Pore-size control of chitin nanofibrous composite membrane using metal-organic frameworks

Carbohydrate Polymers 275 (2022) 118754

Contents lists available at ScienceDirect

Carbohydrate Polymers
journal homepage: www.elsevier.com/locate/carbpol

Research paper

Pore-size control of chitin nanofibrous composite membrane using
metal-organic frameworks
Younghan Song a, b, 1, Jin Young Seo a, c, 1, Hyungsup Kim b, Sangho Cho a, d, *,
Kyung-Youl Baek a, d, e, *
a

Materials Architecting Research Center, Korea Institute of Science Technology, Seoul 02792, Republic of Korea
Department of Organic and Nano System Engineering, Konkuk University, Seoul 05029, Republic of Korea
Department of Chemical and Biological Engineering, Korea University, 5-1 Anam-dong, Seongbuk-gu, Seoul 136-713, Republic of Korea
d
Division of Nano & Information Technology, KIST School, Korea University of Science and Technology, Seoul 02792, Republic of Korea
e
KHU-KIST Department of Converging Science and Technology, Kyung Hee University, Seoul 02447, Republic of Korea
b
c

A R T I C L E I N F O

A B S T R A C T

Keywords:
Chitin nanofiber (ChNF)

Pore size control
Bioderived membrane
Water purification
Rheology

Herein, environmentally benign chitin nanofiber (ChNF) membranes were fabricated by regulating suspension
behavior. The introduction of zeolitic imidazole frameworks (ZIF-8) into the composite membranes led to the
domain formation of ChNF derived by coordinative interaction, resulting in pore size-tunable membranes. Based
on the rheological, morphological, and structural characterizations, the driving force of pore-size control was
studied in the aqueous suspension of ChNF and ZIF-8 according to the relative concentration. At critical con­
centration, the 30-ChNF membrane presents superior water permeance (40 LMH h− 1) while maintaining a high
rejection rate (>80% for all organic dyes). Moreover, the molecular size cut-off of the composite membranes for
dyes can be controlled in the range of less than 1 nm to 2 nm. The experimental results provide a simple strategy
for the preparation of pore tunable ChNF membranes using MOF with high mechanical strength, good durability,
high flux, dye rejection, and antifouling ability.

1. Introduction
In general, nanofibers are defined as fibers with diameters less than
100 nm with a high aspect ratio of more than 100 (Xia et al., 2003; Li &
Xia, 2004). Material-wisely, nanofibers are interesting nanomaterials
due to their unique features from dimensional, optical, and mechanical
characteristics (Lee, An, Kim, Yoon, & Yarin, 2018; Venkatesan et al.,
2020). Therefore, nanofibers have been gained great attention for a wide
range of application areas such as nanocomposites, filtrations, energy
storage, catalyst, and biomedical fields (Cui et al., 2020; Maity &
Mandal, 2018; Seo, Cho, Lee, Lee, & Baek, 2020;). Most of all, the
nanofibrous structure has been widely utilized in membrane technology
for purifying wastewater as an energy-saving process (Subrahmanya
et al., 2021). Among various pollutants in wastewater, organic dyes have
raised great concerns due to their enormous usage, and lethality. Not

only are the dyes toxic and carcinogenic, but even small amounts can
inhibit sunlight transmission, leading to metabolic imbalances (Moradi
& Sharma, 2021; Rajabi, Mahanpoor, & Moradi, 2017; Robati et al.,

2016; Zhu, Chen, & Lou, 2012). As an effective way to remove dyes, the
nanofibrous membrane presents high performances due to the threedimensional inter-connected pore structure (Wu et al., 2015). This
nanofibrous membrane was produced from a polymer solution by
electrospinning and could be controlled by their diameters or mor­
phologies with rheological tuning, voltage, and so on (Chronakis, 2005).
However, the electrospinning process requires a large amount of organic
solvents with low productivity, causing environmental pollution.
With increasing environmental concerns, nanofibers from the hier­
archical structures of biomass have been greatly gained attention by
their environmental friendliness. In nature, many types of biomasses
have nanofibers in their complex hierarchical organization such as cel­
lulose, chitin, etc. (Chen et al., 2010; Xu, Mao, Peng, Luo, & Chang,
2018). These nanofibers can be extracted in physical or chemical ways
(Isogai, Saito, & Fukuzumi, 2011; Rossi et al., 2021). This has triggered
many researchers to explore the application of membrane technology
using biomass-driven nanofibers. The hydrophilicity of biomass-driven
nanofibers can give an advantage for improving the membrane

* Corresponding authors at: Materials Architecting Research Center, Korea Institute of Science Technology, Seoul 02792, Republic of Korea.
E-mail addresses: (S. Cho), (K.-Y. Baek).
1
These authors contributed equally.
/>Received 17 August 2021; Received in revised form 23 September 2021; Accepted 6 October 2021
Available online 16 October 2021
0144-8617/© 2021 The Authors.
Published by Elsevier Ltd.

This is an open
( />
access

article

under

the

CC

BY-NC-ND

license


Y. Song et al.

Carbohydrate Polymers 275 (2022) 118754

properties of anti-fouling ability and high wettability (Ma et al., 2014;
Lotfikatouli et al., 2021). Also, the biomass-driven nanofibrous mem­
brane can be prepared by water with the wet-laid process, so-called
nanopaper fabrication (Yoon, Doh, & Im, 2011). During the wet-laid
process, the pore structure of the nanofibrous membrane can be
blocked by a random distribution of the fibers, resulting in a highly
dense nanopaper (Fukuzumi, Saito, Iwata, Kumamoto, & Isogai, 2009;
Kim, Choi, & Jin, 2020). To avoid blocking the pore, the biomass-driven
nanofibrous membrane was prepared at the thin thickness with the

support membrane. This thin layer membrane with nanofibers showed
high water permanence with a high rejection rate (Zhang, Deng,
Soyekwo, Liu, & Zhu, 2016). Still, there are challenging issues on the
durability and pore-structure control of the membrane.
The pore-size modulation of the nanofibril-based membrane from
biomass is highly required to broaden its applications: for example,
hemofiltration of blood for medical care, purification of industrial
wastewater, and seawater desalination (Shin et al., 2019; Soyekwo et al.,
2017). So far, three different approaches were conducted for pore
modulation. First, the fabrication of nanofibrous membranes with
smaller diameters induces dense packing of nanofiber, resulting in pore
size decrement (Zhao, Zhou, & Yue, 2000). Second, the membrane
morphology with a thinner active layer can induce opening interconnected pore structure, leads to increment of pore size (Chen, Duan,
Shi, & Fang, 2017). For instance, Buehler and coworkers (Ling, Jin,
Kaplan, & Buehler, 2016) demonstrated the control of membrane pore
size by varying active layer thickness (<500 nm) using silk nanofiber
(SNF). However, these approaches were limited to ultra-thin active layer
thickness (<500 nm). Also, the membranes require support layers and
undergo compaction against the hydraulic pressure, resulting in the flux
drop due to pore blocking in the active layer (Stade, Kallioinen, Mikkola,
ănttă
Tuuva, & Ma
ari, 2013). Lastly, the insertion of nano-dimensional
additives could tune the membrane pore size via the regulation of
packing density of nanofibers even with a thick layer (Ling et al., 2017;
Sun et al., 2021). However, the pore-modulation of the membrane using
nanofiber interaction to directly prepare the membrane from the
aqueous dispersion has been less explored.
Recently, the metal-organic frameworks (MOF) have been emerging
nanoparticles for carbohydrate-based nanofibers as a provider of phys­

ical and chemical functions. Their composite showed extraordinary
features in the catalytic degradation, absorbent, and water purifications
due to high surface area, high chemical activity, and physical tolerance
(Abdelhameed, El-Shahat, & Emam, 2020; Abdelhameed & Emam,
2019; Emam, El-Shahat, & Abdelhameed, 2021; Khajeh, Oveisi, Bar­
khordar, Rakhshanipour, & Sargazi-Avval, 2021; Liu et al., 2021; Liu
et al., 2021). However, there was a lack of study on the MOF role for the
interaction changer of pore-size tuning in carbohydrate-based mem­
brane technologies. In our group, we demonstrated that the zeolitic
imidazolate frameworks (ZIF-8) could induce the porous morphologies
of cellulose nanofibers for the water treatment membrane with selective
dye rejection (Song, Seo, Kim, & Beak, 2019). Despite superior perfor­
mance for purifying wastewater, the precise control of pore size could
not be achieved using the in-situ growing ZIF-8 at the nanofibers. The
MOF grown at the fibers can deform the nanofiber shape due to the
strong adhesion rather than changing the fiber-to-fiber interaction.
Moreover, the total amounts, particle sizes, and defect sites of MOF can
be limited by the in-situ synthesis condition of MOF in the matrix.
Therefore, a novel approach is required for the simple utilization of
pore-size controllable membranes using carbohydrate-based nanofibers.
In this study, we focused on the pore-size modulation of the bioderived nanofibrous membrane in the nanoscale. It was assumed that
the regulation of fiber-to-fiber interaction can induce the formation of
domain structures, resulting in the tunable pore size of the membrane.
Among bio-derived nanofibers, the chitin nanofiber (ChNF) was selected
as a platform material due to its superior hydrophilicity, anti-bacterial
and anti-fouling ability (Li et al., 2020; Mousavi, Asghari, & Mah­
moodi, 2020). As a nanofiber interaction changer, the ZIF-8 with a

surface positive charge and its strong coordinative interaction with
ChNF was utilized. The formation of a micro-domain structure with

respect to ChNF and ZIF-8 contents was characterized by rheological,
mechanical, and morphological studies. It led to the successful modu­
lation of pore size by a wet-laid process. More importantly, the mem­
brane performance of ChNF/ZIF-8 composites was evaluated by the
separation of organic dyes. At the boundary concentration of ZIF-8, the
abrupt change of microstructure and membrane performance were
observed, which allows tuning removal performance for the organic
dyes according to the molecular size. This experimental result can give a
fundamental idea of porous structure control using a wet-laid process for
broadening the membrane application of biomass-driven nanofibers.
2. Experimental section
2.1. Materials
Chitin nanofiber (ChNF) suspension (0.5 wt% in water, 3 μm of mean
fiber length, 5 nm of mean diameter) was purchased from ANPOLY
(Republic of Korea). The ChNF suspension was titrated to pH 7.0 with 1
N NaOH(aq) before use. Zinc nitrate hexahydrate (Zn(NO3)2∙6H2O,
98%), 2-methylimidazole (99%), bovine serum albumin (BSA), orange G
(OG, 80%), Janus Green B (JGB, 65%), bromothymol blue (BB, 95%),
vitamin B12 (VB12, 98%), methyl orange (MO, 85%), and methylene
blue (MB, 82%) were purchased from Sigma Aldrich. All chemicals were
used as received. Deionized water (Milli-Q Millipore, 18.2 MΩ cm− 1)
was used in all experiments.
2.2. Hydrothermal synthesis of ZIF-8
ZIF-8 was synthesized according to the previous study (Pan, Liu,
Zeng, Zhao, & Lai, 2011). Briefly, 0.41 g of Zn(NO3)2∙6H2O (0.0013
mol), and 7.92 g of 2-methyl imidazole (0.096 mol) were separately
dissolved in 15 mL of water. After 10 min of sonication, the two solu­
tions were mixed in a 50 mL autoclave reactor and were kept at 70 ◦ C for
24 h. The crude product was washed by centrifugation against water and
dried in a vacuum oven at 30 ◦ C for 24 h.

2.3. Preparation of ChNF/ZIF-8 composite membranes
The fabrication of ChNF/ZIF-8 composite membrane is illustrated in
Fig. 1a. ZIF-8 was blended with ChNF suspension for the concentrations
of ZIF-8 within the resulting composite membranes to be 10, 20, 30, 40,
and 50 wt% (Table S1). The composite dispersions were filtered upon
quantitative filter papers (No. 6, Advantec, Japan) using a dead-end
filtration apparatus (Amicon 8050, Millipore, USA) under 1 bar with a
permeation area of 13.4 cm2. The loading amounts of membranes were
kept to 30 gsm (g⋅m− 2) to the ChNF weight. The sample was named after
the input amount of ZIF-8 in the suspension; for example, 10-ChNF re­
fers to the ChNF composite membrane with the ZIF-8 input fraction of
10 wt%.
2.4. Characterizations
The contents of ZIF-8 in the composite membranes were character­
ized by thermogravimetric analysis (TGA, TA Instrument TGA Q-50)
under N2. X-ray diffraction (XRD, D8 ADVANCE (Sol-X)) was used to
characterize the crystal structure of the samples in the scanning degree
from 5o to 45o of 2θ at 40 kV and 40 mA using Cu Kα radiation (λ = 1.54
Å). The chemical structure of the specimens was investigated with
Fourier-transform infrared spectroscopy (FT-IR, Thermo Nicolet iS10
equipped with attenuated total reflectance (ATR)). The rheological
properties of the ChNF/ZIF-8 aqueous dispersion were measured by a
rotational rheometer (RS-1, HAAKE) using a parallel disk type in steady
mode. The surface zeta-potential and the particle size of the composite
suspensions (0.001 wt% in H2O) were measured at pH 7.0 (ELSZ-2000,
2


Y. Song et al.

Carbohydrate Polymers 275 (2022) 118754

Fig. 1. Schematic illustration for (a) a wet-laid process for membrane fabrication and (b) suggested pore-forming mechanism with ChNF and ZIF-8.

Otsuka electronics). The N2 adsorption behavior was confirmed by
BELSORP-Mini II (MicrotracBEL, Japan) under 77 K. The morphologies
of the composite membranes were observed by field emission scanning
electron microscopy (FE-SEM, FEI Inspect F50). To prohibit the struc­
ture collapse, the composites were freeze-dried for N2 adsorption and
SEM measurements. The nano-indentation (Bruker, TI-950 with Berko­
vich Tip) was used to evaluate the mechanical properties of the mem­
branes. The rejection rate of the organic dyes in the filtered water was
calculated from the absorption intensities using UV–Vis spectroscopy
(JACSO V-670).

3. Results and discussion
3.1. Microdomains formation in composite suspension
Fig. 1 describes the pore-forming process during the membrane
preparation of ChNF composite. In the wet-laid process (Fig. 1a), the
local concentration of the composite suspension became higher in sus­
pension (Liu, Lin, et al., 2021; Liu, Liu, et al., 2021). The membrane
structure of ChNF can be expected as a densely packed membrane with
low water permeance owing to the strong hydrogen bond between
ChNFs (Fig. 1b, top). The flow behavior of ChNF suspension with a
suitable additive could be changed by different fiber-to-fiber in­
teractions, resulting in the formation of a specific domain structure and
pore size modulation in the membrane (Fig. 1b, bottom). In this manner,
we utilized ZIF-8 as domain former for ChNF membrane with its dual
action of attractive interaction by metal-amine interaction and repulsive
interaction by same surface charge with ChNFs (Interaction mechanism

between ZIF-8 and ChNF was illustrated in Scheme S1).
A rheological study was performed to understand the interaction
between ChNF and ZIF-8 in the suspension. Fig. 2a shows the viscosities
versus shear rate for ChNF and ChNF/ZIF-8 composite aqueous disper­
sions. For the ChNF dispersion, two-regime flow behavior was observed.
The viscosity of the ChNF dispersion gradually decreased below the
shear rate of 0.1 s− 1 and monotonically decreased at a constant slope

2.5. Membrane performance
The water permeability of the membranes was measured using a
dead-end filtration (Amicon 8050, Millipore, USA). The effective area
(A) of the membranes was 13.4 cm2, and applied pressure (P) was 1 bar
at 25 ◦ C. The water permeability was calculated by measuring amounts
of permeate (ΔV) with the interval time (Δt).
(
)
Pw LMH∙bar− 1 = L∙m− 2 ∙h− 1 ∙bar− 1 = ΔV∙A− 1 ∙Δt− 1 ∙P− 1
The organic dye solution was prepared as 10 ppm, and the rejection
rate (R) for the organic dye was calculated by the absorbance of feed
(Asf) and permeate (Asp) using UV–Vis spectroscopy.
(
)
R = 1 − Asp ∙Asf − 1 ∙100 (%)

Fig. 2. (a) Viscosity-shear rate curves of ChNF and ChNF/ZIF-8 aqueous dispersion. The inserted images are 30-, 40-ChNF dispersion after leaning. (b) Schematic
illustration for the flow behavior of ChNF according to ZIF-8 loading amounts under shear rate.
3


Y. Song et al.

Carbohydrate Polymers 275 (2022) 118754

(− 0.8 Pa∙s2) with an increase in the shear rate. This shear-thinning
behavior was originated from the change of frictional force in the
deformation of nanofibers. At the low shear rate, the strong inter- and
intra-particular interactions between and within the entangled nano­
fibers inhibit sudden viscosity drop. When shear stress is applied enough
to break the inter- and intra-particular interaction, the nanofibers un­
tangle. Above this point, the only friction between the nanofibers
affected the viscosity, resulting in the viscosity drop at a constant rate
(Fig. 2b, top).
Otherwise, ChNF/ZIF-8 dispersion exhibited similar shear thinning
behavior to ChNF dispersion but in three-regime (Fig. 2a, left). For
ChNF/ZIF-8 dispersion with the ZIF-8 contents up to 30 wt%, the vis­
cosities were similar to that of the ChNF dispersion below the shear rate
of 0.1 s− 1. In such a low shear rate, ZIF-8 not only disperses well but also
does not destroy the entanglement of nanofibers. Above the shear rate
enough for the disentanglement of nanofibers, the viscosities of the
composite dispersion more rapidly decreased with the increase of shear
rate (the slope = − 1.1 Pa∙s2) compared to that of the ChNF dispersion.
Then, the viscosities dropped with a similar rate (− 0.8 Pa∙s2) of the
ChNF dispersion above the shear rate of about 10 s− 1. This three-regime
behavior was a similar phenomenon with the anisotropic dispersion of
nano-rods based on the Onogi and Asada model (Onogi & Asada, 1980;
Li, Revol, & Marchessault, 1996; Wissbrun, 1981). The strong metalamine interaction between ChNF and ZIF-8 could lead to the forma­
tion of micro-domains between ChNF and ZIF-8 with the morphology of
nanofiber-wrapped ZIF-8 (ZIF-8@ChNF) (Fig. 2b, middle). The domain
could be confirmed in the result of a dynamic light scattering (DLS)
analysis, exhibiting additional particles with a size around 100–200 nm

(Fig. S1). The reduced number of free nanofibers gave rise to the addi­
tional viscosity drop with a shear rate between 10− 1 to 10 s− 1.
Furthermore, the shear rate over 10 s− 1 destroys the micro-domain, and
the friction of the nanofibers increases, which leads to the decrease of
the viscosity drop rate similar to that of the ChNF dispersion.
For ChNF/ZIF-8 dispersion with the ZIF-8 contents over 40 wt%, the
initial viscosities were higher than that of the ChNF dispersion (Fig. 2a,
right). The excessively loaded ZIF-8 would be positioned between the
domains and increase the inter-domain interaction, resulting in a gellike structure, as shown in the inserted digital images (Fig. 2a, inset).
In the DLS analysis, the presence of free ZIF-8 was also observed in the
dilute dispersion of ChNF/ZIF-8 with ZIF-8 contents over 40 wt%
(Fig. S1). As the shear rate increased, the viscosity descended with the
rate steeper (− 1.3 Pa∙s2) than that for ChNF/ZIF-8 dispersion with ZIF-8
contents of 0 to 30 wt%. It can be interpreted that the ZIF-8 between the
domains acted as a plasticizer under the shear rate (Fig. 2b, bottom).
Under the shear rate over 10 s− 1, the viscosity drop rate reduced to − 0.6
Pa∙s2, lower than those of the other dispersions. In such a high shear

rate, the excessive free ZIF-8 increased the total number of nanoparticles
and nanofibers in the system, providing additional friction. The flow
behaviors of the dispersion according to the ZIF-8 loading amounts can
explain a major driving force for the pore formation in the composite
membranes. In other words, we can modulate the pore between the
nanofibers in the membrane by simple tuning of the composite ratio.
This will be further discussed in the following morphological study.
3.2. Fabrication and microstructure control of composite membranes
The amounts of ZIF-8 within the composite membranes were
analyzed by TGA (Fig. 3a). ZIF-8 had weight losses at 345 ◦ C and 500 ◦ C,
which are attributed to the decomposition of the organic ligand of 2methylimidazole and the cluster structure of ZIF-8, respectively (Cho
et al., 2018). The pristine ChNF membrane showed three steps of weight

loss at 100, 150, and 250 ◦ C, which were due to residue water, the
degradation of the acetyl groups, and the backbone of chitin, respec­
tively (Zhong, Wolcott, Liu, Glandon, & Wang, 2020). The calculated
ZIF-8 contents for each sample were 8.2, 19.8, 30.7, 38.9, and 53.0 wt%
for 10-, 20-, 30-, 40-, and 50-ChNF, respectively (Table S1). The contents
of ZIF-8 in the composite were not different from the input amounts. It
refers that each structure of ZIF-8 and ChNF did not collapse while the
ChNF holds physically the ZIF-8.
Fig. 3b shows the XRD patterns of the ChNF/ZIF-8 composites. The
prepared ZIF-8 had sharp peaks at 2θ = 7.3, 10.4, 12.7, 14.7, 16.4, and
18◦ (Cho et al., 2018). The ChNF exhibited the typical crystalline
structure of chitin with peaks at 9.1, 19.1, and 23.1◦ (Joseph et al.,
2020). In the composite membranes, ZIF-8 and ChNF maintained their
crystalline structures as both characteristic patterns of ZIF-8 and ChNF.
The peaks associated with ZIF-8 in the composite membranes became
more intense with the higher content of ZIF-8. As coincident with the
XRD result, the characteristic peaks of ChNF and ZIF-8 were retained in
the composite membranes in FT-IR spectra, as shown in Fig. 3c. In the
composite membranes, the characteristic vibration peaks for both ChNF
and ZIF-8 were observed. The ChNF showed ν(O–H), ν(N–H), ν(C=O),
ν(C–N), and ν(C-O-C (1,4-glycosidic bond)) at 3430, 3260, 1655, 1309,
and 1008 cm− 1, respectively. The characteristic peaks of ν(C–H),
ν(C–N, aromatic), ν(Zn–N) at 1417, 1309, 758, 690 cm− 1 corre­
sponded to ZIF-8 (Cho et al., 2018; Tsai, Wang, Chang, & Tsai, 2019).
The XRD and FT-IR results imply that the chemical structure of ChNF
and ZIF-8 did not change even after blending. As a result, both ChNF and
ZIF-8 have excellent stability in the physical and chemical states in the
composite membranes.
The surface morphologies of the prepared membranes were charac­
terized by SEM (Fig. 4). The ChNF membrane without ZIF-8 had a flat

surface with randomly distributed nanofibers. In the composite

Fig. 3. (a) TGA, (b) XRD, and (c) FT-IR spectra of ChNF, ZIF-8, and the composite membranes.
4


Y. Song et al.

Carbohydrate Polymers 275 (2022) 118754

Fig. 4. SEM images of ChNF and ChNF/ZIF-8 composite membranes: (a) 0-ChNF, (b) 10-ChNF, (c) 20-ChNF, (d) 30-ChNF, (e) 40-ChNF, and (f) 50-ChNF (Inset
present SEM images with higher magnification).

membranes, the surface became rougher with the higher amount of ZIF8. The surface of the composite membrane with lower 30 wt% ZIF-8
showed relatively smoother and no noticeable free ZIF-8 nanoparticles
due to the formation of nanofiber-wrapped ZIF-8 (ZIF-8@ChNF) as
discussed above. However, the surface turned coarse distinctively at
above 40 wt% of ZIF-8 loading. These results were supported by DLS and
surface zeta-potential analyses (Figs. S1 and S2). For the composite
dispersions with ZIF-8 loading from 10 to 30 wt%, the surface charges
were steady at 11 to 13 mV, which was close to that of the ChNF
dispersion. On the other hand, the surface charges abruptly increased in
the composite dispersion for 40-ChNF and 50-ChNF due to the free ZIF-8
on the surface. After the filtration process, some free ZIF-8 nanoparticles
laid on the surface of the resulting membranes, leading to a sandpaperlike morphology.
Then, the thickness of the composite membranes using SEM analysis
was measured to study the volume of the composite membranes related
to their microstructure (Fig. 5). The 0-ChNF membrane exhibited a
smooth surface line with a densely packed cross-section. The addition of
ZIF-8 in the composite dispersion led to the thickness expansion of the

membranes. The thickness of the composite membranes less than 30 wt

% of ZIF-8 was well-matched with the calculated values based on the
relative volume expansion per the amount of ZIF-8 added (Arbulu,
Jiang, Peterson, & Qin, 2018). It reveals that the composite membrane
up to 30 wt% loadings of ZIF-8 did not possess any additional pores
except the pores within the ChNF/ZIF-8 domain. Notably, the 40-ChNF
and 50-ChNF membranes were much thicker than the calculated value.
This result indicates the existence of additional pores. Overall, ZIF-8
played a critical role as a binder and a spacer of the nanofibers. Below
30 wt% loading, most of the ZIF-8 was wrapped by ChNF (ZIF-8@ChNF),
as seen in DLS analysis (Fig. S1). The composite membranes exhibited
densely packed structures by the interaction between ZIF-8@ChNF
domain and ChNF as like the ChNF membrane itself. Above 40 wt%
loading of ZIF-8, the additional free ZIF-8 nanoparticles widened the gap
among ChNFs due to the repulsive interaction of the same positive
surface charges for both free ZIF-8 and ChNF, which introduce pores into
the membranes.
The surface area and pore size of the composite membranes were
measured by BET analyses (Fig. 6). The N2 adsorption by the ZIF-8
particles showed a typical type I adsorption by micro-porosity. On the
other hand, the pristine ChNF membrane showed negligible N2

Fig. 5. Cross-sectional SEM images of the prepared membranes: (a) 0-, (b) 10-, (c) 20-, (d) 30-, (e) 40-, and (f) 50-ChNF. (g) Plot of the membrane thickness vs ZIF8 contents.
5


Y. Song et al.

Carbohydrate Polymers 275 (2022) 118754

Fig. 6. (a) N2 adsorption isotherm at 77 K and (b) pore size distribution of ZIF-8, ChNF membrane, and ZIF-8/ChNF composite membranes.

0-ChNF had steady permeance of 6 LMH⋅bar− 1 for 24 h. The permeance
of the ChNF/ZIF-8 membranes was gradually improved as the amounts
of ZIF-8 increased, which is well correlated with the increase of porosity
according to the ZIF-8 contents. The membranes were completely
compressed for 12 h against the hydraulic pressure. The permeance drop
was evaluated to the drop rate (DR = (P0 − P24h) / P0⋅100%) calculated
by comparison of the permeance at initial (P0) and 24 h (P24h) for each
membrane. The DR of the 0-, 10-, 20-, 30-, 40- and 50-ChNF were 17.8,
32.7, 31.5, 25.6, 30.2 and 46.2%, respectively. Among the membranes,
30-ChNF had the lowest value of the drop rate. A well-matched ratio for
the strong interaction between the ChNF and the ZIF-8 could attribute to
the high durability of 30-ChNF. The filtration performance of the
membranes was examed for organic dyes after compaction. The com­
posite membranes were used for the dye rejection of JGB according to
their ZIF-8 contents (Fig. 7b). The membranes of 0-, 10-, 20-, and 30ChNF exhibited a rejection rate of over 99%. For the 40-, and 50ChNF membranes, the rejection rates were decreased. This abrupt
drop in rejection rate was related to the pore size of the membranes. To
evaluate the rejection mechanism, the organic dyes were tested with
various surface charges and different molecule sizes (Table S2). The
chemical structure of used dyes displays in Fig. S4. Fig. 7c displays the
rejection rate of 30-ChNF against the molecular size of organic dyes. For
all the composite membranes, the rejection rate versus the molecular size
of dye shows in Fig. S5. The rejection rate of 30-ChNF only depended on
the molecular size of the dyes. In detail, the dyes (JG, VB, and BBR) with
molecular sizes over 1.4 nm were filtered off with a rejection rate over
98% even with a different charge. The rejection rate of 30-ChNF started
to decrease with the molecular size close to 1.3 nm (MB and MO). A
noticeable drop in rejection rate was observed for dyes smaller than 1.2

nm. The molecular size cut-off of 1.3 nm for the rejection rate is close to

adsorption compared to the ZIF-8. The N2 adsorption behavior of the
composite membranes was dependent on the amount of ZIF-8 loaded.
The surface areas of the composite membranes were calculated by the
BET method. The surface area shows good linearity with ZIF-8 contents,
indicating the successful introduction of ZIF-8 into a composite (Fig. S3).
Pores in the composite membranes could be formed either by ZIF-8 itself
or by the interaction between ChNF and ZIF-8. Fig. 6b shows the pore
size distribution of the prepared membranes using the MP method. The
ZIF-8 had 7 Å sized pores, which is in good agreement with the previous
report (Cho et al., 2018). Interestingly, 10-, 20-, and 30-ChNF had two
sized pores of 0.7, and 1.3 nm, while 40- and 50-ChNF had three sized
pores of 0.7, 1.3, and 1.8 nm. In addition, the more ZIF-8 loaded within
the membranes (40-, 50-ChNF), the higher the fraction of 1.3 and 1.8 nm
pores in all composites. We assume that the 1.3 nm-pores are derived in
the domain region between ChNF and ZIF-8, that is, ZIF-8@ChNF, where
ZIF-8 effectively pulls the nanofibers as a binder. However, the 1.8 nmpores appear only in 40- and 50-ChNF. The excess amounts of ZIF-8 can
be positioned between the domains, resulted in large-sized pores by
repulsive interaction. Table 1 summarizes the total pore volume (Vp)
and pore volume distribution with specific diameters in the composite
membranes. This trend according to the ZIF-8 contents is coincident
with the rheological and morphological studies. These results suggest
that we can modulate the pore structure by controlling ZIF-8 contents.
Furthermore, the homogeneous tuning of the pore structure could
enable a superior separation for the composite membranes.
3.3. Membrane performance
We investigated the water treatment performance of the membranes
with the loading amounts of 30 gsm with respect to ChNF (Fig. 7a). The

Table 1
The summarized data for the pore structure of ChNF and ZIF-8 composite membranes.
Sample name

ZIF-8 contents
(wt%)

Measured thicknessa
(μm)

Calculated thicknessb
(μm)

Total pore volume (Vp)c
(cm3)

Pore volumec (cm3)
0.7 nm pores

1.3 nm pores

1.8 nm pores

0-ChNF
10-ChNF
20-ChNF
30-ChNF
40-ChNF
50-ChNF

0
8.2
19.8
30.7
38.9
53.0

15.6 ± 1.9
16.8 ± 1.7
18.2 ± 1.7
20.3 ± 2.2
27.2 ± 3.0
33.7 ± 3.8

15.6
16.8
18.3
20.3
22.9
26.5

0.0011
0.0017
0.0020
0.0027
0.0096
0.0153

–
0.0006

0.0008
0.0012
0.0023
0.0039

–
0.0011
0.0012
0.0015
0.0023
0.0044

–
–
–
–
0.0050
0.0070

a
b
c

The thickness of the membrane was measured by SEM image.
The thickness of the membrane was calculated based on the relative volume expansion per the amount of ZIF-8 added.
Total pore volume was calculated by N2 adsorption isotherm.
6


Y. Song et al.

Carbohydrate Polymers 275 (2022) 118754

Fig. 7. (a) Pure water flux vs time curves of the prepared membranes. (b) Permeance and rejection rate of the prepared membranes against JGB. (c) Rejection rate vs
molecular size of organic dyes for the composite membranes from the result in Table S2.

the pore size of the dried 30-ChNF analyzed by the N2 absorption
behavior. This result suggests that the porous structure of ChNF/ZIF-8
membranes can be precisely tuned with an inter-particle interaction in
the nanoscale for specific target molecules. Moreover, the JGB rejection
(99%) of 30-ChNF was stably maintained after long-term use of 8 h
(Fig. S6). This high stability of the membrane was characterized by the
FT-IR and XRD analysis (Fig. S7). The chemical and crystal structure did
not change before and after the use of dye rejection. Also, the 30-ChNF
showed good stabilities in the range of natural and basic conditions
(Fig. S8). The acidic condition of pH 4.0 led to the flux drop, which was
originated from the acid-induced crystal collapse of ZIF-8 (Cho et al.,
2018).
In water treatment applications, mechanical strength and hardness
are key factors for the long-term use of the membrane. The mechanical
properties of the prepared membranes were tested by the nanoindention (Figs. 8, S9). The reduced moduli of the composite mem­
branes were kept to that of 0-ChNF at ca. 5.5 GPa up to 30-ChNF. The
low porosity mainly formed within the domain of ChNF and ZIF-8 (ZIF8@ChNF) attributes to an equivalent modulus at these ranges. However,
the reduced modulus and hardness considerably decreased along with
the contents of ZIF-8 for 40- and 50-ChNF. The reduction would be
induced by the enlarged space among inter-nanofibers as described
above.
The anti-fouling characteristics of the prepared membranes were
investigated using BSA protein as a model of organic pollutants (Fig. 9).
The 0-ChNF had a declined flux of ca. 20% within 9 h by the protein.

Then, it recovered the flux to around 92% after the hydrodynamic
rinsing. As well-known, the hydrophilicity of the ChNF membrane can
suppress the attachment of BSA due to the water layer upon the mem­
brane. The fouling of 30-ChNF was slightly more than that of 0-ChNF. It

Fig. 9. Anti-fouling properties of 0-ChNF and 30-ChNF against BSA.

could be interpreted that the rougher surface and the higher surface area
accelerated adhesion of BSA onto the 30-ChNF surface. However, the
flux of 30-ChNF has successfully recovered to 90% after hydrodynamic
rinsing.
4. Conclusion
In this work, we successfully developed the pore size modulated
membranes by regulating of suspension behavior of ChNF using ZIF-8.
The composite ratio of ChNF and ZIF-8 affected the flow behavior of
ChNF/ZIF-8 suspension. At the low concentration of ZIF-8, the strong
amine-metal interaction formed ZIF-8@ChNF micro-domains. Above
the threshold amount of ZIF-8, then, the excess ZIF-8 acted as a spacer.
The ChNF/ZIF-8 ratio-dependent microstructure within the precursor
suspension allowed homogeneous pore size tuning, resulting in superior
water permeance (up to 60 LMH bar− 1) and the precise control of mo­
lecular cut-off of the membrane (in range of 1 nm to 2 nm). In addition to
the tunable pore size, the composite membranes had superior merits on
mechanical strength, long-term usability, filtration performance, and
anti-fouling. These results can provide the fundamental idea for highperformance nanofibrous membranes capable of controlling the pore
size using the particle-to-particle interaction.

Fig. 8. Reduced moduli and hardnesses of the composite membranes.
7

Y. Song et al.

Carbohydrate Polymers 275 (2022) 118754

CRediT authorship contribution statement

Li, P., Wang, B., Liu, Y.-Y., Xu, Y.-J., Jiang, Z.-M., Dong, C.-H., & Zhu, P. (2020). Fully
bio-based coating from chitosan and phytate for fire-safety and antibacterial cotton
fabrics. Carbohydrate Polymers, 237, Article 116173.
Ling, S., Jin, K., Kaplan, D. L., & Buehler, M. J. (2016). Ultrathin free-standing Bombyx
mori silk nanofibril membranes. Nano Letters, 16(6), 3795–3800.
Ling, S., Qin, Z., Huang, W., Cao, S., Kaplan, D. L., & Buehler, M. J. (2017). Design and
function of biomimetic multilayer water purification membranes. Science Advances, 3
(4), Article e1601939.
Liu, K., Liu, N., Ma, S., Cheng, P., Hu, W., Jia, X., & Wang, D. (2021). Highly permeable
polyamide nanofiltration membrane mediated by an upscalable wet-laid EVOH
nanofibrous scaffold. ACS Applied Materials & Interfaces, 13(19), 23142–23152.
Liu, Q., Lin, S., Khan, S., Zhao, Y., Guan, Q., Geng, Z., & Yang, X. (2021). Rapid removal
and mechanism investigation of low-concentration phosphate from wastewater by
CuFe2O4/MIL-101 (Fe) composite. Journal of Nanostructure in Chemistry, 1–11.
Lotfikatouli, S., Hadi, P., Yang, M., Walker, H. W., Hsiao, B. S., Gobler, C., & Mao, X.
(2021). Enhanced anti-fouling performance in membrane bioreactors using a novel
cellulose nanofiber-coated membrane. In (p. 119145). Separation and Purification
Technology.
Ma, H., Burger, C., Hsiao, B. S., & Chu, B. (2014). Fabrication and characterization of
cellulose nanofiber based thin-film nanofibrous composite membranes. J. Membr.
Sci., 454, 272–282.
Maity, K., & Mandal, D. (2018). All-organic high-performance piezoelectric
nanogenerator with multilayer assembled electrospun nanofiber mats for selfpowered multifunctional sensors. ACS Applied Materials & Interfaces, 10(21),

18257–18269.
Moradi, O., & Sharma, G. (2021). Emerging novel polymeric adsorbents for removing
dyes from wastewater: A comprehensive review and comparison with other
adsorbents. Environmental Research, 111534.
Mousavi, S. R., Asghari, M., & Mahmoodi, N. M. (2020). Chitosan-wrapped multiwalled
carbon nanotube as filler within PEBA thin film nanocomposite (TFN) membrane to
improve dye removal. Carbohydrate Polymers, 237, Article 116128.
Onogi, S., & Asada, T. (1980). Rheology and rheo-optics of polymer liquid crystals. In
Rheology (pp. 127–147). Springer.
Pan, Y., Liu, Y., Zeng, G., Zhao, L., & Lai, Z. (2011). Rapid synthesis of zeolitic
imidazolate framework-8 (ZIF-8) nanocrystals in an aqueous system. Chemical
Communications, 47(7), 2071–2073.
Rajabi, M., Mahanpoor, K., & Moradi, O. (2017). Removal of dye molecules from aqueous
solution by carbon nanotubes and carbon nanotube functional groups: Critical
review. RSC Advances, 7(74), 47083–47090.
Robati, D., Mirza, B., Rajabi, M., Moradi, O., Tyagi, I., Agarwal, S., & Gupta, V. (2016).
Removal of hazardous dyes-BR 12 and methyl orange using graphene oxide as an
adsorbent from aqueous phase. Chemical Engineering Journal, 284, 687–697.
Rossi, B. R., Pellegrini, V. O., Cortez, A. A., Chiromito, E. M., Carvalho, A. J., Pinto, L. O.,
& Polikarpov, I. (2021). Cellulose nanofibers production using a set of recombinant
enzymes. Carbohydrate Polymers, 256, Article 117510.
Seo, J. Y., Cho, K. Y., Lee, J.-H., Lee, M. W., & Baek, K.-Y. (2020). Continuous flow
composite membrane catalysts for efficient decomposition of chemical warfare agent
simulants. ACS Applied Materials & Interfaces, 12(29), 32778–32787.
Shin, M. G., Park, S.-H., Kwon, S. J., Kwon, H.-E., Park, J. B., & Lee, J.-H. (2019). Facile
performance enhancement of reverse osmosis membranes via solvent activation with
benzyl alcohol. Journal of Membrane Science, 578, 220–229.
Song, Y., Seo, J. Y., Kim, H., & Beak, K.-Y. (2019). Structural control of cellulose
nanofibrous composite membrane with metal organic framework (ZIF-8) for highly
selective removal of cationic dye. Carbohydrate Polymers, 222, Article 115018.

Soyekwo, F., Zhang, Q., Gao, R., Qu, Y., Lin, C., Huang, X., & Liu, Q. (2017). Cellulose
nanofiber intermediary to fabricate highly-permeable ultrathin nanofiltration
membranes for fast water purification. Journal of Membrane Science, 524, 174–185.
Stade, S., Kallioinen, M., Mikkola, A., Tuuva, T., & Mă
anttă
ari, M. (2013). Reversible and
irreversible compaction of ultrafiltration membranes. Separation and Purification
Technology, 118, 127–134.
Subrahmanya, T., Arshad, A. B., Lin, P. T., Widakdo, J., Makari, H., Austria, H. F. M., &
Hung, W.-S. (2021). A review of recent progress in polymeric electrospun nanofiber
membranes in addressing safe water global issues. RSC Advances, 11(16),
9638–9663.
Sun, X., Xu, W., Zhang, X., Lei, T., Lee, S.-Y., & Wu, Q. (2021). ZIF-67@ cellulose
nanofiber hybrid membrane with controlled porosity for use as Li-ion battery
separator. Journal of Energy Chemistry, 52, 170–180.
Tsai, W.-C., Wang, S.-T., Chang, K.-L. B., & Tsai, M.-L. (2019). Enhancing saltiness
perception using chitin nanomaterials. Polymers, 11(4), 719.
Venkatesan, M., Veeramuthu, L., Liang, F.-C., Chen, W.-C., Cho, C.-J., Chen, C.-W., &
Kuo, C.-C. (2020). Evolution of electrospun nanofibers fluorescent and colorimetric
sensors for environmental toxicants, pH, temperature, and cancer cells–A review
with insights on applications. Chemical Engineering Journal, 397, Article 125431.
Wissbrun, K. F. (1981). Rheology of rod-like polymers in the liquid crystalline state.
Journal of Rheology, 25(6), 619–662.
Wu, N., Wang, Y., Lei, Y., Wang, B., Han, C., Gou, Y., & Fang, D. (2015). Electrospun
interconnected Fe-N/C nanofiber networks as efficient electrocatalysts for oxygen
reduction reaction in acidic media. Scientific Reports, 5(1), 1–9.
Xia, Y., Yang, P., Sun, Y., Wu, Y., Mayers, B., Gates, B., & Yan, H. (2003). Onedimensional nanostructures: Synthesis, characterization, and applications. Advanced
Materials, 15(5), 353–389.
Xu, R., Mao, J., Peng, N., Luo, X., & Chang, C. (2018). Chitin/clay microspheres with
hierarchical architecture for highly efficient removal of organic dyes. Carbohydrate

Polymers, 188, 143–150.

Younghan Song: Conceptualization, Investigation, software, and
Writing – Reviewing and Editing.
Jin Young Seo: Conceptualization, Methodology, Investigation,
Writing – Original draft preparation, Formal analysis.
Hyungsup Kim: Supervision.
Sangho Cho: Writing – Writing – Reviewing and Editing, Funding
acquisition, and Supervision.
Kyung-Youl Baek: Writing, Writing – Reviewing and Editing,
Funding acquisition, and Supervision.
Acknowledgements
This work was supported by National Research Foundation of Korea
(NRF- 2020M3H4A3106354) and KIST institutional program. This work
was partially supported by the Technology Innovation Program
(20008653) funded by the Ministry of Trade, Industry & Energy
(MOTIE, Korea).
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.carbpol.2021.118754.
References
Abdelhameed, R. M., El-Shahat, M., & Emam, H. E. (2020). Employable metal (Ag & Pd)
@ MIL-125-NH2@ cellulose acetate film for visible-light driven photocatalysis for
reduction of nitro-aromatics. Carbohydrate Polymers, 247, Article 116695.
Abdelhameed, R. M., & Emam, H. E. (2019). Design of ZIF (Co & Zn)@ wool composite
for efficient removal of pharmaceutical intermediate from wastewater. Journal of
Colloid and Interface Science, 552, 494–505.
Arbulu, R. C., Jiang, Y. B., Peterson, E. J., & Qin, Y. (2018). Metal–organic framework
(MOF) nanorods, nanotubes, and nanowires. Angewandte Chemie International
Edition, 57(20), 5813–5817.

Chen, T., Duan, M., Shi, P., & Fang, S. (2017). Ultrathin nanoporous membranes derived
from protein-based nanospheres for high-performance smart molecular filtration.
Journal of Materials Chemistry A, 5(38), 20208–20216.
Chen, X., Liu, Y., Lu, H., Yang, H., Zhou, X., & Xin, J. H. (2010). In-situ growth of silica
nanoparticles on cellulose and application of hierarchical structure in biomimetic
hydrophobicity. Cellulose, 17(6), 1103–1113.
Cho, K. Y., An, H., Do, X. H., Choi, K., Yoon, H. G., Jeong, H.-K., & Baek, K.-Y. (2018).
Synthesis of amine-functionalized ZIF-8 with 3-amino-1, 2, 4-triazole by
postsynthetic modification for efficient CO 2-selective adsorbents and beyond.
Journal of Materials Chemistry A, 6(39), 18912–18919.
Chronakis, I. S. (2005). Novel nanocomposites and nanoceramics based on polymer
nanofibers using electrospinning process—A review. Journal of Materials Processing
Technology, 167(2–3), 283–293.
Cui, J., Li, F., Wang, Y., Zhang, Q., Ma, W., & Huang, C. (2020). Electrospun nanofiber
membranes for wastewater treatment applications. Separation and Purification
Technology, 250, Article 117116.
Emam, H. E., El-Shahat, M., & Abdelhameed, R. M. (2021). Observable removal of
pharmaceutical residues by highly porous photoactive cellulose acetate@ MIL-MOF
film. Journal of Hazardous Materials, 414, Article 125509.
Fukuzumi, H., Saito, T., Iwata, T., Kumamoto, Y., & Isogai, A. (2009). Transparent and
high gas barrier films of cellulose nanofibers prepared by TEMPO-mediated
oxidation. Biomacromolecules, 10(1), 162–165.
Isogai, A., Saito, T., & Fukuzumi, H. (2011). 3(1), 71–85.
Joseph, B., Mavelil Sam, R., Balakrishnan, P., Gopi, S., Volova, T., Maria, J. H., &
Thomas, S. (2020). Extraction of nanochitin from marine resources and fabrication
of polymer nanocomposites: Recent advances. Polymers, 12(8), 1664.
Khajeh, M., Oveisi, A. R., Barkhordar, A., Rakhshanipour, M., & Sargazi-Avval, H.
(2021). Ternary NiCuZr layered double hydroxide@ MIL-101 (Fe)-NH 2 metalorganic framework for photocatalytic degradation of methylene blue. Journal of
Nanostructure in Chemistry, 1–11.
Kim, J.-K., Choi, B., & Jin, J. (2020). Transparent, water-stable, cellulose nanofiber-based

packaging film with a low oxygen permeability. Carbohydrate Polymers, 249, Article
116823.
Lee, M. W., An, S., Kim, Y.-I., Yoon, S. S., & Yarin, A. L. (2018). Self-healing threedimensional bulk materials based on core-shell nanofibers. Chemical Engineering
Journal, 334, 1093–1100.
Li, D., & Xia, Y. (2004). Electrospinning of nanofibers: Reinventing the wheel? Advanced
Materials, 16(14), 1151–1170.
Li, J., Revol, J., & Marchessault, R. (1996). Rheological properties of aqueous
suspensions of chitin crystallites. Journal of Colloid and Interface Science, 183(2),
365–373.

8


Y. Song et al.

Carbohydrate Polymers 275 (2022) 118754

Yoon, M. J., Doh, S. J., & Im, J. N. (2011). Preparation and characterization of
carboxymethyl cellulose nonwovens by a wet-laid process. Fibers and Polymers, 12
(2), 247–251.
Zhang, Q. G., Deng, C., Soyekwo, F., Liu, Q. L., & Zhu, A. M. (2016). Sub-10 nm wide
cellulose nanofibers for ultrathin nanoporous membranes with high organic
permeation. Advanced Functional Materials, 26(5), 792–800.
Zhao, C., Zhou, X., & Yue, Y. (2000). Determination of pore size and pore size
distribution on the surface of hollow-fiber filtration membranes: A review of
methods. Desalination, 129(2), 107–123.

Zhong, T., Wolcott, M. P., Liu, H., Glandon, N., & Wang, J. (2020). The influence of prefibrillation via planetary ball milling on the extraction and properties of chitin
nanofibers. Cellulose, 27, 6205–6216.
Zhu, T., Chen, J. S., & Lou, X. W. (2012). Highly efficient removal of organic dyes from

waste water using hierarchical NiO spheres with high surface area. The Journal of
Physical Chemistry C, 116(12), 6873–6878.

9