Separation and Puri cation Technology 359 (2025) 130598

Contents lists available at ScienceDirect

Separation and Purification Technology

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

Multi-stage recovery of ammonia–potassium liquid fertilizer and phosphate

mineral from real human urine
Musie Welldegerima Atsbha a,b,c , Oded Nir a,c,* , Treavor H. Boyer b,*
a
Zuckerberg Institute for Water Research, Blaustein Institutes for Desert Research, Ben Gurion University of the Negev, Sde Boker 84990, Israel
b
School of Sustainable Engineering and the Built Environment (SSEBE), Arizona State University, PO Box 873005, Tempe, AZ 85287-3005, USA
c
The Goldman Sonnenfeldt School of Sustainability and Climate Change, Ben-Gurion University of the Negev, Beer Sheva 84105, Israel

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

Editor: Gaohong He This study explores a novel multi-stage process for recovering valuable nutrients–nitrogen, phosphorus, and
potassium–from real hydrolyzed urine as value-added products. The approach utilizes a combination of mem­
Keywords: brane contactor, zeolite ion exchange, and mineral precipitation techniques. A closed-loop system was estab­
Nutrient recovery lished by reusing the acid regeneration solution from ion exchange as the acid-stripping solution in the hollow
Hydrolyzed human urine
fiber membrane contactor (HFMC), thereby minimizing chemical usage. Ammonia recovery using the HFMC
Hollow fiber membrane contactor
achieved over 90 % removal across three cycles from hydrolyzed urine. Zeolite columns of chabazite and cli­
Ion exchange
Iron phosphate noptilolite demonstrated consistent potassium recovery from HFMC-treated urine, with slightly higher uptake by
chabazite compared with clinoptilolite. This suggests zeolite selection can be based on cost and availability. The
regeneration of the potassium-saturated zeolite columns using sulfuric acid exhibited rapid and substantial
amounts of potassium desorption. Potassium regeneration remained stable over two cycles, with potassium
concentrations reaching up to 14 g/L. The release of other ions, such as sodium, was minor compared with
potassium, highlighting the minimal impact of sodium interference. The combined ammonia–potassium liquid
fertilizer exhibited a favorable N:K mass ratio (3.6 % N and 0.7 % K), with negligible amounts of other ions,
making it suitable for facilitating plant growth. Iron phosphate precipitation, a promising alternative resource for
fertilizer or lithium iron phosphate batteries, was successfully achieved. Iron doses were more effective in
precipitating phosphate at neutral pH than basic pH, reaching over 90 % phosphate removal. This study provides
a promising approach for recovering valuable resources from human urine, promoting a more sustainable
approach to wastewater management and nutrient recycling.

1. Introduction exchange, chemical precipitation, membrane separation, and electro­

chemical processes, have been explored to recover these nutrients as
Adopting non-water urinals in public spaces, such as bus stations, fertilizers. These technologies aim to provide a scalable, portable, and
stadiums, shopping malls, schools, hospitals, and airports, has increased commercial-grade solution for cultivating various plant species. Ion
over time due to its efficient sanitation management. The water con­ exchange, utilizing materials like natural zeolite and hydrated ferric
servation benefits have motivated other interests in using non-water oxide composites, can separate cations (e.g., ammonium (NH+ 4 ) and
urinals to recover valuable resources, such as ammonia and phos­ potassium (K+)) and anions (e.g., phosphate (PO3- 4 )) in urine [8–11].
phate, while reducing the nutrient load to wastewater treatment facil­ Notably, chabazite and clinoptilolite zeolites are effective in ammonium
ities [1–4]. However, direct use of urine as fertilizer can increase soil and potassium removal from human urine and other waste streams [11].
salinity and release pharmaceutical micropollutants [5–7]. Therefore, The zeolite’s affinity for ions is influenced by the wastewater composi­
physicochemical methods are essential for nutrient recovery and the tion, where higher concentrations of competing ions can reduce the
production of value-added products. removal efficiency of the target ion. Despite the higher sodium-to-
Urine nutrients such as nitrogen (N), potassium (K), and phosphorus ammonium or potassium ratio, chabazite demonstrates superior affin­
(P) are vital fertilizers for agriculture. Various methods, such as ion ity for both ammonium and potassium compared to clinoptilolite, which

* Corresponding authors.
E-mail addresses: [email protected] (M.W. Atsbha), [email protected] (O. Nir), [email protected] (T.H. Boyer).

https://doi.org/10.1016/j.seppur.2024.130598
Received 25 September 2024; Received in revised form 7 November 2024; Accepted 17 November 2024
Available online 19 November 2024
1383-5866/© 2024 Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
M.W. Atsbha et al. Separation and Puri cation Technology 359 (2025) 130598

Fig. 1. The HFMC, ion exchange, and precipitation setup used to recover ammonia, potassium, and phosphate (NKP) from hydrolyzed urine. The HFMC experiment
was conducted in batch mode, with hydrolyzed urine recirculated through the membrane in a feed tank.

shows lower selectivity [11]. The selectivity and capacity of zeolite were recover K and P from urine.
significantly enhanced through physicochemical modifications using Integrating processes to recover NPK from urine underscores a ho­
heat and chemicals, improving its mechanical stability [12]. This listic approach that capitalizes on synergies and addresses the limita­
advancement extends the lifespan of zeolite, addressing the key limita­ tions inherent in individual techniques [26]. Numerous studies have
tion of durability for large-scale applications. Phosphate removal from contributed valuable insights into the integrated processes for urine
urine primarily relies on precipitating struvite by adding magnesium nutrient recovery. For instance, research by Wu et al. [27] emphasized
salts [13,14]. Alternative methods using different metal salts have been the advantages of employing ammonia stripping, potassium ion ex­
tested to produce various phosphate minerals, enhancing their economic change, and subsequent struvite precipitation, leading to enhanced re­
value and demand [15,16]. covery rates for NPK as individual nutrients compared to alternative
Membrane separation techniques like reverse osmosis (RO) [17], configurations. Similarly, the work of Yang et al [22] emphasized the
forward osmosis (FO) [18], bipolar membrane electrodialysis [19], potential of incorporating ammonia recovery using hollow fiber mem­
membrane distillation (MD) [20], and hollow fiber membrane con­ brane contactors alongside chemical precipitation for comprehensive
tactors (HFMC) [21–23] have been utilized to extract various nutrients potassium retrieval from urine streams by dosing magnesium and
from wastewater by altering its physicochemical composition through phosphate salt. Additionally, Jagtap et al. [10] demonstrated ammonia
pressure changes, electric potential, heating, and pH adjustments. While stripping–acid absorption for ammonia recovery, and simultaneous
RO effectively concentrates urine nutrients due to its high ion rejection distillation for potassium recovery from urine. The combined process
and permeability to water, it also co-concentrates undesired contami­ removed 70 % of N and recovered approximately 100 % of K and over
nants, necessitating additional post-treatment steps for value-added 50 % of water in urine [10]. The hybrid technology enhances the quality
product generation. Conversely, membrane electrodialysis selectively of the recovered product by selectively recovering N, P, and K while
separates ions based on charge, yielding concentrated cation and anion effectively excluding unwanted ions. The integration of technology for
products but requiring further treatment for value addition. MD and total nutrient recovery from urine must address challenges related to
HFMC are primarily employed for ammonia recovery from urine reducing chemical additions, energy consumption, and byproduct
through physicochemical alterations such as heating or pH adjustment, release, such as salinity (Na+).
producing high-quality ammonia products but leaving other nutrients Zeolite’s effectiveness in removing ammonium and potassium from
unrecovered [23–25]. Thus, additional processes are integrated to hydrolyzed urine offers routes for potential fertilizer applications [11].

2
M.W. Atsbha et al. Separation and Puri cation Technology 359 (2025) 130598

Most previous studies have focused solely on ammonium recovery Table 1

through ion exchange, often overlooking potassium recovery. Building The chemical composition of hydrolyzed stored urine and pH-adjusted stored
on the findings of potassium’s competitive role with ammonium during urine for membrane separation, ion exchange, and precipitation processes to
zeolite cation exchange, further research is needed to optimize the po­ recover NPK.
tassium ion exchange dynamics and regeneration for multicycle use. Chemicals Hydrolized urine (mg/ Hydrolized urine after adding a base (mg/
Existing studies, primarily conducted in batch mode, did not accurately L) L)
represent continuous urine generation. This gap highlights the need for TAN 3140 3010
further research to evaluate zeolite resilience to physicochemical K+ 1070 11,400
changes over extended periods, especially in column studies. Although Na+ 1410 1370
Ca2+ 100 100
ion exchange for ammonium and potassium recovery is promising, Mg2+ 22.5 22.5
achieving highly concentrated N or K fertilizer solutions remains chal­ Cl- 2540 2390
lenging [28]. Maximizing chemical utilization efficiency during acid SO2-4 843 932
regeneration to enhance fertilizer concentration and value through TP 161 157
TOC 138 138
integration with processes like HFMC could address critical gaps in
pH 9 11
current research. To address these gaps, this study evaluated and
compared chabazite and clinoptilolite zeolite columns for potassium
recovery and regeneration with strong acid. It also explored the reuse of hydrolyzed human urine to change the total ammonia nitrogen to an
potassium-rich acid regeneration solutions in HFMC to neutralize ammonia gas state. Sulfuric (H2SO4) acid was used to regenerate the
ammonia and produce concentrated NK fertilizer, optimizing acid use. saturated ion exchange column and subsequently reused to maintain the
Unlike previous approaches that used fresh acid for ammonia recovery pH of the acid-stripping solution in the HFMC process. Ferrous chloride
in HFMC, this study integrated potassium recovery with acid reuse be­ (FeCl2⋅4H2O), ferric sulfate (Fe2(SO4)3), and magnesium chloride
tween ion exchange and HFMC to enhance acid utilization efficiency. hexahydrate (MgCl2⋅6H2O)) were used to test the precipitation of
The overall objective of this work was to develop and evaluate a phosphate to produce iron phosphate and magnesium potassium phos­
multi-stage process for recovering ammonia and potassium as liquid phate (MPP). Hydrochloric acid (HCl) was used to clean the HFMC after
fertilizer (NK) and phosphate as an iron-phosphate mineral from hy­ the end of each experiment. The ion exchange study used clinoptilolite
drolyzed urine using HFMC, ion exchange, and precipitation techniques. and chabazite zeolites from St. Cloud Mining, USA [11]. A 3 M™ Liqui-
The specific objectives were: (1) evaluating the ammonia recovery ef­ Cel™ MM-1.7x5.5 Series Membrane Contactor, with a contact area of
ficiency of the HFMC by reusing K+-rich acid regeneration solution as 0.58 m2, supplied from 3 M™ Liqui-Cel™, USA, was used for ammonia
acid stripping to produce NK liquid fertilizer, (2) comparing K+ recovery recovery from urine [29]. The internal and external diameters (ID/OD)
using chabazite and clinoptilolite in dynamic column studies at different of the fiber, along with other morphological specifications of the HFMC,
flow rates, (3) assessing the regeneration of saturated zeolite columns are provided in Table S1 of the Supplementary Material [30].
(chabazite and clinoptilolite) using strong acid at different flow rates Urine was collected using a urinal-on-wheels system placed in a
and operation modes (full circulation and single pass-through), and (4) restroom at Arizona State University (ASU), as approved by the ASU
investigate iron phosphate precipitation at different pH levels and iron Institutional Review Board [31]. The urine was stored for two years for
sources, comparing the potential with magnesium potassium phosphate hydrolysis, with its composition detailed in Table 1. The separation
(MPP) precipitate. All objectives were achieved using real hydrolyzed system includes an automatic acid dosing system to prevent phosphate
urine in each treatment stage. precipitation in the piping. The total phosphate concentration in hy­
drolyzed urine was lower than that in fresh urine due to the precipitation
2. Materials and methods of phosphate with calcium (Ca2+) and magnesium (Mg2+) ions in the
storage tank, a finding also reported in previous studies [10,28]. Raising
2.1. Description of process the pH from 9 to 11 increases the thermodynamic precipitation potential
for minerals like calcium phosphate and Mg(OH)2. However, after pH
Fig. 1 shows the process investigated in this research. First, HFMC adjustment, no precipitation was visible and no significant changes were
was used to recover ammonia as liquid fertilizer from hydrolyzed urine, measured in phosphorus, calcium, or magnesium concentrations
using KOH to raise the pH. In the second stage, the ion exchange process (Table 1). This lack of evidence for precipitation suggests a kinetic in­
was tested to recover potassium using zeolite ion exchange materials hibition of these processes, possibly influenced by the presence of or­
(chabazite and clinoptilolite) and regenerate using strong acid after ganics and the low concentration of alkaline metal in the urine matrix
saturation for multiple cycles. Ammonium recovery in the first stage that could inhibit or delay the precipitation reactions [32,33]. Never­
enables potassium recovery through ion exchange since it eliminates theless, the feed solution was pre-filtered with a 10-µm filter as a safe­
competition over exchangeable sites at the zeolite surface. The strong guard against membrane clogging. The urine pH was adjusted to 11
acid regeneration solution was used as an acid-stripping solution in using KOH for experiments targeting ammonia recovery via HFMC,
HFMC to neutralize the ammonia and maintain the partial pressure followed by ion exchange and precipitation processes. All experiments
gradient between the shell and lumen. Unlike prior practices using fresh were conducted at room temperature (25 ◦ C).
acid for ammonia recovery in HFMC, our study innovatively integrates
potassium recovery with acid regeneration. This proof of concept en­ 2.3. Hollow fiber membrane contactor setup
hances acid utilization efficiency for NK mixed fertilizer production. In
the last stage, after recovery of NK, the phosphate in the hydrolyzed The membrane contactor system includes a membrane contactor
urine was precipitated as iron phosphate by dosing ferrous/ferric salts to module, a hydrolyzed urine tank and acid-stripping tanks, and a peri­
produce an alternative product in addition to magnesium potassium staltic pump for circulating the urine and the acid-stripping solution
phosphate. This multistep process recovers NPK from hydrolyzed urine through the membrane module (Fig.S1). The hydrolyzed urine and the
as up-cycled products with minimal chemical discharge to the acid-stripping solution were streamed through the shell and lumen side
environment. of the membrane countercurrent at the same flow rate (30 L/hr),
respectively. The hydrolyzed urine was pretreated using the 10-µm filter
2.2. Chemicals and materials to remove suspended and particulate matter, which can clog or damage
the membrane contactor. KOH (0.4 mol/L) base was used to increase the
Potassium hydroxide (KOH) was used to increase the pH of the real pH of the hydrolyzed urine. The ammonia recovery experiment was

3
M.W. Atsbha et al. Separation and Puri cation Technology 359 (2025) 130598

Fig. 2. Ammonia recovery using a hollow fiber membrane contactor (HFMC): (a) Ammonia concentration in the urine tank over time for chabazite treatment, (b)
Ammonia concentration in the urine tank over time for clinoptilolite treatment, (c) Ammonia profile in the acid stripping solution over time using the K+-rich acid
regenerated solution from chabazite, (d) Ammonia profile in the acid stripping solution over time using the K+-rich acid regenerated solution from clinoptilolite, (e)
Volume of the acid stripping solution over three cycle. The HFMC was used for ammonia removal before ion exchange, with similar conditions used for each cycle in
both scenarios. The regeneration solution from the chabazite and clinoptilolite was added to the acid-stripping solution in the 2nd and 3rd cycles of the HFMC.

conducted to generate sufficient volumes of ammonia-depleted solution each scenario. The molar concentration was calculated based on the
to feed chabazite and clinoptilolite separately for subsequent ion ex­ equal molar required proton (H+) for neutralizing the transferred
change (IX) experiments. Three cycles were tested under identical con­ ammonia from the hydrolyzed urine. Then, for the second cycle, the acid
ditions for each scenario (chabazite and clinoptilolite) using 4 L of stripping was reused from the first cycle to increase the N concentration,
hydrolyzed urine per cycle. In the first cycle, as an initial acid-stripping and the strong acid regeneration from the zeolite column was added
solution, 0.4 L with 2.5 M Sulfuric acid (H2SO4) was used separately in separately to maintain the pH. The same condition was applied in the

4
M.W. Atsbha et al. Separation and Puri cation Technology 359 (2025) 130598

third cycle, reusing the stripping solution from the second cycle, and the dissolved. To enhance phosphorus removal kinetics, the precipitation
strong acid regeneration solution was added to the acid tank. Each step was used by adding metal salts, determining the required amount
zeolite column was tested separately in the regeneration, and the acid based on the initial phosphate concentration. While continuous TP
regeneration solution was added to the acid stripping solution in HFMC measurement could optimize chemical usage, however, real-time
for each scenario, respectively. The sample was taken over time from the phosphate measurement methods exceed our laboratory’s capabilities.
hydrolyzed urine and acid stripping tank to monitor the ammonia pro­ Various iron (Fe) sources, such as ferrous chloride (FeCl2⋅4H2O) and
file and other compositions (K+, Na+, etc.). After each cycle, the ferric sulfate (Fe2(SO4)3), were tested at neutral pH (7) and higher pH
ammonia-depleted hydrolyzed urine was transferred to another storage (10) for comparison. Fe(II) and Fe(III) were dosed in a 1.5:1 ratio with
tank and then directly used for the ion exchange experiment to recover the phosphate in the HFMC and ion-exchange-treated urine. MPP pre­
the potassium in each scenario. cipitation was performed by adding Mg2+ (MgCl2⋅6H2O) in a 1.5:1 ratio
at pH 10. The precipitation was carried out using 0.25 L of the treated
2.4. Ion exchange column setup urine in three flasks for each precipitation scenario. The metals (Fe(II),
Fe(III), and Mg2+) were dosed and mixed using a magnetic stirrer at 250
The ion exchange columns were constructed from clear PVC pipes, rpm. After 10 min, the solution was filtered with a vacuum filtration
each 60 cm long and with a 2.54 cm internal diameter. Two columns system using 0.45 µm filter paper. Fe(II) oxidation in this short precip­
were packed with 200 g each of clinoptilolite and chabazite zeolites to itation time was considered insignificant (Simbeye et al., 2023). Samples
test K+ recovery. A peristaltic pump introduced hydrolyzed urine and were taken from treated urine before metal dosing and from the filtrate
the regeneration solution to the columns. Before use, the zeolite columns after precipitation. The solid retained on the filter paper was dried at
were thoroughly rinsed with DI water to remove dust and impurities. room temperature and analyzed using SEM-EDS. Each experiment was
Then, both columns were preconditioned with 2.5 M H2SO4 to convert triplicated, and the average result was reported.
the zeolites from their original Na+ form (Z-Na+) to the protonated form
(Z-H+). The acid was recirculated through the columns for two hours to 2.6. Analytical method
reach equilibrium, after which the columns were rinsed with DI water to
remove any remaining unexchanged protons. Total ammonia nitrogen (TAN) and phosphate concentrations were
K+ recovery experiments were conducted using HFMC-treated hy­ determined using an automated Lachat Quikchem 8500 series 2 flow
drolyzed urine in both zeolite columns operated in up-flow mode at flow injection analysis system (FIA). Samples were analyzed in duplicate with
rates of 30 mL/min and 5 mL/min. At the higher flow rate, two ion standard calibration for accuracy verification. Total organic carbon
exchange cycles were performed with 4 L of urine to evaluate reus­ (TOC) and total nitrogen (TN) were measured using a Shimadzu TOC-L/
ability, while one cycle was conducted at the lower flow rate with 2 L of TNM-L Analyzer. Ion chromatography (Dionex ICS-1000) was employed
urine. Before regeneration, the columns were flushed with DI water to to analyze the ionic composition (K+, Na+, Mg2+, Ca2+, Cl-, SO2- 3-
4 , PO4 ,
remove residual solutions. Regeneration was carried out using 2.5 M NH+ 4 ) of samples collected from HFMC, IX, and precipitation. Scanning
H2SO4, with the solution streamed at the same flow rates and up-flow electron microscopy (SEM/FIB Focussed Ion Beam − Auriga (Zeiss))
mode used for adsorption. Two regeneration modes were tested: full assessed the morphology of dried precipitated products, while EDS
recirculation at 30 mL/min and single pass-through at 5 mL/min to determined their percent elemental composition. During the experi­
evaluate the impact of high H+ ion concentration on overcoming K+ ments, pH values were recorded using Thermo Scientific Orion Versa
selectivity during desorption. Full recirculation required less acid vol­ Star Pro pH probes. The sample was acidified to prevent ammonia
ume (0.4 L) compared to single pass-through (0.6 L), highlighting a volatilization and stabilize it during storage before analysis.
trade-off between acid concentration and regeneration volume, which
impacts freshwater consumption. Ultimately, 0.4 L of acid regeneration 3. Results and discussion
solution was selected based on the column bed volume to minimize
water usage. 3.1. Ammonia recovery using HFMC
The concentration of ions and other compositions through the ion
exchange column inlet and outlet was monitored through sampling over Fig. 2 shows the ammonia profile in the feed tank and acid stripping
time to show the breakthrough curve. Similarly, samples were taken solution over time for three cycles using HFMC. Initially, ammonia was
from the regeneration solution to monitor the concentration of K+ and recovered using HFMC, followed by potassium recovery using zeolite
other chemical compositions. The zeolite column was reused after (chabazite and clinoptilolite) in the subsequent stage. Three cycles of
regeneration with acid for the subsequent ion exchange cycles. The separate hydrolyzed urine ammonia removal experiments were con­
regeneration solution with a high concentration of acid and K+ was then ducted as pretreatment for each ion exchange stage. The ammonia re­
directly reused in the acid-stripping solution to adjust its pH. The vol­ covery from urine using HFMC was consistent across three cycles in both
ume of the regeneration solution was calculated to match the equimolar scenarios, achieving over 90 % removal (Fig. 2a & 2b). The high
proton requirement based on the ammonia concentration in the urine ammonia recovery efficiency of the HFMC with a microporous and gas-
and the targeted 90 % ammonia removal in HFMC. Proton loss during permeable membrane, which allows the transfer of gas while blocking
regeneration was expected to be minimal due to the limited capacity of non-volatile components, was also demonstrated in a previous study
the zeolite compared to the initial acid molarity, ensuring complete using ammonia-rich sources such as hydrolyzed urine [22,24,25]. The
utilization of the excess acid by the HFMC. HFMC process reduced the average TAN concentration to approximately
320 ± 178 mg/L over three cycles, sufficiently low to avoid impacting
2.5. Phosphate precipitation the subsequent ion exchange stage. Effective ammonia removal in HFMC
is crucial, as ammonium can compete with potassium in ion exchange,
The precipitation was conducted using hydrolyzed urine after affecting potassium recovery [27]. Using KOH as the base in the re­
ammonia and potassium recovery via HFMC and a zeolite column, covery process was preferable because K+ can be recycled in ion ex­
respectively. The precipitation step was conducted last, as the optimal change, minimizing chemical loss despite its higher cost compared to
pH for phosphate precipitation as iron phosphate is around neutral a NaOH (KOH is approximately 30 % more expensive than NaOH).
condition achieved after ion exchange due to proton release. Total Additionally, base consumption was minimal due to urine’s pH around
phosphorus (TP) may precipitate during hydrolyzed urine storage, but 9, requiring only a small amount (0.4 mol/L KOH) to reach pH 10 or 11.
this process does not completely remove phosphorus, as only a portion is Although NaOH is cost-effective, it increases urine salinity and com­
precipitated by the existing alkaline metals, leaving significant amounts plicates the recovery of other nutrients like K+ and PO3- 4 , making it

5
M.W. Atsbha et al. Separation and Puri cation Technology 359 (2025) 130598

Table 2 environmentally unfriendly [33]. The consistent recovery of ammonia

The chemical composition of the Liquid fertilizer recovered as NK from the using HFMC ensures a compact and stable treatment technology for
HFMC ammonia stripping with the addition of K+-rich acid regeneration solu­ ammonia-rich urine.
tion from IX. Fig. 2c & 2d show the ammonia profile in the acid stripping tank over
Chemicals (mg/L) HFMC & Chabazite HFMC & Clinoptilolite time across three cycles of ammonia recovery from urine using HFMC.
NH+ 35,800 33,400 As an acid-stripping solution, the potassium-rich acid regeneration so­
4
K+
6,650 4,040 lution was reused in HFMC to maintain a low pH and ensure a sufficient
Na+ 445 236 ammonia partial pressure gradient between the shell and lumen of the
TP 92.2 96.3 membrane. The acid solution was reused for three cycles with the
Mg2+ 33.7 42.8
Ca2+ 105 297
addition of the regeneration solution to maximize the NK concentration
Cl- 26.7 40.7 and utilize the excess acid. Initially, a fresh acid stripping solution was
TOC 11.5 2.70 used, and the %N increased consistently over time in the first cycle for
both scenarios (Fig. 2c&2d). In the second cycle, adding the K+-rich acid
regeneration solution from ion exchange to the reused acid stripping

Fig. 3. Ion breakthrough curves using chabazite and clinoptilolite zeolite columns: (a) potassium ion breakthrough curves at different flow rates; (b) podium ion
profiles in the ion exchange column over time at different flow rates; (c) pH breakthrough over time in both zeolite columns at 5 mL/min flow rate.; (d) ammonium
breakthrough curve at 5 mL/min flow rate; (e) phosphate profile over time in the chabazite column at 5 mL/min flow rate.

6
M.W. Atsbha et al. Separation and Puri cation Technology 359 (2025) 130598

Fig 4. Regeneration of chabazite and clinoptilolite zeolite column at a high flow rate with full recirculation for two cycles. (a) potassium (K+) ion desorption over
time; (b) sodium (Na+) ion release over time through the column. The saturated column was regenerated using 0.4 L of H2SO4 (2.5 M) at 30 mL/min, and the
regeneration solution was recirculated through the column for 2 hr.

solution initially diluted the ammonia concentration by half, yet it plant growth. Overall, the NK liquid fertilizer produced in this study
increased over time, reaching up to 3.2 %wt N. Similarly, in the third could play a significant role, as previous studies mainly focused on ni­
cycle, reusing the K+ acid regeneration solution and the acid stripping trogen recovery from urine. The recovery of both N and K increases the
solution increased the concentration to around 3.6 % N and 3.3 % N for fertilizer’s value and the efficiency of the integrated process by utilizing
chabazite and clinoptilolite, respectively. High ammonia recovery was the acid effectively for both HFMC and IX and maintaining treated urine
maintained regardless of whether fresh or reused acid was used. Overall, pH after treatment without adding acids.
liquid fertilizer quality and consistent production were maintained over
multiple cycles in HFMC by reusing the acid regeneration solution as the 3.2. Potassium recovery using ion exchange
acid stripping solution for cogenerating NK liquid fertilizer from urine.
Fig. 2e shows the volume of the acid-stripping solution over three 3.2.1. Potassium recovery using chabazite and clinoptilolite
cycles with the addition of the K+-rich acid regeneration solution from Fig. 3 shows the ion exchange breakthrough curves for K+, Na+,
the zeolite columns (chabazite and clinoptilolite). The volume of the NH4 , PO3-
+
4 , and pH through the zeolite columns (chabazite and cli­
acid stripping solution significantly increased over cycles due to the noptilolite) over time. The experiment utilized HFMC-treated,
addition of the acid regeneration solution, as shown in Fig. 2e, starting ammonia-depleted urine, and reported the measured relative ion con­
from 0.4 L and reaching 1 L. This indicates a higher and balanced re­ centration (C/C0) in the effluent from each zeolite column. As shown in
covery of liquid fertilizer mass. The slight difference in the obtained %N Fig. 3a, the recovery of K+ using chabazite and clinoptilolite was
in HFMC was mainly due to the volume difference of the regeneration consistent despite the differences in the breakthrough curves observed at
solution. As shown in Fig. 2e, a slightly higher volume was measured in varying flow rates. At a higher flow rate (30 mL/min), the breakthrough
the acid-stripping solution with clinoptilolite than chabazite, indicating curves for both zeolites steepened, while at a lower flow rate, the zeolite
dilution of the %N. The volume of the acid regeneration solution used in column saturated more slowly, enhancing ion exchange dynamics
ion exchange was the same for both zeolite columns. Still, a slight loss of (Fig. 3a). Previous studies have shown consistent flow rate effects for
solution was observed at the end of regeneration, slightly higher with different ions (e.g., NH+4 ), but this study is the first to demonstrate K
+
chabazite due to the density and particle size distribution of the zeolite breakthrough using these two zeolites. This was expected since the
material in the column retaining the solution, even after air draining. higher HRT provides longer contact time in the zeolite column,
Despite these variations, the ammonia profile remained consistent in enhancing the ion exchange [34–36]. Previous batch studies reported
both scenarios, demonstrating the robustness of the treatment process. higher K+ removal rates with chabazite compared to clinoptilolite;
Therefore, the regeneration solution can be reused efficiently without however, the difference was not significant in the column study (Fig. 3a)
performance decline, ensuring effective acid utilization in both HFMC [11]. Batch studies conducted at equilibrium do not reflect the contin­
and ion exchange. uous and dynamic nature of urine generation scenarios. In this study,
Table 2 presents the chemical composition of the NK liquid fertilizer both zeolites performed similarly under dynamic conditions, indicating
produced using HFMC and the reused acid regeneration solution from that selection can be based on cost and availability.
the zeolite column, accumulated over three cycles. The produced NK Fig. 3b shows the Na+ breakthrough curves at different feed flow
liquid fertilizer contained high concentrations of ammonium and po­ rates (30 mL/min and 5 mL/min) through the zeolite column over time.
tassium, with negligible amounts of other ions. The composition ranged As shown in Fig. 3b, sodium ions were mostly desorbed from both zeolite
from 3.3 % to 3.6 % N and 0.4 % to 0.7 % K+, classifying it as a good- columns through ion exchange at higher flow rates ((C/C0) > 1), as the
grade fertilizer due to the favorable N:K ratio, which meets the high K+ breakthrough overcoming selectivity at higher K+ to Na+ ratios. At
nitrogen demand for plant growth (Table 2). The potassium content lower flow rates, Na+ uptake was slightly higher in chabazite than in
from the chabazite acid regeneration solution was higher than that of clinoptilolite, even though the Na+ breakthrough curve was steeper.
clinoptilolite, as discussed in Section 3.2, due to greater K+ uptake and This interference may be due to the initially low K+ ion concentration
regeneration (Fig. 4a). The interference from other ions in the liquid (Fig. 3a), which could induce Na+ exchange at a high Na+/K+ ratio, as
fertilizer was minimal, with Na+ ranging from 0.024 % to 0.045 %, reported in previous studies [11]. Moreover, the slightly adsorbed Na+
having a low impact on product quality [10] (Table 2). Small amounts of was exchanged with K+ after the initial stage, as (C/C0) remained above
Ca2+ and Mg2+ were also present, providing additional nutrients for 1 throughout the experiment (Fig. 3b). Overall, both zeolites

7
M.W. Atsbha et al. Separation and Puri cation Technology 359 (2025) 130598

Fig. 5. Regeneration at a low flow rate with one pass through the chabazite and clinoptilolite zeolite column: (a) potassium (K+) ion desorption over time during
regeneration; (b) sodium (Na+) ion release over time through the column; (c) composition of other ions in the regeneration solution over time.

demonstrated higher selectivity for K+ in the column study, with previous studies [27,42].
negligible selectivity for sodium ions over two cycles (Fig. 3b & Fig.S2),
consistent with previous studies [37–39]. High K+ selectivity is crucial 3.2.2. Regeneration of saturated zeolite
for producing pure NK liquid fertilizer without Na+, which can raise soil Fig. 4a and Fig. 5a show the K+ concentration profiles during the
salinity. regeneration of the zeolite columns (chabazite and clinoptilolite) at high
3-
Fig. 3d-e shows the profiles of NH+
4 , PO4 , and pH in the effluent after flow rates with full recirculation and at low flow rates in single-pass
zeolite column treatment over time. As shown in Fig. 3c, a pH break­ mode, respectively. The regeneration of the saturated zeolite column
through was observed over time as proton (H+), the primary was conducted using 2.5 M sulfuric acid with a low volume of regen­
exchangeable ions, were replaced by K+. The zeolite column was eration solution (6-bed volume). Both full circulation and single-pass
initially preconditioned with concentrated sulfuric acid, leading to modes of the regeneration solution were tested to evaluate how the
proton release and a decrease in the effluent pH, as shown in Fig. 3c. high concentration of proton (H+) outcompete K+ selectivity during
Previous studies have similarly observed the release of exchangeable desorption, leading to a high concentration of K+ ions in the regenera­
ions like Na+ into the effluent when using sodium-form zeolite (Z-Na+) tion solution. As shown in Fig. 4a, and Fig. 5a, desorption of K+ was
[11,40]. Fig. 3d shows that both zeolite column facilitates the uptake of rapid for both zeolite columns, occurring within 20 min using both
residual ammonium after HFMC. The pH reduction in the effluent en­ modes of regeneration. In single-pass mode, the peak K+ concentrations
hances ammonium uptake by converting ammonia to ammonium, as the were similar in both zeolites (51,800 vs 50,100 mg/L). However, with
initial pH for ammonia recovery in the HFMC was 11. However, the full circulation, chabazite showed significantly higher peak concentra­
ammonium desorption occurred after the breakthrough due to high K+ tions (50,000 vs 30,000 mg/L) compared to clinoptilolite, as illustrated
concentration in the feed at the end (C/Co > 1) (Fig. 3d). Zeolite in Fig. 4a and Fig. 5a, respectively. Over time, the amount released K+
selectivity for K+ or NH+ 4 depends mainly on their relative concentra­ decreased due to significant desorption with the excess protons as
tions, with a higher concentration of one ion suppressing the influence of counter ions in both modes (Fig. 4a, and Fig. 5a). With full circulation,
the other [35,38,41]. Phosphate ions also showed a breakthrough over the K+ concentration in the regeneration solution was stabilized over
time, but removal was less than 40 % initially (Fig. 3e). Phosphate time as it was diluted and mixed during recirculation (Fig. 4a). In single-
reduction may occurred through precipitation and uptake by extra­ pass mode, the K+ concentration in the chabazite column declined more
framework cations (Mg2+ and Ca2+) in the zeolite, consistent with rapidly after reaching its peak compared to the clinoptilolite column

8
M.W. Atsbha et al. Separation and Puri cation Technology 359 (2025) 130598

Table 3 mg/L) compared to full recirculation (586 and 891 mg/L) for chabazite
The chemical composition of the accumulated regeneration solution at a high and clinoptilolite columns, respectively (Table 3). Ammonium was also
flow rate with full circulation and a low flow rate with one pass-through. desorbed in small amounts (Fig. 5c), with higher concentrations accu­
Chemicals (mg/L) At a high flow rate (30 mL/ At a low flow rate (5 mL/ mulated in chabazite (223.76 mg/L) than in clinoptilolite (169.51 mg/
min) min) L), consistent with its uptake trend (Fig. 3c) (Table 3). The desorption
with full circulation One Pass through trend and accumulation of other ions (Fig. 5c) and TOC in the regen­
​ Chabizite Clinoptilolite Chabizite Clinoptilolite eration solution were negligible, and the presence of cations such as
K+ 14,000 8140 10,600 11,100 Ca2+ and Mg2+ was also insignificant (Table 3). The Ca2+ and Mg2+ do
Na+ 891 468 586 238
not affect the generated fertilizer and can serve as supplements. The
NH+ 4 37.8 64.1 224 170
PO3- 15.4 14.2 14.9 24.1 final average accumulated K+ concentrations were 14,000 mg/L for
4
Mg 2+
68.7 84.2 21.8 35.8 chabazite and 11100 mg/L for clinoptilolite, while the Na+ concentra­
Ca2+ 223 485 126 218 tions were 891 mg/L and 238 mg/L, respectively (Table 3). The K+ to
Cl- 27.8 90 8.48 16.0 Na+ ratio in the regeneration solution was about 17 %, indicating low
TOC nd nd nd nd
interference, as illustrated in the regeneration trends of both ions.
Moreover, the low interference of Na+ in the regeneration solution was
(Fig. 5a). A faster desorption rate is critical for minimizing the volume of further minimized through dilution when added to the acid-stripping
the regeneration solution and achieving a higher K+ concentration. The solution. The Na+ concentration was reduced by half, as shown in
desorption rate can be influenced by the contact time and the zeolite’s Table 2, compared to its initial concentration in the regeneration solu­
affinity and kinetics for protons. Therefore, single-pass mode is recom­ tion (Table 3), thereby minimizing its impact on the integrated process.
mended for clinoptilolite regeneration, as full recirculation leads to Overall, preconditioning and regenerating the saturated zeolite with
lower K+ concentrations due to increased K+ re-adsorption (Table 3). In acid serves multiple purposes in this integrated process: (1) It neutralizes
contrast, chabazite showed higher K+ concentrations with full recircu­ the high pH effluent from the HFMC, (2) It facilitates further ammonium
lation, likely due to its greater K+ selectivity, and lower volume, as recovery in the IX column. Additionally, lowering the pH creates
indicated in Table 3. Overall, K+ regeneration remained stable across optimal conditions for the subsequent process of phosphate precipita­
two cycles for both zeolite columns, with similar final accumulated K+ tion at neutral pH, saving the need for additional acid for pH adjustment,
concentrations observed (Fig. 4a), demonstrating the process’s appli­ as discussed in section 3.3.
cability for multiple cycles.
Fig. 4b and Fig. 5b show the Na+ profile overtime during the acid
3.3. Precipitation of phosphate
regeneration of the zeolite columns with full recirculation and single-
pass mode, respectively. During regeneration, the release of Na+ fol­
Fig. 6 shows the effect of iron (Fe(II) and Fe(III)) and magnesium
lowed a similar trend to K+ in both columns but with a smaller mass
(Mg) doses on phosphate removal efficiency via precipitation at
released, reflecting its lower uptake in the zeolite columns (Fig. 4b, and
different pH levels. The experiment followed the ion exchange stage
Fig. 5b). The released Na+ was higher in the chabazite column than in
using effluent from a zeolite column, testing phosphate precipitation at
clinoptilolite using both modes of regeneration (Fig. 4b, and Fig. 5b),
various pH values and metal doses. Using Fe(II), phosphate removal
attributed to the slightly higher Na+ uptake in chabazite as shown in the
efficiency was above 90 % at pH 7 but dropped to less than 20 % at pH
breakthrough curve (Fig. 3b). The accumulated Na+ concentrations in
10, highlighting the significant impact of pH on phosphate removal with
the regeneration solution were lower in single pass mode (238 and 468
Fe(II). At higher pH levels, the abundance of carbonate ions increases

Fig. 6. Precipitation of phosphate as Fe-phosphate and MPP using different metal salts and pH range.

9
M.W. Atsbha et al. Separation and Puri cation Technology 359 (2025) 130598

Fig. 7. SEM and EDS results for the precipitated product: (a) Ferrous dose at pH 7; (b) Ferrous dose at pH 10; (c) magnesium dose at pH 10; (d) Ferric dose at pH 7;
and (e) Ferric dose at pH 10.

10
M.W. Atsbha et al. Separation and Puri cation Technology 359 (2025) 130598

due to carbonate speciation kinetics, competing with phosphate for iron CRediT authorship contribution statement
and thereby hindering phosphate precipitation [16,43]. The Fe(III) dose
at pH 7 removed 88 % of phosphate while at pH 10 was 42 %, which Musie Welldegerima Atsbha: Writing – review & editing, Writing –
follows the above-mentioned trend with higher OH– or carbonate original draft, Methodology, Investigation, Data curation, Conceptuali­
interference with phosphate. Neutral pH was optimal for precipitating zation, Visualization. Oded Nir: Conceptualization, Methodology,
phosphate as iron phosphate [16,44–46]. In this integrated approach, Writing – review & editing, Supervision, Project administration, Fund­
the treated effluent’s pH was neutral after ion exchange (Fig. 3c), ing acquisition. Treavor H. Boyer: Conceptualization, Methodology,
facilitating iron phosphate precipitation without acid addition. Simi­ Resources, Writing – review & editing, Supervision, Project adminis­
larly, Mg2+ doses achieved over 90 % phosphate removal, consistent tration, Funding acquisition.
with previous studies on magnesium-potassium phosphate precipitation
[47].
Declaration of competing interest
Fig. 7 shows the SEM and EDS analysis of precipitated solids from Fe
(II), Fe(III), and Mg doses on phosphate precipitation at pH 7 and pH 10.
The authors declare that they have no known competing financial
As shown in Fig. 7a&7b, the phosphate weight percentage in the solid
interests or personal relationships that could have appeared to influence
analysis for the Fe(II) dose was 10.5 % at pH 7 and 0.2 % at pH 10. This
the work reported in this paper.
significant difference aligns with the phosphate removal trends shown in
Fig. 6, where removal was higher at pH 7 using the Fe(II) dose. The
carbon weight percentage was negligible at pH 7 but increased to 13.2 % Acknowledgment
at pH 10, indicating the co-precipitation of carbonate with iron to form
complex iron carbonate precipitates [22,48]. The weight percentage This research was supported by the Israel-US Collaborative Water-
composition of the precipitated product at pH 7 (26 % Fe, 11 % P, 47 % Energy Research Center (Israel-US CoWERC) via the Binational Indus­
O) closely matches that of pure vivianite (33 % Fe, 12 % P, 51 % O, 3 % trial Research and Development Foundation (BIRD) Energy Center (No.
H). The slight differences indicate impurities in the product, which are EC-15). This work was also supported by the Science and Technologies
expected due to the complexity of the urine solution. The weight for Phosphorus Sustainability (STEPS) Center, a National Science
composition for the Mg dose was 3.7 % P, 11.4 % K, 5.2 % Mg, and 40.4 Foundation Science and Technology Center (NSF STC, CBET-2019435).
% O, indicating the expected MgKPO4 (MPP) precipitation (Fig. 7c). The author thanks the Goldman Sonnenfeldt School of Sustainability
Fig. 7d &7e show that the precipitated solid using the Fe(III) dose had a and Climate Change, Ben-Gurion University of the Negev for funding the
higher phosphate presence at pH 7 (7.5 wt%) compared to pH 10 (1.8 wt exchange program at ASU.
%). This is likely due to the high OH– and carbonate ion interference at The author also thanks Michael Serpa and Prof. Paul Westerhoff from
higher pH, which favors precipitation with iron [16]. Silicon (Si) and ASU for their support with the IC analysis of samples for this research.
aluminum (Al) were observed in all precipitated solids, possibly due to We acknowledge the use of facilities within the Eyring Materials Center
their release from the zeolite component. The presence of pharmaceu­ at Arizona State University supported in part by NNCI-ECCS-2025490.
ticals in the precipitated product is expected to be negligible, as
demonstrated in previous studies [49,50]. Testing for these compounds Appendix A. Supplementary data
was beyond the scope of this study. Overall, the experimental results and
SEM-EDS analysis indicate that iron doses (Fe(II) or Fe (III)) favor Supplementary data to this article can be found online at https://doi.
phosphate precipitation at neutral pH. org/10.1016/j.seppur.2024.130598.

4. Conclusion Data availability

This study demonstrates an integrated circular process for recovering Data will be made available on request.
nitrogen and potassium (NK) as liquid fertilizer and phosphate minerals
from hydrolyzed urine using a hollow fiber membrane contactor References
(HFMC), ion exchange, and precipitation. Ammonia recovery using
HFMC was consistent across three cycles, achieving over 90 % removal. [1] L. Crane, A. Merck, S. Delanthamajalu, K. Grieger, A. Marshall, T.H. Boyer,
Benchmarks for urine volume generation and phosphorus mass recovery in
Reusing the acid regeneration solution as the HFMC stripping solution commercial and institutional buildings, Water Res. x. 23 (2024) 100227, https://
sustained NK fertilizer quality. This study addresses the often- doi.org/10.1016/j.wroa.2024.100227.
overlooked topic of K+ ion exchange and regeneration using zeolite [2] N.S. Jagtap, T.H. Boyer, Urine collection in a multi-story building and
opportunities for onsite recovery of nutrients and non-potable water, J. Environ.
columns under various operational modes. Potassium recovery using Chem. Eng. 8 (2020) 103964, https://doi.org/10.1016/j.jece.2020.103964.
chabazite and clinoptilolite was efficient and consistent in the column [3] T.A. Larsen, H. Gruendl, C. Binz, The potential contribution of urine source
study, with better performance at lower flow rates, enhancing ion ex­ separation to the SDG agenda – A review of the progress so far and future
development options†, Environ. Sci. Water Res. Technol. 7 (2021) 1161–1176,
change dynamics. Regenerating K+-saturated zeolite columns with 2.5 https://doi.org/10.1039/d0ew01064b.
M H2SO4 was rapid and efficient, completing within 20 min and [4] L. Crane, D. Saetta, T.H. Boyer, Acid dosing increases recoverable phosphorus
remaining stable over two cycles. The resulting NK liquid fertilizer during different occupancy conditions in full-scale urine diversion, System (2024),
https://doi.org/10.1021/acsestengg.4c00164.
contained high NH+ 4 and K concentrations, with minimal impurities,
+
[5] J. Jermakka, E. Thompson Brewster, S. Freguia, P. Ledezma, M. Kokko, Electro-
achieving an optimal NK ratio for plant growth (3.3 %-3.6 % N and 0.4 concentration of urine designed for separation of sodium from nitrogen, Sep. Purif.
%-0.7 % K by weight). Phosphate precipitation with iron achieved up to Technol. 276 (2021) 119275. https://doi.org/10.1016/j.seppur.2021.119275.
[6] Y.A. Alemayehu, S.L. Asfaw, T.A. Terfie, Nutrient recovery options from human
90 % removal, with SEM-EDS analysis confirming that the precipitate
urine: a choice for large scale application, Sustain. Prod. Consum. 24 (2020)
formed at neutral pH closely resembled vivianite, a potential resource 219–231, https://doi.org/10.1016/j.spc.2020.06.016.
for fertilizer or lithium iron phosphate battery production. This inte­ [7] U. Badeti, J. Jiang, S. Kumarasingham, A. Almuntashiri, N.K. Pathak, A. Chanan,
grated approach offers significant advantages: it recovers valuable nu­ S. Freguia, W.L. Ang, N. Ghaffour, H.K. Shon, S. Phuntsho, Source separation of
urine and treatment: Impact on energy consumption, greenhouse gas emissions,
trients, optimizes chemical use through a closed-loop system, and and decentralised wastewater treatment process, Desalination. 583 (2024) 117633,
enhances the value of phosphate by precipitating it as iron phosphate. https://doi.org/10.1016/j.desal.2024.117633.
This method promotes eco-efficiency by minimizing additional chemical [8] X. Wang, Y. Li, X. Wen, L. Liu, L. Zhang, M. Long, Cooperation of ferrous ions and
hydrated ferric oxide for advanced phosphate removal over a wide pH range:
loads. mechanism and kinetics, Water Res. 249 (2024) 120969, https://doi.org/10.1016/
j.watres.2023.120969.

11
M.W. Atsbha et al. Separation and Puri cation Technology 359 (2025) 130598

[9] X. You, A. Farran, D. Guaya, C. Valderrama, V. Soldatov, J.L. Cortina, Phosphate DEA solution, Ind. Eng. Chem. Res. 63 (2024) 9185–9202, https://doi.org/
removal from aqueous solutions using a hybrid fibrous exchanger containing 10.1021/acs.iecr.4c00525.
hydrated ferric oxide nanoparticles, J. Environ. Chem. Eng. 4 (2016) 388–397, [31] D. Saetta, C. Zheng, C. Leyva, T.H. Boyer, Impact of acetic acid addition on
https://doi.org/10.1016/j.jece.2015.11.032. nitrogen speciation and bacterial communities during urine collection and storage,
[10] N. Jagtap, T.H. Boyer, Integrated decentralized treatment for improved N and K Sci. Total Environ. 745 (2020) 141010, https://doi.org/10.1016/j.
recovery from urine, J. Sustain. Water Built Environ. 6 (2020) 1–9, https://doi. scitotenv.2020.141010.
org/10.1061/jswbay.0000899. [32] A.N. Kofina, K.D. Demadis, P.G. Koutsoukos, The effect of citrate and
[11] U. Regmi, T.H. Boyer, Ammonium and potassium removal from undiluted and phosphocitrate on struvite spontaneous precipitation, Cryst. Growth Des. 7 (2007)
diluted hydrolyzed urine using natural zeolites, Chemosphere. 268 (2021) 128849, 2705–2712, https://doi.org/10.1021/cg0603927.
https://doi.org/10.1016/j.chemosphere.2020.128849. [33] N. Jagtap, T.H. Boyer, Integrated, multi-process approach to total nutrient recovery
[12] S. Guida, C. Potter, B. Jefferson, A. Soares, Preparation and evaluation of zeolites from stored urine, Environ. Sci. Water Res. Technol. 4 (2018) 1639–1650, https://
for ammonium removal from municipal wastewater through ion exchange process, doi.org/10.1039/c8ew00004b.
Sci. Rep. 10 (2020) 1–11, https://doi.org/10.1038/s41598-020-69348-6. [34] J.S. Cyrus, G.B. Reddy, Sorption and desorption of ammonium by zeolite: Batch
[13] J.A. Wilsenach, C.A.H. Schuurbiers, M.C.M. van Loosdrecht, Phosphate and and column studies, J. Environ. Sci. Heal. - Part A Toxic/Hazardous Subst, Environ.
potassium recovery from source separated urine through struvite precipitation, Eng. 46 (2011) 408–414, https://doi.org/10.1080/02773813.2010.542398.
Water Res. 41 (2007) 458–466, https://doi.org/10.1016/j.watres.2006.10.014. [35] Q. Deng, B.R. Dhar, E. Elbeshbishy, H.S. Lee, Ammonium nitrogen removal from
[14] M. Ronteltap, M. Maurer, R. Hausherr, W. Gujer, Struvite precipitation from urine - the permeates of anaerobic membrane bioreactors: economic regeneration of
Influencing factors on particle size, Water Res. 44 (2010) 2038–2046, https://doi. exhausted zeolite, Environ. Technol. (united Kingdom) 35 (2014) 2008–2017,
org/10.1016/j.watres.2009.12.015. https://doi.org/10.1080/09593330.2014.889759.
[15] P. Bucholtz, M. Steele, V. Tripathi, C. Graham, L. Crane, T.H. Boyer, Solar [36] E. Gagliano, M. Sgroi, P.P. Falciglia, C. Belviso, F. Cavalcante, A. Lettino, F.G.
distillation of human urine to recover non-potable water and metal phosphate A. Vagliasindi, P. Roccaro, Removal of ammonium from wastewater by zeolite
mineral, Water Sci. Technol. 88 (2023) 486–501, https://doi.org/10.2166/ synthetized from volcanic ash: batch and column tests, J. Environ. Chem. Eng. 10
wst.2023.218. (2022) 107539, https://doi.org/10.1016/j.jece.2022.107539.
[16] C. Simbeye, C. Courtney, P. Simha, N. Fischer, D.G. Randall, Human urine: a novel [37] M. Sprynskyy, M. Lebedynets, A.P. Terzyk, P. Kowalczyk, J. Namie{\acute{s}}nik,
source of phosphorus for vivianite production, Sci. Total Environ. 892 (2023), B. Buszewski, Ammonium sorption from aqueous solutions by the natural zeolite
https://doi.org/10.1016/j.scitotenv.2023.164517. Transcarpathian clinoptilolite studied under dynamic conditions, J. Colloid
[17] H. Ray, F. Perreault, T.H. Boyer, Ammonia recovery and fouling mitigation of Interface Sci. 284 (2005) 408–415, https://doi.org/10.1016/j.jcis.2004.10.058.
hydrolyzed human urine treated by nanofiltration and reverse osmosis, Environ. [38] A. Damodara Kannan, P. Parameswaran, Ammonia adsorption and recovery from
Sci. Water Res. Technol. 8 (2022) 429–442, https://doi.org/10.1039/ swine wastewater permeate using naturally occurring clinoptilolite, J. Water
d1ew00601k. Process Eng. 43 (2021) 102234, https://doi.org/10.1016/j.jwpe.2021.102234.
[18] F. Volpin, L. Chekli, S. Phuntsho, N. Ghaffour, J.S. Vrouwenvelder, H.K. Shon, [39] Z.G. Chartrand, R.M. Narbaitz, M. Sartaj, J. Downey, Ammonia-Ca-K competitive
Optimisation of a forward osmosis and membrane distillation hybrid system for the ion-exchange on zeolites in mining wastewater treatment: batch regeneration and
treatment of source-separated urine, Sep. Purif. Technol. 212 (2019) 368–375, column performance, J. Sustain. Min. 19 (2020) 58–71. 10.46873/2300-3960.1
https://doi.org/10.1016/j.seppur.2018.11.003. 006.
[19] M.W. Atsbha, R. Liu, O. Nir, Circular hybrid membrane process treating high- [40] M. Maghfiroh, N. Park, H. Chang, H. Lim, W. Kim, Ammonium removal and
salinity ammonium-rich pharmaceutical wastewater, Chem. Eng. J. 497 (2024) recovery from greywater using the combination of ion exchange and air stripping:
154690, https://doi.org/10.1016/j.cej.2024.154690. the utilization of NaOH for the regeneration of natural zeolites, J. Water Process
[20] K. Yang, M. Qin, Understanding ammonia and water transport in direct contact Eng. 52 (2023) 103581, https://doi.org/10.1016/j.jwpe.2023.103581.
membrane distillation toward selective ammonia recovery, ACS ES T Eng. (2023), [41] C.H. Liu, K.V. Lo, Ammonia removal from compost leachate using zeolite. III.
https://doi.org/10.1021/acsestengg.3c00623. Regeneration of zeolite columns, J. Environ. Sci. Heal. - Part A Toxic/Hazardous
[21] M. Welldegerima, L. Farkash, G. Bartal, O. Nir, Extracting liquid fertilizer from Subst, Environ. Eng. 36 (2001) 1825–1843, https://doi.org/10.1081/ESE-
ammonia-laden pharmaceutical wastewater using Hollow fiber membrane 100107432.
contactors, J. Clean. Prod. 418 (2023) 138183, https://doi.org/10.1016/j. [42] E.L. Uzunova, H. Mikosch, Adsorption of phosphates and phosphoric acid in zeolite
jclepro.2023.138183. clinoptilolite: electronic structure study, Microporous Mesoporous Mater. 232
[22] D. Yang, J. Zhang, M. Xie, X. Tong, T. Jiang, W. Yu, D. Qu, An integrated hollow (2016) 119–125, https://doi.org/10.1016/j.micromeso.2016.06.019.
fiber membrane contactor and chemical precipitation to recover N, P and K from [43] J. Liu, X. Cheng, X. Qi, N. Li, J. Tian, B. Qiu, K. Xu, D. Qu, Recovery of phosphate
human urine wastewater, J. Environ. Chem. Eng. 10 (2022) 107844, https://doi. from aqueous solutions via vivianite crystallization: thermodynamics and influence
org/10.1016/j.jece.2022.107844. of pH, Chem. Eng. J. 349 (2018) 37–46, https://doi.org/10.1016/j.
[23] M.M. Damtie, F. Volpin, M. Yao, L.D. Tijing, R.H. Hailemariain, T. Bao, K.D. Park, cej.2018.05.064.
H.K. Shon, J.S. Choi, Ammonia recovery from human urine as liquid fertilizers in [44] F. Yang, S. Chen, S. Liang, M. Wen, J. Quan, X. Shi, S. Yuan, H. Duan, W. Yu,
hollow fiber membrane contactor: effects of permeate chemistry, Environ. Eng. L. Huang, J. Yang, Resources, Conservation & Recycling efficient phosphorus
Res. 26 (2021) 1–9, https://doi.org/10.4491/eer.2019.523. recovery from simulated phosphoric acid wastewater using a tailored dual-module
[24] J. Zhang, M. Xie, X. Tong, S. Liu, D. Qu, S. Xiao, Recovery of ammonium nitrogen flow electrode capacitive deionization (FCDI) strategy, Resour. Conserv. Recycl.
from human urine by an open-loop hollow fiber membrane contactor, Sep. Purif. 211 (2024) 107867, https://doi.org/10.1016/j.resconrec.2024.107867.
Technol. 239 (2020) 116579, https://doi.org/10.1016/j.seppur.2020.116579. [45] J. Cao, Y. Wu, J. Zhao, S. Jin, M. Aleem, Q. Zhang, F. Fang, Z. Xue, J. Luo,
[25] J. Zhang, M. Xie, D. Yang, X. Tong, D. Qu, L. Feng, L. Zhang, The design of multi- Phosphorus recovery as vivianite from waste activated sludge via optimizing iron
stage open-loop hollow fiber membrane contactor and its application in ammonia source and pH value during anaerobic fermentation, Bioresour. Technol. 293
capture from hydrolyzed human urine, Water Res. 207 (2021) 117811, https://doi. (2019) 122088, https://doi.org/10.1016/j.biortech.2019.122088.
org/10.1016/j.watres.2021.117811. [46] R. Priambodo, Y. Shih, Y. Huang, Phosphorus recovery as ferrous phosphate
[26] Q. Liu, W. Sun, Q. Zeng, H. Zhang, C. Wu, E. Lichtfouse, H. Liu, Integrated (vivianite) from wastewater produced in manufacture of thin film transistor-liquid
processes for simultaneous nitrogen, phosphorus, and potassium recovery from crystal displays (TFT-LCD) by a fluidized bed crystallizer (FBC), (2017)
urine: a review, J. Water Process Eng. 59 (2024) 104975, https://doi.org/10.1016/ 40819–40828. https://doi.org/10.1039/C7RA06308C.
j.jwpe.2024.104975. [47] K. Xu, J. Li, M. Zheng, C. Zhang, T. Xie, C. Wang, The precipitation of magnesium
[27] H. Wu, X. Foster, H. KAZEMIAN, C. Vaneeckhaute, N, P, K Recovery from potassium phosphate hexahydrate for P and K recovery from synthetic urine, Water
Hydrolysed Urine by Na-Chabazite Adsorption Integrated with Ammonia Stripping Res. 80 (2015) 71–79, https://doi.org/10.1016/j.watres.2015.05.026.
and (K-)Struvite Precipitation, SSRN Electron. J. 857 (2022) 159277. https://doi. [48] R. Goedhart, S. Müller, M.C.M. van Loosdrecht, D. van Halem, Vivianite
org/10.2139/ssrn.4170677. precipitation for iron recovery from anaerobic groundwater, Water Res. 217
[28] W.A. Tarpeh, I. Wald, M. Wiprächtiger, K.L. Nelson, Effects of operating and design (2022) 118345, https://doi.org/10.1016/j.watres.2022.118345.
parameters on ion exchange columns for nutrient recovery from urine, Environ. [49] M. Ronteltap, M. Maurer, W. Gujer, The behaviour of pharmaceuticals and heavy
Sci. Water Res. Technol. 4 (2018) 828–838, https://doi.org/10.1039/c7ew00478h. metals during struvite precipitation in urine, Water Res. 41 (2007) 1859–1868,
[29] 3M Liqui-CelTM, 3MTM Liqui-CelTM MM-1.7x5.5 Series Membrane Contactor, https://doi.org/10.1016/j.watres.2007.01.026.
(2021). https://multimedia.3m.com/mws/media/1412492O/3m-liqui-cel-mm-1- [50] S.P. Wei, F. van Rossum, G.J. van de Pol, M.K.H. Winkler, Recovery of phosphorus
7x5-5-series-membrane-contactor.pdf (accessed March 11, 2024). and nitrogen from human urine by struvite precipitation, air stripping and acid
[30] G. Pantoleontos, D. Koutsonikolas, A.G. Asimakopoulou, S. Lorentzou, scrubbing: a pilot study, Chemosphere. 212 (2018) 1030–1037, https://doi.org/
G. Karagiannakis, Modeling, simulation, and membrane wetting estimation in gas- 10.1016/j.chemosphere.2018.08.154.
liquid contacting processes including shell-side reaction: biogas upgrading using

12