Phosphate Recovery at “A Campingflight to Lowlands Paradise”: Organic Micropollutant Uptake and Environmental Risk Assessment

Abstract

Despite increasing interest in struvite as a renewable phosphate source and fertilizer, research into real-world cases of organic micropollutant contamination in struvite is limited, with no studies addressing the inclusion of illicit substances. Urine is an ideal matrix to precipitate struvite from and study its organic contamination, as it contains the majority of organic contaminants as well as nutrients excreted by humans. As such, we devised a worst-case scenario in terms of organic contamination by collecting urine at the Dutch festival “A Campingflight to Lowlands Paradise” and precipitating struvite on-site in three batches. This festival setting provided a highly contaminated urine source, offering extreme conditions to evaluate to what extent such contamination translates to struvite and if that struvite would be safe to use as fertilizer. Surveys on consumed pharmaceuticals and illicit drugs by participants guided subsequent analysis of urine and struvite samples, which was performed using liquid chromatography–tandem mass spectrometry (LC–MS/MS). Concentrations of organic contaminants in the urine corresponded well with the survey results and were found in a broad range from <1 to over 34,000 ng mL⁻¹. Concentrations found in the struvite precipitates generally showed a correlation proportional to contaminant K _oc values with those found in their respective urine source and were predominantly in the range of 1–100 ng g⁻¹, with an outlier at 1081 ng g⁻¹. Based on these numbers, the environmental risk associated with using the precipitated struvite as renewable phosphate fertilizer was classified as insignificant in all cases. After 100 years of hypothetical Lowlands-struvite application as fertilizer, only paracetamol and MDMA would be classified as having low risk. All other analytes remain an insignificant environmental risk, showing struvite to be a safe, renewable phosphate source in terms of organic contamination.

Keywords

Nutrient recycling - Struvite recovery - Target analysis - Organic micropollutants - Environmental risk assessment

Significance

• In this study, a worst-case scenario for struvite precipitation in terms of organic contamination is investigated by on-site struvite recovery from urine highly contaminated with pharmaceuticals and illicit drugs at “A Campingflight to Lowlands Paradise” music festival.

• Organic contaminants in festival urine are minimally retained in struvite.

• Substance K _oc values correlate well with the incorporation into struvite.

• Environmental risk assessment shows our festival-sourced struvite to be safe for use.

1

Introduction

Phosphorus (P) is vital for life, being essential for DNA, cell membranes, and ATP [1], [2]. However, it is scarce in the biosphere, often limiting crop growth, and has been dubbed “life’s bottleneck.”[3], [4] Despite the abundance provided by phosphate rock (PR) discovered in the 19th century, PR is finite and unevenly distributed, with most reserves in Morocco and the Western Sahara [5], [6]. Europe has minimal reserves, leading the European Commission to classify PR as well as elemental P as critical raw materials [7].

About 90% of the 21 Mt P mined in 2019 was destined for fertilizer, yet only 16% reached consumed nutrition due to agricultural runoff and losses along the agri-food chain [8] [9] [10]. To make matters worse, less than 10% of P in urban wastewater (WW) is recovered [5]. These glaring inefficiencies contribute to environmental issues like eutrophication [11]. To achieve sustainable P management, it is crucial to recover and recycle P from waste streams, reducing our dependency on finite PR [2]. Urban WW, which loses about 15% of P imported into Europe [12], presents a significant opportunity for P recovery and recycling thanks to existing infrastructure [13].

The most commonly employed technology for the recovery of P from wastewater is struvite (MgNH₄PO₄ ∙ 6 H₂O) precipitation [14], which has been hailed as a slow-release fertilizer [2], [15]. Indeed, while only sparingly soluble, studies have shown that its agronomic efficiency is on par with that of conventional mineral fertilizers [15], [16]. Currently, about 80 struvite production facilities employing 19 different—though similar—technologies are active at wastewater treatment plants (WWTPs) across the world [17]. In the EU, about 11,000 tons of struvite is produced annually [18].

Struvite precipitates according to Eq. ([1]). A pH between 9.0 and 10.7 is considered optimal for effective precipitation, although some industrial processes precipitate struvite at pH 8.0–8.5 [18] [19] [20] [21] [22]. Struvite precipitation ideally occurs at P concentrations of 100 mg L⁻¹, with a minimum of about 50–60 mg L⁻¹ P required [23], [24]. These concentrations are seldomly found in urban WW, and struvite precipitation is therefore oftentimes preceded by Enhanced Biological Phosphorus Removal (EBPR), a technique to increase P concentrations in the liquid phase by utilizing the phosphate metabolism of certain microorganisms [2]. Other important parameters are temperature, Mg:P ratio, and the presence of foreign ions, although none as influential as pH, which governs the speciation of PO₄-P [25] [26] [27]. A major added benefit of struvite precipitation is the simultaneous recovery of N in the form of NH₄ ⁺, also a macronutrient.

To encourage and facilitate the recycling of recovered materials, it is important for them to be as safe as possible to improve market confidence and compatibility [2], [16]. Hence, it is imperative to further our knowledge of the quality of recovered struvite from contaminated sources and map potential environmental risks. However, struvite legislation has predominantly concerned itself with inorganic contaminants thus far, and so literature has followed suit [16], [18], [28]. As such, there has been little research into real-world cases concerning organic micropollutant contamination of struvite and accompanying risks [29], [30], and none including illicit substances.

As urine contributes about 80% of N, 70% of K, and 50% of P found in WW yet only constitutes less than 1% of its volume, struvite precipitation is possible without the need for pre-precipitation processes [31], [32]. Furthermore, urine contributes 64% of organic contaminants in WW [33], [34], making urine the perfect feedstock candidate to investigate struvite sourced from a highly contaminated feedstock. To shed light on the fate of organic contaminants present when struvite is precipitated, a worst-case scenario was envisioned: collecting urine at a music festival—highly contaminated with pharmaceuticals and illicit drugs—to precipitate struvite from and monitor the behavior of said contaminants throughout the process.

And so, we set out to realize this worst-case scenario by attending “A Campingflight to Lowlands Paradise” festival in the Netherlands in 2019. Situated in the festival “Science Lab,” we collected urine from willing festival visitors and precipitated struvite on-site for three days, as well as educated all visitors about the enormous potential of phosphate recovery and recycling. Pharmaceutical- and drug intake of participants was monitored by anonymous surveys, the data of which were subsequently useful in the analysis of urine and struvite samples.

2

Materials and methods

2.1

Urine collection and struvite precipitation

The experiments were conducted at the Lowlands Science Lab, a designated area for scientific experiments on the festival grounds. Operating hours for the Lowlands Science Lab were from 12:00 until 20:00 daily. During this time, we operated a toilet unit with two urinals and one toilet for men and two toilets for women. Here, undiluted urine was collected from the urinals by pumping it directly into a plastic container of 1 m³. While women were allowed to make use of our facilities during the festival, they could unfortunately not be included in the experiment due to us not having source-separated toilets available to divert the urine. Hence, all collected urine was from male participants using our urinals only.

Just over a thousand people participated in our study during the festival, divided about equally over the three days (334, 357, and 336 participants; [Table 1]). This also translated to roughly equal amounts of urine collected each day, with an average amount of 155 L. All volumes have been rounded to 5 L increments due to the large volumes involved.

Table 1 General data on number of participants and urine collected.

Day	1	2	3
a Inoculation volume was left in the urine collection tank for the next day.
Participants (male only)	334	357	336
Urine in collection tank (L)	200	180	175
Inoculation volume (L)a	45	20	25
Urine collected (L)	155	160	150

At 20:00, the collected urine was pumped into a 250 L precipitation reactor with a conical bottom and draining point at the bottom of the cone. All collected urine was used for struvite precipitation each day, except the inoculation volume for the next day ([Table 1]). The purpose of the inoculation volume was to promote urea hydrolysis and ensure that enough NH₄ ⁺ was present to allow for struvite precipitation [35]. Due to uncertainty regarding the daily urine collection volume, a high inoculation volume was used on Day 1, which was halved on Days 2 and 3 to maintain an approximate 15% inoculation volume. This created a carryover effect between days, impacting the correlation between survey responses and measured urine concentrations. Since participants did not specify consumption amounts, this correlation was always an approximation. However, the study's main focus was on contaminant levels in struvite, not the survey to analyte correlation, so no correction for the carryover effect was made.

After urea hydrolysis, phosphate is the limiting factor in struvite precipitation. On-site determination of the PO₄-P concentration in the collected urine was not performed. As such, a concentration of 200 mg L⁻¹ PO₄-P was assumed, in line with numerous findings and estimations made in the literature [29], [35] [36] [37]. The pH of the partially hydrolyzed urine was 8.4–8.7 prior to struvite precipitation. A NaOH aqueous solution (10 M) was used to increase the pH to 9.0, which was monitored using a standard pH meter. Based on the amount of collected urine in the precipitation reactor, an aqueous MgCl₂ solution (1400 ppm) was added to achieve a Mg:P ratio of 1. The mixture was then stirred at 150 rpm for 30 minutes, after which the stirring was stopped overnight to promote crystal growth. See Figure S1 for a picture of the collection tanks and struvite precipitation setup.

2.2

Sample handling

Samples were collected outside the operating hours of the Lowlands Science Lab. Pre-precipitation urine samples of 100 mL each were collected in triplicate from the storage tank as soon as the Science Lab closed for the day at approximately 20:00 h. Samples of both post-precipitation urine and produced struvite were collected each morning from the struvite reactor at approximately 11:00 h. All urine samples were stored at –20 °C immediately after collection and only defrosted prior to analysis. Struvite samples were taken by draining part of the reactor contents into 1 L Schott bottles. Due to the scale of the experiment, the total amount of struvite produced was not collected or recorded. However, visual assessment of precipitation yields indicated increasing struvite yields per day, with Day 1 showing significantly less precipitate than Days 2 and 3. In the week after the festival, the samples of three different batches of struvite were filtrated using a Büchner funnel equipped with standard filter paper (Whatman), rinsed with demineralized water, and left to dry on air. Samples of several grams of struvite were collected for each day (see Figure S2), enough to allow subsequent analyses.

2.3

Anonymous surveys

Every person willing to participate in our experiment was asked to fill in an online anonymous survey on their intake of pharmaceutical substances and illicit drugs over the last three days. For each substance, they were given an option to highlight whether intake was 0–24, 24–48 and/or 48–72 h ago. Substances taken in multiple timeframes were counted as separate hits in all cases. No inquiries were made as to the amount of drugs consumed. After each day, all collected survey data was compounded in Microsoft Excel for subsequent analysis.

The survey results were also guiding as to which analytes were to be selected. Before the experiment, a list of commonly used substances and pharmaceuticals was already drafted. This included cocaine, amphetamine, MDMA, cannabis, ibuprofen and paracetamol. Respondents were able to manually add substances not present in the provided standard list. After the festival, the list of target analytes was then expanded with a number of important metabolites, as well as substances showing up in the surveys. For illicit drugs, added analytes were methamphetamine, ketamine, mephedrone (4-MMC) and 3-MMC. Added pharmaceutical analytes included diclofenac, citalopram, atorvastatin, and tramadol. Note that not all substances that showed up in the survey were included in the analytes list selected for monitoring due to the unavailability of analytical standards.

2.4

Analyte selection

The analyte selection ([Table 2]) was based on commonly used illicit drugs, their metabolites (if any) and commonly used pharmaceuticals in the Netherlands. The selection was further fine-tuned based on the anonymous survey data.

Table 2 List of analytes and relevant physicochemical properties.

Illicit drug analytes	Formula	pK a	Log K ow	Log K oc	Description
Log K oc values of all analytes were estimated using EpiSuite KOCWIN (version 2.00). Values given are the mean of estimations based on the Molecular Connectivity Index (MCI) and log K ow values. See Excel S1 for the separate sets of values. a Predicted by Chemaxon. b Value assumed to be the same as that of 4-MMC due to literature values lacking. c Estimated using EpiSuite KowWIN.
Cocaine	C17H21NO4	8.61 [38]	2.2 [38]	2.65	Illicit stimulant
Benzoylecgonine	C16H19NO4	3.35 and 10.83 [38]	−1.3 [38]	2.00	Cocaine metabolite
Cocaethylene	C18H23NO4	8.77 [39] a	2.64 [39] a	2.90	Cocaine metabolite
Amphetamine	C9H13N	9.9 [38]	1.8 [38]	2.62	Illicit stimulant
Methamphetamine	C10H15N	10.4 [38]	2.2 [38]	2.71	Illicit stimulant
3,4-Methylenedioxy methamphetamine (MDMA)	C11H15NO2	10.14 [40]	1.86 [40]	2.22	Illicit stimulant
Ketamine	C13H16ClNO	7.5 [41]	3.1 [38]	2.95	Illicit anesthetic
norKetamine	C12H14ClNO	7.27a	2.91 [42] a	2.87	Major ketamine metabolite
DehydronorKetamine	C12H12ClNO	7.19a	2.44c	2.79	Minor ketamine metabolite
11-nor-9-carboxy-∆-9-tetrahydrocannabinol (THCCOOH)	C21H28O4	4.2 [40]	5.14 [40]	4.10	THC metabolite
Mephedrone (4-MMC)	C11H15NO	7.4 [38]	2.4 [38]	2.33	Synthetic cathinone
3-methylmethcathinone (3-MMC)	C11H15NO	7.4b	2.39c	2.33	Synthetic cathinone
Pharmaceutical analytes
Ibuprofen	C13H18O2	4.42 [43]	4.13 [43]	2.54	Nonsteroidal anti-inflammatory drug (NSAID)
Paracetamol	C8H9NO2	9.86 [38]	0.27 [38]	1.50	NSAID
Citalopram	C20H21FN2O	9.38 [44]	3.76 [45]	4.17	Antidepressant
Diclofenac	C14H11Cl2NO2	3.99 [43]	4.51 [43]	2.63	NSAID
Atorvastatin	C33H35FN2O5	4.5 [46]	5.39 [47] a	4.16	Statin (Cholesterol drug)
cis-tramadol	C16H25NO2	13.8 [40]	1.34 [40]	2.40	Opioid painkiller
O-desmethyltramadol	C15H23NO2	10.0 [38]	2.45 [48]	2.70	Tramadol metabolite

2.5

Chemical analysis of urine and struvite samples

Analytical methods for urine and struvite samples based on different sample preparation techniques and followed by liquid chromatography–tandem mass spectrometry (LC–MS/MS) determination were developed and validated. Information about analytical standards, chemicals, and reagents is reported in Supplementary Information (SI) (S3).

Two sample preparation procedures adapted from the literature were applied for urine analysis, depending on the expected concentration of target analytes and metabolites [49], [50]. First, urine samples (20 μL) were spiked with internal standard (IS) (2 ng of each labeled compound) and mixed with 460 μL of acetonitrile (ACN) for precipitation. After shaking and centrifugation (5 min, 10,000 rpm), the supernatant was collected and gently evaporated to dryness under nitrogen flow, reconstituted in 200 μL of a mixture of Milli-Q water:MeOH (90:10) and transferred into glass vials for LC–MS/MS analysis. For analytes with expected low concentrations (i.e., methamphetamine, DehydronorKetamine, 3-MMC, 4-MMC, cis-tramadol, O-desmethyltramadol, atorvastatin), and excreted as II-phase metabolites, another procedure based on hydrolysis followed by solid phase extraction (SPE) was applied. One mL of pooled urine sample was spiked with internal standard (2 ng of each labeled compound) and hydrolyzed with β-glucuronidase at 55 °C for 2 h (pH = 4.5–5, acetic acid/ammonium acetate buffer). Then, urine extracts were acidified to pH ~2 with HCl (9%) and extracted by SPE using Oasis® MCX cartridges (60 mg, 3 cc). Before the extraction, MCX cartridges were conditioned with 6 mL of MeOH, 3 mL of Milli-Q water, and 3 mL of Milli-Q water acidified to pH 2. After sample loading, MCX cartridges were vacuum-dried for 5 min and eluted with 1 mL of MeOH and 1 mL of a 2% ammonia solution in MeOH. Eluates were dried under a gentle nitrogen stream, reconstituted in 200 μL of a mixture of Milli-Q water:MeOH (90:10), centrifuged for 2 min at 2500 rpm and transferred into glass vials for LC–MS/MS analysis.

Struvite samples (0.25–0.5 g) were spiked with IS (2 ng of each labeled compound), dissolved in 0.7 mL of HCl (37%), and subsequently diluted in Milli-Q water (500 mL). After pH adjustment to 2–2.5, the aforementioned SPE procedure was used for extraction. MCX cartridges were vacuum-dried for 10 min after sample loading (500 mL) and eluted with 2 mL of MeOH and 2 mL of a 2% ammonia solution in MeOH. Eluates were evaporated to dryness, reconstituted in 200 μL of a mixture of Milli-Q water:MeOH (90:10), centrifuged for 2 min at 2500 rpm, and transferred into vials for LC–MS/MS analysis.

LC–MS/MS methods were adapted from previously published papers [51], [52], using an Agilent 1260 Series HPLC system C (Agilent Technologies, Santa Clara, CA, USA) coupled to a triple quadruple mass spectrometer TripleQuad 5500 ABSciex (Applied Biosystems, Concord, Ontario, Canada) equipped with a TurboIonSpray source operating both in positive and negative ionization mode. For compounds analyzed in positive mode, an Atlantis® T3 (100 × 2.1 mm; 3 μm) column from Waters (Milford, MA) and a dual eluent system consisting of (A) 0.1% formic acid in Milli-Q water and (B) ACN were used with the following elution gradient: 0 min (99% A), 15 min (100% B), 18 min (100% B), 24 min (99% A). The flow rate was 200 μL min⁻¹ and the injection volume was 2 μL. For analytes in negative mode, an XBridge C18 (100 × 2.1 mm; 3.5 μm) column from Waters Corp. (Milford, MA, USA) was selected. Mobile phases were (A) 0.025% triethylamine in Milli-Q water and (B) ACN. The 20-minute elution gradient was as follows: 0 min (95% A), 9 min (100% B), 14 min (100% B), and 20 min (95% A). The flow rate was 200 μL min⁻¹ and the injection volume was 2 μL. A specific LC method was developed for the analysis of isomers 3-MMC and 4-MMC, using a longer column Atlantis® T3 (150 × 2.1 mm; 3 μm) and a dual eluent system consisting of (A) 10 mM NH₄COOCH₃ in Milli-Q water and (B) ACN. More details about this LC–MS/MS method can be found in a previous paper [53].

Selected reaction monitoring (SRM) was chosen as the acquisition mode, selecting the two/three most abundant fragmentation products of the protonated pseudomolecular ions of each analyte and one fragmentation product of each IS. MS/MS parameters of each compound are reported in Table S1. Quantitation of analytes was performed using the isotopic dilution method and a 6-point calibration curve was made freshly before each analytical run. Recovery experiments in urine (n = 3) and struvite (n = 1) samples were performed by spiking 1 mL aliquots with 5 ng mL⁻¹ or 7 ng g⁻¹ of each analyte, respectively. Recovery results in urine were satisfactory, with recoveries between 89% (cocaine) and 111% (citalopram) and relative standard deviations lower than 12%. In the case of struvite, satisfactory recoveries were obtained for all compounds except atorvastatine (23%) and citalopram (156%), which was considered to correct the measured concentrations. Recoveries of all analytes are reported in Tables S2 and S3, respectively. Limits of quantitation (LOQ) in both matrices are shown in Table S4. More detailed information about the analytical method is provided in SI (S4–S7).

2.6

Environmental risk assessment

The environmental risk was determined based on the current Dutch legal annual application limit of 80 kg P₂O₅ ha⁻¹ [54]. Struvite’s legal lower P limit is 7% [18], or 16.03% P₂O₅, much lower than the theoretical value of 28.9%. This lower limit was taken in our calculations to provide a worst-case scenario in terms of contaminant deposition. Following this, the struvite application limit is 499 kg struvite ha⁻¹. Maximum contaminant concentrations in the soil were then determined based on Dutch fertilizer law, which states the top 20 cm of soil should be considered [55], see Excel S2 for the associated calculations.

Predicted no-effect concentration values in water (PNEC_water) were either taken from literature or estimated according to the Technical Guidance Document for Risk Assessment of the European Union [56]. Here, PNEC_water values are derived by dividing known toxicological endpoints by an appropriate assessment factor. This is done as it is assumed that organisms may react more sensitively to substances than the toxicological data suggests [57]. Various endpoints and assessment factors can be used and are given in [Table 3]. Many of the estimated PNEC_water values were derived previously in the literature according to [Table 3] and taken as such [58], [59].

Table 3 Assessment factors used for PNEC_water derivation, determined by the available toxicological data [56], [57].

Available data	Assessment factor
At least one short-term toxicological endpoint (lethal concentration or median effective concentration 50% (LC/EC50)) from each of three trophic levels (fish, Daphnia and algae)	1000
One long-term toxicological endpoint (EC10 or no observed effect concentration (NOEC)) for either fish or Daphnia	100
Two long-term toxicological endpoints from species of two trophic levels	50
Three long-term toxicological endpoints from species of all three trophic levels	10

As no ecotoxicological data could be procured for soil organisms, indicative PNEC_soil values were derived from PNEC_water and K _soil−water values using the equilibrium partitioning method, according to Eq. [(2)].

PNEC soil = K soil − water × PNEC water × 1000 RHO Soil

where K _soil−water is defined as Eq. [(3)].

K soil − water = F air , soil × K air − water + F water , soil + F solid , soil F oc , soil K oc 1000 × RHO solid

Descriptions and standard values of all parameters are given in [Table 4].

Table 4 Relevant parameters for PNEC_soil calculations, their description, and their default values.

Parameter	Description	Default value
K soil−water	Soil−water partitioning coefficient	–
RHOsoil	Bulk density of wet soil	1700 kg m−3
1000 (factor in Eq. [(2)])	Conversion factor from L to m3	1000
F air,soil	Air volume fraction in soil	0.2
K air−water	Air−water partitioning coefficient	0 for nonvolatiles
F water−soil	Water volume fraction in soil	0.2
F solid−soil	Solid volume fraction in soil	0.6
F oc,Soil	Weight fraction OC in soil solids	EU: 0.02 (default)
		NL: 0.059[60]
K oc	Organic carbon-normalized soil−water partitioning coefficient	–
RHOsolid	Bulk density of solid phase	2500 kg m−3

As all default parameter values were used and substances were assumed to be nonvolatile, the equation was simplified to Eq. [(4)].

PNEC soil = 0.1176 + 0.01764 K oc PNEC water

The environmental risk was then assessed based on the risk quotient, which relates exposure to toxicity and is defined according to Eq. [(5)].

Risk quotient RQ = Measured / predicted environmental concentration PEC / MEC Predicted no − effect concentration PNEC

Here, four risk levels are identified, namely Insignificant (RQ < 0.1), Low (RQ = 0.1–1), Moderate (RQ = 1–10) and High (RQ > 10) [61].

3

Results and Discussion

3.1

Anonymous surveys

According to recent numbers on drug use in the Netherlands, cannabis is the most consumed drug [62]. Indeed, the use of cannabis was on average the highest in our survey, with already 90 hits (26% of the respondents) on Day 1 ([Figure 1]). Its consumption remained quite stable over the festival, with an incremental increase of 16% and 4% on Days 2 and 3, respectively. As MDMA is the most popular party drug in the Netherlands (XTC has been grouped with it), a similarly expected result was the steep rise of indicated MDMA use as the festival progressed, going from 5 hits on Day 1, 71 hits on Day 2, to 129 hits on Day 3 [62]. The most often-consumed pharmaceutical was paracetamol, showing 40–63 hits over the three days. Only ibuprofen and cocaine showed a peak in hits on Day 2—all others consumed in significant amounts showed an increasing trend over the three days.

Figure 1 Survey results for participant drug intake for each of the three festival days.

Other substances that were not analyzed are grouped in “other,” which includes a wide array of compounds. Examples of illicit substances grouped here are gamma-hydroxybutanoic acid (GHB), 2-(4-bromo-2,5-dimethoxyphenyl)ethanamine (2C-B), 4-fluoroamphetamine (4-FMP), and lysergic acid diethylamide (LSD). Examples of pharmaceuticals in this category are modafinil, methylphenidate (Ritalin), antihistamine, and prednisone.

Note that the number of hits shown in [Figure 1] does not necessarily correspond to the actually consumed amount of substances and pharmaceuticals. Pharmaceuticals are generally consumed in larger quantities; one pill of paracetamol can contain 500–1000 mg of the active substance whereas illicit drugs are consumed in much smaller quantities. Hence, the numbers presented here cannot be directly correlated to expected analyte concentrations. Still, analyte concentrations of a specific substance in the collected urine were expected to follow roughly the same trends over time as those shown in [Figure 1].

3.2

Urine analysis

Contaminants in pre-precipitation urine were detected in a broad range of concentrations, ranging from < 1 to over 34,000 ng mL⁻¹. Regarding illicit drugs, the five most often-consumed ones were cocaine, amphetamine, MDMA, ketamine, and cannabis ([Figure 1]). Their concentrations—or those of their primary metabolites—were measured at 604.7 ± 34.21 to 862 ± 22.5 ng mL⁻¹ benzoylecgonine (cocaine’s primary metabolite), 98.18 ± 3.21 to 210 ± 17.0 ng mL⁻¹ amphetamine, 81.10 ± 3.38 to 4920 ± 582 ng mL⁻¹ MDMA, 26.38 ± 2.53 to 50.09 ± 0.58 ng mL⁻¹ norKetamine (ketamine’s primary metabolite), and 12.34 ± 0.54 to 44.44 ± 7.76 ng mL⁻¹ THCCOOH (cannabis’ primary metabolite) ([Table 5]). Other drugs, such as 3-MMC, mephedrone, and metabolite parent compounds, showed concentrations roughly between 0.5 and 20 ng mL⁻¹. The two primary pharmaceuticals consumed—NSAIDs ibuprofen and paracetamol—were detected at 640.01 ± 23.13 to 2946 ± 71.5 ng mL⁻¹ and 17490 ± 1737 to 34576 ± 99.8 ng mL⁻¹, respectively, the latter of which was the highest concentration found in all measurements. Cis-tramadol and O-desmethyltramadol were not detected in any sample ([Table 6]).

Table 5 Concentrations of illicit drug analytes in urine pre- and post-struvite precipitation, and the percentage remaining in urine.

Illicit drug analytes	Day 1			Day 2			Day 3
	Pre (ng mL−1)	Post (ng mL−1)	Remaining (%)	Pre (ng mL−1)	Post (ng mL−1)	Remaining (%)	Pre (ng mL−1)	Post (ng mL−1)	Remaining (%)
Cocaine	0.886 ± 0.111	1.22 ± 0.03	137.49% ± 17.61%	2.46 ± 0.17	1.86 ± 0.30	75.46% ± 13.14%	16.96 ± 1.98	2.76 ± 0.54	16.24% ± 3.69%
Benzoylecgonine	604.70 ± 34.21	462.15 ± 13.58	76.43% ± 4.87%	862 ± 22.5	649 ± 21.0	75.26% ± 3.13%	823 ± 130	628 ± 71.2	76.28% ± 14.84%
Cocaethylene	1.95 ± 0.16	2.23 ± 0.09	114.44% ± 10.61%	4.70 ± 0.28	3.65 ± 0.40	77.56% ± 9.63%	6.89 ± 0.79	2.90 ± 0.35	42.03% ± 6.97%
Amphetamine	98.18 ± 3.21	79.98 ± 5.79	81.46% ± 6.47%	178.29 ± 3.79	152.60 ± 16.11	85.59% ± 9.22%	210 ± 17.00	197 ± 11.95	93.73% ± 9.47%
Methamphetamine	<LOQ	<LOQ	–	0.19 ± 0.00	0.16 ± 0.01	83.62% ± 3.23%	0.28 ± 0.01	0.25 ± 0.00	87.44% ± 2.22%
MDMA	81.10 ± 3.38	80.13 ± 6.61	98.79% ± 9.13%	2541 ± 27.37	2072 ± 143	81.53% ± 5.71%	4920 ± 582	4684 ± 240	95.21% ± 12.27%
Ketamine	24.56 ± 0.86	20.91 ± 0.97	85.15% ± 4.96%	33.97 ± 1.41	27.50 ± 0.71	80.94% ± 3.96%	17.71 ± 2.24	16.16 ± 0.78	91.23% ± 12.35%
norKetamine	50.09 ± 0.58	29.27 ± 3.84	58.43% ± 7.70%	44.15 ± 1.89	25.65 ± 3.67	58.09% ± 8.67%	26.38 ± 2.53	25.56 ± 1.49	96.90% ± 10.87%
DehydronorKetamine	2.28 ± 0.03	1.61 ± 0.01	70.56% ± 1.05%	8.91 ± 0.08	6.75 ± 0.11	75.79% ± 1.41%	17.17 ± 0.34	11.32 ± 0.15	65.95% ± 1.55%
THCCOOH	12.45 ± 0.09	11.66 ± 1.18	93.66% ± 9.49%	12.34 ± 0.54	13.74 ± 1.46	111.36% ± 12.80%	44.44 ± 7.76	30.26 ± 2.37	68.09% ± 13.03%
Mephedrone	<LOQ	<LOQ	–	<LOQ	<LOQ	–	0.47 ± 0.03	0.32 ± 0.02	67.77% ± 6.10%
3-MMC	2.92 ± 0.04	1.39 ± 0.05	47.67% ± 1.89%	0.64 ± 0.03	0.54 ± 0.01	85.30% ± 4.62%	0.76 ± 0.06	0.53 ± 0.04	70.26% ± 8.09%

Table 6 Concentrations of pharmaceutical analytes in urine pre- and post-struvite precipitation, and the percentage remaining in urine.

Pharmaceutical analytes	Day 1			Day 2			Day 3
	Pre (ng mL−1)	Post (ng mL−1)	Remaining (%)	Pre (ng mL−1)	Post (ng mL−1)	Remaining (%)	Pre (ng mL−1)	Post (ng mL−1)	Remaining (%)
Ibuprofen	640.01 ± 23.13	559.82 ± 10.82	87.47% ± 3.58%	2946 ± 71.5	2323 ± 160	78.86% ± 5.77%	1566 ± 196	1416 ± 56.6	90.46% ± 11.89%
Paracetamol	30293 ± 2426	27234 ± 2753	89.90% ± 11.59%	34576 ± 99.8	24096 ± 673	69.69% ± 1.96%	17490 ± 1737	17060 ± 732	97.54% ± 10.55%
Citalopram	88.47 ± 4.49	83.25 ± 1.78	94.10% ± 5.18%	46.68 ± 0.32	30.91 ± 1.12	66.22% ± 2.45%	21.19 ± 3.09	18.27 ± 0.46	86.21% ± 12.77%
Diclofenac	<LOQ	<LOQ	–	<LOQ	<LOQ	–	26.86 ± 2.76	26.93 ± 1.70	100.25% ± 12.09%
Atorvastatin	<LOQ	<LOQ	–	0.48 ± 0.02	0.32 ± 0.01	65.47% ± 3.02%	0.70 ± 0.04	0.26 ± 0.01	37.56% ± 2.50%
cis-Tramadol	<LOQ	<LOQ	–	<LOQ	<LOQ	–	<LOQ	<LOQ	–
O-Desmethyltramadol	<LOQ	<LOQ	–	<LOQ	<LOQ	–	<LOQ	<LOQ	–

As expected, these concentrations are significantly higher than those typically found in urban WW. A study on illicit drugs at Dutch WWTPs showed influent concentrations of benzoylecgonine to be 260–3701 ng L⁻¹; 160–230 times lower than found in our urine samples when taking the highest influent concentration [63], [64]. THCCOOH was measured at 75–489 ng L⁻¹—25–90 times lower—while amphetamine and MDMA were found at concentrations of 40–1779 and 12–241 ng L⁻¹, respectively [63], [64]—roughly 55–120 and 335–20400 times lower, respectively. NSAIDs, being more readily accessible, exhibited significantly higher concentrations in WW at about 1–100 μg L⁻¹ [65]. For example, paracetamol concentrations in French municipal WW influent approached 100 μg L⁻¹, with reports of 105.7 μg L⁻¹ in Singapore; just short of 350 times lower than our findings [65]. Considering that commercial struvite is often precipitated from urban WW, it is indeed difficult to imagine a scenario worse for struvite precipitation in terms of organic micropollutant contamination.

Overall, analyte concentrations correlated quite well with the results of the surveys. Hits for cocaine use were 31, 46, and 43 over the three days ([Figure 1]) with respective benzoylecgonine concentrations of 604.7 ± 34.21, 862 ± 22.5, and 823 ± 130 ng mL⁻¹ ([Table 5]). The steep rise of MDMA consumption seen in [Figure 1] is equally pronounced in the actually measured concentrations, increasing by about 2400 ng mL⁻¹ per day ([Table 5]). Diclofenac, with only hits on Day 3, was only detected in Day 3 urine ([Table 6]). As stated, hits do not necessarily match analyte concentrations: ibuprofen received fewer hits but far exceeded the concentrations of others, indicating higher usage ([Table 6]). Some analytes did not show such a clear correlation, however. For example, ketamine hits increased each day, yet it and its metabolite norKetamine showed no such trend ([Table 5]). Citalopram should, according to the survey, only have been detected on Day 1 but was found in each urine sample, albeit at decreasing concentrations—likely at least partially due to the carryover effect induced by the inoculation volumes. As measured concentrations in urine roughly halved every day whereas the inoculation volume was much less than half, another explanation could be that some respondents did not list citalopram on Days 2 and/or 3.

The large majority of contaminants remained in the supernatant after struvite precipitation ([Tables 5] and [6], [Figure 2]). Some analytes present at low concentrations showed erratic behavior, however. For example, the remaining percentages of cocaine and cocaethylene were 16.2 and 42.0% on Day 3 but 137 and 114% on Day 1, respectively. Numbers of over 100% are not unheard of; other studies into the fate of organic micropollutants have reported similar results [29], [66], [67]. To filter out such apparent outliers and allow for a clearer picture, analytes were divided based on measured pre-precipitation concentrations, namely <10, 10–100, and >100 ng mL⁻¹. Indeed, the five analytes present at >100 ng mL⁻¹ (benzoylecgonine, amphetamine, MDMA, ibuprofen, and paracetamol) show a much more consistent picture than less abundant ones. When combining their numbers, remaining percentages of 86.6 ± 7.6, 78.2 ± 5.4, and 90.6 ± 7.5% were found on Days 1, 2, and 3 respectively, with a combined average remaining percentage of 85.2 ± 8.7% taken over three days (see [Figure 2] for individual remaining percentages of analytes present in >100 ng mL⁻¹. Other concentration ranges are visualized in Figure S4). However, no clear trends could be observed correlating remaining percentages with physicochemical properties such as log K _ow, log K _oc or pK _a, which has also been concluded elsewhere [29].

Figure 2 Remaining percentages of analytes present in >100 ng mL⁻¹ after struvite precipitation. For clarity, standard deviations are not shown (see [Tables 6] and [7]).

The remaining percentage of 85.2 ± 8.7% is lower than the numbers reported in other studies investigating pharmaceutical uptake in struvite [28], [29], [66], [68]. In one of these studies, it is noted that a higher Mg:P ratio results in reduced pharmaceutical uptake—specifically tetracyclines [68]. This is owed to the chelating effect of Mg²⁺ ions in solution. Indeed, a different study reports >98% of pharmaceuticals and hormones remaining in the supernatant post-precipitation, using a Mg:P ratio of about 1.5 [28]. Yet another study states deliberately using an Mg:P ratio of 1.2—which is considered low by the authors—to facilitate the uptake of pharmaceuticals [29]. Thus, the equimolar ratio used in this study may well have played a role in the larger differences between pre- and post-precipitation concentrations observed in this study compared to others. Additionally, previously mentioned studies all spiked their solutions of interest with pharmaceuticals, whereas the samples in this study only contain naturally excreted substances.

As experiments were carried out batch-wise, slightly varying parameters such as temperature and Mg addition (urine volumes were rounded to 5 L increments) as well as specific urine composition (i.e., OC content) provide a likely explanation for the observed variation in the remaining percentage between the three days (78.2–90.6%). Hence, when interpreting results of this experiment type it is more prudent to view them as result “bandwidths” more than individual results.

3.3

Struvite analysis

Contaminant concentrations in struvite were found in a wide range of 0.04 to 1081 ng g⁻¹, in a similar way as seen for concentrations in urine. The lowest and highest values found over three days (given as [lowest concentration]–[highest concentration]) were 7.03–8.70 ng g⁻¹ benzoylecgonine, 4.29–8.59 ng g⁻¹ amphetamine, and 18.6–48.9 ng g⁻¹ THCCOOH. Similar to the survey and urine results, MDMA concentrations rose steeply as the festival progressed with 23.1, 62.6, and 208 ng g⁻¹ on Days 1, 2, and 3, respectively. NorKetamine concentrations were measured at 2.14–3.74 ng g⁻¹, which, interestingly, are lower than those of dehydronorKetamine (3.95–16.1 ng g⁻¹) even though the former is the main metabolite and was present in significantly higher concentrations in all parent urine samples. Ketamine itself showed concentrations of 0.97–2.32 ng g⁻¹. Regarding non-metabolized cocaine and cocaethylene—formed when combining alcohol with cocaine use—we see very low concentrations of below 0.072 ng g⁻¹ in all cases ([Table 7]). Methamphetamine, 4-MMC, and 3-MMC were not detected in any struvite sample, which is unsurprising considering their very low concentrations in urine ([Table 5]). Pharmaceutical contaminants were detected in higher concentrations than their illicit drug counterparts, with 34.5–62.6 ng g⁻¹ ibuprofen, 121–273 ng g⁻¹ paracetamol, and 53–1081 ng g⁻¹ citalopram. All analyte concentrations in struvite are listed in [Table 7]. Note that diclofenac, cis-tramadol, and O-desmethyltramadol were not measured in struvite samples due to lack of presence in their parent urine samples.

Table 7 Concentrations of illicit drug- and pharmaceutical analytes in the struvite precipitate.

Illicit drug analytes	Day 1 (ng g−1)	Day 2 (ng g−1)	Day 3 (ng g−1)
Cocaine	<LOQ	0.05	0.039
Benzoylecgonine	8.70	7.03	7.64
Cocaethylene	0.072	0.034	0.041
Amphetamine	8.59	4.29	6.93
Methamphetamine	<LOQ	<LOQ	<LOQ
MDMA	23.1	62.6	208
Ketamine	2.32	1.16	0.97
norKetamine	3.74	2.14	2.97
DehydronorKetamine	16.1	3.95	5.3
THCCOOH	23.7	18.6	48.9
Mephedrone	<LOQ	<LOQ	<LOQ
3-MMC	<LOQ	<LOQ	<LOQ
Pharmaceutical analytes
Ibuprofen	34.5	55.5	62.6
Paracetamol	273	146	121
Citalopram	1081	105	53.0
Atorvastatin	<LOQ	<LOQ	<LOQ

There are only a few literature reports to compare these results with. Analysis of commercial AirPrex struvite from municipal WW showed concentrations of 6–14 ng g⁻¹ for substances such as ciprofloxacin and triclosan (antibiotics), carvedilol (beta blocker), and carbamazepine (anti-epileptic) [69]. Reports regarding struvite from other sources list significantly higher concentrations with 0.5 μg g⁻¹ for the biocide 2,4-dichlorophenol (source-separated gray water) [70], 1.26 μg 4-tert-butylphenol g⁻¹ (digested sludge) [30], and up to 2.0 μg tetracyclines g⁻¹ and 1.1 μg fluoroquinolones g⁻¹ (swine WW) [71]. Our results are thus comparable to those found in the literature, with most illicit drug contaminants (except MDMA and THCCOOH) showing similar concentrations as substances found in AirPrex struvite. The higher concentrations of pharmaceutical contaminants were, with the exception of 1081 ng g⁻¹ citalopram, still lower than numbers listed for contaminants in differently-sourced struvite, showing that the “Lowlands” struvite performs equal to or better than various samples of commercial struvite in terms of organic substance contamination.

While high contaminant concentrations in urine generally lead to higher concentrations seen in struvite, as is the case for ibuprofen and paracetamol for example, this correlation is different for each contaminant. Indeed, citalopram and paracetamol showed the two highest concentrations found in struvite (1081 and 273 ng g ⁻¹, respectively), but concentrations in the parent urine were vastly different (88.5 ± 4.5 ng mL⁻¹ and 30,293 ± 2,426 ng mL⁻¹, respectively). To quantify the relationship between analyte concentrations in struvite and urine, we can define an “incorporation factor” (IF) (Eq. [6]). For citalopram and paracetamol, IFs then become 12.22 and 0.009, respectively. We can explain this large difference by taking into account the K _oc values of the contaminants in question; 14,709 versus 31.88, incidentally the highest and lowest of all analytes. While struvite theoretically does not contain organic carbon, K _oc can be interpreted as a measure of lipophilicity or tendency to prefer a solid phase over an aqueous one in this context.

Incorporation factor mL g − 1 = Concentration in struvite ng g − 1 Concentration in urine ng mL − 1

When comparing IFs with K _oc values for analytes present in every struvite sample, we find a clear correlation for the highest and lowest values: paracetamol and benzoylecgonine (K _oc = 31.88 and 100.07, respectively) show IFs of 0.014 and below in all cases, whereas THCCOOH and citalopram (K _oc = 12,460 and 14,709, respectively) show IFs of 1.10 and higher in all cases. The other analytes present in each struvite sample have less pronounced K _oc values (all in the log 2 order) and do not clearly follow the same trend regarding their IF. However, we can still observe that cocaethylene shows lower IFs (0.006–0.034) than may be expected based on its K _oc value (801.78). Furthermore, IF analysis shows the anomalous nature of Day 1 struvite’s dehydronorKetamine and citalopram concentrations with IFs of 7.061 and 12.22, respectively—far higher than those on Days 2 and 3 (0.443 and 0.309 for dehydronorKetamine, 2.25 and 2.50 for citalopram, respectively). MDMA also shows a relatively high Day 1 IF of 0.285 compared to 0.025 and 0.042 on subsequent days. While IFs are highest on Day 1 in all cases except norKetamine, other analytes do not show such a large difference as these three. It is unclear why this is the case. The higher IFs observed on Day 1 may be due to the lower struvite yield that day. This is likely due to the high inoculation volume, as phosphate may have already (partially) precipitated due to elevated pH caused by urea hydrolysis prior to the experiment. Thus, precipitation amount, along with contaminant concentrations in urine and their respective K _oc values, appears to influence IFs and contribute to inter-experiment variability. Further research is required to state this with more certainty. K _oc values and IFs of analytes detected in all three struvite samples are shown in [Figure 3]. Here, each column is scaled separately to account for inter-experiment variability.

Figure 3 Heatmap correlating analyte incorporation factors (IFs) with their K _oc values. Each column is scaled separately.

It should be noted that K _oc values of the analytes can be adjusted for (partial) ionization at pH 9, lowering them since charged species are more hydrophilic than neutral ones. As this comes with a number of considerations, a more detailed discussion is available in the Supporting Information (S9) as an in-depth analysis here was deemed beyond the scope of this work. The discussed adjustment would only move cocaethylene up two entries in [Figure 3], with no effect on the order of other analytes.

3.4

Legislation on struvite contamination

The legal use of struvite as fertilizer is subject to both national and international legislation. Dutch law requires compliance in terms of fertilizing components (N, P₂O₅, K₂O), heavy metals (HMs), and organic micropollutants—namely, polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs) and other chlorinated hydrocarbons [55]. Struvite from Dutch WWTPs has been shown to meet these criteria [30]. Biological contaminants such as pathogens do not have hard legal limits due to their presence in manure.

The EU Fertilizer Products Regulation imposes similar legal limits regarding HMs (anywhere from 3 to 1500 mg kg⁻¹) and PAHs (Σ 6 mg kg⁻¹ DW), as well as biuret (12 g kg⁻¹ DW), perchlorate (50 mg kg⁻¹ DW). Furthermore, OC content is capped at 3% and some bacteria and phages (E. coli, Enterococcaceae, Clostridium perfringens, Salmonella spp., and Ascaris sp. eggs) are limited (up to 1000 CFU g⁻¹ fresh mass) [16], [18], [72]. A recent study shows that commercial struvite generally stays within all listed legal limits and listed compounds pose no risk [18]. Two findings of note: 3 out of 25 commercial struvite samples did exceed the legal OC limit (10.8–13.2% OC DW) but were all recovered from Upflow Anaerobic Sludge Blanket (UASB) reactor effluents; all other samples came in at 0.12–1.6% OC DW [18]. It is therefore highly likely struvite recovered from urine will meet this standard as well, especially after rinsing/washing with water. Secondly, biological contamination of struvite can possibly exceed legal limits but is often concentrated in organic macro-impurities such as twigs and can be reduced by the removal of such impurities or storage [18], [30]. Tables S5–S7 give a more detailed overview of Dutch and European legislative requirements.

No organic macro-impurities were present in the “Lowlands” struvite due to its urine origin. Furthermore, the maximum concentrations listed for biuret, perchlorate, and PAHs are orders of magnitude higher than concentrations found for analytes considered in this study ([Table 7]). Only the outlying citalopram concentration in Day 1’s struvite (1.08 mg kg⁻¹) comes remotely close. Thus, the “Lowlands” struvite appears to be safe to use from a legislative point of view. However, neither national nor international legislation requires compliance regarding pharmaceutically active substances. To still be able to provide a meaningful discussion on organic microcontaminants in struvite, an environmental risk assessment for “Lowlands” struvite was performed.

3.5

Environmental risk assessment

RQs were determined according to Eqs. [(4)] and [(5)]. Based on the derived indicative PNEC_soil values and maximum concentration in the soil, RQs for all analytes were found to be at or below 0.00. This means the associated risk of applying this particular struvite for one year on Dutch soils is classified as insignificant (RQ < 0.1) for all analytes. When considering 100 years of continuous fertilization using this particular struvite, MDMA and paracetamol would have low RQs (0.24 and 0.25, respectively), and THCCOOH, benzoylecgonine, ibuprofen and citalopram insignificant ones (0.02, 0.01, 0.01, and 0.01, respectively). All other analytes would retain a RQ of (<) 0.00 (see [Table 8]). Note that the maximum concentration in soil is given in ng kg⁻¹, whereas the maximum concentration in struvite is given in μg kg⁻¹. See Excel S1 and S2 for more details regarding PNEC values and other risk assessment calculations.

Table 8 Environmental risk assessment for analytes found in struvite applied to Dutch soils.

Analyte	Highest measured concentration (μg kg−1)	Maximum concentration in soil (ng kg−1)	Indicative PNECsoil for Dutch soils (5.9% OC) (μg kg−1)	Risk quotient	Risk quotient after 100 years
a LOQ values taken.
Cocaine	0.05	0.01	63.00	0.00	0.00
Benzoylecgonine	8.7	1.45	12.94	0.00	0.00
Cocaethylene	0.072	0.01	6.73	0.00	0.00
Amphetamine	8.59	1.43	109.16	0.00	0.00
Methamphetamine	< 0.27*	0.04	40.25	<0.00	<0.00
MDMA	208	34.60	14.40	0.00	0.24
Ketamine	2.32	0.39	33.53	0.00	0.00
norKetamine	3.74	0.62	33.59	0.00	0.00
DehydronorKetamine	16.1	2.68	–	–	–
THCCOOH	48.9	8.13	52.55	0.00	0.02
4-MMC	< 0.16a	0.03	34.73	< 0.00	< 0.00
3-MMC	< 0.16a	0.03	40.11	< 0.00	< 0.00
Ibuprofen	62.6	10.41	74.00	0.00	0.01
Paracetamol	273	45.41	18.46	0.00	0.25
Citalopram	1081	179.81	1225.24	0.00	0.01
Atorvastatin	<0.031a	0.01	64.22	<0.00	<0.00

When translating these numbers to European soils, it is simply a matter of multiplying the risk quotient numbers by factor 2.95: the factor representing the difference in organic carbon in Dutch and European soils (5.9 and 2%, respectively) [60]. In this case, risk quotients of paracetamol and MDMA would go from 0.00 to 0.01, all others would remain (<) 0.00. While risk quotients after 100 years of “Lowlands” struvite application would rise accordingly, all analytes would still fit their previously assigned risk level (insignificant and low).

RQs derived in this study do not take into account the transformation of the contaminants over time. While transformation may lower the environmental risk even further, it should be noted that transformation products may have worse environmental properties than their parent compound, thereby increasing environmental risk instead. For example, methylation of diclofenac significantly increases both bioaccumulation and acute toxicity in two studied aquatic invertebrates [73]. Taking such transformation products into account is outside the scope of this study, but should be considered wherever possible when assessing the impact of organic pollutants in the environment. Still, based on the found RQs, we feel it is safe to say that struvite recovered from urine highly contaminated with pharmaceuticals and illicit drugs would not pose a risk to the environment when being applied as fertilizer—an assessment that holds true even over longer periods of time.

4

Conclusions

For recovered phosphates to reenter the phosphorus cycle, it is imperative that contaminants are kept to a minimum to minimize their reintroduction. This work addresses a knowledge gap on the incorporation of organic microcontaminants—namely pharmaceuticals and illicit drugs—in struvite precipitated from source-separated urine and the associated environmental risk when using this struvite as fertilizer. Measured concentrations in urine correlated well with survey results and were especially high for the illicit drugs MDMA and benzoylecgonine, and the pharmaceuticals ibuprofen and paracetamol. In total, 17 different substances or metabolites thereof were measured in collected urine samples. Twelve of these were subsequently detected in precipitated struvite, albeit at low concentration in most cases (predominantly 0–100 ng g⁻¹) with an outlier at 1.08 μg g⁻¹ for citalopram. The K _oc value of a substance was shown to have a large influence, with higher values translating to higher incorporation factors in struvite. The amount of struvite precipitate appears to have an influence on the incorporation factor as well, but more research is required to postulate this with more certainty. Coprecipitated contaminants would pose an insignificant environmental risk when using “Lowlands” struvite as fertilizer, with only MDMA and paracetamol showing low risk when applied for 100 years. This shows that even in a worst-case scenario, organic micropollutant contamination of struvite is of little concern.

Contributors’ Statement

The following CRediT authorship contribution statement applies: Steven Beijer: Conceptualization, Investigation, Methodology, Formal analysis, Writing – Original draft, Visualization, Project administration; Noelia Salgueiro-Gonzalez: Investigation, Methodology, Formal analysis, Validation, Writing – Review & Editing; Sara Castiglioni: Methodology, Validation, Writing – Review & Editing, Project administration, Funding acquisition; Juan C. Gerlein: Resources, Investigation; Peter Scheer: Resources, Supervision; G. Bas de Jong: Conce ptualization, Resources, Writing – Review & Editing; J. Chris Slootweg: Conceptualization, Writing – Review & Editing, Supervision, Project administration, Funding acquisition.

Conflict of Interest

The authors declare that they have no conflict of interest.

Acknowledgments

Many thanks to Lowlands Science for selecting our team to showcase phosphate recovery at “A Campingflight to Lowlands Paradise 2019” and execute this citizens science research project. Aalke Lida de Jong has our gratitude for graciously sharing her environmental risk assessment method. Jens Tolboom, Maxime Gerber, Anna Butter, and Alba Fonseca Topp are gratefully acknowledged for their help with the preparations prior to- and operations during the festival.

• References

• 1 Alberts B, Johnson A, Lewis J, Raff M, Roberts K, Walter P. Molecular Biology of the Cell. Garland Science; New York: 2008
• 2 Jupp AR, Beijer S, Narain GC, Schipper W, Slootweg JC. Chem. Soc. Rev. 2021; 50: 87-101
• 3 Filipelli GM. Elements 2008; 4: 89-95
• 4 Asimov I. In Mag. Fantasy Sci. Fict.. Mercury Press; Oklahoma City: 1959
• 5 Cordell D, Drangert J, White S. Glob. Environ. Chang. 2009; 19: 292-305
• 6 van Kauwenbergh SJ, Stewart M, Mikkelsen R. Better Crop. 2013; 97: 18-20
• 7 European Commission. Critical Raw Materials Resilience: Charting a Path towards Greater Security and Sustainability. European Commission; 2020
• 8 Rittmann BE, Mayer B, Westerhoff P, Edwards M. Chemosphere 2011; 84: 846-853
• 9 Gustavsson J, Cederberg C, Sonesson U, van Otterdijk R, Meybeck A. Global Food Losses and Food Wastes: Extent, Causes and Prevention. Rome: 2011
• 10 Scherhaufer S, Moates G, Hartikainen H, Waldron K, Obersteiner G. Waste Manage. 2018; 77: 98-113
• 11 Cordell D, White S. Sustainability 2011; 3: 2027-2049
• 12 van Dijk KC, Lesschen JP, Oenema O. Sci. Total Environ. 2016; 542: 1078-1093
• 13 Withers PJ. A, van Dijk KC, Neset TS. S, Nesme T, Oenema O, Rubæk GH, Schoumans OF, Smit B, Pellerin S. Ambio 2015; 44: 193-206
• 14 Egle L, Rechberger H, Krampe J, Zessner M. Sci. Total Environ. 2016; 571: 522-542
• 15 Talboys PJ, Heppell J, Roose T, Healey JR, Jones DL, Withers PJ. A. Plant Soil 2016; 401: 109-123
• 16 Huygens D, Saveyn H, Tonini D, Eder P, Sancho LD. Technical Proposals for Selected New Fertilising Materials under the Fertilising Products Regulation (Regulation (EU) 2019/1009): Process and Quality Criteria, and Assessment of Environmental and Market Impacts for Precipitated Phosphate Salts & Derivates. 2019
• 17 Shaddel S, Bakhtiary-Davijany H, Kabbe C, Dadgar F, Østerhus SW. Sustainability 2019; 11: 1-12
• 18 Muys M, Phukan R, Brader G, Samad A, Moretti M, Haiden B, Pluchon S, Roest K, Vlaeminck SE, Spiller M. Sci. Total Environ. 2021; 756: 143726
• 19 Snoeyink VL, Jenkins D. Water Chemistry. John Wiley & Sons, Inc.; New York: 1980
• 20 Ohlinger KN, Young TM, Schroeder ED. Water Res. 1998; 32: 3607-3614
• 21 Booker NA, Priestley AJ, Fraser IH. Environ. Technol. 1999; 20: 777-782
• 22 Buchanan JR, Mote CR, Robinson RB. Trans. Am. Soc. Agric. Eng. 1994; 37: 617-621
• 23 Kehrein P, Van Loosdrecht M, Osseweijer P, Garfí M, Dewulf J, Posada J. Environ. Sci.: Water Res. Technol. 2020; 6: 877-910
• 24 Cornel P, Schaum C. Water Sci. Technol. 2009; 59: 1069-1076
• 25 Doyle JD, Parsons SA. Water Res. 2002; 36: 3925-3940
• 26 Le Corre KS, Valsami-Jones E, Hobbs P, Parsons SA. Crit. Rev. Environ. Sci. Technol. 2009; 39: 433-477
• 27 Crutchik D, Garrido JM. Chemosphere 2016; 154: 567-572
• 28 Ronteltap M, Maurer M, Gujer W. Water Res. 2007; 41: 1859-1868
• 29 de Boer MA, Hammerton M, Slootweg JC. Water Res. 2018; 133: 19-26
• 30 STOWA. Verkenning van de Kwaliteit van Struviet Uit de Communale Afvalwaterketen. 2015
• 31 Friedler E, Butler D, Alfiya Y. In Source Sep. Decentralization Wastewater Manag. Larsen TA, Udert KM, Lienert J. Eds IWA Publishing; 2013. pp 241-257
• 32 Udert KM, Larsen TA, Gujer W. Water Sci. Technol. 2006; 54: 413-420
• 33 Larsen TA, Gujer W. Water Sci. Technol. 1996; 34: 87-94
• 34 Lienert J, Bürki T, Escher BI. Water Sci. Technol. 2007; 56: 87-96
• 35 Liu Z, Zhao Q, Wang K, Lee D, Qiu W, Wang J. J. Environ. Sci. 2008; 20: 1018-1024
• 36 Etter B, Tilley E, Khadka R, Udert KM. Water Res. 2011; 45: 852-862
• 37 Liu X, Hu Z, Mu J, Zang H, Liu L. Environ. Technol. 2014; 35: 2781-2787
• 38 Petrie B, Youdan J, Barden R, Kasprzyk-Hordern B. J. Chromatogr. A 2016; 1431: 64-78
• 39 Drugbank, webpage on cocaethylene, can be found under https://go.drugbank.com/metabolites/DBMET00645
• 40 Grabic R, Ivanová L, Kodešová R, Grabicová K, Vojs Staňová A, Imreová Z, Drtil M, Bodík I. Water Res. 2022; 220 118651
• 41 Dinis-Oliveira RJ. Forensic Sci. Res. 2017; 2: 2-10
• 42 Drugbank, webpage on norKetamine, can be found under https://go.drugbank.com/metabolites/DBMET00189
• 43 Avdeef A, Berger CM, Brownell C. Pharm. Res. 2000; 17: 85-89
• 44 Settimo L, Bellman K, Knegtel RM. A. Pharm. Res. 2014; 31: 1082-1095
• 45 Malev O, Lovrić M, Stipaničev D, Repec S, Martinović-Weigelt D, Zanella D, Ivanković T, Sindičić Đuretec V, Barišić J, Li M, Klobučar G. Environ. Pollut. 2020; 266 115162
• 46 Ahjel SW, Lupuleasa D. Farmacia 2009; 57: 290-300
• 47 Drugbank, webpage on atorvastatine, can be found under https://go.drugbank.com/drugs/DB01076
• 48 Kirla KT, Erhart C, Groh KJ, Stadnicka-Michalak J, Eggen RI. L, Schirmer K, Kraemer T. Toxicol. Appl. Pharmacol. 2021; 419: 115483
• 49 Bijlsma L, Celma A, Castiglioni S, Salgueiro-González N, Bou-Iserte L, Baz-Lomba JA, Reid MJ, Dias MJ, Lopes A, Matias J, Pastor-Alcañiz L, Radonić J, Turk Sekulic M, Shine T, van Nuijs AL. N, Hernandez F, Zuccato E. Sci. Total Environ. 2020; 725 138376
• 50 Helfer AG, Michely JA, Weber AA, Meyer MR, Maurer HH. Anal. Chim. Acta 2015; 891: 221-233
• 51 Castiglioni S, Zuccato E, Crisci E, Chiabrando C, Fanelli R, Bagnati R. Anal. Chem. 2006; 78: 8421-8429
• 52 Castiglioni S, Davoli E, Riva F, Palmiotto M, Camporini P, Manenti A, Zuccato E. Water Res. 2018; 131: 287-298
• 53 Salgueiro-González N, Zuccato E, Castiglioni S. Sci. Total Environ. 2022; 843: 156982
• 54 Rijksdienst voor Ondernemend Nederland. Hoeveel Mest Gebruiken. Hoe Rekent u Dat Uit?. 2019
• 55 Commissie Deskundigen Meststoffenwet. Protocol Beoordeling Stoffen Meststoffenwet. Versie 3.2. Wageningen: 2016
• 56 European Commission. Technical Guidance Document on Risk Assessment Part III, Version 2, 2003. 2003
• 57 NORMAN. Deriving Environmental Quality Standards for Chemical Substances in Surface Waters. 2022
• 58 Thomaidi VS, Stasinakis AS, Borova VL, Thomaidis NS. J. Hazard. Mater. 2015; 283: 740-747
• 59 Smit CE. Effecten van Drugs Op Het Waterecosysteem. 2015
• 60 Janssen MP. M, Traas TP, Rila J.-P, van Vlaardingen PL. A. 2004: 1-35
• 61 Wennmalm Å, Gunnarsson B. Environ. Int. 2009; 35: 775-777
• 62 Trimbos. Webpage on use of drugs in the Netherlands in 2022, can be found under https://www.trimbos.nl/actueel/nieuws/nieuwe-cijfers-drugsgebruik-in-nederland-in-2022/?vgo_ee=GHN8IKYaWucDAsBHzK4f%2BIpsVVpzmcBI7c0sAXfy0nK99w%3D%3D%3AqDCW%2BlhIAMXLpf6cImugTGNCz3WQ1ezO
• 63 Yadav MK, Short MD, Aryal R, Gerber C, van den Akker B, Saint CP. Water Res. 2017; 124: 713-727
• 64 Bijlsma L, Emke E, Hernández F, De Voogt P. Chemosphere 2012; 89: 1399-1406
• 65 Khasawneh OF. S, Palaniandy P. Process Saf. Environ. Prot. 2021; 150: 532-556
• 66 Kemacheevakul P, Otani S, Matsuda T, Shimizu Y. Water Sci. Technol. 2012; 66: 2194-2201
• 67 Kasprzyk-Hordern B, Dinsdale RM, Guwy AJ. Water Res. 2009; 43: 363-380
• 68 Kemacheevakul P, Chuangchote S, Otani S, Matsuda T, Shimizu Y. Water Sci. Technol. 2015; 72: 1102-1110
• 69 de Boer MA, Kabbe C, Slootweg JC. Environ. Sci. Technol. 2018; 52: 14564-14565
• 70 Butkovskyi A, Leal LH, Zeeman G, Rijnaarts HH. M. Environ. Res. 2017; 156: 434-442
• 71 Ye ZL, Deng Y, Lou Y, Ye X, Chen S. Front. Environ. Sci. Eng. 2018; 12: 1-7
• 72 European Commission. Regulation (EU) 2019/1009 of the European Parliament and of the Council of 5 June 2019 Laying down Rules on the Making Available on the Market of EU Fertilising Products and Amending Regulations (EC) No 1069/2009 and (EC) No 1107/2009 and Repealing Regula. European Commission; 2019
• 73 Fu Q, Fedrizzi D, Kosfeld V, Schlechtriem C, Ganz V, Derrer S, Rentsch D, Hollender J. Environ. Sci. Technol. 2020; 54: 4400-4408

Publication History

Received: 27 November 2024

Accepted after revision: 12 February 2025

Accepted Manuscript online:
14 February 2025

Article published online:
05 March 2025