Eco-friendly, flexible and stretchable printed electronics based on a sustainable elastic substrate and ink

Abstract

The fast-growing demand for electronics is generating a substantial amount of electronic waste, raising serious concerns about its environmental impact. Herein, disintegrable, flexible, and stretchable strain sensors and electrodes are developed by printing sustainable ink onto a biodegradable elastic substrate. The substrate is made of sodium carboxymethyl cellulose (NaCMC) functionalized with glycerol to achieve high stretchability, while the ink is composed of silver nanowires (AgNWs) with NaCMC and glycerol as the binders, which not only contribute to overall environmental sustainability but also enable strong bonding with the substrate. The effects of material composition of both the ink and the substrate on printability, electrical conductivity, as well as mechanical and electromechanical properties were thoroughly examined. The results revealed that the higher the AgNW content, the higher the electrical conductivity attained (highest conductivity of 6.5 S m⁻¹ at 80 wt% of AgNWs). In contrast, the piezoresistive sensitivity first increases with the AgNW content to 50 wt% and decreases thereafter. The printed samples display a constant change in resistance over 1000 cycles, proving their durability. Moreover, printed electronics are found to disintegrate in water at room temperature within one hour, making them an eco-friendly substitute for conventional non-biodegradable electronics. The potential of the printed samples in body motion detection, human–machine interface, and stretchable electrodes has been demonstrated, highlighting their applicability in flexible, stretchable electronics.

---

1. Introduction

Stretchable and flexible electronics are gaining huge attention for their remarkable potential in soft robotics and wearable devices.^1–4 Numerous stretchable sensors and electrodes with various designs have been created over the past few years.^5,6 Synthetic elastomers such as polyurethane, polydimethylsiloxane (PDMS), and conductive nanomaterials such as silver nanowires,⁷ carbon nanotubes,⁸ and graphene⁹ are commonly used to prepare these electronics. Because of their extraordinary flexibility and stretchability, these strain sensors can be seamlessly mounted on irregular surfaces like our body for real-time monitoring of health and body motion, or a human–machine interface.^10–12

The most recently developed stretchable electronics are manufactured using non-biodegradable synthetic polymers such as PDMS,¹³ polyamide,¹⁴ and styrene–ethylene–butylene–styrene.¹⁵ These non-biodegradable electronics pose a serious threat to the environment and our health. Therefore, several sustainable materials such as cellulose,¹⁶ chitosan,¹⁷ polylactic acid,¹⁸ and starch¹⁹ have been researched as alternative materials for making electronics. For example, Chen et al. produced a biodegradable and stretchable double-layer structured strain sensor composed of a porous graphene and gelatin composite combined with a fabric base.²⁰ The sensor has a detection range of 20% and a fast response time of 60 ms. More importantly, the sensor can completely disintegrate in moist soil at 25 °C within 4 days. Poulin et al. prepared conductive ink through the shellac-based solution in which carbon particles were dispersed.²¹ Shellac functions as a water-resistant binder to replace petroleum-based polymers. The ink contains carbon particles that enable electrical conductivity and function as rheology modifiers to form a printable shear-thinning gel. The sheet resistance and conductivity of the ink were measured at 1000 S m⁻¹ and 15 Ω sq⁻¹. In another research, Liu et al. created a wearable pressure sensor that consists of a sandwich design by printing a sensing layer of Ti₃C₂T_x MXene ink and sandwiching it between interdigitated electrodes and polyethylene terephthalate films.²² The sensor shows high sensitivity and can work across a broad pressure sensing range (up to 1200 kPa) while also being capable of degrading in water.

Cellulose, the most abundant natural polymer on earth, is commonly found in the cell walls of plants.^23–25 It is biodegradable, non-toxic, and renewable, making it an environmentally friendly alternative material for creating electronics. However, its high strength (4.9–7.5 GPa)²⁶ and modulus (100–200 GPa) are not desirable for stretchable electronics. Recently, it has been found that its mechanical properties can be modified by adding plasticizers,²⁷ monomer grafting, or blending with a new polymer.²⁸ For example, glycerol has been used as the plasticizer to enhance its stretchability and flexibility, and its strength (74 MPa), modulus (1402 MPa), and failure strain (7.2%) have been successfully tailored to 1 MPa, 0.3 MPa, and 328%, respectively.²⁹ Furthermore, acetylation treatment has been found to be able to increase its elasticity by adding acetyl groups, which can reduce the hydrogen bonding between cellulose molecules.³⁰ Additionally, blending cellulose with an elastomer has been shown to improve the elastic properties, which achieves an elastic stretchability of 200% strain.³¹ Building on these material advances, recent research efforts have focused on translating these improved mechanical properties into functional electronic devices through the development of conductive cellulose composites.^32,33 However, it is still challenging to develop fully disintegrable electronics via an effective and scalable fabrication process. Moreover, there is a lack of understanding of the interplay between material composition, electrical conductivity, mechanical stretchability, and piezoresistivity, which is critical to the development of soft electronics.

Herein, we developed disintegrable, stretchable, and flexible electronics (strain sensors and conductors) by creating sustainable inks printed onto a fully biodegradable substrate. The substrate is made of sodium carboxymethyl cellulose (NaCMC) modified by glycerol, while the ink is composed of silver nanowires (AgNWs) with the same materials as the substrate (i.e., NaCMC and glycerol) as the binder. NaCMC was selected as it is water-soluble and easy to process while glycerol was incorporated as a plasticizer to enhance its stretchability and flexibility. The binder for the ink is made of the same materials as the substrate, which can not only promote strong bonding between the substrate and ink but also ensure full sustainability without using non-biodegradable materials, a distinct advantage in addressing environmental sustainability and the growing challenge of electronic waste. The ink was then printed onto the substrate by a dispensing method. The effects of the material composition of the ink and substrate on the electrical conductivity, mechanical, and electromechanical properties of the printed samples were thoroughly investigated. Their potential applications in the detection of joint bending, respiration, human–machine interface (controlling a robot hand), and stretchable electrodes have been demonstrated. Finally, their disintegrability was studied by immersing them into water at room temperature.

2. Experimental section

2.1. Materials

Carboxymethyl cellulose (NaCMC, M_w ≈ 250 000), glycerol (99%, AR), polyvinylpyrrolidone (PVP, M_w ≈ 40 000), silver nitrate (AgNO₃, p.a., ≥99.5%, AT), and sodium chloride (NaCl, ACS reagent, ≥99.0%) were purchased from Sigma-Aldrich and used without further processing.

2.2. Synthesis of AgNWs

The polyol method was used to synthesize AgNWs, as described in our previous work.³⁴ Firstly, 5.86 g of PVP was dissolved in 190 mL of glycerol at 90 °C, followed by cooling to 50 °C. Then, 1.58 g of AgNO₃ was added to the above solution. At the same time, a solution containing 10 mL of glycerol, 59 mg of NaCl, and 0.5 mL of H₂O was prepared and decanted into the solution. The prepared mixture was then gently stirred (∼50 rpm) and heated up to 210 °C until a hot grey-green solution was obtained. The solution was then transferred to a beaker containing cold Milli-Q water and kept for 24 hours for the complete settling of AgNWs in the solution. AgNWs sedimented at the bottom were collected after removing the supernatant fluid. To further remove the PVP and Ag nanoparticles, the process of washing was repeated at least three times using Milli-Q water.

2.3. Ink and sensor fabrication

Fig. 1 illustrates the process of fabrication of the printed electronics, including the preparation of ink and substrate as well as the printing process.

Fig. 1 Schematic illustrating the fabrication of (a) AgNWs/NaCMC/glycerol ink and (b) NaCMC/glycerol substrate; (c) photos indicating the stretchability of the prepared substrate; (d) schematic of the printing method; (e) photo of a printed sensor, and (f) SEM of the printed conductive track.

The ink was prepared by first dissolving the required amount of NaCMC in deionized water at room temperature. After the complete dissolution of NaCMC, glycerol was added and mixed thoroughly to ensure homogeneity. The next step was introducing AgNWs (10%, 30%, 50%, and 80%) to the NaCMC/glycerol mixture. AgNWs were dispersed in water using probe sonication (Vibra-Cell™, Model VCX-750, Sonics & Materials, Inc., USA) for 5 minutes at 20% amplitude, followed by mixing the dispersed solution into an NaCMC/glycerol mixture. The stirring was done for an additional 30 minutes to ensure the uniform dispersion of AgNWs in the ink matrix. After determining the AgNW content (i.e., 50%) that gives the highest sensitivity, NaCMC/glycerol content was further adjusted to explore its effects on ink performance. Inks containing three different weight ratios of NaCMC to glycerol (NaCMC vs. glycerol: 25 : 75, 50 : 50, and 75 : 25) were prepared.

For substrate preparation, NaCMC was first dissolved in water, followed by the addition of glycerol. The mixture was thoroughly stirred and subsequently poured into molds to dry to obtain the flexible and stretchable film. Three different substrates with varying content of NaCMC and glycerol (NaCMC vs. glycerol: 40 : 60, 50 : 50, and 60 : 40) were prepared.

The specimens were created by using a microelectronic printer (MP-1100A) in dispensing mode, which is a cost-effective technique and enables printing with high precision. This approach is considered cost-effective and user-friendly as it eliminates the need for stencil masks for screen printing and cleanroom facilities for photolithography, which are conventional, commonly used techniques for creating electronic devices. Printing was carried out using a nozzle of 500 μm, a pressure of 1.5 kPa, and a printing speed of 1.5 mm s⁻¹. These parameters were chosen for uniform and consistent ink deposition onto the substrate. The printed sensors were then dried at room temperature for 1 hour before testing.

2.4. Characterization

Viscosities of the prepared ink with different material compositions were measured using a rheometer (Rheoplus – MCR302). Samples were tested at room temperature using a parallel plate (PP25) having a gap of 1 mm between the plates. The tensile properties of the prepared substrates with various amounts of NaCMC and glycerol were tested by using a universal tensile machine. Samples of 25 mm × 10 mm dimensions were sliced for the test at a speed of 1 mm s⁻¹ until broken. Tensile properties of the substrates were evaluated in both quasi-static tension and cyclic tension loading–unloading. Morphological and structural analysis of the AgNWs and printed inks were performed using scanning electron microscopy (SEM – Zeiss Ultra), operating at 5 kV.

The electromechanical properties of the printed samples were analyzed by measuring the change in resistance using a digital multimeter (Keysight E4980 AL) when subjected to a tensile strain. For this purpose, a specimen with a dimension of 25 mm (length) × 10 mm (width) was prepared and tested at a speed of 1 mm s⁻¹. Electromechanical performance was evaluated in both quasi-static tension and cyclic tension loading–unloading. Five cycles were tested with a peak strain ranging from 1% to 20% at a constant speed of 1 mm s⁻¹. To better understand the performance and repeatability, each test was conducted on at least 4 samples. Piezoresistive sensitivity, which is usually measured in terms of gauge factor, is defined by the formula given in eqn (1):

ε is the strain, R and R₀ are the resistance at strain ε and initial resistance, respectively. Samples were also subjected to cyclic tension for 1000 cycles with peak strains of 10% and 20%. The effect of different strain rates was also analyzed by subjecting the specimen to a tensile strain of 10% at strain rates of 0.05 s⁻¹, 0.1 s⁻¹, 0.25 s⁻¹, and 0.5 s⁻¹. The response time was determined by quickly stretching the sensor to 20% strain at a speed of 5 mm s⁻¹ and maintaining this strain for 30 seconds. The sensor was then released back to 0% strain at the same rate and held there for 30 seconds before beginning the subsequent cycle.

The potential use of printed electronics in human body motion detection and human–machine interface was examined. The sensor was first attached to the joint of the wrist and knee, and the resistance changes were recorded during the bending of the joints. Furthermore, the sensor was integrated with a glove, which was then used to control the movement of a robot hand. The resistance changes were used as the input to a microcontroller that can manipulate the robot hand. The voltage signal is passed through a potential divider circuit that used the sensor as one resistor and a 100 kΩ resistor as the second, while operating on a 5 V source voltage. An Arduino UNO R3 microcontroller board with its 10-bit analog-to-digital converter (ADC) received the output voltage from the divider circuit on the pin A7. The Arduino IDE with the DOBOT Arduino library was used to program the microcontroller to convert voltage signals into movement commands. Serial communication (RX/TX pins) transmitted these commands to the robotic arm. Additionally, the sensor has been used to create a chest band, which was attached to the chest for monitoring respiration. To examine the printed specimen's potential as a stretchable electrode, ink was printed onto the cellulose substrate in the form of a straight line, and an LED was attached at the center using conductive adhesive. The illumination of the LED was monitored while stretching it for 100 cycles and bending the substrate. The electrode's stretchability was also assessed by attaching it to the surface of an uninflated balloon. The balloon was then inflated gradually, causing the electrode to stretch along with the balloon. The illumination of the LED was examined and recorded by taking pictures before and after balloon inflation.

The disintegration of the printed sensors/conductors was evaluated by immersing them in water at room temperature. A specimen of 25 mm (length) × 10 mm (width) was placed in a Petri dish containing 10 mL of deionized water. Agitation was provided using an ultrasonicator for 10 seconds every 10 minutes to promote disintegration. A camera was used to capture the change of the degraded sample at a time interval of 5 minutes. The disintegrated solution was allowed to stratify after complete dissolution, and the upper supernatant was removed. Ethanol was then added, and the stratification and separation process were repeated several times to obtain a uniform dispersion of recycled AgNWs.

3. Results and discussion

3.1. Design of the sensors/electrodes

In this work, the sensors/electrodes were designed to achieve high stretchability and disintegrability by using green components, including cellulose, glycerol, and conductive silver nanowires. The substrate consists of cellulose functionalized with glycerol, while the ink uses cellulose and glycerol as the binder and dispersant. This material design is expected to achieve robust mechanical stretchability and excellent bonding between the conductive ink and the substrate. The conductive inks were printed in a U-shaped track onto the flexible substrate, as seen in Fig. 1(e). This geometry was chosen for its functionality and simplicity, allowing effective strain sensing and conductivity measurement. The ends of the printed patterns are connected to widened rectangular pads that assist in securing electrical connections for testing. The use of a flexible substrate enables applications where stretching and bending frequently occur while it is necessary to maintain electrical performance and structural integrity.

3.2. Mechanical properties of the biodegradable substrate

The mechanical characteristics of the substrate used in the printed electronics are important as it determines the mechanical properties of the electronics. This section presents the tensile properties of the cellulose substrate modified with different glycerol contents. Fig. 2(a) presents the tensile stress–strain curves for cellulose/glycerol blends with 40 wt%, 50 wt%, and 60 wt% of glycerol. Young's modulus, tensile strength, and elongation at break calculated from Fig. 2(a) data are shown in Fig. 2(b), (c) and (d), respectively. The blend with a high content of NaCMC, i.e., cellulose – 60% and glycerol – 40%, shows the highest modulus (111.0 MPa) and tensile strength (48.9 MPa), whereas elongation (66%) is the lowest. As the amount of glycerol increases to 50%, the strength and modulus decrease to 21.4 MPa and 44.0 MPa, respectively, while the stretching ability of the substrate increases, resulting in 77% of elongation at break. With a further increase in glycerol content to 60%, the stretchability of the blend film continues to increase while the stiffness reduces. The modulus, strength, and elongation of the film with 60% glycerol are 9.0 MPa, 10.6 MPa, and 107%. The presence of glycerol in the NaCMC substrate may disrupt the hydrogen bonding between polymer chains, resulting in higher chain mobility. This disruption of hydrogen bonding, therefore, enhances the stretching ability of the substrate but decreases the strength and modulus.²⁹

Fig. 2 Tensile properties of NaCMC/glycerol films: (a) quasi-static tensile stress–strain curves for the blends made of NaCMC and glycerol with varying weight ratios and (b) Young's modulus, (c) tensile strength, and (d) elongation at break (%) extracted from (a). (e) Hysteresis in tensile stress–strain curves under cyclic loading for blends made of different contents of NaCMC and glycerol at peak strains of 10%, 20%, and 40%; typical stress–strain curves under cyclic loading (f) with peak strains of 10%, 20%, and 40%; (g) 10^th, 30^th, 60^th, 90^th, and 100^th cycle with peak strain of 20%.

Stretchable electronics require material components that demonstrate mechanical stability and quick recovery when subjected to cyclic loads. The cyclic tensile properties of the NaCMC/glycerol substrates are presented in Fig. 2(e–g). Fig. 2(e) represents the hysteresis for three types of substrates with varying NaCMC and glycerol content at maximum strains of 10%, 20%, and 40% for the first cycle. The hysteresis was calculated using the formula given in eqn (2):³⁵

In this context, Width refers to the horizontal distance between the loading and unloading curves measured at the midpoint of the Y-axis, while Range represents the total span of the X-axis covered by both loading and unloading curves. It is noted that the hysteresis behavior is influenced by the NaCMC and glycerol content with higher glycerol content resulting in higher hysteresis. The glycerol acts as the plasticizer by interrupting the strong and abundant hydrogen bonds between cellulose molecular chains. A lower content of glycerol means less hydrogen bonds are interrupted. So, strong intra- and inter-molecular hydrogen bonds within NaCMC result in a rigid polymer network with lower free volume for polymer chains to move, restricting the reorganization of chains during loading and unloading. Lower hysteresis is thus observed. In contrast, a higher content of glycerol (NaCMC : glycerol = 40 : 60) results in high free volume for cellulose polymer chains to rearrange, allowing more molecular slippage and thereby causing higher hysteresis. Glycerol affects hysteresis by enhancing molecular mobility and reorganization during cyclic loading.³⁶ The substrates containing the highest glycerol content (60%) exhibited the largest hysteresis at all strain levels. Conversely, the substrates containing the lowest glycerol content (40%) are more rigid and demonstrate the least hysteresis. It is also found, from Fig. 2(f), that the substrate shows higher hysteresis when it is stretched to a higher strain. The substrate showed a minimal hysteresis of ∼25% during 10% strain loading. At a strain of 20%, a hysteresis of ∼28% was observed. When the strain increases to a higher strain of 40%, this increased to ∼33%. This indicates that the material undergoes significant molecular slippage,²⁹ when subjected to such a high strain. Fig. 2(g) displays stress–strain curves of the blend containing 40% glycerol in the 10^th, 30^th, 60^th, 90^th, and 100^th cycle with a peak strain of 20%. The NaCMC/glycerol film shows a slight decrease in mechanical hysteresis over its 100 loading/unloading cycles, indicating that the polymer network gradually stabilizes.

3.3. Printability, electrical conductivity, and electromechanical properties of the printed electronics

3.3.1. Effects of AgNW content. The geometry of prepared nanowires was analyzed by using SEM. Based on Fig. 3(a), the AgNWs have an average diameter of approximately 100 nm, while their average length is around 10 μm. The graphs showing the diameter and length distribution of nanowires are presented in Fig. 3(b), indicating that the nanowires possess a uniform diameter. As indicated in Fig. 3(c), the viscosity of the prepared inks decreases with an increase in shear rate, showing a typical shear-thinning behavior. This behavior allows the ink to flow smoothly under pressure during printing while quickly regaining higher viscosity after deposition, which helps in maintaining the pattern shape and prevents uncontrollable spreading.³⁷ Moreover, the content of AgNWs seems to have minimal effects on the viscosity, with similar viscosity being seen in the ink containing 10%, 30%, and 50% AgNWs. However, upon further increasing the AgNW content to 80%, very high viscosity was observed, making it difficult to print due to the risk of nozzle clogging and uneven deposition.⁷

Fig. 3 (a) SEM image of the fabricated AgNWs; (b) frequency histogram showing the length and diameter distribution of AgNWs; (c) and (d) viscosity of ink with varying content of AgNWs. (e) Electrical conductivity, (f) relative resistance changes, (g) GF, and (h) failure strain of the sensors printed from the ink containing varying contents of AgNWs.

The electrical conductivity of the printed specimens was measured, and it is noted that it increases significantly with increasing the AgNW content based on Fig. 3(e). The ink containing 10% AgNWs showed the lowest conductivity (0.001 S m⁻¹). As the nanowires' content increases from 10% to 30%, 50%, and 80%, the conductivity value reaches 0.2 S m⁻¹, 4.4 S m⁻¹, and 6.5 S m⁻¹, respectively. A higher content of AgNWs means a larger number of nanowires present to form more conductive pathways, resulting in a decrease in the overall resistance.³⁸Fig. 3(f) presents the relative resistance changes under tensile strain for the printed specimens with different contents of AgNWs, while Fig. 3(g) shows the gauge factor calculated by linear fitting. When subjected to tensile stretching, all specimens show a rise in resistance, but the degree of resistance increase depends on the content of AgNWs. The sample made from the ink containing 10% AgNWs showed a small change in resistance upon mechanical deformation with a gauge factor of 0.2. Upon increasing the content of AgNWs up to 50%, the change in resistance increases with a gauge factor rising to 0.57. Further increasing the AgNW content to 80% results in a drop in sensitivity (gauge factor ∼ 0.06). The electrical conductivity in the printed track is believed to arise from the formation of conductive pathways by AgNWs. The resistance changes when subjected to tensile strain is likely due to the rearrangement of AgNWs in the printed pattern, which will be discussed in Section 3.3.4. The excessive AgNWs form very dense conductive pathways with nanowires closely packed, which makes it hard to be disrupted and thereby no significant change in resistance upon stretching.³⁹ The sensors printed with different contents of AgNWs also showed excellent linearity with an R² > 0.94, as shown in Fig. 3(f).

All the printed samples show similar failure strain values with minimal variation, as shown in Fig. 3(h). The failure strain for each ink type was calculated as the average value obtained from five independently tested samples. The failure strain here means the mechanical failure strain, i.e., elongation at break, which is predominantly controlled by the substrate. No electric failure (i.e., resistivity increase is so high that the resistance exceeds the capacity of the measuring equipment used in the test) was observed before the mechanical failure, suggesting that the conductive ink developed in this work can be used to create stretchable electronics.

3.3.2. Effects of the ratio of NaCMC to glycerol in the ink. Apart from the AgNW content, NaCMC and glycerol in the ink also play a crucial role in determining the properties of the ink and printed specimens. NaCMC and glycerol not only facilitate the uniform dispersion of AgNWs for creating stable ink but also affect the viscosity and thereby printability as well as the bonding with the substrate. Building on the results discussed above (Section 3.3.1), to study the effects of NaCMC and glycerol content, the AgNW content was kept constant at 50 wt% while the weight ratio of NaCMC to glycerol varied from 25 : 75, 50 : 50, to 75 : 25. It should be noted that the water content is ∼55 wt% for all the inks and kept constant unless otherwise stated. The viscosities of the three inks with varying NaCMC and glycerol content are shown in Fig. 4(a) and (b). All inks exhibit the same behavior, i.e., viscosity decreases with an increase in shear rate. Moreover, the ink containing a higher amount of glycerol (lower NaCMC content) shows a lower viscosity. This may be due to the plasticizing effect of glycerol that disrupts the intermolecular interactions between cellulosic chains.²⁹

Fig. 4 (a) and (b) Viscosity of the inks with varying content of NaCMC and glycerol. (c) Conductivity, (d) relative resistance changes, (e) gauge factor (GF), and (f) failure strain of the printed specimens made from ink containing varying amounts of NaCMC and glycerol.

The dependence of conductivity on the NaCMC to glycerol ratio is shown in Fig. 4(c). The ink containing the highest glycerol content, i.e., 75%, displayed the highest conductivity, which is due to reduced viscosity and higher flowability attributed to the presence of glycerol. Glycerol-rich ink with lower viscosity compared to the ink with high NaCMC content likely promotes better dispersion and alignment of the nanowires.⁴⁰ Highly aligned and uniformly dispersed nanowires are expected to form a larger number of conductive pathways, resulting in higher electrical conductivity. Fig. 4(d) and (e) show the influence of NaCMC to glycerol ratio on the piezoresistivity of the printed conductive tracks. Results of are mirrored in GF vs. NaCMC/glycerol, with the highest GF of 0.57 noted for the specimen printed from the ink containing the highest glycerol content, while the lowest GF of 0.28 is obtained for the specimen made of ink with the highest NaCMC content. Both NaCMC and glycerol played a crucial role in the overall properties of the printed specimens. NaCMC provides structural integrity while glycerol acts as a lubricant and offers flexibility to the printed patterns. High sensitivity at higher glycerol concentrations of 75% may be attributed to the freer movement of AgNWs when subjected to mechanical deformation. The sensor made with ink containing 75%, 50%, and 25% glycerol demonstrated a highly linear response, achieving an R² value of 0.98 and 0.91, respectively, as illustrated in Fig. 4(d). Furthermore, the failure strain is observed to be the same for all printed specimens, determined by the substrate mechanical characteristics. This also suggests that the piezoresistive sensitivity and conductivity of the printed conductive materials depend strongly on the ink composition (i.e., AgNWs, NaCMC, and glycerol content) while the overall stretchability is controlled by the substrate.

3.3.3. Effects of the ratio of NaCMC to glycerol in the substrate. The dependence of substrate composition on the electrical conductivity and electromechanical properties of the printed specimens was also examined by using substrates of three different compositions (NaCMC vs. glycerol: 40 : 60, 50 : 50, and 60 : 40). Interestingly, it is noted that, even using the same conductive ink, the conductivity of the printed pattern is different, that is, the sample made of the substrate containing 60% glycerol displayed a conductivity of 0.25 S m⁻¹, which is much lower than that made of the substrate having 40% glycerol (4.46 S m⁻¹) (Fig. 5(a)) and the higher the glycerol content, the lower the conductivity. The structure of the printed track was investigated by using an optical microscope, and, based on the micrographs shown in Fig. 5(b–d), it is found that the width of the printed track varies when the substrate is different. The width of the printed track is ∼840 μm for a sample made of a substrate having 60% glycerol, which reduces to ∼580 μm when the glycerol content reduces to 40% in the substrate. This indicates that the ink has different wettability and flowability onto substrates having different compositions. The ink can spread more easily onto the substrate with a higher content of glycerol, leading to a thinner and wider printed track. Meanwhile, nanowires flow and spread with the ink when it is printed onto the substrate and are expected to aggregate during this process, resulting in lower conductivity.⁴¹

Fig. 5 (a) Conductivity of the printed track on substrates of different compositions, i.e., the ratio of NaCMC to glycerol is 40 : 60, 50 : 50, and 60 : 40. Optical images showing the width of the printed track on substrates with the ratio of NaCMC to glycerol (b) 40 : 60, (c) 50 : 50, and (d) 60 : 40. (e) Relative resistance changes and (f) GF of the printed specimens fabricated using substrates containing different contents of NaCMC and glycerol.

The piezoresistivity of the printed specimens is shown in Fig. 5(e) and (f). The specimens made of the substrate with the highest glycerol content (60 wt%) are highly stretchable with the highest failure strain of ∼105%. The printed specimen on this substrate also displayed the highest change in resistance upon stretching. This indicates more profound deformation in the conductive network of AgNWs. The substrate possesses lower stiffness based on the results in Fig. 2(a), while AgNWs are much stiffer as compared to the elastic polymer substrate. The mismatch in stiffness between the top conductive AgNW layer and the bottom elastic polymer layer is increased for a substrate having higher glycerol content. Therefore, stress transfer efficiency from the bottom polymer layer to AgNWs should be lowered, resulting in lower changes in the AgNW conductive network and thereby lower piezoresistivity, in good agreement with the performance of the samples consisting of 50 wt% and 40 wt% glycerol. The sensor containing 40% glycerol also exhibited an excellent linear response to the strain, with an R² value of 0.96, while the specimens made of 50% and 60% glycerol show R² values of 0.89 and 0.91, respectively, as shown in Fig. 5(e). This indicates excellent linearity. The substrate made of 60 wt% glycerol, however, showed the highest piezoresistivity, which is an unusual behavior. The very soft, sticky substrate (with 60% glycerol) might have glycerol-rich domains that locally deform in a non-uniform way, which could lead to localized strain concentration at the AgNW network, enhancing the disruption of the conductive paths. It should be noted that high sensitivity in a 60% glycerol substrate comes at the expense of the durability of the sensor because the stickiness and low mechanical properties make the sensor less useable for long periods. Even with a lower sensitivity, the 40% glycerol sensor was chosen for further analysis as it shows a good balance between sensitivity and durability.

3.3.4. Piezoresistivity under cyclic loading and the working mechanism. The piezoresistivity of the printed tracks was also studied by applying cyclic strain of different levels and strain rates, and the results are shown in Fig. 6. Fig. 6(a) shows relative resistance changes when subjected to five cyclic strains of different peak strains (1%, 5%, 10%, and 20%). The resistance increases upon increasing the strain and decreases when reducing the strain. The inset in Fig. 6(a) indicates that the sensor can detect minute strains as low as 1%. As the peak strain increases to 10% and 20%, a more significant change in resistance occurs. A maximum of ∼6% and ∼15% was measured at 10% and 20% peak strain, respectively. The repeatability of under multiple stretching cycles indicates stable piezoresistive performance. Fig. 6(b) displays the corresponding versus strain curves (with the peak strain of 1%, 10%, and 20%), indicating some hysteresis. The hysteresis is likely due to the viscoelastic properties of the substrate, resulting in delay in the nanowire network destruction and reformation during stretching and releasing.⁴² The hysteresis for 20% strain was found to be ∼5% (the hysteresis was determined by a method reported in our previous work^35,43) and it indicates a more prominent change in the conductive network, resulting in significant structural changes. High levels of strain, i.e., 20%, also produce some permanent deformation among the conductive network, which means some of the nanowires do not come back to their initial position, explaining a difference between the initial resistance (∼15 Ω) and the final resistance (∼16 Ω) upon release.

Fig. 6 Piezoresistive properties of the printed conductive tracks: (a) relative resistance change when subjected to cyclic stretching–releasing cycles with different peak strains, (b) the corresponding relative resistance change versus strain curves for one cycle with peak strains of 1%, 10%, and 20%, (c) relative resistance change under five cycles with peak strain of 10% but at different strain rates, (d) response time and recovery time of the strain sensor tested at 20% strain, (e) and (f) relative resistance changes when subjected to 1000 cycles of 10% and 20% strain.

Fig. 6(c) shows the changes in resistance when subjected to 10% applied strain at different strain rates, i.e., 0.05 s⁻¹, 0.1 s⁻¹, 0.25 s⁻¹, and 0.5 s⁻¹. No noticeable difference in values was observed as the displacement rate increased. Fig. 6(d) shows the relative resistance change of the sensor during loading and unloading at 20% strain. The sensor exhibits a fast response time of 0.17 s during stretching and a relaxation time of 0.24 s during release (as highlighted in the insets). However, it should be noted that, due to the intrinsic viscoelasticity of the polymeric materials, it takes some time for the conductive network to respond to the external loading.⁴⁴ Moreover, to check durability, the printed samples were tested at 10% and 20% for 1000 cycles, as shown in Fig. 6(e) and (f). It is noted that an upward drift in the baseline resistance was observed, which may be due to the irreversible microstructural changes that occurred in the conductive network and gradual relaxation of the cellulose–glycerol substrate under cyclic mechanical deformation. This finding has been observed in some previously reported sensors based on viscoelastic conductive polymer nanocomposites,⁴⁵ which indicates that further optimization is required to improve long-term stability and minimize signal drift. The unstretched samples and the samples tested at 10% and 20% strain after 500 and 1000 cycles were also analyzed by using SEM to examine the structural changes for a better understanding of the cyclic behavior. Fig. 7(a) shows the printed electrode before exposure to a tensile stretch. It is noted that AgNWs are highly aligned and no wrinkle was observed. After being tested at 10% strain for 1000 cycles, the AgNW structure remains almost the same, suggesting the stability of the printed track (Fig. 7(b)) which can recover to its initial state after releasing the load. In contrast, sensors tested at 20% strain for 500 cycles showed apparent wrinkles (Fig. 7(c)), which are also observed after 1000 cycles (Fig. 7(d)). The nanowires' alignment seems to be shifted and becomes less aligned. The shift in nanowire alignment and variation in the interparticle spacing (observed in Fig. 6) are likely the reason for the drift in the resistance change vs. cycle curve.

Fig. 7 SEM images of the printed tracks (a) in an unstretched state, (b) after 1000 cycles of stretching–releasing at 10% strain, (c) after 500 and (d) 1000 cycles of stretching–releasing at 20% strain.

Table 1 compares the performance of the as-developed sensors and conductors with some recently reported flexible and stretchable disintegrable electronics, mostly based on cellulose, gluten, guar gum, silk, gelatine, etc. The sensitivity (gauge factor), stretchability (indicated as elongation at break), and conductivity are listed. It is found that biomaterials have been widely used to develop hydrogel-based sensors, generally showing much higher stretchability. However, hydrogels are inherently vulnerable to dehydration, which can compromise their structural integrity and functionality. Although the measured conductivity and GF are somewhat lower than some previously reported high-performance sensors listed in Table 1, the printed electronics created in this study achieves stable performance with a favourable balance of gauge factor, conductivity, and stretchability, which can be tuned by varying the material composition.

Table 1 Summary of the properties of some recently reported disintegrable sensors and conductors

Category	Materials	Gauge factor	Conductivity	Elongation at break	Ref.
Ionically conductive hydrogel	Cellulose/acrylic acid/acrylamide	2.5	5.3 S m^−1	196%	46
	Cellulose/polydopamine/NaOH	3.5	0.35 S m^−1	100%	47
	Cellulose/NaCl	2.53	1.9 S m^−1	1000%	48
	Gluten/guar gum/NaCl	0.61	0.34 S m^−1	665%	49
	Cellulose/Fe^3+	1.9	—	300%	50
	Cellulose/LiCl	3.15	—	300%	51
	Cellulose/NaOH/sodium thiosulfate	—	0.40 S m^−1	124%	52
	Cellulose/polyacrylamide/Fe^3+	1.4	—	30%	53
	Cellulose/LiCl	1.63	2.89 S m^−1	600%	54
	Guar gum/NaCl	8.2	0.2 S m^−1	400%	55
	Cellulose/NaCl	440	—	158%	56
	Cellulose/K3[Fe(CN)6]	—	10.3 S m^−1	80%	57
Electronically conductive hydrogel	Cellulose/CNTs	6.86	2.88 S m^−1	400%	58
	Sodium alginate/MXene	1.54	0.49 S m^−1	100%	59
	Cellulose/pyrrole/acrylamide	—	11.07 S m^−1	700%	60
	PVA/AgNWs	0.34	1.85 S m^−1	900%	61
	Cellulose/CNTs	49.5	3.3 S m^−1	99%	62
	Gelatin/PANI/AgNWs	1.29	2.2 S m^−1	200%	63
	Silk fibroin/MXene	6.04	0.85 S m^−1	500%	64
	Cellulose/Ag particles	0.37	—	800%	65
	Sodium alginate/PEDOT:PSS	0.39	256 S m^−1	334%	66
Electronically conductive thin film	Cellulose/AgNWs	0.24	—	1.5%	67
	Polyurethane/MWCNTs	1.6	19 S m^−1	50%	68
	Cellulose/gold nanoparticles	0.99	—	200%	69
Printed conductive layer on thin film	Cellulose/glycerol/AgNWs	0.57	4.46 S m^−1	52%	Present work
		0.06	6.5 S m^−1	56%

---

The mechanism of the piezoresistive strain sensor is believed to be based on the modulation of the conductive pathways formed by AgNWs. When subjected to mechanical deformation, conductive pathways of AgNWs are expected to be altered, including the changes in spacing, connectivity, contact area, and alignment. The structural changes in AgNWs during the application of tensile strain were observed by using an optical microscope. The optical images and schematic illustrations are presented in Fig. 8 to qualitatively explain the sensing mechanism responsible for resistance change under strain. As the mechanical stretching is applied, the interparticle spacing between AgNWs increases (some examples are marked in Fig. 8(c)), resulting in reduced contacts, hence increasing the overall resistivity. Moreover, the alignment of AgNWs also changes. For example, the two AgNWs which are initially aligned at ∼45° to each other (marked in Fig. 8(a)) reorient to ∼50° (Fig. 8(c)) after the sample is subjected to 10% strain. Upon releasing the applied strain, connections between the nanowires form again, returning mostly to their initial state, indicated by the recovery to R₀, as shown in Fig. 8(e) and (f). These cyclic changes in the AgNW conductive network result in an increase and decrease in the printed sample's resistance, which is the basic principle of piezoresistivity. It should be noted that the formation of conductive networks in polymer nanocomposites and their variation when subjected to mechanical deformation are significantly influenced by the geometry (shape and dimension) of the conductive nanofillers.⁷⁰ It's widely recognized that a higher aspect ratio gives a lower percolation threshold and thereby higher electrical conductivity at the same content of nanofillers.^71,72 On the other hand, it is easier to induce electrical disconnection or variation in contact area by external strain for nanocomposites with fillers of lower aspect ratio, which means higher sensitivity is typically observed.⁷³

Fig. 8 Optical microscopy images and schematic illustrating the structural changes in the printed conductive tracks at (a and b) 0% strain – unstretched, (c and d) 10% strain – stretched, and (e and f) 0% strain – the load released.

3.4. Application demonstration

The potential applications of the printed conductive tracks as sensors in the human–machine interface, body motion monitoring, and as stretchable electrodes are examined and discussed in this section. Firstly, the printed conductive specimen was integrated with a glove which can then be used to control the movement of robotic fingers by putting on the glove and bending the fingers. The change in resistance upon bending the finger joint was used as the input to control the bending of robot fingers. It is found from Fig. 9(a–c) that the robot's finger can replicate the motion of human fingers. When human finger joints bent from 0°, 45°, to 90°, the relative resistance change was ∼14% and ∼21%, corresponding to 19% and 37% strain, nearly consistent with those previously reported.⁴⁴ The relative resistance changes remained consistent for multiple bending cycles (Fig. 9(d and e)). This application demonstrates the printed sensors' potential for human–machine interface and soft robotics. The working of the sensor can be observed in the SI (Video S1).

Fig. 9 Illustration of the sensor's application in controlling a robot hand by integrating it to a glove to create a smart glove: (a–c) a robot hand copying the bending of an index finger at 0°, 45°, and 90°, and (d and e) the corresponding changes in resistance. Sensor's potential in detecting (f) wrist and (g) knee joint movement and (h and i) respiration pattern.

Furthermore, the sensor's capability of detecting wrist bending was tested (Video S2 – SI). The sensor was placed on the wrist while the changes in resistance were measured during wrist bending, as shown in Fig. 9(f). Resistance rises during bending from 0° to 45° with of ∼14% and returns to the initial value while bending was recovered. was recorded for multiple cycles of wrist bending, showing consistent signal outputs. Similarly, the sensor's capability of detecting knee-bending motions was tested. The sensor placed above the knee displayed noticeable changes in its resistance during knee movement from straight to bent positions, as shown in Fig. 9(g). The sensor demonstrated consistent and repeatable resistance changes that proved its ability to detect knee movement strains. The sensor's capacity to track real-time motions of human joints is expected to enable various applications, e.g., hand gesture analysis, gait monitoring, and rehabilitation exercise monitoring. Moreover, the sensor's ability to detect small strains associated with localized skin deformation during breath was evaluated by placing it on the torso (Video S3 – SI). The recorded data in Fig. 9(h) show that periodic variations in relative resistance change upon inhalation and exhalation. The breathing pattern before and after doing exercises was compared and a lower frequency was noted before doing any running (breath per minute is ∼15 bpm, consistent with those reported in the literature⁷⁴). After running as shown in Fig. 9(i), the sensor indicates a substantial increase in the breathing frequency to 47 bpm because the body requires more oxygen and breathes faster. The sensor demonstrates its ability to monitor respiration continuously.

As discussed above, the electrical conductivity and piezoresistivity can be tuned by varying the material composition. When the ink consists of 80% AgNWs, the resulting specimen shows very high electrical conductivity (6.5 S cm⁻¹) with a low gauge factor (0.06). This makes it desirable for the application of stretchable conductors, which was tested by integrating it with different circuit designs to power LED lights. Fig. 10(a–e) show a circuit integrated with an LED light while the printed electrode was unstretched, stretched to 10% and 35% strain, respectively. It is clearly seen that no visible change in illumination occurs even after being stretched to 35% of strain. The durability of the electrode under repeated stretching was also evaluated by repeatedly stretching it for 100 cycles while connecting to an LED. The LED consistently maintained its brightness, confirming stable electrical performance and good mechanical endurance. The durability of the circuit under bending was also studied by directly bending it by hand (Fig. 10(f)). Moreover, the circuit was integrated onto a balloon surface (Fig. 10(g and h)). No physical failure was observed and the LED still lighted up even when the balloon is inflated, which induces significant biaxial deformation. This demonstrates the printed electronics' ability to conform to irregular surfaces without losing their functionality, which is important for stretchable and flexible electronics.

Fig. 10 Demonstration of the printed specimens as stretchable electrodes: (a–c) in lighting an LED light when (a) unstretched, stretched to (b) 10%, (c) 35% of tensile strain; stretched to 35% tensile strain for (d) 10 cycles and (e) for 100 cycles; (f) subjected to bending, and (g and h) on irregular surfaces.

3.5. Disintegration behavior

The key components used to fabricate the sensor are NaCMC, glycerol, and AgNWs. So, it is expected that the printed electronics can be disintegrated in water as a water-soluble matrix, i.e., NaCMC/glycerol, can be dissolved in water, while conductive reinforcement can be collected and reused. This was checked by immersing the printed electronics in deionized water at room temperature, with periodic agitation. The sample starts to disintegrate quickly after around 2 minutes and completely disintegrates within 1 hour, as shown in Fig. 11. The disintegration of the printed electronics primarily involves dissolving the components, i.e., NaCMC and glycerol, which are also biodegradable. Glycerol is a major component of the substrate and readily dissolves in water, leaving behind the cellulose and conductive network of AgNWs. NaCMC, on the other hand, swells and dissolves in water but takes a longer time than glycerol. As the matrix dissolves in water, its ability to support the AgNW network diminishes, resulting in the redistribution of the conductive network. The repeated agitation every 10 minutes played an important role in speeding up the degradation process. The provided mechanical force allows water to penetrate deep into the sensor in less time, enhancing glycerol dissolution and NaCMC swelling. Agitation also produces mechanical stress on the AgNW network, allowing them to detach more rapidly. The AgNWs were collected and tested by using SEM. Fig. 11(g) shows the SEM of the AgNWs, while Fig. 11(h) and (i) give the length distribution and EDX analysis of freshly made and recycled AgNWs. Elemental analysis revealed ∼90% Ag and ∼10% O, compared to ∼70% Ag and ∼30% O in the freshly made AgNWs. This indicates effective removal of surface-adsorbed oxygen-containing compounds, likely due to the dissolution process that involves ultrasonication.⁷⁵ In summary, it is found that the length of AgNWs is slightly reduced after one cycle of usage and disintegration, while their purity improves.⁷⁶

Fig. 11 Disintegration behavior of the printed electronics over a period of 1 hour in deionized water at room temperature. Digital photos of a printed specimen immersed in the water for (a) 0, (b) 2.5, (c) 5, (d) 15, (e) 30, and (f) 60 minutes. (g) SEM image of recycled AgNWs, (h) frequency histogram showing the length distribution, and (i) EDX of freshly made and recycled AgNWs.

4. Conclusions

In conclusion, printed electronics made of sustainable materials have been developed. Biodegradable materials, NaCMC and glycerol, were chosen as the main components of the substrate and as the dispersant as well as the binder for the AgNW ink. The work studied the effects of the material composition of both the ink and substrate on the electrical conductivity, stretchability, and piezoresistivity. It is found that the incorporation of glycerol disrupts hydrogen bonding between NaCMC chains, increasing free volume and enhancing polymer chain mobility. As a result, the mechanical strength and modulus are reduced, while the elongation at break is improved, enabling the creation of cellulose thin film with high elastic stretchability. Stretchable sensors and conductors were fabricated by depositing the AgNWs/NaCMC/glycerol-based ink onto the cellulose elastic substrate. The printed specimen utilizing ink composed of 50% AgNWs, with the remaining 50% consisting of 25% NaCMC and 75% glycerol, achieved a maximum gauge factor of 0.57, a stretchability of up to 50% strain, and excellent repeatability over 1000 stretching–releasing cycles. The ink formulation with 80% AgNWs exhibited the highest conductivity (6.5 S m⁻¹), making it the most suitable candidate for stretchable electrodes. In addition to mechanical and electrical performance, the sensor showed excellent disintegrability in water, fully dissolving in water within one hour at room-temperature. The sensor demonstrates great potential in monitoring various human motions, including wrist and knee joint bending, as well as respiratory patterns. It also demonstrated the ability to recognize hand gestures, underscoring its potential for human–machine interface applications. The stretchable printed electrode successfully powered LEDs during mechanical deformation and on irregular surfaces, confirming its practical utility. Overall, the combination of the sensor's disintegrability, mechanical flexibility, and reliable piezoresistive performance makes it as a strong candidate for applications in wearable electronics, electronic skin, sports performance tracking, and health monitoring systems.

Ethical statement

All subjects gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with Research Ethics and Compliance Support (RECS), University of New South Wales, Sydney, NSW, Australia.

Conflicts of interest

The authors declare no conflict of interest.

Data availability

The data supporting this article (Videos demonstrate the potential applications of the sensors in detecting wrist bending and respiration, as well as enabling a human–machine interface to control a robotic hand) have been included as part of the SI. See DOI: https://doi.org/10.1039/d5ta06546a.

Acknowledgements

The authors S. Wu and S. Peng would like to acknowledge the research funding awarded by ARC through the ARC Discovery Project (DP240101993, DP250100714), ARC Future Fellowship (FT220100094), ARC Industrial Transformation Research Hubs (IH210100040).

References

1. X. Fu, L. Li, S. Chen, H. Xu, J. Li, V. Shulga and W. Han, J. Colloid Interface Sci., 2021, 604, 643–649.
2. Y. Khan, A. Thielens, S. Muin, J. Ting, C. Baumbauer and A. C. Arias, Adv. Mater., 2020, 32, 1905279.
3. W. Cheng, Z. Zheng, X. Li, Y. Zhu, S. Zeng, D. Zhao and H. Yu, Research, 2024, 7, 0305.
4. H. Yu, J. Bian, F. Chen, K. Li and Y. Huang, Research, 2024, 7, 0383.
5. X. Chang, L. Chen, J. Chen, Y. Zhu and Z. Guo, Adv. Compos. Hybrid Mater., 2021, 4, 435–450.
6. S. Peng, S. Wu, Y. Yu, Z. Sha, G. Li, T. T. Hoang, M. T. Thai, T. N. Do, D. Chu and C. H. Wang, J. Mater. Chem. A, 2021, 9, 26788–26799.
7. S. Hemmati, D. P. Barkey and N. Gupta, J. Nanopart. Res., 2016, 18, 1–11.
8. X. Chen, H. Liu, Y. Zheng, Y. Zhai, X. Liu, C. Liu, L. Mi, Z. Guo and C. Shen, ACS Appl. Mater. Interfaces, 2019, 11, 42594–42606.
9. S. Wu, S. Peng, Z. J. Han, H. Zhu and C. H. Wang, ACS Appl. Mater. Interfaces, 2018, 10, 36312–36322.
10. S. Wang, H. Cheng, B. Yao, H. He, L. Zhang, S. Yue, Z. Wang and J. Ouyang, ACS Appl. Mater. Interfaces, 2021, 13, 20735–20745.
11. H. Souri, H. Banerjee, A. Jusufi, N. Radacsi, A. A. Stokes, I. Park, M. Sitti and M. Amjadi, Adv. Intell. Syst., 2020, 2, 2000039.
12. S. Khan, S. Ali, A. Khan and A. Bermak, IEEE Rev. Biomed. Eng., 2023, 16, 152–170.
13. R. Ariati, F. Sales, A. Souza, R. A. Lima and J. Ribeiro, Polymers, 2021, 13, 4258.
14. P. V. T. Weerasinghe, R. N. Wijesena, N. D. Tissera, G. Priyadarshana, N. D. Wanasekara, D. G. K. Dissanayake and K. M. Nalin de Silva, RSC Adv., 2023, 13, 18605–18613.
15. R. Banerjee, S. S. Ray and A. K. Ghosh, Polymers, 2018, 10, 400.
16. J. Hu, Y. Wu, Q. Yang, Q. Zhou, L. Hui, Z. Liu, F. Xu and D. Ding, Carbohydr. Polym., 2022, 275, 118697.
17. Z. Wang, Z. Ma, J. Sun, Y. Yan, M. Bu, Y. Huo, Y. F. Li and N. Hu, Polymers, 2021, 13, 813.
18. M. C. Karlapudi, M. Vahdani, S. M. Bandari, S. Peng and S. Wu, Sensors, 2023, 23, 3245.
19. S. Li, W. Zhou, Y. Hu, C. Huang, Q. Gao and Y. Chen, Int. J. Biol. Macromol., 2022, 207, 365–373.
20. X. Chen, R. Li, G. Niu, M. Xin, G. Xu, H. Cheng and L. Yang, Chem. Eng. J., 2022, 444, 136631.
21. A. Poulin, X. Aeby, G. Siqueira and G. Nyström, Sci. Rep., 2021, 11, 1–9.
22. H. Liu, Q. Zhang, N. Yang, X. Jiang, F. Wang, X. Yan, X. Zhang, Y. Zhao and T. Cheng, ACS Appl. Mater. Interfaces, 2023, 15, 44554–44562.
23. D. Tahir, M. Ramzan, A. Karim, H. Hu, S. Naseem, M. Rehan, M. Ahmad and M. Zhang, Polymers, 2022, 14, 4468.
24. M. R. A. Karim, D. Tahir, E. U. Haq, A. Hussain and M. S. Malik, Polym. Polym. Compos., 2020, 29, 277–300.
25. M. R. Abdul Karim, D. Tahir, K. I. Khan, A. Hussain, E. U. Haq and M. S. Malik, Plast., Rubber Compos., 2022, 52, 114–127.
26. T. Li, C. Chen, A. H. Brozena, J. Y. Zhu, L. Xu, C. Driemeier, J. Dai, O. J. Rojas, A. Isogai, L. Wågberg and L. Hu, Nature, 2021, 590, 47–56.
27. R. Erdmann, S. Kabasci and H. P. Heim, Polymers, 2021, 13, 1356.
28. L. Zhou, K. Ke, M. B. Yang and W. Yang, Compos. Commun., 2021, 23, 100548.
29. M. Vahdani, S. Mirjalali, M. Razbin, S. A. Moshizi, D. Payne, J. Kim, S. Huang, M. Asadnia, S. Peng and S. Wu, Adv. Sustainable Syst., 2024, 8, 2300482.
30. J. F. Sassi and H. Chanzy, Cellulose, 1995, 2, 111–127.
31. C. Lu, S. Xu, D. Tang, C. Huang, D. Zhang, S. Gao, J. Wang, C. Wang, Q. Yong and F. Chu, Chem. Eng. J., 2025, 508, 161024.
32. H. Zhang, X. Wu, Z. Qin, X. Sun, H. Zhang, Q. Yu, M. Yao, S. He, X. Dong, F. Yao and J. Li, Cellulose, 2020, 27, 9975–9989.
33. J. W. Han, J. Park, J. H. Kim, S. A. N. Entifar, A. Prameswati, A. F. Wibowo, S. Kim, D. C. Lim, J. Lee, M. W. Moon, M. S. Kim and Y. H. Kim, Materials, 2022, 15, 5009.
34. S. Azadi, S. Peng, S. A. Moshizi, M. Asadnia, J. Xu, I. Park, C. H. Wang and S. Wu, Adv. Mater. Technol., 2020, 5, 2000426.
35. S. Wu, S. Peng and C. H. Wang, Sens. Actuators, A, 2018, 279, 90–100.
36. J. Tarique, S. M. Sapuan and A. Khalina, Sci. Rep., 2021, 11, 1–17.
37. J. S. Sidhu, A. Misra and A. Bhardwaj, J. Mater. Sci.: Mater. Electron., 2024, 35, 1–11.
38. R. Wang, X. Li and T. Chen, AIP Adv., 2024, 14, 055028.
39. C. Chen, Y. Jia, D. Jia, S. Li, S. Ji and C. Ye, RSC Adv., 2017, 7, 1936–1942.
40. Y. H. Wang, D. X. Du, H. Xie, X. B. Zhang, K. W. Lin, K. Wang and E. Fu, J. Mater. Sci.: Mater. Electron., 2021, 32, 496–508.
41. S. Krainer, C. Smit and U. Hirn, RSC Adv., 2019, 9, 31708–31719.
42. J. R. Zhang, A. Li, Z. J. Sun, Q. Wang, Y. L. Zhang and D. D. Han, Appl. Phys. Lett., 2025, 126, 33505.
43. D. Y. Choi, M. H. Kim, Y. S. Oh, S. H. Jung, J. H. Jung, H. J. Sung, H. W. Lee and H. M. Lee, ACS Appl. Mater. Interfaces, 2017, 9, 1770–1780.
44. S. Wu, J. Zhang, R. B. Ladani, A. R. Ravindran, A. P. Mouritz, A. J. Kinloch and C. H. Wang, ACS Appl. Mater. Interfaces, 2017, 9, 14207–14215.
45. T. Hwang, J. Seo, E. Ko, C. Yang and H. S. Lee, Mater. Adv., 2023, 4, 1326–1334.
46. A. Farooq, H. Wanyan, Q. Li, S. Lu, W. Huang, M. Waqas, B. Hong, L. Huang, L. Chen, X. Zhou and H. Wu, Carbohydr. Polym., 2025, 358, 123534.
47. X. Dang, B. Guo, Y. Fu, Y. Fei and X. Wang, Cell Biomater., 2025, 1, 100076.
48. Z. Jiang, Z. Ji, M. Zhu, W. Ma, S. Gao and M. Xu, Int. J. Biol. Macromol., 2024, 280, 135947.
49. N. Jaroenthai, N. Srikhao, P. Kasemsiri, M. Okhawilai, S. Theerakulpisut, H. Uyama and P. Chindaprasirt, Int. J. Biol. Macromol., 2023, 253, 127401.
50. H. Zhou, Z. Jin, Y. Yuan, G. Zhang, W. Zhao, X. Jin, A. Ma, H. Liu and W. Chen, Colloids Surf., A, 2020, 592, 124587.
51. T. Zhu, Y. Cheng, C. Cao, J. Mao, L. Li, J. Huang, S. Gao, X. Dong, Z. Chen and Y. Lai, Chem. Eng. J., 2020, 385, 123912.
52. W. Sun, X. Liu, W. Hua, S. Wang, S. Wang, J. Yu, J. Wang, Q. Yong, F. Chu and C. Lu, Int. J. Biol. Macromol., 2023, 248, 125900.
53. Y. Cheng, X. Ren, G. Gao and L. Duan, Carbohydr. Polym., 2019, 223, 115051.
54. J. Gao, X. Li, L. Xu, M. Yan, H. Bi and Q. Wang, Carbohydr. Polym., 2024, 329, 121784.
55. R. Ullah, L. A. Shah, M. Khan and L. Ara, Int. J. Biol. Macromol., 2023, 246, 125666.
56. C. Feng, B. Xu, L. Chen, Z. Qiu and J. Guo, Eur. Polym. J., 2024, 214, 113173.
57. R. Zhang, Z. Zhang, P. Xu, J. Xu, Y. Gao and G. Gao, Carbohydr. Polym., 2024, 326, 121654.
58. Y. Jiao, K. Lu, Y. Lu, Y. Yue, X. Xu, H. Xiao, J. Li and J. Han, Cellulose, 2021, 28, 4295–4311.
59. J. Gao, Z. Yu, H. Xue, T. Zhang, J. Gu and F. Huang, Int. J. Biol. Macromol., 2024, 282, 136802.
60. Y. Cheng, X. Ren, L. Duan and G. Gao, J. Mater. Chem. C, 2020, 8, 8234–8242.
61. S. Azadi, S. Peng, S. A. Moshizi, M. Asadnia, J. Xu, I. Park, C. H. Wang and S. Wu, Adv. Mater. Technol., 2020, 5, 2000426.
62. X. Li, S. Xiao, Y. Lao, D. Li, Q. Wei, L. Ye and S. Lu, Int. J. Biol. Macromol., 2024, 266, 131004.
63. J. Cheng, H. Chen, X. Chen, X. Cai, R. Lin, C. Xiong, L. You and S. Wang, Appl. Mater. Today, 2023, 34, 101898.
64. R. Wang, Q. Liu, J. Wei, C. Zhu, Y. Wang, A. Yu, W. Wang, J. Zou, J. Xie and Z. Fu, J. Mater. Sci. Technol., 2025, 210, 195–203.
65. X. Ji, Z. Yu, M. Nan, R. Ge, Y. Gao, H. Zhang, Z. Li and L. Wang, Int. J. Biol. Macromol., 2025, 315, 144405.
66. J. Zheng, J. Zhou, Y. Zhao, C. Wang, M. Fan, Y. Li, C. Yang and H. Yang, Biosensors, 2025, 15, 177.
67. R. Yin, S. Yang, Q. Li, S. Zhang, H. Liu, J. Han, C. Liu and C. Shen, Sci. Bull., 2020, 65, 899–908.
68. W. Ma, Z. Wu, G. Yuan, B. Wang, B. Xu, L. Kong, T. Wan, S. Wang and H. Jiang, Mater. Today Commun., 2025, 46, 112717.
69. P. Li, Z. Ling, X. Liu, L. Bai, W. Wang, H. Chen, H. Yang, L. Yang and D. Wei, Chem. Eng. J., 2023, 466, 143306.
70. O. Tomes, A. Soul, H. Zhang, E. Bilotti and D. G. Papageorgiou, Nanoscale, 2024, 16, 6163–6175.
71. S. Peng, Y. Yu, S. Wu and C. H. Wang, ACS Appl. Mater. Interfaces, 2021, 13, 43831–43854.
72. R. A. Antunes, M. C. L. De Oliveira, G. Ett and V. Ett, J. Power Sources, 2011, 196, 2945–2961.
73. S. Wu, S. Peng, Y. Yu and C. H. Wang, Adv. Mater. Technol., 2020, 5, 1900908.
74. S. Mirjalali, S. Peng, Z. Fang, C. H. Wang and S. Wu, Adv. Mater. Technol., 2022, 7, 2100545.
75. N. Click, I. Teknetzi, E. P. L. Tam, M. Tao and B. Ebin, Sol. Energy Mater. Sol. Cells, 2024, 270, 112834.
76. B. Yoo, Y. Kim, C. J. Han, M. S. Oh and J. W. Kim, Appl. Surf. Sci., 2018, 429, 151–157.

---