Professional Documents
Culture Documents
Adsorption of Cadmium by A High-Capacity Adsorbent Composed of Silicate-Titanate Nanotubes Embedded in Hydrogel Chitosan Beads (2020)
Adsorption of Cadmium by A High-Capacity Adsorbent Composed of Silicate-Titanate Nanotubes Embedded in Hydrogel Chitosan Beads (2020)
Adsorption of Cadmium by A High-Capacity Adsorbent Composed of Silicate-Titanate Nanotubes Embedded in Hydrogel Chitosan Beads (2020)
To cite this article: Roxana Quiroga-Flores, Asma Noshad, Reine Wallenberg & Linda Önnby
(2020) Adsorption of cadmium by a high-capacity adsorbent composed of silicate-titanate
nanotubes embedded in hydrogel chitosan beads, Environmental Technology, 41:23, 3043-3054,
DOI: 10.1080/09593330.2019.1596167
a
Division of Biotechnology, Department of Chemistry, Lund University, Lund, Sweden; bInstituto de Investigaciones Fármaco Bioquímicas,
Universidad Mayor de San Andrés, La Paz, Bolivia; cCentre for Analysis and Synthesis, Department of Chemistry, Lund University, Lund, Sweden;
d
National Centre for High Resolution Electron Microscopy, Lund University, Lund, Sweden
CONTACT Linda Önnby linda.onnby@chem.lu.se Centre for analysis and synthesis, Department of Chemistry, Lund University, P.O. Box 124, SE-223 62
Lund, Sweden
Supplemental data for this article can be accessed at https://doi.org/10.1080/09593330.2019.1596167
© 2019 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/
4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in
any way.
3044 R. QUIROGA-FLORES ET AL.
2.3 Preparation of hydrogel STNTs-Ch beads instrument (USA). The pore size distribution of STNTs-
Ch beads was calculated by the Barret–Joyner–Halenda
A solution of chitosan was prepared (1.7 g chitosan·100
(BJH) model. The samples were degassed at 30°C for
mL−1 of 0.1 M HCl) as reported previously [17].
24 h and the data were collected at a relative pressure
Different amounts of STNTs were added into the sol-
P/Po between 0.0025 and 0.99. The data were processed
ution of chitosan, resulting in STNTs-Ch beads with
by the 3Flex Software version 4.01.
different dosages (expressed as g STNTs·g−1 chitosan):
0.14, 0.28, 0.42 and 0.56. Once homogeneous, the
STNTs-chitosan solution was added drop wise into a
0.1 M NaOH solution. The formed hydrogel beads 2.5 Batch adsorption studies of STNTs and STNTs-
were rinsed with ultrapure water until the pH of the Ch beads
filtrate was neutral. The STNTs-Ch beads were not All batch experiments were carried out using a synthetic
dried but kept as hydrogel beads in order to avoid Cd2+ solution (10 mg · L−1 (89 µM), pH = 6.5 ± 0.2) unless
the changes in the pore structure of chitosan caused otherwise mentioned. The experiments were performed
by normal drying [25]. in duplicates at room temperature (T = 22°C) in polyethy-
lene tubes (volume = 10 mL) under constant agitation
using a rocking table from VWR, USA. Adsorption of
2.4 Characterization of STNTs and STNTs-Ch Cd2+ by the polyethylene tube was found to be negli-
beads gible (data is shown in Table S3).
For determining the morphology and size of STNTs, the
powder of STNTs was analysed by a JEM 2200FS trans-
2.5.1 Adsorption kinetics
mission electron microscope (TEM) working at 200 kV.
For studying adsorption kinetics and the diffusion behav-
The image data were analysed by the Digital Micrograph
iour, kinetic experiments were carried out first for (i)
suite (DM) software (Gatan Inc., USA). A TEM sample of
0.24 g STNTs · L−1, (ii) 15 chitosan beads (0 g STNTs ·
STNTs was also analysed by an energy-dispersive X-ray
L−1) and (iii) 15 STNTs-Ch beads·L−1 (0.28 g STNTs · g−1
spectrometer (TEM-XEDS). The quantification was made
chitosan). Samples were collected between 0 and
using the Oxford INCA software.
2880 min and analysed for Cd2+. The pseudo-first order
The distribution and the elemental composition of
[27], pseudo-second order model [28] and the intraparti-
STNTs in the STNTs-Ch beads were studied by a Jeol
cle diffusion model [29] were used to fit our kinetic data.
6700F-Scanning EM coupled to the Oxford Instruments
Thereafter, to determine the optimal dosage, the second
energy-dispersive X-ray spectrometer (SEM-XEDS). The
batch of kinetics experiments was carried out testing the
acceleration voltage was 5–15 kV. The sample was
following dosages: 0.14, 0.42 and 0.56 g STNTs·g−1
freeze-dried and coated with carbon to allow collection
chitosan.
of data and images from a cross-section of the STNTs-
Ch beads.
Stretch band vibrations corresponding to functional 2.5.2 Adsorption isotherms
groups in STNTs, chitosan beads and STNTs-Ch beads Adsorption isotherms were evaluated using Cd2+ sol-
were studied from freeze-dried samples by using an utions ranging from 0.1 to 400 mg · L−1. Samples were
ALPHA Fourier transformed infrared spectrometer collected at equilibrium. The Langmuir [30] and the
(FTIR), the data were treated by the Opus software, Freundlich [31] adsorption models were used to fit
Bruker, USA. the data of the adsorption isotherms. For calculating
The identification of crystal phases of STNTs, chitosan the capacities of adsorption at equilibrium, the mass
and STNTs-Ch beads was analysed by a Stadi MP Powder of the adsorbent was equal to the mass of the STNTs in
X-ray diffractometer (Stoe, Germany) with a copper tube the STNTs-Ch beads.
working at 40 mA and 40 kV. The scanning was per-
formed over a range of 2θ from 9° to 80° with a step
size of 0.013. The data were processed by the MATCH! 2.5.3 Metal ions competition
and JANA2006 software [26]. More details of the pro- Metal ions competition between Cd2+, Ni2+ and Pb2+ was
cedure are described in Text S1. evaluated by studying adsorption from solutions of
For determining textural properties, the Brunauer– 89 µM of each metal respectively. Three conditions
Emmett–Teller (BET) surface area of STNTs and STNT-Ch were established: (i) single metal ion solution, (ii) bisolute
beads were determined from N2 adsorption–desorption metal ion solution containing Cd2+ and Pb2+ and (iii) tri-
isotherms were measured using a 3Flex Micromeritics solute metal ion solution.
3046 R. QUIROGA-FLORES ET AL.
2.6 Adsorption in continuous mode by STNTs-Ch USA. The data were treated by the WinLab32 software
beads and reported RSD values were <10%. The limits of detec-
tion can be found in Table S4.
A glass column (9 × 170 mm) was packed with 350
Ammonium (NH+4 , LCK303), nitrate (NO− 3 , LCK339), sul-
STNTs-Ch beads with a bed volume of 6 mL. In continu-
phate (SO4 , LCK153) and phosphate (PO3−
2−
4 , LCK049)
ous mode, the column was used to treat (i) a synthetic
were analysed using analytical kits from Dr. Hach
Cd2+ solution (1400 mL or 233 bed volumes of 0.5 mg
Lange (Sweden). Presented data are average values
Cd2+·L−1) and (ii) real leachate (210 mL or 35 bed
from 3 measurements with an RSD <5%. The limits of
volumes of 0.034 mg Cd2+ · L−1) with an upward flow
detection were 1.0 mg · L−1 for NO− 3 , 2.5 mg · L
−1
for
rate of 0.32 mL · min−1. Both experiments were con-
NH4+, 5 mg · L−1 for PO3−
4 , and 40 mg · L
−1
for SO2−
4 .
ducted in duplicates at 22°C. Samples of 15 mL (interval
of 46.8 min) were collected using a fraction collector
FC204, Gilson, USA. 3. Results and discussion
3.1 Characterization of STNTs, STNTs-Ch beads
2.7 Recycling and reuse of STNTs-Ch beads and effects of pH on STNTs dispersion
The STNTs-Ch beads were recycled and reused in batch 3.1.1 STNTs and STNTs-Ch beads characterization
experiments. The recycling procedure consisted of deso- Figure 1(a)–(c) depicts TEM images of STNTs showing
rption and regeneration. The conditions for desorption hollow nanotubes and the elemental analysis obtained
were optimized by comparing two acids (HCl and by EDS. The diameter of the nanotubes was approxi-
H2SO4) as desorbing agents. The dosage for desorption mately 6–8 nm with 3–4 layers in the walls. The multi-
and regeneration was the same used in a previous walls had an interplanar distance of approximately
study for TNTs (i.e. 0.08 g TNTs·100 mL−1) [32]. The 0.7 nm. Figure 1(c) shows the co-existence of Si and Ti
optimal desorbing agent was selected according to Cd2 in the nanotubes determined by TEM-XEDS analysis of
+
desorption efficiency (%) and weight loss quantification a pair of nanotubes where the quantification showed a
(% mass loss). Once selected the desorbing agent, deso- Ti:Si molar ratio of 1.04:1.
rption was performed by immersing 15 STNTs-Ch beads Figures S1(a)–(c) show the infrared spectra of STNTs,
in 3 mL of the desorbing agent for 3 h. For regeneration, chitosan beads and STNTs-Ch beads, respectively. In
the 15 Cd2+-desorbed STNTs-Ch beads were immersed in Figure S1(a) the observed bands for STNTs at 411 cm−1
3 mL of 0.2 M NaOH for 3 h, thereafter the beads were
rinsed 3 times with 1 mL of ultrapure water. The recycled
STNTs-Ch beads were then reused under the same con-
ditions as virgin beads. Recycling was carried out 7 times
following the above-mentioned steps. For every cycle,
adsorption capacity at equilibrium (Qe) was measured.
With the newly Qe measured in every cycle, Qe recovery
(%) was calculated as the ratio between the Qe of each
cycle and Qe of virgin beads.
and 905 cm−1 were assigned to the TiO6 octahedron and corresponded to O–H and N–H stretching [36]. Finally,
to the four-coordinated Ti–O of (TiO(OH)2) respectively. Figure S1(c) shows that STNTs-Ch beads exhibited the
Both bands have been reported in studies for TNTs same bands observed in the other two spectra (Figure
before [7,8,33]. The absorption band at 1033 cm−1 was S1(a)–(b). The absence of new bands in STNTs-Ch
assigned to Si–O–Si. The same band has been previously beads suggests that STNTs and chitosan did not form
ascribed to Si–O3–Si in the case of polygorskite new chemical bonds.
embedded in chitosan beads [21,34]. The two remaining Figure 2(a) shows the XRD spectra of STNTs, chitosan
bands observed at 1634 cm−1 and 3345 cm−1 were beads and STNTs-Ch beads. The spectrum of STNTs
attributed to the O–H and H–O–H binding vibrations, exhibited diffraction peaks at 2θ value of 10°, 24° and
respectively. Overall, the infrared results confirmed the 48° corresponding to the planes (0 2 0), (1 1 0) and (2
presence of –OH, the TiO6 octahedron and the Si–O–Si 0 0) respectively indicating the presence of nanotubes.
bond in STNTs. Figure S1(b) shows bands observed in The obtained spectrum clarified the low intensity
chitosan beads. The bands in the range from 989 cm−1 peaks, proposing a low crystalline nature of the STNTs
to 1148 cm−1 were assigned to the C–O and C–N particles. The spectrum of chitosan beads displayed
bonds. The bands in the range from 1561 cm−1 to one weak diffraction peaks at 2θ value of 20° indexed
1649 cm−1 corresponded to the secondary amide and with (1 0 0) plane which is consistent with the reports
primary amide of the N-acetylated group [35,36]. The for chitosan [37]. The possible reason behind could be
bands in the range from 3285 cm−1 to 3362 cm−1, the presence of inter and intra-molecular H-bonding
which has been reported earlier [38]. This is consistent
with our discussed observations from the FTIR spectrum
(Figure S1(b)). In the STNTs-Ch beads spectrum, some
interesting differences have been observed such as (i) a
location shift at 2θ value of 20° towards a higher value
of intensity indicating high quantity of chitosan beads
and (ii) all the characteristic peaks of STNTs disappeared
in the STNTs-Ch beads spectrum which could be due to
the dominating existence of a high quantity of chitosan
and a weak crystalline nature of STNTs.
The BET surface area of STNTs and STNTs-Ch beads
was calculated to be 216.7 ± 0.7 and 12.9 ± 0.02 m2 · g−1,
respectively. The high surface area of STNTs and similar
materials, e.g. between 240 and 302 m2 · g−1 for TNTs
[6–9] and 411 m2 · g−1 for SNTs [10] could be attributed
to its small inner diameters and multilayer morphology
as observed by TEM in Figure 1(a). In contrast, the
surface of the STNTs-Ch beads was 16 times lower as com-
pared to the STNTs-Ch beads, which was probably due to
the relatively low load of STNTs in the STNTs-Ch beads,
which was 22% of the total mass. A higher load of particles
would increase the surface area of the composite as
observed when the load of TNTs was increased to γ-
Fe2O3/TNTs [9].
The pore size and pore volume STNTs-Ch beads were
calculated to be 11.6 nm and 0.0035 cm3 · g−1. Figure 2
(b) shows the nitrogen adsorption–desorption isotherms
and it is seen that the isotherm exhibited an IV-type iso-
therm indicative of capillary condensation occurring in a
mesoporous adsorbent [39,40]. The H3-type hysteresis
loop indicates that it does not exhibit limiting adsorption
while approaching P/Po = 1 and is often seen in aggre-
Figure 2. (a) XRD spectra of STNTs, chitosan beads and STNTs-Ch gates of materials with flat- or slit-shaped pores [39].
beads. (b) Nitrogen adsorption–desorption isotherms of STNTs- The pore size distribution showed in Figure 2(c) shows
Ch beads. (c) Pore size distribution of STNTs-Ch beads. a bimodal distribution. The small pores (2–3 nm) may
3048 R. QUIROGA-FLORES ET AL.
Figure 3. SEM images and SEM-XEDS mapping of STNTs-Ch beads, respectively. In (a) the cross-section of a STNTs-Ch bead shows both
the inner and outer part of the bead. The selected area is magnified in (b) and a group of white aggregates are shown in (c). The SEM-
XEDS mapping of a STNTs aggregate from the outer part is shown in (d), and from the cross-section is shown in (e). Both mapped areas
in (d,e) showed the presence of Na, Si, Ti and O in the aggregates.
correspond to the pores of STNTs; similar values are for –OH groups present in the STNTs have been reported
reported for TNTs [9,41] and the larger pores (10– to be in the range from 9.5 to 13 [43,44]. Thus, for the
12 nm) might correspond to sites with higher aggrega- prepared chitosan-STNTs suspension at pH = 1 (cf.
tion degree [8,9] of STNTs or probably correspond to Section 2.3), both the –NH2 and –OH groups present in
chitosan pores. chitosan and STNTs respectively would be protonated.
This, in turn, would yield highly dispersed STNTs in the
produced STNTs-Ch beads as similar charges result in
3.1.2 STNTs dispersion in STNTs-Ch beads and
repulsion and hence absent interaction between STNTs
effects from pH
and chitosan. Our results are supported with previous
Figure 3(a)–(f) shows the SEM imaging and the elemental
mapping of the STNTs-Ch beads obtained by SEM-XEDS.
Figure 3(a) shows the cross-section of an STNTs-Ch bead Table 1. Obtained model constants from the pseudo-second
and its hollow inner area. The magnification of the order model and the intraparticle diffusion model for STNTs
and STNTs-Ch beads, respectively.
selected area in Figure 3(a) is shown in Figure 3(b),
Kinetic model STNTs STNTs-Ch beads
which shows white aggregates in the outer and inner
Pseudo-second order
wall of the bead. Higher magnification of a group of Qe (mg · g−1) 31.4 ± 1.2 38.8 ± 1.6
aggregates is shown in Figure 3(c). Figure 3(d,e) shows k2 × 10−3 (g · mg−1 · min−1) 2.5 ± 0.6 0.8 ± 0.2
h (mg · g−1 · min−1) 2.5 ± 0.2 1.2 ± 0.1
the SEM-XEDS mapping of two aggregates; one of t1/2 (min) 12.7 ± 0.5 32.2 ± 1.3
them was from the cross-section and the other one R2 0.9224 0.9604
Intraparticle diffusion
from the outer wall. Both selected sites (Figure 3(d,e))
Kd1 (mg · g−1 · min−0.5) 3.4 ± 0.7 3.4 ± 0.5
showed the presence of Na, Si and Ti, confirming that C1 (mg · g−1) 3.9 ± 3.0 0.8 ± 2.1
the aggregates were indeed STNTs. R21 0.9550 0.9772
Kd2 (mg · g−1 · min−0.5) 0.9 ± 0.0 1.8 ± 0.5
The degree of protonation could affect the distri- C2 (mg · g−1) 16.3 ± 0.2 9.5 ± 5.0
bution of the particles inside the chitosan bead. The R22 0.9996 0.9237
Kd3 (mg · g−1 · min−0.5) 0.2 ± 0.0 0.3 ± 0.1
pKa of chitosan ranges from 6.17 to 6.54 for a degree C3 (mg · g−1) 28.0 ± 0.9 28.7 ± 2.3
of deacetylation from 94% to 73% [42]. In our study, R23 0.9560 0.9280
the degree of deacetylation was 70%. The pKa values Note: Numbers were rounded off to one decimal.
ENVIRONMENTAL TECHNOLOGY 3049
observations done for TNTs where a higher degree of layer, were higher for the first and second region of
reorganization of aggregates followed by an increased STNTs than that of STNTs-Ch beads. This can be corre-
surface area was observed in an acidic environment [45]. lated with the higher diffusion resistance observed for
STNTs. Overall, STNTs-Ch beads showed improved
diffusion than the STNTs alone.
3.2 Adsorption studies
3.2.1 Kinetic modelling of Cd2+ adsorption
3.2.2 Optimization of the dosage of STNTs in STNTs-
The kinetic data were best described by the pseudo-
Ch beads
second order and intraparticle diffusion models. Both
For a batch adsorption process, the overall removal can
models are described in Text S2. Table 1 shows the con-
be a guiding tool for the selection of optimal dosage,
stants obtained from modelling. The highest Qe was
whereas the kinetic parameters are those directing for
obtained for STNTs-Ch beads (38.8 mg · g−1), followed
a continuous process. We studied three additional
by STNTs (31.4 mg · g−1). The kinetic data of chitosan
dosages of STNTs in our STNTs-Ch beads (g STNTs ·
did not fit to any of the studied models (Table S5). For
g−1 chitosan): 0.14, 0.42 and 0.56, that were compared
that reason, we only calculated the experimental Qe
with the dosage of 0.28. The kinetic data were studied
using Equation S4, and was found to be low (Qe =
by using the pseudo-second order model. The model
5.21 mg · g−1). The higher Qe found for STNTs-Ch beads
constants are listed in Table S6. Figures 4(a)–(b)
is hypothesized to be due to (i) the chelation of Cd2+
depict the changes of k2, h, Qe, Cd2+ removal (%) and
by the duplet of electrons on nitrogen atoms present
t1/2 with increasing STNTs dosage. Figure 4(a) shows
in chitosan, and (ii) the dispersion of STNTs in the compo-
the low variation of h, which indicates that the adsorp-
site as confirmed by SEM-XEDS (cf. Figure 3(d)–(e), and
tion at the initial stage was independent of the dosage.
Section 3.1.2). Overall, the STNTs were shown to be the
In contrast, k2 showed an increasing trend due to the
main component involved in the adsorption of Cd2+ by
availability of more adsorption sites for Cd2+ as the
STNTs-Ch beads.
In Table 1, the initial adsorption rate (h) indicates that
adsorption of Cd2+ by STNTs was faster than STNTs-Ch
beads (2.5 versus 1.2 mg g−1 · min−1). In contrast, the
chemical rate (k2) for STNTs was slower as compared
with STNTs-Ch beads (2.5 × 10−3 versus 0.8 × 10−3 g
mg−1 min−1), which explains the higher resulting Qe for
STNTs-Ch beads. The advantage of a faster k2 and
higher Qe of STNTs-Ch beads indicates an improvement
in kinetics adsorption for STNTs-Ch beads.
Table 1 shows the Kd rates and C values from the
linear regressions of the intraparticle diffusion model.
The linear regressions (c.f. Figure S2) of the three com-
pared adsorbents: STNTs, chitosan beads and STNTs-Ch
beads, showed three regions that did not pass through
the origin, which means that diffusion is not the only
rate-liming step. The Kd rates and C values for each adsor-
bent were ranked as follows: Kd1 > Kd2 > Kd3 and C3 > C2 >
C1. As described by Weber [46], the first region is associ-
ated with the transport of Cd2+ ions towards the surface
of the adsorbent. Both STNTs and STNTs-Ch beads had
equal Kd1 indicating an identical pace. The second Figure 4. Relation between dosage of STNTs (g STNTs·g−1 chit-
region is associated with the diffusion of Cd2+ ions osan) present in STNTs-Ch beads and kinetic constants k2, h, t1/
through the pores of the adsorbent. The Kd2 of STNTs- 2 and Qe obtained from pseudo-second order kinetic modelling
Ch beads was twice the value of Kd2 of STNTs, indicating and total Cd2+ removal (%). In (a) k2 (g · mg−1 · min−1) and h
on a higher diffusion resistance for STNTs. Finally, the (mg · g−1 · min−1). In (b) Qe (mg · g−1), total Cd2+ removal (%)
and t1/2 (min) as a function of the STNTs dosage in STNTs-Ch
third region that is associated with the final equilibrium
beads. The dashed lines show the trends for each constant
stage showed the lowest and similar Kd3 values for with increased STNTs dosage. The experimental conditions
both adsorbents. One the other hand, C values, which were: pH = 6.5 ± 0.2, T = 22°C, Ci = 10 mg Cd2+ · L−1. Kinetic
are associated with the thickness of the boundary modelling was done with Graphpad prism®.
3050 R. QUIROGA-FLORES ET AL.
STNTs dosage increased. The changes in k2 influenced to the single solute solutions. Nevertheless, the overall
the resulting Qe that gradually decreased with increas- removal did not change and was ranked as: Pb2+ > Cd2
+
ing dosage of STNTs. In another study, similar >> Ni2+ for all cases. Similar removal ranking was
changes in Qe were observed for an increased dosage observed for TNTs when Cd2+, Pb2+, Ni2+ and Cr3+ were
of TNTs used for adsorption of Pb2+ [8]. Finally, the together in solution [5,8,33]. As there were differences
highest overall removal (82–87%) was obtained by the in the kinetics constants, we can hypothesize that the
dosages from 0.28 to 0.56. We chose 0.28 as the differences in the total removal would be larger if a
optimal dosage due to the small-observed variations higher concentration of metals close to a limiting
in this interval. threshold had been evaluated.
3.3 Continuous removal of metal ions present in showed the beads were hollow, which means that
synthetic and real leachate water could have occupied the core of the beads. The
longitudinal diffusion rates of water molecules inside
We treated a synthetic Cd2+ solution (500 µg Cd2+·L−1)
the bead could have affected the Cd2+ adsorption
and real wastewater leachate by STNTs-Ch beads
during phase I. Phase II showed that Cd2+ adsorption
packed in columns. Figure S5 shows the fractions of
maintained a threshold of 0.9 for the removed fractions
Cd2+ removed (expressed as 1 − (C·C−1o )) from the syn-
until 107 bed volumes.
thetic solution as a function of bed volumes. Two
Figure 5(a)–(c) shows the fractions (1 − (C·C−1 o )) of
phases were observed during the treatment, phase I
removed metals from the real leachate as a function of
comprised 0 < bed volumes <15 and phase II comprised
bed volumes. We divided the column run in phase I
15 < bed volumes <275. During phase I, an outlier value
and II, based on the observations done for Cd2+ adsorp-
(1 − (C·C−1
o ) = 0.76) was observed at 7.5 bed volumes,
tion in Figure S5. Figure 5(a,b) shows metals that were
indicating that poor adsorption occurred. Figure 3(a)
removed during phase I. Figure 5(a) shows that alkali
and alkali-earth metals such as Mg2+, Ca2+and K+ are
removed up to 1 − (C·C−1 o ) < 0.5. Figure 5(b) shows that
transition metal ions such as Zn2+, Co2+, Ni2+ and Cu2+
are removed up to 1 − (C·C−1 o ) > 0.5. During phase II,
Figure 5(c) shows the removal of Cd2+ (transition
metal) and Pb2+ (post-transition metal) and their respect-
ive removed fractions that were found in the range of 1
− (C·C−1
o ) > 0.5.
In order to study the ranking of the fractions removed
in both phases, we plotted the data of the fractions
removed versus the ionic radii, hydrated radii, hydration
energy and electronegativity, respectively (the theoreti-
cal values are listed in Table S9). From our comparisons,
we found that R 2 values ≥ 0.7 occurs only for post-tran-
sition and transition metal ions plotted versus ionic
radii and hydration energy. The graphs can be found in
Figures S6a-d. Similar accordance between hydration
energy and adsorption capacities for Pb2+, Cd2+, Cu2+
and Cr3+ has been found for TNTs in batch experiments
[5,6]. The fractions removed of other ionic compounds
−
such as SO2− 3− +
4 , PO4 , NH4 and NO3 are shown in Figure S7.
Figure 6. Scheme of STNTs-Ch beads production from synthesis of STNTs to waste disposal without (upper area) or with (lower area)
recycling. The non-recycling balance shows single use to solid waste shown as steps 1 to 3 for 4 times of STNTs-Ch beads use. The
recycling process includes steps 1 to 7 where recycling of STNTs-Ch beads was performed three times. The pre-treatment was only
conducted for the 1st cycle.
3.4 Recycle and reuse of STNTs-Ch beads for the production of virgin beads. The additional
inputs for recycling were H2SO4 (step 5) and NaOH
The recycle of STNTs-Ch beads was composed of deso-
(steps 4 and 6). As a result, 1.9 times more of NaOH is
rption of Cd2+ and regeneration of the STNTs. The data
needed for a recycling process of 4 cycles as compared
for the optimization of desorption conditions can be
to 4 times of single use of STNTs-Ch beads.
found in Table S10. The results showed that a pre-treat-
To further reduce the amount of NaOH needed during
ment of STNTs-Ch beads with 0.1 M NaOH prior deso-
recycling, the optimization of the NaOH dosage needed
rption with H2SO4 increased the tolerance of the beads
in Steps 4 and 6 is more advantageous than the reuse of
from 0.01 M to 0.5 M H2SO4. The highest desorption
the supernatant in Step 1; the latter would only save
efficiencies were 71% and 75% for 0.1 and 0.5 M
3.3% of the total amount of NaOH. Note that the optim-
H2SO4, respectively. Hence, 0.1 M H2SO4 was selected
ization will be restricted by the alkalinity needed for
as a desorbing agent.
forming round-shaped chitosan beads and the amount
Table 3 shows the total removal (%) and desorption of
2+ of Na+ and OH− that needs to be replenished in regener-
Cd (%), the equilibrium capacity for each adsorption
ation [50,51].
cycle (Qe′ ) and the recovered capacity (Qe,recov, %). The
Recycling has advantages over a non-recycling
quantified mass loss was measured on dried beads and
process in terms of both consumed chemicals and the
found to be negligible. Satisfactory values of Qe,recov
lifespan of the adsorbent. To give an example, 89
(>86%) were obtained for up to four cycles with more
columns would be able to perform 445 treatments if
than 85% of Cd2+ removed. However, it was not possible
the adsorbent is recycled 4 times.
to desorb all Cd2+ in each desorption step; only 69% was
desorbed during the 1st cycle.
4. Conclusion
3.4.1 Comparison and balance of the consumed
chemicals for recycled STNTs For the sake of obtaining more secure water, the technol-
Figure 6 shows the different steps from STNTs-Ch bead ogies for treating complex waters need to be highly selec-
production followed by (i) single use or (ii) four times tive, but preferably also affordable and feasible. Although
reuse. Steps 1–3 comprise the synthesis of STNTs till some adsorbents have shown to be highly selective; there
the use of STNTs-Ch beads for adsorbing Cd2+, which is still a notable gap between lab research and pilot appli-
thereafter go to waste disposal. Steps 1–7 include a cations. In this study, we have chosen to work with a com-
four-times repeated use of recycled STNTs-Ch beads. posite made of a silicate-titanate nanotube that was
The detailed calculations for the respective scenarios embedded in a biodegradable hydrogel polymer (chito-
are shown in Table S11. Thus, for four times recycled san). The adsorbent was used for adsorbing Cd2+ from
STNTs-Ch beads, the amount saved for TiO2, SiO2, chito- synthetic water and real leachate. By dispersing the
san and HCl was three times the initial amount needed STNTs in chitosan beads, the BET surface area decreased
ENVIRONMENTAL TECHNOLOGY 3053
pH, ionic strength and natural organic matter. Colloids Spectrochimica Acta Part A: Molecular and Biomolecular
Surf A: Physicochem Eng Aspects. 2013;434(0):319–328. Spectroscopy. 2012;97(Suppl. C):1052–1057. doi:10.1016/
doi:10.1016/j.colsurfa.2013.05.010. j.saa.2012.07.085.
[17] Miller SM, Zimmerman JB. Novel, bio-based, photoactive [35] Nair V, Panigrahy A, Vinu R. Development of novel chito-
arsenic sorbent: TiO2-impregnated chitosan bead. Water san–lignin composites for adsorption of dyes and metal
Res. 2010;44(19):5722–5729. doi:10.1016/j.watres.2010. ions from wastewater. Chem Eng J. 2014;254:491–502.
05.045. doi:10.1016/j.cej.2014.05.045.
[18] Shukla SK, Mishra AK, Arotiba OA, et al. Chitosan-based nano- [36] Rusmin R, Sarkar B, Liu Y, et al. Structural evolution of chit-
materials: a state-of-the-art review. Int J Biol Macromol. osan–palygorskite composites and removal of aqueous
2013;59:46–58. doi:10.1016/j.ijbiomac.2013.04.043. lead by composite beads. Appl Surf Sci. 2015;353(Suppl.
[19] Lee DW, Yoo BR. Advanced silica/polymer composites: C):363–375. doi:10.1016/j.apsusc.2015.06.124.
materials and applications. J Ind Eng Chem. 2016;38:1– [37] Gokila S, Gomathi T, Sudha P, et al. Removal of the heavy
12. doi:10.1016/j.jiec.2016.04.016. metal ion chromiuim (VI) using chitosan and alginate nano-
[20] He J, Bardelli F, Gehin A, et al. Novel chitosan goethite bio- composites. Int J Biol Macromol. 2017;104:1459–1468.
nanocomposite beads for arsenic remediation. Water Res. [38] Zhou JQ, Wang JW. Immobilization of alliinase with a
2016;101:1–9. doi:10.1016/j.watres.2016.05.032. water soluble–insoluble reversible N-succinyl-chitosan
[21] Rusmin R, Sarkar B, Liu Y, et al. Structural evolution of chit- for allicin production. Enzyme Microb Technol. 2009;45
osan–palygorskite composites and removal of aqueous (4):299–304.
lead by composite beads. Appl Surf Sci. 2015;353:363– [39] Sing KSW. Reporting physisorption data for gas/solid
375. doi:10.1016/j.apsusc.2015.06.124. systems with special reference to the determination of
[22] Önnby L, Harald K, Nges IA. Cryogel-supported titanate surface area and porosity (Recommendations 1984).
nanotubes for waste treatment: Impact on methane pro- Pure and applied chemistry 1985. p. 603.
duction and bio-fertilizer quality. J Biotechnol. 2015;207 [40] Brunauer S, Deming LS, Deming WE, et al. On a theory of
(0):58–66. doi:10.1016/j.jbiotec.2015.05.014. the van der Waals adsorption of gases. J Am Chem Soc.
[23] Du AJ, Sun DD, Leckie JO. Selective sorption of divalent 1940;62(7):1723–1732.
cations using a high capacity sorbent. J Hazard Mater. [41] Zhang L, Du AJ, Sun DD, et al. Effects of inorganic anions
2011;187(1):96–100. doi:10.1016/j.jhazmat.2010.12.120. on cadmium sorption behaviours on titanate nanotube
[24] Kasuga T, Hiramatsu M, Hoson A, et al. Formation of tita- surfaces. Environ Technol. 2013. doi:10.1080/09593330.
nium oxide nanotube. Langmuir. 1998;14(12):3160–3163. 2013.798035.
doi:10.1021/la9713816. [42] Wang QZ, Chen XG, Liu N, et al. Protonation constants of
[25] Toshihiko T, Mitsuaki M, Nobufumi A. Mechanism of resist chitosan with different molecular weight and degree of
pattern collapse during development process. Jpn J Appl deacetylation. Carbohydr Polym. 2006;65(2):194–201.
Phys. 1993;32(12S):6059. [43] Volesky B. Biosorption and me. Water Res. 2007;41
[26] Petříček V, Dušek M, Palatinus L. Crystallographic comput- (18):4017–4029. doi:10.1016/j.watres.2007.05.062.
ing system JANA2006: general features. Zeitschrift für [44] Ripin D. Evans pKa table. 2005.
Kristallographie – Cryst Mater. 2014;229(5):345–352. [45] Grigorieva AV, Yuschenko VV, Ivanova II, et al. Chemical
[27] Yuh-Shan H. Citation review of Lagergren kinetic rate tuning of adsorption properties of titanate nanotubes. .
equation on adsorption reactions. Scientometrics. 2012;2012:134.
2004;59(1):171–177. [46] Weber WJ, Smith EH. Simulation and design models for
[28] Ho YS, McKay G. A comparison of chemisorption kinetic adsorption processes. Environ Sci Technol. 1987;21
models applied to pollutant removal on various sorbents. (11):1040–1050.
Process Saf Environ Prot. 1998;76(4):332–340. doi:10.1205/ [47] Pirveysian M, Ghiaci M. Synthesis and characterization of
095758298529696. sulfur functionalized graphene oxide nanosheets as
[29] Weber WJ, Morris JC. Kinetics of adsorption on carbon efficient sorbent for removal of Pb2+, Cd2+, Ni2+ and
from solution. J Sanit Eng Div. 1963;89(2):31–60. Zn2+ ions from aqueous solution: a combined thermo-
[30] Langmuir I. The adsorption of gases on plane surfaces of dynamic and kinetic studies. Appl Surf Sci. 2018;428:98–
glass, mica and platinum. J Am Chem Soc. 1918;40 109.
(9):1361–1403. doi:10.1021/ja02242a004. [48] Ordoñez-Regil E, Granados-Correa F, Ordoñez-Regil E,
[31] Freundlich H. Über die adsorption in lösungen. Zeitschrift et al. Nanoparticles of KFeP2O7 implanted on silica gel
für Physikalische Chemie. 1907;57(1):385–470. beads for Cd2+ ion adsorption. Environ Technol.
[32] Wang T, Liu W, Xu N, et al. Adsorption and desorption of 2015;36(2):188–197.
Cd(II) onto titanate nanotubes and efficient regeneration [49] Lin S, Yang Y, Chen G, et al. Study on the influence of thio-
of tubular structures. J Hazard Mater. 2013;250–251 lation on the adsorption and magnetic recovery of super-
(Suppl. C):379–386. doi:10.1016/j.jhazmat.2013.02.016. paramagnetic nanoadsorbents for Cd2+ removal. Appl
[33] Wang T, Liu W, Xiong L, et al. Influence of pH, ionic Surf Sci. 2017;425:141–147.
strength and humic acid on competitive adsorption of [50] Zhang S, Chen Q, Peng LM. Structure and formation of
Pb(II), Cd(II) and Cr(III) onto titanate nanotubes. Chem H2Ti3O7 nanotubes in an alkali environment. Phys Rev B.
Eng J. 2013;215–216(Suppl. C):366–374. doi:10.1016/j.cej. 2005;71(1):014104. doi:10.1103/PhysRevB.71.014104.
2012.11.029. [51] Du AJ, Sun DD, Leckie JO. Sequestration of cadmium ions
[34] Yan W, Liu D, Tan D, et al. FTIR spectroscopy study of the using titanate nanotube. J Hazard Mater. 2011;187(1–
structure changes of palygorskite under heating. 3):401–406. doi:10.1016/j.jhazmat.2011.01.053.