Piezoelectricity in monolayer MXene for nanogenerators and piezotronics

Highlights

• The first study on the piezoelectricity of monolayer Ti₃C₂Tx MXene is reported, showing the efficient piezoelectric outputs.

• The functional groups on the MXene break the inversion symmetry of lattice structures to possess piezoelectric properties.

• The response current of Ti₃C₂Tx MXene is 0.3 nA, the power density is 6.5 mW/m² and the conversion efficiency is 11.15%.

• The discovery of piezo-MXene can lay the foundation for understanding and applying Ti₃C₂Tx MXene in self-powered nanodevices.

Abstract

The piezoelectric properties of two-dimensional (2D) materials have been widely studied due to their broad application prospects. MXene, one of the well-known 2D material members, is predicted to be a highly directional piezoelectric material with a non-centrosymmetric lattice structure. Here, the first experimental study of piezoelectric responses of the monolayer Ti₃C₂Tx MXene is reported, showing that the cyclic strain excites stable oscillating piezoelectric voltage and current outputs. The functional groups on the surface of the MXene break the inversion symmetry of lattice structures to possess piezoelectric properties. The piezoelectricity in the armchair direction of the Ti₃C₂Tx MXene sheet exhibits an intrinsic current output of 0.3 nA at 1.08% tensile strain, corresponding to the 6.5 mW/m² power density and the 11.15% conversion efficiency, which are all higher than that of previous reported 2D materials. Further, theoretical calculations of the MXene have explained the origin of piezoelectric polarizations among the multi-atomic structure and the surface functional groups. The discovery of the piezo-MXene can lead to the foundation for the understanding and applications of the Ti₃C₂Tx MXene in powering nanodevices and stretchable electronics.

Introduction

The asymmetric structure of the crystal determines the natural interaction with the external stimuli [1], [2], [3], [4]. As represented by ferroelectrics and wurtzite structure materials, the piezoelectricity generated by the unique asymmetric structure makes them to have broad potentials in the fields of wearable devices, energy conversion devices, electronics, and sensors [5], [6], [7], [8], [9], [10], [11]. Recent studies have demonstrated that the destructions of the inversion symmetry in the plane of layered materials, such as multilayer black phosphorous (BP) [12], [13] and single-atomic-layer molybdenum disulfide (MoS₂) [14], [15], exhibit inherent piezoelectric properties. On the other hand, the strain-caused lattice distortion can also produce the piezoelectricity and the related ion charge polarization to make the applications of 2D nanomaterials in nanoscale electromechanical devices, which is confirmed in hexagonal boron nitride (h-BN) [16], [17], [18], [19], [20], [21], [22] and transition-metal dichalcogenides (TMDS) [23], [24], [25], [26], [27], [28], [29], [30]. MXene, a type of the transition metal carbide/nitride with a 2D layered structure, has lots of advanced properties as high conductivity, high surface area, and abundant surface functional groups [31], [32], [33], [34], [35], [36]. The unique properties of the Ti₃C₂Tx MXene are able to provide distinct application capabilities, such as dynamic sensors, stress-strain sensors, nanogenerators, piezoelectric tuned transistors, etc. Several calculated properties of the MXene have already revealed the asymmetric mechanisms of the upper and lower surfaces of the functionalized MXene [37], [38], [39], [40], and the crystal structure of the Ti₃C₂Tx MXene can be analyzed according to the non-neutralized polarity such as the piezoelectric properties of oxygen-functionalized MXene (M₂CO₂, M = Sc, Y and La) have been reported in previous studies, revealing the asymmetric and piezoelectric mechanisms of the upper and lower surfaces of the functionalized MXene [37]. However, the piezoelectricity of the monolayer Ti₃C₂Tx MXene device has not been elaborated from the experimental testament [40], [41].

In this paper, an asymmetric atomic structure model is proposed to indicate the mechanism of the piezoelectric polarization, and the anisotropic Raman spectra of the deformed Ti₃C₂Tx MXene are systematically analyzed. Then, a piezoelectric measuring device, based on the monolayer Ti₃C₂Tx MXene, is prepared on a polyethylene terephthalate (PET) substrate, and the corresponding strains are loaded on the Ti₃C₂Tx MXene with different bending stresses. For the piezo-MXene, the transport of carriers on the MXene-metal potential barrier can be regulated by the strain-induced polarization charges, which enhance the strain sensing effect. These studies can afford the potential of the Ti₃C₂Tx MXene in energy conversion units, adaptive bio-probes, and other flexible electronic devices.

Fig. 1a presents the atomic lattice structure of the monolayer Ti₃C₂-based MXene (left) and Ti₃C₂Tx with end-capping group (right). Three kinds of color balls are used to identify different atoms and groups; blue and purple as titanium (Ti) and carbon (C) atoms respectively, and green as functional groups (Tx) [42]. According to the fabricated samples in the experiments, the Tx functional group relates to the -OH [42], [43], [44], [45]. (The detailed synthetic strategy of the monolayer Ti₃C₂Tx MXene, X-ray photoelectron spectroscopy (XPS) and the x-ray diffraction (XRD) analysis are shown in Supporting part 1.) In the fabrication, the monolayer Ti₃C₂Tx MXene solution with deionized water is coated on the PET substrate and kept in a vacuum heating oven at 60 °C to remove moisture and avoid oxidation. The monolayer Ti₃C₂Tx MXene sheets are identified by an optical microscope (The details in Supporting Fig. S2) and tested by an atomic force microscope (AFM). As demonstrated in Fig. 1b, the Ti₃C₂Tx MXene with a thickness of ~ 1.5 nm is measured by non-contact AFM, which is confirmed the monolayer parameter in previous works [46], [47], [48], [49]. Then, the crystal orientation of the monolayer Ti₃C₂Tx MXene can be determined by the optical second-harmonic method (SHG), and Fig. 1c shows the polar plot of the second harmonic (SH) signal intensity from the monolayer Ti₃C₂Tx MXene as a function of the crystal azimuth [50], [51]. The measured SH component is perpendicular to the induced polarization. The crystal orientation intensity is determined by fitting the angle dependence of SH and $I=I_0{\mathit{sin}}^2(3\theta )$, where θ is the angle between the armchair direction and the polarized laser, and I₀ is the maximum intensity of SH response [52]. (The schematic diagrams of the SHG and the detailed measurement methods are shown in Supporting part 3.) After the optical characterizations, the monolayer Ti₃C₂Tx MXene can be prepared as the piezoelectric measurement device. Fig. 1d shows the optical microscope image of a piezoelectric response, measuring the device with a typical monolayer Ti₃C₂Tx MXene. The metal-MXene interface is constructed by physical deposition of Au electrodes. The monolayer Ti₃C₂Tx MXene sheet is outlined by the black dotted line, and the inset illustration is an optical photo of the measurement device prepared on a PET substrate. The crystal lattice of the Ti₃C₂Tx MXene in this device is superimposed on the optical image of the MXene, and then the strain can be applied in the armchair direction. These well-connected electrode-MXene compositions can not only achieve stable mechanical stability, but also transmit electrical signals completely. With the bending of the PET substrate, the uniaxial strain can be loaded to the prepared Ti₃C₂Tx MXene. (The detailed fabrication process of the flexible device based on monolayer Ti₃C₂Tx MXene is described in Supporting part 4.) With the process in Fig. 1e, the piezoelectric responses with an external resistor are investigated by applying strains. Such a structure makes the sample subjected to strain, and the strain induces charge polarization, driving electrons to flow through the external circuit. And, the corresponding electrons will flow back in the opposite direction as the substrate is released.

Next, the three-dimensional (3D) structure is simplified to x-y and x-z planes for the in-plane deformation analysis in order to describe the piezoelectric effect of the monolayer Ti₃C₂Tx MXene more intuitively. Fig. 1f is a top view of the MXene crystal structure, where the monolayer Ti₃C₂Tx MXene consists of Ti and C atoms with a honeycomb lattice structure and surface groups at both ends [53], [54]. (The crystal structure of Ti₃C₂-based MXene is given in Supporting part 5.) The bare Ti₃C₂-based MXene is a crystal structure composed of Ti atomic layers and C atomic layers. The center atomic layer is composed of Ti atoms, with a layer of C atoms on each side, and the outermost is a layer of Ti atoms on both sides. These Ti atoms can provide active bonds to combine with functional groups [42], [55], and the non-equivalent positions of these functional groups can break the inversion symmetry of the crystal structure. The crystal structure, atomic composition, and arrangement of Tx as a 2D hexagonal structure lack inversion centers in and out of the plane in the direction of the x-y plane, which causes the monolayer Ti₃C₂Tx MXene to produce in-plane piezoelectric properties [56], [57]. In the x-y plane as presented in Fig. 1g, the Ti₃C₂Tx MXene can be divided into two kinds of hexagonal structures composed of Ti-Tx atoms and Ti-C atoms, in which the x-axis is defined as the direction of the armchair. When the tension is applied along the x-axis, the TiTx bonds will be subjected to the force to produce spatial motion along the armchair direction in the x-y plane. The functional groups of Tx will destroy the symmetry of the structure, and generate charge polarization in the hexagonal element, thus generating the piezoelectric field. Before loading external forces, Ti and Tx atoms form electric dipole moments P₁, P₂ and P₃ (the left illustration in Fig. 1g), and positive and negative charge centers overlap present a neutral state as P₁ + P₂ + P₃ = 0. When the external strain acts on the Ti₃C₂Tx MXene (the right of Fig. 1g), the positive and negative charges polarize with increased P₁ and decreased P₂ and P₃, and produce a piezoelectric field in the corresponding direction. Similarly, the x-axis tension on the TiC bond will also cause structural deformation, the neutral state of the electric dipole moments P₄, P₅, and P₆ are destroyed, resulting in charge polarization phenomenon, and generating piezoelectric fields in the same direction. It is noted that the piezoelectric efficiency generated by Ti-C polarization is secondary [37], [58], and the piezoelectric effect of the monolayer Ti₃C₂Tx MXene is mainly due to the non-centrosymmetric lattice structure caused by the functional groups at both ends of the x-y plane [37], [40]. Fig. 1h shows a side view of the Ti₃C₂Tx MXene, where the multilayer atomic structure can be divided into Ti-Tx atom-functional group and Ti-C atom-atom combinations with the central Ti atoms layer as the symmetry axis. With the stretching along the x-axis in Fig. 1i, the asymmetric structure of Ti-Tx is under tension and deformed in the x-z plane, resulting in electrical polarization and leading to the generation of the piezoelectric field. The large number of surface groups can be stably adsorbed on both sides of the Ti₃C₂Tx MXene to produce a strong piezo-response. Different from 2D BP, the piezoelectric effect of the Ti₃C₂Tx MXene is caused by the break of the inversion symmetry from uneven distribution of functional groups, and lays a good foundation for the preparation of the Ti₃C₂Tx MXene piezoelectric devices [12]. Because the monolayer Ti₃C₂Tx MXene is a repeating unit of the Ti-Tx and Ti-C structure, the monolayer Ti₃C₂Tx MXene is reasonable to own in-plane piezoelectric properties under external strains [17], [27], [37], [59].

To study the corresponding atomic structures, the Raman spectra of the monolayer Ti₃C₂Tx MXene are systematically analyzed with different situations. The Raman spectra of the monolayer Ti₃C₂Tx MXene with a 633 nm wavelength laser are shown in Fig. 2a, which can be divided into four main regions to relate with different atomic structures [60]. The first region is the plasma peak coupled resonant peak, and the second region is the flake region consisting of E_g (Ti, C, O) and A_1g (Ti, C, O) modes, which represents vibrations of Ti, C, and surface groups. The 230–470 cm⁻¹ region represents the vibrations of functional groups on the surface attached to Ti atoms in the plane (E_g), which is only affected by the atoms at the surface. The areas between 580 cm⁻¹ and 730 cm⁻¹ are mainly used for C vibrations (E_g and A_1g) [61]. For the Raman spectra of the Ti₃C₂Tx MXene, a peak as the transition metal and surface functional group (M-Tx) appears around 599 cm⁻¹, standing for the vibrational relationship between the transition metal and functional group bonds [60]. To prove the piezoelectric effect caused by the deformation of the crystal structure, the Raman spectra under the different strains are measured and recorded. As shown in Fig. 2b, different degrees of strains (from 0% to 1.5%) are exerted on the Ti₃C₂Tx MXene by the PET substrate, and the corresponding peak values put forward the different drift. (The strain calculation method and schematic diagram are shown in Supporting part 6.) With obtaining the Raman spectra of the monolayer Ti₃C₂Tx MXene under the 633 nm laser, the resonance Raman effect begin to form at 120 cm⁻¹ (dash line 1 in Fig. 2b). However, such resonance, limited by the weak signal of the monolayer, can be reasonably ignored [62], [63]. The vibration of the flake region of the monolayer Ti₃C₂Tx MXene is the hardest, as a group vibration containing Ti, C, and surface groups with the maximum atomic number unit information of the MXene. And, there is no Raman displacement when the external strain is applied [60], [64], and the peak value of A_1g (Ti, O, C) at 220 cm⁻¹ (dash line 2 in Fig. 2b) remains stable during the straining process. The limited vibration outside the plane under the action of the substrate restricts the whole motion, which causes little impact on the vibration of the whole atoms [60]. For the Tx region (dash line 3 in Fig. 2b), the vibration of functional groups begins to become sensitive, as shown in Fig. 2c. Under the action of the tension, the Raman wave peaks located near 378 cm⁻¹ moved to a lower direction with the increased tensions, indicating that the Tx bond has a spatial deflection under the action of the stress. This wave peak shows the vibration mode of the layer of functional groups on the Ti atoms [60]. The microscopic position deflection of the functional groups leads to the shift of the corresponding peak value in the Raman spectra, which directly indicates the importance of the functional groups in the generation process of the piezoelectric effect. The crest of 599 cm⁻¹ (dash line 4 in Fig. 2b) represents the change of the vibration mode of M-Tx as the change of the vibration mode of Ti-Tx [60]. Therefore, the deformation of M-Tx is the main part of piezoelectric polarizations by the inversion of the symmetry failure. In Fig. 2d, the displacement of the 599 cm⁻¹ wave peaks shows the change of the vibration mode of M-Tx under the stress, as shown in the simplified deformation model of the x-y plane and x-y plane above. The process is caused by the spatial movement, resulting in a change of the vibration mode with the angular deflection and the atomic displacement, and leading to the formation of piezoelectric polarizations. The C region contains two important parts, one is the C-Tx vibration peaks near 620 cm⁻¹, the other is the C atomic layer resonance peaks of A_1g (C) near 720 cm⁻¹ [65]. As shown in Fig. 2e, the TiC bond (dash line 5 in Fig. 2b) forms in-space deflection and atomic displacement from the atomic bond displacement and the spatial angle change of laminated atomic structures. In Fig. 2f, the resonance state located at 580–730 cm⁻¹ is used to describe the vibration of the C atomic layer [60]. The vibration state of the C atoms centered on the Ti atomic layer is presented in the form of E_g and A_1g at the same time. The change of their vibration pattern is shown in the peak shift of the Raman spectra (dash line 6 in Fig. 2b). The Raman spectroscopy analysis of the monolayer Ti₃C₂Tx MXene confirm the existence of the displacement of the original structure, and the deflection of the bond under the strain proves the formation of the piezoelectric polarization and the generation of the piezoelectric phenomenon, which is caused by the change of the atomic structure in the non-centrosymmetric Ti₃C₂Tx MXene crystal structure.

To investigate the piezoelectricity of the Ti₃C₂Tx MXene, the electrodes composed of Au are deposited on the PET substrate to prepare a piezoelectric nanogenerator [47], [66], [67], [68]. (The preparation process of the piezoelectric nanogenerator is detailed in the Supporting part 7.) Fig. 3a exhibits the generation of the piezoelectric phenomenon and the principle of energy conversions in the monolayer Ti₃C₂Tx MXene. When elastic deformations are exerted to the Ti₃C₂Tx MXene through the bent PET substrate, the deformation of the non-centrosymmetric lattice can produce piezoelectric polarizations [12], [14], [15]. The Ti₃C₂Tx MXene with OH functional group as a narrow bandgap of 0.40 eV, can produce Schottky contact with Au electrode during contacts [69], [70], and the piezo-charges generated by piezoelectric polarization can accumulate in the Schottky contact during the deformation process. In the process of the charge generation, the energy band decreases with the accumulation of the positive piezoelectric charges and increases with the accumulation of negative charges. Thus, the Schottky barrier changes asymmetrically at the source and drain ends [59]. Therefore, the carrier transport can be tuned as a control gate by using a shaped piezoelectric potential. When the deformed Ti₃C₂Tx MXene device is connected to an external circuit, electrons will flow into the external circuit because of the unbalanced state, which is formed by the piezoelectric potential at the Fermi level [71], [72]. And, the piezoelectric potential disappears as the strain is set free, and the electrons will flow back to the original state. The strain in the armchair direction induces an asymmetric change in the current. As the strain variable increases, the I-V curve shifts downward, which is shown in Fig. 3b. With the action of the piezoelectric effect, the current under the positive bias voltage and negative bias voltage will produce asymmetric changes, that is the piezoelectric potential to the Schottky barrier height asymmetric modulation. (The more details of the corresponding piezoelectric current of the I-V curve shows in Fig. S7) The piezoelectric polarization in the armchair direction enables the Ti₃C₂Tx MXene to play the role of regulating the carrier transport like a transistor gate modulation, such as BP [12], zinc oxide (ZnO) [73], [74], [75], and MoS₂ [58], [76], [77], [78]. The change of the Schottky barrier height (${∆\varnothing }_{\mathit{SB}}$) in the armchair direction is estimated in Fig. 3c. (Supporting part 9 for the calculation process). The regulation of the Schottky barrier height by the piezoelectric potential is an asymmetric regulation, the variation of the Schottky barrier increases gradually with the increase of strains. For the Ti₃C₂Tx MXene, the piezoelectric property in the zigzag direction exists a weak piezo-performance as that of the armchair direction. The change of the electric dipole moment along the direction of the armchair causes the piezoelectric polarization to generate the piezoelectric potential. Since the electric dipole moment vector direction in the hexagonal Ti-Tx and Ti-C atomic structure changes along the direction of the armchair in the process of stretching or compression, the piezoelectric polarization phenomenon cannot occur in the zigzag direction, so the piezoelectric effect in the zigzag direction is weak. As presented in Fig. 3d, the strain in the zigzag direction of the Ti₃C₂Tx MXene causes the I-V curve to move downward while the tensile strain increases. In Fig. 3e, the Schottky barrier change (${∆\varnothing }_{\mathit{SB}}$) in the zigzag direction also presents the corresponding proportional strain trend, which is weaker than that in the armchair direction.

Then, the piezoelectric nanogenerator of the Ti₃C₂Tx MXene along with the armchair and zigzag directions are prepared to characterize the piezoelectric responses. In the piezoelectric devices, the strain causes a current of piezoelectric charges to enter the external circuit. When the deformation of the substrate is set free, the charge will flow back. Fig. 3f reveals the piezoelectric response signal of the Ti₃C₂Tx MXene device with the piezoelectric output current under the periodic strain, and the Ti₃C₂Tx MXene with the period (1.3 s) of the compression and tensile strain (1.08%). (Supporting part 10 for the cycle control system.) It is obvious from Fig. S9 that the piezoelectric output is almost the same under different bending cycle times. The piezoelectric nanogenerator based on Ti₃C₂Tx MXene obtains energy through cyclic bending, and its efficiency is determined by the cycle time. The output current is negative when the compressive strain is applied, and the output current is positive when the strain is released. The piezoelectric current in the armchair direction is greater than the piezoelectric current in the zigzag direction, which is consistent with the previous prediction. The output voltage is measured with 1 MΩ load resistance. The related current response signal of the piezoelectric nanogenerator in the armchair direction is shown in Fig. 3g, and the loading strain of the constant peak current signal is 1.08% with ~ 5000 cycles in a single period of ~ 1.3 s. The peak response current reaches ~ 0.3 nA (maximum current density ~ 0.4 × 10⁷ A/m²), and the current signal shows a little damping (< 5%), which may be related to the oxidation of Ti₃C₂Tx MXene and mechanical fatigue/structural damage [58], [66]. The maximum power transferred to the load is 7.83 × 10⁻¹⁴ W at 1.08% strain, and the corresponding power density is 6.5 mW/m². The conversion efficiency of the monolayer Ti₃C₂Tx MXene nanogenerator can be estimated to be ~ 11.15% (MoS₂ as 5.08% and BP as 5.60%) [12], [14], that is, the ratio of the electrical energy transmitted to the load to the total mechanical deformation energy stored after stretching the monolayer Ti₃C₂Tx MXene. (The relevant calculation methods are detailed in Supporting part 12. When the same electrodes are prepared on the PET without Ti₃C₂Tx MXene, no interference current signal is generated under the strain, as shown in Supporting part 13, which also ensured the accuracy of piezoelectric response current measurement.) Fig. 3h is the detailed variation of the piezoelectric output current during the strain cycle. The piezoelectric electronics and elastic theories evident that the electrical output of the piezoelectric is proportional to its elasticity [20]. In the strain cycle, as the strain increases, the piezoelectric charge is generated and accumulated continuously until reaching the strain limit or saturation state. With the decrease of the strain, the piezoelectric charges lose the regulation of the piezoelectric potential, and the piezoelectric charges return to the original state gradually [2]. As shown in Fig. 3i, the response currents of the piezoelectric nanogenerator can reach a saturation value of 0.3 nA at 1.08% strain, and the piezoelectric currents decrease with the decreased strains. The magnitude of the current tends to level off under 1.08% strain because the piezoelectric charges recombine with free holes in Ti₃C₂Tx MXene with increasing strain. When the polarity of the connection is reversed, the signal direction and amplitude change significantly, indicating that the majority of the output current comes from monolayer MXene and only a small part comes from noise interference. (The piezoelectric response before and after polarity reversal is shown in Supporting part 14.) Fig. 3j exhibits the histogram of the current outputs and voltage responses when the strains are applied with the 1 MΩ resistance. The piezoelectric properties of the Ti₃C₂Tx MXene are characterized at different strains and measured at a constant strain for a long time. (For the Ti₃C₂Tx MXene in zigzag direction, the Supporting part 15 shows the relevant piezoelectric current responses process in detail.) In conclusion, this energy conversion phenomenon, occurring in the Ti₃C₂Tx MXene, proves the existence of in-plane piezoelectricity and can promote the function of the piezoelectric nanogenerator.

Finally, the monolayer Ti₃C₂Tx MXene array integrated as a chip device is demonstrated to improve the piezoelectric output energy conversion. By combining two slices of the monolayer Ti₃C₂Tx MXene in parallel or series, the increase or decrease of the outputs can be treated as superposition. As shown in Fig. 4a, the two MXene devices with the same polarization direction are arranged in parallel to determine the working state of the piezoelectric current. The response current of devices 1 and 2 are slightly different under the same 0.50% strain state, and the current presents linear assembly under parallel conditions (Fig. 4b). Fig. 4c reveals the Ti₃C₂Tx MXene devices with the same polarization direction in a series state. Under a strain of 0.50%, the individual voltage responses of devices 1 and 2 also exhibit an added state with the series voltage in Fig. 4d. In contrast, when the parallel or series Ti₃C₂Tx MXene devices with the reverse polarity, the corresponding signals can also be derived from the experiments. Fig. 4e shows the schematic diagram of the parallel MXene devices with opposite polarization directions, and the piezoelectric current is the difference set between devices 1 and 2 in Fig. 4f. Fig. 4h indicates the corresponding piezoelectric voltage with the opposite piezoelectric polarization directions in series. These piezoelectric voltages in series present the difference between the device 1 and 2 in Fig. 4g. In addition, the piezoelectric distribution in multilayer Ti₃C₂Tx MXene is shown in Supporting part 16. The interactions between the piezoelectric current and voltage of the combined devices create the possibility of realizing a larger scale tunable 2D piezoelectric nano-devices based on the MXene, including the fields of wearable electronic devices, sensors, energy collection devices, and so on.

In conclusion, the piezoelectric properties of the monolayer Ti₃C₂Tx MXene have been proved by various methods, including the analysis of the deformation principle of the asymmetric lattice structure, the analysis of the Raman spectra, the provability of the piezoelectric effect, and the measurement of piezoelectric signals by a variety of 2D device designs. The discussion of the piezoelectric polarity and the proof of the superposition theory permit the foundation for the further development of multifunctional piezoelectric devices. As a new 2D material, the piezo-performance of the Ti₃C₂Tx MXene has positive significance for the development of energy harvesting devices, electromechanical sensors, energy storage devices, and transistors.