You are on page 1of 8

Cleaner Environmental Systems 2 (2021) 100026

Contents lists available at ScienceDirect

Cleaner Environmental Systems


journal homepage: www.journals.elsevier.com/cleaner-environmental-systems

Life cycle assessment of electrodialysis for sidestream nitrogen recovery in


municipal wastewater treatment
Donald Vineyard a, Andrea Hicks a, K.G. Karthikeyan b, Christy Davidson c, Phillip Barak c, *
a
University of Wisconsin-Madison, Department of Civil and Environmental Engineering, 1415 Engineering Drive, Madison, WI, 53706, USA
b
University of Wisconsin-Madison, Department of Biological Systems Engineering, 228 Agricultural Engineering Building, 460 Henry Mall, Madison, WI, 53706, USA
c
University of Wisconsin-Madison, Department of Soil Science, 1525 Observatory Drive, Madison, WI, 53706, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Nitrogen is removed during municipal wastewater treatment to reduce eutrophication of waterways and preserve
Ammonium drinking water quality. Nitrification-denitrification and sidestream partial nitrification-anammox are state-of-the-
Anaerobic digest art municipal N removal technologies, but they require energy for aeration and can release nitrous oxide as a
Modeling
fugitive greenhouse gas. An emerging sidestream electrodialysis technology is intended to both remove and
Anammox
Nitrification
recover N as ammonium-based fertilizer. This midpoint life cycle assessment compiles literature values to
Denitrification compare the state-of-the-art technologies of municipal wastewater nitrogen removal to the new electrodialysis
nitrogen removal and recovery technology, accounting for the offset of Haber-Bosch-derived ammonia. Electro-
dialysis is projected to be environmentally favorable compared to the state-of-the-art, predicting electricity sav-
ings similar to anammox and, with the offset of industrial ammonia manufacture, net negative emissions in five of
ten midpoint environmental impact categories, including global warming potential.

1. Introduction and biomass; this C is typically derived from extant organic molecules
and thus serves to decrease biological oxygen demand (USEPA, 1973).
The United States consumes approximately 12 million tons of nitro- Some reactors add an external organic substrate, such as methanol, to
gen (N) fertilizer each year (USDA, 2018), which leads to a significant improve N removal (USEPA, 1973; Purtschert et al., 1996; Theis and
load of nutrient pollution to surface waters through surface runoff and Hicks, 2012).
wastewater discharge. Environmental policy is shifting in order to Stoichiometrically, anammox requires only 50% of NHþ 4 to be
discourage nutrient pollution of waterways (USEPA, 2009) as the nation oxidized to nitrite (NO2 ) but 57% is considered ideal (Magri et al., 2012).
becomes more conscious of the deleterious effects of eutrophication and Anammox requires roughly one-third of the oxygen as
mitigating technologies mature. Point sources such as municipal waste- nitrification-denitrification and no C amendment for the same NHþ 4
water utilities are easy and effective targets for regulation and control. removal, but is incapable of processing nitrate (NO 3 ) (Eini, 2012).
Partial removal of N as part of municipal wastewater treatment is Anammox reactors are well-suited for side stream processing (USEPA,
currently a common practice to reduce nutrient load to surface water 2007) of anaerobic digester effluent liquids and can achieve high N
(USEPA, 2007). Existing targeted N removal methods primarily facilitate removal rates (Lu et al., 2011; Nifong et al., 2013). Both technologies run
the decomposition by microbial consortia of complex nitrogenous com- the risk of releasing powerful greenhouse gases in the form of methane
pounds into ammonium (NHþ 4 ) and then oxidation to atmospheric ni- (CH4) and nitrous oxide (N2O), depending on reactor conditions
trogen gas (N2) (Clark et al., 2010); nitrification-denitrification is the (Kampschreur et al., 2009; Law et al., 2011).
most common example while sidestream partial nitrification-anammox is Electrodialysis with monovalent-selective membranes is a newly
an emerging technology (van Dongen et al., 2001, Fux et al., 2002, patented technology (Barak, 2018) to recover NHþ 4 directly from waste-
Kampschreur et al., 2009; Clark et al., 2010, Eini 2012). The water and concentrate it for reuse as fertilizer. Electrodialysis is a mature
nitrification-denitrification and anammox processes require energy in desalination technology that uses an electric field to transport ions
the form of electricity for aeration and secondary processes like mixing. through ion-permeable membranes (Pruyn et al., 1969). The electrodi-
Nitrification-denitrification also requires a carbon (C) source for energy alysis stack typically consists of alternating cation- and anion-exchange

* Corresponding author.
E-mail address: phillip.barak@wisc.edu (P. Barak).

https://doi.org/10.1016/j.cesys.2021.100026
Received 30 November 2020; Received in revised form 15 March 2021; Accepted 19 March 2021
2666-7894/© 2021 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
D. Vineyard et al. Cleaner Environmental Systems 2 (2021) 100026

membranes (CEM, AEM) separated with spacers, with inert electrodes on environmental impacts of new and existing technologies to achieve
each end by which an electrical field is applied to the stack. As aqueous increased N removal goals; while technologies have immediate practical
solution flows between these specialized membranes, ions of the influent considerations on-site, their environmental benefits ripple out.
stream are separated into an enriched concentrate stream and a partially We used the OpenLCA 1.8 software package (GreenDelta, Berlin,
desalinated diluate stream. Electrodialysis can be an energy-efficient Germany) and the ecoinvent 3.5 database (Wernet et al., 2016) to build
method of concentrating ions but solutions with high electrical resis- the life cycle inventory and apply impact assessment methods. OpenLCA
tance or large desalting goals can affect stack performance. was chosen for flexibility and ease of use. Where possible, U.S.-specific
Electrodialysis has been previously explored as a method of removing processes were chosen. To interpret the inventory, we used the
NHþ 4 ions from municipal wastewater (Eisenmann and Smith, 1967; Ali midpoint categories of the USEPA's Tool for the Reduction and Assess-
et al., 2004; Ippersiel et al., 2012; Gurreri, 2020; Vineyard et al., 2020), ment of Chemicals and other environmental Impacts (TRACI) version 2.1
but the process was hampered by accumulation of calcium (Ca2þ) and (Bare, 2011) packaged with OpenLCA. TRACI was chosen for its high
magnesium (Mg2þ) scale on membranes that blocked ion flow. To regard and ubiquity in LCA literature and for its comprehensive range of
minimize energy costs, this sidestream electrodialysis technology is midpoint environmental impacts for comparison, but it should be noted
intended to be applied to high-NHþ 4 , low-C feeds such as filtered liquids that the characterizations were also U.S.-specific.
from biosolids after anaerobic digestion. In contrast to other N removal When wastewater specifics were necessary, data were taken from the
technologies, electrodialysis recovers an aqueous NHþ 4 product that has Nine Springs Wastewater Treatment Plant in Madison, Wisconsin, a mid-
potential use as an N fertilizer. size treatment plant that serves a population of 360,000 over a 180 sq.
Typical N fertilizers are derived from the Haber-Bosch process, a re- mile (466 km2) service area, with a design capacity of 50 million gallons
action that converts atmospheric N2 gas and hydrogen gas into ammonia per day (50 MGD; 189,000 m3/day) (AECOM, 2015). This plant contains
(NH3) using a metal catalyst and high temperatures. This reaction re- a mesophilic anaerobic digester and dewaters its anaerobic digestor
quires use of fossil fuel in the form of methane, a large industrial facility, sludge to produce biosolids using a gravity belt thickener (GBT). The
and a broad scale transportation network. By recapturing NHþ 4 at treat- liquid stream separated out by GBT for recycling to the headworks is high
ment plants for use in local agriculture, an electrodialysis stack can in nutrients and well-suited for side-stream treatment processes such as
theoretically offset the costs and environmental impacts of producing an anammox or electrodialysis, but can also be returned to headwaters or
equivalent amount of industrial NH3. Electrodialysis will also recover fed directly to nutrient removal reactors. Our technologies are assumed
other monovalent ions such as potassium (Kþ), sodium (Naþ), and to be fed the high nutrient liquid stream from this stream. The total GBT
chloride (Cl) in amounts that will vary depending on the composition of filtrate flow is 0.25 MGD, or 946 m3/day, and averages 1437 mg/L
the feed and selectivity of the ion exchange membranes. Of note, these NHþ þ þ
4 -N, 214 mg/L K , and 256 mg/L Na for 103 meq/L, 5.47 meq/L,
membranes will also exclude polyvalent heavy metal ions, previously and 11.1 meq/L, respectively. Electricity is assumed to be of low voltage
identified as a key risk in municipal nutrient recovery technologies (Fang from the U.S. characterized by data from the Midwest Reliability Orga-
et al., 2016; Lam et al., 2019). nization (MRO), one of the eight regional electric reliability councils
Life cycle assessment (LCA) has been used previously to evaluate the under the North American Electric Reliability Corporation (NERC).
environmental impacts associated with wastewater treatment (Guest In the literature search, publications were mined for reactor perfor-
et al., 2009; Byrne et al., 2017). The wastewater treatment sector is mance data to be incorporated into a master list. If a single publication
exploring the benefits of resource recovery as part of the global shift reported data from multiple reactors, those reactors were given equal
towards sustainability and circular economies (Guest et al., 2009). While weight in the list. When publications did not directly report the values
some studies focus on water (Fang et al., 2016) or energy recovery (Hao needed, values were instead calculated from available data within the
et al., 2019), the myriad of emerging nutrient recovery technologies are same publication; missing data were filled with the aggregate average of
an ongoing area of research (Lebuf et al., 2012; de Faria et al., 2015; publications that reported the needed upstream data. In condensing this
Malila et al., 2019; Sena and Hicks, 2018; Lam et al., 2019). distribution to singular unit processes for an OpenLCA input, the geo-
Nitrogen-specific recovery technologies (Lin et al., 2016) receive less metric mean of values was used for the process inventory. Full values are
focus in the literature as most technologies recover a combined N and P contained within the SI for further exploration. The intention of this
product (Lam et al., 2019). While other N recovery technologies exist in method was to create a process representative of the real world by taking
various stages of implementation, we are aware of none that are wide- surveyed data from multiple publications. This assumes that studied and
spread or industry standard. A literature review found no previous LCA surveyed plants are representative of plants as a whole and that the data
studies of electrodialysis for nitrogen recovery. collection methods have comparable results.
We performed an uncertainty analysis for discussion characterizing
2. Methods the potential impacts of the variability in literature values and addressing
possible alternative assumptions. High and low values from literature are
2.1. Goal and scope definition used to put boundaries on emissions and energy and material con-
sumptions for traditional technologies. Electrodialysis alternatives
The goal of this study was to evaluate an emerging wastewater N address varied operating conditions and necessary material inputs as
recovery technology and compare the life cycle impacts to existing N suggested by Vineyard et al. (2020) from electrochemical modeling and
removal methods. We evaluate four different municipal N management literature review. We use these values to create new inventories and
strategies: (1) Do nothing; (2) Nitrification-denitrification; (3) Anam- reassess them for the same environmental impacts. Differences between
mox; and (4) Electrodialysis for NHþ 4 recovery. We built life cycle in- scenarios are highlighted with comment to potential causes, effects, and
ventories for the purpose of profiling their environmental impacts using probabilities.
LCA.
This study is an expansion of Vineyard et al. (2020), which focused on 2.2. Functional unit
comparative economic and energy performance of electrodialysis for
these same technologies. Though cost is typically the major deciding Our work assumed that new technology would be implemented with
factor for municipalities mandated to implement nutrient removal, such the intention of meeting a nutrient removal goal and as such uses a mass
mandates are ultimately written with the goal of improving environ- basis for N removed; this allows decision-makers to directly compare the
mental quality, which can be evaluated using LCA with standard prac- estimated impacts incurred by meeting their goals and then size their
tices outlined in ISO 14040 (ISO, 2006) and 14044 (ISO, 2006). This reactors appropriately. The functional unit for comparison in this study
study is intended for wastewater engineers and policymakers to consider was a single kilogram (kg) of NHþ4 -N removed, a metric with precedent in

2
D. Vineyard et al. Cleaner Environmental Systems 2 (2021) 100026

nutrient recovery LCA (Byrne et al., 2017; Lin et al., 2016). Comparisons
were performed based on total supply chain environmental impacts, with
highlights on the local energy consumption and fugitive emissions due to
their high impacts.

2.3. System Boundaries

All three technologies embody high energy consumption, material


consumption, and/or fugitive emissions, have relatively simple con-
struction and disposal, and are expected to have useful lifetimes in excess Fig. 2. Intervention point.
of 20 years. Construction and disposal impacts are expected to be
negligible compared to 20 years of operation impacts (see Supplemental
water supplies, fisheries, and recreation usability, a targeted IPCC report
Information). As such, this was a gate-to-gate study beginning with the
estimated that 0.01 kg N2O-N is emitted to the atmosphere for every kg of
entry of wastewater N into the reactor and ending with the N leaving that
N released to rivers (IPCC, 2000); as TRACI does not include a global
reactor to the environment (Fig. 1). The intervention point for the side
warming characterization factor for NHþ 4 to surface water, we supple-
stream technologies is the high-nutrient liquid effluent from an anaerobic
mented the IPCC value. One kg of NHþ 4 -N, for our purposes equated to
digester (Fig. 2). We assumed that removal of N from a small side stream
1.29 kg of NHþ4 ions and 0.0157 kg of N2O gas.
is equivalent to removal from a large stream, effectively disregarding
scalability and total nutrient removal potential in favor of marginal
2.4.2. Nitrification-denitrification
impact of nutrient removal. This assessment allocated all impacts of the
The second scenario, nitrification-denitrification, is a traditional N
process to nutrient removal, disregarding the effects of any secondary
removal technology that uses large cisterns and pumps to move and
benefits and externalities within other plant processes, such as
aerate process water through a series of aerated (oxic) and anoxic zones.
decreasing biochemical oxygen demand (USEPA, 1973) or reduced
We used a value of 3.20 kWh per kg of N removed for electricity to run
digester inhibition by ammonium (Hansen et al., 2004; Calli et al., 2005).
pumps and aerators (SI, Table 1A). For methanol, we used the stoichio-
While there are no existing nitrogen recovery technologies in place to
metric ratio for heterotrophic digestion of 2.44 kg methanol per 1 kg of
compare, comparing removal to recovery processes is made possible by
NO 3 -N, which is metabolized to 2.40 kg of CO2 (Rittmann and McCarty,
the inclusion of the fertilizer offset (ISO, 2006). The production of
2001; Theis and Hicks, 2012). These values are similar to the 3.5 kWh
fertilizer-quality N is accounted as an offset of its nearest industrial
and 3.1 kg of methanol per kg N calculated by the Chesapeake Bay Pro-
analog, ammonia.
gram for marginal reduction of the effluent from 8 to 5 ppm N (2002).
In accounting for fugitive N2O emissions, the nitrification-
2.4. Scenario description
denitrification model considered values from a number of studies and
metastudies (SI, Table 2A). Many publications reported fugitive emis-
2.4.1. Do nothing
sions per N influent in various units, requiring conversion. Two publi-
The default scenario is a simple release of treated wastewater N to
cations (Wicht and Beier, 1995; Foley et al., 2010) contained a combined
surface water through effluent discharge, the bar against which other
70 data points, but were treated as a single data point due to inaccessi-
scenarios are compared. In this scenario, the N leaves the sidestream
bility of associated data. The uncertainty distribution removed two out-
untreated and re-enters the main treatment train where it will eventually
liers from both the top and bottom of the emissions profile. Despite being
be discharged from the plant. This N will then be utilized by plants and
an anoxic, metabolically active, and C-containing environment, we did
microbiota in natural waters. No energy consumption is required in this
not expect methane (CH4) emissions from the denitrification process
scenario. In addition to eutrophic effects of excess nutrients on drinking

Fig. 1. System boundaries.

3
D. Vineyard et al. Cleaner Environmental Systems 2 (2021) 100026

(Westermann and Ahring, 1987; Islas-Lima et al., 2004; Ettwig Katharina Considering the outputs, electrodialysis generated no fugitive N2O
et al., 2008). emissions; emissions were three times higher for an anammox reactor
than a denitrification reactor and twice as high for the do nothing sce-
2.4.3. Anammox nario, which also was the only scenario with an ammonium ion output.
Energy consumption data were obtained from Maurer et al. (2003), Electrodialysis is the most environmentally favorable technology in
WSHD (2012), and Lackner et al. (2014), through Hauck et al. (2016), six of ten environmental impact categories considered (Table 2), with
where the geometric mean was 1.44 kWh/kg N (SI, Table 3A). The N2O anammox having a narrow advantage in the other four. The offset of
emission data for anammox reactors were taken from literature values industrial NHþ 4 production in the electrodialysis model led to a net
(SI, Table 4A) and averaged 0.0209 kg N2O/kg N; anammox was not negative impact, i.e., an environmental benefit, in five of the impact
expected to yield fugitive CH4 emissions. It should be noted that the categories, while the combination of high electricity usage with meth-
relationship between reactor conditions and fugitive emissions is not yet anol consumption in the nitrification-denitrification model led to the
well understood and might be improved by alternative reactor schemas largest impact in every TRACI category tested except for eutrophication.
and aeration regimes (Domingo-Felez et al., 2014). All N removal strategies generated lower eutrophication than the Do-
Nothing scenario by two orders of magnitude. While allowing the N to
2.4.4. Electrodialysis discharge freely into the environment did generate global warming po-
An electrodialysis stack is a relatively simple machine composed of a tential through the generation of N2O, both nitrification-denitrification
pump, electrodes, and hydrocarbon-based membranes. The electrodial- and anammox generated higher total global warming impacts from the
ysis membranes are predicted to have a lifetime of 10 years (Vineyard combination of fugitive emissions and energy consumption due to aera-
et al., 2020; AECOM, 2015). Other components are not expected to tion. Fossil fuel depletion, acidification, and smog production were
require unusual maintenance or replacement. The impacts of this stan- strongly impacted by methanol production and the fertilizer offset. The
dard upkeep are expected to be negligible compared to the high elec- toxicity emission categories of carcinogens, non-carcinogens, ecotoxicity,
tricity consumption and fertilizer offset. The manufacture and disposal of and respiratory effects were electricity-dominated effects.
the unit is expected to be similarly negligible. In an electrodialysis stack, The four scenarios had significant differences in their greenhouse gas
the total energy consumption is the product of the voltage and the (GHG) profiles (Fig. 3). NHþ 4 discharge to surface water bodies in the do-
electric current. The relationship between these stack conditions and nothing scenario was predicted to result in significant N2O generation.
resultant voltage is a complicated one. Vineyard et al. (2020), a com- The nitrification-denitrification process had double the electricity-
panion article to this study, predicted that an economically optimal stack related emissions of anammox, as well as additional methanol produc-
of in our assumed conditions would consume 2.36 kWh/kg N. This value tion and respiration emissions, but it emitted a third of the N2O of
represented an optimal profit method of current selection and was pre- anammox and less than half the N2O of a surface water discharge.
dicted to remove roughly 41.6% of NHþ 4 from the stream in a single stage Fugitive N2O generation dominated the global warming profile of the
of treatment. anammox process. Despite requiring no C substrate and less oxygenation
As electrodialysis is also a N recovery technology that produces than nitrification-denitrification, anammox showed only an 11% savings
fertilizer-quality NHþ 4 , this process contained an offset of 1 kg of N in against nitrification-denitrification. By contrast, electrodialysis was not
NHþ 4 for each kg of N removed, manifested here as an input of negative projected to generate N2O and the offset of NHþ 4 production eliminated
production. Ammonia was chosen as the offset fertilizer because it is the methane-related emissions from the supply chain. Greenhouse gas
direct chemical analogue to the NHþ 4 salts produced. No weight, positive emissions from its moderate energy consumption were outweighed by
or negative, was given to K, Na, bicarbonate or chloride byproducts. The offsets in NHþ4 production.
ecoinvent library's market option for NH3 was chosen for the inventory
flow. No N2O or CH4 emissions from the electrodialysis process were 4. Discussion
expected.
4.1. Uncertainty analysis
3. Inventories and impacts
Each of the four scenarios had uncertainty in at least one input or
The life cycle inventories of the four scenarios (1) Do nothing; (2) output due to the range of literature reported values. In order to address
Nitrification-denitrification; (3) Anammox; and (4) Electrodialysis for this uncertainty, we selected literature values as alternative LCA inputs
NHþ 4 recovery are shown in Table 1. Looking at the inputs, electrodialysis and outputs (Table 3) to assess the effects on predicted environmental
used 74% of the electricity of an equivalent capacity nitrification- impacts (Tale 4). These alternative models were intended to help
denitrification reactor; an anammox reactor used 40%. Only conceptualize the ranges of possible emissions profiles from real world
nitrification-denitrification required a chemical methanol input and its implementation. The SI contains tables with a more detailed accounting
conjugate respiration to CO2 as an output. Only electrodialysis generated of the data sources and uncertainty (Tables 1A–4A) and modeling of
a fertilizer offset, shown here as a negative ammonia liquid input. additional scenarios (Tables 5A–12A).

Table 1
Flows of the models of scenarios removing 1 kg of N.
I/O Flow Do Nitrification- Anammox Electro- Provider/Category
Nothing Denitrification dialysis

Inputs Ammonia, liquid (kg) – – – 1.21 market for ammonia, liquid j ammonia, liquid j APOS, U - RER
Electricity, low voltage – 3.20 1.44 2.36 market for electricity, low voltage j electricity, low voltage j APOS,
(kWh) U - MRO, US only
Methanol (kg) – 2.44 – – market for methanol j methanol j APOS, U - GLO
Outputs Ammonium, ion (kg) 1.29 – – – Emission to water/unspecified
Dinitrogen monoxide 0.0157 0.00647 0.0209 – Emission to air/unspecified
(kg)
Carbon dioxide, fossil – 2.40 – – Emission to air/unspecified
(kg)

4
D. Vineyard et al. Cleaner Environmental Systems 2 (2021) 100026

Table 2
Environmental impact results per kg NH4-N removed.
Impact category Unit Do nothing Nitrification-denitrification Partial nitrification-anammox Electro-dialysis
-3
Acidification kg SO2 Eq (10 ) – 12.2 3.02 2.63
Carcinogenics CTUh (109) – 250 91.7 107
Ecotoxicity CTUe – 52.8 19.7 23.7
Eutrophication kg N Eq (10-3) 1002 20.9 8.27 11.2
Fossil Fuel Depletion MJ surplus – 14.1 0.715 6.05
Global Warming kg CO2 Eq 4.68 8.10 7.23 0.810
Non-Carcinogenics CTUh (109) – 1100 335 218
Ozone Depletion kg CFC-11 Eq (10-9) – 775 83.6 440
Respiratory Effects kg PM2.5-Eq (103) – 7.38 3.04 3.85
Smog kg O3 Eq (10-3) – 133 27.0 23.6

Fig. 3. Comparison of global warming potential by energy consumption and fugitive emissions between the three N removal technologies for the removal of 1 kg of
NH4-N.

appropriately included in the impacts. This alternative accounting would


make doing nothing a greater acidifier than any other option by multiple
orders of magnitude.
Nitrification-denitrification had a range of alternative energy con-
sumptions and fugitive emissions reported in literature. Both high and
low reported values were tested alongside a nitrification-denitrification
basin that did not require methanol as a C substrate. In a standard
nitrification-denitrification reactor, methanol was found to account for
over half of the global warming impacts and two-thirds of acidification,
fossil depletion, and non-carcinogenic human health impacts due to both
the industrial production itself and the eventual oxidation to CO2;
reducing the usage of methanol equivalently reduced these emissions.
A recent metastudy (Hu et al., 2013) strongly suggested that the IPCC Methanol sourced from a non-fossil C substrate such as food waste or
(2000) estimate of fugitive N2O from surface water discharged NHþ 4 is an sludge cycled from another plant stream might also lower the global
overestimate and that a more representative emission factor might be as warming impacts as the eventual decomposition would not result in a net
low as 0.0016 kg. If the Hu et al., 2013 data were correct, then the GHG increase; an alternative C source might even be allocated an offset of
Do-Nothing scenario might have a much smaller global warming impact the disposal costs for such a substrate. Fugitive N2O emissions ranged
of 0.75 kg CO2-Eq/kg N, an 84% reduction. This would reduce the from insignificant (<0.5 kg CO2-Eq/kg N) to completely dominant over
Do-Nothing scenario to the lowest global warming impact among the the GHG profile (>28 kg CO2-Eq/kg N). Variability in reported electricity
options, with the exception of the electrodialysis stack that recycles the N usage did not dramatically affect results; the difference between the
to fertilizer to offset NHþ4 production. The oxidation reaction of NH4 to
þ
lowest (2.5 kWh/kg N) and highest (4.17 kWh/kg N) reported energy
NO 3 in a natural waterway stoichiometrically results in a net two consumptions was about 1 kg CO2-Eq/kg N.
hydrogen ions produced per NHþ 4 ion oxidized. Literature estimated that Anammox had a distribution of literature values similar to
half of this NO 3 will undergo plant uptake with an equivalent amount of nitrification-denitrification. The lowest N2O emitting reactors showed
alkalinity production upon assimilation (Havlin et al., 2014), leading to a fugitive emissions as only 20% of the total GHG profile (0.26 kg CO2-Eq),
potential net of 1 mol of Hþ per mole of NHþ 4 ; as the acidification impact while the highest emitting reactors showed fugitive GHG emissions forty
category includes effects on natural waterways, this pH change might be times higher than energy based GHG emissions (41 kg CO2-Eq). The

5
D. Vineyard et al. Cleaner Environmental Systems 2 (2021) 100026

highest estimate for an anammox reactor (41.8 kg CO2-Eq/kg N) pre-

Recovery
dicted more greenhouse gases than the highest estimate for nitrification-
denitrification (34.6 kg CO2-Eq/kg N). The highest electricity consump-

2.36
No
tion in an anammox reactor (4.17 kWh/kg N) was equivalent to the



highest electricity consumption in a nitrification-denitrification reactor.
Govern This suggested that reactor performance was of particular importance to

1.21

2.04t
Mc-

anammox reactors as poor function leading to fugitive emissions can



eliminate expected technological gains from energy savings. It was also
possible that the mesophilic bacteria require additional energy for heat
Limiting
Current

pumping in some reactors where prior digestion does not create warm
1.21

3.67
enough conditions, though data of this level of detail were not provided
Electrodialysis



in-source.
Default

Electrodialysis with the economically optimal current as predicted by


1.21

2.36

Tanaka (2015) (2.36 kWh/kg N) (Vineyard et al., 2020) was the scenario


considered, but four alternative energy consumption values were tested:
0.00088

energy consumption at the limiting current density model (3.67 kWh/kg


N) (Vineyard et al., 2020); energy consumption at maximum hourly
1.44
N2O
Low


profitability (3.09 kWh/kg N); energy consumption at optimal cost as


predicted by an alternative model (2.04 kWh/kg N) (McGovern et al.,
0.137
High

1.44
N2 O

2014); and energy consumption for a chloride targeting analogue



considered in a vendor proposal (2.42 kWh/kg N) (AECOM, 2015),


0.0209
Energy

covered in the SI. The vendor proposal included acid usage to remove
Low

0.8

mineral scale.


Each of these scenarios showed changes in impacts proportionate to


0.0209
Energy

the change in electricity usage; acid addition played a qualitatively


High

4.17

insignificant role in the vendor scenario. Three alternative energy mixes



were also tested in the SI: the Midwest Reliability Organization, the
Anammox

Default

0.0209

Western Electric Coordinating Council, and the Northeast Power Coor-


1.44

dinating Council. The Midwestern mix was found to be the greatest GHG


emitter (0.577 kg CO2-Eq/kg N after methanol offset), likely owing to


Methanol

the disproportionate use of coal in the Midwest region.


0.00647

The offset of industrial ammonia production is a large contributor to


3.2
No

the emissions profile of electrodialysis. Photochemical oxidation, acidi-



fication, and global warming impacts are the most sensitive to the offset.
0.00023

If captured ammonia is not used to offset industrial fertilizer, modeled in


2.44
N2 O
Low

3.2

2.4

the No-Recovery scenario, global warming impacts would be positive and


more than double the reported magnitude; acidification and oxidation


would be four and six times their reported magnitudes, respectively. In
0.0953
High

2.44
N2 O

the ozone depletion, eutrophication, and non-carcinogenic human health


3.2

2.4

categories, the impacts of energy consumption are negligible compared


to the offset, while in the respiratory effects and ecotoxicity categories
0.00647
Energy

the offsets are of a lower magnitude compared to the impacts of energy


2.44
Low

2.5

2.4

consumption. Without the ammonia recovery offset, N removal by elec-


Nitrification-Denitrification

trodialysis would have a larger impact than anammox in every category


except global warming. As a final test, ammonia fertilizer was replaced
0.00647
Energy
High

4.17

2.44

with the common fertilizer urea (SI), which resulted in higher offsets in
2.4

each category and net negative emissions of greater magnitude in the


cumulative results.
0.00647
Default

2.44
3.2

2.4

4.2. Additional considerations


0.00251

Electrodialysis run at optimal cost efficiency was projected to


Lower
Alternative flows of N management processes.

1.29
N2O

consume more electricity per kg of N than an anammox reactor, but the



Do Nothing

increased energy consumption was offset in the supply chain by the


avoided industrial ammonia production; though a utility will not expe-
Default

0.0157
1.29

rience these energy savings itself, it may be able to pursue a grant or


subsidy for benefits to the local market.


An electrodialysis stack can theoretically concentrate NHþ 4 with
Ammonia, liquid

Ammonium, ion

Carbon dioxide,
Electricity, low

equimolar chloride and bicarbonate up to the saturation point of about


monoxide (kg)
voltage (kWh)
Methanol (kg)

4 M, 6% N by mass, before precipitation and scaling begin at 20  C


Dinitrogen

fossil (kg)

(Barak, 2018; unpublished data). The product would be a dilute liquid


Flow

fertilizer composed of NHþ


(kg)

(kg)

4 with chlorides and bicarbonates. Though this


product would avoid typical waste-derived fertilizer concerns about the
over application of phosphorus, there will be concerns regarding excess
Outputs
Table 3

Inputs

chloride application. Compared to anhydrous ammonia (82%), urea


I/O

(46%), or even aqua ammonia (20%), a 6% solution would incur greater

6
D. Vineyard et al. Cleaner Environmental Systems 2 (2021) 100026

transportation, storage, and application costs as a fertilizer. These costs sensitive to fugitive N2O emissions, suggesting that proper reactor
might be offset by a much shorter transportation distance between pro- function should be a top priority for operators seeking to reduce global
duction and application. Comparing a 6% solution to a 46% urea load warming impacts.
suggests 7.8 times the transportation costs both environmentally and
economically. With a rough, ecoinvent-based estimate of 0.22 kg CO2-Eq
per 1000 kg-kilometers transported, the kg of electrodialyzed N incurs Declaration of competing interest
3.19 extra kg of CO2-Eq per 1000 km compared to urea. It would take an
average facility-to-farm transport distance of over 750 km by truck to The authors declare the following financial interests/personal re-
invalidate the industrial ammonia offset in the comparison. Volatilizing lationships which may be considered as potential competing interests:
and condensing recovered N as ammonia or ammonium bicarbonate One of the authors, Phillip Barak, is a co-owner of Nutrient Recovery and
would increase electricity costs but reduce transportation costs. Though Upcycling, LLC, which is rights holder to US Patent 10,125,428 (issued
construction-phase impacts were expected to be negligible, other studies Nov. 13, 2018) for the removal of nitrogen from wastewater using
of different scopes and methods (Lin et al., 2016; Hao et al., 2019) have monovalent-selective electrodialysis, with international patent applica-
suggested the high concrete, steel, and fuel consumption in construction tions pending. This work was funded by the United States Department of
may be significant. Accounting for these impacts would likely widen the Agriculture/National Institute of Food and Agriculture through Hatch
gap between electrodialysis and traditional technologies. Project No. WIS01920 and INFEWS/T3 (Project No. 2017-67003-26055).
Short and long term effects of membrane scale and fouling remain an We thank the Nine Springs Wastewater Treatment Plant and the Madison
unresolved issue here. Eisenmann and Smith (1967) also found that Metropolitan Sewerage District for providing plant and reject water
membranes rapidly scaled with Ca and Mg under operating conditions. specifications.
Electrodialysis also faces potential membrane fouling from trace organic
material. Wastewater, even post-treatment, contains high levels of Appendix A. Supplementary data
complex biologically-derived C molecules. These molecules tend to
behave as if they were strongly negatively charged and so interact with Supplementary data to this article can be found online at https://
an electrodialysis current, forming a layer of sorbed organic molecules on doi.org/10.1016/j.cesys.2021.100026.
the membranes that will impede ion flow (Oztekin and Altin, 2016).
Scale, whether organic or inorganic, may dramatically increase energy References
costs by increasing electrical resistance. Prevention and remediation of
scale within the electrodialysis chamber might require pretreatment of AECOM, 2015. Chloride compliance study nine Springs wastewater treatment plant final
the solution either to remove solutes before precipitation or prevent their report. Prepared for: Madison metropolitan sewerage District, nine Springs
wastewater treatment plant, Madison, WI.
detrimental effects on the electrodialysis process. Pretreatment could https://www.madsewer.org/Portals/0/ProgramInitiatives/Ch
have impacts to both the reactor's environmental life cycle inventory and lorideReduction/MMSD%20Chloride%20Compliance%20Study%20Report%20-%
to the function of the whole treatment plant. Monovalent-selective 20Final%206-19-15 bookmarks.pdf.
Agriculture, United States Department of, 2018. Fertilizer use and price, in: workbook, All
membranes are expected to vigorously exclude the large, highly fertilizer use and price tables in a single workbook. https://www.ers.usda.gov/data
charged organic molecules in a fashion that should prevent fouling and -products/fertilizer-use-and-price.aspx.
reject the Ca and Mg ions that cause mineral scaling. Future work should Ali, M., Rakib, M., M Laborie, S., Viers, P., Durand, G., 2004. Coupling of bipolar
membrane electrodialysis and ammonia stripping for direct treatment of wastewaters
investigate the performance of reactors in broader operating conditions, containing ammonium nitrate. J. Membr. Sci. 244 (1–2), 89–96. https://doi.org/
particularly with respect to scale formation and membrane fouling. 10.1016/j.memsci.2004.07.007.
Given the complex interrelations of wastewater treatment processes, Barak, P., 2018. Electrodialysis stacks, systems, and methods for recovering ammonia and
monovalent salts from anaerobic digestate. U.S. Patent 10 (125), Ш, 13 Nov 2018.
there may be unforeseen externalities to implementation of these tech-
(US Patent Application 20150308001. Filed: 24 Apr 2015, Publication date: 29 Oct
nologies not captured in this study. Some wastewater treatment systems 2015).
may not have suitable influent for processing, may have side stream Bare, J., 2011. TRACI 2.0: the tool for the reduction and assessment of chemical and other
environmental impacts 2.0 Clean Technol. Environ. Pol. 13 (5), 687–696. https://
processes incompatible with new nutrient removal, or may have existing
doi.org/10.1007/s10098-010-0338-9.
products degraded or reactors impeded by nutrient removal. It is also Byrne, D.M., Lohman, H.A., Cook, S.M., Peters, G.M., Guest, J.S., 2017. Life cycle
possible that N removal in-plant can reduce the occurrence of ammonia- assessment (LCA) of urban water infrastructure: emerging approaches to balance
toxicity in municipal sludge digesters (Desloover et al., 2012; Desloover objectives and inform comprehensive decision-making. Environ. Sci.: Water Res.
Technol. 3 (6), 1002–1014.
et al., 2015); improved digester performance can decrease end waste Calli, B., Mertoglu, B., Inanc, B., Yenigun, O., 2005. Effects of high free ammonia
mass and cost, while yielding higher biogas recovery for on-site energy concentrations on the performances of anaerobic bioreactors. Process Biochem. 40
generation (Karim et al., 2005; Gungor and Karthikeyan, 2008; Holly (3–4), 1285–1292. https://doi.org/10.1016/j.procbio.2004.05.008.
Chesapeake Bay Program, 2002. Nutrient Reduction Technology Cost Estimations for
et al., 2017). Current allocation methods disregard any costs or benefits Point Sources in the Chesapeake Bay Watershed. The Nutrient Reduction Technology
of pathogen reduction or C removal in the reaction. Future analysis Cost Task Force A Stakeholder Group of the Chesapeake Bay Program.
would benefit greatly from a characterization of the interaction between Clark, D.L., Hunt, G., Kasch, M.S., Lemonds, P.J., Moen, G.M., Neethling, J.B., 2010. HDR
Engineering Inc. Water Environment Research Foundation Nutrient Management
N removal processes and overall treatment plant performance, particu- Report, p. 978, 1-84339-617-8/1-84339-617-3.
larly in the emission of fugitive GHGs and aeration energy requirements Corominas, L., Larsen, H.F., Flores-Alsina, X., Vanrolleghem, P.A., 2013. Including life
of main stages. An expanded scope would be within the recommenda- cycle assessment for decision-making in controlling wastewater nutrient removal
systems. J. Environ. Mgmt. 128, 759–767. https://doi.org/10.1016/
tions of Corominas et al. (2013).
j.jenvman.2013.06.002.
de Faria, A.B., Sperandio, M., Ahmadi, A., Tiruta-Barna, L., 2015. Evaluation of new
5. Conclusions alternatives in wastewater treatment plants based on dynamic modelling and life
cycle assessment (DM-LCA). Water Res. 84, 99–111.
Desloover, J., Woldeyohannis, A.A., Verstraete, W., Boon, N., Rabaey, K., 2012.
Based on these theoretical LCA results, N removal and recovery by Electrochemical resource recovery from digestate to prevent ammonia toxicity during
unimpeded electrodialysis appear to be far and away the preferable anaerobic digestion. Environ. Sci. Technol. 46, 12209–12216. https://doi.org/
10.1021/es3028154.
environmental choice of N removal technologies, projecting net negative
Desloover, J., De Vrieze, J., Van de Vijver, M., Mortelmans, J., Rozendal, R., Rabaey, K.,
impacts in key categories including GHG emissions and fossil fuel 2015. Electrochemical nutrient recovery enables ammonia toxicity control and biogas
depletion. Electrodialysis is predicted to consume 64% more electricity desulfurization in anaerobic digestion. Environ. Sci. Technol. 49, 948–955. https://
than an equivalent capacity anammox reactor but 26% less than a doi.org/10.1021/es504811a.
Domingo-Felez, C., Mutlu, A.G., Jensen, M.M., Smets, B.F., 2014. Aeration strategies to
nitrification-denitrification reactor. Models of aeration-based N removal mitigate nitrous oxide emissions from single-stage nitritation/anammox reactors.
technologies, particularly anammox, were found to be extremely Environ. Sci. Technol. 48 (15), 8679–8687. https://doi.org/10.1021/es501819n.

7
D. Vineyard et al. Cleaner Environmental Systems 2 (2021) 100026

Eini, E.J., 2012. Reducing Operating Cost with Anammox in Wastewater Treatment – A Lebuf, V., Accoe, F., Vaneeckhaute, C., Meers, E., Michels, E., Ghekiere, G., 2012. Nutrient
Simulation Study. Chemical Engineering. Ryerson Univ., Engineering Commons, Recovery from Digestates: Techniques and End-Products. In 4th International
p. 113 (Toronto, Canada). Symposium on Energy from Biomass and Waste.
Eisenmann, J.L., Smith, J.D., 1967. Electrodialysis in Advanced Waste Treatment. Lin, Y., Guo, M., Shah, N., Stuckey, D.C., 2016. Economic and environmental evaluation
Environmental Protection Agency Old Reports, p. 218. of nitrogen removal and recovery methods from wastewater. Bioresour. Technol.
Ettwig, K.F., Shima, T., Van De Pas-Schoonen, K., Kahnt, J., Marnix, H.M., Op Den Camp 215, 227–238.
Huub, J.M., Jetten, S.M., Strous, M., 2008. Denitrifying bacteria anaerobically oxidize Lu, H., Ding, S., Zheng, P., 2011. Central metabolism of anammox bacteria–a review. Acta
methane in the absence of Archaea. Environ. Microbiol. 10 (11), 3164–3173. https:// Microbiol. Sin. 51, 1014–1022.
doi.org/10.1111/j.1462-2920.2008.01724.x. Magri, A., Vanotti, M.B., Szogi, A.A., Cantrell, K.B., 2012. Partial nitritation of swine
Fang, L.L., Valverde-Perez, B., Damgaard, A., Plosz, B.G., Rygaard, M., 2016. Life cycle wastewater in view of its coupling with the anammox process. J. Environ. Qual. 41,
assessment as development and decision support tool for wastewater resource 1989–2000. https://doi.org/10.2134/jeq2012.0092.
recovery technology. Water Res. 88, 538–549. Malila, R., Lehtoranta, S., Viskari, E.L., 2019. The role of source separation in nutrient
Foley, J., de Haas, D., Yuan, Z., Lant, P., 2010. Nitrous oxide generation in fullscale recovery–comparison of alternative wastewater treatment systems. J. Clean. Prod.
biological nutrient removal wastewater treatment plants. Water Res. 44 (3), 219, 350–358.
831–844. https://doi.org/10.1016/j.watres.2009.10.033. Maurer, M., Schwegler, P., Larsen, T.A., 2003. Nutrients in urine: energetic aspects of
Fux, C., Boehler, M., Huber, P., Brunner, I., Siegrist, H., 2002. Biological treatment of removal and recovery. Water Sci. Technol. 48 (1), 37–46.
ammonium-rich wastewater by partial nitritation and subsequent anaerobic McGovern, R.K., Weiner, A.M., Sun, L., Chambers, C.G., Zubair, S.M., Lienhard, V.,J.H.,
ammonium oxidation (anammox) in a pilot plant. J. Biotechnol. 99, 295–306. 2014. On the cost of electrodialysis for the desalination of high salinity feeds. Appl.
https://doi.org/10.1016/S0168-1656(02)00220-1. Energy 136, 649–661. https://doi.org/10.1016/j.apenergy.2014.09.050.
Guest, J.S., Skerlos, S.J., Barnard, J.L., Beck, M.B., Daigger, G.T., Hilger, H., Jackson, S.J., Nifong, A., Nelson, A., Johnson, C., Bott, C.B., 2013. Performance of a full-scale
Karvazy, K., Kelly, L., Macpherson, L., Mihelcic, J.R., 2009. A New Planning and sidestream DEMON® deammonification installation. Proc. Water Environ. Fed.
Design Paradigm to Achieve Sustainable Resource Recovery from Wastewater. 3686–3709. https://doi.org/10.2175/193864713813685700, 2013,2013.
Gungor, K., Karthikeyan, K.G., 2008. Phosphorus forms and extractability in dairy Oztekin, E., Altin, S., 2016. Wastewater treatment by electrodialysis system and fouling
manure: a case study for Wisconsin on-farm anaerobic digesters. Bioresour. Technol. problems. Turk. J. Sci. Technol. 6 (1) (Online).
99, 425–436. https://doi.org/10.1016/j.biortech.2006.11.049. Pruyn, K.T., Harrington, J.J., Smith, J.D., 1969. Mathematical Model of the
Gurreri, L., Tamburini, A., Cipollina, A., Micale, G., 2020. Electrodialysis applications in Electrodialysis Process. Department of the Interior. Federal Water Quality Admin.,
wastewater treatment for environmental protection and resources recovery: a Cincinnati, Ohio.
systematic review on progress and perspectives. Membranes 10 (7), 146. ISSN: 2077- Purtschert, I., Siegrist, H., Gujer, W., 1996. Enhanced denitrification with methanol at
0375. WWTP Zürich-Werdh€ olzli. Water Sci. Technol. 33, 117–126. https://doi.org/
Hansen, T.L., Schmidt, J.E., Angelidaki, I., Jansen, E.M., Mosbaek, H., Christensen, T.H., 10.1016/0273-1223(96)00465-9.
2004. Method for determination of methane potentials of solid organic waste. Waste Rittmann, B.E., McCarty, P.L., 2001. Environmental Biotechnology: Principles and
Mgmt 24 (4), 393–400. https://doi.org/10.1016/j.wasman.2003.09.009. Applications, International. McGraw-Hill, Singapore.
Hauck, M., Maalcke-Luesken, F.A., Jetten, M.S.M., Huijbregts, M.A.J., 2016. Removing Sena, M., Hicks, A., 2018. Life cycle assessment review of struvite precipitation in
nitrogen from wastewater with side stream anammox: what are the trade-offs wastewater treatment. Resour. Conserv. Recycl. 139, 194–204.
between environmental impacts? Resour. Conserv. Recycl. 107, 212–219. https:// Tanaka, Y., 2015. Ion Exchange Membranes: Fundamentals and Applications. Elsevier,
doi.org/10.1016/j.resconrec.2015.11.019. Amsterdam. https://doi.org/10.1016/B978-0-444-63319-4.00001-8.
Havlin, J., Tisdale, S., Nelson, W.R., Beaton, J., 2014. Soil Fertility and Fertilizers: an Theis, T., Hicks, A., 2012. White Paper: Methanol Use in Wastewater Denitrification.
Introduction to Nutrient Management, eighth ed. Pearson Publ., New York. Exponent, Inc. for the Methanol Institute, Alexandria, Virginia. Doc. no. 1105602.000
Holly, M.A., Larson, R.A., Powell, J.M., Ruark, M.D., Aguirre-Villegas, H., 2017. 0101 0712 PT01.
Greenhouse gas and ammonia emissions from digested and separated dairy manure United States Environmental Protection Agency USEPA, 1973. Nitrification and
during storage and after land application. Agric. Ecosyst. Environ. 239, 410–419. Denitrification Facilities Wastewater Treatment. Technology Transfer. August.
https://doi.org/10.1016/j.agee.2017.02.007. Revised February 1974. NEPIS publication number 625473004A. https://nepis.epa.
Hu, Z., Lotti, T., de Kreuk, M., Kleerebezem, R., van Loosdrecht, M., Kruit, J., gov/Exe/ZyPURL.cgi?Dockey&equals;20008RFG.txt.
Jetten, M.S.M., Kartal, B., 2013. Nitrogen removal by a nitritation-anammox United States Environmental Protection Agency USEPA, 2007. Wastewater Technology
bioreactor at low temperature. Appl. Environ. Microbiol. 79, 2807–2812. https:// Fact Sheet: Side Stream Nutrient Removal. Office of Water. NEPIS publication
doi.org/10.1128/aem.03987-12. number 832F07017. https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey&equals;P100
International Standard Organization, 2006a. Environmental Management–Life Cycle IL7Z.txt.
Assessment: Principles and Framework. ISO Report 14040. United States Environmental Protection Agency USEPA, 2009. An urgent call to action –
Ippersiel, D., Mondor, M., Lamarche, F., Tremblay, F., Dubreuil, J., Masse, L., 2012. report of the state-EPA nutrient innovations task group. State-EPA Nutrient
Nitrogen potential recovery and concentration of ammonia from swine manure using Innovations Task Group. https://www.epa.gov/nutrient-policy-data/nutrient-inn
electrodialysis coupled with air stripping. J. Environ. Manag. 95, S165–S169. ovations-task-group-documents.
Islas-Lima, S., Thalasso, F., Gomez-Hernandez, J., J, 2004. Evidence of anoxic methane van Dongen, U., Jetten, M.S., van Loosdrecht, M.C., 2001. The SHARON-Anammox
oxidation coupled to denitrification. Water Res. 38 (1), 13–16. https://doi.org/ process for treatment of ammonium rich wastewater. Water Sci. Technol. 44,
10.1016/j.watres.2003.08.024. 153–160. https://doi.org/10.2166/wst.2001.0037.
Kampschreur, M.J., Poldermans, R., Kleerebezem, R., Van Der Star, W.R.L., Haarhuis, R., Vineyard, D., Hicks, A., Karthikeyan, K.G., Barak, P., 2020. Economic analysis of
Abma, W.R., Jetten, M.S.M., Van Loosdrecht, M.C.M., 2009a. Emission of nitrous electrodialysis, denitrification, and anammox for nitrogen removal in municipal
oxide and nitric oxide from a full-scale single-stage nitritation-anammox reactor. wastewater treatment. J. Clean. Prod. 262, 121145. ISSN: 0959-6526.
Water Sci. Technol. 60 (12), 3211–3217. Wernet, G., Bauer, C., Steubing, B., Reinhard, J., Moreno-Ruiz, E., Weidema, B., 2016.
Kampschreur, M.J., Temmink, H., Kleerebezem, R., Jetten, M.S.M., van The Ecoinvent Database Version 3 (Part I): Overview and Methodology. The
Loosdrecht, M.C.M., 2009b. Nitrous oxide emission during wastewater treatment. International Journal of Life Cycle Assessment. Springer, New York. http://link.
Water Res. 43 (17), 4093–4103. https://doi.org/10.1016/j.watres.2009.03.001. springer.com/10.1007/s11367-016-1087-8.
Karim, K., Hoffmann, R., Thomas Klasson, K., Al-Dahhan, M.H., 2005. Anaerobic Westermann, P., Kiær Ahring, B., 1987. Dynamics of methane production, sulfate
digestion of animal waste: effect of mode of mixing. Water Res. 39, 3597–3606. reduction, and denitrification in a permanently waterlogged alder swamp. Appl.
https://doi.org/10.1016/j.watres.2005.06.019. Environ. Microbiol. 53 (10), 2554–2559.
Lackner, S., Gilbert, E.M., Vlaeminck, S.E., Joss, A., Horn, H., van Loosdrecht, M.C.M., Wicht, H., Beier, M., 1995. N2O emission aus nitrifizierenden unddenitrificierenden Kla
2014. Full-scale partial nitritation/anammox experiences – an application survey. €ranlagen. Korresp. Abwasser 42 (3), 411–413, 404–406.
Water Res. 55, 292–303. https://doi.org/10.1016/j.watres.2014.02.032. WSHD, 2012. Waterketenrapportage 2011. In: Waterschap Hollandse Delta, Ridderkerk,
Law, Y., Lant, P., Yuan, Z., 2011. The effect of pH on N2O production under aerobic p. 116.
conditions in a partial nitritation system. Water Res. 45 (18), 5934–5944. https://
doi.org/10.1016/j.watres.2011.08.055.

You might also like