 2007). Low salinity wastewater effluent enables pressure driven membrane processes to achieve more than 80%
water recovery with substantially lower pressure and less
energy demand than desalination of seawater (Wilf, 2007).
Overall, osmotic pressure plays an important role in treatment of impaired water.
1.2.
Osmotic dilution
In two previous studies (Cath et al., 2010 and Hancock et al.,
2011b) we investigated the technical merits, advantages, and
limitations of the hybrid osmotic dilution (ODN) process,
which is a forward osmosis (FO) process operated without
draw solution reconcentration (Cath et al., 2006) e ODN was
integrated with SWRO to perform simultaneous seawater
desalination and wastewater reclamation. The hybrid ODNSWRO process is shown schematically in Fig. 1. Water from
Fig. 1 e Hybrid OND-SWRO process for simultaneous
seawater desalination and wastewater reclamation. A PRO
variant of the ODN-SWRO process to achieve energy
savings from the mixing of saline seawater and low
salinity wastewater effluent sources can be utilized by
replacing the ODN processes with PRO.
impaired stream diffuses spontaneously through an FO
membrane in ODN1 to dilute seawater before SWRO desalination. The diluted seawater has a lower osmotic pressure
and enables production of more water at the same applied
hydraulic pressure; thus, reducing the energy required to
produce a unit volume of product water.
Previous studies demonstrated that osmotically driven
membrane processes, and specifically ODN and FO, have very
high rejection of contaminants common to impaired water
and that they experience low irreversible membrane fouling
even when treating highly impaired water (Cath et al., 2010).
Consequently, FO membranes require less pretreatment of
impaired water and less periodic cleaning than pressure
driven membrane processes.
Additional energy savings may be achieved by implementing the pressure retarded osmosis process (PRO) (Achilli
and Childress, 2010; Yip and Elimelech, 2011) in place of the
ODN process (Fig. 1). PRO provides a method of converting
chemical energy into mechanical energy by depressurizing
diluted draw solution through a turbogenerator. Current
generation FO membranes used in PRO mode demonstrate
power density (i.e., a measure of electric power saved or
produced by a unit surface area of membrane) output of up to
5 W per m2 of membrane at the bench-scale and less than 1 W/
m2 at the pilot scale (Patel, 2010). Results in this study will be
compared to the current state of PRO performance.
1.3.
Investigating environmental impacts of
desalination technologies with life cycle assessment
methodology
Life cycle assessment (LCA) is a methodology to holistically
evaluate the environmental impacts associated with the
development of products for human consumption or use (ISO,
2006a). In recent years studies were published that explored
the use of LCA methodology to evaluate the environmental
impact of water treatment technologies (Biswas, 2009; Muñoz
and Fernández-Alba, 2008; Raluy et al., 2005a, 2005b;
Vlasopoulos et al., 2006). The majority of these studies
focused on the construction and operations phases of SWRO
plants. Each study developed a life cycle inventory (LCI) to
account for various reference flows such as construction
materials, chemicals required for operation (primarily antiscaling and disinfection chemicals), and materials used for
membrane fabrication. A primary conclusion from these
studies is that energy consumption during the operation
phase of SWRO plants is the single greatest contributor to its
negative environmental impacts, accounting for greater than
85% of the environmental impact (Raluy et al., 2005a).
Thus, the main objective of this study was to investigate
the comparative environmental impacts of established and
novel processes for seawater desalination and water reclamation. A specific focus of this study was to explore if the
hybrid ODN-SWRO system provides intrinsic benefits over
established SWRO desalination or other hybrid technologies
for seawater desalination and wastewater reclamation when
environmental impacts are considered. Through system
optimization analysis, a secondary objective was to identify
required improvements to the hybrid ODN-SWRO process that
will make it more favorable from a life cycle perspective.
w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 1 4 5 e1 1 5 4
2.
1147
Material and methods
LCA consists of four phases, including goal and scope definition, LCI analysis, life cycle impact assessment (LCIA), and
interpretation (ISO, 2006b). The first two phases are described
in this section while the LCIA and interpretation are discussed
in the Results and Discussion section.
2.1.
Goal and scope
The goal of this study was to evaluate and compare the
environmental impacts of the hybrid ODN-SWRO process to
those of competitive processes during the desalination of
seawater and reclamation of wastewater. A comparative LCA
of three functionally equivalent seawater desalination and
wastewater reclamation processes was chosen to achieve this
goal (Fig. 2). The first scenario (A) is a baseline case of seawater
desalination using SWRO and traditional wastewater treatment with discharge to the environment. The second scenario
(B) employs a dual pass NF-RO membrane process for wastewater reclamation to augment the water supplied by a SWRO
desalination plant. The third scenario (C) uses the hybrid
ODN-SWRO process to achieve simultaneous seawater desalination and wastewater reclamation.
It is assumed that treating seawater with the SWRO
membrane produces sufficient product water quality for use
(permeate total dissolved solids concentration of 200 60 mg/L),
while purifying the wastewater requires two tight membrane
barriers (e.g., nanofiltration-RO or ODN-RO) to achieve a functionally equivalent product water quality. The environmental
impacts associated with each scenario are normalized to the
production of 1 m3 of water produced, which is defined as the
functional unit.
The scope of this study includes three gate-to-gate LCIs
based on primary chemical inputs for clean-in-place systems,
materials consumed during membrane replacement, and
energy required for operation and heating of fluids during
clean-in-place. This study is not limited to one geographic
location, and is therefore unable to include a comprehensive
inventory based on a bill of materials for the construction
phase. Although the capital construction costs may vary
substantially between the different scenarios, the environmental impacts resulting from the construction and decommissioning phase of a membrane based treatment process are
typically negligible over a 30 year operation lifetime (as used in
this study) when compared to the operation phase (Raluy
et al., 2005a, 2005b). Without a specific site to consider, the
logistics and impacts of transporting chemicals and materials
to the site were not considered for this LCA. LCI boundaries for
each scenario are shown in Fig. 2. Each LCI encompasses the
core technologies and an auxiliary pretreatment process that
includes a 100 mm disk strainer for all feed streams and an
additional treatment stage with ultrafiltration membranes for
suspended solids removal prior to RO and NF. Waste streams
generated during the operation phase (i.e., landfill disposal of
membranes at the end of their assumed 5 year service life and
discharge of clean-in-place chemicals to the sewer) are
included in each LCI. Initial treatment of wastewater feed,
post-treatment of the product water, and the distribution
Fig. 2 e Gate-to-gate LCI system boundaries for (a) baseline
scenario (A) SWRO desalination and wastewater discharge
to the environment, (b) alternative scenario (B) NF/RO
treatment of wastewater and blending with SWRO
permeate, and (c) alternative scenario (C) simultaneous
OND/SWRO treatment of seawater and wastewater.
Acronym: SW-seawater, WW-wastewater, SDW-solid
waste landfill, Pr-T-pretreatment, Po-T-post treatment.
Note that only the portion of the wastewater treatment
plant that treats wastewater from the three scenarios was
considered in the LCA.
1148
w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 1 4 5 e1 1 5 4
system of each scenario is assumed to be equivalent between
all scenarios and is therefore irrelevant for a process-based
comparative LCA (Hancock, 2011).
Energy and material requirements for each scenario were
estimated based on systems designed to produce 3785 m3/day
(1 million gallons per day (MGD)) of drinking water. The relevant flow rates for each process are summarized in Table 1.
Water recovery of the disk filter and ultrafiltration membrane
process were assumed to be 99% and 90%, respectively. The
SWRO system was designed to achieve 50% water recovery in
all cases. A high water recovery design was assumed for the
dual pass NF-RO system for wastewater treatment in scenario
B (80% for the first pass (NF) and 85% for the second pass (RO))
based on suggested values from the literature (Wilf, 2007). The
ODN process in scenario C was initially assumed to achieve
50% recovery of the wastewater stream.
The energy consumption and system design of unit
processes for each scenario was either calculated from known
principles of hydraulic flow or derived from manufacturer data
(see Table S1 in Supplementary Content). Design of the ODN
process was facilitated with a Microsoft Excel based software
program developed by the authors and described in our
previous publication (Cath et al., 2010). Energy for wastewater
pumping to the system in each scenario was not considered
because it was assumed to already be flowing under pressure
to a pre-existing discharge point. Energy demand of seawater
intake pumping was calculated assuming that seawater with
a density of 1024 kg/m3 is pumped 1.6 km (1 mile) and requires
30 m (100 ft) of lift using a pump with a standard motor and
pump efficiencies of 94% and 86%, respectively (Wilf, 2007).
Environmental impacts associated with the energy
consumed in each treatment scenario were estimated from
the Ecoinvent 2000 Union for the Co-ordination of Transmission Energy (UCTE) energy mix (Swiss Center for Life Cycle
Inventories, 2011). The UCTE energy mix is composed of
approximately 40% fossil fuels (i.e., hard coal and natural gas),
30% nuclear, 10% hydroelectric, and the remaining 20% is
Table 1 e Flow rates into specific unit processes and RO
product waters in each treatment scenario.
Flow m3/min (MGD)
Unit process
Scenarios
A
Seawater
Disk filter
UF
SWRO
SWRO Permeate
Wastewater
Disk filter
UF
WW NF
WW RO
WW RO Permeate
ODN feed
ODN Permeate
Total product
a After dilution.
5.8
5.8
5.3
2.6
B
(2.2)
(2.2)
(2.0)
(1.0)
2.9
2.9
2.6
1.3
e
e
e
e
e
e
e
2.6 (1.0)
2.1
2.1
1.8
1.6
1.3
(1.1)
(1.1)
(1.0)
(0.5)
(0.8)
(0.8)
(0.7)
(0.6)
(0.5)
e
e
2.6 (1.0)
C1
2.9 (1.1)
2.9 (1.1)
5.3 (2.0)a
2.6 (1.0)
5.3 (2.0)
e
e
e
e
5.3 (2.0)
2.6 (1.0)
2.6 (1.0)
derived from several other sources including biomass, wind,
and photovoltaic.
Both seawater and wastewater are assumed to have
a temperature of 20 C. Systems were designed with an
assumed total dissolved solids concentrations of 35,000 mg/L in
the seawater and 850 mg/L in the wastewater (Tchobanoglous
et al., 2003; Wilf, 2007). Concentrations of chemical species
are provided in Table S2 in the Supplementary Content.
To further reduce mineral scaling, a scale inhibitor is
frequently added to SWRO feed and is essential for operation
of high water recovery membrane applications (Asano et al.,
2007; Sauvet-Goichon, 2007). The ultrafiltration process
requires pre-coagulation to achieve enhanced removal of
membrane foulants (Voutchkov, 2010). Each unit process also
requires periodic clean-in-place at an elevated fluid temperature with one or more chemicals such as HCl and H2SO4
(acid), NaOH (base), NaOCl (oxidizing agent), sodium dodecylbenzenesulfonate (SDBS) (surfactant), and tetrasodium
ethylenediaminetetraacetic acid (Na4EDTA) (chelating agent)
added to maintain long-term process performance.
Materials used for the fabrication of membrane modules
are another important component for the comparative LCA of
the different water treatment scenarios. The types of materials used in ultrafiltration, nanofiltration, and RO membrane
modules are generally well known (Baker, 2004; Beerya et al.,
2011; Mulder, 1997); yet, membranes and modules for FO
and ODN are not standardized. More detailed description is
provided in the Supplementary Content. The inventory analysis was performed based on the scope of the study presented
above. The LCI for three scenarios is summarized in Table S3
in the Supplementary Content.
2.2.
Life cycle impact assessment methodology
SimaPro LCA software (Pré Consultants, 2011) was employed to
aggregate the chemical and material requirements for each
scenario and translate that data into values associated with ten
environmental impact categories. Characterizing chemical
and material use into impact categories was performed with
the Leiden University Institute of Environmental Sciences
(CML) 2001 method (CML, 2011). The ten baseline impact categories provided by the CML method include abiotic depletion
potential (ADP), acidification potential (AP), eutrophication
potential (EP), fresh water aquatic ecotoxicity potential
(FAETP), global warming potential (GWP) (100 years), human
toxicity potential (HTP), marine aquatic toxicity potential
(MAETP), ozone layer depletion potential (ODP), photochemical
oxidation potential (POCP), and terrestrial ecotoxicity potential
(TETP). The categories and their descriptions are summarized
in Table S4 in the Supplementary Content.
SimaPro was also used to perform interpretation of the
LCIA data. Based on findings from the interpretation phase,
the software was also used to perform a perturbation analysis
of system design parameters that were found to have the most
significant impact on the comparative performance of the
hybrid ODN-SWRO process relative to the other two scenarios.
The hybrid ODN-SWRO process was exclusively used for the
perturbation analysis because the design and material
requirements of the unit processes in the other two scenarios
followed industry standards or were designed and optimized
w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 1 4 5 e1 1 5 4
with system design software from major vendors (i.e., DOW
Filmtec ROSA (Dow Filmtec, 2010) and ERI Power Model
(ERI, 2011)).
3.
Results and Discussion
3.1.
Baseline comparative LCIA of water treatment
scenarios
The LCIA of the inventories (Table S3) was performed with the
CML 2001 method. The ten impact categories (Table S4) were
used to investigate the relative environmental impact of the
different treatment scenarios.
The normalized impacts of the three scenarios across the ten
categories are shown in Fig. 3. Each scenario is normalized by the
maximum value observed between the three scenarios, such
that the dominant scenario records 100% of the relative impact
compared to the other processes. Results in Fig. 3 indicate that
the SWRO system (scenario A) and the ODN-SWRO system
(scenario C1 with membrane Aw0 ¼ 0.36 L/m2 h bar) produce
a similar degree of environment impact, while the dual barrier
NF-RO and SWRO system (scenario B) achieves the design
objective with substantially less environmental impact. Scenario
C1 has the greatest impact potential in the acidification, eutrophication, ozone layer depletion, human toxicity, terrestrial
ecotoxicity, and photochemical oxidation categories. For these
categories, the dominant product flows associated with the
greatest contribution to the impact were the Na4EDTA and scale
Fig. 3 e Relative impacts of scenarios A, B, and C1
evaluated using ten impact categories. Scenario C1 is for
the ODN membrane with the baseline membrane Aw0
value (0.36 L/m2$h$bar). The maximum values observed for
each impact category are: ADP: 9.9$10L3 kg Sb, AP:
6.9$10L3 kg SO2, EP: 5.5$10L3 kg PO4, GWP: 1.4 kg CO2, ODP:
8.1$10L8 kg CFC-11, HTP: 0.72 kg 1,4-DB, FAETP: 0.74 kg
1,4-DB, MAETP: 1600 kg 1,4-DB, TETP: 5.3$10L3 kg 1,4-DB,
POCP: 2.9$10L4 kg C2H4.
1149
inhibitor required for the monthly clean-in-place of the ODN
membranes. These impacts are the result of the manufacturing
processes required to produce these chemicals. The treatment of
the clean-in-place waste at the wastewater treatment plant also
contributed substantially to the eutrophication, human toxicity,
and terrestrial ecotoxicity categories of scenario C1. Thus, the
baseline ODN-SWRO scenario does not appear favorable for
reducing the environmental impacts associated with seawater
desalination and water reclamation; however, relatively simple
improvement to the process that are readily achievable at the
time of this publication can decrease chemical consumption of
scenario C and reduce the environmental impacts of the process.
Normalized scoring of the impact categories in Fig. 3 are
provided in Figure S1 in the Supplemental Content document.
Normalization scoring divides the results in each impact category by a reference value to provide a measure of their relative
importance. The figure shows the relative importance of fresh
and marine aquatic ecotoxicity potentials for the three
scenarios, but especially for scenario A and C1. Note that specific
impact categories such as ozone layer depletion, human
toxicity, terrestrial ecotoxicity, fresh and marine ecotoxicity,
and photochemical oxidation potentials are difficult to quantify
because their local impacts make them difficult to aggregate
with the traditional global impact categories used (Guinée, 2002).
3.2.
FO membrane optimization
Unique aspects of the hybrid ODN-SWRO process were identified while collecting data for the LCI of scenario C. One aspect
is that membrane modules for ODN are not yet optimized with
respect to their materials and membrane packaging compared
to RO membranes. The water permeability coefficient of
the FO membrane used to construct the ODN design
model (Cath et al., 2010), and used in the baseline LCIA
(Hydration Technology Innovations (HTI) membrane cast in
2005 (Hancock and Cath, 2009)), was relatively low
(Aw0 ¼ 0.36 L/m2$h$bar) and the packing density of an 8-inch
membrane module was 9.5 m2. These parameters, in addition to the inherently low flux through FO membranes when
seawater is the draw solution, resulted in a packing density
ratio of FO membrane modules to SWRO membrane module of
14:1 to achieve the same water recovery described in scenario
C1. Newer HTI membranes that were cast in 2008 and 2010
have permeability coefficients 1.4 and 2 times higher
(Aw1 ¼ 0.50 and Aw2 ¼ 0.71 L/m2$h$bar, respectively) (Hancock
et al., 2011a), and a corresponding membrane packaging ratio
of 4.8:1 and 3.4:1, respectively. In the near future, HTI
membranes with three times higher permeability will be
available (Aw3 ¼ 1.08 L/m2$h$bar) which will further reduce the
ratio to approximately 2:1. New innovative thin-film
composite FO membranes might further improve this ratio
(Yip et al., 2010; Yip et al., 2011). Significantly higher packing
density of FO modules, approaching that of 8-inch RO
membranes, is unlikely because of the need to use feed spacers
that provide wider flow channels and allow flushing out of
solids from the feed channels. Yet, 8-inch diameter FO
membrane modules containing 20 m2 of membrane area are
already commercialized (Hutchings et al., 2010).
A second observation is that the power density of the ODN
may be calculated by multiplying the specific energy savings
1150
w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 1 4 5 e1 1 5 4
(kWh/m3) of the SWRO process due to seawater feed dilution
by the nominal water flux through the FO membrane used in
the ODN process. The energy demands and power densities
for the SWRO process in scenarios A and C are summarized in
Table 2. Results from the energy balance in scenario C1 show
that the power density of the ODN-SWRO process (with
Aw0 ¼ 0.36 L/m2$h$bar) is 3.6 W/m2 when an energy recovery
device is used in the SWRO process. However, the values
reported in Table 2 are not the maximum achievable energy
density by the hybrid ODN-SWRO process under the design
specifications listed in Table 1. A plot of the potential power
density as a function of the fraction of water recovered from
the wastewater achievable by the ODN is shown in Fig. 4 for FO
membranes with different permeability coefficients. With an
energy recovery device operating in conjunction with the
SWRO the maximum potential power density of the ODN
process is 3.7, 5.2, 7.4, and 11.1 W/m2 for Aw values of 0.36,
0.50, 0.71, and 1.08 L/m2$h$bar, respectively. Without the use
of energy recovery devices (unlikely in most SWRO desalination cases) the maximum power density achieved by the ODN
could be 7.5, 10.5, 15.0, and 22.4 W/m2, respectively. For the
design conditions listed in Table 1 the maximum power
density occurs at approximately 61% water recovery from the
7570 m3/day (2 MGD) wastewater stream (Fig. 4).
The potential power density of the ODN-SWRO process is
significantly greater than any values measured for PRO
membranes to date. Using a membrane with an Aw value of
0.67 L/m2 h bar, Achilli and Childress (2010) measured a power
density of 5.1 W/m2 with a 60 g/L NaCl draw solution. A similar
membrane used in the ODN-SWRO process would achieve
power density in excess of 7.4 W/m2 with an energy recovery
device or 15.0 W/m2 without an energy recovery device when
using 35 g/L sea salt draw solution.
3.3.
New comparative LCIAs of water treatment
scenarios
Before attempting to optimize the ODN-SWRO process to
reduce its environmental impact, it is important to revisit the
operating conditions to achieve peak power density. Operating
the hybrid system at peak power density represents a conservative approach to chemical demand. This is because more
Table 2 e Energy demands and power densities for the
SWRO process in scenarios A and C1. The nominal water
flux determined by our design program (Cath et al., 2010)
at 50% water recovery from a 7570 m3/day (2 MGD)
wastewater feed is 3.8 L/m2$h.
Description
SWRO energy consumption
in scenario A, kWh/m3
SWRO energy consumption after
ERD in scenario A, kWh/m3
SWRO Energy consumption after
dilution in scenario C1, kWh/m3
Energy recovered with ERD in
scenario C1, kWh/m3
Power density including ERD, W/m2
Power density excluding ERD, W/m2
Energy value
4.0
2.4
2.1
0.6
3.6
7.2
Fig. 4 e Power density of the FO membrane in the hybrid
ODN-SWRO process as a function of the fraction of water
recovered from wastewater. A baseline water permeance
(Aw0) of 0.36 L/m2$h$bar was assumed for model
development, and sensitivity analysis was performed with
1.43, 23, and 33 the membrane permeance. The model
assumed utilization of ERD in the SWRO process (Table 2).
membrane area is required, and therefore creates more system
volume to clean, to achieve 11% more water recovery from the
wastewater stream. To decrease chemical demand of the ODNSWRO system, the membrane packing density and membrane
permeability coefficient were varied. Doubling the membrane
packing density reduced by approximately 50% the number of
pressure vessels that would otherwise need to be filled with
heated cleaning solution during clean-in-place. Increasing Aw
of the FO membrane will reduce the total surface area of
membrane required to achieve the desired water recovery.
The relative impact of various improvements to the ODN
membrane (i.e., permeance coefficient) and module (i.e.,
packing density) in scenario C are summarized with respect to
the ten impact categories in Fig. 5. The greatest environmental
impact is observed when the ODN water recovery is 61% and
membrane Aw0 is utilized (scenario C2). Increasing the water
recovery required an additional 14,612 m2 of FO membrane
area (35% increase) due to the reduction in the average water
flux when the seawater draw solution is diluted. This increase
results in significantly more membrane area to clean. The
amount of membrane area required to perform the dilution is
reduced by a maximum of 33% when Aw increased by a factor
of three (Aw3) (scenario C6). The substantial reduction in
membrane area achieves significant reductions across all
environmental impact categories. Terrestrial ecotoxicity
potential is most sensitive to the reduction of chemicals
required for system operation and maintenance. The two
primary factors that contribute to terrestrial ecotoxicity
potential are energy production and treatment of clean-inplace chemicals by a wastewater treatment plant.
w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 1 4 5 e1 1 5 4
Fig. 5 e Relative impact of various improvements to the
ODN membrane and module in scenario C evaluated using
the ten impact categories. %R represent percent ODN water
recovery, PD1 represent packing density of 9.5 m2/module,
PD2 represent 20 m2/module, and Aw represent the
different FO membrane permeance coefficients. The
maximum values observed for each impact category are:
ADP: 0.01 kg Sb, AP: 7.3$10L3 kg SO2, EP: 5.9$10L3 kg PO4,
GWP: 1.4 kg CO2, ODP: 9$10L8 kg CFC-11, HTP: 0.78 kg
1,4-DB, FAETP: 0.76 kg 1,4-DB, MAETP: 1400 kg 1,4-DB,
TETP: 6.1$10L3 kg 1,4-DB, POCP: 3.1$10L4 kg C2H4.
Decreasing the volume of the FO membrane modules
(scenarios C3 through C6) decreases the volume of cleaning
solution that requires heating and reduces the mass of
chemicals required to prepare the clean-in-place solution.
Further reduction in the clean-in-place chemicals used in
scenario C may be achieved by reducing the frequency of
clean-in-place. It is unrealistic to expect that high flux FO
membranes would be able to continually treat impaired water
(e.g., secondary effluent) without frequent cleaning. Adding
ultrafiltration pretreatment to the wastewater stream (similar
to scenario B) is a simple method that may reduce the expected frequency of clean-in-place to twice per year.
Normalized scoring of the impact categories in Fig. 5 are
provided in Figure S2 in the Supplemental Content document.
Figure S2 also shows the relative importance of fresh and
marine aquatic ecotoxicity potentials for all the scenarios.
The effect of adding an ultrafiltration pretreatment to the
wastewater stream prior to the ODN process is shown in Fig. 6
and normalized scoring of the impact categories are provided
in Figure S3 in the Supplemental Content document. Adding
ultrafiltration pretreatment before the ODN process with FO
membrane Aw2 (scenarios C5) further reduces the environmental impacts by achieving a reduction in the total amount of
chemicals required during clean-in-place and reducing the
volume of clean-in-place wastewater requiring treatment
1151
Fig. 6 e Relative impact of UF pretreatment of the ODN feed
stream prior to the FO membranes across the ten impact
categories of scenarios C5 and C6. In all cases the ODN
water recovery is 61% and the membrane packing density
is 20 m2/module. The maximum values for each impact
category are: ADP: 7.5$10L3 kg Sb, AP: 5.2$10L3 kg SO2, EP:
3.8$10L3 kg PO4, GWP: 1.0 kg CO2, ODP: 5.3$10L8 kg CFC-11,
HTP: 0.52 kg 1,4-DB, FAETP: 0.56 kg 1,4-DB, MAETP:
1100 kg 1,4-DB, TETP: 3.2$10L3 kg 1,4-DB, POCP:
2.1$10L4 kg C2H4.
(both by a factor of 6 by performing biannual rather than
monthly clean-in-place events). Conversely, adding ultrafiltration pretreatment to the ODN process when membrane Aw3
is used (scenarios C6) slightly increases the environmental
impact of this scenario relative to a scenario with more
frequent (monthly) clean-in-place cleaning intervals. The
increase in nine out of the ten categories may be primarily
attributed to an increase in the energy required by the process
to operate the additional ultrafiltration system (an increase in
0.07 kWh/m3 over the scenario with more frequent clean-inplace events). These results are significant because they show
that the high power density of the FO membrane (with Aw3)
provides enough energy savings to over compensate for the
increase in negative impacts associated with the greater
chemical demand of this scenario. Yet, none of the scenario C
variants in Fig. 6 achieved substantially less environmental
impact than the other variants. This relative indifference in
environmental impacts raises the issue of tradeoffs in optimization of the process and will be discussed later in this section.
The comparative analysis of the three scenarios presented
in Fig. 3 is revisited in light of the FO membrane module design
modifications, performance improvements, and operation
adjustments discussed above. A comparative LCA for scenarios
A and B with two of the modifications to scenario C are shown in
Fig. 7 and normalized scoring of the impact categories are
provided in Figure S4 in the Supplemental Content document.
1152
w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 1 4 5 e1 1 5 4
These improvements include doubling the FO membrane
packing density from 9.5 to 20 m2 per 8-inch membrane module
and either doubling Aw and adding ultrafiltration pretreatment
of the wastewater stream prior to the ODN or tripling Aw and
maintaining a monthly clean-in-place frequency (C5 with
pretreatment of ODN feed and C6). With these two modified
cases of scenario C, the ODN-SWRO process is shown to have
lower environmental impact than scenario A for all impact
categories and all but two categories (ozone depletion and
terrestrial ecotoxicity potentials) when compared to scenario B.
The contribution of Na4EDTA manufacturing increased the
contribution of scenario C variants relative to the scenario B
ODP impact category. The contribution of NaOH production and
treatment of spent clean-in-place chemicals account for the
slightly greater terrestrial ecotoxicity potential of the scenario C
variant with monthly chemical cleaning.
3.4.
LCIA contribution analysis for hybrid ODN variants
An analysis was performed to determine the contribution of
individual unit processes to the overall impact category values
for two scenario C5 variants presented in Fig. 6 e scenario C5
with or without ultrafiltration pretreatment of the ODN feed.
Contribution plots for the two variants are presented in Fig. 8a
and b for the variant with and without ultrafiltration
pretreatment, respectively. The contribution of the UCTE
electricity mix is shown to be the dominant factor for both
scenarios and across all impact categories. For the scenario C
variant without ultrafiltration pretreatment, chemicals
required for the monthly clean-in-place are the second most
significant contribution and the scale inhibitor for the SWRO
membranes is primarily responsible for it being the third most
significant. When ultrafiltration pretreatment of the
Fig. 8 e Contribution analysis of various unit processes to
the ten impact categories for scenario C5 without UF
pretreatment of ODN feed (a) and with UF pretreatment of
ODN feed (b). DF indicates the 100 mm disk filter. In the legend,
C [ Chemicals, M [ Materials, and EM [ Energy mix.
wastewater stream is included and the CIP frequency reduced
to twice per year, the consumables (i.e., FeCl3 and scale
inhibitor) of the wastewater ultrafiltration and SWRO
membrane processes alternate between the second and third
greatest contributions among different impact categories.
3.5.
Contribution of energy source to environmental
impacts
Fig. 7 e Comparative LCIA for scenarios A, B, C5 with UF
pretreatment of ODN feed, and C6. Legend descriptors are
provided in the captions of Figs. 5 and 6. Maximum impact
category values are the same as in Fig. 3.
The UCTE electricity mix chosen to study the comparative life
cycle impacts of the various water treatment scenarios
contains a balanced mix of energy derived from fossil fuel
sources, renewable sources, and nuclear power. If the electricity mix were composed of less polluting energy sources,
the contribution of each unit process to an individual impact
w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 1 4 5 e1 1 5 4
1153
(scenario B) produces the least environmental impact
compared to the baseline ODN-SWRO scenario. Currently
achievable improvements to the water permeability of FO
membranes and membrane packing density for an ODN
process can substantially reduce the environmental impacts
of the hybrid ODN-SWRO process across all impact categories;
yet, at the current state of the technology, its environmental
impact is only marginally lower than that of established
technologies such as hybrid nanofiltration-RO for water
reclamation. Enhancement of water pretreatment to protect
the FO membranes and reduce the environmental impact of
ODN due to high cleaning frequency may produce additional
environmental impacts associated with operation of
advanced pretreatment processes, making the environmental
impact difference between the two options insignificant.
Analysis of the effective power density of the ODN
membrane in the hybrid process showcases the remarkable
ability of the process to improve the utilization of energy
resources. The ODN-SWRO hybrid process can achieve power
densities that are higher than levels currently demonstrated
by the pressure retarded osmosis process.
Acknowledgements
We acknowledge the support of California Department of
Water Resources (Grant 46-7446-R-08) and the AWWA Abel
Wolman Fellowship. Special thanks to Prof. David Muñoz at
Colorado School of Mines, Dr. Jean Debroux at Kennedy/Jenks
Consultants, and Hydration Technologies Innovations for
providing supporting information.
Appendix. Supplementary material
Fig. 9 e Contribution analysis similar to Fig. 8, but using
renewable electricity mix instead of the UCTE electricity
mix. The ten impact categories are assessed for scenario
C5 without UF pretreatment of ODN feed (a) and with UF
pretreatment of ODN feed (b). In the legend, C [ Chemicals,
M [ Materials, and EM [ Energy mix.
Details on energy use, water quality, materials used in
membrane manufacturing, and life cycle inventory are available via the Internet at http://sciencedirect.com/science/
journal/00431354, doi:10.1016/j.watres.2011.12.004.
category changes considerably compared to data presented in
Fig. 8a and b. By modeling the LCIA of scenario C variants with
an electricity mix composed of largely renewable energy
sources (e.g., Switzerland with 50% hydroelectric, 40%
nuclear, and 10% others), new distributions of impacts can be
obtained as can be seen in Fig. 9a and b. The contribution of
the electricity mix is reduced substantially and other impacts
associated with materials and chemicals consumed by the
systems produce greater relative impacts.
Achilli, A., Childress, A.E., 2010. Pressure retarded osmosis: from
the vision of Sidney Loeb to the first prototype installation e
review. Desalination 261 (3), 205e211.
Asano, T., Burton, F., Leverenz, H., Tsuchihashi, R.,
Tchobanoglous, G., 2007. Water Reuse: Issues, Technologies,
and Applications. McGraw-Hill, New York.
Baker, R.W., 2004. Membrane Technology and Applications. John
Wiley & Sons, Ltd, New York, NY.
Beerya, M., Hortopa, A., Woznya, G., Knopsb, F., Repkec, J., 2011.
Carbon footprint of seawater reverse osmosis desalination
pre-treatment: initial results from a new computational tool.
Desalination and Water Treatment 31, 164e171.
Biswas, W.K., 2009. Life cycle assessment of seawater
desalination in western Australia, vol. 56. World Academy of
Science, Engineering and Technology. 369e375.
Cath, T.Y., Childress, A.E., Elimelech, M., 2006. Forward osmosis:
principles, applications, and recent developments. Journal of
Membrane Science 281, 70e87.
4.
Conclusions
The comparative LCIA of the three different scenarios shows
that the nanofiltration-RO process to augment SWRO
references
1154
w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 1 1 4 5 e1 1 5 4
Cath, T.Y., Hancock, N.T., Lundin, C.D., Hoppe-Jones, C.,
Drewes, J.E., 2010. A multi-barrier osmotic dilution process for
simultaneous desalination and purification of impaired water.
Journal of Membrane Science 362 (1e2), 417e426.
CML, 2011. CML-IA Characterization Factors. http://cml.leiden.
edu/software/data-cmlia.html.
Dow Filmtec, 2010. Reverse Osmosis System Analysis (ROSA).
http://www.dow.com/liquidseps/design/rosa.htm.
Energy Recovery Incorporated, 2011. ERI Power Model. http://
www.energyrecovery.com/index.cfm/0/0/56-Power-Model.
html (San Leandro, CA).
Guinée, J.B., 2002. Handbook on Life Cycle Assessment: Operational
Guide to the ISO Standards. Springer, New York, NY.
Hancock, N.T., 2011. Engineered Osmosis: Assessment of Mass
Transport and Sustainable Hybrid System Configurations for
Desalination and Water Reclamation. Colordo School of
Mines, Golden, CO.
Hancock, N.T., Cath, T.Y., 2009. Solute coupled diffusion in
osmotically driven membrane processes. Environmental
Science and Technology 43 (17), 6769e6775.
Hancock, N.T., Phillip, W.A., Elimelech, M., Cath, T.Y., 2011a.
Bidirectional permeation of electrolytes in osmotically driven
membrane processes. Environmental Science and
Technology 45 (24), 10642e10651.
Hancock, N.T., Xu, P., Heil, D.M., Bellona, C., Cath, T.Y., 2011b. A
comprehensive bench- and pilot-scale investigation of trace
organic compound rejection by forward osmosis.
Environmental Science and Technology 45 (19), 8483e8490.
Hutchings, N.R., Appleton, E.W. and McGinnis, R.A., 2010. Making
High Quality Frac Water Out of Oilfield Waste. In: Proceedings
of the International Society of Petrolume Engineers 2010
Annual Technical Conference and Exhibition, Florence, Italy,
September 19e22, 2010.
ISO, 2006a. Standard 14040 e Environmental Management e Life
Cycle Assessment e Principles and Framework. International
Organization of Standardization, Ganeva.
ISO, 2006b. Standard 14044 e Environmental Management e Life
Cycle Assessment e Requirements and Guidelines.
International Organization of Standardization, Ganeva.
Lattemann, S., Höpner, T., 2008. Environmental impact and
impact assessment of seawater desalination. Desalination 220
(1e3), 1e15.
Letterman, R., 1999. Water Quality and Treatment. McGraw-Hill,
New York.
Mulder, M., 1997. Basic Principles of Membrane Technology.
Kluwer Academic Publishers, Dordrecht, The Netherlands.
Muñoz, I., Fernández-Alba, A.R., 2008. Reducing the environmental
impacts of reverse osmosis desalination by using brackish
groundwater resources. Water Research 42 (3), 801e811.
Patel, S., 2010. Norway Inaugurates Osmotic Power Plant.
powermag.com. http://www.powermag.com/issues/
departments/global_monitor/Norway-Inaugurates-OsmoticPower-Plant_2425.html.
Pré Consultants, 2011. SimaPro Version 7.2 (Amersfoort, The
Netherlands).
Raluy, R.G., Serra, L., Uche, J., 2005a. Life cycle assessment of
water production technologies Part 1: life cycle assessment of
different commercial desalination technologies (MSF, MED,
RO). International Journal of Life Cycle Assessment 10 (4),
285e293.
Raluy, R.G., Serra, L., Valero, A., 2005b. Life cycle assessment of
water production technoloiges part 2: reverse osmosis
desalination versus the ebro river water transfer.
International Journal of Life Cycle Assessment 10 (5),
346e354.
Sauvet-Goichon, B., 2007. Ashkelon desalination plant e
a successful challenge. Desalination 203 (1e3), 75e81.
Stover, R.L., 2007. Seawater reverse osmosis with isobaric energy
recovery devices. Desalination 203 (1e3), 168e175.
Swiss Center for Life Cycle Inventories, 2011. Ecoinvent 2.0,
St-Gallen, Switzerland. http://www.ecoinvent.org/.
Tchobanoglous, G., Burton, F.L., Stensel, D., 2003. Wastewater
Engineering. Treatment, Disposal and Reuse. McGraw-Hill,
New York.
Vlasopoulos, N., Memon, F.A., Butler, D., Murphy, R., 2006. Life
cycle assessment of wastewater treatment technologies
treating petroleum process waters. Science of the Total
Environment 367 (1), 58e70.
Voutchkov, N., 2010. Seawater Pretreatment. Water Treatment
Academy, Bangkok.
WHO, 2003. Total Dissolved Solids in Drinking-water: Background
Document for Preparation of WHO Guidlines for Drinkingwater Quality. WHO, Geneva.
Wilf, M., 2007. The Guidebook to Membrane Desalination
Technology. Balaban Deslination Publications, L’Aquila.
Yip, N.Y., Elimelech, M., 2011. Performance limiting effects in
power generation from salinity gradients by pressure retarded
osmosis. Environmental Science and Technology 45 (23),
10273e10282.
Yip, N.Y., Tiraferri, A., Phillip, W.A., Schiffman, J.D., Elimelech, M.,
2010. High performance thin-film composite forward osmosis
membrane. Environmental Science & Technology 44 (10),
3812e3818.
Yip, N.Y., Tiraferri, A., Phillip, W.A., Schiffman, J.D., Hoover, L.A.,
Kim, Y.C., Elimelech, M., 2011. Thin-film composite pressure
retarded osmosis membranes for sustainable power
generation from salinity gradients. Environmental Science &
Technology 45 (10), 4360e4369.