SCIENCE ADVANCES | RESEARCH ARTICLE

MATERIALS SCIENCE Copyright © 2020

The Authors, some
Lightweight, tough, and sustainable cellulose rights reserved;
exclusive licensee
nanofiber-derived bulk structural materials with low American Association
for the Advancement
thermal expansion coefficient of Science. No claim to
original U.S. Government
Works. Distributed
Qing-Fang Guan1*, Huai-Bin Yang1*, Zi-Meng Han1*, Li-Chuan Zhou2, Yin-Bo Zhu2, under a Creative
Zhang-Chi Ling1, He-Bin Jiang2, Peng-Fei Wang2, Tao Ma1, Heng-An Wu2, Shu-Hong Yu1† Commons Attribution
NonCommercial
Sustainable structural materials with light weight, great thermal dimensional stability, and superb mechanical License 4.0 (CC BY-NC).
properties are vitally important for engineering application, but the intrinsic conflict among some material prop-
erties (e.g., strength and toughness) makes it challenging to realize these performance indexes at the same time
under wide service conditions. Here, we report a robust and feasible strategy to process cellulose nanofiber (CNF)
into a high-performance sustainable bulk structural material with low density, excellent strength and toughness,
and great thermal dimensional stability. The obtained cellulose nanofiber plate (CNFP) has high specific strength
[~198 MPa/(Mg m−3)], high specific impact toughness [~67 kJ m−2/(Mg m−3)], and low thermal expansion coeffi-
cient (<5 × 10−6 K−1), which shows distinct and superior properties to typical polymers, metals, and ceramics,
making it a low-cost, high-performance, and environmental-friendly alternative for engineering requirement, es-
pecially for aerospace applications.

Downloaded from https://www.science.org on May 13, 2023

INTRODUCTION
Since the dawn of human existence, materials have been fundamen­
tal to the development of society. In all kinds of materials, structural
materials, such as metals, ceramics, and polymers, are the most widely
used (1). In recent years, designing structural materials with high
performance on mutually exclusive properties (e.g., strength and
toughness) at the same time, especially based on nanoscale building
blocks, attracted more and more interest (2–8). When these nano­
scale building blocks are assembled into macroscale materials, many
extraordinary nanoscale properties can be scaled to macroscopic level
and new macroscopic properties that are ascribable to the assembly
of individual units emerge. In particular, it is of great importance
and remains a huge challenge to construct high-performance, all-green
bulk structural materials from renewable and sustainable nanoscale
building blocks (9–15).
Cellulose nanofiber (CNF), which can be derived from plant or
produced by bacteria, is one of the most abundant all-green resources
on Earth (16, 17). Many attractive properties of CNF including low
density, low thermal expansion coefficient, high strength, high stiff­
ness, and easily modifiable surface make CNF an ideal nanoscale
building block for constructing macroscopic high-performance ma­
terials. Although a variety of efforts have been made to scale those
extraordinary nanoscale properties of CNFs to macroscopic level,
only macro fibers and films can be prepared by different strategies up
until now. For example, macro fiber was obtained from wood CNF
by flow-assisted organization, with a Young’s modulus of 86 GPa and
a tensile strength of 1.57 GPa, exceeding any known natural or syn­
thetic biopolymers (18). Macro fiber with high performance was also
1
Division of Nanomaterials and Chemistry, Hefei National Laboratory for Physical
Sciences at the Microscale, Chinese Academy of Sciences (CAS) Center for Excel-
lence in Nanoscience, Hefei Science Center of CAS, Department of Chemistry, Insti-
tute of Biomimetic Materials & Chemistry, University of Science and Technology of
China, Hefei 230026, China. 2CAS Key Laboratory of Mechanical Behavior and De-
sign of Materials, Department of Modern Mechanics, CAS Center for Excellence in
Complex System Mechanics, University of Science and Technology of China, Hefei
230027, China.
*These authors contributed equally to this work.
†Corresponding author. Email: shyu@ustc.edu.cn
prepared by a wet-drawing and wet-twisting process of ultralong
bacterial CNF (19). In addition, films consisting of CNF with high
strength, high transparency, and low thermal expansion coefficient
have been designed and used in many fields such as electron devices
and flexible display (20–24). However, challenges remain in scaling those
extraordinary nanoscale properties of CNFs to three-dimensional
bulk structural materials. If sustainable high-performance bulk struc­
tural materials can be constructed, it will certainly promote the devel­
opment of CNF, expand its application fields, and provide more
materials selection in engineering design.
Here, we report a robust and feasible strategy to process CNFs
into a high-performance bulk structural material with low density,
outstanding strength and toughness, and great thermal dimensional
stability. The obtained CNF plate (CNFP) has high specific strength
[~198 MPa/(Mg m−3)], high specific impact toughness [~67 kJ m−2/
(Mg m−3)], and low thermal expansion coefficient (<5 × 10−6 K−1),
which shows distinct and superior properties to typical polymers,
metals, and ceramics, making it a low-cost, high-performance, and
environmental-friendly alternative for engineering requirement, es­
pecially as spacecraft materials. Moreover, CNFP also shows good
serviceability under extreme temperature or rapid thermal shock
and high energy absorption properties, and its mass production can
be achieved in large scale at low cost.
RESULTS
Material fabrication and characterization
Figure 1 shows a schematic of our bottom-up approach to press multi­
layered pretreated CNF hydrogels into high-performance CNFP.
CNF hydrogels with a robust three-dimensional nanofiber network
are produced from glucose by biosynthesis and then treated with a
different polymer solution or by surface modification before hot-­
pressing (Fig. 1, A to C). The CNFPs prepared from untreated, poly­
vinyl alcohol (PVA)–treated, and silicic acid solution–treated bacterial
CNF hydrogels and polyacrylic acid (PAA)–treated surface-oxidized
bacterial CNF hydrogels are marked as CNFP-0, CNFP-1, CNFP-2,
and CNFP-3, respectively (table S1).

Guan et al., Sci. Adv. 2020; 6 : eaaz1114 1 May 2020 1 of 8

SCIENCE ADVANCES | RESEARCH ARTICLE

Downloaded from https://www.science.org on May 13, 2023

Fig. 1. Fabrication and structure analysis of CNFP. (A) CNF hydrogel can be produced by biosynthesis. (B) CNF hydrogel and its robust three-dimensional nanofiber
network. (C) Numerous layers of treated CNF hydrogels are pressed at 80°C to fabricate CNFP. (D) The diagrammatic drawing of CNFP. (E) The multilayer structure of CNFP.
(F) The robust three-dimensional nanofiber network of one layer in CNFP. (G) Cellulose molecular chains are tightly bonded together by hydrogen bonds and expose lots
of –OH groups on the surface of CNF to form interfiber hydrogen bonds. (H) Photograph of large-sized CNFP with a volume of 320 mm by 220 mm by 27 mm. (I) Parts with
different shapes of CNFP produced by a milling machine. Scale bar, 1 cm (I). (Photo credit: Zi-Meng Han, Department of Chemistry, University of Science and Technology
of China, Hefei 230026, China.)

Fig. 2. Structural characterization of CNFP. (A) SEM image of the marked area in the inset photograph, clearly showing the multilayer structure where each layer is
about 20 m. The inset is a photograph of CNFP. (B) Magnified SEM image of a layer in CNFP, showing the microscopic layered structure of CNFs. (C) SEM image of the
marked area in the inset photograph, showing the robust three-dimensional nanofiber network. Numerous CNFs are intertwined with each other and combined together
by a strong hydrogen bond. The inset is a photograph of CNFP. (D) Profile of the fractured CNFP-0 showing the sliding between about 20-m layers. (E) Numerous CNFs
are intertwined with each other and combined together between layers [enlarged micrograph of the marked area in (D)]. (F) SEM image of an oblique section of CNFP. Be-
tween different layers, a large number of CNFs are pulled out from the layer and intertwine with each other. (Photo credit: Huai-Bin Yang, Department of Chemistry,
University of Science and Technology of China, Hefei 230026, China.)

Guan et al., Sci. Adv. 2020; 6 : eaaz1114 1 May 2020 2 of 8

SCIENCE ADVANCES | RESEARCH ARTICLE

Our method can achieve large-scale preparation directly by larger
press equipment. A large-sized CNFP was fabricated with a volume
of 320 mm by 220 mm by 27 mm and with a weight of 2560 g (Fig. 1H).
Since bacterial cellulose hydrogel is a kind of industrial product and
its market price is lower than $0.01 per kilogram, CNFP can be pre­
pared at low cost, which is lower than many typical structural materials
(tables S2 and S3). Furthermore, CNFP can be processed into desired
shape and size; for example, different shapes of parts were obtained by
a milling machine, which shows its good processability (Fig. 1I).
The scanning electron microscopy (SEM) images reveal the multi­
layer structure of CNFP, and each layer is about 20 m, which is deter­
mined by the thickness of CNF hydrogels (Fig. 2A). Inside the layer,
there are densified and robust three-dimensional nanofiber networks
(Fig. 2, B and C and fig. S1, G to I). Numerous CNFs are intertwined
with each other and combined together by strong hydrogen bonds.
The interfacial failure of CNFP under external loading shows a large
amount of connected CNFs, demonstrating the tightly hydrogen-­
bonded adjacent layers in CNFP (Fig. 2, D to F, and fig. S1, A to F).
Thermal and mechanical properties
Through this strategy, many remarkable nanoscale properties of
CNF can be successfully scaled to macro level on CNFPs, including
low thermal expansion coefficient and high strength. Low thermal
expansion coefficient is a vitally important feature for materials in
many application areas, especially for aerospace applications. From
−120° to 150°C, the average thermal expansion coefficients of CNFP
are lower than 5 × 10−6 K−1 (parallel to layer) and 7 × 10−6 K−1 (per­
pendicular to layer), respectively, which is close to those of ceramic
materials and much lower than those of typical polymers and metals
(Fig. 3A and fig. S2). Meanwhile, the excellent mechanical properties

Downloaded from https://www.science.org on May 13, 2023

Fig. 3. Superb thermal and mechanical properties of CNFP. (A) Thermal expansion of CNFP (parallel to layer), polyamide (PA), Al alloy (7075 Al), and Al2O3. (B) Compar-
ison of flexural strength and stiffness of different kinds of CNFPs. (C) Comparison of Charpy impact toughness of CNFP-0 with other widely used polymer-based materials.
(D) Flexural stress–strain curves of CNFP-0 at different temperatures. (E and F) Comparison of CNFP-0 with other widely used polymer-based materials at (E) 30°C and (F)
200°C. (G) Schematic of rapid thermal shock for 10 times. (H) Flexural stress–strain curves of CNFP-0 before and after rapid thermal shock for 10 times. PMMA, polymethyl
methacrylate; PVC, polyvinyl chloride; ABS, acrylonitrile butadiene styrene; PC, polycarbonate; PF, phenolic resin; POM, polyformaldehyde; PP, polypropylene. (Photo
credit: Zi-Meng Han, Department of Chemistry, University of Science and Technology of China, Hefei 230026, China.)

Guan et al., Sci. Adv. 2020; 6 : eaaz1114 1 May 2020 3 of 8

SCIENCE ADVANCES | RESEARCH ARTICLE
of CNF are also successfully scaled to macro level on CNFPs
(Fig. 3B; see also figs. S3 to S6). With a low density of ~1.35 g cm−3,
the ultimate flexural strength and the flexural modulus of CNFP can
be up to ~269 MPa and ~17 GPa, respectively (Fig. 3B). The organi­
zation of CNFs also provides other intriguing macroscopic proper­
ties. The Charpy impact test of notched CNFP yields an impact
toughness of 87.6 ± 4.3 kJ m−2, which is much higher than those of
typical plastics (Fig. 3C). Moreover, it shows no visible change at
200°C, indicating that the thermostability of CNFP is much better
than those of other widely used polymer-based materials (Fig. 3,
E and F, and fig. S7). After enduring rapid thermal shock for 10 times
between two extreme temperature conditions (−196° and 120°C),
CNFP still maintains a similar mechanical performance (Fig. 3, G
and H). Notably, at 120° and −50°C, the flexural strength of CNFP
does not change notably (Fig.  3D), which is vitally important for
practical application in extreme environments. In addition, after
having been subjected to 95% relative humidity for 60 hours, the
thickness and flexural strength of CNFP change only slightly, show­
ing good stability under moisture attack (fig. S8).
CNFP has multiple outstanding features in one material, includ­
ing low thermal expansion coefficient, lightweight, high strength, and
high toughness. To demonstrate the overall performance of CNFP,
we put its properties into two Ashby maps to compare with different
kinds of traditional structural materials (Fig. 4). In Ashby maps, de­
spite the wide range of properties of metals (as an example), the clusters
occupy a field that is distinct from that of polymers or that of ceramics.
This fact shows that for different kinds of materials, the fields may
overlap, but they always have a characteristic place within the whole
picture (1). According to this pattern, CNFP appears on the maps as
a whole new kind of material, which occupies a distinct field with low
thermal expansion coefficient (lower than 5 × 10−6 K−1), high specific
strength [up to 198 MPa/(Mg m−3)], and high specific impact toughness
[up to 67 kJ m−2/(Mg m−3)]. The specific strength and specific impact
toughness of CNFP are higher than those of traditional metals and
alloys, making it an all-green and high-performance alternative for
engineering design. The position of the fields on Ashby maps can be
understood in simple physical terms, as the nature of the fundamental
building blocks and how they combine with each other determine the
Fig. 4. Comparison of thermal and mechanical properties of CNFP with typical polymers, metals, and ceramics. (A) Ashby diagram of thermal expansion versus
specific strength for CNFP compared with typical polymers, metals, and ceramics (1, 41–46). (B) Ashby diagram of thermal expansion versus specific impact toughness for
CNFP compared with typical polymers, metals, and ceramics (1).
Guan et al., Sci. Adv. 2020; 6 : eaaz1114 1 May 2020
position of the fields that different kinds of materials occupy. Consider­
ing CNFP, CNFs are its basic building blocks, while hydrogen bonds
are the main interaction that bind them together. The strength (at
least 2 GPa) and Young’s modulus (138 GPa) of individual CNFs can
be almost as high as those of steel and Kevlar (22), and those superior
nanoscale properties can be scaled to macro level (tables S4 and S5),
mainly due to the strong interaction between CNFs (25–29).
Mechanism analysis of thermal and mechanical properties
According to previous researches (16, 22, 30, 31), since the CNFs are
aggregates of semi-crystalline extended cellulose chains, its thermal
expansion coefficient is as small as 1 × 10−7 K−1, which is lower than
that of silica glass. When those CNFs are bound by strong interfiber
hydrogen bonds, a low thermal expansion coefficient is achieved.
The ultrafine nanofiber network structure in CNFP results in more
extensive hydrogen bonding, the high in-plane orientation, and “three-­
way branching points” of the microfibril networks (Fig. 2F) (17).
Those structural features allow the CNFP to withstand high stress
without breaking and also disperse stress and suppress crack forma­
tion and crack propagation. According to a photograph of a broken
sample (fig. S9A), the main fracture mode of CNFP is interlaminar
Downloaded from https://www.science.org on May 13, 2023
debonding, which means that the interlaminar shear strength be­
tween layers is the limiting factor of the strength of CNFP. Thus, we
tune the interlaminar shear strength by treating the CNF hydrogel
surface with different polymer solutions. For untreated CNFP-0, the
interaction between layers is hydrogen bond, while PVA-treated
CNFP-1 has weaker interlaminar shear strength, since PVA forms
weaker hydrogen bonds between layers than cellulose itself. For
CNFP-2 treated by silicic acid solution, silicic acid dehydrates and
provides covalent strong cross-links during the hot-press process
(32), which effectively improves the interlayer interaction compared
to CNFP-0 and CNFP-1. For CNFP-3, the oxidation process intro­
duces an amount of carboxyl groups on the surface of CNF. Those
carboxyl groups of CNFs are further cross-linked by the calcium ions
and polyacrylic acid. The calcium ions and carboxyl groups form
strong ionic bond and further improve the interaction between layers.
The nuclear magnetic resonance (NMR)–13C and Fourier transform
infrared spectroscopy (FTIR) results verify the mechanism above,
4 of 8
SCIENCE ADVANCES | RESEARCH ARTICLE

Downloaded from https://www.science.org on May 13, 2023

Fig. 5. Superb toughness mechanism and impact resistance properties of CNFP. (A and B) FEM simulations of (A) laminated structure and (B) monolithic bulk for the
SENB test. (C) Stress-strain curves of different kinds of interfaces for the SENB test. (D) Force-displacement curve of CNFP-0 for the drop hammer impact test. (E) Schematic
of the drop hammer impact tester. (F) Photograph of CNFP-0 after the drop hammer impact test. Scale bar, 1 cm. (G) Schematic of the SHPB. (H) Compressive stress-strain
curves of CNFP-0 for the SHPB test under different strain rates. (Photo credit: Huai-Bin Yang, Department of Chemistry, University of Science and Technology of China,
Hefei 230026, China.)

Guan et al., Sci. Adv. 2020; 6 : eaaz1114 1 May 2020 5 of 8

SCIENCE ADVANCES | RESEARCH ARTICLE
while the x-ray diffraction (XRD) result shows that the crystallinity of
CNFs has no visible change after pretreatment (figs. S10, S11, and S12)
(33, 34). By tuning the interaction between layers, it can be seen that
when the interaction is enhanced, the interlaminar shear strength and
flexural strength increase simultaneously (Fig. 3B; see also fig. S9B).
A long-standing challenge in engineering material design is the
conflict between strength and toughness, because these properties are,
in general, mutually exclusive (25, 35–37). The high ultimate flexural
strength of CNFP is not accompanied by low impact toughness
(Fig. 3, B and C). This unique toughening mechanism can be under­
stood at multiscale. At microscale, bending force initiated the sliding
between ~20-m layers (Figs. 1E and 2D; see also fig. S1, A to C),
and the sliding dispersed the stress, avoiding stress concentration.
At nanoscale, tightly hydrogen-bonded CNFs at interface are pulled
out from the opened layer during the interlaminar sliding (Figs. 1F
and 2, E and F; see also fig. S1, D to F), which further disperse the
stress and prevent the generation and propagation of crack. At mo­
lecular scale, owing to the rich hydroxyl groups in cellulose molec­
ular chains, when CNFPs are subjected to deformation, relative sliding
among CNFs involves an enormous number of hydrogen-bond for­
mation, breaking, and reformation (Fig. 1G) (25, 38). Owing to the
above mechanisms, deformation of the CNF network can absorb a
large amount of energy. To demonstrate the role of interfacial prop­
erties on toughness, we further investigated the failure behavior of
CNFP under single-edge notched bend (SENB) simulation. As a com­
parison, the monolithic bulk shows the stress concentration around
the crack tip and soon displays the brittle damage through the frac­
ture section (Fig. 5B). With the decrease of interfacial strength, it is
clear that laminar bulk illustrates brittle-to-tough transition from
the force-displacement curves (Fig. 5C). For an appropriate inter­
facial strength, the detailed failure behavior is given in Fig.  5A,
where the deformation energy is dissipated by interfacial sliding and
opening instead of brittle damage of laminae. For CNFP, therefore,
suitable interfacial modification can enormously enhance the tough­
ness and retain considerable strength at the same time.
Impact resistance
To further evaluate the impact resistance of CNFP, dynamic me­
chanical tests including drop hammer impact and split Hopkinson
pressure bar (SHPB) were carried out. The CNFP sample can gen­
erate a high resist force during the impact and remarkably decrease
the velocity of hammer and absorb its energy (Fig. 5, D and E). The
absorption energy in the drop hammer test can be calculated by the
force-displacement curve, and for CNFP, it is ~2.7 kJ m–1 (Fig. 5D).
CNFP under high-velocity impact can maintain its shape, instead of
shattering or deforming. The damage pattern of CNFP demonstrates
the stress localization of energy adsorption under high-velocity im­
pact (Fig. 5F). SHPB is a powerful tool to study the behavior of mate­
rial under high–strain rate deformation. Its structure is shown in
Fig. 5G, and the strain gauge on the incident bar and the transmis­
sion bar can convert mechanical wave into electronic signal and yield
the stress-strain curve. The highest compression stress of CNFP can
reach ~1600 MPa during high–strain rate compression (14,000 s−1),
and the energy absorption of compression is ~387.5, ~89.3, and
~37.0 MJ m−3 under the strain rates of 14,000, 10,000, and 7500 s−1,
respectively (Fig. 5H). The stress-strain curves in the SHPB experi­
ments illustrate anomalous decline after the first peak compared
with the common plastic platform area of traditional metal materials
and energy absorbing materials. Under initial low strain in the impact,
Guan et al., Sci. Adv. 2020; 6 : eaaz1114 1 May 2020
the CNF network and the layered plate as the structural frameworks
resist the shock stress. Then, the following large strain would trigger the
deformation of the CNF network at microscale and even the shear
behavior of cellulose nanocrystals at nanoscale. Recent simulation
studies demonstrated that breaking and reformation of interfiber
hydrogen bonds as well as the dihedral rotation are non-negligible
mechanisms in the origin of outstanding mechanical properties of
cellulose nanocrystals (38, 39). At last, the larger compression strain re­
sults in the densification of CNFP (CNF network at microscale). While
the compression speed increases, adsorption energy increases. The light
weight and high energy absorption properties of CNFP indicate that
it can be a potential armored material for shock waves from blast (40).
DISCUSSION
In summary, we have developed a robust and feasible strategy to
process CNF into an environmentally friendly bulk structural mate­
rial, CNFP. It has multiple unique features in one material, which
are low thermal expansion coefficient, light weight, high strength and
toughness, good impact resistance, and easy realization of large-scale
production, making it a low-cost and high-performance alternative
Downloaded from https://www.science.org on May 13, 2023
for engineering design. For example, it can be a strong competitor
for the lightweight material used for automobiles and aircraft, and
remarkably for aerospace application, such as optical lens bracket
for the lunar rover where light weight, high strength, and low thermal
expansion coefficient are all vitally essential. From the perspective of
environmental life cycle impact assessment (10), CNFP contributes
almost nothing to extra greenhouse gas emissions, human health
impairment, ecosystem toxicity, or resource depletion, indicating
that it is a good example of an all-green structural material. Further­
more, CNFP can be designed at multiscale according to application
requirements. Various functional bulk structural nanocomposites
can be achieved, and thus, the performance of this emerging sus­
tainable structural material can be further explored.
MATERIALS AND METHODS
Fabrication of CNFPs
All reagents and raw materials were commercially available. Bacte­
rial cellulose hydrogels were produced by Gluconacetobacter xylinus
1.1812 (China General Microbiological Culture Collection Center)
on the liquid nutrient media at 30°C (24). One liter of liquid nutrient
media consisted of glucose (50 g liter−1), yeast extract (5 g liter−1),
citric acid (2 g liter−1), Na2HPO4·12H2O (4 g liter−1), and KH2PO4
(2 g liter−1). A part of bacterial cellulose hydrogels were surface-­
oxidized by the 2,2,6,6-Tetramethylpiperidine-1-oxyl-oxidized method.
They were then immersed into CaCl2 solution (0.1 M) for 24 hours
and rinsed three times with deionized water. The bacterial cellulose
hydrogels (untreated/surface oxidation) were cut into sheets, and
they were immersed in the surface treatment solution for 6 hours.
Then, a certain number of bacterial cellulose sheets were stacked by
hydrogel layer-­by-layer (HLBL) and were pressed with a pressure of
~1 MPa. Last, a hot-pressing step with a pressure of 100 MPa was
applied at 80°C until CNFP was completely dry. The treatment method
of bacterial cellulose hydrogels, surface treatment solution, and the
corresponding numbers are shown in table S1. For a typical CNFP-0
with a thickness of 6 mm, it was pressed with a pressure of ~1 MPa
for about 3 hours and then was hot-pressed with a pressure of ~100 MPa
at 80°C for about 1 hour.
6 of 8
SCIENCE ADVANCES | RESEARCH ARTICLE
Characterization
SEM images were taken with a Carl Zeiss Supra 40 field emission
scanning electron microscope (5 kV), and all samples were gold-­
sputtered for 30 s at a constant current of 30 mA before observation.
XRD data were measured by a PANalytical X’pert PRO MRD X-ray
diffractometer equipped with Cu K radiation ( = 1.54056 Å), and
the samples were prepared by cutting down a ~0.5-mm layer of CNFPs.
Thermogravimetric analysis data were measured in air atmosphere
with a TA Instruments SDT Q600 thermogravimetric analyzer, and
the samples were prepared by grinding CNFPs into powders. FTIR
spectra were acquired by a Bruker Vector-22 FTIR spectrometer
in attenuated total reflectance mode, and the samples were pre­
pared by cutting down a ~0.5-mm layer of CNFPs. 13C-NMR cross-­
polarization magic angle spinning spectra were recorded at 100.62 MHz
on a Bruker Avance III 400 WB spectrometer using a 90° pulse of 4 s,
an acquisition time of 33.9 ms, a contact time of 3 ms, a recycle delay of
5 s, and a spin rate of 14 kHz.
Thermal and mechanical testing
Three-point bending test was performed on an Instron 5565A uni­
versal testing machine according to ASTM D790-15e1. The speci­
mens were carefully cut with a size of about 25  mm by 2  mm by
2 mm. The test was carried out at a loading rate of 1.0 mm min−1
with a support span of 12.5 mm. For all the mechanical tests, the
applied loading direction was perpendicular to the layers, and each
kind of specimen was tested at least five times, unless noted other­
wise. Tensile test and compression test were performed on an MTS
809 material testing machine. The dimensions for tensile specimens
(dog-bone-shaped specimens) were approximately 100  mm long,
10 mm wide, and 2 mm thick. The specimens were clamped at both
ends and stretched along the specimen length direction with a con­
stant test speed of 1.0 mm min−1 at room temperature. The dimen­
sions for compression specimens were about 9 mm by 9 mm by 4.5 mm,
and the specimens were compressed along the thickness direction
(4.5 mm) at room temperature with a loading rate of 1 mm min−1.
The Charpy impact test of CNFPs was performed on a Chengde Bao
Hui XJJY-5 pendulum impact tester, and the dimensions for Charpy
V-notch specimens were about 50 mm long, 10 mm wide, and 2 mm
thick with a notch depth of 1 mm. The drop hammer impact test of
CNFPs was performed on an Instron CEAST 9340 drop hammer
tester. The hammer with a weight of 15 kg (total mass) was falling
freely from a height of 30 cm, and the dimensions of the tested spec­
imens were about 50 mm by 50 mm by 2 mm. For the SHPB test,
the dimensions of the tested specimens were about 6 mm by 6 mm
by 2 mm.
Thermal expansion coefficient of the CNFP was measured by
NETZSCH TMA 402F3. The specimens were carefully cut with a
size of about 20 mm by 4 mm by 2 mm, and the test was carried out
from −130° to 150°C. Thermal expansion coefficient () was calcu­
lated by the equation  = L/(L T) K−1. Density () of the CNFP
was calculated by first treating the material into a cuboid and then
using the equation  = mass/volume. The specific values of the im­
pact fracture toughness and the ultimate flexural strength were cal­
culated by dividing the density. The short-beam shear strength can be
calculated by three-point bending test according to ASTM D2344/
D2344M-16. Short-beam shear strength (F sbs) was calculated by
​  ​​
0.75 ​Pm
​​F​​  sbs​ = ​─
    ​​     (1)
bh
Guan et al., Sci. Adv. 2020; 6 : eaaz1114 1 May 2020
where Pm is the maximum load in three-point bending test, b is the
measured specimen width, and h is the measured specimen thickness.
Computational methods
The SENB simulation for CNFPs is implemented by finite element
method (FEM) with the software ABAQUS. The geometric model
consists of three components: a plate specimen, an indenter, and
two symmetrical fixtures. The plate specimen is pre-cracked with a
notch in the middle of the specimen. The thickness of the notch is
half that of the specimen. The indenter and the fixtures are set as the
rigid body in the simulation. The contact constraints are defined be­
tween the component surface pairs.
The failure behavior of monolithic bulk without lamination is
simulated first. The property of CNFP elements is defined by the
user-defined field subroutine. The elements are linear elastic at the
initial stage. When the element stress is greater than the material
strength max, the element Young’s modulus is degraded to 2 × 10−9 of
the initial value as an approximate simulation of fracture damage. In
the simulation, the indenter pushes down with a uniform speed. The re­
action force and falling displacement of indenter are compared and
adopted to plot the stress-strain curve (Fig. 5C). It can be seen that
Downloaded from https://www.science.org on May 13, 2023
the curve grows linearly at first and then decreases rapidly after
reaching the peak value. The overall failure of the specimen is due to the
stress concentration at the pre-crack tip. The crack expands upward rap­
idly along the initial crack path, causing the specimen to break (Fig. 5B).
For multilayered CNFPs, the specimen is divided into isopachous
multiple layers to consider the effect of interface on the bending
property. Contact constraints with cohesive behavior are defined be­
tween the layers. The maximum shear stress max and fracture energy
G of cohesive elements are adjusted with different values to observe
the change of failure mode. When the values of max and G are big
enough, the failure mode is almost unchanged, and the specimen is
still broken from the middle. The flexural modulus is reduced due to
interlayer slippage. With the decrease of interfacial strength, we can
see that the failure mode varies, in which the interface reaches the
critical strength earlier than the elements during the indenter falling
process. The cracks appear at the middle layer first and expand to
both ends gradually (Fig. 5A). Then, more and more interlaminar
cracks appear inside the specimen, whereas the prefabricated cracks
cannot expand upward because energy is dispersed through layers. It
is found that the toughness of specimen is much higher according
to the stress-strain curve. As the interface strength continues to de­
crease, the failure mode remains the same, while the total strength
and toughness are gradually reduced.
In summary, the interface property has a significant impact on
the SENB test of CNFPs. The toughness of CNFPs can be improved
by adjusting the interface strength.
SUPPLEMENTARY MATERIALS
Supplementary material for this article is available at http://advances.sciencemag.org/cgi/
content/full/6/18/eaaz1114/DC1
REFERENCES AND NOTES
1. M. Ashby, Materials Selection in Mechanical Design (Elsevier, Elsevier-Butterworth-
Heinemann, 2011).
2. X. C. Liu, H. W. Zhang, K. Lu, Strain-induced ultrahard and ultrastable nanolaminated
structure in nickel. Science 342, 337–340 (2013).
3. H. L. Gao, S. M. Chen, L. B. Mao, Z. Q. Song, H. B. Yao, H. Cölfen, X. S. Luo, F. Zhang, Z. Pan,
Y. F. Meng, Y. Ni, S. H. Yu, Mass production of bulk artificial nacre with excellent
mechanical properties. Nat. Commun. 8, 287 (2017).
7 of 8
SCIENCE ADVANCES | RESEARCH ARTICLE
4. E. Munch, M. E. Launey, D. H. Alsem, E. Saiz, A. P. Tomsia, R. O. Ritchie, Tough, bio-inspired
hybrid materials. Science 322, 1516–1520 (2008).
5. R. Weinkamer, P. Fratzl, Solving conflicting functional requirements by hierarchical
structuring—Examples from biological materials. MRS Bull. 41, 667–671 (2016).
6. A. Picker, L. Nicoleau, Z. Burghard, J. Bill, I. Zlotnikov, C. Labbez, A. Nonat, H. Cöelfen,
Mesocrystalline calcium silicate hydrate: A bioinspired route toward elastic concrete
materials. Sci. Adv. 3, e1701216 (2017).
7. P. Podsiadlo, E. M. Arruda, E. Kheng, A. M. Waas, J. Lee, K. Critchley, M. Qin, E. Chuang,
A. K. Kaushik, H. S. Kim, Y. Qi, S. T. Noh, N. A. Kotov, LBL assembled laminates with
hierarchical organization from nano- to microscale: High-toughness nanomaterials
and deformation imaging. ACS Nano 3, 1564–1572 (2009).
8. J. Zhu, D. Watts, N. A. Kotov, Gelation-assisted layer-by-layer deposition of high
performance nanocomposites. Z. Phys. Chem. 232, 1383–1398 (2018).
9. E. A. Olivetti, J. M. Cullen, Toward a sustainable materials system. Science 360, 1396–1398
(2018).
10. R. E. Kirchain Jr., J. R. Gregory, E. A. Olivetti, Environmental life-cycle assessment. Nat.
Mater. 16, 693–697 (2017).
11. Y. Zhu, C. Romain, C. K. Williams, Sustainable polymers from renewable resources. Nature
540, 354–362 (2016).
12. A sustainable material world. Nat. Mater. 16, 691 (2017).
13. M. A. Hillmyer, The promise of plastics from plants. Science 358, 868–870 (2017).
14. A. K. Mohanty, S. Vivekanandhan, J. M. Pin, M. Misra, Composites from renewable
and sustainable resources: Challenges and innovations. Science 362, 536–542 (2018).
15. P. Podsiadlo, S. Y. Choi, B. Shim, J. Lee, M. Cuddihy, N. A. Kotov, Molecularly engineered
nanocomposites: Layer-by-layer assembly of cellulose nanocrystals. Biomacromolecules
6, 2914–2918 (2005).
16. K. Abe, S. Iwamoto, H. Yano, Obtaining cellulose nanofibers with a uniform width
of 15 nm from wood. Biomacromolecules 8, 3276–3278 (2007).
17. S. Yamanaka, K. Watanabe, N. Kitamura, M. Iguchi, S. Mitsuhashi, Y. Nishi, M. Uryu, The
structure and mechanical-properties of sheets prepared from bacterial cellulose.
J. Mater. Sci. 24, 3141–3145 (1989).
18. N. Mittal, F. Ansari, V. K. Gowda, C. Brouzet, P. Chen, P. T. Larsson, S. V. Roth, F. Lundell,
L. Wågberg, N. A. Kotov, L. D. Söderberg, Multiscale control of nanocellulose assembly:
Transferring remarkable nanoscale fibril mechanics to macroscale fibers. ACS Nano 12,
6378–6388 (2018).
19. S. Wang, F. Jiang, X. Xu, Y. Kuang, K. Fu, E. Hitz, L. Hu, Super-strong, super-stiff
macrofibers with aligned, long bacterial cellulose nanofibers. Adv. Mater. 29, 1702498
(2017).
20. S. Ifuku, M. Nogi, K. Abe, K. Handa, F. Nakatsubo, H. Yano, Surface modification
of bacterial cellulose nanofibers for property enhancement of optically transparent
composites: Dependence on acetyl-group DS. Biomacromolecules 8, 1973–1978 (2007).
21. S. Wang, T. Li, C. Chen, W. Kong, S. Zhu, J. Dai, A. J. Diaz, E. Hitz, S. D. Solares, T. Li, L. Hu,
Transparent, anisotropic biofilm with aligned bacterial cellulose nanofibers. Adv. Funct.
Mater. 28, 1707491 (2018).
22. H. Yano, J. Sugiyama, A. N. Nakagaito, M. Nogi, T. Matsuura, M. Hikita, K. Handa, Optically
transparent composites reinforced with networks of bacterial nanofibers. Adv. Mater. 17,
153–155 (2005).
23. M. Nogi, H. Yano, Transparent nanocomposites based on cellulose produced by bacteria
offer potential innovation in the electronics device industry. Adv. Mater. 20, 1849–1852
(2008).
24. Q.-F. Guan, Z.-M. Han, T.-T. Luo, H.-B. Yang, H.-W. Liang, S.-M. Chen, G.-S. Wang, S.-H. Yu,
A general aerosol-assisted biosynthesis of functional bulk nanocomposites. Natl. Sci. Rev.
6, 64–73 (2019).
25. J. Song, C. Chen, S. Zhu, M. Zhu, J. Dai, U. Ray, Y. Li, Y. Kuang, Y. Li, N. Quispe, Y. Yao,
A. Gong, U. H. Leiste, H. A. Bruck, J. Y. Zhu, A. Vellore, H. Li, M. L. Minus, Z. Jia, A. Martini,
T. Li, L. Hu, Processing bulk natural wood into a high-performance structural material.
Nature 554, 224–228 (2018).
26. M. Frey, D. Widner, J. S. Segmehl, K. Casdorff, T. Keplinger, I. Burgert, Delignified
and densified cellulose bulk materials with excellent tensile properties for sustainable
engineering. ACS Appl. Mater. Interfaces 10, 5030–5037 (2018).
27. C.-H. Fang, N. Mariotti, A. Cloutier, A. Koubaa, P. Blanchet, Densification of wood veneers
by compression combined with heat and steam. Eur. J. Wood Wood Prod. 70, 155–163
(2012).
28. P. Bekhta, S. Hiziroglu, O. Shepelyuk, Properties of plywood manufactured
from compressed veneer as building material. Mater. Design 30, 947–953 (2009).
29. E. V. Kultikova, Structure and Properties Relationships of Densified Wood (Virginia
Polytechnic Institute and State University, 1999).
30. T. Nishino, I. Matsuda, K. Hirao, All-cellulose composite. Macromolecules 37, 7683–7687
(2004).
Guan et al., Sci. Adv. 2020; 6 : eaaz1114 1 May 2020
31. J. Zhu, C. M. Andres, J. Xu, A. Ramamoorthy, T. Tsotsis, N. A. Kotov, Pseudonegative
thermal expansion and the state of water in graphene oxide layered assemblies.
ACS Nano 6, 8357–8365 (2012).
32. J. Shi, L. Lu, W. Guo, J. Zhang, Y. Cao, Heat insulation performance, mechanics
and hydrophobic modification of cellulose–SiO2 composite aerogels. Carbohydr. Polym.
98, 282–289 (2013).
33. R. Demir-Cakan, N. Baccile, M. Antonietti, M.-M. Titirici, Carboxylate-rich carbonaceous
materials via one-step hydrothermal carbonization of glucose in the presence of acrylic
acid. Chem. Mater. 21, 484–490 (2009).
34. W. Yang, X. Wang, P. Gogoi, H. Bian, H. Dai, Highly transparent and thermally stable
cellulose nanofibril films functionalized with colored metal ions for ultraviolet blocking
activities. Carbohydr. Polym. 213, 10–16 (2019).
35. R. O. Ritchie, The conflicts between strength and toughness. Nat. Mater. 10, 817–822
(2011).
36. Y. Wang, M. Chen, F. Zhou, E. Ma, High tensile ductility in a nanostructured metal.
Nature 419, 912–915 (2002).
37. L. B. Mao, H. L. Gao, H. B. Yao, L. Liu, H. Cölfen, G. Liu, S. M. Chen, S. K. Li, Y. X. Yan, Y. Y. Liu,
S. H. Yu, Synthetic nacre by predesigned matrix-directed mineralization. Science 354,
107–110 (2016).
38. H. Zhu, S. Zhu, Z. Jia, S. Parvinian, Y. Li, O. Vaaland, L. Hu, T. Li, Anomalous scaling law
of strength and toughness of cellulose nanopaper. Proc. Natl. Acad. Sci.U. S. A. 112,
8971–8976 (2015).
39. G. Molnár, D. Rodney, F. Martoïa, P. J. J. Dumont, Y. Nishiyama, K. Mazeau, L. Orgéas,
Cellulose crystals plastify by localized shear. Proc. Natl. Acad. Sci. U. S. A. 115, 7260–7265
(2018).
Downloaded from https://www.science.org on May 13, 2023
40. S. L. Chinke, I. S. Sandhu, D. R. Saroha, P. S. Alegaonkar, Graphene-like nanoflakes
for shock absorption applications. ACS Appl. Nano Mater. 1, 6027–6037 (2018).
41. M. Ashby, Hybrid materials to expand the boundaries of material-property space.
J. Am. Ceram. Soc. 94, s3–s14 (2011).
42. M. Guazzato, M. Albakry, S. P. Ringer, M. V. Swain, Strength, fracture toughness
and microstructure of a selection of all-ceramic materials. Part II. Zirconia-based dental
ceramics. Dent. Mater. 20, 449–456 (2004).
43. R. C. Munro, Material properties of a sintered -SiC. J. Phys. Chem. Ref. Data 26, 1195–1203
(1997).
44. S. Kanzaki, H. Tabata, T. Kumazawa, S. Ohta, Sintering and mechanical properties
of stoichiometric mullite. J. Am. Ceram. Soc. 68, C-6–C-7 (1985).
45. M. Munro, Evaluated material properties for a sintered alpha-alumina. J. Am. Ceram. Soc.
80, 1919–1928 (1997).
46. C. Aksel, B. Rand, F. L. Riley, P. D. Warren, Mechanical properties of magnesia-spinel
composites. J. Eur. Ceram. Soc. 22, 745–754 (2002).
Acknowledgments
Funding: This work was supported by the National Natural Science Foundation of China
(Grant 51732011), the Foundation for Innovative Research Groups of the National Natural
Science Foundation of China (Grant 21521001), Key Research Program of Frontier Sciences,
Chinese Academy of Sciences (Grant QYZDJ-SSW-SLH036), the National Basic Research
Program of China (Grant 2014CB931800), the Fundamental Research Funds for the Central
Universities (WK2090050043), the Users with Excellence and Scientific Research Grant of Hefei
Science Center of Chinese Academy of Sciences (2015HSC-UE007), the Strategic Priority
Research Program of the Chinese Academy of Sciences (XDB22040402), the National Natural
Science Foundation of China (11525211), and the USTC Research Funds of the Double
First-Class Initiative (YD2480002002). Author contributions: S.-H.Y. and Q.-F.G. conceived the
idea and designed the experiments. S.-H.Y. supervised the project. Q.-F.G., H.-B.Y., Z.-M.H., and
Z.-C.L. carried out the synthetic experiment and analysis. H.-A.W., L.-C.Z., and Y.-B.Z.
contributed to both mechanical simulations and analysis. H.-B.J. and P.-F.W. contributed to
SHPB test and analysis. T.M. contributed to the 3D illustrations. Q.-F.G., H.-B.Y., Z.-M.H., and
S.-H.Y. wrote the paper, and all authors discussed the results and commented on the
manuscript. Competing interests: The authors declare that they have no competing interests.
Data and materials availability: All data needed to evaluate the conclusions in the paper
are present in the paper and/or the Supplementary Materials. Additional data related to this
paper may be requested from the authors.
Submitted 13 August 2019
Accepted 10 February 2020
Published 1 May 2020
10.1126/sciadv.aaz1114
Citation: Q.-F. Guan, H.-B. Yang, Z.-M. Han, L.-C. Zhou, Y.-B. Zhu, Z.-C. Ling, H.-B. Jiang, P.-F. Wang,
T. Ma, H.-A. Wu, S.-H. Yu, Lightweight, tough, and sustainable cellulose nanofiber-derived bulk
structural materials with low thermal expansion coefficient. Sci. Adv. 6, eaaz1114 (2020).
8 of 8
 Lightweight, tough, and sustainable cellulose nanofiber-derived bulk structural
materials with low thermal expansion coefficient
Qing-Fang Guan, Huai-Bin Yang, Zi-Meng Han, Li-Chuan Zhou, Yin-Bo Zhu, Zhang-Chi Ling, He-Bin Jiang, Peng-Fei
Wang, Tao Ma, Heng-An Wu, and Shu-Hong Yu

Sci. Adv., 6 (18), eaaz1114.

DOI: 10.1126/sciadv.aaz1114

Downloaded from https://www.science.org on May 13, 2023

View the article online
https://www.science.org/doi/10.1126/sciadv.aaz1114
Permissions
https://www.science.org/help/reprints-and-permissions

Use of this article is subject to the Terms of service

Science Advances (ISSN 2375-2548) is published by the American Association for the Advancement of Science. 1200 New York Avenue
NW, Washington, DC 20005. The title Science Advances is a registered trademark of AAAS.
Copyright © 2020 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim
to original U.S. Government Works. Distributed under a Creative Commons Attribution NonCommercial License 4.0 (CC BY-NC).