Introduction

MXenes are an emerging family of two-dimensional transition metal carbides and nitrides with a unit formula of Mn+1XnTx (n = 1–3), where M represents an early transition metal (such as Sc, Ti, V, Cr, Mn, Y, Zr, Nb, Mo, Hf, Ta, et al.), X corresponds to nitrogen or carbon, Tx stands for the surface functional groups (such as -OH, -F, =O)1,2,3. By virtue of the high conductivity and solution processability, MXenes show great potential in Joule heating, energy conversion/storage, electromagnetic interference (EMI) shielding, etc. For instance, compared to the conventional dense metallic films and graphene with hydrophobic surface4,5, self-assembled ultrathin MXene film (a 24-layer film of ≈55 nm thickness) is capable of providing 99% reflection shielding (EMI SE of 20 dB) in microwave band6. Notably, besides the traditional microwave band, MXenes are demonstrated as highly efficient shielding or absorbing material for higher-frequency terahertz waves (THz, frequency ranging from 0.1 to 10 THz), which is assigned for 6 G wireless communication and has many potential applications in aerospace industry7,8,9,10. However, the good performance of MXenes is always achieved at room temperature or low-temperature due to their nature of poor antioxidation properties in air and humid environment at elevated temperature, attributed to thermodynamically metastable surface11. This may greatly limit their further application towards practical high-temperature scenarios (for instance, a heater is required to work under hundreds of degrees in some cases, EMI shielding of engine casting of aircraft needs to endure high temperatures above 380 °C12, etc.).

Previously, the oxidation mechanisms of MXenes, especially for Ti3C2Tx, have been investigated13,14,15,16,17. In general, both oxygen and water molecules are responsible for the oxidation of the Ti3C2Tx, even Ti3C2Tx preserved in degassed water would be oxidized after a while14. This oxidation process would be accelerated at elevated temperatures with the presence of oxygen or water molecules18. Accordingly, numerous efforts have been focusing on the antioxidation of Ti3C2Tx, mainly by eliminating O2 and/or H2O from contacting Ti3C2Tx, which could be summarized into three strategies. First, changing the storage condition of Ti3C2Tx (such as ultralow freezing temperature19, inert or reductive atmosphere13,20, organic solvents21) is an effective way to protect the flakes from oxidation. Second, applying antioxidants (such as ascorbate22 and inorganic salts23) can prolong the storage lifetime of Ti3C2Tx MXene in solutions. Third, a homogeneous surface encapsulation strategy has been used to isolate Ti3C2Tx from O2 and/ or H2O24,25. However, the above strategies mainly concentrate on the antioxidation of Ti3C2Tx in an aqueous solution or room temperature. The high-temperature resistant property of Ti3C2Tx in the air or humid environment for a long operation time remains unexplored while highly important.

Herein, we report an air-stable Ti3C2Tx MXene-based composite that could endure high-temperature annealing in the air. Restacked film composed of pristine Ti3C2Tx flakes with metallic conductivity up to 24,000 S cm−1 is a good candidate for applications requiring high conductivity. However, the oxidation process of Ti3C2Tx into TiO2 with the presence of oxygen under high temperature largely decreases its conductivity, as shown in Fig. 1. To solve this issue, we propose a competition mechanism by compositing Ti3C2Tx with extracted bentonite (EB). EB is a layered material consisting of one Al octahedral sheet sandwiched between two Si tetrahedral nanosheets with Na+ ions to offset the charge imbalance26 (bottom left corner in Fig. 1). Interestingly, EB shows stronger O2 adsorption compared to Ti3C2Tx. Even after saturated adsorption of O2 on EB, further O2 molecules are still repelled onto the surface of Ti3C2Tx due to the coupling between Ti3C2Tx and EB. In this context, Ti3C2Tx/EB (MEB) composite could endure annealing at above 400 °C for several hours in an air and humid environment.

Fig. 1: Schematic shows the oxidation process of pristine Ti3C2Tx (starting from the edge) and suppressed oxidation of MEB under high-temperature annealing with the presence of oxygen.
figure 1

The top row shows the oxidation of pristine Ti3C2Tx, which is induced by interacting with oxygen molecules, and Ti3C2Tx finally transforms to TiO2. The bottom row shows the suppressed oxidation of MEB, which is induced by the prior adsorption of oxygen molecules on EB compared to Ti3C2Tx. Even after saturated adsorption of O2 on EB, further O2 molecules are still repelled onto the surface of Ti3C2Tx due to the coupling between Ti3C2Tx and EB.

Full size image

Results and discussion

Fabrication and characterization of MEB

Ti3C2Tx was obtained by selectively etching the Ti3AlC2 MAX phase (see details in Methods). EB was achieved by extracting and delaminating sodium bentonite powder (see details in Methods). The successful synthesis of those two materials was confirmed by X-ray diffraction (XRD) patterns (Fig. S1), scanning electron microscopy (SEM), and transmission electron microscopy (TEM) images (Fig. S2), which show typical nanosheet morphology with lateral size of ∼8 μm for Ti3C2Tx and 2–5 μm for EB, respectively1,26,27. As shown in Fig. S3a, the mixed aqueous dispersions of Ti3C2Tx/EB are aggregated when the mass percentage of EB is more than 50 wt%. With the Zeta potential of −42 mV (Fig. S3b) close to sole Ti3C2Tx or EB, the homogeneously mixed dispersion (Fig. S3c) with an EB mass ratio of 50 wt% is chosen to fabricate Ti3C2Tx/EB composite film for the following study and be denoted as MEB. MEB maintains a similar lamellar structure and flexibility compared to the restacked Ti3C2Tx and EB films (Fig. S4), which is further confirmed by the images of scanning transmission electron microscopy (STEM). As shown in Fig. 2a, layered and homogeneous distributions of Si and Ti elements indicate the stacking of EB and Ti3C2Tx.

Fig. 2: Characterization of pristine Ti3C2Tx and MEB under high-temperature annealing with the presence of oxygen.
figure 2

a Cross-sectional STEM images of MEB with EDX mapping. b Optical image of pristine Ti3C2Tx film and MEB film before and after annealing at 400 °C for 2 h in synthetic air. c The tensile stress-strain curves of MEB-RT and MEB-Air-400C-2. Cross-sectional SEM images of the annealed Ti3C2Tx films (d) and MEB films (e). f Raman spectra of Ti3C2Tx films before and after treatment under synthetic air at 400 °C for 2, 4, and 6 h. g Raman spectra of MEB films before and after treatment under synthetic air at 400 °C for 2, 4, and 6 h. Source data are provided as a Source Data file.

Full size image

As mentioned previously, O2 contributes to the oxidation of Ti3C2Tx in the air, which would be accelerated at elevated temperatures. Therefore, we take the annealing time and temperature as variables to investigate the antioxidation of MEB under synthetic air (with volume fractions of 21% O2 and 79% N2). The samples treated at different conditions are denoted as sample-air-T-t, where the sample is Ti3C2Tx, EB, or MEB, T is annealing temperature, and t is annealing time. The pristine Ti3C2Tx and MEB at room temperature are the control samples denoted as Ti3C2Tx-RT and MEB-RT.

As shown in Fig. 2b, MEB-Air-400C-2 reserves the flexibility of fresh MEB and keeps its integrity after ultrasonication for 1 min in DI water (Fig. S5b). By contrast, Ti3C2Tx-Air-400C-2 tends to be fragile (Fig. 2b and Fig. S5a). According to the tensile stress-strain test (Fig. 2c and Fig. S6), MEB-Air-400C-2 exhibits a 10.6% decrease in tensile stress and a 1.0% decrease in strain. By contrast, Ti3C2Tx-Air-400C-2 exhibits a 57.7% decrease in tensile stress and a 5.5% decrease in strain, which is caused by the degradation of Ti3C2Tx in high-temperature treatment. The cross-sectional SEM images (Fig. 2d, e) show good retention of layered structure for MEB-Air-400C-2 and partial transformation from nanosheets to particles in Ti3C2Tx-Air-400C-2, demonstrating that the presence of EB nanosheets can delay the oxidation of Ti3C2Tx. Further measurements are utilized to investigate the high-temperature stability of MEB under the atmosphere of synthetic air. Raman spectra (Fig. 2f, g) display that typical Eg (Ti, C) mode at 289, 369 cm−1 and A1g (Ti, C) mode at 201, 599, and 722 cm−1 of Ti3C2Tx28,29 remain high strength in MEB-Air-400C-2 and A1g peak (201 cm−1) is still obvious in MEB-Air-400C-6. In sharp contrast, no typical peaks of Ti3C2Tx remain, while the peaks of TiO2 at 145, 400, 519, and 637 cm−1 appear in Ti3C2Tx-Air-400C-2, showing the transformation of Ti-C bonding to Ti-O bonding to the great extent30,31. Interestingly, the morphology of MEB remains the lamellar structure even after treatment at 400 °C for 6 h in the air, as shown in cross-sectional SEM images (Fig. S7c), while pristine Ti3C2Tx flakes transform into thicker chunk under the same treatment condition (Fig. S8c). The crystalline structure analyses of annealed MEB (Fig. S9b) and Ti3C2Tx (Fig. S9a) further confirm the suppression of oxidation for MEB at high temperatures in the air, even for a long treatment time.

Mechanism of high-temperature resistant property in MEB

The mechanism of high-temperature resistant property in MEB is investigated experimentally and theoretically. The thermogravimetry-mass spectrum (TG − MS) was used to reveal detailed insight into the thermal stability of Ti3C2Tx, EB, and MEB in synthetic air or argon (Ar). As shown in Fig. 3a, Ti3C2Tx exhibits four stages of weight change at 100, 360, 480, and 600 °C, respectively. At 100 °C, the removal of intercalated water leads to the appearance of the first H2O peak and weight loss. The second weak H2O peak is observed at 360 °C due to the slow dissociation of hydroxyl surface terminations32. The weight loss at 100–360 °C is in agreement with the TG − MS of Ti3C2Tx in Ar (Fig. S10a), which indicates that the degradation of Ti3C2Tx in synthetic air starts at～360 °C. This is also the reason why we set the temperature variable above 400 °C in this work. Around 480 °C, the weight of Ti3C2Tx quickly increases due to the interaction with O2, which evolves into TiO2 accompanied by the generation of CO233. The oxidation process accelerates with the increase of annealing temperature and approaches completion at 600 °C. For EB, as shown in Fig. 3b and S10b, only H2O is detected from the mass spectrum in the heating process under air and Ar atmospheres, suggesting the high thermal stability of EB with further confirmed by XRD and SEM analyses in Figs. S11–S13. The degradation of EB starts at 700 °C because of the dissociation of -OH and =O terminations26. Notably, EB exhibits a weight loss of 3.7% at 700–1040 °C in air, larger than that in Ar (2.5%). We assume that EB may interplay with O2, thereby inducing the change in weight loss during the O2 adsorption and desorption. This is evident that the rate of weight loss of EB in synthetic air changes at around 200 °C (Fig. 3b), while a negligible change in the rate of weight loss can be observed from the TG curve in Ar (Fig. S10b). Interestingly, as shown in Fig. 3c, MEB exhibits much better resistance to oxidation than pure Ti3C2Tx, as demonstrated by (i) the weight of MEB starts to increase at 650 °C with the appearance of CO2, which is much higher than pure Ti3C2Tx of 480 °C; (ii) Ti3C2Tx in MEB transforms into TiO2 completely at 920 °C with the strongest signal of CO2, while 600 °C is enough for Ti3C2Tx. Since EB is thermally stable in synthetic air during annealing, and the interaction between the surface of Ti3C2Tx and O2 plays a key role in the oxidation process, we reasonably speculate that the induction of EB may affect the interaction between adsorbed O2 and Ti3C2Tx by the Ti3C2Tx/EB interface coupling.

Fig. 3: Mechanism of high-temperature resistance in MEB.
figure 3

Thermal gravimetric (TG) curves in the air with mass spectrometry analysis (MS) for the atomic mass unit (amu) of 18/H2O and 44/CO2 for (a) Ti3C2Tx, (b) EB, and c MEB. d The charge density difference plots for the stable configurations of one O2 adsorbed on Ti3C2O2, EB, and Ti3C2O2/EB heterostructure. In the heterostructure, EB is subject to saturated oxygen adsorption. The isosurface level is set to be 0.0002 e/Å3 except for O2 adsorbed on EB with a value of 0.0006 e/Å3. The yellow area indicates charge accumulation, and the green region represents charge depletion. e X-ray photoelectron spectroscopy (XPS) of O 1 s of Ti3C2Tx-RT and MEB-RT. Source data are provided as a Source Data file.

Full size image

To further reveal the mechanism of the high-temperature resistant property of MEB, theoretical calculations were performed based on density functional theory (DFT). The adsorption energy (Ead) and Bader charge states of O2 molecule adsorbed on Ti3C2O2 and EB substrates are listed in Table 1. Ti3C2O2 was chosen as the prototype model due to the O-termination is the dominant way in experimentally synthesized Ti3C2Tx MXene (as shown in Fig. 4d). The more negative Ead for the oxygen adsorbed on EB indicates the stronger coupling when compared with that on Ti3C2O2. Meanwhile, the Bader charge analysis34 was performed in order to quantitatively assess the amount of charge transfer within the adsorbed systems, as shown in Fig. 3d and S14a, b. The adsorbed O2 obtains 0.421 electrons on EB, much larger than that on Ti3C2O2 (0.068 electrons), further confirming the stronger binding ability between O2 and EB.

Table 1 The calculated adsorption energy Ead and Bader charge state that transferred from substrates to adsorbed O2 molecules
Full size table
Fig. 4: Different termination ratios of MILD-Ti3C2Tx and HF-Ti3C2Tx, and high-temperature resistant property of MILD MEB with the presence of oxygen.
figure 4

a–c XPS of O 1s, F 1s, and Cl 2p signals, respectively, for the pristine HF-Ti3C2Tx and MILD-Ti3C2Tx films. d The termination ratio of HF-Ti3C2Tx and MILD-Ti3C2Tx was measured by XPS with a table list (inserted). e Raman spectra of MILD Ti3C2Tx before and after treatment under synthetic air at 400 °C for 2, 4, and 6 h. f Raman spectra of MILD MEB before and after treatment under synthetic air at 400 °C for 2, 4, and 6 h. Source data are provided as a Source Data file.

Full size image

Besides, we calculated the Ead of oxygen adsorbed on Ti3C2O2 that is interfaced with EB (denoted as Ti3C2O2/EB heterojunction). Considering that EB would firstly couple with oxygen, four O2 molecules were firstly placed on the surface of EB to mimic the saturated adsorption of O2 and then EB was interfaced with Ti3C2O2 in the heterojunction model (Fig. S14c). The Ead of one O2 molecule on Ti3C2O2/EB heterojunction (Fig. 3d) is −0.126 eV, which is less negative than that of pure Ti3C2O2 (−0.254 eV) (Table 1). This indicates a significant decrease in the interaction of Ti3C2O2 with oxygen in Ti3C2O2/EB heterojunction. The Bader charges of adsorbed O2 on heterostructure is 0.056 electrons, smaller than that on pure Ti3C2O2 (0.068 electrons), in line with the charge density difference plot (Fig. 3d). In addition, the charge density difference plot shows that the adsorbed O2 on freestanding Ti3C2O2 mainly obtains electrons from the nearest-neighbor Ti-O sub-layer of Ti3C2O2. Upon Ti3C2O2 is interfaced with EB (O2 molecules saturatedly adsorbed), the existence of the interface produces a smooth channel for charge transfer through the whole heterostructure along the c direction (Fig. S14c). This leads to a charge accumulation around the saturated O2 layer (four O2 molecules) and a deficient domain of electron states in the upper Ti3C2O2 layer. When one more O2 is adsorbed on the outermost Ti3C2O2 layer of Ti3C2O2/EB heterostructure, the amount of electron loss in the intermediate region is greatly reduced, thus weakening the coupling of O2 with the Ti3C2O2 layer. This is associated with the weakened hybridization of O-p orbitals of adsorbed O2 molecules and Ti-d orbitals. All of these indicate that the formation of the Ti3C2O2/EB interface greatly inhibits the further adsorption of O2 and thus enhances environmental stability.

The F-terminated MXene is also calculated as F termination accounts for a high proportion of Tx (Fig. 4d). The calculated Ead and Bader charge states of Ti3C2F2 before and after adsorption of O2 show that Ti3C2F2 exhibits more inferior interaction with O2 compared to EB (Fig. S15), suggestive of similar behavior to that of Ti3C2O2.

To further confirm the proposed mechanism, a 2D h-BN nanosheet with high thermal stability35 (Fig. S16) is chosen to composite with Ti3C2Tx. Theoretically, h-BN shows inferior adsorption of O2 to Ti3C2O2, which is proved by higher negative Ead values (−0.143 eV) and less amount of Bader charge transferred (0.055 electrons) of h-BN (Fig. S17 and Table 1). The Ti3C2Tx/h-BN composite was prepared by mixing two materials with a mass ratio of 1/1 (denoted as MBN). After annealing at 400 °C for 2 h in synthetic air, the layered structure of MBN transforms into a nanoparticle@nanosheet structure, as shown in Fig. S18. The peaks of TiO2 at 145 cm−1 appear with high intensity in the Raman spectrum (Fig. S19), indicating the oxidation of MBN. Since both of h-BN and EB exhibit good heat retardant property, the different thermal stability performance of their composites with Ti3C2Tx excludes the possibility that the high-temperature resistant property of MEB comes from the heat retardant property of EB. Actually, as MEB was treated at a high temperature for more than 2 h, the heat distribution around/in MEB should be highly uniform. In this context, we believe the high-temperature resistant property of MEB in the air should be attributed to the superior O2 adsorption on EB and the coupling between EB and Ti3C2Tx that largely weakens the further adsorption of O2 on Ti3C2Tx.

In addition, we found another phenomenon that may be beneficial for a high-temperature resistant property of MEB. As shown in X-ray photoelectron spectroscopy (XPS, Fig. 3e and S20), compared to pure Ti3C2Tx (44% Ti-O, 35% Ti-OH), the ratio of Ti-O species reduced to 21% while Ti-OH species increased to 55% in MEB. It reveals that EB with abundant hydroxy (confirmed by Fourier transform infrared spectroscopy in Fig. S21) interacts with Ti-O terminations of Ti3C2Tx to form hydrogen-bond, inducing compact layered structure as confirmed by SEM (Fig. S4b) and XRD analyses with the shift of (002) peak from 2θ = 5.8° (Ti3C2Tx) to 6.0° (MEB) (Fig. S9a, b). This is further evidenced by the higher permeability of O2/H2O of Ti3C2Tx than MEB, as shown in Table S1, which reveals the better barrier property of MEB possibly benefitted by the compact layer structure. As a result, the MEB composite could suppress the diffusion of O2 to some extent, lowering the amount of O2/H2O around Ti3C2Tx.

As we mentioned, H2O is another factor for oxidizing Ti3C2Tx at elevated temperatures. Here, we also study the high-temperature resistant property of MEB in the H2O atmosphere (argon with 90% relative humidity, denoted as RH 90%). The Raman spectra in Fig. S22 show that the intensities of Ti3C2Tx peaks in Ti3C2Tx-RH 90%-400C-2 become weak. Further prolonging of annealing time results in more severe oxidation which can be confirmed by the appearance of TiO2 Eg peak at 145 cm−1 in pure Ti3C2Tx film after annealing for 4 h. However, negligible change of Ti3C2Tx fingerprints is detected in MEB-RH 90%-400C-4. Along with the results of SEM (Figs. S23, S24) and XRD (Fig. S9c, d), it verifies the good thermal stability of MEB in the H2O atmosphere. In addition, the antioxidant capability of MEB is detected at higher temperatures (500 and 600 °C) under RH 60% synthetic air (simulated atmospheric environment). The layered structure of MEB retains after annealing for 2 h at 600 °C; on the contrary, the layered Ti3C2Tx transforms into amorphous TiO2 completely, which is in agreement with the XRD analysis (Fig. S25). All the results confirm that MEB demonstrates high-temperature resistant property even under long-time treatment with the presence of both oxygen and water molecules.

The effect of termination ratio on the high-temperature resistant property of MEB

To detect the influence of terminations on the high-temperature resistant property of Ti3C2Tx, two approaches are adopted to synthesize Ti3C2Tx, which are minimally intensive layer delamination method (denoted to MILD-Ti3C2Tx, Fig. S26) and HF etching method (HF-Ti3C2Tx), respectively. By fitting the relative intensities of the O 1s, F 1s, Cl 2p peaks36,37 as shown in Fig. 4a–d and S27, we determine the termination of HF-Ti3C2Tx to be 38.7% = O, 31.6% -F, 7.7% -Cl, 22.0% -OH and MILD-Ti3C2Tx with 43.8% = O, 24.4% -F, 13.1% -Cl, 18.7% -OH, which reveals the different termination ratios for two samples. For clear description, a composite consisting of MILD-Ti3C2Tx and EB is denoted as MILD MEB in order to differentiate with MEB (HF-Ti3C2Tx/EB). Following the same thermal treatment of MEB, MILD MEB-Air-400C-2 shows high-intensive Eg (Ti, C) peaks at 289, 369 cm−1 and A1g (Ti, C) peaks at 201, 599, 722 cm−1 of Ti3C2Tx, and A1g peak (201 cm−1) is still obvious in MILD MEB-Air-400C-6 as shown in Fig. 4f. By contrast, MILD Ti3C2Tx-Air-400C-2 (Fig. 4e) shows the transformation of Ti-C bonding to Ti-O bonding, which is indicated by the appearance of TiO2 peaks at 145, 400, 519, 637 cm−1. Along with the SEM analysis (Fig. S28), the results confirm the high-temperature resistant property for MILD MEB in air. The similar behavior in MEB and MILD MEB reveals our strategy for suppressing the oxidation of Ti3C2Tx is independent of the termination ratio of Ti3C2Tx.

The potential applications of MEB

High-temperature resistant electromagnetic interference (EMI) shielding is of importance for the aerospace industry, for scenarios including but not limited to the EMI shielding of engine casing of aircraft (higher than 380 °C), etc. Here we tested the EMI shielding efficiency of MEB in the frequency range of 0.2–1.3 THz after annealing at different temperatures for a long time in RH 60% synthetic air (simulated atmospheric environment in a practical scenario), which are denoted as MEB-Atmos-T-t (T is temperature, t is time) for different annealing conditions. Pristine Ti3C2Tx is taken for comparison. In order to truly reflect the performance comparison, MEB and Ti3C2Tx are controlled to obtain the same amount of 11 mg Ti3C2Tx, as the EMI shielding originates from the high electrical conductivity of Ti3C2Tx and EB is noneffective for THz shielding (Fig. S29), which is further explained by the proposed EMI shielding mechanism of MEB in Fig. S30. As shown in Fig. S31, the transmitted THz signals of Ti3C2Tx-RT and MEB-RT are too weak to be detected due to the excellent THz shielding performance of Ti3C2Tx, and the average THz shielding efficiencies (THz SEs) are obtained around 47 dB as shown in Fig. 5a–c. After annealing, MEB-Atmos-400C-6 have a good shielding capacity retention, even higher than pristine material (∼50 dB). This can be attributed to the increase of the electrical conductivity from 850 to 1000 S cm−1, which is caused by more compact nanosheet stacking of MEB-Atmos-400C-6 with the removal of intercalated water. The XRD analysis validates the reduction of the interlayer (d-) spacing, characterized by the peak of (002) shifting from 2θ = 6.0° to ~8.2° after annealing (Fig. S9b). By contrast, Ti3C2Tx-Atmos-400C-6 performs high transmittance and ultralow THz SE (approach to zero) owing to its complete degradation. Besides, THz transmittance and shielding efficiency of MEB after annealing at a higher temperature for 2 h are investigated, as shown in Fig. 5b, c. With the annealing temperature increasing, the transmissions of THz waves through MEB-Atmos-500C-2 and MEB-Atmos-600C-2 remain at 0.001 and 0.0015%. In other words, the THz SE of the two samples can reach about 50 dB at 500 °C and 48 dB at 600 °C (Fig. 5b, c) because of good electric conductivity (1090 S cm−1 for MEB-Atmos-500C-2h and 960 S cm−1 for MEB-Atmos-600C-2h), which is consistent with the XRD analysis (Fig. S25d). On the contrary, THz SE of Ti3C2Tx decreases to～0 dB at 600 °C because of the oxidation of Ti3C2Tx. All the results suggest that the introduction of EB can suppress the oxidation-induced deterioration of Ti3C2Tx, therefore making MEB promising high-temperature resistant THz shielding materials. In addition, a comprehensive literature review is summarized to compare the performance of MEB with published THz shielding materials. So far, most THz shielding materials are utilized at room temperature and few researchers focus on their applications at high temperature or hash environment, as shown in Fig. S32. Therefore, MEB with high-temperature resistant property in air and humid environment may meet the demands of working in a harsh environment (further high-temperature THz test is necessary to finally evaluate the feasibility of MEB for the practical scenario).

Fig. 5: Potential applications of MEB.
figure 5

a–c THz EMI shielding property of MEB. a EMI SE in 0.2–1.3 THz of Ti3C2Tx-RT, MEB-RT, Ti3C2Tx-Atmos-400C-6, MEB-Atmos-400C-6. b EMI SE in 0.2–1.3 THz of Ti3C2Tx-Atmos-500C-2, 600C-2, and MEB-Atmos-500C-2, 600C-2. c The average THz SE of the samples showed in Fig. 5a (left) and Fig. 5b (right). d Long-term Joule heating performance of MEB and Ti3C2Tx films driven by an applied voltage of 3.0 V. e The thermal cycling performance of MEB film, which is realized by switching the applied voltage (3.0 V) on and off repeatedly. Source data are provided as a Source Data file.

Full size image

To further explore the potential application of MEB by virtue of the high-temperature resistant property, we also measure the Joule heating performance of MEB. Joule heating device is required to provide fast heating and high-temperature output (in some cases, hundreds of degrees are needed). In general, the metallic conductivity and high thermal conductivity of Ti3C2Tx meet the demand for the Joule heating device. The Ti3C2Tx and MEB freestanding films were used to make the devices of the Joule heater (Fig. S33). As shown in Fig. 5d, pristine Ti3C2Tx film quickly reaches 190 °C under a driving voltage of 3.0 V. However, it is stable for only 0.57 h, then the temperature begins to decline due to the oxidation of Ti3C2Tx. As a comparison, the heating rate of MEB film can be 20 °C s−1, with a steady-state temperature of 198 °C without large temperature fluctuation for more than 3.5 h, which reflects the rapid thermal response and stable high-temperature electrothermal performance. In addition, the thermal cycling performance of MEB is tested as shown in Fig. 5e. The steady-state temperatures of MEB under driving voltage being on/off is basically identical during 30 cycles, which reveals the stable thermal cycling of MEB. The above results suggest that the capabilities of fast thermal response and thermal cycling stability make MEB potentially suitable for thermal/photothermal catalysis and beyond. In addition, a comparison of the steady temperature and heating rate for various Joule heating materials is listed in Table S3.

In conclusion, we present a Ti3C2Tx-based composite that is capable of suppressing oxidation at high temperatures (higher than 400 °C) in air. We demonstrate that oxygen is preferentially adsorbed on EB due to superior oxygen adsorption. The saturated adsorption of oxygen on EB further inhibits more oxygen molecules to be absorbed on the surface of Ti3C2Tx due to the weakened p-d orbital hybridization between adsorbed O2 and Ti3C2Tx that is induced by the Ti3C2Tx/EB interface coupling. In addition, our strategy for suppressing the oxidation of Ti3C2Tx is independent of the termination ratio of Ti3C2Tx. Utilized as THz shielding material, MEB shows a high THz EMI SE of 48 dB after annealing under 600 °C in an atmospheric environment, which validates MEB as a promising candidate for THz shielding material that works in a high-temperature scenario. Joule heating and thermal cycling properties are also tested to expand the potential applications. The strategy of designing high-temperature resistant MXene-based composite may be applicable to other two-dimensional materials and beyond.

Methods

Synthesis of pristine Ti3C2Tx and EB

we prepared Ti3C2Tx by two approaches in this study. HF-Ti3C2Tx (denoted as Ti3C2Tx in the paper for readability): Ti3C2Tx was synthesized by selectively etching Al of Ti3AlC2 MAX (characterized by SEM and XRD38, Fig. S34, S35) using HF and HCl10,39. About 1 g of Ti3AlC2 powder (11 Technology Co.) was added to the etchant solution, which includes 12 mL 38 wt% hydrochloric acid (HCl, Sigma-Aldrich), 2 mL 49 wt% hydrofluoric acid (HF, Sigma-Aldrich), and 6 mL deionized (DI) water. Next, the Ti3AlC2 powder was etched for 24 h by string at 400 rpm at 35 °C. Then, the multilayer Ti3C2Tx was obtained by washing the reaction mixture with DI water via centrifugation at 684.78 × g (5 min per cycle). After washing 5–6 times (until the pH ≥ 6), the sediment was dispersed into the delamination solution, which was prepared by adding 1 g of LiCl (Sigma-Aldrich) into 40 mL DI water. Then, the mixture was stirred for 4 h at room temperature. The dispersion was washed using DI water and the single layer Ti3C2Tx aqueous solution was obtained by centrifugation at 684.78 × g for 5 min. MILD-Ti3C2Tx (prepared by minimally intensive layer delamination method): LiF (1.6 g; Sigma-Aldrich) was added to 9 M HCl (20 mL; HCl, 38 wt%, Sigma-Aldrich) and continuously stirred at 400 rpm for 30 min at 40 °C. About 1 g Ti3AlC2 (11 Technology Co.) was added to the mixture solution and stirred at 400 rpm for 24 h at 40 °C. Then the dispersion was washed with DI water 4–6 times until the pH ≥ 6. After washing, the sediment was dispersed in 50 mL of fresh DI water and the single layer Ti3C2Tx aqueous solution was obtained by centrifugation at 684.78×g for 5 min40.

To prepare the EB aqueous solution, sodium bentonite powder (San Ding Technology Co.) was extracted and delaminated. The pristine powder was dissolved in DI water. Then, the homogeneous EB supernatant was collected after ultrasonic treatment for 3 h and centrifugation at 684.78 × g for 5 min. The concentrations of Ti3C2Tx and EB dispersions were measured by filtering a specific amount of colloidal solution through a polypropylene filter (3501 Coated PP, Celgard LLC, Charlotte, NC), followed by drying under vacuum at 60 °C overnight and measuring the weight of the solid residue.

Preparation of freestanding Ti3C2Tx film, EB film, and MEB

About 11 mg of freestanding pristine Ti3C2Tx film (with a thickness of ∼5 μm) and 22 mg of freestanding pristine EB film (with a thickness of 11.2 μm) were fabricated by vacuum filtration of aqueous solutions, then the films were dried under vacuum at 60 °C overnight. The MEB solutions (with 11 mg Ti3C2Tx and 11 mg EB) were prepared by adding EB aqueous solution into the Ti3C2Tx solution and continued stirring for 2 h. The obtained homogeneous solution was filtrated by vacuum-assisted filtration and dried under vacuum at 60 °C for 24 h to obtain MEB (with a thickness of 11 μm).

Meanwhile, MILD Ti3C2Tx film and MILD MEB are fabricated by MILD Ti3C2Tx solution in a similar process.

Preparation of freestanding MBN

h-BN nanosheets were prepared by delaminating h-BN powder (Aladdin) through ultrasonication in DI water for above 8 hours35. Then, the dispersion was centrifugated at 684.78×g for 5 min to collect the homogeneous layered h-BN supernatant and the concentration was measured by the same method of Ti3C2Tx supernatant. MBN film was prepared by vacuum-assisted filtration of Ti3C2Tx/BN mixture, in which the weight percentage of BN was 0.5.

Annealing treatment of materials

The sample was put into a quartz boat, then annealed in a tube furnace (OTF-1200X, Kojing Material Technology Co.) under a specific atmosphere (synthetic air, RH 90% argon, or simulated atmospheric environment) with a programmed heating rate of 5 °C min−1. During the annealing process, the bottom surface of the sample may interact with less oxygen/water molecules than the upper surface due to the contact between the bottom surface of the sample and the quartz boat. This may result in less oxidation of the bottom surface than the top layer.

Material characterization

Scanning electron microscopy (SEM) images were acquired at 15 kV by using a JSM-7600F SEM. X-ray diffraction (XRD) was performed on a Bruke D8 advanced powder X-ray diffractometer equipped with Copper Kα radiation (λ = 1.540598 Å). Transmission electron microscope (TEM) images were obtained from Tecnai G2 F20 S-TWIN, FEI (200 kV). Raman spectra were acquired using HORIBA’s Raman system with a 532 nm excitation laser at 5% laser power (Jobin Yvon S.A.S, France). XPS was conducted using PHI VersaProbe 5000 instrument (Physical Electronics) with a 100 µm and 25 W monochromatic Al-Kα (1486.6 eV) X-ray source. Thermal gravimetric with mass spectrometry analysis (TG-MS) curves were obtained under a synthetic air/Ar atmosphere using STA449F3 and QMS 403 (Netzsch, Germany) with a mass spectrometer (110/220 V) and a temperature scan rate of 5 °C min−1 from room temperature to 1100 °C. Electrical conductivities were tested by an ST 2253 four-probe test instrument (Suzhou Jingge Electronic Co., Ltd.). The samples for cross-sectional scanning transmission electron microscopy (STEM) were prepared by an FEI HELIOS NanoLab 600i Focused Ion Beam (FIB) system and the local element distribution (mapping) was analyzed by highly efficient energy dispersive X-ray (EDX) spectroscopy at 200 kV by FEI Talos F200X. Mechanical tests of films were performed on a dynamic mechanical analyzer WDW-10D and DS2-50N-XB (Baichuan Instruments Co.). The Ti3C2Tx film samples were cut into 10 mm × 30 mm strips using a steel razor blade. Uniaxial tensile tests were performed at a strain rate of 5 mm/min. The sample thickness was estimated by the helical micrometer. The O2 permeability was measured using a permeability tester (OX-TRAN 2/12 R, MOCON Co., USA). The water vapor transmission rate was measured using a moisture permeability testing apparatus (C360, Jinan Languang Electromechanical Technology Co.).

Computational details

The theoretical calculations were performed by density functional theory (DFT)41, as implemented in PWmat code42,43. The norm-conserving pseudopotentials (NCPP-SG15)44 with Perdew–Burke–Ernzerhof (PBE) exchange-correlation functional were adopted45. The slab model of EB was built based on its (001) surface. A vacuum layer of 20 Å was utilized in the z direction to eliminate the coupling between the substrate sheet and its periodic replica. 2 × 2, 3 × 3, 3 × 3, and 4 × 4 supercells of (001)-plane EB, Ti3C2O2, Ti3C2F2 MXenes, and BN monolayer were utilized, respectively, so as to guarantee the close surface area in xy plane. Moreover, a \(2\sqrt{3}\times 3\) supercell of (001)-plane MXene was interfaced with a 2 × 2 supercell of (001)-plane EB (denoted as Ti3C2O2/EB) to model the inhibiting effect of O2 adsorbed on MXene34. The lattice mismatch of the heterojunction is only about 1.5%. The plane-wave basis set cutoff energy was set to 60 Ry, and all atoms within systems were fully relaxed until the force on each atom was less than 0.02 eV/Å. The PBE-D3 method of Grimme was adopted to correct the long-range van der Waals interaction in all the calculations46.

The adsorption energy \({E}_{{{{{{{\mathrm{ad}}}}}}}}\) were quantitatively determined by the following equation:

$${E}_{{{{{{{\mathrm{ad}}}}}}}}={E}_{({{{{{{\mathrm{Slab}}}}}}} \!+\! {{{{{\mathrm{M}}}}}})}-{E}_{{{{{{{\mathrm{Slab}}}}}}}}-{E}_{{{{{{\mathrm{M}}}}}}}$$
(1)

Where E(Slab+M) is the total energy of the M molecule adsorbed on the slab. ESlab and EM correspond to the energy of the slab model and M molecule, respectively.

Terahertz shielding measurements

Terahertz time-domain spectroscopy (THz-TDS) was measured by titanium sapphire with a center wavelength of 800 nm and a pulse width <50 fs as the excitation source9,10. The measurements were performed at room temperature. The humidity in the test room was controlled in a low range (<5%). The effective spectrum range is 0.2−1.3 THz, and the repetition frequency is 1 kHz. The testing time step was about 0.067 ps throughout the experiment. The EMI SE values were calculated using the following equation:

$${{{{{{\mathrm{EMI}}}}}}}\,{{{{{{\mathrm{SE}}}}}}}=10\,{{{{{\rm{log }}}}}}(1/{{{{{\rm{T}}}}}})=-20{{{{{\rm{log }}}}}}({E}_{t}/{E}_{a})$$
(2)

Where T is the transmittance of terahertz wave; \({E}_{t}\) and \({E}_{a}\) refer to the amplitudes of transmission terahertz pulses for the samples and the air cavity, respectively.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.