Water Research 162 (2019) 170e179

Contents lists available at ScienceDirect

Water Research
journal homepage: www.elsevier.com/locate/watres

White bean (Phaseolus vulgaris L.) as a sorbent for the removal of zinc
from rainwater
Mo
nica P.S. Ferreira a, 1, Patrícia S.M. Santos a, *, 1, Maria T. Caldeira a, Ana C. Estrada b,
~o P. da Costa a, Teresa Rocha-Santos a, Armando C. Duarte a
Joa
a
CESAM & Department of Chemistry, University of Aveiro, Campus Universita
rio de Santiago, 3810-193, Aveiro, Portugal
b
rio de Santiago, 3810-193, Aveiro, Portugal
CICECO & Department of Chemistry, University of Aveiro, Campus Universita

a r t i c l e i n f o a b s t r a c t

Article history: The present work aimed to assess the sorption capacity of the common white bean (Phaseolus vulgaris L.)
Received 19 March 2019 to remove Zn(II) from rainwater, rendering it suitable for use in buildings, and the efficiency of the
Received in revised form process was evaluated for two initial Zn(II) concentrations, representing high (100 mg L-1) and very high
21 June 2019
(500 mg L
1) levels of Zn(II) in rainwater. The effects of the amount of beans (1, 5 and 10 beans per 50 mL),
Accepted 24 June 2019
Available online 26 June 2019
as well as the initial pH values of the zinc solution [acid (4), neutral (5.6) and basic (7) for atmospheric
waters] were also assessed. The removal of Zn from water was affected by the change in pH values. When
5 and 10 beans were used, after 4 h and 2 h of contact time, respectively, the accumulated Zn(II) on the
Keywords:
Sorption
beans was released back into the solution, and this release occurred first for the highest tested pH value.
Zinc The sorption rate of Zn(II) from the solution increased with the increasing amount of beans, but for 5 and
White bean 10 beans this only took place up to 4 h and 2 h, respectively. Furthermore, the removal percentages of
Biosorbent Zn(II) increased with the increase of the initial concentrations of the metal in water. Kinetic studies
Rainwater revealed that a pseudo-first-order model provided the best fitting for the experimentally obtained
values. Fourier transform infrared spectroscopy coupled with attenuated total reflectance (FTIR-ATR)
analyses of the bean shells and cores indicated that contact with a Zn(II) solution did not cause notable
alterations to the chemical structures of these bean components. Scanning electron microscopy (SEM)
and energy dispersive X-ray spectroscopy (EDX) analyses suggested that the process of sorption occurred
at the surface of the beans (shells). The results obtained in this study also suggest that the matrix of
rainwater samples did not interfere with the removal of metal, and that the process of the removal of
Zn(II) by the white beans can be applied to real samples. On the whole, results indicate that for the
removal of Zn(II) from rainwater, 1 bean up to 6 h, or 5 and 10 beans up to 2 h can be used per 50 mL for
the removal of up to 60% of Zn(II) present in water, thus constituting a viable solution for the effective
reduction of this metal in rainwater.
© 2019 Elsevier Ltd. All rights reserved.

1. Introduction suggested in the past (Angrill et al., 2017). In fact, various studies
have evidenced the potential of harvesting rainwater for domestic
Climate change, urban development and higher water con- purposes (Simmons et al., 2001; Busset et al., 2011). Moreover, the
sumption rates are commonly acknowledged factors presently rooftops represent an important percentage of impermeable area in
putting stress on existing fresh water resources. Thus, the scarcity urban zones, offering a significant possibility for rainwater collec-
of fresh water all over the world (Hoekstra et al., 2012) highlights tion, which may be reused in buildings. However, although rain-
the need to find alternative water resources and the possibility of water may be considered in the first instance as distilled water,
reusing rainwater as a way to overcome this issue has been several types of chemical contaminants have been found in har-
vested rainwater, such as polycyclic aromatic hydrocarbons and
pesticides (Basheer et al., 2003), as well as heavy metals (Santos
et al., 2008). Thus, for reusing rainwater in buildings, the removal
* Corresponding author.
of such contaminants should be performed in order to safeguard
E-mail address: [email protected] (P.S.M. Santos).
1
Equal contributors. human health. The European Union (EU) does not have specific

https://doi.org/10.1016/j.watres.2019.06.064
0043-1354/© 2019 Elsevier Ltd. All rights reserved.
 M.P.S. Ferreira et al. / Water Research 162 (2019) 170e179 171

legislation or guidelines for the reuse of rainwater in buildings, but the sorbent amount (three masses); the initial concentrations of
the legislation for drinking water (Council Directive 98/83/EC) can metal in water (two concentrations, representing very high and
be used as a reference point, certainly ensuring the safety of such high levels of zinc in rainwater); and the contact time between
treated waters. sorbent and water containing zinc. Some experiments were also
From the potentially toxic elements, zinc has been recurrently performed on real rainwater samples. The variation of zinc con-
found as a major metal present in rainwater (Chon et al., 2015; centration in the water was evaluated using atomic absorption
Omrani et al., 2017). Furthermore, various studies in different parts spectroscopy (AAS), and kinetic sorption studies were carried out.
of the world have showed that rainwater frequently exhibits high Chemical, structural and morphological features of the white beans
levels of zinc (e.g. Santos et al., 2008; Chon et al., 2015; Omrani tested along the process of sorption were also evaluated using
et al., 2017), some of those much higher than the levels recom- Fourier transform infrared (FTIR-ATR) spectroscopy, scanning
mended for drinking water, namely, in Portugal. According to the electron microscopy (SEM) and electron dispersive X-ray spec-
Portuguese legislation, water for human consumption must have a troscopy (EDX).
Zn(II) concentration lower than 5.0 mg L
1 (Portuguese Decree-Law
n.º 236/98). However, the EU legislation for drinking water (Council 2. Material and methods
Directive 98/83/EC) does not specify the recommended maximum
level for zinc, allowing this parameter to be legislated by each 2.1. Biosorbent preparation
member state. As zinc may have toxic effects on human health,
including fever and gastrointestinal distress (World Health The white beans used were Phaseolus vulgaris L., and were ac-
Organization, 1996), its presence in rainwater becomes a matter quired from a large commercial retailer. The need to wash the beans
of concern when its intended use is in buildings, highlighting that prior to their use, due to the possibility of releasing organic matter
rainwater must undergo treatment for metal removal before its use. into the water, was assessed analyzing the water after different
The removal of contaminants from rainwater requires a meth- contact periods with the beans (1 h, 2 h and 4 h), by UVeVis.
odology that makes use of materials able to retain the contami- Washed beans were rinsed with ultrapure water three times, and
nants. To the best of our knowledge, few materials have been tested then dried in an oven at 30  C during 2 h, 4 h, or 24 h. In the other
for the retention of contaminants from rainwater: peat (Cho et al., hand, the unwashed white beans were not subjected to any treat-
2014), Tecomella leaves (Brahman et al., 2016), Fe3O4@MOF- ment following their removal from the original packaging. Then,
235(Fe)eOSO3H (Moradi et al., 2016), Zeolite(Na)@nano-Fe (Gan 2.6 g of sorbent (equivalent to 5 beans) were added to 50 mL of
et al., 2012), and activated carbon (Kus et al., 2013a,b) for sorp- ultrapure water (at 100 mL Erlenmeyer flasks), and inserted in an
tion of inorganic contaminants; and magnetic (Heidari et al., 2012; incubator at 22  C, with agitation at 150 rpm, and in the absence of
Wang et al., 2015; Yang et al., 2014, 2019; Zhang et al., 2016, 2018) light. The UVeVis results (see Supplementary Material (SM),
and gold (Feng et al., 2010; Yang et al., 2016) nanoparticles for Fig. SM-1) showed that up to 4 h of contact of the beans with the
sorption of aromatic pollutants. However, only activated carbon water, the washed white beans dried at 30  C during 4 h and 24 h
was tested for the removal of zinc from rainwater, and showed not were the most suitable for the purpose of this work. For this reason,
to be efficient (Kus et al., 2013b). throughout the posterior experiments concerning the removal of
Biosorption is a biotechnological innovation with the potential zinc from water, white beans that had washed and dried at 30  C
for the removal of contaminants from water, due to the ability of during 24 h were used.
biomaterials to accumulate contaminants via metabolically-
independent mediated or physicochemical pathways (Fomina and 2.2. Sorption experiments
Gadd, 2014). Furthermore, biomaterials can be non-hazardous
and readily available from cheap sources. Various biosorbents Batch sorption studies were performed in an incubator at 22  C,
have been studied for the removal of contaminants from water, with agitation at 150 rpm and in the absence of light, using a known
such as bacteria (including both Gram-positive and Gram-negative mass of Zn(II) in 50 mL of ultrapure water (in 100 mL Erlenmeyer
bacteria, as well as cyanobacteria), fungi (e.g. mushrooms, yeast), flasks). The Zn(II) solutions used had final concentrations of
algae (e.g. micro-algae and macro-algae), industrial wastes (e.g. 0.1 mg L
1 or 0.5 mg L
1, and were freshly prepared on a daily basis
fermentation and food wastes, activated and anaerobic sludges), from a stock solution of Zn(NO3)2.6H2O (1000 mg L
1 in 1% of nitric
agricultural wastes (e.g. fruit/vegetable wastes, rice straw, wheat acid (65% suprapure)). Such concentrations were selected to
bran, soybean hulls), natural residues (e.g. plant residues, sawdust, represent a high level (100 mg L
1), and a very high level
tree barks, weeds) and other materials (e.g. chitosan, cellulose) (500 mg L
1) of zinc found in rainwater. The masses of biosorbent
(Park et al., 2010). Furthermore, some biomaterials, such as fungi, used to perform the experiments were 0.5 g (1 bean), 2.6 g (5
bacteria and algae, were already tested for the removal of zinc from beans) and 5.3 g (10 beans), corresponding to a bean mass-to-
residual, coastal and esturine waters (e.g. Aftab et al., 2013; Mameri volume ratio (m/V) of 10 g L
1, 52 g L
1 and 106 g L
1, respectively.
et al., 1999; Saiano et al., 2005; Suzuki et al., 2005). However, as Since the pH value of water directly affects metal solubility and the
rainwater is considered a “clean” water, the biomaterial chosen for dissociation degree of functional groups located on the surface of
the removal of contaminants should not introduce additional biosorbent, the effect of pH on the removal of Zn(II) from water by
organic and/or inorganic compounds, other than those already the white beans was evaluated for three initial pH values: 4.0 (acid
present. pH value for atmospheric waters); 5.6 (neutral pH for atmospheric
The aim of this study was to assess the sorption efficiency of the waters), and 7.0 (basic pH value for atmospheric waters) (Seinfeld
common white bean (Phaseolus vulgaris L.) to remove Zn(II) from and Pandis, 1998). The pH of solutions was adjusted with nitric
rainwater, under realistic conditions and concentrations. This acid and sodium hydrogencarbonate solutions.
biomaterial was chosen because it is cheap and can be easily Experiments started when a known mass of white beans was
separated from water after treatment, allowing its use as a home- added to a solution of known concentration of Zn(II) at a defined
made method for rainwater treatment and subsequent reuse of this pH, and inserted in the incubator. The experiments were stopped at
type of water in buildings, namely, for domestic purposes. In this specific times of 0.5 h, 1 h, 1.5 h, 2 h, 4 h, 6 h and 24 h. For such,
study, the following effects were evaluated: the initial pH of water aliquots of 20 mL of solution were collected for each period of
containing zinc (acid, neutral and basic pH values for rainwater); interaction time, and then 20 mL of nitric acid 65% were added
172 M.P.S. Ferreira et al. / Water Research 162 (2019) 170e179

(suprapure; to obtain pH < 2), for posterior analysis of the Zn(II) radiation source at 45 kV/40 mA.
concentration by atomic absorption spectroscopy. The initial con-
centration of Zn(II) solution (at 0 h) was also determined. For each 2.4. Evaluation of sorption and kinetic models
of the periods of interaction time, an independent flask was used,
and at least three replicates were made. The normalized average The amount of Zn(II) sorbed by white beans at a given time t (qt,
concentrations of Zn(II) (Ct/C0) obtained for different pH values, mg g
1) was inferred using the mass balance between the initial
masses of biosorbent and initial Zn(II) concentrations were concentration of Zn(II) in solutions (C0, mg L
1) and the concen-
compared using one sided Student's t-test, at 5% significance level. tration remaining after the time of contact t (Ct, mg L
1) with the
The controls were performed with ultrapure water and beans in a biosorbent, as follows:
similar way to the remaining experiments.
To evaluate the chemical, morphological and structural features ðC0
Ct Þ V
qt ¼ (1)
of white beans along the sorption process, bean samples of each m
sampling time were saved, and dried at 30  C, for posterior analysis
using FTIR-ATR, SEM, EDX and XRD. For SEM analyses, images of the where V (L) is the volume of the solution, and m (g) is the mass of
bean shells and bean cores before, and after 1, 4 and 6 h were ac- dried white beans. For most cases, with the Zn(II) removal, Ct seems
quired resorting to experiments using a zinc aqueous solution of to decrease, while qt values increase, until equilibrium values (qe)
10 mg L
1, in order to amplify the sorption behavior of zinc on the are achieved.
beans, as no changes were observed with the lower concentration The results were also compared using the removal percentage
of 500 mg L
1. for each contact time t (Rt):
Sorption experiments were also performed with five real rain-
ðC0
Ct Þ
water samples. Nevertheless, due to the volume of samples avail- Rt ð%Þ ¼  100 (2)
able, the experiments were only performed once for each sample.
C0
Moreover, the experiments were performed at rainwater's natural Experimental data were fitted to pseudo-first-order (Eq. (3)) and
pH (pH 5.1 for sample 1; pH 5.8 for sample 2; pH 6.4 for sample 3; pseudo-second-order (Eq. (4)) kinetic models through the
pH 6.1 for sample 4; and pH 5.5 for sample 5), for the bean mass of following equations (Tran et al., 2017):
2.6 g (5 beans) only, and were stopped at 0.5 h, 1 h, 2 h, 4 h and 6 h.  
The rainwater samples were spiked with 100 mg L
1 (samples: 1, 2, qt ¼ qe 1
e
k1 t (3)
3, 4 and 5) and/or 500 mg L
1 (samples: 1, 2 and 4) of Zn(II), ac-
cording to the volume of the sample available for analysis. The
concentration of Zn(II) in the rainwater samples before and after q2e k2 t
qt ¼ (4)
the period of interaction time with the white beans was also 1 þ qe k2 t
measured by atomic absorption spectroscopy. Regarding the rain-
water samples, they were collected in the city of Estarreja, Portugal, where k1 (h
1) and k2 (g mg
1 h
1) are the pseudo-first-order and
in May 2016 (between 4th and 10th of May). The sample collection the pseudo-second-order rate constants, respectively. Furthermore,
was carried out at 70 cm above ground, through glass funnels with qe (mg/g) and qt (mg/g) are the amount of solute sorbed per gram of
a 30 cm diameter, into 5 L glass bottles. These containers were left sorbent at equilibrium and at any t (h), respectively.
out open to collect the bulk deposition (both wet and dry de-
positions) during approximately 24 h (Santos et al., 2014). After 3. Results and discussion
collection, the samples were transported to the laboratory, and
were immediately filtered through mixed cellulose esters filters 3.1. Effects of the initial concentration of Zn(II), the initial pH and
(Millipore) of 0.45 mm pore size, and were refrigerated at 4  C until the amount of beans on the sorption process
analysis.
Prior to use, all glass material was immersed in a solution of Fig. 1 shows the variation of the normalized concentration (Ct/
NaOH (0.1 M) for 30 min, then rinsed with distilled water, followed C0) of Zn(II) in water with time, t (h), after contact with 1, 5 and 10
by further immersion in a solution of HNO3 (4 M) for 24 h, and then beans, for the initial Zn(II) concentrations of 100 mg L
1 and
rinsed first with distilled water, and, finally, with ultrapure water. 500 mg L
1, and for the initial pH values of 4.0, 5.6 and 7.0. As shown
in Fig. 1, when using 1 and 5 beans per solution, and for both initial
2.3. Analytical methods concentrations, the normalized concentration of Zn(II) tends to
decrease with time, suggesting an uptake of this metal by the
UVeVis spectra, in the range of 200e500 nm, were recorded on beans. However, when 5 beans were used, for the initial concen-
a Shimadzu (Dusseldorf, Germany) Model UV 210 PC spectropho- tration of 100 mg L
1, at a pH value of 7.0, and with a contact time of
tometer using quartz cells of 1 cm path lengths. Ultrapure water 6 h, a clear increase of Zn(II) concentration in the solution
was used as reference in order to obtain the baseline. Atomic ab- (p ¼ 0.001) was observed, which was also slightly observed for the
sorption spectroscopy analysis was carried out using a PerkinElmer initial Zn(II) concentration of 500 mg L
1 (p ¼ 0.04). This suggests
AAnalyst 100 atomic absorption spectrometer (air-acetylene flame) that the Zn(II) previously sorbed by the beans was released back
operated at the 213.9 nm (Zn) analytical line using a Zinc Hallow into the solution. Nevertheless, at 24 h the concentration of Zn(II) in
Cathode Lamp. FTIR-ATR analyses were performed on a Perki- the solution decreased again, suggesting that the sorption process
nElmer (USA) Spectrum BX FTIR instrument. The samples were restarted. For the experiments of 1 and 5 beans, and up to an hour
analyzed within the 4000-550 cm
1 range at a 4 cm
1 resolution, and a half, a slower sorption capacity for water at pH 4.0 than at pH
and sixty-four scans were averaged. For the background spectrum, 7.0 (p < 0.05) was observed, suggesting that at the initial times the
air was used. SEM analyses were carried out using a scanning pH of the water affects the sorption capacity.
electron microscope Hitachi SU-70 operating at an accelerating When 10 beans were used, and for the initial concentrations of
voltage of 15 kV. EDX examinations were performed using an EDX 100 mg L
1 and of 500 mg L
1, the normalized concentration of Zn(II)
Bruker. XRD data were collected using a PAN analytical Empyrean decreased with time up to 2 h. However, the concentration of Zn(II)
X-ray diffractometer equipped with Cu-Ka monochromatic in solution increased after 2 h for all pH values and for both initial
 M.P.S. Ferreira et al. / Water Research 162 (2019) 170e179 173

Fig. 1. Variation of the normalized concentration (Ct/C0) with time (t, h) for different values of pH values (4.0, 5.6 and 7.0), and for the following conditions: C0 Zn(II) of 100 mg L
1
(upper line graphics), and C0 Zn(II) of 500 mg L
1 (bottom line graphics); and m/V of 10 g L
1 (1 bean; left column), 52 g L
1 (5 beans; middle column), and 106 g L
1 (10 beans; right
column). The controls were performed with ultrapure water and beans in a similar way to the remaining experiments.

concentrations, and was more pronounced for the initial concen- initial concentrations of 100 mg L
1 and 500 mg L
1, respectively. For
tration of 100 mg L
1, while for the concentration of 500 mg L
1 the 5 beans, higher percentages of Zn(II) removal were also obtained
increase was less pronounced, tending to stabilize. Moreover, at pH for a contact time of 24 h, and were higher than 71% and 86%, for
7.0, the normalized concentration of Zn(II) at 2 h was statistically the initial concentrations of 100 mg L
1 and 500 mg L
1, respectively.
lower than at 4 h (p < 0.05) for both the initial tested concentrations These results are in line with other findings, which have showed
(100 mg L
1 and 500 mg L
1). The increase of Zn(II) concentration in the uptake of zinc by Phaseolus vulgaris when irrigated with metal
the solution indicated that part of the Zn(II) sorbed by beans was polluted waters (e.g. Tom et al., 2014), or during phytoremediation
then released back into the solution, and this effect seems to have of metal contaminated soil (e.g. Poniedziałek et al., 2010), corrob-
increased with the increasing pH values. orating the tendency of this specie to retain zinc.
Table 1 presents the removal of Zn(II) (Rt, %), the residual con-
centration of Zn(II) (CZn(II), mg L
1) and the calculated amount of 3.2. Kinetic modelling of results
Zn(II) sorbed per gram of bean (qt, mg g
1) for different contact
times, under the tested conditions (two initial concentrations, The amount of Zn(II) sorbed by 1 and 5 white beans at a time t
100 mg L
1 and 500 mg L
1; three initial pH values, 4.0, 5.6, and 7.0; (qt, mg g
1) versus the time, and the fittings of the experimental
three amounts of beans, 1, 5 and 10 beans), which complements data to pseudo-first and pseudo-second order kinetic models are
Fig. 1. Thus, as can be surmised from Table 1 and Fig. 1, results presented in Fig. 2. Since for 10 beans, the fitting of the experi-
suggest that an increase of the amount of beans increases the speed mental data to the kinetic models was not as satisfactory as that
of the removal of Zn(II) from the solution. This can be observed for observed for 1 and 5 beans, these are presented in the SM (Fig. SM-
the contact time of 2 h (considering that up to this time there is no 2). In addition, the fitting parameters obtained for the three
release of Zn(II) into the solution), and depending on the pH used: 1 quantities of white beans used are also summarized in the SM
bean removed between about 31% and 40% for an initial concen- (Table SM-1). Results from the kinetic data obtained for 1 and 5
tration of 100 mg L
1, and 26% and 35% for an initial concentration of beans showed to be satisfactory considering the values of R square
500 mg L
1; using 5 beans, between about 62% and 68% for an initial (R2), which ranged between 0.853 and 0.998 for the pseudo-first-
concentration of 100 mg L
1, and 71% and 73% for an initial con- order model, and between 0.826 and 0.995 for the pseudo-
centration of 500 mg L
1; using 10 beans between about 65% and second-order model. However, as the values of the residual sum
83% for an initial concentration of 100 mg L
1, and 79% and 89% for of squares (RSS) were between 0.036 and 4.712 for the pseudo-
an initial concentration of 500 mg L
1. On the other hand, the in- first-order model, and between 0.075 and 19.296 for the pseudo-
crease of the amount of beans accelerated the release of the Zn(II) second-order model, the results indicate that the pseudo-first-
from the beans back into the solution, highlighted in the case of 5 order model presents a better adjustment to the experimental
beans with a contact time of 6 h, and in the case of 10 beans, with data. Moreover, the results also showed to be kinetically adequate
contact times of 2 h and 4 h for the initial concentrations of for the removal of Zn(II) from water, either using 1 bean or 5 beans,
100 mg L
1 and 500 mg L
1, respectively. Thus, overall, these results as well as, for both initial concentrations tested.
suggest that for 2 h of contact time, higher removal percentages
were attained using larger amounts of beans. On the other hand, 3.3. Characterization of the bean in the sorption process
using a smaller number of beans, more contact time was needed for
the removal of a higher quantity of Zn(II) (high percentages of Zn(II) Fig. 3 shows the SEM images of the bean shells and bean cores
removal were attained for 1 and 5 beans, with no or a small release before, and after 1, 4 and 6 h in a zinc aqueous solution of 10 mg L
1.
of Zn(II) from the beans into the solution, respectively). Therefore, As shown in Fig. 3, the bean shells and bean cores presented
for 1 bean, higher percentages of Zn(II) removal were obtained for a different morphological traits and levels of organization. Before
contact time of 24 h and were higher than 83% and 92%, for the contact with the Zn(II) solution, the bean shells showed a
174 M.P.S. Ferreira et al. / Water Research 162 (2019) 170e179

Table 1
Removal of Zn(II) (Rt), residual concentration of Zn(II) in solution (CZn(II)), and amount of Zn(II) sorbed per gram of white bean (qt), at different times (t), for different initial
concentrations (C0 Zn(II)), for different amounts of white beans (m/V), and for different pH values. The values are expressed as the mean of, at least, three replicate experiments.

pH 4.0 pH 5.6 pH 7.0

C0 Zn(II) (mg L
1) m/V (g L
1) t (h) Rt (%) CZn(II) (mg L
1) qt  103 (mg g
1) Rt (%) CZn(II) (mg L
1) qt  103 (mg g
1) Rt (%) CZn(II) (mg L
1) qt  103 (mg g
1)

100 10 0.5 7 87 0.82 13 78 1.4 16 82 0.18

(1 bean) 1 11 83 1.2 23 70 2.2 25 73 1.0
1.5 21 75 2.0 30 63 2.8 30 68 1.5
2 31 65 2.9 40 54 3.6 34 64 1.8
4 54 43 5.0 61 35 5.5 59 39 4.2
6 72 26 6.7 64 33 5.7 63 36 4.6
24 90 10 8.3 83 16 7.4 86 13 6.8

52 0.5 29 71 0.52 35 82 0.37 36 62 0.68

(5 beans) 1 48 53 0.87 48 52 0.94 57 42 1.0
1.5 54 46 1.0 66 34 1.3 67 32 1.3
2 62 38 1.2 67 33 1.3 68 30 1.3
4 78 22 1.5 73 27 1.4 67 32 1.3
6 44 34 1.2 70 29 1.4 46 51 0.89
24 71 28 1.3 80 20 1.6 79 20 1.5

106 0.5 50 51 0.44 56 45 0.53 48 51 0.44

(10 beans) 1 72 29 0.65 71 30 0.67 68 31 0.63
1.5 79 21 0.72 75 26 0.71 71 28 0.66
2 83 17 0.77 73 27 0.70 65 33 0.61
4 80 19 0.74 59 41 0.57 57 41 0.53
6 61 38 0.57 59 41 0.56 38 59 0.36
24 44 54 0.42 33 67 0.32 48 49 0.46

500 10 0.5 9 440 4.5 10 431 5.1 13 418 6.3

(1 bean) 1 15 410 7.5 17 397 8.1 22 374 11
1.5 21 383 10 22 372 11 31 331 15
2 26 359 12 33 321 16 35 312 16
4 49 248 23 55 216 25 58 203 27
6 65 170 31 70 142 33 71 137 33
24 97 13 46 92 40 42 94 28 44

52 0.5 27 343 2.4 34 311 3.2 46 255 4.3

(5 beans) 1 42 267 3.8 59 199 5.3 55 216 5.1
1.5 55 211 4.9 67 158 6.1 65 166 6.0
2 71 133 6.4 72 133 6.6 73 128 6.7
4 89 50 8.0 84 74 7.7 77 109 7.1
6 86 65 7.7 85 73 7.7 73 131 6.7
24 92 38 8.1 88 55 8.0 86 70 7.8

106 0.5 44 261 1.92 51 235 2.276 59 193 2.7

(10 beans) 1 74 121 3.2 72 135 3.2 74 126 3.3
1.5 83 80 3.6 80 94 3.6 81 88 3.7
2 89 52 3.9 81 89 3.6 79 100 3.6
4 85 70 3.7 80 94 3.6 70 144 3.2
6 85 72 3.7 77 111 3.4 71 141 3.2
24 73 136 3.1 77 111 3.4 79 104 3.5

continuous structure with an irregular porosity and high inter- amorphous silica, which favors the retention of metals (Rocha et al.,
connectivity, while the cores showed an irregular structure with 2013), and that may have been involved in the sorption of zinc from
protuberances. Although after 6 h of contact with the aqueous so- the solution. Fig. 4 shows the EDX spectra of bean shells after 1 h,
lution of zinc morphological alterations on the bean shells were not 4 h and 6 h in a zinc solution. The presence of a Zn peak in the bean
observed, some alterations were noticeable on the bean cores. The shells after contact with the zinc solution is evident, contrary to the
SEM images of the bean cores after 6 h of contact with the aqueous results pertaining to either the bean shells before contact with the
solution of zinc highlight the swelling of protuberances on the core, solution or in the bean cores (see SM, Fig. SM-5). Thus, results
possibly due to the occurrence of the absorption of water. Table 2 indicate that the Zn(II) in the solution may have been adsorbed in
shows the results of the EDX analyses of the bean shells and the bean shells, which suggests the occurrence of an adsorption
cores, and results evidenced that both present a very high content process.
of oxygen, and contained minor amounts of aluminum, iron, In addition, in order to evaluate possible structural changes
magnesium, potassium, and silicon, and also minor quantities of during the sorption process, the FTIR-ATR spectra of bean cores and
calcium for the shells, and phosphorus and sulphur for the cores. bean shells were obtained, before and after 1 bean was placed in
The level of potassium in the core was one order of magnitude contact during 1 h, 2 h, 4 h and 6 h with the Zn(II) solution of
higher than in the shells, while the levels of aluminium, iron and 500 mg L
1 (see SM, Fig. SM-6). FTIR-ATR spectra show the presence
silicon found in the shells were one order of magnitude higher than of some similar bands for the cores and shells of the beans during
those found in the cores (Table 2). In addition, XRD analyses of the the sorption (Stuart, 2004), and show that no chemical structural
bean shells and cores (see SM, Fig. SM-4) evidenced a broad band changes were observed either on the cores or on the shells of the
from 15 to 35 2q diffraction angles, suggesting the presence of beans when subject to up to 6 h of contact with the Zn(II) solution.
 M.P.S. Ferreira et al. / Water Research 162 (2019) 170e179 175

Fig. 2. Kinetic modelling of the sorption process of Zn(II) onto 1 (m/V of 10 g L
1) and 5 (m/V of 52 g L
1) beans, fitting to pseudo-first (solid line) and pseudo-second (dotted line)
order models. Experimental conditions: C0 Zn(II) of 100 mg L
1 (upper line) and 500 mg L
1 (bottom line), and pH values of 4.0 (left column), 5.6 (middle column) and 7.0 (right
column).

3.4. Evaluation of organic matter release by beans Thus, it is possible that the release of organic matter from the beans
into the water may have acted as a vector for the release of Zn(II)
Fig. 5 shows the UVeVis spectra of organic matter released by 1, from the beans back into the solution after 4 h and 2 h of contact
5 or 10 beans into ultrapure water, after 2 h, 4 h and 6 h (the UVeVis time when 5 and 10 beans were used, respectively.
spectra after 1 h and 24 h are presented in SM, Fig. SM-3). As shown
in Fig. 6, increasing the contact time of beans with the water results 3.5. Removal of Zn(II) from rainwater samples
in the appearance of one absorption band located between 250 and
380 nm. Moreover, the band intensity increases with the increasing In order to evaluate the behavior of the Zn(II) removal process in
number/mass of beans used. Thus, considering that the removal of real rainwater, experiments were performed with five different
Zn(II) from rainwater by white beans will not result in the intro- rainwater samples, at their natural pH: pH 5.1, pH 5.8, pH 6.4, pH 6.1
duction of more organic compounds than those already found in and pH 5.5, corresponding to samples 1e5, respectively. The vari-
the water (absorbance values lower than 0.05 A.U.), results suggest ations of the normalized concentration of Zn(II) with the time for
that for each 50 mL 1 bean (m/V of 10 g L
1) up to 6 h, or 5 beans (m/ the two initial concentrations, 100 mg L
1 and 500 mg L
1, and for
V of 52 g L
1) and 10 beans (m/V of 106 g L
1) up to 2 h can be used. the sorbent amount of 5 beans are shown in Fig. 6.
176 M.P.S. Ferreira et al. / Water Research 162 (2019) 170e179

Fig. 3. SEM images of bean shell (upper images) and bean core (bottom images) before, and after 1, 4 and 6 h in a zinc aqueous solution of 10 ppm.

Table 2 ultrapure water at the pH value of 5.6 (Fig. 1) they were, on average,
The unnormalised concentration in weight percent of the elements found in shell about 72%. Thus, the removal percentages obtained with ultrapure
and core of bean by EDX analysis.
water are in the range and near to values obtained with rainwater
Shell (wt.%) Core (wt.%) samples for the initial concentration of 100 mg L
1 and 500 mg L
1,
Oxygen 10.8 20.0 respectively. It should be noted, however, that the rainwater sam-
Magnesium 0.18 0.24 ples contained other metals and ions (data not shown) and the
Aluminium 0.22 0.03 results suggest that they did not affect the zinc adsorption by the
Silicon 0.38 0.02
beans. Furthermore, the concentrations obtained after 2 h were
Potassium 0.34 3.0
Calcium 0.38 e lower than 42 mg L
1 and 175 mg L
1, for the initial concentration of
Iron 0.21 0.07 100 mg L
1 and 500 mg L
1, respectively. Although the concentra-
Phosphorus e 0.59 tions of Zn(II) obtained after the removal were higher than the
e
Sulfur 0.14
levels recommended for drinking water (5 mg L
1; Decree-Law n.º
The values are average of two EDX analysis. 236/98), they correspond to the removal of more than 58%.
Hence, the levels of Zn(II) obtained in rainwater after the herein
described removal process, indicate that the treated water is more
The pH values of the samples are near to the neutral pH value for suitable for use in buildings, namely for domestic purposes (e.g.
rainwater (5.6), and the pH value of sample 3 (6.4) tends towards bathing, washing clothes and dishes, brushing teeth, watering the
the basic pH of rainwater (Seinfeld and Pandis, 1998). Results of the garden, washing pets). Therefore, this process works well under
removal of Zn(II) from the rainwater samples were compared with real conditions and may constitute a viable alternative towards a
the results obtained for the removal of Zn(II) from ultrapure water more efficient use of water resources.
at a pH value of 5.6 and for the sorbent amount of 5 beans (Fig. 1).
Overall, for both initial concentrations, results indicate the same
trend up to 4 h, that is, the concentration of Zn(II) in solution
4. Conclusions
decreased with time. However, after 4 h and up to 6 h of contact, for
both initial concentrations, results showed a decrease of concen-
The removal of Zn(II) from rainwater by the biosorbent white
tration of Zn(II) in the solution, except for samples 3, 4 and 5 at the
initial concentration of 100 mg L
1, where the concentration bean (Phaseolus vulgaris L.) at environmentally relevant concen-
trations and conditions was evaluated, and results led to the
increased. This increase evidenced the release of Zn(II) from the
beans back into the solution, which was also slightly noted for the following conclusions:
same initial concentration (100 mg L
1) in ultrapure water at pH
values of 4.0 and 5.6, and clearly at pH 7.0 (Fig. 1). These findings a) White beans were able to reduce the low Zn(II) levels (100
and 500 mg L
1) present in water, and the removal percent-
highlight that, after 6 h of contact time between the beans and the
sample, the release of Zn(II) from the beans back into the solution is ages of Zn(II) increased with the increase of the initial con-
affected by the initial concentration of Zn(II). However, taking into centration of the metal in water.
account that 5 beans for the removal of Zn(II) from rainwater were b) The sorption rate of Zn(II) from the solution increased with
used, and that no more than 2 h of contact time should be used increasing sorbent mass (1, 5 or 10 beans), but for 5 and 10
beans this occurred only up to 4 h and 2 h, respectively.
(Fig. 5), the percentages of the removal of Zn(II) obtained for the
initial concentration of 100 mg L
1 were between 58% and 85%, c) The removal of Zn(II) from water was not affected by its
initial pH value only when 1 bean per 50 mL of water was
while for the experiments in ultrapure water for the pH value of 5.6
(Fig. 1), the percentages of the removal of Zn(II) were on average used. When 5 and 10 beans per 50 mL of water were used,
the initial pH of the water affected the removal of Zn(II) after
about 66%. Regarding the percentages of the removal of Zn(II) from
rainwater samples for the initial concentration of 500 mg L
1, they 4 h and 2 h of contact time, with the release of zinc from the
beans back into the solution occurring first for the water
were between 65% and 69% after 2 h of contact time, while in
samples with the basic pH for atmospheric waters.
 M.P.S. Ferreira et al. / Water Research 162 (2019) 170e179 177

Fig. 4. EDX spectra of bean shell before, and after 1, 4 and 6 h in a zinc aqueous solution of 10 ppm.

Fig. 5. UVeVis spectra of organic matter released by 1 (m/V of 10 g L
1), 5 (m/V of 52 g L
1) and 10 (m/V of 106 g L
1) beans to ultrapure water (50 mL), after 2 h, 4 h and 6 h.

d) The kinetics of the Zn(II)/white bean system suggested that f) The release of the organic matter of beans into the water was
generally the pseudo-first-order reaction model was the best evaluated by UVeVis, and the results highlighted that, for 1
prediction for experimental data. bean in 50 mL of water, the amount of organic matter
e) SEM and FTIR-ATR studies did not present any clear alter- released by the bean was acceptable up to 6 h of contact time,
ations on the morphology and on the chemical structures of while for 5 and 10 beans the release of organic matter was
the bean shells during the sorption process, but SEM showed satisfactory only up to 2 h, which in the last cases may have
morphologic changes in the cores, possibly due to the ab- contributed to the release of Zn(II) from the beans back into
sorption of water. EDX analyses revealed that Zn(II) is the solution observed after 4 h and 2 h, respectively.
retained on the bean shells, suggesting that the removal of g) On the whole, the results suggest that for the removal of
Zn(II) from the water is the result of an adsorption process. Zn(II) from rainwater 1 bean (per 50 mL of water) up to 6 h
178 M.P.S. Ferreira et al. / Water Research 162 (2019) 170e179

Appendix A. Supplementary data

Supplementary data to this article can be found online at

https://doi.org/10.1016/j.watres.2019.06.064.

References

Aftab, K., Akhtar, K., Jabbar, A., Bukhari, I.H., Noreen, R., 2013. Physico-chemical
study for zinc removal and recovery onto native/chemically modified Asper-
gillus flavus NA9 from industrial effluent. Water Res. 47, 4238e4246.
Angrill, S., Petit-Boix, A., Morales-Pinzo
n, T., Josa, A., Rieradevall, J., Gabarrell, X.,
2017. Urban rainwater runoff quantity and quality - a potential endogenous
resource in cities? J. Environ. Manag. 189, 14e21.
Basheer, C., Balasubramanian, R., Kee, H., 2003. Determination of organic micro-
pollutants in rainwater using hollow fiber membrane/liquid-phase micro-
extraction combined with gas chromatography e mass spectrometry.
J. Chromatogr. A 1016, 11e20.
Brahman, K.D., Kazi, T.G., Afridi, H.I., Baig, J.A., Abro, M.I., Arain, S.S., Ali, J., Khan, S.,
2016. Simultaneously removal of inorganic arsenic species from stored rain-
water in arsenic endemic area by leaves of Tecomella undulata: a multivariate
study. Environ. Sci. Pollut. Res. 23, 15149e15163.
Busset, G., Vialle, C., Montre
jaud Vignoles, M., Huau, M., Jacobs, S., Sablayrolles, C.,
2011. Health risk assessment case study of trace metals in collected rainwater
for domestic uses. Fresenius Environ. Bull. 20 (9), 2277e2283.
Cho, Y., Kim, S., Park, H., Komarneni, S., Hong, Y., 2014. Removal of inorganic pol-
lutants in rainwater by a peat-derived porous material. J. Porous Mater. 21,
387e394.
Fig. 6. Variation of the normalized concentration (Ct/C0) of Zn(II) in five rainwater Chon, K., Kim, Y., Bae, D.H., Cho, J., 2015. Confirming anthropogenic influences on
samples, with time (t, h), at pH values of samples (pH 5.1, pH 5.8, pH 6.4, pH 6.1 and pH the major organic and inorganic constituents of rainwater in an urban area.
5.5, for sample 1, sample 2, sample 3, sample 4 and sample 5, respectively), and for the Drink. Water Eng. Sci. 8, 35e48.
following conditions: C0 Zn(II) of 100 mg L
1 (upper graphic), and C0 Zn(II) of 500 mg L
1 Council Directive 98/83/EC of 3 November 1998 on the quality of water intended for
(bottom graphic); and m/V of 52 g L
1 (5 beans). The average values of normalized human consumption. Off. J. 330, 0032-0054.
concentration of Zn(II) in ultrapure water for the pH 5.6, m/V of 52 g L
1, and for both Decree-Law n.º 236/98, Ministry of the Environment. In: Di
ario da República, I


SERIE-A, vol 176, 1 August 1998, pp. 3676e3722.
initial concentrations of Zn(II) are presented in the line.
Feng, J., Suna, M., Liua, H., Li, J., Liua, X., Jianga, S., 2010. Au nanoparticles as a novel
coating for solid-phase microextraction. J. Chromatogr., A 1217, 8079e8086.
Fomina, M., Gadd, G.M., 2014. Biosorption: current perspectives on concept, defi-
nition and application. Bioresour. Technol. 160, 3e14.
can be used, with an average Zn(II) removal rate of 63%. Using Gan, L., Zuo, J., Xie, B., Li, P., Huang, X., 2012. Zeolite (Na) modified by nano-Fe
particles adsorbing phosphate in rainwater runoff. J. Environ. Sci. 24,
5 and 10 beans (per 50 mL of water) up to 2 h of contact time, 1929e1933.
the average removal rates of 60% and 65% are obtained, Heidari, H., Razmi, H., Jouyban, A., 2012. Preparation and characterization of
respectively. The removal of Zn(II) from rainwater samples by ceramic/carbon coated Fe3O4 magnetic nanoparticle nanocomposite as a solid-
phase microextraction adsorbent. J. Chromatogr., A 1245, 1e7.
the white beans showed that the matrix of the sample did Hoekstra, A.Y., Mekonnen, M.M., Chapagain, A.K., Mathews, R.E., Richter, B.D., 2012.
not interfere with the removal of metal, which means that Global monthly water scarcity: blue water footprints versus blue water avail-
the process can be used in real samples. ability. PLoS One 7 (2) , e32688. https://doi.org/10.1371/journal.pone.0032688.
Kus, B., Kandasamy, J., Vigneswaran, S., Shon, H.K., Moody, G., 2013a. Gravity driven
membrane filtration system to improve the water quality in rainwater tanks.
Further research is required in order to ascertain and render the Water Sci. Technol. Water Supply 13 (2), 479e485.
process described in this work more sustainable. As such, it is Kus, B., Kandasamy, J., Vigneswaran, S., Shon, H.K., Moody, G., 2013b. Household
rainwater harvesting system e pilot scale gravity driven membrane-based
necessary to investigate whether it is possible to obtain the same
filtration system. Water Sci. Technol. Water Supply 13 (3), 790e797.
results using only the bean shells in the sorption process, avoiding Mameri, N., Boudries, N., Addour, L., Belhocine, D., Lounici, H., Grib, H., Pauss, A.,
waste of the bean cores, which may be used for food. On the other 1999. Batch zinc biosorption by a bacterial nonliving Streptomyces rimosus
biomass. Water Res. 33, 1347e1354.
hand, using the entire bean, the possibility of removing zinc from
Moradi, S.E., Shabani, A.M.H., Dadfarnia, S., Emami, S., 2016. Sulfonated metal
rainwater and its posterior use in composting should be explored. organic framework loaded on iron oxide nanoparticles as a new sorbent for the
Research should also move towards studying the suitability of magnetic solid phase extraction of cadmium from environmental water sam-
beans to remove other potentially toxic elements from rainwater. ples. Anal. Methods 8, 6337e6346.
Omrani, M., Ruban, V., Ruban, G., Lamprea, K., 2017. Assessment of atmospheric
Moreover, in the future, it will be important to disclose how the trace metal deposition in urban environments using direct and indirect mea-
organic structures of beans contribute to the adsorption of zinc. surement methodology and contributions from wet and dry depositions.
Atmos. Environ. 168, 101e111.
Park, D., Yun, Y.-S., Park, J.M., 2010. The past, present, and future trends of bio-
sorption. Biotechnol. Bioproc. Eng. 15, 86e102.
Acknowledgments Poniedziałek, M., Se˛ kara, A., Je˛ drszczyk, E., Ciura, J., 2010. Phytoremediation effi-
ciency of crop plants in removing cadmium, lead and zinc from soil. Folia
Hortic. Ann. 22/2, 25e31.
Thanks are due for the financial support provided by CNRS-INEE, Rocha, L.S., Lopes, C.B., Borges, J.A., Duarte, A.C., Pereira, E., 2013. Valuation of un-
LabEx DRIIHM, CESAM (UID/AMB/50017/2019), FCT/MCTES modified rice husk waste as an eco-friendly sorbent to remove mercury: a study
using environmental realistic concentrations. Water, Air, Soil Pollut. 224, 1599.
through national funds (IF/00407/2013/CP1162/CT0023), and also https://doi.org/10.1007/s11270-013-1599-9.
the co-funding by the FEDER, within the PT2020 Partnership Saiano, F., Ciofalo, M., Cacciola, S.O., Ramirez, S., 2005. Metal ion adsorption by
Agreement and Compete 2020. Ana C. Estrada and Joa ~o P. da Costa Phomopsis sp. biomaterial in laboratory experiments and real wastewater
~o treatments. Water Res. 39, 2273e2280.
are funded through national funds (OE), through FCT e Fundaça
Santos, P.S.M., Pereira, M.E., Duarte, A.C., 2008. Source assessment of metals in
para a Cie^ncia e a Tecnologia, I.P., in the scope of the framework rainwater events at a town in western Portugal. Fresen. Environ. Bull. 17 (12b),
contract foreseen in the numbers 4, 5 and 6 of the article 23, of the 2232e2239.
Decree-Law 57/2016, of August 29, changed by Law 57/2017, of July Santos, P.S.M., Santos, E.B.H., Duarte, A.C., 2014. Dissolved organic and inorganic
matter in bulk deposition of a coastal urban area: an integrated approach.
19. Patrícia S.M. Santos thanks support by the FCT Scientific J. Environ. Manag. 145, 71e78.
Employment Stimulus - Individual 2017 (CEECIND/01835/2017). Seinfeld, J.H., Pandis, S.N., 1998. Atmospheric Chemistry and Physics: from Air
 M.P.S. Ferreira et al. / Water Research 162 (2019) 170e179 179

Pollution to Climate Change, first ed. John Wiley & Sons Inc, New York (United ed., vol 2. Health criteria and other supporting information, Geneva.
States of America). Yang, J., Li, J.-y., Qiao, J.-q., Lian, H.-z., Chen, H.-y., 2014. Solid phase extraction of
Simmons, G., Hope, V., Lewis, G., Whitmore, J., Gao, W., 2001. Contamination of magnetic carbon doped Fe3O4 nanoparticles. J. Chromatogr., A 1325, 8e15.
potable roof-collected rainwater in Auckland, New Zealand. Water Res. 35, Yang, X., Yin, Y., Zong, Y., Wan, T., Liao, X., 2019. Magnetic nanocomposite as sorbent
1518e1524. for magnetic solid phase extraction coupled with high performance liquid
Stuart, B.H., 2004. Infrared Spectroscopy: Fundamentals and Applications, first ed. chromatography for determination of polycyclic aromatic hydrocarbons.
John Wiley & Sons, U.K. Microchem. J. 145, 26e34.
Suzuki, Y., Kametani, T., Maruyama, T., 2005. Removal of heavy metals from aqueous Yang, L., Zhang, J., Zhao, F., Zeng, B., 2016. Electrodeposition of self-assembled
solution by nonliving Ulva seaweed as biosorbent. Water Res. 39, 1803e1808. poly(3,4-ethylenedioxythiophene) @gold nanoparticles on stainless steel
Tom, M., Fletcher, T.D., McCarthy, D.T., 2014. Heavy metal contamination of vege- wires for the headspace solid-phase microextraction and gas chromatographic
tables irrigated by urban stormwater: a matter of time? PLoS One 9 (11), determination of several polycyclic aromatic hydrocarbons. J. Chromatogr. A
e112441. https://doi.org/10.1371/journal.pone.0112441. 1471, 80e86.
Tran, H.N., Youb, S.-J., Hosseini-Bandegharaei, A., Chao, H.-P., 2017. Mistakes and Zhang, S., Yao, W., Ying, J., Zhao, H., 2016. Polydopamine-reinforced magnetization
inconsistencies regarding adsorption of contaminants from aqueous solutions: of zeolitic imidazolate framework ZIF-7 for magnetic solid-phase extraction of
a critical review. Water Res. 120, 88e116. polycyclicaromatic hydrocarbons from the air-water environment.
Wang, H., Zhao, X., Meng, W., Wang, P., Wu, F., Tang, Z., Han, X., Giesy, J.P., 2015. J. Chromatogr., A 1452, 18e26.
Cetyltrimethylammonium bromide-coated Fe3O4 magnetic nanoparticles for Zhang, M., Huang, G., Huang, J., Chen, W., 2018. Three-dimensional multi-walled
analysis of 15 trace polycyclic aromatic hydrocarbons in aquatic environments carbon nanotubes@g-C3N4@ Fe3O4 nanocomposites-based magnetic solid
by ultraperformance, liquid chromatography with fluorescence detection. Anal. phase extraction for the determination of polycyclic aromatic hydrocarbons in
Chem. 87, 7667e7675. water samples. Microchem. J. 142, 385e393.
World Health Organization, 1996. Guidelines for Drinking-Water Quality, second