Synthesis of 2D Ti₃C₂T_x MXene and MXene-based composites for flexible strain and pressure sensors

Abstract

As an important device in flexible and wearable microelectronics, flexible sensors have gained a lot of attention due to their wide application in human motion monitoring, human–computer interactions and healthcare fields. The preparation of flexible sensors with superior sensing performance and a simple process is still a challenging goal pursued by scientific researchers all over the world. The emerging two-dimensional (2D) Ti₃C₂T_x MXene material, having the characteristics of high metallic conductivity, good flexibility, excellent dispersibility and hydrophilicity, is suitable for flexible sensors as a conductive sensing material. In this review, the preparation strategies of Ti₃C₂T_x are summarized. Combined with its research progress in flexible sensors, the preparation methods, sensing performance, working mechanism and applications of Ti₃C₂T_x flexible sensors with different device architectures are reviewed.

---

Yuping Zeng

Yuping Zeng received her BS degree from Wuhan University in 2019. She is currently pursuing her master's degree in the School of Printing and Packaging at Wuhan University under the supervision of Prof. Wei Wu. Her research involves the preparation of functional nanomaterials and their applications in flexible sensors.

Wei Wu

Wei Wu received his PhD degree from Wuhan University in 2011. He then joined the group of Prof. Daiwen Pang at Wuhan University (2011) and Prof. V. A. L. Roy at the City University of Hong Kong (2014) as a postdoctoral fellow. Currently, he is the full professor and Dean of the School of Printing and Packaging, Wuhan University. He received the STAM Best Paper Award in 2017 and Hong Kong Scholars Award in 2014. He has published over 90 papers, which have received over 7000 citations. His research interests include the synthesis and application of printable functional materials, printed electronics, wearable electronics and intelligent packaging.

---

1. Introduction

Human beings are pursuing a higher quality of life. Rigid electronics based on traditional silicon materials gradually cannot fully satisfy people's requirements for portable, foldable, and wearable miniature electronic devices.^1–3 At the same time, the scientific community's unremitting exploration and research on organic electronic materials and inorganic nanomaterials has led to the continuous development and progress of flexible electronic materials, thus giving birth to flexible electronic technology.⁴ Flexible electronic devices manufactured by flexible electronic technology can bear bending, stretching, compression or folding.⁵ Flexible sensors, as an important branch of flexible electronic devices, can be used on a variety of irregular surfaces due to their unique flexibility and biocompatibility, and can be used in application scenes that rigid sensors cannot detect, expanding more application scenarios. For example, by directly attaching flexible sensors to the surface of human skin, or directly using technologies such as weaving to integrate flexible electronics into clothes, various functions such as monitoring human motions, sensing physical states and detecting drugs can be achieved, making wearable electronic devices become a reality.^6–13 Flexible sensors can detect changes in the outside environments, such as strain,^14–16 humidity,^17–19 temperature,^20–22 harmful gases,^23–26 the concentration of biological macromolecules,^27,28etc. These physical and biochemical signals are converted into electrical signals and outputs. In addition to shining in the fields of wearable electronics such as artificial skin, motion tracking, medical care and soft robots, flexible sensors also occupy a place in the field of Internet of Things.

An ideal flexible sensor should have the characteristics of good flexibility, high sensitivity, stable performance, low detection limit, good linearity and low hysteresis. The key to preparing a flexible sensor with superior performance lies in the flexibility of conductive materials. The flexibility of the commonly used nanoscale conductive materials (metal nanomaterials, carbon materials, and conductive polymers) alone often fails to meet the requirements of practical applications. Therefore, a conductive sensing material with poor stretchability and a flexible polymer elastomer is usually combined to form a flexible sensor with both flexibility and conductive sensing functions.^29,30

Since 2004, Geim's group at the University of Manchester has used mechanical exfoliation to prepare atomically thick graphene for the first time. 2D materials have gradually entered people's field of vision and have aroused great interest. Scientists have conducted a lot of theoretical and applied research on 2D materials. In 2011, Gogotsi's research group of Drexel University in the United States discovered a new type of 2D material – MXene, that is, 2D transition metal carbide, nitride or carbonitride.³¹ They used hydrofluoric acid (HF) to etch Ti₃AlC₂ to prepare the 2D layered compound Ti₃C₂, and then used the same method to etch the ternary carbide and nitride MAX phase materials with similar structures to Ti₃AlC₂. The corresponding 2D transition metal carbide nanosheets were successfully obtained, which proved that the etchant HF has a selective etching effect on the A element in the MAX phase materials.^32,33 MAX is an anisotropic laminated structure consisting of alternate M_n+1X_n layers and A layers. Its general chemical formula is M_n+1AX_n (n = 1, 2 or 3), where M is the transition metal, A represents the IIIA or IVA group element, and X stands for C and/or N. There are nearly 30 MAX phases that have been discovered so far.³⁴ The MXene can be formed by means of selectively etching the A layer of the MAX phases, and its chemical formula is M_n+1X_nT_x, where T is the surface functional group (−OH, −F, −O), which is introduced under etching conditions, and x represents the number of end groups.³⁵ Among them, Ti₃C₂T_x is considered to be the most promising MXene material. Its unique 2D structure, high metallic conductivity and rich surface functional groups make it exhibit excellent electrochemical, electrical, optical and mechanical properties.³⁶ When Ti₃C₂T_x is used as a conductive sensing material for a flexible sensor, its 2D sheet architecture can have both high flexibility and conductivity at the same time, and it can be easily combined with other nanomaterials as a composite or hybrid material, further greatly improving the ductility. Compared with other 2D materials such as graphene, the metal-like conductivity of Ti₃C₂T_x facilitates the rapid transfer of electrons in the sensing process. In addition, Ti₃C₂T_x has negative surface functional groups, the surface is negatively charged, and it exhibits good hydrophilicity due to the hydrogen bond between T_x and water, which helps in dispersing uniformly in the aqueous solution and is beneficial to the manufacture of flexible sensors.^34,37–39 This review summarizes the preparation methods, sensing performance and specific applications of Ti₃C₂T_x-based flexible sensors in recent years, and details on the future development trend.

2. Synthesis methods for MXene

The Ti₃AlC₂ compound crystal belongs to the layered hexagonal structure, which is formed by alternately arranged Ti₃C₂ lamellae connected with densely packed Al atomic layers. Among them, the Ti–C atomic layers are mainly composed of strong covalent bonds and ionic bonds, and the Ti–Al atomic layers are mainly connected by relatively weak metallic bonds. Therefore, the Al atoms are relatively reactive and easy to be exfoliated, but the Ti₃C₂T_x MXene phase cannot be obtained directly by a classical mechanical peeling method, and Al atoms can only be etched away from Ti₃AlC₂ by a chemical liquid-phase etching method to obtain Ti₃C₂T_x with a 2D stacked and layered structure (Fig. 1a).⁴⁰ At present, the top-down method is generally used to prepare Ti₃C₂T_x sheets, and this process mainly includes four steps: synthesis of Ti₃AlC₂, etching, intercalation and delamination. According to the type of etchant used, it can be divided into fluorine-containing etching method (FCEM) and fluorine-free etching method (FFEM).

Fig. 1 (a) Atomic composition model of Ti₃C₂T_x.⁴⁵ Reproduced with permission from American Chemical Society. Copyright 2019. (b) Schematic illustration of the process for preparing MXene by HF etching and delaminating Ti₃AlC₂.⁴⁶ Reproduced with permission from John Wiley and Sons. Copyright 2017. (c) The structure of Ti₃C₂T_x.³⁹ Reproduced with permission from American Chemical Society. Copyright 2019.

2.1 Fluorine-containing etching method

The FCEM of Ti₃C₂T_x mainly includes HF etching, and hydrochloric acid-lithium fluoride (HCl + LiF) for in situ etching. Fig. 1b shows the schematic diagram of a common Ti₃C₂T_x prepared by the HF etching strategy. Under the strong corrosive action of HF, the Al atoms in Ti₃AlC₂ are corroded by HF to form AlF₃. After washing with distilled water many times, excess HF and AlF₃ are removed. A product with a 2D layered structure was obtained (Fig. 1c). A large amount of gas was generated during the etching process. The chemical reactions can be expressed as follows:³¹

Ti₃AlC₂ + 3HF = AlF₃ + 3/2H₂↑ + Ti₃C₂ (1)

Ti₃C₂ + 2H₂O = Ti₃C₂(OH)₂ + H₂↑ (2)

Ti₃C₂ + 2HF = Ti₃C₂F₂ + H₂↑ (3)

When etching using the HCl + LiF etching method, the chemical reaction equation can be expressed as follows:⁴¹

2Ti₃AlC₂ + 6LiF + 6HCl = 2Ti₃C₂ +Li₃AlF₆ + AlCl₃ + 3LiCl + 3H₂↑ (4)

The major difference between the two etching methods is that the HCl + LiF etching will insert the lithium ions (Li⁺), which can increase the number of hydroxyl functional groups, the distance of the Ti₃C₂T_x interlayers, and the area of the Ti₃C₂T_x flakes.³⁵ In order to take advantage of the structure and performance of the 2D material, the multi-layer loose accordion-like structure Ti₃C₂T_x is often delaminated into single-layer or few-layer Ti₃C₂T_x nanosheets. With the help of an intercalating agent and ultrasonic processing, the multi-layer Ti₃C₂T_x can be further delaminated into single or few layers. But the as-obtained Ti₃C₂T_x has small hole defects by using the HF etching method, which increases the contact of active titanium ions with air and aggravates oxidation. Therefore, TiO₂ is formed and the conductivity of Ti₃C₂T_x decreases. The Ti₃C₂T_x prepared by the etching method of HCl + LiF is hydrophilic due to its rich hydroxyl groups, and the intercalation effect of lithium ions expands the spacing between the multilayer clay Ti₃C₂T_x sheets, and it is easier to be delaminated into a single sheet.⁴² Moreover, the HCl + LiF etching method is more gentle than the HF etching method. The single-layer Ti₃C₂T_x sheet can be obtained directly by hand-shaking without ultrasonic treatment during the delamination process. Therefore, the size of the obtained Ti₃C₂T_x nanosheet will be larger, and the surface properties and mechanical stability will be more excellent. Additionally, the morphology as well as the performance of Ti₃C₂T_x will change with the different etching temperature, reaction time, and intercalation agent. For example, increasing the etching temperature can accelerate the etching rate and degree, and promote the conversion of raw Ti₃AlC₂ to multilayer Ti₃C₂T_x MXene, but the increase of the etching temperature will reduce the integrity of the material structure. Extending the etching time will increase the degree of Al atoms in Ti₃AlC₂ being etched, which will broaden the interlayer spacing of the product, but if the etching time is excessive, it will cause the layers to restack. During the preparation process with the intercalation agent, organic molecules and impurity ions will be inserted into the gaps of the MXene sheets, which significantly increases the delamination rate of Ti₃C₂T_x MXene.^41,43,44

2.2 Fluorine-free etching method

In the HCl + LiF method, the HF is in situ synthesized and used for etching Ti₃AlC₂. This method reduces the delamination steps and avoids the direct use of harmful HF, but it still requires the use of toxic fluorine-containing reagents. Currently, several FFEMs have been developed to fabricate the MXene, including electrochemical etching methods in various electrolytes, high-temperature alkali treatment, and molten salt synthesis strategies. For example, the Ti₃AlC₂ precursor is etched with KOH in the presence of a small amount of water to successfully prepare monolayer Ti₃C₂ nanosheets with larger lateral dimensions. During the etching process, the Al layer is replaced by the −OH group in the Ti₃AlC₂ structure. Furthermore, Ti₃C₂(OH)₂ nanosheets can be easily and effectively obtained by simple cleaning.⁴⁷ The high-temperature hydrothermal method was also used to etch Ti₃AlC₂ powder in NaOH solution with a high pH value to synthesize a high-purity multi-layer MXene without fluorine-containing functional groups.⁴⁸ Another Ti₃C₂T_x was prepared by the electrochemical etching method using Ti₃AlC₂ as the anode in NH₄Cl and tetramethylammonium hydroxide (TMAOH) electrolyte.⁴⁹ FFEMs avoid the generation of fluorine-containing and highly toxic waste liquids, eliminate some hidden safety hazards, and pave the way for the scalable synthesis and application of MXene materials.

The FCEMs and FFEMs have different effects on the properties of MXene while etching Ti₃AlC₂. The Ti₃C₂T_x prepared by the HF etching method and the HCl + LiF in situ etching method contains −O, −F and −OH active functional groups on the surface. Because hydrogen bonds will be formed between T_x and water, the Ti₃C₂T_x becomes hydrophilic and can be uniformly dispersed in aqueous solutions to obtain the Ti₃C₂T_x suspension with good dispersibility. This surface functionalization will also have an important influence on the electron transport performance of Ti₃C₂T_x MXene, which affects the conductivity of MXene.⁵⁰ However, the Ti₃C₂T_x sheet prepared in this way is easy to stack and oxidize, and the FFEM can be used to prepare Ti₃C₂T_x while modifying it, which broadens the application fields of Ti₃C₂T_x.⁵¹

Obviously, the preparation method is very important for the properties of Ti₃C₂T_x. And exploring a green, harmless, simple and quick strategy to prepare Ti₃C₂T_x materials with unique properties will be an important direction for the industrialization of Ti₃C₂T_x in the future.

3. Ti₃C₂T_x-based flexible sensors

Compared with graphene, the emerging 2D material Ti₃C₂T_x contains metallic elements. Theoretically, Ti₃C₂T_x is metallic, but the presence of surface functional groups leads to dipolar polarization on the surface of MXene, and reduces the conductivity to the semiconductor level, making Ti₃C₂T_x a narrow band gap semiconductor.^52,53 The conductivity of the self-supporting pure Ti₃C₂T_x film prepared by the HF etching method is as high as 2.4 × 10⁵ S m⁻¹ (the thickness is 3.3 μm), and that of the pure Ti₃C₂T_x film prepared by the HCl + LiF etching method is 2 × 10⁵ S m⁻¹ (the thickness is 5 μm). Obviously, Ti₃C₂T_x shows excellent electrical conductivity.^41,54 Ti₃C₂T_x, with a variety of unique properties such as high conductivity, easy dispersion and functional regulation, is suitable for constructing a conductive network for highly sensitive flexible sensors.

Currently, the flexible sensors based on Ti₃C₂T_x that have been reported are classified into stretchable strain sensors, pressure sensors and other sensors according to the different external stimuli applied. When the Ti₃C₂T_x sensor receives an external stimulus, the sensor can convert the detected stimulus signal into a recordable electrical signal, and realizes the sensing function by recording the relationship between the input signal and the electrical signal. Due to the different sensing mechanisms, sensors can be divided into piezoelectric,^55,56 resistive,^57–60 capacitive sensors,^61–63etc. Although each type of sensor has different characteristics, they are all developing towards the goal of realizing an ideal flexible sensor with high sensitivity, wide working range, low detection limit, good stability and fast response.^64,65 In order to manufacture a Ti₃C₂T_x flexible sensor with superior performance, it is necessary to carefully select an appropriate elastic substrate, prepare the conductive filler with excellent performance, and adopt a reasonable preparation process.^66–68 There are many methods for preparing flexible sensors, which can be roughly divided into two types: one is to use a mixing method to directly disperse conductive materials into a solid or liquid polymer matrix, and then form a 3D conductive network by freeze–drying or thermal curing. Another one is to directly deposit the conductive layer onto the polymer substrate by printing, spraying, spin coating, dip coating, etc.⁶⁹

3.1 Strain sensors

As one of the important electronic devices used in the fields of healthcare, soft robots and human–machine interactions, strain sensors can convert mechanical deformations into electrical signals (such as the changes of resistance or capacitance).⁷⁰ The conventional sensing mechanism of the resistive strain sensor is to convert tensile strain into resistance change, and its sensitivity is expressed by the gauge factor (GF), which is the ratio of the relative change in resistance to the tensile strain (GF = (R – R₀)/R₀ε, where R represents the real-time resistance, R₀ represents the initial resistance, and ε represents the tensile strain).⁶⁶ The following will introduce the sensing mechanism and the latest research progress of Ti₃C₂T_x tensile strain sensors.

Ti₃C₂T_x is usually composed of closely stacked 2D layers. The possible interaction between adjacent layers makes pure Ti₃C₂T_x-based flexible sensors unable to achieve effective slippage between the layers under external strains. Therefore, a pure Ti₃C₂T_x-based sensor can only generate cracks or microcracks to disperse strain.^71,72 But the cracks usually block the conductive path of the conductive layer and severely damage the conductive transmission layer. When the external strain increases, the crack gradually expands until the conductive path is completely cut off.⁷³ Although this sensing mechanism allows the Ti₃C₂T_x-based flexible sensor to possess ultra-high sensitivity in a small strain range, it greatly limits the improvement of the working range of strain sensors.⁷⁴ In addition, the delaminated Ti₃C₂T_x MXene sheet has a high aspect ratio. Without compromising their mechanical and electrical properties, it is hard to build up an orderly macroscopic geometric structure with good connection from individual Ti₃C₂T_x MXene nanosheets.⁷⁵ Therefore, the preparation of Ti₃C₂T_x composite strain sensors has become a reasonable and effective choice. The common Ti₃C₂T_x composite flexible strain sensor is generally composed of a stretchable polymer substrate and a conductive sensing layer attached to it.⁷⁶ As the carrier of the sensor, the polymer substrate is mainly responsible for absorbing strain and generating deformation, and the conductive sensing layer is responsible for generating electrical signal changes.⁷⁷ The elastic substrate and the conductive layer work together to correlate strain changes with electrical signal changes to achieve sensing purposes. Therefore, for the purpose of facilitating the overall performance of the strain sensor, the sensing layer and the flexible substrate are usually designed with microstructures.

Designing the microstructure of the conductive MXene layer, changing the individual 2D layer structure of MXene, and designing and constructing a more tightly connected conductive network, could enhance the sensing performance of strain sensors based on Ti₃C₂T_x. For instance, the Ti₃C₂T_x strain sensor is fabricated by transferring the Ti₃C₂T_x nanoparticle-nanosheet conductive layer with a hybrid network structure to a stretchable polydimethylsiloxane (PDMS) substrate. Fig. 2a shows the working mechanism of this strain sensor. In the hybrid sensing layer of Ti₃C₂T_x nanoparticles and nanosheets, the nanosheets wrap and bridge the nanoparticles, effectively hindering the propagation of cracks to maintain the connectivity of the conductive path, which greatly increases the strain region of the sensor. The Ti₃C₂T_x nanoparticle–nanosheet hybrid network helps make the microstructure fully function and maximize synergy, and improves the comprehensive sensing performance of the sensor. In the 0–5%, 5–35% and 35–53% strain ranges, the value of GF is 178.4, 505.1 and 1176.7, respectively.⁷⁸ In addition, other dimensions of materials can be added to the pure 2D Ti₃C₂T_x material to build a new conductive network. For example, the Ti₃C₂T_x MXene/CNT sensing layer with a sandwich structure is prepared with delaminated Ti₃C₂T_x nanosheets and one-dimensional (1D) single-walled carbon nanotubes (SWCNTs) with hydrophilicity by a layer-by-layer (LBL) spraying–coating method. The manufacturing process is shown in Fig. 2b. The 1D CNTs have a high aspect ratio and are dispersed between the Ti₃C₂T_x sheets. The loose Ti₃C₂T_x sheets are woven into a integrated structure, which improves the orderliness of the hierarchical structure and electronic pathways, constructing the entire conductive networks, and finally forming a highly stretchable polymer film. Under large strain, the hair-like CNTs played a bridging role, connecting the Ti₃C₂T_x MXene nanosheets at both ends of the gap, maintaining a part of the conductive path, making the sensor have high sensitivity (GF value is up to 772.6) in a large sensing range (as shown in Fig. 2c).⁷⁵ Because Ti₃C₂T_x has negative surface charge, hydrophilicity and excellent dispersing ability, it is suitable for preparing printed electronic ink to quickly realize the mass production of sensors by printing methods.⁷⁹ Silver nanowires (Ag NWs), polydopamine (PDA) and nickel ions (Ni²⁺) are added to Ti₃C₂T_x for building a “brick-and-mortar” architecture by simulating the structure of the nacre. A composite strain sensor is fabricated by the screen printing method, and have a wide strain range (83%) as well as high sensitivity (GF > 8700 in the strain region of 76–83%).⁷³ Because of the special hierarchical “brick-and-mortar” structure of the conductive layer, this sensor exhibits excellent sensing performance with high sensitivity and good stretchability, as shown in Fig. 2d. 2D Ti₃C₂T_x MXene and 1D Ag NWs are used as the “brick” to provide mechanical brittleness and electrical conductivity for the entire sensing layer. The PDA and Ni²⁺ work as a “mortar”, triggering synergistic toughening effects through interfacial interactions, layer slippage, polymer chain stretching, etc. This synergy effect endows the sensor with remarkable comprehensive sensing performance.

Fig. 2 (a) Schematic diagram of the working mechanism of the Ti₃C₂T_x nanoparticle–nanosheet hybrid conductive film.⁷⁸ Reproduced with permission from John Wiley and Sons. Copyright 2019. (b) Diagram for the preparation of Ti₃C₂T_x/CNT conductive layer by the LBL spraying–coating method. (c) The GF and strain range of the Ti₃C₂T_x/CNT/latex sensor.⁷⁵ Reproduced with permission from American Chemical Society. Copyright 2018. (d) Schematic of the “brick-and-mortar” structure of a Ti₃C₂T_x-Ag NW-PDA/Ni²⁺ sensor.⁷³ Reproduced with permission from American Chemical Society. Copyright 2019.

Furthermore, the Ti₃C₂T_x is utilized as a functional material for preparing conductive Ti₃C₂T_x MXene ink. Then, the MXene/ink is assembled to a nylon fabric film substrate by a spraying–coating technique to manufacture a MXene/ink strain sensor, which showed a high sensitivity (GF values are 27.2 and 170.9 under 0–25% and 25–29.7% strain, respectively), good linearity, and excellent reliability and stability (>2950 cycles).⁸⁰ Another approach is to composite nanomaterial fillers into the polymer matrix. For example, Ti₃C₂T_x MXene nanosheets are composited into the polyacrylamide-sodium alginate matrix to construct a double networked hydrogel.⁸¹ As a strain sensor, the MXene-based hydrogel achieved high tensile sensitivity (18.15 in GF) and extremely excellent stretchable properties (3150%), demonstrating great significance in the fields of speech recognition, expression recognition and handwriting verification. It can be seen that the microstructure design of the conductive layer of the Ti₃C₂T_x strain sensor is very important to improve the overall performance. The difference is that the flexible sensor prepared by printing has fewer consumables than other methods, simple operation, large-scale production, and fine conductive patterns can be achieved. More importantly, the printing method can realize the rapid integration of electronic devices and truly realize the flexibility of the entire integrated system. With further deepening of research, more constructions and designs of Ti₃C₂T_x-based sensing networks are proposed, endowing the sensing network robustness, stability, large tensile property and sensitivity, improving the practicability of Ti₃C₂T_x-based strain sensors and expanding their application.

The microstructure design of the substrate also plays a very important role because of the cooperative interaction of the elastic substrate and the conductive layer which will alter the quality of the strain sensor. For example, a network structure of polyurethane (PU) can be obtained by the electrospinning technology, and then the Ti₃C₂T_x solution is dripped on the PU surface (Fig. 3a), and finally, a MXene/PU strain sensor with a network structure is prepared.⁸² Since the stretchable PU is a highly interlocked network structure, MXene nanosheets penetrate into it freely and are assembled on the PU mat through electrostatic interactions or hydrogen bonding to form a 3D sensing network, as shown in Fig. 3b. The synergistic effect of the PU networks and the MXene sheet makes the strain sensor perform well. The GF values are 22.9 and 228 in the strain range of 0–10% and 10–100%, respectively, with an ultralow detection limit of 0.1%, and the sensing range is as high as 150%. In addition, a flexible MXene/PANIF strain sensor is prepared by spraying the MXene solution and the polyaniline fiber (PANIF) solution layer by layer on the pre-stretching elastic rubber substrate, showing superior comprehensive sensing performance (the preparation process is shown in Fig. 3c).⁸³ Compared with the MXene/PANIF sensor prepared without pre-stretching, the sensor with tile-like stacked hierarchical microstructures broadens its working range (80%), improves its sensitivity (GF is up to 2369.1), and also has an ultra-low detection limit (as low as 0.1538%).

Fig. 3 (a) Assembling of Ti₃C₂T_x sheets on the network PU mat. (b) SEM images of the network-MXene/PU mat with a wrinkled surface.⁸² Reproduced with permission from The Royal Society of Chemistry. Copyright 2019. (c) Schematic diagram of the preparation process of the MXene/PANIF strain sensors with pre-stretching.⁸³ Reproduced with permission from Elsevier. Copyright 2020.

Most of the Ti₃C₂T_x-based sensors reported so far are planar or sheet-shaped, resulting in the need for using tape and other complex substrates when used for human activity monitoring. However, it is difficult to fix them on human skin and be woven into user clothes or accessories, which seriously hinders their use in wearable electronic devices.⁸⁴ The MXene material is directly combined with the yarn to form a conductive yarn, which can be used as a single sensor or woven into clothes and accessories. It is used to manufacture smart fabrics so that the physiological conditions and sports activities can be monitored continuously and closely without any interference or restriction, which is more comfortable and convenient.⁸⁵ As shown in Fig. 4a, the Ag NW/WPU-MXene fiber strain sensor is prepared by alternately immersing hydrophilic polyurethane fibers (PUF) into Ag NW/water-based polyurethane (WPU) mixed solution and Ti₃C₂T_x MXene solution in turn.⁷¹ In the multi-layer structure of the fiber sensor, the PUF as the carrier is ultrasonically treated in a HCl solution to obtain a hydrophilic surface, which can be tightly combined with the sensing layer, and there are hydrogen bonds between the layers. The interaction makes the multilayer structure uniform, strong and stable. This fiber sensor exhibits excellent comprehensive performance; the gauge factors are 128, 800, 1553, 3.2 × 10⁵, 3 × 10⁶, and 1.6 × 10⁷ which correspond to six strain ranges of 0–15%, 15–25%, 25–50%, 50–70%, 70–85% and 85–100%, respectively; the detection limit is as low as 0.4% with a very short response time (344 ms). Subsequently, integrating this sensor into clothes can create smart fabrics (as shown in Fig. 4b), which can be used in human movement monitoring, and can be further expanded to the field of human healthcare monitoring. In addition to using strong acid and oxygen plasma⁷³ to make the surface of the substrate hydrophilic, it is also feasible to use materials with good biocompatibility, such as PDA, to chemically modify the substrate to strengthen the adhesion between the sensor material and the substrate. For example, a stretchable strain sensor is prepared by dipping the PDA-modified yarn into the Ti₃C₂T_x MXene/Ag NP mixed suspension and Ag NW solution in turn.³⁹ The composite yarn-based strain sensor combines 0D, 1D and 2D materials. The interconnection and synergy effect of the 3D nanomaterials greatly enhance the comprehensive sensing capability of this yarn-based strain sensor. This sensor shows high gauge factor (872.79) and wide strain range (200%) (Fig. 4c). Smart gloves made with yarn sensors can also be used to recognize gestures (Fig. 4d), which would play a vital valuable role in the lives of deaf-mute people. Moreover, Ti₃C₂T_x MXene@cotton fabric (MCF) strain sensors are fabricated by immersing cotton fabric@polyethyleneimine (PEI) into Ti₃C₂T_x dispersions.⁶⁷ Due to the strong electrostatic interaction between the negatively charged Ti₃C₂T_x MXene sheets and the positively charged PEI, the Ti₃C₂T_x nanosheets were wrapped on the cotton fabric, forming MCF strain sensors which possessed a gauge factor reaching up to 4.11 within the strain range of 15% and high durability (>500 cycles) and a low strain detection limit of 0.3%.

Fig. 4 (a) Fabrication process of Ag NW/WPU-MXene fiber strain sensor. (b) The picture of a smart sweater woven with six fiber sensors.⁷¹ Reproduced with permission from The Royal Society of Chemistry. Copyright 2019. (c) The gauge factor of MXene/Ag composite yarn strain sensor. (d) The smart glove used for sign language recognition.³⁹ Reproduced with permission from American Chemical Society. Copyright 2019.

Although it is a facile and scalable method to fabricate wearable Ti₃C₂T_x/textile strain sensors by coating the fibers, yarns, or fabrics with conductive Ti₃C₂T_x, there are still some big challenges such as linearity, cycling stability and washing durability.⁸⁶ One solution is to strengthen the interaction between the material and the substrate. Ti₃C₂T_x MXene infiltrated nylon and polyurethane (PU) nanofiber yarns (nanoyarns) were produced by a one-step bath electrospinning technique.⁶⁰ By using MXene dispersions as both a current collector (electrospinning) and a coagulation bath (wet-spinning), Ti₃C₂T_x flacks infiltrate into nanoyarns, maximizing interactions between the conductive material and the substrate. The obtained MXene/PU nanoyarn strain sensor achieved the stretchability of up to 263% (MXene/PU) and demonstrated high sensitivity (GF ≈ 17 in the range of 20–50% strain) and low drift, showing great potential application in human body movements monitoring. Apart from this, integrating Ti₃C₂T_x nanosheets within an elastomeric host to fabricate Ti₃C₂T_x composite fiber sensors was also efficient.⁸⁶ Ti₃C₂T_x MXene/PU composite fibers were produced by a scalable wet-spinning technique.⁵⁹ This fiber strain sensor achieved high stretchability (≈152%) and sensitivity (GF ≈ 12900). Finally, the MXene/PU composite fiber were knitted into a one-piece elbow sleeve, tracking the wearer's various movements, providing an effective strategy to achieve Ti₃C₂T_x-based textile sensors. It can be seen that the combination of the Ti₃C₂T_x MXene material and the textile not only extends the tensile range of the strain sensor, but also gives the as-obtained conductive fiber and yarn softness, skin-friendly properties, good biocompatibility and stretching stability. Multifunctional smart fabrics can be directly made by weaving and other methods, which enhances the use of flexible strain sensors as wearable electronics, and provides a more comfortable and non-implantable platform for healthcare monitoring.

Obviously, taking advantage of the material with a good structural design helps in preparing a Ti₃C₂T_x MXene-based flexible strain sensor possessing both wide sensing range and high sensitivity.⁷⁵ A large number of strategies have been used, such as designing the microstructure of MXene materials, changing the individual 2D layer structure of MXene, adding other materials to build more stable conductive networks, or tailoring the microstructure of the substrate. Making full use of the synergy effect of the materials in the hybrid system is very effective for better overall performance of Ti₃C₂T_x-based flexible strain sensors.

3.2 Pressure sensors

A flexible pressure sensor (FPS) is an electronic device that can convert applied pressure changes into electrical signal changes. It has an extensive range of applications, from human physiological condition detection to facial expression recognition,^87–89 from voice recognition to human motion monitoring,^90,91 from robotic tactile perception to remote human–computer interaction,^92,93 demonstrating that the key role of FPSs is indispensable. According to different sensing mechanisms, pressure sensors can be divided into piezoelectric sensors, piezocapacitive sensors and piezoresistive sensors. Among them, a piezoresistive sensor can convert pressure into resistance change, which has been widely studied due to its simple measurement circuit, low manufacturing cost and high sensitivity.^94,95 The sensitivity is defined as S = (ΔR/R₀)/P, where ΔR stands for the relative change of resistance, R₀ stands for the initial resistance, P is the pressure, usually kPa⁻¹ is used as the sensitivity unit, and sometimes S = (ΔI/I₀)/P, ΔI represents the relative change of current, I₀ represents the initial current.⁴⁵ The preparation process of FPSs based on Ti₃C₂T_x can be divided into two types: coating method and mixing method. The working range, sensitivity and stability of FPSs prepared by different processes are different.

3.2.1 The coating method. A Ti₃C₂T_x FPS prepared by the coating method is usually made by directly attaching the active material layer to the elastic polymer matrix through electrostatic force, van der Waals force, hydrogen bonding and other forces by printing, spraying, spin coating, drip coating, dipping, etc.⁹⁶ Finally, a flexible conductive polymer composite material is formed, and then the electrodes are adhered to it to produce a FPS. For instance, the Ti₃C₂T_x MXene solution is drip-coated on the PI film (printed with interdigital electrodes) to prepare a Ti₃C₂T_x MXene piezoresistive sensor with a sandwich structure (Fig. 5a). Under the external pressure, the distance between the two adjacent interlayers with a larger initial distance in MXene is reduced, thereby reducing the internal resistance and increasing the conductivity. This MXene pressure sensor monitors external pressures by the changes in resistivity derived from the changes in the distance between interlayers. In addition, this sensor makes full use of the intrinsic characteristics of multi-layer MXene, showing good stability and fast response (<30 ms), with a GF value of up to 180.1. But it is very precise because the sensor relies on the change of tiny layer spacing to reflect the pressure instead of the macroscopic geometric effect that it shows in a small range of compressible strain (0.19–2.13%).⁹⁷ This sensor plays an important role in the monitoring of small human activities and other weak strain. However, the vigorous development of the wearable electronics market requires devices to be suitable for full-range human motion detection and human–computer interactions, which puts forward higher requirements for the expansion of the working range of pressure sensors. To enable pressure sensors to measure a larger range of pressure, it is urgent to introduce a flexible matrix with better mechanical properties into the pure active layer of functional conductive materials to withstand more pressure deformation.⁹⁸ For example, one biodegradable PLA sheet and one PLA sheet coated with interdigital electrodes sandwich a porous tissue paper, which has been immersed in the MXene solution, to make a piezoresistive sensor. The piezoresistive sensor reflects the external pressure through the contact resistance, which is caused by the change of the distance between the porous conductive Ti₃C₂T_x MXene/tissue paper and the interdigital electrode. It is characterized by wide pressure measurement range (up to 30 kPa), rapid response (11 ms), low detection limit (10.2 Pa), excellent stability and outstanding degradability. This sensor is very practical in human–machine interactions and wireless movement monitoring of robots (Fig. 5b).⁴⁵

Fig. 5 (a) Working mechanism, equivalent circuit diagram and sensing performance of MXene-material for piezoresistive sensor.⁹⁷ Reproduced with permission from Springer Nature. Copyright 2017. (b) Pressure sensing capability of flexible sensor based on the MXene/tissue-paper and its application in remote movement monitoring of robots.⁴⁵ Reproduced with permission from American Chemical Society. Copyright 2019.

It is a simple and efficient method to prepare a composite conductive active layer by depositing conductive materials on an flexible substrate which possesses excellent mechanical properties.⁹⁹ Therefore, research on porous flexible substrate materials has expanded from tissue and cellulose papers to porous fabrics and foam sponge materials with good mechanical stability, great elasticity, biocompatibility and large specific surface area.¹⁰⁰ As for porous fabrics, it is an appropriate substrate material because of its various properties including low cost, comfort and high breathability. Compared with other soft substrates, sensors based on fabrics and textiles can not only detect complicated deformations, but can also achieve a seamless connection with a variety of fabric products, implementing an integrated design easily.¹⁰¹ For instance, a Ti₃C₂T_x MXene-coated cotton fabric pressure sensor is fabricated by dip-coating, which exhibited a high gauge factor (7.67 kPa⁻¹), a rapid response and relaxation speed (<35 ms), good stability (>2000 cycles), and washability.¹⁰² Another dip-coated conductive MXene/cotton fabric is sandwiched between the PDMS film and an interdigitated electrode for preparing a piezoresistive pressure sensor.¹⁰³ The MXene/cotton fabric based pressure sensor possessed high sensitivity (5.30 kPa⁻¹ in the pressure range of 0–1.30 kPa), broad sensing range (0–160 kPa) and rapid response/recovery time (50 ms/20 ms). As for elastic 3D porous substrates, a highly porous melamine sponge is soaked in the Ti₃C₂T_x dispersion to absorb MXene nanosheets. After drying, the MXene nanosheets are fixed on the sponge skeleton to form a 3D network structure (Fig. 6a). Under pressure, the microfibers in the MXene coated sponge deform and link each other, creating new conductive paths, which naturally brings about a corresponding current raise. After the pressure is released, the MXene sponge gradually returns to its original shape. This sponge sensor shows great sensitivity in a broad pressure range (147 kPa⁻¹ and 442 kPa⁻¹ in the range of 0–5.37 kPa and 5.37–18.56 kPa, respectively), and the detection limit is as low as 9 Pa; it can sense even the surface deformation when blowing the balloon.⁵⁸ The PU sponge skeleton treated with chitosan (CS) is immersed in the Ti₃C₂T_x solution to make a piezoresistive sensor (Fig. 6b). In this sensing system, the PU sponge is used as a supporting framework to provide high compressive elasticity, and CS (positively charged) is used as a binder to enhance the interface interaction between the negatively charged MXene sheet and the PU sponge through the electrostatic action. The stability of the sensing structure enables the MXene@CS@PU sensor to achieve a compression rate of up to 85% under a pressure of 245.7 kPa, and the detection limit is as low as 30 μN.¹⁰⁴ Furthermore, a Ti₃C₂T_x/PU foam based physical pressure sensor is fabricated by employing a dip-coating method for detecting human gestures and locating the positions of unknown objects, which displayed a remarkable sensitivity of ∼34.24 kPa⁻¹ and a high gauge factor of ∼323.59. This Ti₃C₂T_x/PU device can also sense strain and temperature which further finds exciting applications in the field of wearable electronics.⁹⁹ Moreover, to prepare porous MXene-PDMS composites (MPCs) with a hollow structure, nickel foam as a 3D substrate is dip-coated with Ti₃C₂T_x sheets first, then immersed in PDMS, and finally etched away, as shown in Fig. 6c. The MPC flexible piezoresistive sensor demonstrates a large working range of bending angle from 0° to 180°, good reliability and stable durability, and also has an extremely low detection limit (10 mg) and fast response time (20 ms).¹⁰⁵ The above examples show that the dipping method of loading the Ti₃C₂T_x material on the porous elastic substrate by the adsorption effect is very convenient and effective in constructing the 3D conductive networks for piezoresistive sensors. However, because the conductive material is adsorbed on the elastic substrate by van der Waals force or electrostatic actions, the content of the conductive material can only be roughly controlled by adjusting the concentration of the conductive material mixed solution, the dipping time and the number of times. As a result, the Ti₃C₂T_x MXene sheet is unevenly distributed in the flexible matrix, which might affect the uniformity and stability of the device performance. Therefore, firmly bonding Ti₃C₂T_x MXene to the flexible substrate involves many influencing factors (e.g., surface modification, interface adhesion and stability), which require further exploration before manufacturing pressure sensors.

Fig. 6 (a) Fabrication process of MXene sponge based sensor and its microstructure model of press-release dynamic process.⁵⁸ Reproduced with permission from Elsevier. Copyright 2018. (b) Fabrication process, digital image, sensing performance of MXene@CS@PU sponge sensor.¹⁰⁴ Reproduced with permission from Elsevier. Copyright 2019. (c) Diagram of fabrication and response time of hollow-structured MPC sensor, and the MPC electronic skin with pixel arrays for pressure sensing mapping.¹⁰⁵ Reproduced with permission from Elsevier. Copyright 2019.

3.2.2 The mixing method. The Ti₃C₂T_x material is dispersed into other flexible materials by a simple mixing method, and then a conductive composite material with a 3D conductive network structure is formed by means of hot pressing, thermal curing or freeze–drying. Hydrogel with a 3D network structure, composed of a large amount of water and ionic liquid, shows great stretchability, self-healing and biocompatibility. It is a polymer material widely used in flexible electronic products in recent years.^61,106 The combination of Ti₃C₂T_x and hydrogel not only endows the Ti₃C₂T_x material with great softness and flexibility, but also enables the sensor device to have self-healing properties similar to human skin, which expands its application range. For example, the Ti₃C₂T_x nanosheets are mixed with a commercial hydrogel made up of poly(vinyl alcohol) (PVA), anti-dehydration additives and water to prepare an MXene-based hydrogel (M-hydrogel) sensor. This sensor shows excellent stretchable properties (3400%) and self-healing ability, with a gauge factor of 25 under tensile strain and 80 under compressive strain. This anisotropic response to compressive strain and stretchable strain and viscous deformation add a new dimension to the sensing ability of hydrogels, which can be used to detect the direction and speed of movements on the surface of M-hydrogel (Fig. 7a and b). In addition, its unique adhesiveness allows it to be directly stuck to the body surface to monitor body motions without an adhesive (Fig. 7c).⁸⁹

Fig. 7 The response of the M-hydrogel sensor to the movements on its surface (a) vertical and (b) parallel to the current direction. (c) The response of M-hydrogel sensor to facial expressions.⁸⁹ Reproduced with permission from the American Association for the Advancement of Science. Copyright 2018. (d) A diagram demonstrating the preparation process and structure of the light MXene/PI aerogel.¹⁰⁷ Reproduced with permission from John Wiley and Sons. Copyright 2018. (e) Fabrication process and sensing mechanism of MXene/rGO aerogel-based sensor and its response to the applied pressure.¹⁰⁸ Reproduced with permission from American Chemical Society. Copyright 2018.

Aerogel, as a hydrogel derivative, is a material that also has a 3D network structure, and has the characteristics of high porosity, low density, and lightweight features. However, the 2D MXene material is difficult to support it as the solid phase of the framework of the aerogel network due to its brittleness. Therefore, it is often required to add other highly elastic materials to construct a 3D porous structure, improving the mechanical strength of MXene-based aerogels and increasing the pressure sensing range.⁷⁴ The Ti₃C₂T_x dispersion and the poly(amic acid) (PAA) solution are uniformly mixed, freeze–dried to form MXene/PAA aerogels, and then annealed to induce the polymerization of PAA to form polyimide (PI) macromolecules, and a stable 3D structure of MXene/PI aerogel is obtained, as shown in Fig. 7d. The lightweight aerogel uses PI macromolecules to bridge the MXene sheets, and has high elasticity and good mechanical stability. It can endure deformations of 180% bending, 80% compression and torsion, and it is ultra-light and can even be put on the dandelion.¹⁰⁷ The 3D network of hydrogels and aerogels is usually composed of several chemical (covalent) and physical (such as hydrophobic and electrostatic) interactions of polymer chains. Other gelators such as 2D materials are also used to construct the network, which also effectively improve the characteristics of the formed gels.¹⁰⁹ The Ti₃C₂T_x solution and the graphene oxide (GO) solution are uniformly mixed, and then followed by ice-template-based freezing and low-temperature annealing to prepare an ultra-light and super-elastic MXene/rGO aerogel (Fig. 7e). Among them, rGO with a larger surface serves as the network skeleton of the aerogel. While providing mechanical strength for the aerogel, it partially covers the Ti₃C₂T_x nanosheets to partly avoid the oxidation of Ti₃C₂T_x. The conductivity of MXene is good. The synergy effect of rGO and Ti₃C₂T_x meets the geometric and resistance effects of the piezoresistive sensor, resulting in the pressure sensor that has excellent sensitivity (22.56 kPa⁻¹) and detection limit as low as 10 Pa.¹⁰⁸ Therefore, filling MXene materials into other materials with greater mechanical strength is an effective way to enhance the sensing capabilities of Ti₃C₂T_x-based pressure sensors.

Obviously, the Ti₃C₂T_x FPS prepared by combining Ti₃C₂T_x with a flexible substrate can overcome the problem of insufficient mechanical performance of individual MXene sheets, and endow the pressure sensor with higher sensitivity and compressibility, which significantly improves the sensor's comprehensive sensing performance. It can also add the characteristics of other materials to the sensor, providing an effective and usable idea for the manufacturing of new multifunctional devices.

4. Conclusions and perspectives

In recent years, the 2D material Ti₃C₂T_x MXene has gained widespread attention owing to its advantages such as great flexibility, high conductivity, and hydrophilicity. Flexible strain and pressure sensors based on Ti₃C₂T_x materials have broad application prospects in sports, medical care, human–computer interactions and other fields. They can not only respond to large strain changes such as joint activities, but can also be used to detect subtle strain changes such as pulse and sound. After investigating the progress of flexible sensors based on Ti₃C₂T_x materials, it is found that good progress has been made in performance analysis and working principle elaboration. Two strategies can be used to further improve the sensing performance: one is to add other materials to the conductive networks, and use the synergy effect between the materials to improve the overall performance of the sensor, which usually makes the sensor maintain the connectivity of the conductive path under large strain, greatly improving the working range of the MXene sensor. The second is to control the microscopic morphology of Ti₃C₂T_x and design the microstructure of the conductive network to improve the sensing performance.

For the purpose of meeting the different requirements of various flexible sensor devices, many manufacturing methods have been adopted to prepare various forms of Ti₃C₂T_x sensors, which greatly enriches the types of flexible sensors. However, to truly realize the Ti₃C₂T_x-based flexible sensor from laboratory preparation to actual industrial production, there are still many challenges. (1) The preparation process of MXene is not green and pollution-free because it requires the use of fluorine-containing reagents, which is dangerous, and will produce a large amount of acidic waste liquid and organic waste liquid. These waste liquids need to undergo strict treatment before they can be discharged, otherwise it is easy to cause great harm to the human body or the environment. (2) MXene has low yields and high production cost, which is far from the goal of manufacturing cheap flexible sensors. (3) The oxidation resistance of MXene is not good. It is easy to oxidize under heating conditions and in air, and the conductivity decreases, which leads to the reduction of sensor performance and the service life of the sensor.

In addition to the above challenges, Ti₃C₂T_x-based flexible strain and pressure sensors have the following development trends in the future. (1) In terms of sensor manufacturing, green, environmentally friendly and energy-saving processes such as the printing method will be more suitable to reduce environmental pollution, raw material waste, and production costs. (2) In terms of the function of the sensor, in addition to improving the sensitivity and working range of the sensor, the stability and life of the sensor will also be considered. For improving the user experience, the readability and the wireless sensing capability should also be valued. (3) In terms of sensor's installation and application, future sensors are expected to be smaller and more portable, and it is inevitable that more wearable sensors and sensing systems that integrate multiple functions will be developed. Therefore, some features (e.g., self-powering, self-healing, self-cleaning, water resistance and transparency) should be considered to be incorporated into Ti₃C₂T_x-based flexible sensors. And the integration and packaging technology of sensors and other components such as power supply or signal processing units will also be further studied. (4) For the sensor recycling process, in order to realize the harmonious symbiosis of man and nature and reduce the generation of garbage, the sensor will move towards development in the direction of easy recycling and degradability.

In this paper, we summarized the latest developments in the synthesis of 2D Ti₃C₂T_x MXene and flexible strain and pressure sensors based on Ti₃C₂T_x, and introduced in detail the methods to improve the performance of Ti₃C₂T_x-based sensors. At the same time, we put forward the current challenges and future development trends of Ti₃C₂T_x-based flexible force sensors. We hope that this review can provide effective help for broad researchers, encourage extensive research and promote the commercialization of Ti₃C₂T_x-based sensors.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

This work was supported by the National High-Level Talents Special Support Program, the Natural Science Foundation of Hubei Province for Distinguished Young Scholars (2019CFA056), the Fundamental Research Funds for the Central Universities (2042021kf0226), and the Guangdong-Hong Kong-Macao Joint Innovation Funding Project of Guangdong Science and Technology Program (2020A0505140004).

References

1. H. Yang, W. R. Leow and X. Chen, Small Methods, 2018, 2, 1700259.
2. X. Hu, Y. Dou, J. Li and Z. Liu, Small, 2019, 15, 1804805.
3. W. Gao, H. Ota, D. Kiriya, K. Takei and A. Javey, Acc. Chem. Res., 2019, 52, 523–533.
4. M. Gong, L. Zhang and P. Wan, Prog. Polym. Sci., 2020, 107, 101279.
5. H. Gwon, J. Hong, H. Kim, D.-H. Seo, S. Jeon and K. Kang, Energy Environ. Sci., 2014, 7, 538–551.
6. J. Oh, J. C. Yang, J. O. Kim, H. Park, S. Y. Kwon, S. Lee, J. Y. Sim, H. W. Oh, J. Kim and S. Park, ACS Nano, 2018, 12, 7546–7553.
7. C. Luo, B. Tian, Q. Liu, Y. Feng and W. Wu, Adv. Mater. Technol., 2020, 5, 1900925.
8. M. Ren, Y. Zhou, Y. Wang, G. Zheng, K. Dai, C. Liu and C. Shen, Chem. Eng. J., 2019, 360, 762–777.
9. X. Wang, J. Li, H. Song, H. Huang and J. Gou, ACS Appl. Mater. Interfaces, 2018, 10, 7371–7380.
10. D. Wasserberg and P. Jonkheijm, Chem., 2017, 3, 531–533.
11. Y. Zhao, M. Ren, Y. Shang, J. Li, S. Wang, W. Zhai, G. Zheng, K. Dai, C. Liu and C. Shen, Compos. Sci. Technol., 2020, 200, 108448.
12. B. Tian, W. Yao, P. Zeng, X. Li, H. Wang, L. Liu, Y. Feng, C. Luo and W. Wu, J. Mater. Chem. C, 2019, 7, 809–818.
13. S. Yao, P. Swetha and Y. Zhu, Adv. Healthcare Mater., 2018, 7, 1700889.
14. Z. Lv, J. Liu, X. Yang, D. Fan, J. Cao, Y. Luo and X. Zhang, ACS Appl. Mater. Interfaces, 2020, 12, 22163–22169.
15. B. Wang, K. Yang, H. Cheng, T. Ye and C. Wang, Chem. Eng. J., 2021, 404, 126393.
16. J. Chen, H. Wen, G. Zhang, F. Lei, Q. Feng, Y. Liu, X. Cao and H. Dong, ACS Appl. Mater. Interfaces, 2020, 12, 7565–7574.
17. H. An, T. Habib, S. Shah, H. Gao, A. Patel, I. Echols, X. Zhao, M. Radovic, M. J. Green and J. L. Lutkenhaus, ACS Appl. Nano Mater., 2019, 2, 948–955.
18. J. Yang, R. Shi, Z. Lou, R. Chai, K. Jiang and G. Shen, Small, 2019, 15, 1902801.
19. Y. Wang, L. Zhang, J. Zhou and A. Lu, ACS Appl. Mater. Interfaces, 2020, 12, 7631–7638.
20. T. Q. Trung, S. Ramasundaram, B.-U. Hwang and N.-E. Lee, Adv. Mater., 2016, 28, 502–509.
21. C. Yan, J. Wang and P. S. Lee, ACS Nano, 2015, 9, 2130–2137.
22. J. Yang, D. Wei, L. Tang, X. Song, W. Luo, J. Chu, T. Gao, H. Shi and C. Du, RSC Adv., 2015, 5, 25609–25615.
23. E. Singh, M. Meyyappan and H. S. Nalwa, ACS Appl. Mater. Interfaces, 2017, 9, 34544–34586.
24. W. Li, R. Chen, W. Qi, L. Cai, Y. Sun, M. Sun, C. Li, X. Yang, L. Xiang, D. Xie and T. Ren, ACS Sens., 2019, 4, 2809–2818.
25. S. Li, A. Liu, Z. Yang, J. He, J. Wang, F. Liu, H. Lu, X. Yan, P. Sun, X. Liang, Y. Gao and G. Lu, Sens. Actuators, B, 2019, 299, 126970.
26. T. Zhang, Q. Cheng, B. Jiang and Y. Huang, Composites, Part B, 2020, 196, 108131.
27. Q. Zhang, D. Jiang, C. Xu, Y. Ge, X. Liu, Q. Wei, L. Huang, X. Ren, C. Wang and Y. Wang, Sens. Actuators, B, 2020, 320, 128325.
28. H.-B. Lee, M. Meeseepong, T. Q. Trung, B.-Y. Kim and N.-E. Lee, Biosens. Bioelectron., 2020, 156, 112133.
29. R. Ma, S.-Y. Chou, Y. Xie and Q. Pei, Chem. Soc. Rev., 2019, 48, 1741–1786.
30. T. Q. Trung and N.-E. Lee, Adv. Mater., 2017, 29, 1603167.
31. M. Naguib, M. Kurtoglu, V. Presser, J. Lu, J. Niu, M. Heon, L. Hultman, Y. Gogotsi and M. W. Barsoum, Adv. Mater., 2011, 23, 4248–4253.
32. M. Naguib, J. Halim, J. Lu, K. M. Cook, L. Hultman, Y. Gogotsi and M. W. Barsoum, J. Am. Chem. Soc., 2013, 135, 15966–15969.
33. M. Naguib, V. N. Mochalin, M. W. Barsoum and Y. Gogotsi, Adv. Mater., 2014, 26, 992–1005.
34. Y. Z. Zhang, Y. Wang, Q. Jiang, J. K. El-Demellawi, H. Kim and H. N. Alshareef, Adv. Mater., 2020, 32, 1908486.
35. X. Zang, J. Wang, Y. Qin, T. Wang, C. He, Q. Shao, H. Zhu and N. Cao, Nano-Micro Lett., 2020, 12, 77.
36. H. Kim and H. N. Alshareef, ACS Mater. Lett., 2019, 2, 55–70.
37. J. Yan, C. E. Ren, K. Maleski, C. B. Hatter, B. Anasori, P. Urbankowski, A. Sarycheva and Y. Gogotsi, Adv. Funct. Mater., 2017, 27, 1701264.
38. X. Xie, M.-Q. Zhao, B. Anasori, K. Maleski, C. E. Ren, J. Li, B. W. Byles, E. Pomerantseva, G. Wang and Y. Gogotsi, Nano Energy, 2016, 26, 513–523.
39. H. Li and Z. Du, ACS Appl. Mater. Interfaces, 2019, 11, 45930–45938.
40. J.-F. Zhang, H.-Y. Cao and H.-B. Wang, J. Inorg. Mater., 2017, 32, 561–570.
41. T. Zhang, L. Pan, H. Tang, F. Du, Y. Guo, T. Qiu and J. Yang, J. Alloys Compd., 2017, 695, 818–826.
42. M. Ghidiu, J. Halim, S. Kota, D. Bish, Y. Gogotsi and M. W. Barsoum, Chem. Mater., 2016, 28, 3507–3514.
43. O. Mashtalir, M. Naguib, B. Dyatkin, Y. Gogotsi and M. W. Barsoum, Mater. Chem. Phys., 2013, 139, 147–152.
44. J.-C. Lei, X. Zhang and Z. Zhou, Front. Phys., 2015, 10, 276–286.
45. Y. Guo, M. Zhong, Z. Fang, P. Wan and G. Yu, Nano Lett., 2019, 19, 1143–1150.
46. L. Ding, Y. Wei, Y. Wang, H. Chen, J. Caro and H. Wang, Angew. Chem., Int. Ed., 2017, 56, 1825–1829.
47. G. Li, L. Tan, Y. Zhang, B. Wu and L. Li, Langmuir, 2017, 33, 9000–9006.
48. T. Li, L. Yao, Q. Liu, J. Gu, R. Luo, J. Li, X. Yan, W. Wang, P. Liu, B. Chen, W. Zhang, W. Abbas, R. Naz and D. Zhang, Angew. Chem. Int. Ed., 2018, 57, 6115–6119.
49. S. Yang, P. Zhang, F. Wang, A. G. Ricciardulli, M. R. Lohe, P. W. M. Blom and X. Feng, Angew. Chem. Int. Ed., 2018, 57, 15491–15495.
50. A. Sinha, Dhanjai, H. Zhao, Y. Huang, X. Lu, J. Chen and R. Jain, TrAC, Trends Anal. Chem., 2018, 105, 424–435.
51. W. Huang, L. Hu, Y. Tang, Z. Xie and H. Zhang, Adv. Funct. Mater., 2020, 30, 2005223.
52. Q. Tang, Z. Zhou and P. Shen, J. Am. Chem. Soc., 2012, 134, 16909–16916.
53. M. Han, X. Yin, H. Wu, Z. Hou, C. Song, X. Li, L. Zhang and L. Cheng, ACS Appl. Mater. Interfaces, 2016, 8, 21011–21019.
54. Z. Ling, C. E. Ren, M.-Q. Zhao, J. Yang, J. M. Giammarco, J. Qiu, M. W. Barsoum and Y. Gogotsi, Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 16676–16681.
55. S. Wang, H.-Q. Shao, Y. Liu, C.-Y. Tang, X. Zhao, K. Ke, R.-Y. Bao, M.-B. Yang and W. Yang, Compos. Sci. Technol., 2021, 202, 108600.
56. D. Wang, D. Zhang, P. Li, Z. Yang, Q. Mi and L. Yu, Nano-Micro Lett., 2021, 13 DOI:10.1007/s40820-020-00580-5.
57. L. Li, X. Fu, S. Chen, S. Uzun, A. S. Levitt, C. E. Shuck, W. Han and Y. Gogotsi, ACS Appl. Mater. Interfaces, 2020, 12, 15362–15369.
58. Y. Yue, N. Liu, W. Liu, M. Li, Y. Ma, C. Luo, S. Wang, J. Rao, X. Hu, J. Su, Z. Zhang, Q. Huang and Y. Gao, Nano Energy, 2018, 50, 79–87.
59. S. Seyedin, S. Uzun, A. Levitt, B. Anasori, G. Dion, Y. Gogotsi and J. M. Razal, Adv. Funct. Mater., 2020, 30, 1910504.
60. A. Levitt, S. Seyedin, J. Zhang, X. Wang, J. M. Razal, G. Dion and Y. Gogotsi, Small, 2020, 16, 2002158.
61. J. Zhang, L. Wan, Y. Gao, X. Fang, T. Lu, L. Pan and F. Xuan, Adv. Electron. Mater., 2019, 5, 1900285.
62. S. Sharma, A. Chhetry, M. Sharifuzzaman, H. Yoon and J. Y. Park, ACS Appl. Mater. Interfaces, 2020, 12, 22212–22224.
63. L. Zhang, S. Zhang, C. Wang, Q. Zhou, H. Zhang and G.-B. Pan, ACS Sens., 2021, 6, 2630–2641.
64. L. Wang, M. Zhang, B. Yang, J. Tan, X. Ding and W. Li, Small Methods, 2021, 5, 2100409.
65. H. Riazi, G. Taghizadeh and M. Soroush, ACS Omega, 2021, 6, 11103–11112.
66. H. Souri, H. Banerjee, A. Jusufi, N. Radacsi, A. A. Stokes, I. Park, M. Sitti and M. Amjadi, Adv. Intell. Syst., 2020, 2, 2000039.
67. L. Liu, L. Wang, X. Liu, W. Yuan, M. Yuan, Q. Xia, Q. Hu and A. Zhou, Nanomaterials, 2021, 11, 889.
68. T. Sekine, M. Abe, K. Muraki, S. Tachibana, Y.-F. Wang, J. Hong, Y. Takeda, D. Kumaki and S. Tokito, Adv. Intell. Syst., 2020, 2, 2000179.
69. L. Duan, D. R. D'Hooge and L. Cardon, Prog. Mater. Sci., 2020, 114, 100617.
70. M. Amjadi, K.-U. Kyung, I. Park and M. Sitti, Adv. Funct. Mater., 2016, 26, 1678–1698.
71. J.-H. Pu, X. Zhao, X.-J. Zha, L. Bai, K. Ke, R.-Y. Bao, Z.-Y. Liu, M.-B. Yang and W. Yang, J. Mater. Chem. A, 2019, 7, 15913–15923.
72. A. Ahmed, M. M. Hossain, B. Adak and S. Mukhopadhyay, Chem. Mater., 2020, 32, 10296–10320.
73. X. Shi, H. Wang, X. Xie, Q. Xue, J. Zhang, S. Kang, C. Wang, J. Liang and Y. Chen, ACS Nano, 2019, 13, 649–659.
74. Y.-N. Yang, R.-R. Wang and J. Sun, J. Inorg. Mater., 2019, 35, 8–18.
75. Y. Cai, J. Shen, G. Ge, Y. Zhang, W. Jin, W. Huang, J. Shao, J. Yang and X. Dong, ACS Nano, 2018, 12, 56–62.
76. Y. Lu, M. C. Biswas, Z. Guo, J. W. Jeon and E. K. Wujcik, Biosens. Bioelectron., 2019, 123, 167–177.
77. S. Zhao, J. Li, D. Cao, G. Zhang, J. Li, K. Li, Y. Yang, W. Wang, Y. Jin, R. Sun and C.-P. Wong, ACS Appl. Mater. Interfaces, 2017, 9, 12147–12164.
78. Y. Yang, L. Shi, Z. Cao, R. Wang and J. Sun, Adv. Funct. Mater., 2019, 29, 1807882.
79. S. Abdolhosseinzadeh, R. Schneider, A. Verma, J. Heier, F. Nüesch and C. Zhang, Adv. Mater., 2020, 32, 2000716.
80. Z. Ma, K. Ma, S. Lu, S. Wang, X. Liu, B. Li, L. Zhang and X. Wang, Sens. Actuators, A, 2020, 315, 112304.
81. W. Yuan, X. Qu, Y. Lu, W. Zhao, Y. Ren, Q. Wang, W. Wang and X. Dong, Chin. Chem. Lett., 2021, 32, 2021–2026.
82. K. Yang, F. Yin, D. Xia, H. Peng, J. Yang and W. Yuan, Nanoscale, 2019, 11, 9949–9957.
83. M. Chao, Y. Wang, D. Ma, X. Wu, W. Zhang, L. Zhang and P. Wan, Nano Energy, 2020, 78, 105187.
84. G. M. N. Islam, A. Ali and S. Collie, Cellulose, 2020, 27, 6103–6131.
85. J. S. Heo, J. Eom, Y.-H. Kim and S. K. Park, Small, 2018, 14, 1703034.
86. S. Seyedin, P. Zhang, M. Naebe, S. Qin, J. Chen, X. Wang and J. M. Razal, Mater. Horiz., 2019, 6, 219–249.
87. X. Li, Y. J. Fan, H. Y. Li, J. W. Cao, Y. C. Xiao, Y. Wang, F. Liang, H. L. Wang, Y. Jiang, Z. L. Wang and G. Zhu, ACS Nano, 2020, 14, 9605–9612.
88. J. Kim, E. F. Chou, J. Le, S. Wong, M. Chu and M. Khine, Adv. Healthcare Mater., 2019, 8, 1900109.
89. Y.-Z. Zhang, K. H. Lee, D. H. Anjum, R. Sougrat, Q. Jiang, H. Kim and H. N. Alshareef, Sci. Adv., 2018, 4, eaat0098.
90. Z. Li, B. Zhang, K. Li, T. Zhang and X. Yang, J. Mater. Chem. C, 2020, 8, 3088–3096.
91. Y. Xiong, Y. Shen, L. Tian, Y. Hu, P. Zhu, R. Sun and C.-P. Wong, Nano Energy, 2020, 70, 104436.
92. G. Li, S. Liu, L. Wang and R. Zhu, Sci. Robot., 2020, 5, eabc8134.
93. S. Lee, S. Franklin, F. A. Hassani, T. Yokota, M. O. G. Nayeem, Y. Wang, R. Leib, G. Cheng, D. W. Franklin and T. Someya, Science, 2020, 370, 966–970.
94. A. Georgopoulou and F. Clemens, ACS Appl. Electron. Mater., 2020, 2, 1826–1842.
95. Z. Peng, X. Zhang, C. Zhao, C. Gan and C. Zhu, Mater. Chem. Phys., 2021, 271, 124729.
96. X. Tang, D. Cheng, J. Ran, D. Li, C. He, S. Bi, G. Cai and X. Wang, Nanotechnol. Rev., 2021, 10, 221–236.
97. Y. Ma, N. Liu, L. Li, X. Hu, Z. Zou, J. Wang, S. Luo and Y. Gao, Nat. Commun., 2017, 8, 1207.
98. M.-Y. Liu, C.-Z. Hang, X.-F. Zhao, L.-Y. Zhu, R.-G. Ma, J.-C. Wang, H.-L. Lu and D. W. Zhang, Nano Energy, 2021, 87, 106181.
99. V. Adepu, V. Mattela and P. Sahatiya, J. Mater. Chem. B, 2021, 9, 4523–4534.
100. H. Liu, M. Dong, W. Huang, J. Gao, K. Dai, J. Guo, G. Zheng, C. Liu, C. Shen and Z. Guo, J. Mater. Chem. C, 2017, 5, 73–83.
101. E. Ismar, S. Kurşun Bahadir, F. Kalaoglu and V. Koncar, Global Challen., 2020, 4, 1900092.
102. R. Liu, J. Li, M. Li, Q. Zhang, G. Shi, Y. Li, C. Hou and H. Wang, ACS Appl. Mater. Interfaces, 2020, 12, 46446–46454.
103. Y. Zheng, R. Yin, Y. Zhao, H. Liu, D. Zhang, X. Shi, B. Zhang, C. Liu and C. Shen, Chem. Eng. J., 2021, 420, 127720.
104. X. P. Li, Y. Li, X. Li, D. Song, P. Min, C. Hu, H. B. Zhang, N. Koratkar and Z. Z. Yu, J. Colloid Interface Sci., 2019, 542, 54–62.
105. D. Song, X. Li, X. P. Li, X. Jia, P. Min and Z. Z. Yu, J. Colloid Interface Sci., 2019, 555, 751–758.
106. Y. Zhang, K. Chen, Y. Li, J. Lan, B. Yan, L. Shi and R. Ran, ACS Appl. Mater. Interfaces, 2019, 11, 47350–47357.
107. J. Liu, H. B. Zhang, X. Xie, R. Yang, Z. Liu, Y. Liu and Z. Z. Yu, Small, 2018, 14, 1802479.
108. Y. Ma, Y. Yue, H. Zhang, F. Cheng, W. Zhao, J. Rao, S. Luo, J. Wang, X. Jiang, Z. Liu, N. Liu and Y. Gao, ACS Nano, 2018, 12, 3209–3216.
109. Y.-Z. Zhang, J. K. El-Demellawi, Q. Jiang, G. Ge, H. Liang, K. Lee, X. Dong and H. N. Alshareef, Chem. Soc. Rev., 2020, 49, 7229–7251.

---