Synthesis of glycopolymer nanosponges with enhanced adsorption performances for boron removal and water treatment

Article abstract of DOI:10.1039/C8TA06802J

The high-affinity interactions between cis-diols and boric/boronic acid have been widely employed as a tool for carbohydrate analysis, protein separation and boron removal. Herein we report the design and synthesis of cyclodextrin-scaffolded glycopolymers as bifunctional nanosponges for boron removal and water treatment for the first time. Different glycopolymer nanosponges (GNs) have been successfully synthesized from monosaccharides and β-cyclodextrin via a combination of a cross-linking reaction, Fischer glycosylation and a click reaction. Such functional GNs are mesoporous polymer frameworks with cis-diol-containing saccharides immobilized on the surface, which have exhibited selective adsorption behaviour towards boric acid depending on the structure of the GNs and the loaded saccharides. The GNs have also shown remarkable adsorption rates and capacities for an organic dye as a model pollutant in this work. Secondary bonding, such as hydrogen bonding and van der Waals forces between the immobilized saccharides and the adsorbates is believed to be responsible for the significantly enhanced adsorption rates and capacities. Such bifunctional materials may exhibit potential applications in seawater desalination and water treatment.

Full text of DOI:10.1039/C8TA06802J

Journal of
Materials Chemistry A
View Article Online
View Journal | View Issue
PAPER
Synthesis of glycopolymer nanosponges with
enhanced adsorption performances for boron
removal and water treatment†
Cite this: J. Mater. Chem. A, 2018, 6,
21193
ab
Xueping Liao,ab Bingyu Wangab and Qiang Zhang
*
The high-affinity interactions between cis-diols and boric/boronic acid have been widely employed as
a tool for carbohydrate analysis, protein separation and boron removal. Herein we report the design and
synthesis of cyclodextrin-scaffolded glycopolymers as bifunctional nanosponges for boron removal and
water treatment for the first time. Different glycopolymer nanosponges (GNs) have been successfully
synthesized from monosaccharides and b-cyclodextrin via a combination of a cross-linking reaction,
Fischer glycosylation and a click reaction. Such functional GNs are mesoporous polymer frameworks
with cis-diol-containing saccharides immobilized on the surface, which have exhibited selective
adsorption behaviour towards boric acid depending on the structure of the GNs and the loaded
saccharides. The GNs have also shown remarkable adsorption rates and capacities for an organic dye as
a model pollutant in this work. Secondary bonding, such as hydrogen bonding and van der Waals forces
between the immobilized saccharides and the adsorbates is believed to be responsible for the
significantly enhanced adsorption rates and capacities. Such bifunctional materials may exhibit potential
applications in seawater desalination and water treatment.
Received 14th July 2018
Accepted 8th October 2018
DOI: 10.1039/c8ta06802j
rsc.li/materials-a
hybrid nanomaterials for selective removal of boron.8–11 Natural
carbohydrate polymers (cellulose bres etc.) and articial poly-
Introduction
meric nanomaterials have been used to achieve great success in
water treatment and engineering, such as the efficient adsorp-
tion of heavy metal ions and organic dyes.12–17 Synthetic glyco-
polymers could reversibly complex with boronic acids by
forming 5- or 6-membered ring cyclic borate esters, which is
useful for the construction of glucose biosensors and smart
materials.18–20 Compared with the chelating diol groups of
pyranoses, borates have shown much higher affinity towards
furanoses and form very stable boron esters with D-ribose.21,22
Most reports in the literature have focused on the study of
interactions between pyranose-containing glycopolymers and
phenylboronic acids (PBA) rather than boric acids, however,
boron in nature mainly exist in the form of boric acid or boro-
nate salts rather than PBA.23 It is worth noting that even a slight
excess of boric acid was reported to have toxic effects on the
growth of some boron-sensitive agricultural crops.24,25 Thus it is
imperative to develop an effective treatment technology for the
removal of boric acid in order to match the specic restrictions
of drinking and irrigation. To the best of our knowledge, few
relevant studies on the synthesis of furanose-containing glyco-
polymers for bonding with boric acids have been performed
thus far.
Water scarcity has become one of the most serious challenges
facing the world today and there is great potential to address the
worldwide water-shortage problem by converting nearly inex-
haustible seawater into fresh water using desalination tech-
nologies.1–3 Although most salts could be removed by reverse
osmosis (RO) desalination technology, it is still a challenge to
remove boron. This is mainly due to the existence of boron as
neutrally charged boric acid in seawater and its ability to effi-
ciently diffuse through RO membranes like water molecules,
which will result in unsatisfactory levels of residual boron in the
permeate water.4,5 Since the discovery by Hermans, Boeseken
¨
et al. that cis-diol-containing molecules could coordinate with
boric acid to form stable borate esters, polyols have been
extensively applied for the removal of boron from water.6,7
Functional polyols such as N-methylglucamine, imino-bis-
propane diol and phenol hydroxyl groups have been immobi-
lized on the surface of ion exchange resins, membranes or
aJiangsu Key Laboratory of Chemical Pollution Control and Resources Reuse, School of
Environmental and Biological Engineering, Nanjing University of Science and
Technology, Nanjing 210094, P. R. China
bInstitute of Polymer Ecomaterials, School of Environmental and Biological
Engineering, Nanjing University of Science and Technology, Nanjing 210094, P. R.
China. E-mail: zhangqiang@njust.edu.cn
Activated carbons are the most widespread adsorbents for
water purication; however, some deciencies such as the slow
uptake rate and poor removal of hydrophilic pollutants make
them have poor performances in removing emerging organic
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c8ta06802j
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A, 2018, 6, 21193–21206 | 21193

View Article Online
Journal of Materials Chemistry A
Paper
micropollutants from water.26 Cyclodextrins (CDs) are well-
known supermolecules with hydrophobic interior cavities,
which could encapsulate thousands of different organic
pollutants via host–guest interactions and thus become poten-
tial candidates for adsorption and separation. Chemical cross-
linking of CDs with bifunctional crosslinking agents in alkaline
media resulted in the formation of polymeric networks, which
have already shown extraordinary abilities in the selective
removal of heavy metal ions or cationic dyes from aqueous
solutions.27–29 Studies using tetrauoroterephthalonitrile or
decauorobiphenyl etc. as novel crosslinking agents have raised
great attention recently, which is mainly attributed to the abil-
ities of the generated highly porous CD polymers with ultrafast
adsorption rates of organic micropollutants from water.26,30–32
However, most hydroxyl groups in the CD polymers are either
substituted, coordinated or not in a cis conguration, thus no
boron adsorption ability could be attained.
Position- and face-selective modications of CDs with func-
tional saccharides, which were mainly through click-type liga-
tion chemistry, have generated a wide family of CD-scaffolded
glycoclusters over the past few decades.33–36 These glycoclusters
have been well developed, especially in the research of opti-
mizing carbohydrate–protein interactions, drug delivery and
anti-adhesion therapy.37–40 Recently mannose or glucose func-
tionalized CD glycoclusters have been used for the construction
of molecular recognition surfaces via host–guest interac-
tions.41–45 Specic capture and release of proteins and bacteria
could be realized due to the presence of a multivalency effect
between the glycoclusters and the biomacromolecules.42,43,45 The
dynamic covalent bonding between PBA and secondary
hydroxyls of CD-functionalized supermolecules was proven to be
sugar-responsive, however, this system has not been applied to
the adsorption of boric acid due to the specic preference for cis-
diols.46 Due to the combination of hydrophobic interior cavities
and functional saccharides, it could be reasonable to hypothe-
size that such CD-scaffolded glycoclusters could act as potential
candidates for the simultaneous encapsulation of organic
pollutants and adsorption of boron. However, these glyco-
clusters are generally highly water-soluble due to the presence of
saccharides and it is hard to remove them from aqueous solu-
tion aer water treatment. We believe that cross-linked CD
polymers with persubstituted multivalent saccharides could
theoretically combine hydrophobic cavities and glycoligand
arrays together, while the insoluble polymer networks could
endow the materials with good recyclability.
Experimental
Materials
b-Cyclodextrin (b-CD, 96%, Aladdin, China) was recrystallized
twice from water in order to improve the purity and dried in
a vacuum oven at 100 ꢀC for two days before use. Heptkis(6-
deoxy-6-azido)-b-cyclodextrin (b-CD-(N3)7) and H2SO4-silica
catalyst were synthesized according to previous reports.40,47–49 1-
(20-Propargyl)D-glucose, 1-(20-propargyl)D-mannose, 1-(20-prop-
argyl)N-acetyl-D-glucosamine and 1-(20-propargyl)D-ribose were
prepared according to procedures described in previous litera-
ture.48,49 D-(+)-Gluconic acid d-lactone (99%), propargylamine
(98%), terephthaloyl chloride (TCL, 99%), copper(II) sulfate
pentahydrate (CuSO4$5H2O, 99%), (+)-sodium L-ascorbate
(NaAsc, 98%), boric acid (99.5%) and methylene blue (MB, dye
content $90% (HPLC)) were obtained from Aladdin (China) and
used directly. All the other reagents and anhydrous solvents
were obtained at analytical grade (>99%) from Aladdin (China)
and used without further purication unless otherwise stated.
Characterization techniques
1H and 13C NMR spectra were recorded on a Bruker Avance 500
MHz spectrometer using deuterated solvents obtained from
Aladdin. The NMR spectrometer had a BBFO-Plus forward
broadband liquid probe. The probe temperature was 25 ꢀC during
the test, the number of 1H scans was 16, and the number of 13
C
scans was 300. Fourier Transform Infrared Spectroscopy (FTIR)
measurements were carried out using a Nicolet iS5 FTIR spec-
trometer using an iD7 diamond attenuated total reectance
optical base and thirty-two scans were taken for each spectrum at
a normal resolution of 2 cmꢁ1. Mass spectrometric data was
measured using a Bruker Daltonics APE III ESI spectrometer.
Surface area and pore structure parameters of the adsorbents
were measured from N2 adsorption/desorption isotherms at
ꢁ196 ꢀC (77 K) using a TriStar II 3020 V1.03 system (Micromeritics
Instrument Co., Norcross, GA, USA). Prior to N2 adsorption, each
sample (25–50 mg) was degassed at 90 ꢀC under vacuum for 24 h
and then backlled with N2. A UV/Vis spectrophotometer (SHI-
MADZU UV-2600) was utilized to measure the concentration of
sample during adsorption at dened wavelengths. Elemental
analysis (EA) was performed using an Elementar Vario MICRO.
Thermal gravimetric analysis (TGA, Mettler Toledo, Switzerland)
was performed at a heating rate of 10 ꢀC minꢁ1 from room
temperature to 800 ꢀC under N2 protection. Differential scanning
calorimetry (DSC 823e, Mettler Toledo, Switzerland) was per-
formed at a heating rate of 10 ꢀC minꢁ1 from room temperature to
This study herein develops the synthesis of CD-scaffolded
glycopolymers and investigates the possibility of using such
glycopolymers as bifunctional nanosponges to remove boron
and organic pollutants from aqueous solution. To the best of
our knowledge, this is the rst report of CD glycopolymers with
the ability to adsorb boric acids. A facile synthetic route via the
combination of a cross-linking reaction, Fischer glycosylation
and a copper-mediated azide-alkyne cycloaddition reaction
(CuAAC) was developed for the construction of varied GNs with
controlled CD compositions and various functional saccha-
ꢀ
550 C under N2 protection (note: azide-containing compounds
should be handled with care and the amount for TGA/DSC anal-
ysis was limited below 5 mg). Scanning electron microscope
(SEM) images were taken using a eld emission SEM (FEI Quant
250FEG, USA) aer the synthesised cross-linked CD polymer
samples were coated with a 20 nm gold layer.
Synthesis of cross-linked CD polymers
rides. The properties of such GNs for the adsorption of boron For the synthesis of the CD copolymer, b-CD (2.00 g, 1.76
and organic pollutants were also evaluated respectively.
mmol), (b-CD-(N3)7) (2.00 g, 1.53 mmol) and 50 mL anhydrous
21194 | J. Mater. Chem. A, 2018, 6, 21193–21206
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
pyridine were added to a ame dried three-neck round bottom- repeated at least six times in order to remove any unreacted
ask and stirred for 1 h to afford a homogeneous solution monosaccharides or residual catalyst. The cross-linked CD-
ꢀ
before being cooled to 0 C via an ice/water bath. TCL (9.35 g, scaffolded glycocluster, poly(b-CD)-co-(b-CD-(N3)7)@ribose,
46.06 mmol) dissolved in 20 mL anhydrous THF was then added was obtained as a white powder aer lyophilization.
dropwise to the pyridine solution. Subsequently, the reaction
mixture was stirred at 70 ꢀC for 4 h. Aer cooling down, 100 mL
water was slowly added into the suspension and the resulting
Adsorption and regeneration experiments
Briey, each 50 ml vial with a polytetrauoroethylene-lined
screw cap received 100 mg or 10 mg of the GN samples and
DI water as a background solution. The sorbent samples were
pre-wetted for 12 h and a pre-determined volume of a sorbate
stock solution (prepared in DI water) was added to achieve an
initial concentration of 300 mg Lꢁ1 for boron and 400 mg Lꢁ1
for methylene blue (MB). The samples were covered with
aluminium foil and shaken by an orbital shaker incubated at
25 ꢀC and 100 rpm for 24 h. This duration was sufficient to reach
an apparent sorption equilibrium (no further uptake of sorbate)
according to the adsorption kinetics data as discussed later.
Aerwards, 1 mL of boron or 0.5 mL of MB was withdrawn from
the vials, ltered through a 0.22 mm MCE lter head to remove
the sorbents, and analyzed by UV-vis spectrometry (see below).
The lter membrane was proved to have a negligible sorption
amount aer ltration of the adsorbates compared without
ltration. To take into account any potential adsorbate loss to
experimental apparatus (e.g., glassware and septum) or through
volatilization during the experiments, calibration curves were
obtained from control treatments that underwent the same
experimental procedure without any sorbent added. Calibration
curves included at least 6 standard concentration levels to cover
the test concentration range. The sorbed mass of adsorbate was
calculated from the difference between the initial and nal
aqueous-phase masses, and was calculated using the following
equation:
precipitates were ltered off and washed with excess water and
acetone repeatedly. The nal product, poly(b-CD)-co-(b-CD-
(N3)7), was obtained as a white solid (9.69 g) aer drying under
ꢀ
vacuum at 100 C for one day.
For the synthesis of the CD homopolymers, only b-CD-(N3)7
(2.00 g, 1.53 mmol) or b-CD (2.17 g, 1.91 mmol) was used and
the reaction was performed via the same procedure with only
decreased amounts of TCL (4.34 g, 21.37 mmol for b-CD-(N3)7;
5.43 g, 26.78 mmol for b-CD) and solvents (pyridine, 30 mL;
THF, 10 mL). The poly(b-CD-(N3)7) and poly(b-CD) were ob-
ꢀ
tained as white solids aer drying under vacuum at 100 C for
one day.
Synthesis of N-(prop-2-yn-1-yl)D-gluconamide
In a 250 mL round-bottom ask equipped with a magnetic
stirrer bar, D-(+)-gluconic acid d-lactone (10 g, 56.2 mmol) and
propargylamine (5.32 g, 96.6 mmol) were dissolved in methanol
(100 mL). The solution was stirred under ambient temperature
and aer 0.5 h a occulent precipitate started to appear in the
ask. The amount of precipitate stopped increasing aer ꢂ5 h
and thus the reaction was stopped. The suspension was
concentrated via rotary evaporation to a quarter of its original
volume and then poured into an excess of DCM. The formed
white precipitate was ltered off and washed with DCM and 2-
propanol alcohol repeatedly. The nal product, (2R,3R,4R,5S)-
2,3,4,5,6-pentahydroxy-N-(prop-2-yn-1-yl)hexanamide (N-(prop-
2-yn-1-yl)D-gluconamide) was obtained as white crystals aer
drying under vacuum (5.89 g, yield: 45%).
ðC0 ꢁ CtÞ
Qe ¼
V
(1)
M
1H NMR (D2O, 500 MHz, 298 K): d 4.32 (d, J ¼ 3.81 Hz, 1H,
C]O–C–H), 3.60–4.15 (multiple peaks, sugar residues), 2.59 (t,
J ¼ 2.50 Hz, 1H, C^C–H) ppm. 13C NMR (D2O, 125 MHz, 298 K):
d 173.5 (C]O), 78.5 (HC^C–CH2), 72.3 (HC^C–CH2), 71.0 (C]
O–CH–), 70.8 (CH–CH2–OH), 70.1 (C]O–CH–CH–), 69.5 (CH–
CH–CH2–OH), 61.7 (–CH2–OH), 27.5 (HC^C–CH2) ppm. FTIR n:
3498, 3325 (C^C–H), 3250, 2117 (C^C), 1656 (C]O), 1525,
1039 cmꢁ1. ESI-MS m/z calcd for C9H15NO6H+ (M + H) 234.09,
found 234.07.
where C0 and Ct are the initial and nal concentrations
(mg Lꢁ1), respectively. M (g) and V (L) represent the weight of the
adsorbent and the volume of solution.
The MB adsorption data were tted with two commonly used
nonlinear adsorption models:
The Freundlich model,
n
Qe ¼ KFCe
(2)
where KF (mg1ꢁn Ln gꢁ1) is the Freundlich affinity coefficient
and n (unitless) is the Freundlich linearity index;
The Langmuir model,
Synthesis of CD-scaffolded GNs via click chemistry
The typical reaction procedure was carried out as below. To
a vial with a magnetic stirrer bar, poly(b-CD)-co-(b-CD-(N3)7)
(500 mg), 1-(20-propargyl)D-ribose (376 mg, 2 mmol), CuSO4-
KLCe
Qe ¼
Qmax
(3)
1 þ KLCe
$5H2O (37 mg, 0.15 mmol) and water (5 mL) were added and the where KL (L mgꢁ1) is the Langmuir affinity coefficient and Qmax
suspension was degassed via nitrogen bubbling for 10 min. (mg gꢁ1) is the maximum adsorption capacity.
Subsequently, 1 mL of a degassed aqueous solution containing
Single batch experiments were conducted to assess the
NaAsc (44 mg, 0.22 mmol) was added into the suspension and adsorption kinetics with repeated sampling (1 mL of boron or
reacted at 50 ꢀC for one day. Aer the reaction, the solution was 0.5 mL of MB each time for analysis) using 100 mL vials. The
removed via centrifugation and the solids were re-dispersed in initial concentration of boron was 300 mg Lꢁ1 with a pH value of
water by ultra-sonication. The centrifugation-wash cycle was 9.2 and the initial concentration of MB was 400 mg Lꢁ1 with
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A, 2018, 6, 21193–21206 | 21195

View Article Online
Journal of Materials Chemistry A
Paper
a pH value of 6.8. The kinetic data were subsequently described the cross-linking reaction of b-CD-(N3)7 to form an insoluble
by pseudo-rst-order and pseudo-second-order kinetics models, polymer framework with a CD core moiety and persubstituted
azide groups on the primary face of the b-CD. Unmodied b-
k1t
logðQe ꢁ QtÞ ¼ logðQeÞ ꢁ
(4)
(5)
CD was selectively incorporated into the reaction to manipu-
late the ratio of azide functionality and the spatial structure of
the GNs. Aer that, the CuAAC reactions of the CD polymer
with alkyne-functionalized polyols were performed in order to
generate the desired functional glycoclusters. The immobi-
lized multivalent saccharides with cis-diols were used for
coordination with boron or interaction with adsorbates via
secondary bonding etc., while the hydrophobic CD core could
encapsulate organic molecules through the typical host–guest
interactions.
TCL-cross-linked b-CD-(N3)7 in the presence and absence of
b-CD, named as poly(b-CD)-co-(b-CD-(N3)7) and poly(b-CD-(N3)7)
separately, have been synthesized as the scaffolds for further
immobilization with cis-diol-containing saccharides. As shown
in Fig. 1, FTIR spectra revealed the expected characteristic
bands for the functional groups of the CD polymers, such as the
azide groups at ꢂ2100 cmꢁ1, the residual hydroxyl groups at
ꢂ3300 cmꢁ1 and the important ester linkage (C]O) at
1718 cmꢁ1. It is worth noting that the intensity of the absorp-
tion band from the azide group of poly(b-CD)-co-(b-CD-(N3)7) is
much weaker than that of the poly(b-CD-(N3)7), probably due to
the incorporation of b-CD which unavoidably decreased the
relative content of azide functional groups in the polymer.
Subsequent elemental analysis (EA, Table 1) revealed that
the weight content of nitrogen for the poly(b-CD-(N3)7) was
8.14%, which was 2 times higher than that of the poly(b-CD)-co-
(b-CD-(N3)7) (3.51%) and further proved the presence of
unmodied b-CD in the poly(b-CD)-co-(b-CD-(N3)7).
2:303
t
1
t
¼
þ
:
2
Qt
Qe
k2Qe
To evaluate the reusability of the adsorbents for the
adsorption of boric acid, regeneration was performed by acid
leaching. The regenerated sample was rinsed in HCl solution
(pH ¼ 3) and shaken overnight, then neutralized with DI water,
and dried via lyophilization for re-adsorption. The adsorbents,
aer adsorption of MB, were regenerated by efficient washing
with solution. Firstly, the sample was immersed in MeOH
containing 5% of 1 M HCl and was shaken overnight, then the
sample was washed with plenty of DI water, and was nally
dried under vacuum.
The boron concentration was determined by UV-vis spec-
trometry using the Azomethine-H acid method.50 The MB
concentration was analyzed by UV-vis spectrometry in full
spectrum mode (185–800 nm) or single wavelength mode
(610 nm). Unless otherwise stated, error bars in all gures
represent standard deviations calculated from triplicates.
Results and discussion
Synthesis and characterization of bifunctional GNs
The total synthesis route of the cross-linked GNs is shown in
Scheme 1. Firstly, terephthaloyl chloride (TCL) was used for
Scheme 1 Schematic representation of the synthesis of bifunctional GNs.
21196 | J. Mater. Chem. A, 2018, 6, 21193–21206
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
Fig. 2 N2 adsorption (blue)/desorption (red) isotherms and pore
volume vs. pore diameter (inset) for poly(b-CD-(N3)7) and poly(b-CD-
(N3)7)@mannose. SBET is the Brunauer–Emmett–Teller (BET) surface
area (in units of m2 gꢁ1) calculated from the N2 adsorption isotherm; p
and p0 are the equilibrium and saturation pressures of N2 at 77 K,
respectively. The Barrett–Joyner–Halenda (BJH) method was used to
estimate the pore volume and pore diameter from the adsorption
isotherm.
Fig. 1 FTIR spectra of the GNs obtained by click reaction.
Table 1 Elemental compositions for typical GNs
Elemental analysis (%)
GNs
C
H
N
Oa
Poly(b-CD)-co-(b-CD-(N3)7)
Poly(b-CD)-co-(b-CD-(N3)7)@ribose
Poly(b-CD-(N3)7)
54.87
51.05
54.47
47.67
4.11
4.87
4.11
4.75
3.55
2.74
8.14
7.43
37.47
41.34
33.28
40.15
the pore structure was from textural mesoporosity rather than
the b-CD (ꢂ0.7–0.8 nm in diameter). All these results proved the
successful synthesis of mesoporous polymers with a CD core
moiety and azide functional groups.
Poly(b-CD-(N3)7)@gluconolactone
a
Oxygen content calculated from measured C, H, and N values.
Alkyne-functionalized polyols, including typical pyranoses,
furanoses and chain-type polyol compounds, were then
synthesized via a one-pot and one-step reaction as shown in
Scheme 2. Different pyranoses and furanoses have been modi-
ed with alkyne groups via Fischer glycosylation for the
synthesis of glycopolymers through a “post-polymerization
modication” method.40,49,54,55 These glycosides were generally
anomeric mixtures and were directly used without further
separation. The mole ratio of an anomeric mixture (e.g. 1-(20-
propargyl)D-ribose) could be dened by the adjacent triple
peaks corresponding to the terminal alkynyl hydrogen at
ꢂ2.5 ppm in the 1H NMR spectrum (a : b ¼ 1 : 1, Fig. S2†). The
FTIR spectrum (Fig. S3†) also revealed the signicant appear-
ance of C–H and C^C stretches of the terminal alkyne at
ꢂ3281 cmꢁ1 and ꢂ2117 cmꢁ1 separately.
Thermogravimetric analyses (TGA, Fig. S1†) of poly(b-CD)-co-
(b-CD-(N3)7) and poly(b-CD-(N3)7) both showed one sharp
degradation step, centred at ꢂ321 ꢀC and 310 ꢀC separately,
which accounted for ꢂ40% weight loss. Thus the incorporation
of pristine b-CD will increase the thermal stability of the CD
scaffold, possibly due to the higher crosslinking degree and the
presence of less unstable b-CD-(N3)7 in the poly(b-CD)-co-(b-CD-
(N3)7).
N2 adsorption/desorption isotherms were then used for the
characterization of the specic surface area (SSA) and pore
parameters of the cross-linked CD polymers. As shown in Fig. 2,
an obvious hysteresis loop in the range of p/p0 > 0.8 and notably
a sharp rise in the adsorbed volume for p/p0 > 0.9 were observed
in the N2 adsorption/desorption isotherms. The SSA estimated
for the poly(b-CD-(N3)7) was 21.02 m2 gꢁ1, which is at the same
level as those of typical urethane-based CD polymers.51 This
value is relatively smaller than those of the activated carbons
and certain porous b-CD-containing polymers such as the ones
cross-linked by tetrauoroterephthalonitrile etc., however, it is
larger than the SSA of CD–calixarene nanosponges with triazole
linkers.26,52,53 Due to their limited N2 adsorption, such polymers
were considered as non-porous materials in previous reports;
however, the SSA is much larger than that of the crystalline b-CD
(0.635 m2 gꢁ1 as previously reported) and thus the interstitial
framework mesopores should not be overlooked.26,51 The
average pore diameter calculated by the BJH adsorption
isotherm was ꢂ35 nm and the pore size distribution (Fig. 2,
inset) suggests that the major contribution to the majority of
To expand the library of cis-diol-containing compounds
rather than pyranose and furanose, an open-chain polyol
Scheme 2 Synthesis of alkyne-functionalized polyols.
J. Mater. Chem. A, 2018, 6, 21193–21206 | 21197
This journal is © The Royal Society of Chemistry 2018

View Article Online
Journal of Materials Chemistry A
Paper
compound was synthesized via the amidation reaction of D- poly(b-CD)-co-(b-CD-(N3)7) (up to ꢂ200 ꢀC, Fig. S1†). The
(+)-gluconic acid d-lactone with propargylamine. This mild differences in the thermal stabilities of the GNs are possibly
reaction could give the target product without the need for owed to the varied chemical stabilities of the different saccha-
chromatography purication. 1H and 13C NMR spectra were rides. However, N2 adsorption/desorption isotherms (Fig. 2)
used to conrm the purity of the product and the numbering of revealed a dramatic decrease of the SSA for poly(b-CD-(N3)7)
the hydrogen and carbon atoms used for the NMR peak @mannose aer the click reaction, from 21.02 m2 gꢁ1 to
assignments is shown in Fig. 3. The FTIR spectra (Fig. S5†) also 3.34 m2 gꢁ1, which indicated that the addition of more
demonstrated the disappearance of absorbance from the ester saccharides may increase the packing density and attenuate the
bond (C]O) at ꢂ1750 cmꢁ1 as well as the appearance of C–H SSA of the framework. The pore size also signicantly decreased
and C^C stretches of the terminal alkyne at ꢂ3325 and to less than 5 nm (Fig. 2, inset) and revealed a change of pore
ꢂ2117 cmꢁ1 and the amide bond (C]O) at ꢂ1650 cmꢁ1. All of structure. Interestingly, it was found that poly(b-CD)-co-(b-CD-
these reactions used inexpensive reagents and could be easily (N3)7)@mannose showed a similar SSA as poly(b-CD) and
scaled up to 100 gram scale in the lab.
poly(b-CD)-co-(b-CD-(N3)7) ranging from 15.2 to 22.5 m2 gꢁ1
Subsequently, these cis-diol-containing compounds were (Fig. S6†). It was hypothesized that the existence of unmodied
immobilized to the CD polymers by CuAAC click reactions using CD may reduce the occupation of the pore structure due to the
CuSO4/NaAsc as the catalyst (Scheme 1). The successful click reaction.
synthesis of cross-linked glycoclusters could be proved by
Finally, the morphological properties of the GNs were
characterization using FTIR, EA, TGA and N2 adsorption/ investigated using scanning electron microscopy (SEM). As
desorption isotherms. As shown in Fig. 1 and S5,† the absor- shown in Fig. 4, SEM images of poly(b-CD)-co-(b-CD-(N3)7)
bance at ꢂ2100 cmꢁ1 totally vanished, which indicated the revealed the main presence of aggregated sphere-like nano-
complete consumption of the azide groups. The broad absor- particles with diameters ranging from ꢂ0.1 mm to ꢂ0.3 mm.
bance at 3000–3600 cmꢁ1 increased dramatically due to the Aer the click reaction, the sphere-like morphology was
additional hydroxyl groups from the immobilized saccharides. retained for poly(b-CD)-co-(b-CD-(N3)7)@mannose; however, the
For the products aer the click reaction, the elemental surfaces of the spheres became rougher. Interestingly, aggre-
compositions (Table 1) revealed a decrease in the nitrogen and gated irregular particles rather than spheres were found for the
carbon content accompanied with an increase of the hydrogen poly(b-CD-(N3)7) and poly(b-CD-(N3)7)@mannose. Pores and
and oxygen content, which was mainly due to the higher tunnels were found between these particles and may favour the
content of hydrogen and oxygen in the immobilized saccha- efficient transport of small molecules into the GNs.
rides. It is also worth noting that the nitrogen contents would
In summary, cross-linked GNs have been facilely synthesized
theoretically decrease to 3.06% and 7.47% if the azide groups from monosaccharides and b-CD via a combination of a cross-
were totally reacted for poly(b-CD)-co-(b-CD-(N3)7)@ribose and linking reaction, glycosylation and click chemistry. These
poly(b-CD-(N3)7)@gluconolactone respectively, which are very mesoporous polymer frameworks have multiple b-CD cores as
close to the experimental values (2.74% and 7.43% respectively). hydrophobic cavities and are covalently modied with different
TGA analysis revealed
a different thermal-decomposition cis-diol-containing saccharides.
behaviour aer the click reaction. As shown in Fig. S1,† the
GNs showed smoother traces aer the click reaction compared
with that of the TCL-cross-linked b-CD-(N3)7. Aer immobili-
zation with more saccharides, the degradation of the GNs
started at ꢂ120 ꢀC, which was much lower than that of the
Boron removal performance of the varied functional GNs
Boron could be removed from water in the form of a borate
ester, which was formed by the complexation reaction with
Fig. 4 SEM images of poly(b-CD)-co-(b-CD-(N3)7) (A), poly(b-CD)-
Fig. 3 1H and 13C NMR spectra of N-(prop-2-yn-1-yl)D-gluconamide co-(b-CD-(N3)7)@mannose (B), poly(b-CD-(N3)7) (C) and poly(b-CD-
in D2O.
(N3)7)@mannose (D).
21198 | J. Mater. Chem. A, 2018, 6, 21193–21206
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
polyol groups immobilized on the surface of the adsorbents. suggested that the diols from these two saccharides could
Poly(b-CD)-co-(b-CD-(N3)7) and poly(b-CD-(N3)7) were rst used coordinate with borate more efficiently than the residual
for the boron adsorption and a commercially available boron- hydroxyls from CD. When these functional pyranoses were
specic ion exchange resin (Amberlite IRA743) was used as covalently linked to poly(b-CD-(N3)7), which theoretically
a control test. The boron adsorption capacities of poly(b-CD)-co- contains more azide groups and thus more saccharides could
(b-CD-(N3)7) and poly(b-CD-(N3)7) polymers were ꢁ1.86 ꢃ be incorporated, the adsorption capacities of the obtained
0.62 mg gꢁ1 and ꢁ1.24 ꢃ 1.60 mg gꢁ1 respectively, which reveal poly(b-CD-(N3)7)@mannose, poly(b-CD-(N3)7)@glucose and
that the unmodied CD polymers had almost no ability to poly(b-CD-(N3)7)@glucosamine showed an unusual decreasing
adsorb boron, while under the same conditions, Amberlite resin trend to almost zero, at 8.16 ꢃ 2.74, ꢁ2.78 ꢃ 7.78 and 1.79 ꢃ
showed a sorption capacity as high as 55.83 ꢃ 1.86 mg gꢁ1 1.65 mg gꢁ1 (Table 1) respectively. This unusual decrease may
(Table 2).
be due to the signicantly decreased SSA compared with the
The presence of residual hydroxyl groups in the azide- other GNs (Fig. 2), which may prevent the efficient contact of
functionalized CD polymers has been revealed clearly by the boron with the saccharides inside the GNs.
FTIR spectra (Fig. 1), however these hydroxyl groups of b-CD
Aer that, a typical furanose, D-ribose, was covalently linked
were either single hydroxyl groups or trans-diols and thus may to the primary face of b-CD for boron adsorption. The adsorp-
not be able to coordinate with boron and form stable borate tion capacity showed a signicant increase to 19.04 ꢃ 4.39 mg
esters. Subsequently, different functional polyols were immo- gꢁ1 for poly(b-CD)-co-(b-CD-(N3)7)@ribose, which was much
bilized in order to supply cis-diols for complexation with boron. higher than the sorption capacity of the glucose or N-acetyl-D-
Aer poly(b-CD)-co-(b-CD-(N3)7) was covalently immobilized glucosamine-functionalized glycoclusters. It has been reported
with mannose groups, the adsorption capacity tended to be very that the local stability constants of the borate esters comprised
low (0.87 ꢃ 1.67 mg gꢁ1). Although the hydroxyl groups from the of cis-1,2-diol furanose (2500–45 000) are hundreds or even
carbons two, three, four and six are in adjacent positions, they several thousand times higher than those of the cis-1,2-diol
orient in different directions and may not favour the formation pyranoses (10–20) and no coordination was observed when
of tetrahedrally coordinated structures. It has been reported using trans-1,2-diol pyranoses/furanoses.22,58 Thus the coordi-
that the 1, 2-diol or 1, 3-diol from glucose and N-acetyl-D- nation of the hydroxyl groups from carbons two and three of
glucosamine could efficiently coordinate with phenylboronic ribose with boron should be much more stable than that of the
acid or its derivatives to form stable phenylborates.20,56,57 For other pyranoses, which could explain the higher adsorption
poly(b-CD)-co-(b-CD-(N3)7)@glucose and poly(b-CD)-co-(b-CD- capacity. Interestingly, the poly(b-CD-(N3)7)@ribose containing
(N3)7)@glucosamine, the adsorption capacities reached 4.45 ꢃ more ribose also showed a low sorption capacity (2.29 ꢃ
4.14 and 6.86 ꢃ 0.62 mg gꢁ1 (Table 2), respectively, which 1.92 mg gꢁ1, Table 2) and accounted for only ꢂ10% of the
Table 2 Boron adsorption capacity of the different GNsa
Sorption capacity (Qe, mg gꢁ1) for
Immobilized polyols (X)b
Structure
Poly(b-CD)-co-(b-CD-(N3)7)@X
Poly(b-CD-(N3)7)@X
b-CD residual
Mannose
ꢁ1.86 ꢃ 0.62
0.87 ꢃ 1.67
4.45 ꢃ 4.14
ꢁ1.24 ꢃ 1.60
8.16 ꢃ 2.74
Glucose
ꢁ2.78 ꢃ 7.78
Glucosamine
6.86 ꢃ 0.62
1.79 ꢃ 1.65
Ribose
19.04 ꢃ 4.39
6.18 ꢃ 3.15
2.29 ꢃ 1.92
Gluconolactone
31.35 ꢃ 2.04
Amberlitec
55.83 ꢃ 1.86
a
All the tests were performed by shaking a suspension of 100 mg solids in 50 mL boron solution (initial concentration of boron: 300 mg Lꢁ1; pH ¼
9.20) at 25 ꢀC. All tests were performed in triplicate. b X represents the functional groups immobilized onto the surface of poly(b-CD)-co-(b-CD-(N3)7)
and poly(b-CD-(N3)7). Amberlite (IRA 743) is a commercial resin modied with N-methyl-D-glucamine and was used in a control experiment.
c
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A, 2018, 6, 21193–21206 | 21199

View Article Online
Journal of Materials Chemistry A
Paper
adsorption capacity of poly(b-CD)-co-(b-CD-(N3)7)@ribose (19.04
ꢃ 4.39 mg gꢁ1), which is in accordance with the above-
mentioned speculation that separated spatial conguration
will not favour the formation of stable borate esters.
Finally, GNs immobilized with chain-type polyols were used
for boron adsorption. The poly(b-CD)-co-(b-CD-(N3)7)@gluco-
nolactone showed a similar sorption capacity (6.18 ꢃ 3.15 mg
gꢁ1) as the pyranose-containing polymers and only 1/3 of that of
the ribose-containing GNs, which further proved that cis-diols
of the pyranoses could coordinate with boron more efficiently
and stably. To our surprise, the poly(b-CD-(N3)7)@glucono-
lactone showed an adsorption capacity of 31.35 ꢃ 2.04 mg gꢁ1
,
which was 5 times higher than that of poly(b-CD)-co-(b-CD-
(N3)7)@gluconolactone. It is worth noting that theoretically
poly(b-CD-(N3)7)@gluconolactone contains only ꢂ2.3 times
more N-(prop-2-yn-1-yl)D-gluconamide than poly(b-CD)-co-(b-
CD-(N3)7)@gluconolactone, which was calculated according to
the content of azide by elemental analysis. It was hypothesized
that the polyols tend to have a longer and soer chain than the
pyranoses or furanoses with rigid ring structures and this may
also facilitate the match of cis-diols with boron. All these results
proved that the immobilization of functional polyols could
endow the glycopolymers with boron adsorption abilities.
To further prove the chemical adsorption nature of the
boron removal, TGA and DSC were used to investigate the
changes in the thermal degradation of the GNs before and aer
the adsorption of boron. As shown in Fig. 5A, the degradation of
poly(b-CD)-co-(b-CD-(N3))@ribose started slowly aer ꢂ150 ꢀC
Fig. 5 TGA curves (A) and DSC scans (B) of poly(b-CD)-co-(b-CD-
(N3))@ribose and poly(b-CD-(N3))@gluconolactone before and after
the adsorption of boric acid.
ꢀ
and accelerated aer ꢂ250 C, which led to a weight retention
ratio of 20% at 700 ꢀC. While aer boron adsorption, the
degradation of the poly(b-CD)-co-(b-CD-(N3))@ribose-borate
Polyols have strongly pH-dependent adsorption capacities
product only started aer ꢂ240 ꢀC and accelerated aer for boron since polyols react preferentially with boronate
ꢂ280 ꢀC and nally led to a higher weight retention ratio anions. The optimised pH is around 9.0–9.5, at which the
(ꢂ25%) at 700 C. Further TGA characterization of poly(b-CD- dominant species is the boronate anion (B(OH)4ꢁ) rather than
ꢀ
(N3))@gluconolactone also revealed
a higher degradation boric acid (H3BO3) thus the coordination with cis-diols would be
starting temperature and a higher weight retention ratio aer more efficient.4,59,60 The pH effect on the adsorption of boron by
boron adsorption. This phenomenon suggested that the coor- two typical GNs, poly(b-CD)-co-(b-CD-(N3)7)@ribose and poly(b-
dination of cis-diols with boron made the GNs more stable CD-(N3)7)@gluconolactone was systematically studied over a pH
during thermal degradation. The higher weight retention ratio range from 5.0 to 11.0. As shown in Fig. 6, the Qe of each gly-
could be due to the higher stability of the inorganic boron salt copolymer adsorbent increased pronouncedly with increasing
compared with the degradable organic polymers.
pH until pH 9.0 and subsequently decreaseꢁd from pH 9.0 to
Subsequent DSC analysis, as shown in Fig. 5B, revealed the 11.0. This is ascribed to the fact that B(OH)4 is the dominant
presence of large and broad exothermic peaks above 200 ꢀC, species, according to the pKa of 9.25, under slightly alkaline
which were typical signs of side reaction such as oxidation, conditions, which favours the generation of boronate esters via
cross-linking or solidifying during thermal degradation. The coordination with polyols. It is worth noting that the glyco-
exothermic peak temperature of poly(b-CD)-co-(b-CD-(N3)) clusters still showed considerable adsorption capacity around
@ribose signicantly increased from 298 ꢀC to 310 ꢀC aer the neutral pH, however, the adsorption capacity decreased signif-
adsorption of boric acid, which was in accordance with the shi icantly at relatively extreme pH values such as 5.0 or 11.0.
of the degradation temperature in the TGA analysis. The
The variation of boron adsorption as a function of contact
melting temperature (Tm) for poly(b-CD)-co-(b-CD-(N3))@ribose, time is shown in Fig. 7 using poly(b-CD)-co-(b-CD-(N3)7)@ribose
which was revealed by the endothermic peaks below 100 ꢀC, also and poly(b-CD-(N3)7)@gluconolactone as the adsorbents. In
shied from 55 ꢀC to 65 ꢀC aer boron adsorption. The Tm and both cases, the boron sorption is a relatively slow process and
degradation temperature also changed due to adsorption of needs four hours to reach ꢂ90% of the total capacity. It was
boron for poly(b-CD-(N3))@gluconolactone. All of these results interesting to nd that the ribose-containing glycoclusters
successfully proved the chemical coordination of boron to the could reach ꢂ37% of their total boron capacity in 20 min, which
cis-diol-containing GNs, which unavoidably changed the struc- was faster than the glycoclusters immobilized with more chain-
tures and properties of the nal adsorbent-borate products.
type polyols, which only reached ꢂ29% of their total efficiency
21200 | J. Mater. Chem. A, 2018, 6, 21193–21206
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
The regeneration of GNs for boron removal was subsequently
investigated. Regeneration performed by a combination of acid
elution (such as 1.0 M HCl solution) and base neutralization
(such as 0.1 M NaOH solution), which would destroy the borate
esters and reactivate the amine groups separately, was the most
frequently-used strategy in previous reports.8,60 In our system, no
amine groups exist in the GNs and thus no base was needed for
the regeneration. Thus efficient elution using acid solution and
deionized water was utilized to regenerate the GNs. At rst, 1.0 M
HCl solution was trialled to destroy the borate ester, however, the
acid was so strong that the ester type cross-linkers were also
destroyed and hydrolysis of the GNs occurred. Then, milder acid
solution with a pH of 3.0 was utilized for the regeneration. As
shown in Fig. 8, poly(b-CD)-co-(b-CD-(N3)7)@ribose and poly(b-
CD-(N3)7)@gluconolactone were utilized in ve adsorption/
desorption cycles. During the tests, a gradual decrease of
adsorption capacity was observed and only ꢂ20% of the pristine
adsorption capacity was le over the ve regeneration cycles. It
was hypothesized that the regeneration was signicantly affected
by the pH of the elution solution and desorption may not have
completely happened under such mild conditions.
Fig. 6 Effects of pH on the boron adsorption of poly(b-CD)-co-(b-
CD-(N3)7)@ribose and poly(b-CD-(N3)7)@gluconamide (100 mg GNs in
50 mL boron solution, C0 ¼ 300 mg Lꢁ1; 25 ꢀC).
In summary, the GNs have shown varied adsorption capacities
for boron under different conditions, which were mainly depen-
dant on the structure of the GNs and the immobilized polyols as
well as the pH value of the solutions. The kinetic studies show
that the adsorption ts a pseudo-second-order kinetic model and
is mainly controlled by a chemical reaction, however, the small
SSAs of the mesoporous materials resulted in a relatively slow
adsorption rate. The decrease of the adsorption capacity found
during the repeated adsorption/desorption tests was possibly due
to the mild reaction conditions utilized. Different crosslinking
reagents such as more stable tetrauoroterephthalonitrile etc. or
more violent regeneration conditions could be an option to
improve the adsorption properties in future research.
Fig. 7 Boron adsorption kinetics expressed as adsorbed concentra-
tion (Qe, mg gꢁ1) at a given time (T, h) for different GN adsorbents
(100 mg GNs in 50 mL boron solution, C0 ¼ 300 mg Lꢁ1, pH ¼ 9.20; 25
ꢀC).
Adsorption of methylene blue (MB) by the GNs
Cross-linked CD polymers and GNs were studied to determine
the adsorption isotherms, adsorption kinetics and regeneration
in 20 min. This suggests that ribose-containing glycoclusters
have higher efficiency, especially in the early period of boron
adsorption. Previously reported mesoporous resins or silicas
containing polyols have shown faster adsorption rates with
a total efficiency of up to 90% obtained in several minutes.60,61
TCL-cross-linked GNs in this research have shown much
smaller SSAs (<20 m2 gꢁ1) than previously mentioned meso-
porous materials (even up to 877.7 m2 gꢁ1), which may prevent
the efficient contact of boron with the functional saccharides
within the pores, thus resulting in a slower adsorption rate.
The kinetic data were subsequently described by pseudo-
rst-order and pseudo-second-order kinetics models. The
pseudo-second-order kinetics model generally ts the kinetic
data better (as indicated by the R2 values), which justies the
adsorption mechanism with a high correlation coefficient
(Table S1†) for GN adsorbents. This indicates that the exchange-
adsorption rate is mainly controlled by a chemical process and
thus supports the adsorption of boron via the formation of
boronate esters.
Fig. 8 Regeneration of GNs for the adsorption of boron (dose of 2 g
L
ꢁ1; HCl acid solution, pH 3.0; equilibrium time of 12 h; 100 mg GNs in
50 mL boron solution, initial concentration of 300 mg Lꢁ1; 25 ꢀC).
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A, 2018, 6, 21193–21206 | 21201

View Article Online
Journal of Materials Chemistry A
Paper
using MB as a model organic pollutant. MB is a cationic dye based materials) in previous reports, which have been listed
with wide applications in the printing industry and has been in Table S3.†62–66 The effects of pH or salt ion strength on the
studied as a typical compound for encapsulation by CD poly- adsorption of MB by the GNs are shown in Fig. 9B. The effect of
mers. Fig. 9A compares the adsorption isotherms of MB for the pH on the removal efficiency of MB was not signicant, with
different GN adsorbents, plotted as the adsorbed-phase only a slightly increase in alkaline solutions compared with that
concentration (Qe, mg gꢁ1) vs. the aqueous-phase concentra- under acidic conditions. Addition of 0.1 M salt had no obvious
tion (Ce, mg Lꢁ1) at equilibrium. The adsorption data were tted inuence on the MB adsorption, which indicated that the GNs
with two commonly used nonlinear adsorption models: the showed good adsorption potential in water treatment.
n
Freundlich model, Qe ¼ KFCe , where KF (mg1ꢁn Ln gꢁ1) is the
The adsorption capacities (Qe) for MB of the different GNs
Freundlich affinity coefficient and n (unitless) is the Freundlich are listed in Table 3. Compared with poly(b-CD) which showed
a capacity up to 1159 ꢃ 271 mg gꢁ1, poly(b-CD)-co-(b-CD-(N3)7)
KLCe
linearity index; and the Langmuir model, Qe ¼
where KL (L mgꢁ1) is the Langmuir affinity coefficient and Qmax
(mg gꢁ1) is the maximum adsorption capacity.
The model tting parameters are summarized in Table S2.†
The Freundlich model generally ts the adsorption data better
in the concentration range measured (as indicated by the R2
values >0.95). The adsorption capacity for MB could reach an
astonishing 1155 ꢃ 124 mg gꢁ1 for the poly(b-CD)-co-(b-CD-
(N3)7), which is much higher than the most frequently-used
adsorbents (activated carbon or other reported cyclodextrin-
Qmax;
had a similar Qe for MB (1155 ꢃ 124 mg gꢁ1) under the same test
conditions, while poly(b-CD-(N3)7) showed a much smaller Qe
(616 ꢃ 112 mg gꢁ1). Although the slight difference in the
molecular weights (MW) of b-CD and b-CD-(N3)7 may induce
different Qe values for the GNs, the main reason for the
dramatic difference in Qe should be attributed to the structure
of poly(b-CD-(N3)7). Due to the totally substituted primary face
of b-CD, crosslinking could only occur between the hydroxyl
groups derived from the secondary face and this will unavoid-
ably cause overlap between the bottom portals of the macrocy-
clic cavity as shown in Scheme 3. This may further prevent the
efficient encapsulation of MB into the cavities, and thus
signicantly decrease the adsorption of MB. Besides, the mes-
oporous structure would become more clogged aer immobi-
lization with more saccharides and this may not favour the
encapsulation of MB.
1 þ KLCe
For the poly(b-CD)-co-(b-CD-(N3)7)@saccharides, the Qe
values ranged from 821 ꢃ 21 to 1152 ꢃ 218 mg gꢁ1 (Table 3). It is
worth noting that there was 0.85 mmol of azide groups per gram
of poly(b-CD)-co-(b-CD-(N3)7) according to the nitrogen content
by EA, which means that the weight content of the functional
monosaccharides in the GNs could range from 14% (for ribose)
to 18% (for glucosamine). Thus theoretically there should be
fewer CD core cavities and a lower Qe aer the click reaction.
However, for the poly(b-CD-(N3)7)@saccharides, the Qe gener-
ally increased even aer immobilization of the mono-
saccharides or polyols, ranging from 553 ꢃ 35 to 978 ꢃ
90 mg gꢁ1. We believe this could be attributed to the secondary
bond interactions between MB and the saccharides, most likely,
hydrogen bonding and van der Waals forces. Considering there
was 1.94 mmol of azide groups per gram of poly(b-CD-(N3)7), the
weight content of the functional monosaccharides in the GNs
could range from 27% (for ribose) to 33% (for glucosamine).
With much fewer CD core cavities, poly(b-CD-(N3)7)@glucos-
amine showed a much higher Qe (978 ꢃ 90 mg gꢁ1, almost half
as much again, Table 3) and this indicates that the secondary
bonding between MB and the saccharides contributes even
more for the adsorption of MB rather than the host–guest
interactions. The adsorption capacity for cationic MB was
generally much higher compared with that of the anionic
borate. It needs to be noted that there are no other efficient ways
to remove boron from water other than coordination with cis-
diol-containing compounds. However, MB could be encapsu-
lated into the hydrophobic cavities of the porous CD polymers.
The presence of hydroxyl-rich saccharides could enhance the
adsorption via hydrogen bonding and van der Waals forces etc.
Fig.
9 Adsorption isotherms plotted as the adsorbed-phase
concentration (Qe) vs. the aqueous-phase concentration (Ce) at
equilibrium for the different GN adsorbents ((A), Freundlich fitting) and
the effect of pH and salt ions (B) on MB adsorption by the GNs
(10 mg GNs in 50 mL MB solution, C0 ¼ 400 mg Lꢁ1; equilibrium time,
24 h; 25 ꢀC).
21202 | J. Mater. Chem. A, 2018, 6, 21193–21206
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
Table 3 Adsorption capacities for MB obtained by the GNsa
Sorption capacity (Qe, mg gꢁ1) for
Poly(b-CD)-co-(b-CD-(N3)7)@X
Immobilized polyols (X)b
Structure
Poly(b-CD-(N3)7)@X
b-CD residual
Mannose
1155 ꢃ 124
941 ꢃ 116
821 ꢃ 21
616 ꢃ 112
553 ꢃ 35
583 ꢃ 65
Glucose
Glucosamine
1152 ꢃ 218
978 ꢃ 90
Ribose
862 ꢃ 262
833 ꢃ 90
723 ꢃ 49
Gluconolactone
1026 ꢃ 153
Poly(b-CD)
1159 ꢃ 271
a
All the tests were performed by shaking the suspension of 10 mg solids in 50 mL MB solution (initial concentration: 400 mg Lꢁ1; pH ¼ 6.8) under
25 ꢀC. All tests were performed in triplicate. b X represents the functional groups immobilized onto the surface of poly(b-CD)-co-(b-CD-(N3)7) and
poly(b-CD-(N3)7).
could reach more than 70% of the adsorption capacity within
10 min and the equilibrium time was signicantly decreased to
ꢂ5 h (Fig. 10). The kinetics data of MB adsorption were also
described by pseudo rst-order and second-order kinetics
models. As shown in Table S4,† similar to the boron kinetics
data, the pseudo-second-order kinetics model generally ts the
kinetic data better (the R2 values were close to 0.99). The
adsorption rate constants (k2, g mgꢁ1 hꢁ1, Table S4†) for the GNs
were signicantly higher than the values for the unmodied CD
polymers, which further revealed the higher adsorption rate of
the GNs. These results strongly prove that the adsorption of MB
by the GNs obeys second-order adsorption kinetics and may be
Scheme 3 Schematic representation for the encapsulation of MB by
the GNs with different structures.
mainly controlled by chemical processes (hydrogen bonding etc.).
Compared with other adsorbents such as activated carbon,
functionalized graphene oxide, b-CD-based bres and b-CD-
containing nanocomposites, which showed adsorption capac-
ities for MB from 56.5 to 826.5 mg gꢁ1 (Table S3†), the GNs have
shown more astonishing adsorption capacities as high as 1152
ꢃ 218 mg gꢁ1. To the best of our knowledge, this is the highest
adsorption capacity obtained by CD-containing polymeric
adsorbents so far.
To evaluate the adsorption properties of the GNs, the MB
adsorption kinetic data for different GNs are compared in
Fig. 10. It can be seen that the MB adsorption rates for poly(b-
CD), poly(b-CD)-co-(b-CD-(N3)7) and poly(b-CD-(N3)7) were rapid
in the rst 10 min, reaching ꢂ1/3 of the adsorption capacity,
followed by a long adsorption equilibrium time for the next 12 h
(Fig. 10). Surprisingly, aer immobilization with more saccha-
rides the MB adsorption rates of the GNs were dramatically
accelerated. It can be seen that the MB adsorption by these GNs
In order to test the regeneration of the GNs, MB adsorption was
investigated over ve regeneration cycles. Aer adsorption, the
GNs were desorbed by washing rst using methanol containing
5% 1 M HCl and then using DI water, and were then dried under
vacuum for reuse. As shown in Fig. 11, the MB adsorption capac-
ities of the GNs had no signicant decrease over the ve adsorp-
tion–regeneration cycles, which proved excellent reusability.
In summary, the GNs have demonstrated excellent adsorp-
tion rates and capacities for typical organic pollutants.
Hydrogen bonding and van der Waals forces, derived from the
interactions between the immobilized saccharides and MB,
accelerate the adsorption rates and increase the adsorption
capacities, which contribute even more than the host–guest
interactions. These GNs also showed outstanding reusability
during repeated adsorption/desorption cycling experiments
and could potentially be used for water decontamination to
remove different organic pollutants.
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A, 2018, 6, 21193–21206 | 21203

View Article Online
Journal of Materials Chemistry A
Paper
frameworks with multiple b-CD cores as hydrophobic cavities
and persubstituted saccharides containing functional cis-diols
on the primary face of the b-CD. The cross-linked glycopolymers
have shown varied adsorption capacities for boron under
different conditions, mainly affected by the structure and
content of the immobilized polyols as well as the solution pH.
The adsorption kinetics for boron t a pseudo-second-order
kinetic model and the adsorption is mainly controlled by
a chemical reaction, however, the small SSAs of the mesoporous
materials resulted in relatively slow adsorption rates. The GNs
have shown excellent adsorption rates and capacities for typical
pollutants using an organic dye (MB) as the adsorbate in this
research. Hydrogen bonding and van der Waals forces between
the immobilized saccharides and substrates signicantly
accelerated the adsorption rate and increased the adsorption
capacity, which contributed even more than the host–guest
interactions. These GNs demonstrated outstanding reusability
for MB adsorption during repeated adsorption/desorption
experiments, however, a decrease in the adsorption capacity
for boron was observed, possibly due to the mild reaction
conditions utilized.
Further optimization of the saccharide type and the spatial
structure may improve the adsorption performance, stability
and reusability of these materials as high-performance adsor-
bents. Utilization of specic cross-linking reagents such as tet-
rauoroterephthalonitrile or decauorobiphenyl etc. may
signicantly increase the SSA of the GNs as well as the adsorp-
tion performance towards micropollutants. Amine-containing
polyols and saccharides such as N-methylglucamine etc. could
serve as alternatives to pyranose sugars to form more stable
borate esters and thus could favour the adsorption of boron.
Efforts should be put into increasing the saccharide density and
chain exibility such as by using glycopolymers to replace the
monosaccharides in order to increase the adsorption capacity
for boron. We believe such novel bifunctional nanosponges may
nd potential applications in seawater desalination, boron
production and water purication.
Fig. 10 Batch adsorption kinetics of MB (10 mg GNs in 50 mL MB
solution, C0 ¼ 400 mg Lꢁ1, 25 ꢀC) expressed as adsorbed concen-
tration (Qe, mg gꢁ1) at a given time (T, h) for different GN adsorbents.
Poly(b-CD) and poly(b-CD)-co-(b-CD-(N3)7)@saccharides (A); poly(b-
CD-(N3)7)@saccharides (B).
Conflicts of interest
There are no conicts to declare.
Acknowledgements
This project was conducted with nancial support from the
Natural Science Foundation of China (21504044), the China
Postdoctoral Science Foundation (157453) and the Natural
Science Foundation of Jiangsu Province for Distinguished
Young Scholars (BK20180017).
Fig. 11 Regeneration of the GNs for the adsorption of MB (dose of
0.2 g Lꢁ1; methanol containing 5% 1.0 M HCl; equilibrium time of 24 h;
10 mg of the GNs in 50 mL MB solution, initial concentration of
400 mg Lꢁ1; 25 ꢀC).
Notes and references
Conclusions
1 S. J. Kim, S. H. Ko, K. H. Kang and J. Han, Nat. Nanotechnol.,
2010, 5, 297.
2 G.-D. Kang and Y.-M. Cao, Water Res., 2012, 46, 584–600.
3 M. A. Shannon, P. W. Bohn, M. Elimelech, J. G. Georgiadis,
In conclusion, we have successfully synthesized different func-
tional GNs from monosaccharides and b-CD via a combination
of a cross-linking reaction, Fischer glycosylation and a CuAAC
click reaction. Such glycopolymers are mesoporous polymer
˜
B. J. Marinas and A. M. Mayes, Nature, 2008, 452, 301.
21204 | J. Mater. Chem. A, 2018, 6, 21193–21206
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
4 N. Hilal, G. J. Kim and C. Somereld, Desalination, 2011, 273, 31 Y. Ling, M. J. Klemes, L. Xiao, A. Alsbaiee, W. R. Dichtel and
23–35.
D. E. Helbling, Environ. Sci. Technol., 2017, 51, 7590–7598.
´
¨
5 E. Guler, C. Kaya, N. Kabay and M. Arda, Desalination, 2015, 32 D. M. Alzate-Sanchez, B. J. Smith, A. Alsbaiee,
356, 85–93.
J. P. Hinestroza and W. R. Dichtel, Chem. Mater., 2016, 28,
8340–8346.
6 P. H. Hermans, Z. Anorg. Allg. Chem., 1925, 142, 83–110.
¨
7 J. Boeseken, in Advances in Carbohydrate Chemistry, ed. 33 A. Martinez, C. Ortiz Mellet and J. M. Garcia Fernandez,
W. W.Pigm and M. L.Wolfro, Academic Press, 1949, vol. 4,
pp. 189–210.
Chem. Soc. Rev., 2013, 42, 4746–4773.
´
´
34 J. M. Benito, M. Gomez-Garcıa, C. Ortiz Mellet, I. Baussanne,
´
´
8 H. Mishra, C. Yu, D. P. Chen, W. A. Goddard, N. F. Dalleska,
M. R. Hoffmann and M. S. Diallo, Environ. Sci. Technol., 2012,
46, 8998–9004.
J. Defaye and J. M. Garcıa Fernandez, J. Am. Chem. Soc., 2004,
126, 10355–10363.
´
´
´
35 M. Gomez-Garcıa, J. M. Benito, D. Rodrıguez-Lucena,
´
J.-X. Yu, K. Chmurski, C. Ortiz Mellet, R. Gutierrez Gallego,
9 Y.-T. Wei, Y.-M. Zheng and J. P. Chen, Langmuir, 2011, 27,
6018–6025.
´
´
A. Maestre, J. Defaye and J. M. Garcıa Fernandez, J. Am.
10 J. Meng, J. Cao, R. Xu, Z. Wang and R. Sun, J. Mater. Chem. A,
2016, 4, 11656–11665.
Chem. Soc., 2005, 127, 7970–7971.
36 C. Ortiz-Mellet, J. M. Benito, J. M. G. Fernandez, H. Law,
´
11 S. Morisada, T. Rin, T. Ogata, Y.-H. Kim and Y. Nakano,
Water Res., 2011, 45, 4028–4034.
K. Chmurski, J. Defaye, M. L. O’Sullivan and H. N. Caro,
Chem.–Eur. J., 1998, 4, 2523–2531.
12 M. Jabli, N. Tka, K. Ramzi and T. A. Saleh, J. Mol. Liq., 2018, 37 M. Massaro, S. Riela, P. Lo Meo, R. Noto, G. Cavallaro,
249, 1138–1144.
S. Milioto and G. Lazzara, J. Mater. Chem. B, 2014, 2, 7732–
13 M. Jabli, T. A. Saleh, N. Sebeia, N. Tka and R. Khiari, Sci. Rep.,
2017, 7, 14448.
7738.
´
38 S. Andre, H. Kaltner, T. Furuike, S.-I. Nishimura and
14 T. A. Saleh and I. Ali, J. Environ. Chem. Eng., 2018, 6, 5361–
5368.
15 T. A. Saleh, I. B. Rachman and S. A. Ali, Sci. Rep., 2017, 7,
4573.
16 Advanced Nanomaterials for Water Engineering, Treatment,
and Hydraulics, ed. A. S.Tawk, IGI Global, Hershey, PA,
USA, 2017.
17 T. A. Saleh and V. Gupta, Nanomaterial and Polymer
Membranes: Synthesis, Characterization, and Applications,
2016.
18 M.-S. Steiner, A. Duerkop and O. S. Woleis, Chem. Soc. Rev.,
2011, 40, 4805–4839.
19 D. Li, Y. Chen and Z. Liu, Chem. Soc. Rev., 2015, 44, 8097–
8123.
20 R. Ma, H. Yang, Z. Li, G. Liu, X. Sun, X. Liu, Y. An and L. Shi,
Biomacromolecules, 2012, 13, 3409–3417.
21 S. Chapelle and J.-F. Verchere, Tetrahedron, 1988, 44, 4469–
4482.
22 R. van den Berg, J. A. Peters and H. van Bekkum, Carbohydr.
Res., 1994, 253, 1–12.
H.-J. Gabius, Bioconjugate Chem., 2004, 15, 87–98.
39 Y. Liu, C. Yu, H. Jin, B. Jiang, X. Zhu, Y. Zhou, Z. Lu and
D. Yan, J. Am. Chem. Soc., 2013, 135, 4765–4770.
40 Q. Zhang, L. Su, J. Collins, G. Chen, R. Wallis, D. A. Mitchell,
D. M. Haddleton and C. R. Becer, J. Am. Chem. Soc., 2014,
136, 4325–4332.
41 X. Shi, G. Chen, L. Yuan, Z. Tang, W. Liu, Q. Zhang,
D. M. Haddleton and H. Chen, Mater. Horiz., 2014, 1, 540–
545.
42 Y. Qu, T. Wei, W. Zhan, C. Hu, L. Cao, Q. Yu and H. Chen,
J. Mater. Chem. B, 2017, 5, 444–453.
43 W. Zhan, T. Wei, L. Cao, C. Hu, Y. Qu, Q. Yu and H. Chen,
ACS Appl. Mater. Interfaces, 2017, 9, 3505–3513.
44 H. Xiang, G. Jinbo, L. Yan, W. Ting, D. Yishi, C. Gaojian and
C. Hong, Macromol. Rapid Commun., 2017, 38, 1700434.
45 C. Hu, J. Wu, T. Wei, W. Zhan, Y. Qu, Y. Pan, Q. Yu and
H. Chen, J. Mater. Chem. B, 2018, 6, 2198–2203.
46 W. Zhan, Y. Qu, T. Wei, C. Hu, Y. Pan, Q. Yu and H. Chen,
ACS Appl. Mater. Interfaces, 2018, 10, 10647–10655.
47 K. Chmurski and J. Defaye, Supramol. Chem., 2000, 12, 221–
224.
23 R. Ma and L. Shi, Polym. Chem., 2014, 5, 1503–1518.
24 Z. L. L. He, X. E. Yang and P. J. Stoffella, J. Trace Elem. Med. 48 B. Roy and B. Mukhopadhyay, Tetrahedron Lett., 2007, 48,
Biol., 2005, 19, 125–140. 3783–3787.
25 T. Sutton, U. Baumann, J. Hayes, N. C. Collins, B. J. Shi, 49 Q. Zhang, S. Slavin, M. W. Jones, A. J. Haddleton and
T. Schnurbusch, A. Hay, G. Mayo, M. Pallotta, M. Tester
and P. Langridge, Science, 2007, 318, 1446–1449.
D. M. Haddleton, Polym. Chem., 2012, 3, 1016–1023.
¨
50 M. YurdakoÇ, M. Karakaplan and H. Ho¸SGORen, Sep. Sci.
26 A. Alsbaiee, B. J. Smith, L. Xiao, Y. Ling, D. E. Helbling and
W. R. Dichtel, Nature, 2015, 529, 190.
27 F. van de Manakker, T. Vermonden, C. F. van Nostrum and
W. E. Hennink, Biomacromolecules, 2009, 10, 3157–3175.
28 G. Crini, Dyes Pigm., 2008, 77, 415–426.
Technol., 1999, 34, 2615–2625.
51 L. D. Wilson, M. H. Mohamed and J. V. Headley, J. Colloid
Interface Sci., 2011, 357, 215–222.
52 H. Tamai, M. Kouzu and H. Yasuda, Carbon, 2003, 41, 1678–
1681.
29 A. Z. M. Badruddoza, Z. B. Z. Shawon, W. J. D. Tay, K. Hidajat 53 P. Lo Meo, G. Lazzara, L. Liotta, S. Riela and R. Noto, Polym.
and M. S. Uddin, Carbohydr. Polym., 2013, 91, 322–332. Chem., 2014, 5, 4499–4510.
30 L. Xiao, Y. Ling, A. Alsbaiee, C. Li, D. E. Helbling and 54 J. S. Basuki, L. Esser, H. T. T. Duong, Q. Zhang, P. Wilson,
W. R. Dichtel, J. Am. Chem. Soc., 2017, 139, 7689–7692.
M. R. Whittaker, D. M. Haddleton, C. Boyer and
T. P. Davis, Chem. Sci., 2014, 5, 715–726.
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A, 2018, 6, 21193–21206 | 21205

View Article Online
Journal of Materials Chemistry A
Paper
55 C. Jie, F. Zhi-jie, W. Guang-hui, J. Yu-hua and Z. Bao-hui, 61 C. Sanfeliu, R. Martinez-Manez, F. Sancenon, J. Soto,
Chin. J. Synth. Chem., 2009, 17, 603–605.
56 F. Cheng and F. Jakle, Polym. Chem., 2011, 2, 2122–2132.
V. Puchol, P. Amoros and M. D. Marcos, J. Mater. Chem.,
2012, 22, 25362–25372.
¨¨
57 A. P. Vogt, V. Trouillet, A. M. Greiner, M. Kaupp, U. Geckle, 62 F. Zhao, E. Repo, D. Yin, Y. Meng, S. Jafari and M. Sillanpaa,
L. Barner, T. Hofe and C. Barner-Kowollik, Macromol.
Rapid Commun., 2012, 33, 1108–1113.
58 J. Rohovec, T. Maschmeyer, S. Aime and J. A. Peters, Chem.–
Eur. J., 2003, 9, 2193–2199.
Environ. Sci. Technol., 2015, 49, 10570–10580.
63 R. Zhao, Y. Wang, X. Li, B. Sun and C. Wang, ACS Appl. Mater.
Interfaces, 2015, 7, 26649–26657.
64 J. Liu, G. Liu and W. Liu, Chem. Eng. J., 2014, 257, 299–308.
ˇ
59 R. Boncukcuoglu, A. Erdem Y11maz, M. Muhtar Kocakerim 65 X. Zhang, H. Li, M. Cao, L. Shi and C. Chen, Sep. Sci.
and M. Çopur, Desalination, 2004, 160, 159–166.
Technol., 2015, 50, 947–957.
¨
¨
¨
60 N. Ozturk and T. E. Kose, Desalination, 2008, 227, 233–240.
66 A. M. M. Vargas, A. L. Cazetta, M. H. Kunita, T. L. Silva and
V. C. Almeida, Chem. Eng. J., 2011, 168, 722–730.
21206 | J. Mater. Chem. A, 2018, 6, 21193–21206
This journal is © The Royal Society of Chemistry 2018