Process Safety and Environmental Protection 158 (2022) 486–494

Contents lists available at ScienceDirect

Process Safety and Environmental Protection

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

Valorisation of drinking water treatment sludge as substrate in

subsurface flow constructed wetlands for upgrading treated wastewater ]]
]]]]]]
]]

⁎
Carmen Hernández-Crespo a, , Núria Oliver b, María Peña b, Miguel Añó a,c, Miguel Martín a
a
Instituto Universitario de Ingeniería del Agua y Medio Ambiente, Universitat Politècnica de València, Camino de Vera s/n, 46022 Valencia, Spain
b
Global Omnium Medioambiente S.L, Spain
c
Empresa Mixta Valenciana de Aguas S.A, Spain

a r t i cl e i nfo a bstr ac t

Article history: Drinking water treatment sludge (DWTS) is the main waste produced in drinking water treatment plants
Received 28 September 2021 (DWTPs). Its valorisation as substrate for constructed wetlands (CWs) aimed at upgrading treated urban
Received in revised form 14 December 2021 wastewater is presented. Keeping a holistic approach in mind, this study looks for nutrient and organic
Accepted 14 December 2021
matter removal but also contaminants of emerging concern (CECs) and pathogens. Three pilot subsurface
Available online 16 December 2021
flow CWs (1 m2) were installed under outdoor conditions in real WWTPs. Different operation modes (se­
quential: S-CW and continuous saturated flow: C-CW, CC-CW), different nutrient influent concentrations (S-
Keywords:
Alum sludge CW and C-CW: 0.6 mg TP/l, 12.7 mg TN/l; CC-CW: 6.5 mg TP/l, 48 mg TN/l) and high hydraulic loading rates
Treatment wetlands (HLRs, 0.9–5.1 m3/m2/d) were tested. C-CW presented higher removal efficiencies than S-CW for TP (C-CW:
Emerging pollutants 56–86%; S-CW: 32–66%), total nitrogen (C-CW: 23–38%; S-CW: −3 to 6%) and E. coli (C-CW: 94%; S-CW:
Upgrading treatment 84%), while S-CW performed better for ammonium (C-CW: 29–45%; S-CW: 72–86%) and CECs removal.
Phosphorus removal Among fifteen CECs monitored, most pharmaceuticals, four were significantly reduced in C-CW and nine in
Pathogens removal S-CW, which had more aerobic conditions. CC-CW reduced nutrients and organic matter by 62% (TP), 8%
(TN), 23% and 40% (chemical and biochemical oxygen demands, respectively). The potential release of
aluminium was negligible. Novel values for the first-order reaction coefficient of P-k-C* model are provided
for the TP removal process using DWTS (0.6–1.0 h−1). The main conclusion is that DWTS is a suitable
substrate to significantly upgrade WWTP effluents, even at high HLRs. A hybrid system combining
sequential and continuous flow modes could optimize the upgrading treatment. A proposal for the full
valorisation of the sludge produced in one DWTP is presented.
© 2021 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

1. Introduction WWTPs whose receiving water body is the Lake Albufera ( RD 1/,
2016), a shallow hypereutrophic lagoon located in Valencia (Spain).
Constructed wetlands (CWs) constitute a nature-based solution Although this is an important step to recover the good status of the
for wastewater treatment, particularly suitable for small towns. They lake, according to the Water Framework Directive, earlier water
can efficiently remove organic matter and suspended solids, whereas quality modelling studies reported that it would be necessary a
specific configurations or substrates are needed when nutrients maximum of 0.1 mg/l for all the influents to the lake, for achieving
must be removed, to meet the requirements of Urban Wastewater this goal (Martín et al., 2013). However, it is difficult to achieve such
European Directive (UWWTD) (Directive 91/271/EEC). low concentration in conventional WWTP since the biological
According to the Redfield ratio, the discharge limits fixed in treatment could be negatively affected because of nutrient limitation
UWWTD for phosphorus (1–2 mg TP/l) are not enough to prevent or bulking problems (Jenkins et al., 2004). At this point, CWs can
eutrophication when WWTP effluents represent a high percentage of play a key role in further phosphorus removal by using adsorbent
the water inputs to a water body, especially in case of lentic aquatic substrates, especially in small WWTPs, where other kind of treat­
ecosystems, such as natural wetlands or lakes. Accordingly, Jucar ments for nutrient removal complicates its management.
River Basin Plan has lowered the discharge limit to 0.6 mg TP/l for Drinking water treatment sludge (DWTS) is the residual of po­
table water treatment process, not harmful and without toxic ele­
ments such as heavy metals (Zhao et al., 2021). Moreover, it has a
⁎
Corresponding author. high phosphorus adsorption capacity, as well as other pollutants
E-mail address: [email protected] (C. Hernández-Crespo).

https://doi.org/10.1016/j.psep.2021.12.035
0957-5820/© 2021 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
C. Hernández-Crespo, N. Oliver, M. Peña et al. Process Safety and Environmental Protection 158 (2022) 486–494

such as metals and semimetals, thus representing a low-cost 2.2. Chemical analyses
adsorbent material (Zhao et al., 2021).
To the best of our knowledge, only few studies have focused on Water quality (WQ) variables analysed, and methods (in par­
the potential reuse of DWTS for treating wastewater with low enthesis), were: turbidity (TN100 Eutech turbidity-meter), total
phosphorus concentrations (Gao et al., 2020). Regarding to con­ suspended solids (TSS) (UNE-EN 872, UNE 77034), chemical oxygen
taminants of emerging concern (CECs), few studies have dealt with demand (COD) (Hach test: ISO 6060), biochemical oxygen demand
the potential of DWTS as substrate for removing pesticides (Hu et al., (BOD5) (respirometry test through OxiTop® control system), total
2011; Zhao et al., 2013), surfactants (Jangkorn et al., 2011), tetra­ nitrogen (TN) (Hach test: ISO 11905–1 +photometry), ammonium
cyclines or dyes (Devi and Saroha, 2017). A recent review (Wang (N-NH4+) (Hach test: ISO 7150–1), nitrites (N-NO2-) (Hach test: ISO
et al., 2020), did not report any study regarding pharmaceuticals nor 26777), nitrates (N-NO3-) (Hach test: ISO 7890–1–2), total phos­
faecal indicator removal using DWTS. Accordingly, no study has phorus (TP) (Hach test: digestion + ISO 6878), orthophosphates (P-
addressed the role of DWTS as substrate in CWs for removing PO43-) (Hach test: ISO 6878), aluminium (Al) (ultraviolet–visible
pharmaceuticals, thus it is interesting to cover this knowledge gap. spectrophotometry). The analyses frequency was 3 times a week for
Previous studies have demonstrated that CW are able to remove the physicochemical parameters and soluble compounds, and one a
CECs, at least as efficient as conventional activated sludge, because of week for total compounds. Escherichia coli (ISO 9308–1) was mon­
their higher sludge retention time and the coexistence of aerobic, itored weekly from Apr to Sep 2019 and monthly afterwards (n = 24).
anoxic and anaerobic microenvironments (Dotro et al., 2017). Intestinal nematode eggs (modified Bailinger’s Method) were
In the present study, pilot-scale CWs were tested under outdoor evaluated once a month.
weather conditions and treating real wastewater coming from two A total of 15 CECs, belonging to different therapeutic classes,
different secondary treatments. The objective of the study was to were monitored monthly. The CECs were selected because of their
assess the ability of DWTS-based CWs to upgrade the quality of widespread use and high frequency in urban wastewater. Chemical
secondary effluents, from a holistic point of view, looking at organic analyses were performed by an accredited laboratory, which used
matter and nutrient removal, but also emerging pollutants and the following methods: CSN EN ISO 18857–2 for Bisphenol A; DIN
faecal indicators. 38407–35 for caffeine, ibuprofen, salicylic acid and triclosan; US EPA
The following hypothesis were put forward: (1) secondary ef­ 1694 for atenolol, metoprolol, diclofenac, naproxen, paracetamol,
fluents from WWTPs with nutrient removal can be substantially carbamazepine, sulfamethoxazole, trimethoprim, gemfibrozil, and
upgraded using CWs with DWTS as reactive substrate; (2) total furosemide.
phosphorus concentrations around 0.1 mg/l can be reached working DWTS was analysed using standard methods for TP, Al and other
at high hydraulic loading rates using these intensified CWs; (3) other metals (ISO 11885), organic matter (loss on ignition according to
pollutants like CECs and faecal coliforms can be significantly reduced UNE EN 15169) and total organic carbon (TOC) (EN 15936).
in this upgrading treatment.
2.3. Statistical analyses

2. Materials and methods Statistical analyses were performed using Statgraphics XVII
centurion. Three different periods were established according to
2.1. Pilot plant and monitoring description HLR. Influent and effluents (S-CW and C-CW) were compared for WQ
variables, assuming that samples were related because were taken
Three pilot CWs were installed in two WWTPs in Valencia on the same dates. For this, t-student test was used if normality
(Spain). Two of them, called S-CW and C-CW (Fig. S1), were placed in was satisfied and a non-parametric test (W de Mann-Whitney
a large WWTP (129000 PE) which has conventional activated sludge (Wilcoxon)) otherwise. Correlation coefficients were calculated to
followed by tertiary treatment, with nutrient removal and disin­ assess the influence of some variables on the performance of the
fection. The influent to these CWs was the effluent from secondary pilot CWs. Multiple linear regression models (MLRM) were
treatment. The third pilot wetland (CC-CW) was placed in the estimated using the stepwise method. Statistical significance was
WWTP of a small town (100 PE), which consists of two Imhoff tanks indicated by a probability of type I error < 5% (p < 0.05).
in series followed by two horizontal subsurface flow CWs in series.
The pilot systems consisted of containers with a surface 1 m2, filled
2.4. Mathematical modelling
with DWTS from “La Presa” DWTP (Valencia) (37 cm depth;
0.83–19.0 mm grain size), and two layers of coarse gravel (10–11 mm
The mathematical model P-k-C* , based on a tanks-in-series
grain size) at bottom (10 cm depth) and surface (5 cm depth).
model (Dotro et al., 2017), has been calibrated for TP. The objective
Common reed (Phragmites australis) was planted with an initial
was to provide suitable values of the constant rates to be used for
density of 10 stems/m2 in C-CW and S-CW, and 3 stems/m2 in
designing intensified treatment wetlands aimed at phosphorus re­
CC-CW.
moval. The calibration was focused on the saturated systems (C-CW
Two operation modes were tested: continuously saturated flow
and CC-CW) because the hydraulic model of tanks-in-series is suited
(C-CW and CC-CW) and sequential operation (S-CW). The sequential
only for them (Dotro et al., 2017). The P-k-C* approach has the
operation consisted of loading-contact period-draining batches. The
following equation:
draining operation was controlled by a solenoid valve. Three dif­
ferent hydraulic loading rates (HLR) in different periods were tested Cin C *
Cef = C * +
in S-CW and C-CW (Table 1). (1 + k P ) P (1)
The performance of pilot CWs was monitored for one year and a
half. Grab samples (2 l) of the influent and the effluent from each Where Cef is the effluent concentration (mg P/l) to be achieved, Cin
pilot CW were collected for physical-chemical analysis. Dissolved the influent concentration (mg P/l), C* is the background con­
oxygen (DO) was measured in situ, with a Hach HQ40D portable centration, k is the first-order reaction coefficient (h−1), τ is the
multiparameter. Electrical conductivity and pH were measured with theoretical retention time (h), and P is the apparent number of
a Hach440d multiparameter. The samples were transported to the tanks-in-series (dimensionless). For its calibration, data were di­
laboratory in an icebox and kept in refrigerator (4ºC) until chemical vided into groups with the following ranges of Cin: < 1, 1–2, 2–3.5,
analysis, within 24 h. 3.5–9 mg P/l. The Cin ranges below 3.5 mg P/l correspond to C-CW

487
C. Hernández-Crespo, N. Oliver, M. Peña et al. Process Safety and Environmental Protection 158 (2022) 486–494

Table 1
Mean ± standard deviations are indicated for HLR (hydraulic loading rate) and HRT (hydraulic retention time), and range of variation for temperature (T) for different study
periods. N is the number of data for each period, for small WWTP the monitoring period was extended only for TP, which N is in parenthesis.

WWTP Large Small

Variable Period 1 (Apr – Aug 2019) Period 2 (Sep – Dec 2019) Period 3 (Dec 2019 – Sep 2020) (Nov 2019 – Feb 2021)

Abbreviature S-CW C-CW S-CW C-CW S-CW C-CW CC-CW

HLR (m3/m2/d) 3.4 ± 0.3 2.8 ± 0.7 5.1 ± 0.8 4.7 ± 1.6 0.9 ± 0.1 1.3 ± 0.6 2.6 ± 1.5
HRT (h) 0.9 ± 0.1 1.8 ± 0.7 0.5 ± 0.1 1.3 ± 1.3 2.4 ± 0.0 4.6 ± 3.0 2.3 ± 1.3
T (ºC) range 17.6–28.9 17.1–28.7 16.8–27.8 15.6–26.0 11.6–29.5 14.4–30.0 5.7 –25.9
N 56 52 23 21 37 32 8 (12)

and the range 3.5–9 mg P/l correspond to CC-CW system. Then the or a partially saturated system would allow the optimization of
data of Cef were fitted to Ec. 1 by varying the value of k, seeking for nitrogen removal.
the minimum root-mean-square error (RMSE) between the simu­ In the case of CC-CW, the low removal efficiencies reached could
lated and experimental data. The value of P was fixed in constant be related with the much lower vegetation development in com­
values (P = 3) and C* was set in 0.015 mg P/l, the half of the lowest parison to C-CW.
value measured in the full study period.

3.1.2. Phosphorus removal efficiencies and saturation degree

3. Results and discussion All three pilot CWs significantly reduced the concentration of
total phosphorus (p < 0.05) (Fig. 1 and Table 3). The CWs treating
3.1. Performance as upgrading treatment system low concentrations of TP reached average effluent concentrations
ranging from 0.18 to 0.44 mg P/l in S-CW and from 0.08 to 0.25 mg P/l
3.1.1. Organic matter, total suspended solids, pH, dissolved oxygen and in C-CW. C-CW removed TP more efficiently than the S-CW in all
nitrogen three periods (p < 0.05). The best performances were found during
COD was significantly reduced in the three pilot systems period 1, when efficiency mean values of 66% and 86% for the S-CW
(p < 0.05). It decreased from values between 26 and 38 mg/l to va­ and C-CW were reached, working at very high HLR (around
lues around 20 mg/l (Table 2) for S-CW and C-CW, and from a mean 3 m3/m2/d).
value of 40 mg/l to an average of 31 mg/l in CC-CW (Table 3). All After these good results, it was decided to raise the HLR
systems contributed to a further reduction of BOD5 as well (5 m3/m2/d) to test the boundaries of the pilot systems. However, in
(p < 0.05). The effluents presented very low values, around 4 mg/l, this second period both CWs decreased their efficiencies, with
which is very beneficial for the receiving water bodies. TSS were also average values of 32% (S-CW) and 56% (C-CW). Probably HLR was too
significantly improved by systems S-CW and C-CW (p < 0.05), pro­ high and, consequently, HRT was too short to achieve a good effi­
ducing an effluent with less than 5 mg/l and high transparency. ciency of the sorption process. In addition, the influent concentra­
Conversely, CC-CW did not achieve a further decrease of TSS, as they tion was lower than in the other periods, which could also have
reached 5 mg/l in the influent. contributed to the lower efficiency, because lower influent con­
The second period presented a lower efficiency than the others, centration implies lower efficiency (Martín et al., 2013). In fact, the
likely due to the extremely high HLR. Contrariwise to other para­ influent concentration (Cin) influenced positively on mass removal
meters discussed later, non-significant differences were found be­ rate (Fig. S4). Temperature can also influence the efficiency of the
tween the operation modes for COD, BOD5 or TSS (p > 0.05). In the process; however, it was not significantly different among the
case of organic matter, the more aerobic conditions in S-CW could periods.
have compensated the shorter HRT, thus achieving removal effi­ During the third period, the HLR was reduced to 1 m3/m2/d,
ciencies like those of C-CW. For TSS, the filtration process would giving rise to an enhancement of the efficiencies, 45% for S-CW and
have been equally efficient in both systems due to the same char­ 69% for C-CW. However, these efficiencies were lower than those
acteristics of the filter media. Interestingly, the effluents generally registered in the first period, despite the HLR reduction. A possible
presented less standard deviation than the influent (Tables 2 and 3), reason could be the accumulation of organic matter inside the sys­
highlighting the buffer capacity of CWs. tems. On several occasions, surface clogging symptoms were ob­
The pH did not vary significantly from the influent to effluent served after cleaning operations of the secondary clarifiers of the
(Table 2), in contrast to other adsorbent materials that cause sig­ WWTP. This organic matter could have released phosphorus during
nificant pH rises (Dotro et al., 2017). Dissolved oxygen (DO) in the its biodegradation. Moreover, it could have hindered the diffusion of
effluent from C-CW was significantly lower than the influent and S- phosphorus to the sorbent surface.
CW effluent (p < 0.05) (Table 2), because of the lower reaeration of In general, orthophosphates presented a trend similar to TP. For
C-CW. both variables, C-CW presented higher efficiencies in all periods.
The ammonium concentration was significantly reduced in sys­ Despite working at relatively similar HLR, the HRT was significantly
tems S-CW and C-CW (p < 0.05), from influent values around lower in the S-CW because the draining stage spends time, which is
4 mg N/l to average effluent concentrations of 0.8 and 2.6 mg N/l in only partially considered as HRT. Overall, the C-CW was more
S-CW and C-CW, respectively (Table 2), which is very beneficial for resilient to HLR and Cin changes.
the receiving aquatic environment, especially in areas under water Regarding the buffer capacity, the variability of influent is re­
stress and droughts. The removal efficiency was significantly higher duced in the upgrading treatment in the pilot CWs. During the third
in the S-CW, because of its more aerobic conditions (Table 2). Con­ period some peak concentrations in the influent were indeed sub­
versely, total nitrogen and nitrates were only significantly reduced in stantially reduced in both pilot systems (Fig. 1), with removal
system C-CW (p < 0.05). In this system, aerobic and anoxic en­ efficiencies around 65%.
vironments coexisted, allowing nitrification and denitrification The CW prototype treating higher phosphorus concentration
processes, while in S-CW the denitrification could have been in­ (CC-CW) presented a mean removal efficiency of 62%, significantly
hibited due to higher DO concentrations. Based on these results, a reducing the TP concentration from 6.15 to 2.36 mg/l (p < 0.05)
combination of both operation modes to create a two-stage system (Table 3). More interestingly, just with 1 m2, CC-CW treated 17% of

488
C. Hernández-Crespo, N. Oliver, M. Peña et al. Process Safety and Environmental Protection 158 (2022) 486–494

Table 3

23.42 ± 6.43(38%)
Mean concentration (mg/l) ± standard deviation and removal efficiency (R.E.) of the

2.49 ± 2.53(39%)
8.56 ± 6.95(38%)
2.71 ± 4.00(45%)

3.47 ± 1.55(70%)
4.11 ± 2.58(65%)
pilot CC- for the full period. N and P forms in mg N/l and mg P/l.

2.71 ± 1.20
Influent Effluent R.E. (%)
Ef. C-CW

7.4 ± 0.2
TP 6.15 ± 2.77 2.36 ± 1.21 62%
PO43- 5.89 ± 2.58 2.12 ± 1.07 62%
TN 48.06 ± 19.02 44.06 ± 13.41 8%
NH4+ 37.34 ± 15.45 34.19 ± 16.23 8%
NO2- 0.21 ± 0.53 0.36 ± 0.38 -68%
9.39 ± 4.44(−132%)

NO3- 1.17 ± 2.59 2.47 ± 3.33 -111%

21.90 ± 4.61(42%)

8.27 ± 10.74(30%)
0.87 ± 1.43 (82%)
13.57 ± 6.73 (1%)

4.88 ± 1.83(58%)

COD 40 ± 11 31 ± 6 23%
BOD5 6 ±9 3±6 40%
Mean concentration (mg/l) ± standard deviation and removal efficiencies (%) of S-CW and C-CW for each study period. N and P forms in mg N/l and mg P/l, respectively. Inf: influent, Ef: effluent.

3.48 ± 1.11

TSS 5.38 ± 1.99 5.75 ± 2.91 -7%

Ef. S-CW

7.4 ± 0.2

the total WWTP’s flow, and had a noticeable effect on its final ef­
fluent, which decreased from 6.12 mg/l of TP (average of 18 months
37.67 ± 10.41

before) to an average of 5.2 mg/l during the research period. Thus,

13.77 ± 8.04

11.59 ± 4.54
11.86 ± 4.99
4.05 ± 2.30
4.93 ± 4.57

4.17 ± 1.47

the full-scale CW that receives the effluent of the CC-CW increased

7.4 ± 0.3
Period 3

its TP removal efficiency from 8% to 16%.

Inf.

The results obtained in this study agree with earlier research.

Babatunde et al. (2010) found very high removal efficiencies in the
first of a four-stage treatment, but lower efficiencies in the sub­
19.20 ± 3.38(26%)

sequent treatment stages. Under a HLR 1.27 m3/m2/d and an influent

2.66 ± 3.24(40%)
8.29 ± 2.40(23%)

1.76 ± 0.90(59%)
4.43 ± 1.13(44%)
4.99 ± 2.39(8%)

concentration of 21 mg P/l (SRP), the treatment reached a removal

1.57 ± 0.30

efficiency of 89% and an effluent concentration of 2.3 mg P/l in the

Ef. C-CW

7.4 ± 0.2

first stage, and subsequent reductions to 1.2 mg P/l (48%), 0.86 mg P/l
(28%) and 0.47 mg P/l (45%) in the second, third and fourth stages,
respectively.
Mass removal rates (MRR) varied between −0.17 and 5.0 g/m2/d
in the systems treating low concentrations. On average, the mass
10.60 ± 4.33(−96%)
18.94 ± 4.46(27%)
10.44 ± 2.05 (4%)
1.33 ± 2.69 (70%)

4.57 ± 1.27(42%)
3.11 ± 6.52(27%)

loading rate (MLR) was reduced by 59% in S-CW and by 69% in C-CW
(Fig. S2). The obtained multiple linear regression models (MLRM)
3.06 ± 1.22

showed that MRR depended on Cin, HLR and temperature (Figs. S4

Ef. S-CW

7.4 ± 0.2

and S5).
CC-CW reached a removal rate of 14.7 g/m2/d when the HLR was
3.3 m3/m2/d, on average the MLR was removed by 58%.
The cumulative phosphorus removed during the whole study
period, 285, 396 and 3013 g/m2 in S-CW, C-CW and CC-CW respec­
26.06 ± 3.79
10.82 ± 2.07
4.44 ± 3.64
5.42 ± 2.46

4.28 ± 1.27
3.21 ± 1.29
7.88 ± 1.46

7.4 ± 0.2

tively, was estimated by interpolating linearly for the periods be­

Period 2

tween samplings (Fig. S3). These data represent a 7%, 10% and 111% of
Inf.

the sludge adsorption capacity, respectively, calculated from the

mass of DWTS inside the pilot systems and its maximum sorption
capacity, which was calculated experimentally applying the Lang­
20.81 ± 4.49(36%)
5.95 ± 2.23 (33%)
2.38 ± 1.60 (27%)

muir model (Table S1). Phosphorus can be bound to the DWTS

2.66 ± 1.54(32%)

2.27 ± 1.29(69%)
5.21 ± 2.53(51%)

through adsorption, chemical precipitation with iron, aluminium, or

1.89 ± 0.47

calcium, and coprecipitation with calcium carbonate (Babatunde and

Ef. C-CW

7.5 ± 0.2

Zhao, 2009). Considering the average removal rates of 0.53 and

0.73 g/m2/d (S-CW and C-CW respectively), they could keep working
for more than ten years. Remarkably, CC-CW exceeded the maximum
adsorption capacity and, even though the efficiency has decreased to
32% in the last three samplings, it continues removing phosphorus.
8.11 ± 2.68(−108%)
20.91 ± 4.52(36%)
0.29 ± 0.45(91%)

2.72 ± 1.50(63%)
5.21 ± 2.97(51%)

Zhao et al. (2008) also noticed that the mass of P immobilized ex­
8.70 ± 2.73(2%)

ceeded the maximum adsorption capacity.

3.20 ± 1.38
Ef. S-CW

7.4 ± 0.2

Compared with other materials, the efficiencies achieved in the

present study can be considered high, considering the high HLR
tested. Martín et al. (2013) obtained a mean efficiency of 75% for TP
in a full-scale vertical flow CW filled with a mixture of sand and iron
oxides but working with HRT of 24 h and Vera et al. (2014) obtained
32.59 ± 8.30
10.70 ± 3.97

a removal efficiency around 70% using zeolites with HRT of 3–4 days.
3.28 ± 2.43
8.90 ± 3.04

2.54 ± 0.75
7.26 ± 5.85
3.89 ± 1.34

7.5 ± 0.2
Period 1

If non-reactive materials, such as gravel, are used, typical efficiencies

of 10–20% are obtained (Dotro et al., 2017).
Inf.

In summary, the results obtained indicate that it is possible to

work with high HLR without giving up the system efficiency. This
allows upgrading large flows with relatively small areas. The effi­
Table 2

BOD5
NH4+
NO3-
COD

ciency can be enhanced enlarging the depth of DWTS, to increase the

TSS
DO
pH
TN

HRT, or working in multi-stage systems.

489
C. Hernández-Crespo, N. Oliver, M. Peña et al. Process Safety and Environmental Protection 158 (2022) 486–494

Fig. 1. Box-and-whisker plot of TP (left) and ortophosphates (right) in the three studied periods. Cin is the influent concentration, Cef_S and Cef_C are the effluent concentrations
from the sequential and continuous flow CWs. aindicates significant difference between Cin and Cef, bmeans significant difference between Cef_S and Cef_C.

3.1.3. Aluminium effluents. It usually presents very high removal efficiencies, indeed
The concentration of aluminium in DWTS is very high, because of Verlicchi et al. (2012) reported an average removal efficiency of 93%
the coagulant used in the water treatment. In this study the content in conventional activated sludge, so it was probably removed in the
of Al in the DWTS was on average 34.8% (Table S2). Other authors previous treatment.
measured Al contents varying between 6.1% and 44.05% (Hou et al., Regarding the studied antibiotics, trimethoprim was reduced in
2018). Looking at the values compiled by those authors, it appears to both wetland systems, more significantly in S-CW (43%). The effi­
exist a certain positive correlation between Al content and max­ ciencies are within the range of variation reported by Ilyas and van
imum adsorption capacity. Hullebusch (2020). In contrast, sulfamethoxazole increased sig­
Given this aluminium high content, it is important to check that nificantly, about a 42% (p < 0.05). Previous studies have also re­
it is not leached from the substrate to the treated water. During the ported variable and mainly poor, even negative, removal efficiency
whole study period, the Al concentration in the effluent was below for sulfamethoxazole (Auvinen et al., 2017). They hypothesized that
0.2 mg/l, value established for human consumption (Directive 98/83/ it could be due to the potential retransformation of the metabolite to
EC). Indeed, it was mostly below the detection limit (0.05 mg/l). the parent compound.
These results agree with previous studies and corroborates that Among beta-blockers, metoprolol was undetectable neither in
DWTS can be safely used as reactive media in CWs (Zhao et al., 2008; the influent nor in the effluents. Whereas, atenolol was significantly
Babatunde et al., 2011). Moreover, the aluminium content in reed reduced in both pilot systems (p < 0.05), reaching similar mean
was measured in a previous microcosm experiment where two concentrations of 0.122 µg/l (S-CW) and 0.108 µg/l (C-CW). Sorption
substrates were compared (gravel and DWTS). The results showed is recognized as the dominant removal mechanism for this CEC
that reed growing on DWTS presented an Al content (48 and 11 mg/ (Table 4). This could explain the fact that both pilot systems achieved
kg dry weigh (d.w.) in dead and fresh stems respectively) similar to efficiencies statistically non different (p > 0.05).
reed growing on gravel (52 and 34 mg/kg d.w. respectively) The diuretic furosemide presented negative removal efficiencies.
(Naranjo-Ríos et al., 2018). The removal efficiency reported in literature is widely variable, from
values of 35% (Verlicchi et al., 2013) to 80–96% (Chen et al., 2016) or
3.1.4. Contaminants of emerging concern (CECs) −10.5–98.8% (Vymazal et al., 2017). Hydrolysis and photolysis are
In this study, the concentrations of CECs in the influent are low usually considered the major removal pathways in water environ­
because it comes from the secondary treatment of the WWTP. Most ment, and hydrolysis in subsurface flow CWs (Chen et al., 2016;
CECs were below 1 µg/l, except bisphenol A, which reached a mean Vymazal et al., 2017). Olvera-Vargas et al. (2016) demonstrated mi­
concentration of 1.301 µg/l (Table 4). These low concentrations were crobial transformation of furosemide and identified three different
further reduced in the pilot systems for most CECs. A significant metabolites, two of which exhibited higher toxicity than furosemide.
decrease was achieved in the S-CW for seven of the CECs analysed, Nevertheless, CECs with chlorine in their molecular structure are
while four CECs were significantly removed in the C-CW (p < 0.05). considered recalcitrant and their removal efficiency is inversely
Hence, the results indicate that the more aerobic conditions in the S- proportional to the molecular weight (Ilyas et al., 2020). Furosemide
CW favoured a more efficient removal of CECs, with significant re­ has a high molecular weight (330.7 g/mol), which could explain the
moval efficiencies varying between 11% (carbamazepine) and 85% low removal efficiencies. Other authors obtained good removal ef­
(bisphenol A). This positive effect of aerobic conditions agrees with ficiencies in adsorption experiments using light expanded clay ag­
previous studies (Kahl et al., 2017; Ilyas and van Hullebusch, 2020). gregates (LECA) and cork granulates as sorbent materials (Machado
Matamoros et al. (2008) measured similar levels of analgesics in et al., 2017). It is difficult to explain the increase observed in the pilot
a secondary treatment effluent, which were further reduced in a CWs, feasible hypotheses could be the retransformation of meta­
surface flow CW with moderate to high efficiencies (carbamazepine: bolites to the parent compound or desorption of previously adsorbed
0.37 µg/l, 30–47%; ibuprofen: 0.04 µg/l, 95–96%; naproxen: 0.34 µg/l, contaminant, if the substrate was rapidly saturated. The monitoring
52–92%; diclofenac: 1.25 µg/l, 73–96%). Their efficiencies were con­ of CECs started two months after the operation onset and the sub­
siderably higher than those found in this study, probably due to strate could have been saturated during this period. In this sense, the
higher HRT (1 month). In addition, the photooxidation process, valorisation of exhausted activated carbon, which is another waste
feasible in surface flow CW, is not active in subsurface flow CW. In produced in DWTPs, to generate a mixed reactive media (DWS +
this sense, removal efficiencies are highly variable among different activated carbon) would enhance the removal efficiency of this and
studies. For instance, Ilyas and van Hullebusch (2020) compiled other CECs.
mean efficiencies of 53%, 39% and 63% for the analgesics ibuprofen, Carbamazepine was reduced with a low but significant percen­
diclofenac and naproxen, respectively, with relative standard de­ tage yield in both systems (16% in C-CW; 11% in S-CW) (p < 0.05).
viations varying between 40% and 60%. In this study, paracetamol The recalcitrance of this compound is well documented (Kahl et al.,
was under the detection limit in most cases, both in influent and 2017; Delgado et al., 2020; Ilyas and van Hullebusch, 2020) and

490
C. Hernández-Crespo, N. Oliver, M. Peña et al. Process Safety and Environmental Protection 158 (2022) 486–494

Mean concentrations ( ± standard deviation) of CECs (µg/l) in influent and effluent of CW pilot systems, from June 2019 to June 2020 (N = 12). RE: removal efficiency. The arrows indicate the direction of the variation, double arrows in bold
indicate significant variation (p < 0.05). The detection limit is assigned when the concentration is below it. 1Based on Ilyas and van Hullebusch (2020); 2(Chen et al., 2016; Vymazal et al., 2017); 3Olvera-Vargas et al. (2016); 4Machado et al.
Delgado et al. (2020) found that its removal was mainly linked to

Hydrolisis, photolysis2, Biodegradation3, Adsorption4

plant absorption.

Biodegradation (aerobic) and photodegradation1

The concentration of the lipid regulator gemfibrozil decreased a
35% in S-CW and 3% in C-CW. It agrees with previous studies where
the removal is mainly attributed to aerobic degradation (Ilyas and

Biodegradation (aerobic), Plant uptake1

van Hullebusch, 2020).

Biodegradation (aerobic, anaerobic) 1

Adsorption, sorption, plant uptake1

Salicylic acid was substantially reduced in both pilot systems
(p < 0.05), from 0.452 µg/l in the influent to 0.271 µg/l in the effluent

Biodegradation and sorption2

Biodegradation (anaerobic) 1
of S-CW and 0.187 µg/l in C-CW, which achieved a more significant

Biodegradation (aerobic) 1

Biodegradation (aerobic) 1

Biodegradation (aerobic) 1
Biodegradation (aerobic) 1
Main removal mechanism

Biodegradation (aerobic)1
decrease (p < 0.05). The main removal pathway reported in the lit­
erature is aerobic biodegradation (Table 4). Plant uptake could be
also possible because of its low molecular weight joined to its high-

Biodegradation5
Biodegradation5
water solubility and hydrophilic nature (Ilyas et al., 2020). In this

Sorption1
study, it could be more efficiently removed in the C-CW due to its
higher HRT.
Caffeine was reduced in both pilot CWs with similar and high
efficiencies (56% in C-CW and 58% in S-CW), in line with other stu­

↓↓

↑↑

↓↓

↓↓

↓↓
dies where caffeine is described as a readily biodegradable com­

↓
↓

↓
↓
↓
pound (Ilyas and van Hullebusch, 2020).

RE (%)
The antiseptic triclosan was below the detection limit in all

-43%

-20%
44%
30%

59%
16%
29%

58%

63%
13%
3%

3%
samples. It could have been removed in the previous activated
sludge process, as high removal efficiencies are usually found

0.062

0.050

0.023

0.024
0.339

0.759
0.056

0.045

0.271
0.333

0.134
(Verlicchi et al., 2012). Ávila et al. (2015) reported an average re­

0.113
moval efficiency of 61% in a full-scale vertical flow CW, attributing
the high efficiency to the prevailing aerobic conditions, which was

C-CW effluent

±
±
±

±
±
±

±
±
±
±
±

±
further increased to 79% in subsequent wetlands (horizontal flow +
free water surface flow), decreasing the concentration from 0.15 to

< 0.05

< 0.40
< 0.10

0.088
0.064

0.648

0.226

0.483
0.144

0.108
0.514
0.153

0.187
0.127
0.118
0.03 µg/l. These authors also demonstrated a decrease of the treat­
ment performance when HLR was increased (Ávila et al., 2014).
Bisphenol A was efficiently reduced in both pilot CWs, more
↓↓
↓↓
↓↓

↑↑

↓↓

↓↓

↓↓

↓↓
↓↓

↓
significantly in S-CW (85%) (Table 4). High removal efficiencies have
been found for this CEC, especially in aerobic conditions, and not
RE (%)

-45%

-76%
dependent on HLR by Ávila et al., (2014, 2015). This high biode­
28%
35%
46%

40%

85%
37%

11%
43%

35%

56%
gradability would explain the good efficiencies obtained in this
study despite the high HLR applied to the pilot systems.
0.038

0.029

0.052

0.026
0.004

0.048
0.221

0.021
0.095
0.411

0.161
1.314
Interestingly, it was observed that CEC concentrations in influent
and effluents were negatively correlated with temperature (Table
S-CW effluent

S3). This correlation can be explained because most processes im­

±
±
±

±
±
±

±
±
±
±
±

±
plied in CEC removal depend on temperature (biodegradation, ad­
< 0.05

< 0.40
sorption, photolysis, hydrolysis). Additionally, another influencing
< 0.10
0.342

0.093
0.953

0.271
0.051
0.126

0.120

0.122

0.195
0.123
0.151
0.110

factor could be the consume pattern (Chen et al., 2016), as it seems

reasonable that some of the pharmaceuticals are more consumed in
0.007

3.656
0.359

0.632
0.438
0.029
0.055

winter. In this sense, the monitoring for periods of one year, or

0.457
0.072

0.057

0.016
0.167
0.102

longer, are essential to allow a robust evaluation of the dependence

of removal performance on operational conditions (Kahl et al., 2017).
±
±
±
±
±
±
±

±
±
±
±
±

These authors also found a dependence of removal efficiencies on

temperature.
Influent

< 0.05

< 0.40
0.529

0.233
0.452
0.205

0.083

0.540
0.090

0.291

1.301
0.193
0.102

0.105
0.176

Finally, the environmental risk was assessed through the calcu­

lation of the risk quotient (Fig. 2), according to Chen et al. (2016). In
general, CECs levels presented low environmental risk in the in­
Sulfamethoxazole

fluent. For eleven of the monitored CECs the Risk Quotient (RQ) was
Carbamazepine
Trimethoprim

Salicylic acid

below 0.1, indicating low environmental risk. Three CECs (gemfi­

Paracetamol

Bisphenol A
Gemfibrozil
Furosemide
Metoprolol
Diclofenac
Ibuprofen

Naproxen

brozil, salicylic acid, and bisphenol A) presented a medium en­

Triclosan
Atenolol

Caffeine

vironmental risk in the secondary treatment effluent (RQ between

CEC

0.1 and 1), which was considerably reduced in the wetland pilot
systems. Only sulfamethoxazole presented a high-risk level (RQ > 1),
Analgesical / anti-inflammatories

despite its concentration was relatively low (0.1 µg/l), but the pre­
(2017); 5Ávila et al., (2014, 2015).

dicted no-effect concentration (PNEC) is even much lower (0.027 µg/

l). This highlights the importance of bearing in mind the toxicity and
non-toxicity levels of each substance, to evaluate the risk posed by
Endocrine disrupter

their presence in treated wastewater.

Therapeutic class

Psychiatric drug
Lipid regulators

Stimulant drug
Beta-blockers

Antiseptics
Keratolytic
Antibiotics

Diuretics

3.1.5. Microbiological indicators

Table 4

The main microbiological indicators considered in Spanish and

European regulation on water reuse, Escherichia coli and helminth

491
C. Hernández-Crespo, N. Oliver, M. Peña et al. Process Safety and Environmental Protection 158 (2022) 486–494

Fig. 3. Escherichia coli (log CFU/100 ml) measured in the influent and effluents of the
CW pilot systems. Temperature is displayed in the secondary y-axis.

The values obtained for k were 1.0, 0.8, 0.6 and 0.6 h−1 for the
Fig. 2. Environmental risk assessment by CECs monitored. Risk quotient (RQ) is the ranges < 1, 1–2, 2–3.5, 3.5–9 respectively, thus the value of k de­
ratio between the measured concentration in each site and the predicted no-effect creases as the Cin increases. This trend was also reported for BOD5
concentration (PNEC). PNEC values have been taken from Ilyas et al. (2020), Isidori (Dotro et al., 2017), which means that when Cin is high more HRT is
et al. (2006) and ECHA website. needed for reaching the same Cef. The RMSE values obtained are
shown in Table S4. The k values resulted far higher than those listed
by Rousseau et al. (2004) (0.14–0.28 d−1) for gravel beds, because of
eggs, were monitored. Helminth eggs were not detected neither in
the greater adsorption capacity of DWTS, and also higher than those
the influent nor the effluents, in any sampling campaign.
obtained for other sorbent materials (0.26 h−1, Delgado-González
Regarding E. coli, a significant decrease was obtained in both pilot
et al., 2021).
systems (Fig. 3), from an average concentration of 2.2·105 CFU/
Once obtained the k values, the equation for estimating the
100 ml to 3.9·104 and 1.6·104 CFU/100 ml in S-CW and C-CW, re­
necessary area is:
spectively. The mean efficiencies were 84.4% (S-CW) and 94.1% (C-
CW). Such difference could be related with the higher retention time 1/ P
PQ Cin C*
in C-CW or the operation mode, maybe more predator organisms can A= 1
k h Cef C*
develop inside the continuously saturated CW. Other studies have (2)
also reported better efficiencies in horizontal flow CWs (99.7%) re­ 2
spect to vertical flow CWs (82.0%) (Ávila et al., 2015). Although it Where A is the area (m ) and ϕ is the porosity (Dotro et al., 2017). By
should be mentioned that the studies are not directly comparable way of example, for a town with a population of approximately 1000
because in that study the CWs worked in series (VFCW+HFCW) and inhabitants, with an average daily flow of 150 m3/d and a Cin equal to
their HLR was much lower. 6 mg P/l, the necessary area to reach a Cef equal to 0.5 mg P/l would
C-CW efficiency presented a strong correlation with temperature be 234 m2. A wetland water depth of 0.5 m, a porosity of 0.35, P
(r = 0.59, p < 0.05) and HLR (r = −0.62, p < 0.05), so high tempera­ equal to 3, C* equal to 0.015 mg P/l and the k corresponding to this
tures and low HLR favoured the removal of E. coli. In contrast, S-CW Cin (0.6 h−1) have been assumed. Additionally, from a management
efficiency did not correlate with these variables (r = 0.25, p > 0.05 for point of view, the lifespan of the intensified CW should be estimated.
T; r = 0.07, p > 0.05 for HLR). For both systems, the influent con­ For this, the quantity of DWTS inside the CW can be calculated as
centration strongly influenced the effluent concentration (r = 0.78, follows. The density of the material is around 750 kg dw/m3. The
p < 0.05 for C-CW; r = 0.73, p < 0.05 for S-CW). The relation between total volume of the CW would be 117 m3, thus the mass of DWTS
E. coli and TSS was also evaluated to assess if TSS could be used as would be 82 tons. Multiplying the mass per the maximum adsorp­
indicator of faecal contamination, but the correlation was very weak. tion capacity (14.5 g P/kg DWTS), the total quantity of TP to be re­
On an average basis the effluents would not meet the require­ moved is 1188 kg P. Applying a safeguard of 25% because the
ments for any use foreseen in the European Regulation on water
reuse (Regulation (EU)/741, 2020). Nevertheless, it should be noted
that in most cases (81% of samples) the effluent from the C-CW
accomplished the requirements for the use called D (crop category:
industrial, energy and seeded crops). The compliance rate in case of
S-CW was lower (43%), which could be enhanced by decreasing HLR.
The addition of a subsequent free-water surface CWs would help to
further reduce the concentration of E. coli and could provide an
effluent suitable for several irrigation uses (Ávila et al., 2015).

3.2. Design criteria for phosphorus removal and water quality

upgrading

The P-k-C* model has been calibrated as a tool for wetland de­
sign. This model was selected because the influence of the HRT was
observed, so that the hypothesis is that the sorption of phosphorus
on this material follows a first-order kinetics (Fig. 4). Fig. 4. Calibration of the P-k-C* model for total phosphorus.

492
C. Hernández-Crespo, N. Oliver, M. Peña et al. Process Safety and Environmental Protection 158 (2022) 486–494

efficiency decreases as the DWTS becomes saturated, the mass of P • Continuous flow operation performs better for TP, nitrates, TN
would be 890 kg P. This would allow treating the flow for approxi­ and pathogens, whereas a sequential setup is more efficient for
mately 3 years. In addition, as the efficiency decreases as the ma­ ammonium and CECs removal. Therefore, a hybrid system, con­
terial becomes saturated, it is advisable to propose a design based on sisting of a sequential layer on the top and a continuously satu­
a sequence of cells in series, in order to make maximum use of the rated bottom layer, could optimize the upgrading treatment.
sorption capacity of the DWTS and renew them as they become • The combination of DWTS with exhausted activated carbon, an­
completely saturated. other waste produced in DWTPs, could enhance the efficiency for
CECs removal.
3.3. Proposal for the full valorisation of sludge produced at “La • Our results confirm previous literature statements, DWTS is
Presa” DWTP presented as a low-cost adsorbent substrate with good effi­
ciencies despite the variability in its composition, usually found
“La Presa” DWTP (Manises, Valencia), with a production capacity in waste materials. Therefore, it should be valorised whenever
of 0.15 Hm3/d, supplies drinking water to a population of 0.8 million possible to integrate the principles of circular economy in the
inhabitants. The purification process produces between 3 and 4 m3/d urban water cycle. A widespread implementation of systems for
of sludge cake, which is equivalent to approximately 1670 tons/year, upgrading treated wastewater, based on wetlands with reactive
with an average content of 26% of dry matter. This dryness must be media, can contribute to significantly improve the water quality
increased to 90% to proceed with grinding. In the grinding process to of the receiving environments, thus helping to meet the objec­
obtain the desired granulometry, there were losses of 15% of the tives of the Water Framework Directive, and making them more
processed material, because the size was too fine or too coarse. resilient to the effects of climate change. This is especially im­
Considering these losses, a production of 370 tons/year of dry portant for removing emerging pollutants such as those eval­
granulated sludge is estimated. uated in this study, but also other like microplastics or organic UV
Assuming that the annual production could be stored, a selection filters.
of towns near the DWTP has been made with the aim of proposing a • These low cost, effective and environmentally friendly technol­
full valorisation of the sludge annually generated. The following ogies are highly recommended for upgrading treatments.
criteria have been applied: population less than 2000 inhabitants,
WWTP without phosphorus removal, distance from DWTP less than Declaration of Competing Interest
100 km, proximity between towns, and discharge of their treated
effluents in the Turia river basin. This last criterion has been con­ The authors declare that they have no known competing fi­
sidered since upgrading the wastewater discharged in the Turia river nancial interests or personal relationships that could have appeared
basin will contribute to improve its water quality, which in turn can to influence the work reported in this paper.
benefit the purification process, since it is an important water source
for the DWTP. Acknowledgements
Another variable to be fixed is the lifespan of the upgrading CWs.
As a premise, it was set for at least 5 years. Further considering as­ The authors thank the partial funding by the Generalitat
pects such as transport costs and associated emissions, it was con­ Valenciana-IVACE (Valencian Institute of Business Competitiveness)
cluded that 7 years provided the best combination of groups and and the European Regional Development Fund (through the ERDF
individual municipalities to valorise the annual production of dry Operational Program for the Valencian Community 2014–2020). The
granulated sludge, so that every 7 years they would have to change authors also thank the Public Entity for Wastewater Sanitation of the
the sludge-based substrate in the upgrading CWs (Fig. S6). The ne­ Valencian Community (EPSAR), attached to the Conselleria de
cessary mass of sludge to satisfy this life span was calculated from Agricultura, Desarrollo Rural, Emergencia Climática y Transición
the maximum adsorption capacity determined and the phosphorus Ecológica. As well as the Cathedra “Aguas de Valencia- Universitat
emitted by the towns proposed. This group of municipalities has a Politècnica de València”, Department of Hydraulic Engineering and
total population near 10,000 inhabitants and produces a total was­ the Environment.
tewater flow of 4800 m3/d. Therefore, a relation between drinking
water produced and treated wastewater to be upgraded can be es­ Appendix A. Supporting information
timated, giving a ratio of 32 litres upgraded wastewater per cubic
meter of drinking water produced. Supplementary data associated with this article can be found in
Once exhausted, the sludge could be used in agriculture as it the online version at doi:10.1016/j.psep.2021.12.035.
meets the requirements set out in the European Directive on the
protection of the environment and soil, when sewage sludge is used References
in agriculture (86/278/EEC) (Table S2).
Auvinen, H., Havran, I., Hubau, L., Vanseveren, L., Gebhardt, W., Linnemann, V., Van
Oirschot, D., Du Laing, G., Rousseau, D.P.L., 2017. Removal of pharmaceuticals by a
4. Conclusions pilot aerated sub-surface flow constructed wetland treating municipal and hos­
pital wastewater. Ecol. Eng. 100, 157–164. https://doi.org/10.1016/j.ecoleng.2016.
The main conclusions are: 12.031
Ávila, C., Matamoros, V., Reyes-Contreras, C., Piña, B., Casado, M., Mita, L., Rivetti, C.,
Barata, C., García, J., Bayona, J.M., 2014. Attenuation of emerging organic con­
• DWTS is a good substrate for upgrading treated urban waste­ taminants in a hybrid constructed wetland system under different hydraulic
water, being able to produce a significant enhancement of the loading rates and their associated toxicological effects in wastewater. Sci. Total
Environ. 470–471, 1272–1280. https://doi.org/10.1016/j.scitotenv.2013.10.065
effluent even under high hydraulic loading rates.
Ávila, C., Bayona, J.M., Martín, I., Salas, J.J., García, J., 2015. Emerging organic con­
• Effluent concentration depends on operation conditions. A level taminant removal in a full-scale hybrid constructed wetland system for waste­
of TP around 0.1 mg/l is achievable working with a continuous water treatment and reuse. Ecol. Eng. 80, 108–116. https://doi.org/10.1016/j.
ecoleng.2014.07.056
flow and 3 m3/m2/d if influent concentration is around 0.5 mg/l.
Babatunde, A.O., Zhao, Y.Q., 2009. Forms, patterns and extractability of phosphorus
When the influent concentration is around 6 mg P/l, an effluent retained in alum sludge used as substrate in laboratory-scale constructed wetland
concentration of 0.5 mg/l can be reached working at 0.8 m3/m2/d. systems. Chem. Eng. J. 152 (1), 8–13. https://doi.org/10.1016/j.cej.2009.03.020
Babatunde, A.O., Zhao, Y.Q., Zhao, X.H., 2010. Alum sludge-based constructed wetland
• Novel values for the first-order reaction coefficient are provided
system for enhanced removal of P and OM from wastewater: concept, design and
for TP removal.

493
C. Hernández-Crespo, N. Oliver, M. Peña et al. Process Safety and Environmental Protection 158 (2022) 486–494

performance analysis. Bioresour. Technol. 101 (16), 6576–6579. https://doi.org/10. Machado, A.I., Dordio, A., Fragoso, R., Leitão, A.E., Duarte, E., 2017. Furosemide removal
1016/j.biortech.2010.03.066 in constructed wetlands: comparative efficiency of LECA and cork granulates as
Babatunde, A.O., Kumar, J.L.G., Zhao, Y., 2011. Constructed wetlands using aluminium- support matrix. J. Environ. Manag. 203 (Pt 1), 422–428. https://doi.org/10.1016/j.
based drinking water treatment sludge as P-removing substrate: should alumi­ jenvman.2017.08.002
nium release be a concern? J. Environ. Monit. 13 (6), 1775–1783. https://doi.org/ Martín, M., Gargallo, S., Hernández-Crespo, C., Oliver, N., 2013. Phosphorus and ni­
10.1039/c1em00001b trogen removal from tertiary treated urban wastewaters by a vertical flow con­
Chen, Y., Vymazal, J., Březinová, T., Koželuh, M., Kule, L., Huang, J., Chen, Z., 2016. structed wetland. Ecol. Eng. 61. https://doi.org/10.1016/j.ecoleng.2013.09.046
Occurrence, removal and environmental risk assessment of pharmaceuticals and Matamoros, V., García, J., Bayona, J.M., 2008. Organic micropollutant removal in a full-
personal care products in rural wastewater treatment wetlands. Sci. Total Environ. scale surface flow constructed wetland fed with secondary effluent. Water Res. 42
566–567, 1660–1669. https://doi.org/10.1016/j.scitotenv.2016.06.069 (3), 653–660. https://doi.org/10.1016/j.watres.2007.08.016
Delgado, N., Bermeo, L., Hoyos, D.A., Peñuela, G.A., Capparelli, A., Marino, D., Navarro, Naranjo-Ríos Y., Hernández-Crespo C., Fernández M., Macián J., Añó M., Moscardó C.,
A., Casas-Zapata, J.C., 2020. Occurrence and removal of pharmaceutical and per­ Martín M. Availability of drinking water treatment sludge as substrate in con­
sonal care products using subsurface horizontal flow constructed wetlands. Water structed wetlands for phosphorus removal. In 16th IWA International Conference
Res. 187, 116448. https://doi.org/10.1016/j.watres.2020.116448 on Wetland Systems for Water Pollution Control Conference Proceedings. ISBN:
Delgado-González, L., Prost-Boucle, S., Troesch, S., Molle, P., 2021. Granulated apatite 978–84-17098–52-0.
filters for phosphorous retention in treatment wetlands: experience from full- Olvera-Vargas, H., Leroy, S., Rivard, M., Oturan, N., Oturan, M., Buisson, D., 2016.
scale applications. J. Water Process Eng. 40, 101927. https://doi.org/10.1016/J. Microbial biotransformation of furosemide for environmental risk assessment:
JWPE.2021.101927 identification of metabolites and toxicological evaluation. Environ. Sci. Pollut. Res.
Devi, P., Saroha, A.K., 2017. Utilization of sludge based adsorbents for the removal of 23 (22), 22691–22700. https://doi.org/10.1007/s11356-016-7398-2
various pollutants: a review. Sci. Total Environ. 578, 16–33. https://doi.org/10. RD 1/2016. Real Decreto 1/2016, de 8 de enero, por el que se aprueba la revisión de los
1016/j.scitotenv.2016.10.220 Planes Hidrológicos de las demarcaciones hidrográficas del Cantábrico Occidental,
Directive 91/271/EEC of 21 May, 1991 concerning urban waste-water treatment. Guadalquivir, Ceuta, Melilla, Segura y Júcar, y de la parte Española de las de­
Directive 98/83/EC of 3 November, 1998 on the quality of water intended for human marcaciones hidrográficas del Cantábrico Oriental, Miño-Sil, Duero, Tajo,
consumption. Guadiana y Ebro.
Dotro, G., Langergraber, G., Molle, P., Nivala, J., Puigagut, J., Stein, O., von Sperling, M., Regulation (EU) 2020/741 of the European Parliament and of the Council of 25 May,
2017. Treatment Wetlands - Biological Wastewater Treatment Series. Vol. 7 IWA 2020 on minimum requirements for water reuse.
Publishing. Rousseau, D.P.L., Vanrolleghem, P.A., De Pauw, N., 2004. Model-based design of hor­
Gao, J., Zhao, J., Zhang, J., Li, Q., Gao, J., Cai, M., Zhang, J., 2021. Preparation of a new izontal subsurface flow constructed treatment wetlands: a review. Water Res. 38
low-cost substrate prepared from drinking water treatment sludge (DWTS)/ (6), 1484–1493. https://doi.org/10.1016/j.watres.2003.12.013
bentonite/zeolite/fly ash for rapid phosphorus removal in constructed wetlands. J. Vera, I., Araya, F., Andrés, E., Sáez, K., Vidal, G., 2014. Enhanced phosphorus removal
Clean. Prod. 261, 121110. https://doi.org/10.1016/j.jclepro.2020.121110 from sewage in mesocosm-scale constructed wetland using zeolite as medium
Hou, Q., Meng, P., Pei, H., Hu, W., Chen, Y., 2018. Phosphorus adsorption characteristics and artificial aeration. Environ. Technol. 35 (13), 1639–1649. https://doi.org/10.
of alum sludge: adsorption capacity and the forms of phosphorus retained in alum 1080/09593330.2013.877984
sludge. Mater. Lett. 229, 31–35. https://doi.org/10.1016/j.matlet.2018.06.102 Verlicchi, P., Al Aukidy, M., Zambello, E., 2012. Occurrence of pharmaceutical com­
Hu, Y.S., Zhao, Y.Q., Sorohan, B., 2011. Removal of glyphosate from aqueous environ­ pounds in urban wastewater: removal, mass load and environmental risk after a
ment by adsorption using water industrial residual. Desalination 271 (1–3), secondary treatment-a review. Sci. Total Environ. 429, 123–155. https://doi.org/
150–156. https://doi.org/10.1016/j.desal.2010.12.014 10.1016/j.scitotenv.2012.04.028
Ilyas, H., van Hullebusch, E.D., 2020. Performance comparison of different types of Verlicchi, P., Galletti, A., Petrovic, M., Barceló, D., Al Aukidy, M., Zambello, E., 2013.
constructed wetlands for the removal of pharmaceuticals and their transforma­ Removal of selected pharmaceuticals from domestic wastewater in an activated
tion products: a review. Environ. Sci. Pollut. Res. 27 (13), 14342–14364. https:// sludge system followed by a horizontal subsurface flow bed - analysis of their
doi.org/10.1007/s11356-020-08165-w respective contributions. Sci. Total Environ. 454–455, 411–425. https://doi.org/10.
Ilyas, H., Masih, I., van Hullebusch, E.D., 2020. Pharmaceuticals’ removal by con­ 1016/j.scitotenv.2013.03.044
structed wetlands: a critical evaluation and meta-analysis on performance, risk Vymazal, J., Dvořáková Březinová, T., Koželuh, M., Kule, L., 2017. Occurrence and re­
reduction, and role of physicochemical properties on removal mechanisms. J. moval of pharmaceuticals in four full-scale constructed wetlands in the Czech
Water Health 18 (3), 253–291. https://doi.org/10.2166/wh.2020.213 Republic – the first year of monitoring. Ecol. Eng. 98, 354–364. https://doi.org/10.
Isidori, M., Nardelli, A., Parrella, A., Pascarella, L., Previtera, L., 2006. A multispecies 1016/j.ecoleng.2016.08.010
study to assess the toxic and genotoxic effect of pharmaceuticals: furosemide and Wang, Y., Cai, Z., Sheng, S., Pan, F., Chen, F., Fu, J., 2020. Comprehensive evaluation of
its photoproduct. Chemosphere 63 (5), 785–793. https://doi.org/10.1016/j. substrate materials for contaminants removal in constructed wetlands. Sci. Total
chemosphere.2005.07.078 Environ. 701, 134736. https://doi.org/10.1016/j.scitotenv.2019.134736
Jangkorn, S., Kuhakaew, S., Theantanoo, S., Klinla-or, H., Sriwiriyarat, T., 2011. Zhao, Y., Wang, C., Wendling, L.A., Pei, Y., 2013. Feasibility of using drinking water
Evaluation of reusing alum sludge for the coagulation of industrial wastewater treatment residuals as a novel chlorpyrifos adsorbent. J. Agric. Food Chem. 61 (31),
containing mixed anionic surfactants. J. Environ. Sci. 23 (4), 587–594. https://doi. 7446–7452. https://doi.org/10.1021/jf401763f
org/10.1016/S1001-0742(10)60451-2 Zhao, Y., Nzihou, A., Ren, B., Lyczko, N., Shen, C., Kang, C., Ji, B., 2021. Waterworks
Jenkins, D., Richard, M.G., Daigger, G.T., 2004. Manual on the Causes and Control of sludge: an underrated material for beneficial reuse in water and environmental
Activated Sludge Bulking, Foaming, and Other Solids Separation Problems, third engineering. Waste Biomass-.-. Valoriz. 12, 4239–4251. https://doi.org/10.1007/
ed. CRC Press: IWA Publishing, Cop., London. s12649-020-01232-w
Kahl, S., Nivala, J., van Afferden, M., Müller, R.A., Reemtsma, T., 2017. Effect of design Zhao, Y.Q., Babatunde, A.O., Razali, M., Harty, F., 2008. Use of dewatered alum sludge
and operational conditions on the performance of subsurface flow treatment as a substrate in reed bed treatment systems for wastewater treatment. J. Environ.
wetlands: emerging organic contaminants as indicators. Water Res. 125, 490–500. Sci. Health Part A Toxic. Hazard. Subst. Environ. Eng. 43 (1), 105–110. https://doi.
https://doi.org/10.1016/j.watres.2017.09.004 org/10.1080/10934520701750264

494