Environmental Earth Sciences (2021) 80:607

https://doi.org/10.1007/s12665-021-09903-0

ORIGINAL ARTICLE

Anthropogenic gadolinium as a tracer of raw sewage in surface water

Thaís De Paula Marteleto1 · Jacinta Enzweiler1

Received: 5 December 2020 / Accepted: 15 August 2021 / Published online: 26 August 2021
© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2021

Abstract
Anthropogenic gadolinium ­(Gdanth) is a frequent contaminant in surface water that receives wastewater treatment (WWT)
effluents due to the high stability of Gd-based contrast agents (Gd-CAs) used in magnetic resonance imaging (MRI) exams
and excreted by patients. This study describes the presence of G ­ danth in surface water of a small drainage area (circa 4 k­ m2),
which does not receive WWT effluents. The unexpected ­Gdanth anomalies were attributed to raw sewage leaking from under-
ground effluent removal pipes. The study area is within a university campus with a hospital that regularly runs MRI exams
employing Gd-CAs. Besides Gd, the whole lanthanide (Ln) series was measured and except for Gd, all remaining Ln data
were correlated, implicating that besides the natural Gd, a distinct Gd species contributes up to 95% to its whole content
in the water samples. Besides surface water, samples of the local sewage and pluvial water networks were collected and
analyzed. The ratio between measured and expected Gd values (Gd/Gd*) in surface water samples ranged between 1 and
46, the first corresponding to pristine locations and absence of Gd anomaly and the second nearby the sewage and pluvial
water networks. The G ­ danth anomalies presented transient values, probably associated with the application of the Gd-CAs
to patients and their permanence in the hospital. Other proxies commonly associated to sewage discharge also presented
transient values, not necessarily correlated ­Gdanth, because of the input of distinct anthropogenic sources in the study area.

Keywords Anthropogenic gadolinium · Gd anomaly · Surface water · Tracer · Sewage

Introduction and their transference to the environment occurs through the

discharged effluents.
Anthropogenic gadolinium (­ Gdanth) anomalies have been Under natural conditions, Gd-CAs are highly stable and
reported in water bodies around the world, especially in soluble (Bau et al. 2006; Möller and Dulski 2010; Birka
urban rivers and lakes that receive effluents from wastewa- et al. 2016; Beck et al. 2017), and differently from geogenic
ter treatment plants (WWTP) (Bau and Dulski 1996; Möller Gd ­(Gdgeo), ­Gdanth behaves conservatively (Verplanck et al.
et al. 2000; Elbaz-Poulichet et al. 2002; Zhu et al. 2004; 2005; Kulaksiz and Bau 2007; Dulski et al. 2011). ­Gdanth is
Knappe et al. 2005; Rabiet et al. 2005; Lawrence et al. 2006, mainly associated with surface waters of highly populated
2009; Merschel et al. 2015; Campos and Enzweiler 2016; urban areas with a well-developed health care system (high
Alkan et al. 2020). Such ­Gdanth anomalies result from the number of MRI facilities) and that receive water discharge
use of chelates of this rare earth element (REE) as contrast from WWTPs (Lawrence et al. 2009; Birka et al. 2013; Song
agents (Gd-CAs) in magnetic resonance imaging (MRI) et al. 2017). Discrete G­ danth anomalies were also found in
exams. After administration to patients, the Gd-CAs remain rural catchment areas without any MRI facility but influ-
unmetabolized, and their renal excretion is fast (Birka et al. enced by WWTP effluents (Rabiet et al. 2009). Generally,
2013). During regular sewage treatment the Gd-CAs are effluents from hospitals and medical installations using Gd-
mostly unaffected (Heberer 2002; Kulaksiz and Bau 2011b) CAs in MRI are the primary sources of G ­ danth to the aquatic
environment (Kümmerer and Helmer 2000; Kulaksiz and
Bau 2011a; Goullé et al. 2012; Lerat-Hardy et al. 2019).
* Jacinta Enzweiler Because of the described characteristics, ­Gdanth can be
[email protected] an excellent tracer of wastewater, effluents, and recycled
1
Institute of Geosciences, University of Campinas (Unicamp),
water (Knappe et al. 1999; Möller et al. 2000; Brünjes et al.
Rua Carlos Gomes 250, Campinas, São Paulo 13083‑855, 2016) and is an emerging contaminant in aquatic systems
Brazil

13
Vol.:(0123456789)
607 Page 2 of 18 Environmental Earth Sciences (2021) 80:607

(Kulaksiz and Bau 2013; Rogowska et al. 2018). Compared well as granites, coarse-grained metamorphized rocks, and
to other sewage tracers in surface water, such as B, C ­ l −, paragneisses (IG 1993; Perrotta et al. 2005) (Fig. 1). First-
−
­NO3 , artificial sweeteners, and drugs (Massmann et al. order streams consists of small springs, with a few upstream
2004; Rabiet et al. 2006), ­Gdanth is considered reliable, its damns. On a 2.7 km SE–NW oriented section ending at the
quantification by ICP-MS is relatively straightforward and Chico Mendes Lake, the main brook is covered. The lake
of acceptable cost (Knappe et al. 1999; Bichler et al. 2016). water flows to the Pedras Creek, then to Anhumas Creek, a
Boester and Rüde (2020) concluded that G ­ danth is a better tributary of Atibaia River (Fig. 1).
environmental tracer for municipal sewage in water than five Surface water was sampled at drainages (Mn 1, LC, LB,
other anthropogenic pollutants. RRA, and AR, Fig. 2) and outlet pipes (LI, LM, LS and
MRI contrast agents based on Gd complexes are among RSL, Fig. 2) repeatedly from August 2016 to October 2017.
the most frequently used and their applications have con- Some extra samples at the outlet pipes of surface water,
tinuously increased worldwide (Thomsen 2017). In Brazil, storm and sanitary sewers and drinking water were also col-
Campos and Enzweiler (2013, 2016) and Mortatti and Enz- lected in May 2019. Most surface water sampling occurred
weiler (2019) describe Gd anomalies in riverine water from along a little more than one year with a frequency of three
the Atibaia and Jaguari watersheds. Merschel et al. (2015) to six samples per location. Effluents and tap water were
and Amorim et al. (2019) also determined G ­ danth anomalies sampled two and one years later, respectively. The sanitary
in Lake Paranoá, Brasilia. Pedreira et al. (2018) and Andrade sewers were sampled at the university’s hospital (EF-UH),
et al. (2020) characterized ­Gdanth in coastal water samples the women's hospital (EF-WH) and at medical laboratories
influenced by effluent discharges. In the above cases, water (EF-ML). Drinking water supplied by the municipality was
bodies were affected by either raw or treated sewage or both. collected directly from the main entrance pipeline, which
In this study, we quantified ­Gdanth anomalies in surface supplies water for about 55% of the campus (SE part of the
water of a small catchment area that does not receive any dis- university). Table 1 contains a brief description of the sam-
charge from WWTP. The study area comprises a university pling points with respective geographic coordinates. The
campus and its neighborhood, where hospitals have facilities main stream and the CML sites were sampled three and four
that employ contrast agents like Gd-chelates. Despite that, times, respectively. Overall, we used 54 samples in the study,
­Gdanth anomalies should be absent in surface water because including five field blanks and five duplicates.
all raw sewage is piped to a WWTP. The unexpected ­Gdanth
indicates surface water contamination by raw sewage due to Analytical procedures
leaking in the sewer pipelines. The data can help to map the
leak locations and assist in improving the local wastewater Measurements of temperature, pH values, electrical con-
transference system. ductivity (EC), and turbidity of surface water samples
were carried on site with a multiparameter portable probe
(Horiba U-52). Surface water samples were collected in
Materials and methods acid-cleaned 1 L low-density polyethylene bottles (LDPE,
Nalgene), transported to the laboratory in an ice cooler, and
Study area and sampling sites immediately processed. Alkalinity was determined by aci-
dimetric titration using an automatic titrator (G20 Compact,
The studied watershed (ca. 4 ­km2) comprises small streams Mettler Toledo). Filtered water samples (< 0.22 µm pore
and reservoirs of the University of Campinas campus and membranes) were used for the measurement of all analytes
its neighbor Chico Mendes Lake (CML) in Barão Geraldo described below. Concentrations of major ions were deter-
District, Campinas, São Paulo State, Brazil. The district mined by ion chromatography (IC) (Dionex ICS 2500). Dis-
hosts several educational, research, and healthcare institu- solved organic carbon (DOC) was measured using a 15 mL
tions, as well as residential areas, commercial and indus- acidified aliquot with HCl (2 mol ­L−1) by an NPOC (non-
trial establishments. In the area, three hospitals have MRI purgeable organic carbon) method (Multi N/C 2100S, Ana-
facilities that employ Gd-based contrast agents to examined lytik Jena) following ISO 8245 (1999).
patients. The largest one is the University’s Hospital (UH), Boron and REE concentrations were determined in the
a reference institution for the region, annually performing acidified aliquots (1% v/v ultrapure H­ NO3) on a sector field
more than 8000 MRI exams, including interned and daily inductively coupled plasma—mass spectrometer (ICP-
patients. Daily, about 60,000 people circulate on the uni- SFMS, Element XR™, Thermo Fisher Scientific). Polya-
versity campus, which sewage is piped to a WWTP circa tomic interferences were corrected mathematically following
7 km downstream. the procedures described by Cotta and Enzweiler (2012).
The complex regional geology comprises sedimentary Certified standard solutions (AccuStandard, USA) were
rocks, basalts, and dikes of the Serra Geral Formation as diluted gravimetrically to calibrate the ICP-MS instruments.

13
Environmental Earth Sciences (2021) 80:607 Page 3 of 18 607

Fig. 1  Simplified geological map of study area, showing the flow direction of watercourses and the topographic elevation contours (Modified
from Perrota et al. 2005)

Fig. 2  Sampling sites of the study area

13
607 Page 4 of 18 Environmental Earth Sciences (2021) 80:607

Table 1  Location and Sample ID UTM X UTM Y Description

description of the sampling sites
of the study area Mn 1 288,622 7,474,613 Spillway of a small dam
LC 288,506 7,474,603 Impounded NE headwater stream
LB 288,402 7,474,437 Impounded SE headwater stream
RRA​ 288,203 7,474,781 Mainstream formed by confluence of NE and SE headwaters
AR 288,169 7,474,799 Small tributary of the RRA mainstream
LS 287,671 7,475,797 Outlet of the transference water pipe between the upper and
lower impoundments of Chico Mendes Lake
LM 287,509 7,475,680 Outlet of the water pipes of the channelized the main drainage
(SE–NW transect) to Chico Mendes Lake
LI 287,248 7,475,739 Outlet of the pipe of a secondary channelized headwater stream
RSL 286,983 7,476,032 Discharge outlet of the Chico Mendes Lake
EF-UH 288,169 7,474,390 Sewage sewer pipeline from the university hospital
EF-WH 287,961 7,474,091 Sewage sewer pipeline from the women’s hospital
EF-ML 288,304 7,474,062 Sewage sewer pipeline from the medical laboratories
ESB 287,571 7,474,686 Drinking water supplied by the municipality

The sewage samples were analyzed by ICP-MS after Quantification of REE anomalies and anthropogenic
digesting a 2 mL filtered aliquot with 4 mL ultrapure con- gadolinium
centrate ­HNO3 + 2 mL ultrapure deionized H ­ 2O on a Mul-
tiwave PRO (Anton Paar, Graz, Austria). The total dura- The Ce, Eu, and Gd anomalies were estimated by Eq. (1)
tion of the digestion run was 30 min, divided into a ramp (Lawrence et al. 2006) and Eq. (2) (Bau and Dulski 1996),
of 20 min (0–1300 W) plus 10 min at 1300 W. In these respectively. The subscript SN denotes shale normalization
conditions, the maximum temperature ranged between 170 using PAAS (Post Archean Australian Shale, Pourmand
and 227 °C, with a pressure of 38.1 bar. et al. 2012) values and the asterisk indicates the geogenic
Field blanks prepared from deionized water and dupli- background obtained by interpolation.
cate samples were used for quality control purposes. For [ ]
most analytes, the blank samples results were below the CeSN CeSN
respective detection limits (DL), both by IC and ICP-MS.
= ( )0.5 , (1)
Ce∗SN La × Pr
Only the values of ­K+, ­Ca2+ and C
­ l−, obtained by IC in one
SN SN

field blank (BR-5) exceeded the DL detection limit 5 times

for ­K+ and 10 times for C­ a2+ and C
­ l−. The DOC values of GdSN GdSN
∗ = ( ). (2)
five blank samples ranged from 0.64 to 0.99 mg·L−1, while GdSN 0.33SmSN + 0.67TbSN
the DL of the method was 0.35 mg·L−1.
The electroneutrality parameter (EN%), given by the A positive REE anomaly is defined by a ratio above unity.
relative difference between the totals of dissolved with Due to measurement uncertainties, which increase for water
cations and anions concentrations in meq·L −1, resulted samples with low REE concentrations, only ratios ­REESN/
in the EN values between − 9.8 and + 4.3%, considered REESN* > 1.2 were considered as anomalies.
acceptable for surface water samples (APHA 1998). The total Gd concentration measured in the water samples
The REE concentrations measured in six duplicate sam- has two components: geogenic Gd ­(Gdgeo) and anthropo-
ples from the same sampling location (RSL) had a median genic Gd (­ Gdanth), which were estimated by Eqs. (3) and (4)
coefficient of variation of 3.4%. The REE values obtained (Kulaksiz and Bau 2007), where Gd*SN comes from Eq. (2)
for the certified reference material (CRM) of river water and, ­[GdPAAS] is the Gd mass fraction in the normalizing
SLRS-5 (NRC, Canada) were in acceptable agreement shale (PAAS). The combination of Eqs. (3) and (4) results
(Table S1, Supplementary Material) with the respective in Eq. (5), where the 1.2 multiplying factor is the same men-
compiled REE data (Yeghicheyan et al. 2013). The median tioned above (Lawrence et al. 2009).
value of the expanded measurement uncertainty, estimated
(3)
[ ] [ ] [ ]
Gdanth = Gdmeas − Gdgeo ,
according to Magnusson et al. (2017), on a longer term
SLRS-5 REE set of results is about 25% (at the 95% con-
fidence interval).

13
Environmental Earth Sciences (2021) 80:607 Page 5 of 18 607

(4) Gd concentrations ­([Gd anth]) and their relative propor-

[ ] [ ]
Gdgeo = Gd∗SN × GdPAAS ,
tion compared to total Gd. The ΣREE in the surface water
samples presented a large interval, varying among sam-
(5)
[ ] [ ] ( [ ])
Gdanth = Gdmeas − 1.2 × Gd∗SN × GdPAAS . pling locations and campaigns (Mn 1–2 = 15.2 ng ­L−1 to
LM-6 = 1747 ng ­L−1, Table 3). Some of the lower ΣREE
values were obtained in the upstream sampling points (i.e.,
Mn 1–1, Mn 1–2, and Mn 1–3), which are the most pristine
Results and discussion in the area. At the outlets of the water transference pipes
(LS, LM and LI samples) to the Chico Mendes Lake, the
Chemical composition of the water samples measured ΣREE concentrations ranged between 178 ng ­L−1
(LI-4) and 1747 ng ­L−1 (LM-6).
The water temperature at the sampling sites ranged from The wide range of measured ΣREE in the “dissolved”
20.5 to 28.6 °C, with lower values corresponding to colder fraction of the water samples at each sampling site is prob-
weather periods (Table 2). The pH of the water increased ably controlled by seasonal effects (precipitation, rainfall
from upstream (5.5, LC-3) to downstream (8.1, LS-5) along events) and diversity of land uses, which can implicate in
the SE–NW transect of the main drainage. Turbidity showed variable amounts of colloidal particles in the filtered water
a wide range, varying from 2.4 NTU (LI-2) to 229 NTU samples. The strong correlation (r = coefficient of corre-
(LM-6). The highest turbidity values correspond to the peri- lation = 0.8, p = p value = p < 0.01, n = sample size = 39)
ods of more intense rainfall, in May and October 2017 (35.2 between the ΣREE concentrations and turbidity values in
NTU and 41.1 NTU, respectively). The accumulated precipi- the surface water samples (Fig. 3) suggests an association
tation immediately before the four sampling campaigns was between these variables, despite the first being measured in
37.3 mm (Aug 2016), 94.9 mm (Jan 2017), 97.8 mm (May the filtered sample and the second in natural water. Campos
2017) and 61.2 mm (Oct 2017) (Fig. S1—Supplementary and Enzweiler (2016) and Mortatti and Enzweiler (2019)
Material). However, considering only the accumulated pre- also showed correlation between natural water turbidity
cipitation immediately before sampling, the highest meas- with the REE contents in water samples. Many authors
ured turbidity values followed 24 h of high rainfall. On Jan (e.g., Deberdt et al. 2002) demonstrated the role of the col-
10th, 2017 it rained 29.5 mm, approximately one-third of loidal fraction in retaining most of the REE measured as
the total volume accumulated in January of that year. Simi- “dissolved” in surface water.
larly, on Oct 1st the rainfall was 25.4 mm, corresponding to Figures 4 and 5 present the REE shale-normalized distri-
almost 41% of that month’s precipitation. Compared to other butions in surface water samples collected along the main
fieldwork periods, May 2017 had accumulated precipitation, drainage and at the Chico Mendes Lake and the sewage pipe-
spread along 5 days before the sampling on May 25th, 2017 line, respectively. The diagrams of samples Mn 1–1, Mn
(Fig. S1—Supplementary Material) (Cepagri 2017). 1–2, AR-1, and ESB are not included in Fig. 4 because many
Electrical conductivity (EC) values and, correspond- of their REE measured concentrations were below or very
ingly, the dissolved major ions concentrations increased close to the method DL. The almost flat general pattern of
from upstream to downstream along the main drainage. the shale-normalized diagrams (Figs. 4, 5) indicates a lack
EC ranged from 53 µS·cm−1 at sampling point LC-3 to 336 of fractionation between light (LREE) and heavy (HREE)
µS·cm−1 (LI-1), while at the lake’s outlet (RSL) a median rare earth elements. Exceptions are samples LI (Fig. 5C)
value of 115 µS·cm−1 was obtained. Ammonium, chloride, with slightly enriched HREE patterns. This feature is usu-
and sodium concentrations at the mouth area are between ten ally attributed to the relatively higher solubilities of the
to sixty times those of the most upstream samples. The DOC HREE complexes over the LREE ones in aqueous phases
values ranged from 0.85 mg ­L−1 (Mn 1 and LC sites) to (Sholkovitz 1995; Byrne and Sholkovitz 1996; Gaillardet
20.5 mg ­L−1 (LI site). The DOC median value (3.2 mg ­L−1) et al. 2014).
for all samples is lower than that of Chico Mendes Lake sites Some samples collected during the first campaign
(4.7 mg ­L−1), which is close to the average world DOC con- (RRA-1, LS-1, LM-1, and RSL-1) showed higher nor-
tent (5 mg ­L−1) in surface water (Meybeck 1982). However, malized Yb compared to its REE neighbors, as observed
considering the small area represented by the catchment, the in Figs. 4B and 5A, B, D, respectively. These samples
DOC contents indicate anthropogenic input. were analyzed by ICP-QMS and ICP-SFMS, and in both
sets of results, high Yb values were obtained. As already
Rare earth elements mentioned, at the REE concentrations in natural water,
expanded measurement uncertainties may achieve one
Table 3 shows the measured concentrations of REE, their fourth of the obtained values. Besides the measurement
sum (ΣREE), Ce and Gd anomalies, estimated anthropogenic uncertainty, those samples with high Yb results are in

13
607 Page 6 of 18 Environmental Earth Sciences (2021) 80:607

Table 2  Sample identification, date of sampling, physicochemical parameters measured on site and selected analytical results
Sample Day/month/ Tempera- pH EC Turbidity DOC Na+ NH4+-N K+ Mg2+ Ca2+
year ture (°C) (µS ­cm−1) (NTU) (mg ­L−1) (mg ­L−1) (mg ­L−1) (mg ­L−1) (mg ­L−1) (mg ­L−1)

Mn 1 24/08/2016 21.0 5.8 65 18.4 2.11 2.34 0.07 0.14 2.58 4.00
Mn 1–2 11/01/2017 23.5 7.2 78 18.4 0.85 2.38 0.11 0.19 2.49 3.80
Mn 1–3 25/05/2017 21.8 5.7 74 44.6 0.95 2.52 0.06 0.29 2.44 3.89
LC-1 24/08/2016 20.6 5.8 62 20.0 0.85 2.38 0.09 0.32 2.68 4.25
LC-2 11/01/2017 26.4 7.0 56 17.1 1.68 2.31 0.04 1.14 2.44 3.89
LC-3 25/05/2017 22.4 5.5 53 32.2 1.08 2.35 0.12 0.47 2.43 3.89
LB-1 24/08/2016 20.8 6.7 154 76.0 2.72 6.75 0.30 2.15 4.91 13.8
LB-2 11/01/2017 26.2 6.9 116 79.1 3.26 5.34 0.56 3.09 3.60 10.8
LB-3 25/05/2017 21.4 5.9 114 112 2.20 6.20 0.21 1.81 3.80 11.0
RRA-1 24/08/2016 20.8 6.7 135 33.3 2.89 6.25 1.13 1.84 3.87 11.7
RRA-2 11/01/2017 25.9 7.0 141 43.8 2.95 7.20 0.54 2.86 3.75 14.0
RRA-3 25/05/2017 22.6 5.9 162 133 3.93 5.99 0.80 2.47 4.04 13.5
AR-1 24/08/2016 21.1 6.7 210 7.00 2.08 20.5 0.14 2.80 2.22 13.9
AR-2 11/01/2017 24.2 6.9 188 21.0 2.36 21.9 0.25 2.97 1.49 5.45
AR-3 25/05/2017 21.4 6.1 131 58.9 1.58 23.7 0.13 2.31 1.38 4.57
LS-1 17/08/2016 21.0 7.8 54 24.0 3.69 5.09 0.04 0.90 0.79 3.51
LS-2 04/01/2017 27.9 6.3 54 39.4 5.21 4.76 0.05 1.12 0.77 3.54
LS-3 11/01/2017 28.6 7.2 45 42.4 4.79 4.63 < DL 1.42 0.73 3.33
LS-4 18/05/2017 21.6 7.9 67 77.5 3.19 5.58 0.67 1.24 0.75 3.75
LS-5 25/09/2017 26.1 8.1 104 76.2 4.91 12.3 < DL 2.62 0.95 4.43
LS-6 02/10/2017 23.2 6.4 99 61.5 5.35 11.1 0.08 2.44 0.98 4.83
LM-1 17/08/2016 20.5 7.4 142 33.5 3.19 8.13 1.19 1.46 2.69 9.27
LM-2 04/01/2017 24.9 6.6 159 26.5 3.30 10.5 1.79 2.64 3.15 13.8
LM-3 11/01/2017 24.8 6.9 183 36.0 5.99 12.5 2.88 3.51 3.24 15.7
LM-4 18/05/2017 21.5 7.4 138 35.2 2.52 6.96 0.92 1.65 3.15 10.8
LM-5 25/09/2017 22.9 6.5 268 32.2 4.70 17.1 < DL 4.90 3.15 13.1
LM-6 02/10/2017 21.5 6.3 150 229 4.61 6.86 1.21 3.99 3.07 11.2
LI-1 17/08/2016 22.6 7.3 336 22.9 20.5 32.8 < DL 3.88 3.65 11.2
LI-2 04/01/2017 24.2 6.7 147 2.40 1.43 15.7 0.16 0.57 3.24 8.21
LI-3 11/01/2017 25.0 7.0 175 30.8 5.13 18.4 1.76 2.13 3.08 10.3
LI-4 18/05/2017 22.8 7.5 158 7.20 4.79 15.5 0.97 1.22 2.78 7.99
LI-5 25/09/2017 24.1 7.2 276 18.2 3.95 20.5 < DL 4.16 3.14 12.7
LI-6 02/10/2017 21.6 6.2 126 64.0 8.91 8.46 2.46 4.34 1.40 8.11
RSL-1 17/08/2016 21.0 7.0 127 22.3 3.87 8.18 1.18 1.46 2.71 9.32
RSL-2 04/01/2017 27.3 7.3 90 22.2 3.84 6.13 0.31 2.04 1.61 7.62
RSL-3 11/01/2017 28.0 7.0 77 31.6 3.55 5.41 0.38 2.00 1.38 6.75
RSL-4 18/05/2017 21.4 7.5 103 19.9 2.48 6.96 1.32 1.68 1.78 7.33
RSL-5 25/09/2017 26.5 7.9 144 36.8 3.83 11.8 1.23 2.64 2.43 8.91
RSL-6 02/10/2017 22.7 6.1 129 41.1 4.72 9.78 1.07 3.73 2.19 8.52
EF-UH 10/05/2019 – – – – – 95.9 – 35.6 3.50 17.7
EF-WH 10/05/2019 – – – – – 124 – 32.7 4.00 18.6
EF-ML 10/05/2019 – – – – – 60.6 – 34.1 4.09 20.3
ESB 27/11/2018 24.3 7.11 164 2.8 2.84 10.2 < DL 4.09 1.40 22.4
DL* – – – – 0.35 0.02 0.02 0.05 0.03 0.05
Sample F− (mg ­L−1) Cl− (mg ­L−1) NO2−-N (mg ­L−1) SO42−-S (mg ­L−1) NO3−-N (mg ­L−1) HCO3− (mg ­L−1) B (µg·L−1)

Mn 1 < DL 3.31 < DL 0.10 1.02 26.3 < DL

Mn 1–2 < DL 3.55 < DL 0.05 1.30 24.8 1.26
Mn 1–3 < DL 3.37 < DL 0.12 1.02 26.8 1.46

13
Environmental Earth Sciences (2021) 80:607 Page 7 of 18 607

Table 2  (continued)
Sample F− (mg ­L−1) Cl− (mg ­L−1) NO2−-N (mg ­L−1) SO42−-S (mg ­L−1) NO3−-N (mg ­L−1) HCO3− (mg ­L−1) B (µg·L−1)
LC-1 < DL 3.32 < DL 0.12 0.81 26.2 < DL
LC-2 < DL 3.67 < DL 0.08 0.74 25.7 2.72
LC-3 < DL 3.20 < DL 0.16 0.82 25.6 1.86
LB-1 0.07 12.3 0.02 0.86 0.29 67.7 3.46
LB-2 0.09 8.72 0.02 0.68 0.18 56.5 7.14
LB-3 0.06 9.95 0.01 1.02 0.40 55.6 4.10
RRA-1 0.03 10.4 0.03 0.73 0.58 58.0 3.38
RRA-2 0.13 11.8 0.04 0.79 0.53 65.1 7.02
RRA-3 0.03 9.59 0.02 1.24 0.64 66.4 5.09
AR-1 0.06 35.7 0.01 0.81 0.28 49.8 5.46
AR-2 0.02 34.9 < DL 0.35 0.49 24.6 3.93
AR-3 0.02 38.6 < DL 0.31 0.70 17.1 3.08
LS-1 0.03 4.83 < DL 0.02 0.02 21.7 2.24
LS-2 0.03 4.30 < DL 0.04 0.01 22.2 4.74
LS-3 0.03 4.14 < DL 0.04 0.01 21.9 5.08
LS-4 0.03 5.16 < DL 0.28 0.08 26.7 4.14
LS-5 0.11 13.0 < DL 0.61 0.01 25.4 4.24
LS-6 0.06 11.9 < DL 0.58 0.03 28.5 4.53
LM-1 0.03 10.3 0.05 0.50 0.86 51.1 3.20
LM-2 0.12 15.7 0.07 0.61 0.80 67.6 6.43
LM-3 0.13 18.5 0.06 0.98 0.77 78.0 8.31
LM-4 0.07 10.0 0.06 0.62 0.87 54.0 4.70
LM-5 0.15 25.2 0.05 1.15 0.42 71.1 5.68
LM-6 0.04 12.3 0.04 1.80 0.45 52.3 6.04
LI-1 0.07 39.5 0.28 3.02 0.52 114 4.85
LI-2 0.02 17.2 0.03 0.41 5.82 28.0 1.71
LI-3 0.03 20.1 0.07 1.06 3.91 52.7 3.36
LI-4 0.08 18.0 0.16 0.57 4.28 34.3 4.03
LI-5 0.22 27.7 0.16 1.76 1.21 88.0 5.07
LI-6 0.09 10.5 0.04 1.62 1.08 38.1 5.44
RSL-1 0.06 10.3 0.06 0.52 0.43 49.2 3.30
RSL-2 0.04 7.46 0.02 0.29 0.26 38.1 7.34
RSL-3 0.04 6.55 0.01 0.27 0.25 34.6 7.30
RSL-4 0.06 8.66 0.03 0.45 0.37 41.8 4.98
RSL-5 0.11 14.7 0.02 0.49 0.18 53.3 4.37
RSL-6 0.05 12.2 0.03 0.62 0.10 50.7 5.33
EF-UH – – – – – – 157
EF-WH – – – – – – 199
EF-ML – – – – – – 339
ESB 0.73 34.8 < DL 1.65 1.14 24.6 12.4
DL* 0.001 0.015 0.005 0.01 0.003 – 0.3

*DL—method detection limit

some way connected. RRA-1 is the sampling point of the and Er in samples AR-2 and RRA-2, Fig. 4B) also present
main drainage, before its outlets LS and LM to the CML, isolated normalized values that differ slightly from their
which outlet is RSL (Fig. 2). Few other elements (e.g., Sm neighbors, for which we still have no explanation.

13
607 Page 8 of 18 Environmental Earth Sciences (2021) 80:607

Fig. 3  Relation between surface

water turbidity values at each
sampling site and the measured
∑REE in respective water
samples at A upstream sites
(Mn 1, LC, LB, RRA and AR)
and B downstream sites at the
CML (LS, LM, LI) and its
outlet (RSL)

Fig. 4  REE shale normalized

REE distributions of surface
water samples from A upstream
(Mn 1, LC and LB) sampling
locations and B the middle sites
of the mainstream (RRA and
AR). Diagrams of samples Mn
1–2 and AR-1 are not shown
because their REE measured
concentrations were mostly
below or close to ten times the
method DL. See text for discus-
sion about general and specific
features of the diagrams

Rare earth elements anomalies et al. 2008). Positive Ce anomalies can also be produced
by siderophores (Neilands 1995), which easily oxidize
The REE shale-normalized distributions (Figs. 4, 5) indi- ­Ce3+ and complex ­Ce4+ keeping it in solution (Kraemer
cate that many of the water samples contained Ce, Eu, et al. 2015). Positive Ce anomalies were also previously
and Gd concentrations decoupled from their neighbors, reported in surface water samples of regional watersheds
characterizing anomalies. The values of Ce and Gd anoma- (Campos and Enzweiler 2016; Mortatti and Enzweiler
lies, quantified according to Eqs. 1 and 2, are in Table 3. 2019).
Europium anomalies, visible in many patterns, were not Some samples (LI site and tap water) presented negative
quantified. They are natural due to weathering and dissolu- Ce anomalies (Ce/Ce* < 0.8). Negative Ce anomalies are
tion of minerals in which E ­ u2+ replaces C
­ a2+. A previous characteristic of seawater and have also been observed in riv-
lithogeochemical study (Bulia and Enzweiler (2015) iden- erine (Goldstein and Jacobsen 1988; Leybourne et al. 2000;
tified positive Eu anomalies in diabase located upstream Campos and Enzweiler 2016) and groundwater (Smedley
from the main drainage. 1991; Dia et al. 2000). Negative Ce anomalies were also
Most samples exhibited discrete positive Ce anomalies, reported in tap water in Germany (Tepe et al. 2014; Schmidt
with Ce/Ce* ≥ 1.2 (Table 3), which we consider of natural et al. 2019). The slight negative anomalies of the first four
source. The Ce decoupling from its neighbors results from LI samples are attributed to the preferential association of
oxidation of C ­ e3+ to C
­ e 4+, which is very insoluble and ­Ce4+ to particulate phases, compared with the trivalent REEs
can associate to colloidal phases present in the < 0.22 μm (Nelson et al. 2003; Leybourne and Johannesson 2008;
fraction of the filtered water samples. Such phases can be Kraemer et al. 2015), which then flocculate and settle (or
cerianite and finely dispersed inorganic Mn or Fe (oxy- are filtered) leaving the aqueous phase depleted in Ce. Simi-
hydr-)oxides or organic particles (Bau et al. 1997; Pédrot larly, the negative Ce anomaly in the tap water (ESB) sample

13
Environmental Earth Sciences (2021) 80:607 Page 9 of 18 607

Fig. 5  REE shale normalized distributions of Chico Mendes Lake water samples A at the LS site, B at the LM site, C at the LI site and, D at the
RSL site. Diagram of panel E are of the hospitals’ effluents samples. See text for discussion about general and specific features of the diagrams

is a result of water treatment and solid flocculation, which the downstream following samples RRA-1 and AR-1 pre-
removes particulate material from the aqueous phase. The sented Gd/Gd* = 3.4 and 4.6, respectively. The Gd anomaly
Ce and Eu anomalies are attributed to natural sources, while vanished at the RRA site and dropped to one-third or less at
Gd anomalies are the human-related source, and it will be the AR point (Gd/Gd* = 1.3 and 1.5) due to improvements
discussed in detail below. in the sewer network in the vicinity of those sampling loca-
tions. The samples collected at the outlet of the transference
Spatial and temporal variations of ­Gdanth anomalies water pipes to and out the Chico Mendes Lake presented a
wide interval of Gd/Gd* ratios, ranging from absence of Gd
Many samples of this study presented measured Gd concen- anomaly (Gd/Gd* = 1.0 for LI-1) to very high values (up to
trations (Table 3 and Figs. 4, 5) attributed to anthropogenic 46 at RSL-5).
input. Its non-natural source is confirmed by the absence or The increase in the Gd/Gd* values from upstream to
weak correlation (r = up to 0.20, p > 0.01) between Gd and downstream indicates that sewage leaking occurs at the
the other REE, which presented strong correlation coeffi- sewer pipeline, inside the university campus. The regis-
cients among themselves (r = 0.60–1.00, p < 0.01). tered Gd/Gd* values for the CML (LS and LM sampling
The Gd/Gd* ratios between 0.9 and 1.2 of the upstream locations) are transient, presenting variable values and an
drainage and reservoirs samples (Mn 1, LC, and LB sites) absence of a pattern for the sample series. The highest Gd
indicate the absence of Gd anomalies in the surface water in anomalies were registered for three of the lake locations (LS,
this area of the campus. During the first sampling campaign, LM, and RSL). At the LS sampling site, LS-5 and LS-6

13
607 Page 10 of 18 Environmental Earth Sciences (2021) 80:607

samples presented Gd/Gd* = 42 and 37, respectively, while not hospitalized are more numerous. Under such circum-
the RSL-5 sample achieved Gd/Gd* = 46. For the LM site, stances, the expected anthropogenic signal Gd can vary
LM-2, LM-3, and LM-5 exhibited their highest Gd anoma- accordingly. The Gd-chelate dose administered per patient
lies (Gd/Gd* = 19–21). As shown in Table 3, the first sam- is between 0.1 and 0.3 mmol ­kg−1, and its half-life in the
ples of these three CML sites presented similar Gd/Gd* body is approximately 1.5 h. A patient excretes more than
values (Gd/Gd* = 3.0 (LS-1), 2.8 (LM-1) and 3.4 (RSL-1)). 90% of the injected dose (1.1–3.3 g, in a patient with 70 kg
Interestingly, these Gd/Gd* ratios are of the same order than body weight) (Tweedle et al. 1988; Oksendal and Hals
those at the RRA and AR sites (Gd/Gd* = 3.4 (RRA-1) and 1993; Shellock and Kanal 1999) while still in the hospi-
4.6 (AR-1)). These results suggest that, at the beginning of tal. The remaining dose is eliminated within 24 h, and yet
the study, the Gd concentrations in excess were almost the a smaller amount gradually for more than 40 days after
same along the main drainage and probably related to the the MRI exam (Kümmerer and Helmers 2000). Therefore,
same leaking source, unlike the ­Gdanth anomalies identified most of the primary sources of anthropogenic Gd are hos-
in the subsequent sampling campaigns. The proportion of pitals and clinics sewers (Kümmerer and Helmers 2000;
­Gdanth at the AR sampling location, estimated by Eq. 5 and Kümmerer 2001; Künnemeyer et al. 2009; Ort et al. 2010;
included in the last column of Table 3, oscillated from 74 to Goullé et al. 2012).
9% and then to 21%, whereas for the RRA site, only the first A plan obtained from the administration (Fig. S2—Sup-
sample (RRA-1) showed anomalous ­Gdanth, attaining 65%. plementary Material, SIC (2019)), indicates storm sewers,
The ­Gdanth concentrations estimated in the six water sam- as well as the campus and municipal sewer pipelines, the
ples collected at the LS, LM, and RSL locations, shown in ­ danth anomalies relate
last two running parallel. As local G
Fig. 6, were almost the same for the first sampling campaign to point sources, interventions to remove leaking in the
(~ 10 ng ­L−1), close to 60% of measured Gd. The ­Gdanth campus sewer pipeline are not necessarily effective. As
of the LS site showed increasing concentrations, reaching suggested by the measured Gd anomalies, a significant
between 500 and 600 ng ­L−1 for its last two samples. The source of raw sewage to the CML seems to have moved
second and third samples of the LM site contained close to from the LM to the LS outlet during the time of this study.
250 ng ­L−1 and afterward it decreased to original levels and Besides the university’s hospital, the contribution of other
lastly had none G ­ danth. The outlet of the CML (RSL site) sources to the increase in the absolute ­Gdanth anomalies
represents the combined the ­Gdanth contributions of the LS, at LS and RSL during the last two sampling campaigns,
LM and LI sites. The change in Gd anomalies values for the cannot be discarded, because other two hospitals with
three sampling sites (Fig. 6), especially the huge increase MRI facilities operate at NE of the Chico Mendes Lake.
for LS-4 in May 2017 and the simultaneous decrease for To out rule such possibility, other sampling sites require
LM-4 also suggest interventions in the sewer network. These investigation.
results were communicated to the technical staff in charge The results obtained for the sewage samples collected
of the infrastructure, but they were not able to relate them in the pipes originating from the two local hospitals and
to any specific event. medical laboratories as well as drinking water also con-
The Gd anomalies in surface water of the study area firmed the presence of anthropogenic Gd (Table 3). The
were attributed to leaking in the local sewer network. The highest Gd anomaly corresponds to the EF-UH sam-
university’s hospital regularly uses Gd MRI enhancers in ple (Gd/Gd* = 8.3) of the university hospital, while the
exams on both admitted and daily patients, but their num- other samples (EF-WH and EF-ML) had much smaller
bers are variable. Between the two groups of patients, the Gd anomalies, of 1.4 and 1.6, respectively. Such Gd/
Gd* values equate to ­Gdanth concentrations of 189 ng ­L−1
(1.2 nmol ­L −1 , 86% of measured Gd), 3.3 ng ­L −1
800 (0.02 nmol ­L−1, 17%) and 6.1 ng ­L−1 (0.04 nmol ­L−1, 24%)
585 Day/month/year for EF-UH, EF-WH and EF-ML samples, respectively. The
600 518 drinking water sample (ESB) also presented a Gd anomaly
[Gdanth] ng·kg-1

25/08/2016

400
04/01/2017 (Gd/Gd* = 3.6) with a G­ danth concentration of 7.7 ng ­L−1
−1
243 263
11/01/2017
(0.05 nmol ­L , 67% of measured Gd). The drinking water
195 18/05/2017
200
101
126
152
25/09/2017
supplied by the municipal sanitary company is extracted
11 28 24 9.5 7.1 2.0 8.6 9 12
41
02/10/2017 from Atibaia river, which along its course receives both
0
LS LM RSL treated wastewater and raw effluents (Campos and Enz-
Sampling sites weiler 2016; Mortatti and Enzweiler 2019). The Gd pre-
sent in drinking water may contribute to the measured Gd
Fig. 6  Gdanth concentrations of the samples from LS, LM and RSL anomalies in the sewage pipelines, but the highest values
sites

13
Environmental Earth Sciences (2021) 80:607 Page 11 of 18 607

such as those registered for the EF-UH and in the CML ­(Gdanth) and geogenic ­(Gdgeo) for surface water samples
surface water samples are locally added. with Gd anomalies.
To summarize such findings, Fig. S3 (Supplemen-
tary Material) indicates the proportion of anthropogenic

Fig. 7  Spatial and temporal variability of Gd/Gd* values and B, ­Cl− and ­Na+ concentrations in surface water samples inside university campus
(A–D) and at the CML sampling locations, including its outlet pipe RSL (E–H)

13
 Table 3  Dissolved concentrations of rare earth elements in surface water, effluent and drinking water samples (ng L
­ −1)
607

Sample La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er

13
Mn 1–1 8.36 16.0 1.55 7.79 1.64 1.35 1.20 0.20 1.20 0.28 0.84
Mn 1–2 2.71 3.40 2.00 2.59 2.01 < DL 1.99 < DL 0.50 < DL < DL
Page 12 of 18

Mn 1–3 24.7 66.6 7.06 25.3 6.83 1.92 6.61 0.85 4.02 < DL 1.93
LC-1 26.2 73.0 6.09 27.7 5.60 2.83 4.05 0.67 3.47 0.64 1.99
LC-2 34.9 145 10.1 37.7 8.01 3.10 9.18 1.16 4.97 < DL 2.20
LC-3 76.4 227 21.6 90.5 19.0 5.40 18.4 2.69 13.2 2.23 6.44
LB-1 111 288 25.4 110 22.9 10.0 16.6 2.71 15.8 2.81 8.89
LB-2 164 446 44.2 189 40.6 9.97 41.2 6.04 30.1 5.59 16.5
LB-3 232 673 62.8 259 53.4 13.9 56.5 8.73 42.0 7.65 23.2
RRA-1 52.9 135 11.9 51.4 10.5 8.11 31.2 1.35 7.30 1.28 4.07
RRA-2 85.9 240 22.6 93.1 20.0 5.74 21.6 2.87 14.1 2.36 8.33
RRA-3 106 299 27.9 116 26.0 7.64 27.1 3.72 17.8 3.25 10.1
AR-1 6.36 10.8 0.84 5.62 1.38 8.96 4.33 0.12 1.15 0.36 2.20
AR-2 69.7 216 18.5 68.9 17.5 3.01 21.1 2.41 12.3 2.35 7.20
AR-3 44.3 144 12.5 52.2 12.0 2.72 16.8 1.67 9.27 1.44 5.03
LS-1 49.0 127 11.2 46.0 8.19 2.30 17.8 0.80 4.91 0.83 2.45
LS-2 82.5 247 20.1 79.3 15.5 4.05 44.1 1.92 8.19 1.52 4.61
LS-3 79.5 241 20.8 80.2 14.7 4.00 39.9 2.00 8.55 1.58 5.35
LS-4 50.4 151 13.8 50.5 9.26 2.05 111 1.23 5.11 < DL 3.10
LS-5 93.5 266 22.0 88.8 16.1 4.36 602 2.12 13.0 2.60 6.91
LS-6 91.5 275 22.5 91.0 16.6 4.18 535 2.13 12.7 2.43 7.10
LM-1 41.1 90.4 8.39 35.9 7.24 3.46 16.4 0.83 4.84 0.90 2.80
LM-2 59.3 165 15.5 57.8 13.3 3.60 258 2.00 9.30 1.56 5.27
LM-3 57.6 159 15.1 60.9 13.1 2.85 279 2.08 9.73 1.78 5.38
LM-4 44.8 116 11.6 45.2 10.1 2.66 18.7 1.51 6.26 < DL 4.65
LM-5 30.3 89.6 7.98 36.0 7.65 2.24 134 1.06 6.29 1.24 3.67
LM-6 257 746 74.7 334 70.1 21.4 82.1 10.5 63.8 12.7 33.8
LI-1 106 128 20.2 84.7 14.8 4.58 12.1 1.86 11.2 2.18 6.93
LI-2 53.3 51.7 11.7 43.3 8.55 2.14 11.1 1.50 8.23 1.89 7.49
LI-3 72.2 105 17.7 68.3 15.7 3.51 16.8 2.39 13.3 2.96 12.0
LI-4 39.1 57.1 9.38 32.6 7.94 < DL 11.3 1.18 5.91 < DL 4.79
LI-5 37.5 78.0 10.3 43.1 8.39 2.38 13.7 1.41 10.2 2.67 10.5
LI-6 124 285 31.4 136 26.2 6.77 50.6 3.62 22.5 4.91 14.0
RSL-1 23.1 50.0 4.91 21.0 4.33 2.66 12.9 0.56 2.99 0.58 1.77
RSL-2 18.8 44.0 6.11 19.6 5.65 1.93 15.7 0.75 2.96 < DL 1.97
RSL-3 75.6 193 20.4 80.8 17.5 5.42 30.2 2.45 11.9 2.35 6.57
Environmental Earth Sciences (2021) 80:607
 Table 3  (continued)
Sample La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er

RSL-4 20.9 58.0 6.49 21.6 5.98 < DL 47.4 0.76 3.54 < DL 2.00
RSL-5 18.9 50.7 4.90 22.5 4.59 1.42 200 0.67 4.01 0.89 2.60
RSL-6 90.4 242 24.6 107 22.0 6.66 176 3.40 21.1 4.22 11.9
EF-UH 201 465 46.1 199 35.3 4.30 221 3.63 18.0 3.37 7.80
EF-WH 80.6 145 15.5 59.1 12.0 1.56 19.1 2.13 13.9 2.72 7.99
EF-ML 219 469 47.2 173 26.1 1.76 25.7 1.92 9.30 1.81 4.86
ESB 4.50 2.56 0.88 3.49 9.48 1.52 11.5 0.09 0.48 < DL < DL
DL* 0.60 0.70 0.11 0.54 0.13 0.05 0.14 0.02 0.10 0.02 0.05
Environmental Earth Sciences (2021) 80:607

Sample Tm Yb Lu ΣREE Ce/Ce* Gd/Gd* [Gdanth] % ­Gdanth

Mn 1–1 0.15 1.07 0.18 41.9 – – – –

Mn 1–2 < DL < DL < DL 15.2 – – – –
Mn 1–3 < DL 1.98 < DL 148 1.2 1.1 – –
LC-1 0.28 1.99 0.28 155 1.4 0.9 – –
LC-2 < DL 1.51 < DL 258 1.9 1.2 – –
LC-3 < DL 5.79 0.98 490 1.3 1.0 – –
LB-1 1.19 8.41 1.12 624 1.3 0.9 – –
LB-2 2.23 13.0 2.05 1011 1.3 1.0 – –
LB-3 2.97 18.8 2.90 1458 1.3 1.0 – –
RRA-1 0.55 6.71 0.58 323 1.3 3.4 20 65
RRA-2 1.01 6.62 0.91 525 1.3 1.1 – –
RRA-3 1.26 8.37 1.14 655 1.3 1.1 – –
AR-1 0.51 6.77 1.20 50.6 – 4.6 3.2 74
AR-2 1.11 6.33 1.08 448 1.5 1.3 1.9 9
AR-3 < DL 5.69 0.98 309 1.5 1.5 3.5 21
LS-1 0.34 3.36 0.37 274 1.3 3.0 11 59
LS-2 < DL 3.45 < DL 512 1.5 3.3 28 64
LS-3 < DL 3.61 < DL 501 1.4 3.0 24 60
LS-4 < DL 2.02 < DL 400 1.4 13 101 91
LS-5 0.95 6.27 0.94 1125 1.4 42 585 97
LS-6 1.03 6.83 1.03 1069 1.5 37 518 97
LM-1 0.44 4.23 0.48 217 1.2 2.8 9.5 57
LM-2 < DL 4.60 0.76 597 1.3 20 243 94
LM-3 < DL 5.23 0.73 612 1.3 21 263 94
LM-4 < DL 3.43 0.62 265 1.2 1.9 6.9 37
LM-5 0.64 3.96 0.60 325 1.4 19 125 94

13
Page 13 of 18 607
 Table 3  (continued)
607

Sample Tm Yb Lu ΣREE Ce/Ce* Gd/Gd* [Gdanth] % ­Gdanth

13
LM-6 4.77 31.6 4.66 1747 1.3 1.2 0.7 0.9
LI-1 1.10 9.01 1.43 404 0.7 1.0 – –
Page 14 of 18

LI-2 1.16 9.56 1.86 213 0.5 1.2 – –

LI-3 2.03 16.9 3.07 351 0.7 1.1 – –
LI-4 < DL 7.01 1.29 178 0.7 1.5 2.1 19
LI-5 1.96 16.0 3.39 240 1.0 1.6 3.1 23
LI-6 2.18 15.3 2.69 726 1.1 2.1 22 43
RSL-1 0.32 3.33 0.43 129 1.1 3.4 8.4 65
RSL-2 < DL 2.01 < DL 120 1.0 3.1 9.6 61
RSL-3 < DL 5.35 0.85 452 1.2 1.8 10 35
RSL-4 < DL 2.13 < DL 169 1.2 9.1 41 87
RSL-5 0.40 3.01 0.58 315 1.3 46 163 97
RSL-6 1.72 11.9 1.79 725 1.2 7.9 148 85
EF-UH 0.95 6.56 1.02 1213 1.2 8.3 189 86
EF-WH 0.98 5.96 1.01 368 1.0 1.4 3.3 17
EF-ML 0.65 4.22 0.75 985 1.1 1.6 6.1 24
ESB 0.08 0.60 0.10 35.3 0.1 3.6 7.7 67
DL* 0.02 0.03 0.01 – – – – –

*DL—method detection limit

Environmental Earth Sciences (2021) 80:607
Environmental Earth Sciences (2021) 80:607 Page 15 of 18 607

Comparison with other wastewater tracers Conclusions

Several dissolved constituents associated with effluents Anthropogenic Gd anomalies (Gd/Gd* = 1.3–46) detected
such as ­Cl−, ­NO2−, ­NO3−, ­HCO3−, ­NH4+, ­Na+, ­Ca2+, B and in surface water of an isolated small drainage basin without
DOC are frequently used as tracers of sewage effluents in any WWTP was associated to the contamination of storm
aquatic systems (Alhajjar et al. 1990; Hatt et al. 2004; Bar- pipes and local creeks by raw effluents leaking from the local
ber et al. 2006; Richards et al. 2017; Song et al. 2017). For sewage pipelines. The G ­ danth in wastewater was associated
the study area, Frias et al. (2017) showed a close relation to the application of Gd based contrast agents to patients of
between surface water hydrogeochemical data and land an MRI facility at the local hospital.
use, with several parameters increasing from upstream to Expressive spatial and temporal variability observed for
downstream (Fig. 2). several hydrogeochemical constituents as well as the Gd/
Figure 7A–D show the variation in Gd anomalies (Gd/ Gd* values, were attributed to differences in land use among
Gd*), and of B, C ­ l− and ­Na+ concentrations, respectively, sampling locations, to the intermittent application of contrast
at upstream of the main drainage (Mn 1, LC, LB, RRA, and agents at the MRI facility, and to operational interventions
AR sites). The Mn 1 and LC correspond to the most pris- performed by the local administration during the study.
tine sample sites. Despite the absence of Gd anomalies at Gadolinium was confirmed as a highly sensitive tracer of
the LB site, the relatively high concentrations of C ­ l−, ­Na+, sanitary sewage in surface water, with relatively simple and
2−
­SO4 -S and B (average values of 10.3 mg ­L , 6.1 mg ­L−1,
−1
low-cost determination, compared to some emerging con-
0.85 mg ­L−1 and 4.9 µg ­L−1, respectively) probably origi- taminants. Also, anthropogenic Gd behaves conservatively
nate from domestic effluents (wastewater) from an upstream but its signal may be transient. Particularly in this study, the
neighborhood, SE of the study area. estimated variability in G­ danth was probably influenced by
As discussed above, the Gd/Gd* values decreased at the number of Gd-CAs applications for MRI exams and the
RRA and AR sites after the first sampling season. The same patients' permanence in the hospital.
occurred for B concentrations at the AR site, but not at the
RRA, where B increased and then diminished, similarly to Supplementary Information The online version contains supplemen-
tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 12665-0​ 21-0​ 9903-0.
LB samples. For these sampling sites, ­Cl− and ­Na+ concen-
trations almost did not change during the study, suggesting
Acknowledgements TPM acknowledges a scholarship from the Coor-
a constant source. dination for the Improvement of Higher Education Personnel—Bra-
At CML sampling sites, the concentrations of the dis- zil (CAPES)—Finance Code 001. JE acknowledges the support from
solved constituents were highly variable (Fig. 7E–H), except the National Council for Scientific and Technological Development
(CNPq, Grant number 312507/2013-5) and The São Paulo Research
for ­Cl− and N­ a+. At the LI site discrete Gd anomalies were
Foundation (FAPESP, Grant number 2012/05024-2). TPM acknowl-
present in all samples, while other anthropogenic proxies edges Francisco Ferreira de Campos for clarifying discussions along
changed along the study. The same occurred for other con- this research. The kind support of DSIS (Department of Systems) and
stituents (DOC, N ­ O3−, ­SO42−, ­NO2−, Table 2). The patterns CEMA (Center of Animal Monitoring) staff of University of Campinas
was essential to obtain information about the campus infrastructure and
of recorded Gd anomalies and B, C ­ l− and N
­ a+ concentra-
to access restrict sampling locations.
tions for LS and RSL sites are similar during time of study.
At the LM sampling location, all four proxies show similar Author contributions Both authors contributed at all stages of the
tendencies despite the temporal variability. research and in the manuscript’s writing.
The effluent samples contained high concentrations of
several major dissolved components (Table 2) while Gd Funding All funding is mentioned in the Acknowledgements.
anomalies were only expressive for EF-UH. Among them, B
Availability of data and materials All used data are presented either in
results were high in all three samples (157 µg ­L−1 (EF-UH),
the manuscript or in the Supplementary Material. All collected water
199 µg ­L−1 (EF-WH), and 339 µg ­L−1 (EF-ML)), possibly samples were discarded two years after analysis.
because its use as a component of detergents and bleaching
agents (Waggott 1969; Hasenmueller and Criss 2013). Code availability Not applicable.
Overall, the results of this study suggest distinct human-
related source inputs to local surface water. Some sampling Declarations
locations exhibited natural compositions (Mn 1 and LC),
others can receive wastewater from outside the university Conflict of interest The authors declare that they have no competing
interest.
campus (LB and LI), some display as a mixture (RRA, AR),
and other ones contain Gd anomalies (LS, LM, and RSL)
which we attribute to contamination mostly from hospital
effluents.

13
607 Page 16 of 18 Environmental Earth Sciences (2021) 80:607

References Campinas (SP). Revista Águas Subterrâneas 29(2):224–243.

https://​doi.​org/​10.​14295/​ras.​v29i2.​28409
Byrne RH, Sholkovitz ER (1996) Marine chemistry and geochemistry
Alhajjar BJ, Chesters G, Harkin JM (1990) Indicators of chemical pol-
of the lanthanides. Handb Phys Chem Rare Earths 23:497–593.
lution from septic systems. Groundwater 28(4):559–568. https://​
https://​doi.​org/​10.​1016/​S0168-​1273(96)​23009-0
doi.​org/​10.​1111/j.​1745-​6584.​1990.​tb017​12.x
Campos FF, Enzweiler J (2013) Gadolinium anomalies in Atibaia River
Alkan A, Alkan N, Yanar B (2020) Investigation of pollution levels
water (SP, Brazil)—Goldschmidt abstracts 2013. Mineral Mag
originated from anthropogenic gadolinium in Ankara Stream.
77:815. https://​doi.​org/​10.​1180/​minmag.​2013.​077.5.3
Environ Sci Pollution Res 27:23677–23685. https://​doi.​org/​10.​
Campos FF, Enzweiler J (2016) Anthropogenic gadolinium anomalies
1007/​s11356-​020-​08702-7
and rare Earth elements in the water of Atibaia River and Anhu-
Amorim AM, Sodré FF, Rousseau TCC, Maia PD (2019) Assessing
mas Creek, Southeast Brazil. Environ Monit Assess 188:281.
rare-earth elements and anthropogenic gadolinium in water sam-
https://​doi.​org/​10.​1007/​s10661-​016-​5282-7
ples from an urban artificial lake and its tributaries in the Brazilian
Cepagri—CENTRO DE PESQUISAS METEOROLÓGICAS E
federal district. Microchem J 148(March):27–34. https://​doi.​org/​
CLIMÁTICAS APLICADAS A AGRICULTURA (2017) Pre-
10.​1016/j.​microc.​2019.​04.​055
visões do tempo e precipitação—Estação meteorológica Uni-
Andrade RLB, Hatje V, Pedreira RMA, Böning P (2020) REE frac-
camp. https://​www.​agrit​empo.​gov.​br/​agrit​empo/​jsp/​Pesqu​isaCl​
tionation and human Gd footprint along the continuum between
ima/​index.​jsp?​sigla​UF=​SP. Accessed 20 Oct 2017
Paraguaçu River to coastal South Atlantic waters. Chem Geol
Cotta AJB, Enzweiler J (2012) Classical and new procedures of whole
532:119303. https://​doi.​org/​10.​1016/j.​chemg​eo.​2019.​119303
rock dissolution for trace elements analysis by ICP-MS. Geostand
APHA (1998) Standard Methods for the Examination of Water and
Geoanal Res 36:27–50. https://​doi.​org/​10.​1111/j.​1751-​908X.​
Wastewater, 20th edn. American Public Health Association,
2011.​00115.x
American Water Works Association and Water Environmental
Deberdt S, Viers J, Dupré B (2002) New insights about the rare
Federation, Washington DC
earth elements (REE) mobility in river waters. Bull La Société
Barber LB, Murphy SF, Verplanck PL, Sandstrom MW, Taylor HE,
Géologique Fr 173:147–160. https://​doi.​org/​10.​2113/​173.2.​147
Furlong ET (2006) Chemical loading into surface water along a
Dia A, Gruau G, Olivié-Lauquet G, Riou C, Molénat J, Curmi P (2000)
hydrological, biogeochemical, and land use gradient—a holistic
The distribution of rare earth elements in groundwaters: assessing
watershed approach. Environ Sci Technol 40:475–486. https://d​ oi.​
the role of source rock composition, redox changes and colloidal
org/​10.​1021/​es051​270q
particle. Geochim Cosmochim Acta 64:4131–4151. https://​doi.​
Bau M, Dulski P (1996) Anthropogenic origin of positive gadolinium
org/​10.​1016/​S0016-​7037(00)​00494-4
anomalies in river waters. Earth Planet Sci Lett 143:245–255.
Dulski P, Möller P, Pekdeger A (2011) Comparison of gado-
https://​doi.​org/​10.​1016/​0012-​821X(96)​00127-6
pentetic acid (Gd-DTPA) and bromide in a dual-tracer field
Bau M, Möeller P, Dulski P (1997) Yttrium and lanthanides in east-
experiment. Hydrogeol J 19:823–834. https://​doi.​org/​10.​1007/​
ern Mediterranean seawater and their fractionation during redox
s10040-​011-​0713-6
cycling. Mar Chem 56:123–131. https://​doi.o​ rg/​10.​1016/​S0304-​
Elbaz-Poulichet F, Seidel J, Othoniel C (2002) Occurrence of an
4203(96)​00091-6
anthropogenic gadolinium anomaly in river and coastal waters
Bau M, Knappe A, Dulski P (2006) Anthropogenic Gadolinium as
of Southern France. Water Res 36:1102–1105. https://​doi.​org/​10.​
a micropollutant in River Waters in Pennsylvania and in Lake
1016/​S0043-​1354(01)​00370-0
Erie, Northeastern United States. Chemie Der Erde Geochemistry
Frias MG, Enzweiler J, Marteleto TP (2017) Hidrogeoquímica das
66:143–152. https://​doi.​org/​10.​1016/j.​chemer.​2006.​01.​002
águas superficiais do campus da Unicamp (SP) e sua relação com-
Beck S, Trog S, Knizia S, Linscheid MW (2017) Fragmentation behav-
posicional com o uso e a ocupação do solo. In: XXV Congresso
ior of DOTA complexes under different activation conditions. J
de Iniciação Científica da Unicamp. https://​doi.​org/​10.​19146/​
Mass Spectrom 52:442–451. https://​doi.​org/​10.​1002/​jms.​3942
pibic-​2017-​78495
Bichler A, Muellegger C, Brünjes R, Hofmann T (2016) Quantification
Gaillardet J, Viers J, Dupré B (2014) Trace element in river waters.
of river water infiltration in shallow aquifers using acesulfame and
In: Drever JI (ed) Surface and ground water, weathering, erosion
anthropogenic gadolinium. Hydrol Process 30:1742–1756. https://​
and soils. Treatise on geochemistry, vol 5. Elsevier, Pergamon.
doi.​org/​10.​1002/​hyp.​10735
https://​doi.​org/​10.​1016/​B0-​08-​043751-​6/​05165-3 (eds: Holland
Birka M, Wehe CA, Telgmann L, Sperling M, Karst U (2013) Sensitive
HD, Turekian KK)
quantification of gadolinium-based magnetic resonance imaging
Goldstein SJ, Jacobsen SB (1988) Rare earth elements in river waters.
contrast agents in surface waters using hydrophilic interaction liq-
Earth Planet Sci Lett 89:35–47. https://​doi.​org/​10.​1016/​0012-​
uid chromatography and inductively coupled plasma sector field
821X(88)​90031-3
mass spectrometry. J Chromatogr A 1308:125–131. https://​doi.​
Goullé JP, Saussereau E, Mahieu L, Cellier D, Spiroux J, Guerbet
org/​10.​1016/j.​chroma.​2013.​08.​017
M (2012) Importance of anthropogenic metals in hospital and
Birka M, Wehe CA, Hachmöller O, Sperling M, Karst U (2016) Trac-
urban wastewater: its significance for the environment. Bull
ing gadolinium-based contrast agents from surface water to
Environ Contam Toxicol 89:1220–1224. https://​doi.​org/​10.​1007/​
drinking water by means of speciation analysis. J Chromatogr
s00128-​012-​0829-y
A 1440:105–111. https://​doi.​org/​10.​1016/j.​chroma.​2016.​02.​050
Hasenmueller EA, Criss RE (2013) Multiple sources of boron in urban
Boester U, Rüde TR (2020) Utilize gadolinium as environmental
surface waters and groundwaters. Sci Total Environ 447:235–247.
tracer for surface water-groundwater interaction in Karst. J Con-
https://​doi.​org/​10.​1016/j.​scito​tenv.​2013.​01.​001
tam Hydrol 235:103710. https://​doi.​org/​10.​1016/j.​jconh​yd.​2020.​
Hatt BE, Fletcher TD, Walsh CJ, Taylor SL (2004) The influence of
103710
urban density and drainage infrastructure on the concentrations
Brünjes R, Bichler A, Hoehn P, Lange FT, Brauch H-J, Hofmann T
and loads of pollutants in small streams. Environ Manage 34:12–
(2016) Anthropogenic gadolinium as a transient tracer for inves-
24. https://​doi.​org/​10.​1007/​s00267-​004-​0221-8
tigating riverbank filtration. Sci Total Environ 571:1432–1440.
Heberer T (2002) Occurrence, fate, and removal of pharmaceutical res-
https://​doi.​org/​10.​1016/j.​scito​tenv.​2016.​06.​105
idues in the aquatic environment: a review of recent research data.
Bulia IL, Enzweiler J (2015) Elementos Terras Raras e Urânio Em
Toxicol Lett 131:5–17. https://​doi.​org/​10.​1016/​s0378-​4274(02)​
Águas Subterrâneas sob Influência de Aquíferos Distintos em
00041-3

13
Environmental Earth Sciences (2021) 80:607 Page 17 of 18 607

IG—Instituto Geológico (1993) Subsídios do meio físico-geológico ao gadolinium anomalies. Sci Total Environ 656:409–420. https://​
planejamento do município de Campinas (SP), São Paulo. https://​ doi.​org/​10.​1016/j.​scito​tenv.​2018.​11.​343
www.​infra​estru​turam​eioam​biente.​sp.​gov.​br/​insti​tutog​eolog​ico/​ Leybourne MI, Johannesson KH (2008) Rare earth elements (REE)
1993/​01/​insti​tuto-​geolo​gico-​sp-2/. Accessed 27 Jan 2018 and yttrium in stream waters, stream sediments, and Fe–Mn oxy-
ISO 8245 (1999) Water quality—Guidelines for the determination of hydroxides: Fractionation, speciation, and controls over REE + Y
total organic carbon (TOC) and dissolved organic carbon (DOC). patterns in the surface environment. Geochim Cosmochim Acta
https://​www.​iso.​org/​stand​ard/​29920.​html. Acessed 24 Apr 2014 72:5962–5983. https://​doi.​org/​10.​1016/j.​gca.​2008.​09.​022
Knappe A, Sommer-von Jarmerstedt C, Pekdeger A, Bau M, Dulski Leybourne MI, Goodfellow WD, Boyle DR, Hall GM (2000) Rapid
P (1999) Gadolinium in aquatic systems as indicator for sewage development of negative Ce anomalies in surface waters and
water contamination. In: Armannsson H (ed) Geochemistry of contrasting REE patterns in groundwaters associated with Zn–Pb
the earth’s surface. Proceedings of the 5th international sympo- massive sulphide deposits. Appl Geochem 15:695–723. https://​
sium, Reykjavic, 16–20 August. 1st edn. CRC Press, Rotterdam, doi.​org/​10.​1016/​S0883-​2927(99)​00096-7
pp 187–190 Magnusson B, Näykki T, Hovind H, Krysell M, Sahlin E (2017) Hand-
Knappe A, Möller P, Dulski P, Pekdeger A (2005) Positive gadolinium book for calculation of measurement uncertainty in environmental
anomaly in surface water and ground water of the urban area Ber- laboratories, Nordtest Report TR 537 (ed. 4). www.​nordt​est.​info.
lin, Germany. Chem Erde 65:167–189. https://​doi.​org/​10.​1016/j.​ Accessed 17 Oct 2019
chemer.​2004.​08.​004 Massmann G, Knappe A, Richter D, Pekdeger A (2004) Investigat-
Kraemer D, Kopf S, Bau M (2015) Oxidative mobilization of cerium ing the influence of treated sewage on groundwater and surface
and uranium and enhanced release of “immobile” high field water using wastewater indicators in Berlin, Germany. Acta
strength elements from igneous rocks in the presence of the bio- Hydrochimica Et Hydrobiological 32(4–5):336–350. https://​
genic siderophore desferrioxamine B. Geochim Cosmochim Acta doi.​org/​10.​1002/​aheh.​20040​0543
165:263–279. https://​doi.​org/​10.​1016/j.​gca.​2015.​05.​046 Merschel G, Bau M, Baldewein L, Dantas EL, Walde D, Bühn B
Kulaksız S, Bau M (2007) Contrasting behavior of anthropogenic gado- (2015) Tracing and tracking wastewater-derived substances
linium and natural rare earth elements in estuaries and the gado- in freshwater lakes and reservoirs: anthropogenic gadolinium
linium input into the North Sea. Earth Planet Sci Lett 260:361– and geogenic REEs in Lake Paranoá. Brasilia c r Geoscience
371. https://​doi.​org/​10.​1016/j.​epsl.​2007.​06.​016 347:284–293. https://​doi.​org/​10.​1016/j.​crte.​2015.​01.​004
Kulaksiz S, Bau M (2011a) Rare earth elements in the Rhine River, Meybeck M (1982) Carbon, nitrogen, and phosphorus transport by
Germany: first case of anthropogenic lanthanum as a dissolved world rivers. Am J Sci 282:401–450
microcontaminant in the hydrosphere. Environ Int 37:973–979. Möller P, Dulski P (2010) Transmetallation of Gd-DTPA by Cu,
https://​doi.​org/​10.​1016/j.​envint.​2011.​02.​018 Y, and lanthanides and its impact on the hydrosphere. Appl
Kulaksiz S, Bau M (2011b) Anthropogenic gadolinium as a micro- Geochem 25:48–59. https://​doi.​org/​10.​1016/j.​apgeo​chem.​2009.​
contaminant in tap water used as drinking water in urban areas 09.​027
and megacities. Appl Geochem 26:1877–1885. https://d​ oi.o​ rg/1​ 0.​ Möller P, Dulski P, Bau M, Knappe A, Pekdeger A, Sommer-von
1016/j.​apgeo​chem.​2011.​06.​011 Jarmersted C (2000) Anthropogenic gadolinium as a conservative
Kulaksiz S, Bau M (2013) Anthropogenic dissolved and colloid/nano- tracer in hydrology. J Geochem Explor 69–70:409–414. https://​
particle-bound samarium, lanthanum and gadolinium in the Rhine doi.​org/​10.​1016/​S0375-​6742(00)​00083-2
River and the impeding destruction of the natural rare Earth ele- Mortatti BC, Enzweiler J (2019) Major ions and rare earth elements
ment and distribution in rivers. Earth Planet Sci Lett 632:43–50. hydrogeochemistry of the Atibaia and Jaguari rivers subbasins
https://​doi.​org/​10.​1016/j.​epsl.​2012.​11.​033 (Southeast Brazil). Appl Geochem. https://​doi.​org/​10.​1016/j.​
Kümmerer K (2001) Pharmaceuticals in the environment—sources, apgeo​chem.​2019.​104461
fate, effects and risks editors: Kümmerer, K. Springer, New York. Neilands JB (1995) Siderophores: structure and function of microbial
https://​doi.​org/​10.​1007/​978-3-​662-​04634-0 iron transport compounds. J Biol Chem 270:26723–26726. https://​
Kümmerer K, Helmers E (2000) Hospital effluents as a source of gado- doi.​org/​10.​1074/​jbc.​270.​45.​26723
linium in the aquatic environment. Environ Sci Technol 34:573– Nelson BJ, Wood SA, Osiensky JL (2003) Partitioning of REE between
577. https://​doi.​org/​10.​1021/​es990​633h solution and particulate matter in natural waters: a filtration study.
Künnemeyer J, Terborg L, Nowak S, Telgmann L, Tokmak F, Krämer J Solid-State Chem 171:51–56. https://​doi.​org/​10.​1016/​S0022-​
BK, Günsel A, Wiesmüller GA, Waldeck J, Bremer C, Krast U 4596(02)​00145-7
(2009) Analysis of the contrast agent magnevist and its transmeta- Oksendal A, Hals P (1993) Biodistribution and toxicity of MR imaging
lation products in blood plasma by capillary electrophoresis/elec- contrast media. J Magn Reson Imaging 3:157–165. https://d​ oi.o​ rg/​
trospray ionization time-of-flight mass spectrometry. Anal Chem 10.​1002/​jmri.​18800​30128
81(9):3600–3607. https://​doi.​org/​10.​1021/​ac802​7118 Ort C, Lawrence MG, Rieckermann J, Joss A (2010) Sampling for
Lawrence MG, Greig A, Collerson KD, Kamber BS (2006) Rare earth pharmaceuticals and personal care products (PPCPs) and illicit
element and yttrium variability in South East Queensland water- drugs in wastewater systems: are your conclusions valid? A criti-
ways. Aquat Geochem 12(1):39–72. https://​doi.​org/​10.​1007/​ cal review. Environ Sci Technol 44(16):6024–6035. https://​doi.​
s10498-​005-​4471-8 org/​10.​1021/​es100​779n
Lawrence MG, Ort C, Keller J (2009) Detection of anthropogenic gado- Pedreira MA, Pahnke K, Böning P, Hatje V (2018) Tracking hospital
linium in treated wastewater in South East Queensland. Australia effluent-derived gadolinium in Atlantic coastal waters off Brazil.
Water Res 43:3534–3540. https://​doi.​org/​10.​1016/j.​watres.​2009.​ Water Res 145:62–72. https://​doi.​org/​10.​1016/j.​watres.​2018.​08.​
04.​033 005
Lerat-Hardy A, Coynel A, Dutruch L, Pereto C, Bossy C, Gil-Diaz Pédrot M, Dia A, Davranche M, Bouhnik-Le Coz M, Henin O, Gruau
T, Capdeville MJ, Blanc G, Shäfer J (2019) Rare earth element G (2008) Insights into colloid-mediated trace element release at
fluxes over 15 years into a major European Estuary (Garonne- the soil/water interface. Journal of Colloid and Interfac Science
Gironde, SW France): hospital effluents as a source of increasing 325:187–197. https://​doi.​org/​10.​1016/j.​jcis.​2008.​05.​019

13
607 Page 18 of 18 Environmental Earth Sciences (2021) 80:607

Perrotta MM, Salvador ED, Lopes RC, D'Agostino LZ, Peruffo N, Smedley PL (1991) The geochemistry of rare earth elements in ground-
Gomes SD, Sachs LLB, Meira VT, Garcia MGM, Lacerda Filho water from the Carnmenellis área, southwest England. Geochim
JV (2005) Mapa Geológico do Estado de São Paulo, escala Cosmochim Acta 55:2767–2779. https://​doi.​org/​10.​1016/​0016-​
1:750.000. Programa Geologia do Brasil—PGM, CPRM, São 7037(91)​90443-9
Paulo. https://​geosgb.​cprm.​gov.​br/. Accessed 17 June 2019 Song H, Shin WJ, Ryu JS, Shin HS, Chung H, Lee KS (2017) Anthro-
Pourmand A, Dauphas N, Ireland TJ (2012) A novel extraction chroma- pogenic rare earth elements and their spatial distributions in the
tography and MC-ICP-MS technique for rapid analysis of REE, Sc Han River, South Korea. Chemosphere 172:155–165. https://​doi.​
and Y: revising CI-chondrite and Post-Archean Australian Shale org/​10.​1016/j.​chemo​sphere.​2016.​12.​135
(PAAS) abundances. Chem Geol 291:38–54. https://​doi.​org/​10.​ Tepe N, Romero M, Bau M (2014) High-technology metals as emerg-
1016/j.​chemg​eo.​2011.​08.​011 ing contaminants: strong increase of anthropogenic gadolinium
Rabiet M, Brissaud F, Seidel JL, Pistre S, Elbaz-Poulichet F (2005) levels in tap water of Berlin, Germany, from 2009 to 2012. Appl
Deciphering the presence of wastewater in a medium-sized Medi- Geochem 45:191–197. https://d​ oi.o​ rg/1​ 0.1​ 016/j.a​ pgeoc​ hem.2​ 014.​
terranean catchment using a multitracer approach. Appl Geochem 04.​006
20:1587–1596. https://​doi.​org/​10.​1016/j.​apgeo​chem.​2005.​04.​005 Thomsen HS (2017) Are the increasing amounts of gadolinium in sur-
Rabiet M, Togola A, Brissaud F, Seidel JL, Budzinski H, Elbaz- face and tap water dangerous? Acta Radiol 58:259–263. https://​
Poulichet F (2006) Consequences of treated water recycling as doi.​org/​10.​1177/​02841​85116​666419
regards pharmaceuticals and drugs in surface and ground waters Tweedle MF, Eaton SM, Eckelman WC, Gaughan GT, Hagan JJ,
of a medium-sized Mediterranean catchment. Environ Sci Technol Wedeking PW, Yost FJ (1988) Comparative chemical structure
40(17):5282–5288. https://​doi.​org/​10.​1021/​es060​528p and pharmacokinetics of MRI contrast agents. Invest Radiol
Rabiet M, Brissaud F, Seidel JL, Pistre S, Elbaz-Poulichet F (2009) 23(Suppl):236–239
Positive gadolinium anomalies in wastewater treatment plant efflu- Verplanck PL, Taylor HE, Nordstrom DK, Barber LB (2005) Aque-
ents and aquatic environment in the Herault watershed (South ous stability of gadolinium in surface waters receiving sewage
France). Chemosphere 75:1057–1064. https://​doi.​org/​10.​1016/j.​ treatment plant effluent, Boulder Creek, Colorado. Environ Sci
chemo​sphere.​2009.​01.​036 Technol 39:6923–6929. https://​doi.​org/​10.​1021/​es048​456u
Richards S, Withers PJA, Paterson E, McRoberts CW, Stutter M (2017) Waggott A (1969) An investigation of the potential problem of increas-
Potential tracers for tracking septic tank effluent discharges in ing boron concentrations in rivers and water courses. Water Res
watercourses. Environ Pollut 228:245–255. https://​doi.​org/​10.​ 3(10):749–765. https://​doi.​org/​10.​1016/​0043-​1354(69)​90039-6
1016/j.​envpol.​2017.​05.​044 Yechicheyan D, Bossy C, Le Coz MB, Douchet C, Granier G, Heim-
Rogowska J, Olkowska E, Ratajczyk W, Wolska L (2018) Gadolinium burger A, Lacan F, Lanzanova A, Rousseau TCC, Seidel JC et al
as a new emerging contaminant of aquatic environment. Environ (2013) A compilation of silicon, rare earth element and twenty-
Toxicol Chem 37:1523–1534. https://​doi.​org/​10.​1002/​etc.​4116 one other trace element concentrations in the natural river water
Schmidt K, Bau M, Merschel G, Tepe N (2019) Anthropogenic gado- reference material SLRS-5 (NRC-CNRC). Geostand Geoanal Res
linium in tap water and in tap water-based beverages from fast- 37:449–467. https://​doi.​org/​10.​1111/j.​1751-​908X.​2013.​00232.x
food franchises in six major cities in Germany. Sci Total Environ Zhu Y, Hoshino M, Yamada H, Itoh A, Haraguchi H (2004) Gadolin-
687:1401–1408. https://​doi.​org/​10.​1016/j.​scito​tenv.​2019.​07.​075 ium anomaly in the distributions of rare earth elements observed
Shellock FG, Kanal E (1999) Safety of magnetic resonance imaging for coastal seawater and river waters around Nagoya City. Bull
contrast agents. J Magn Reson Imaging 10(3):477–484. https://​ Chem Soc Jpn 77:1835–1842. https://​doi.​org/​10.​1246/​bcsj.​77.​
doi. ​ o rg/ ​ 1 0. ​ 1 002/ ​ ( sici) ​ 1 522- ​ 2 586(199909) ​ 1 0:3% ​ 3 c477:: ​ a id-​ 1835
jmri33%​3e3.0.​co;2-e
Sholkovitz ER (1995) The aquatic chemistry of rare earth elements in Publisher's Note Springer Nature remains neutral with regard to
rivers and estuaries. Aquat Geochem 1:1–34. https://​doi.​org/​10.​ jurisdictional claims in published maps and institutional affiliations.
1007/​BF010​25229
SIC—Serviço de Informação ao Cidadão (Unicamp) (2019). http://​
www.i​ nform ​ acoes.u​ nicam ​ p.b​ r/i​ ndex.p​ hp/o-s​ ic/. Accessed 25 June
2019

13