Background

After the intensive studies of graphene [1] and transition metal dichalcogenides(TMDs) [2,3,4,5] for a decade, two-dimensional (2D) metal carbides (MXenes) have been drawing much attention recently owing to their extraordinary electrical properties [6, 7]. The Ti3C2 MXene is a layered material with Van der Waals stacked structure, in which each layer contains two carbon atom planes sandwiched among three Ti atom planes. In the energy band structure of Ti3C2 MXene, the conduction band touches the valence band at the Γ point, which indicates that Ti3C2 MXene is a half-metallic material [8]. More interestingly, the band structure of Ti3C2 Mxene can be opened up slightly by the surface functional groups (which are annotated as “T” in the formula) [8], such as –F, –O, and –OH [9]. These functional groups are generated in the solution etching of MnAlCn + 1 phase [9], which then forms a Ti3C2Tx MXene with tunable electrical properties. The electrical conductivity of Ti3C2Tx MXene was measured to be 4.2 × 10−4S/m [10], which is superior to most TMDs. To date, Ti3C2Tx MXenes have shown potential applications in advanced supercapacitors [11], Li-batteries [12], electromagnetic shielding [10], antibacterial [13], and light emission [14].

In addition to high electrical conductivity, the elastic property of Ti3C2Tx MXene also attracts much attention. Theoretical calculations predicted that this ultrathin carbide has a Young’s modulus of as high as ~ 500 GPa [15,16,17]. Lipatov et al. recently used nanoindentation to determine that the Young’s modulus of monolayer Ti3C2Tx MXene was 330 Gpa [18]. These values are much higher than those of MoS2 [18] and are comparable to those of monolayer graphene [19]. Recently, Ghidiu et al. measured the high-pressure X-ray diffraction (XRD) spectra of Ti3C2Tx MXene up to 3 GPa and observed no phase transformation [20]. However, as the pressure loaded in ref. [20] was too low, the phase stability and lattice deformation of Ti3C2Tx at higher pressure are still unknown.

Raman spectroscopy acts as a useful non-destructive tool to investigate the crystal structure and phonon vibration of 2D materials such as graphene [21] and TMDs [2]. The composition of Ti2CTx [22] and phase stability of Ti3C2Tx Mxene at different annealing conditions [10] can be probed by using confocal Raman measurements. Recently, the phonon dispersion of Ti3C2Tx MXene was theoretically calculated by Hu et al. [23, 24], thus enabling a further understanding of the Raman spectra of this material. However, the high-pressure Raman spectroscopy of Ti3C2Tx is still lacking. Moreover, the phonon response of Ti3C2Tx as a function of pressure is unknown.

In this paper, we prepared Ti3C2Tx thin flakes and measured their pressure-dependent XRD and Raman spectra up to 26.7 GPa. The elastic constants of Ti3C2Tx were calculated from XRD diffraction peak shifts by the Murnaghan equation. The positive Grüneisen parameters of out-of-plane phonons were obtained from their pressure-dependent Raman shift and lattice parameter deformation ratio. The obtained results would benefit the further understanding of the mechanical and phonon-vibrational behavior of Ti3C2Tx MXene.

Results and Discussions

Before conducting high-pressure measurements, we first investigated the basic material properties of the exfoliated Ti3C2Tx Mxene flakes. An optical image of the exfoliated Ti3C2Tx flakes deposited on Si/SiO2 (300 nm) substrate is shown in Fig. 1a. Light green contrast can be observed for the exfoliated flakes. As reported by Miranda et al., the optical contrast of Ti3C2Tx flakes depends on the flake thickness strongly, where thicker flakes always show higher contrast, while thin flakes exhibit low contrast [25]. The light green contrast of most of the flakes in Fig. 1b indicates their thin thickness. An atomic force microscope (AFM) topographic image of exfoliated Ti3C2Tx flakes is shown in Fig. 1b. The flakes in the mapping area show surfaces with high roughness, which is typical for Ti3C2Tx flakes [26]. The thickness of a typical thin flake can be determined from its line profile (inset of Fig. 1b) across the marked position in Fig. 1b to be 170 nm. A scanning electron microscope (SEM) image of an exfoliated flake is shown in Fig. 1c. The laminated structure of Ti3C2Tx can be seen clearly, indicating the successful preparation of Ti3C2Tx layered samples [10].

Fig. 1
figure 1

a Optical image of ultrasonically exfoliated Ti3C2Tx flakes; b AFM topographic image of ultrasonically exfoliated Ti3C2Tx flakes, and a line profile across the marked dashed line is shown as an inset, indicating the Ti3C2Tx flake thickness of 170 nm; c SEM image of ultrasonically exfoliated Ti3C2Tx flakes; d XRD spectra of Ti3C2Tx raw powder

Full size image

We further measured the XRD spectra of raw Ti3C2Tx powder, as shown in Fig. 1d. This XRD pattern is an analogy to previous reports [10]; thus, the prominent peaks at 8.95°, 18.28°, and 27.7° can be assigned to the diffractions of (002), (004), and (006) planes. Compared to the prominent peaks, the intensity of diffraction peaks of minor phases (anatase TiO2(101) at 25.3°, JCPDS Card No. 71-1116) is relatively weak, indicating the high purity of Ti3C2Tx phase in the obtained powder. The (002) peak emerges at an angle slightly lower than that reported by Han et al. (9.21°) [10]. The calculated c lattice parameter, 19.66 Å, is larger than the reported value (19.2 Å) [10]. It should be noted that since the interlayer space may be tuned by different densities of chemical groups and ions, such as –F, –OH, and Li+, the c lattice parameter varies substantially from 19.2 Å to 58.8 Å in different studies [10, 20, 26, 27]. The c lattice parameter for our sample is very close to the low value that was measured for Ti3C2Tx powder simply using HF as etchant [10].

The XRD spectra of Ti3C2Tx flakes measured at different pressures until 26.7 GPa are shown in Fig. 2a. It can be seen that the spectra measured at different pressures are similar to each other, while no new diffraction peak can be found. This finding indicates no phase transformation occurs with the pressure up to 26.7 GPa. In Fig. 2a, all the diffraction peaks shift to large angles with increasing pressure, indicating the shrinkage of the Ti3C2Tx lattice. Such pseudo-negative compressibility has also been observed for Ti3C2Tx [20] and other low dimensional materials with a layered structure, such as graphite [28], graphene oxide [29, 30], MoS2 [31], clay [32], and titanates [33]. The (002) peak shifts from 2.883° to 3.162° as the pressure increased from 1.8 GPa to 26.7 GPa. The deformation ratio of lattice parameter c, c/c0, as a function of pressure, can be calculated from the shift of (002) peak. Moreover, the deformation ratio of a, a/a0, can be calculated from the shift of (110) peak. As shown in Fig. 2b, the lattice parameters c and a are deformed by 9.1% and 2.4%, respectively, at a pressure of 26.7 GPa. In the low-pressure region at ~ 3 GPa, the compressive ratio of lattice parameter c is 3%. In the previous high-pressure XRD measurement of Ti3C2Tx flakes, a slightly larger c compressive ratio of 4% for dry Ti3C2Tx flakes was reported by Ghidiu et al. [11](Fig. 2b). This difference could be induced by a larger lattice parameter c(25.1 Å) of the sample used by Ghidiu et al. [11] with respect to that of ours(19.66 Å).

Fig. 2
figure 2

a XRD spectra of Ti3C2Tx at different loaded pressures. Note that the unit is GPa for the pressure annotated on each spectrum. Peaks are assigned according to ref. [26]; b experimental (dots) and calculated compressive ratio (solid line) along c and a directions. The solid lines are the fitted results using the equation \( r(P)/{r}_0={\left[\left({\delta}_0/{\delta}^{\prime}\right)P+1\right]}^{\delta^{\prime }} \)

Full size image

To obtain the elastic constants, the c and a compressive ratios in Fig. 2b are further fitted by using the Murnaghan equation [34].

$$ r(P)/{r}_0={\left[\left({\beta}^{\prime }/{\beta}_0\right)P+1\right]}^{-1/{\beta}^{\prime }} $$
(1)

where r represents the lattice constants along c and a axes, \( {\beta}_0^{-1}=-{\left(\frac{dlnr}{lnP}\right)}_{P=0} \) is the linear compressibility, and β′ is the pressure derivate of β.

The fitted r/r0 curves of a and c are plotted as solid lines in Fig. 2b. It can be seen that the experimental results fit the equation expectation very well. The best fitting generates β0 and β′ for c as 67.7 GPa and 25.5, respectively. While for lattice parameter a, β0 and β′ are calculated to be 387.4 GPa and 72.1, respectively (Table 1). For ultrathin 2D materials such as graphene, Young’s modulus (1TPa) is very close to the β0 of thick graphite [19, 28]. Therefore, β0 can be used as a substitution to evaluate the elastic constant of Ti3C2Tx. Young’s modulus of Ti3C2Tx was recently measured by Lipatov et al. to be 330 GPa [18], which is consistent with the β0 in our study. Our measured values are also comparable to the elastic constants of Ti3C2 that were calculated in other studies [15, 17] (Table 1). The β0 at c axis is larger than that of graphite (β0 = 35.7 GPa), while the β0 at a axis is smaller than that of graphite (β0 = 1250 GPa) [28]. The β0 of Ti3C2Tx is higher than the bulk modulus of MoS2 (270 Pa) [35] and is also comparable to that of graphene oxide (210 GPa) [36], indicating a high elastic constant of Ti3C2Tx Mxene among 2D materials.

Table 1 The first and second order compression coefficients of Ti3C2Tx, which are calculated by fitting Fig. 2b through Eq. 1
Full size table

High-pressure Raman spectra of Ti3C2Tx samples were measured at different compressive pressures up to 25.5 GPa, as shown in Fig. 3a. The Raman spectra obtained at different decompressive pressures are shown in Fig. 3b. At low compressive pressures, Ti3C2Tx Mxene exhibits three major Raman bands at ~ 210, ~ 500, and 700 cm−1. It should be noted that the Raman spectra of Ti3C2Tx MXene vary significantly in different kinds of literature. Hu et al. [23] reported strong Raman peaks at ~ 200 cm−1 and 720 cm−1, while other bands at 400 cm−1 were quite broad. However, Han et al. [10] and Zhu et al. [37] observed a sharp peak at ~ 200 cm−1, but other bands were all broad. Xue et al. [14] only observed broad peaks from 100 to 700 cm−1. The Raman spectra in Fig. 3 are different from those in ref. [10, 14, 23, 37]. This difference could be induced by different types and concentrations of chemical groups on Ti3C2Tx MXene. Further interpretation of these different Raman bands needs to recall the phonon dispersion of Ti3C2Tx that was theoretically calculated by Hu et al. [23, 24]. The space group of Ti3C2Tx was P63/mmc [23]. The number of atoms (N) in a primitive cell of Ti3C2Tx was calculated to be 7, 7, and 9 for T = −O, −F, and −OH, respectively, given x = 2. At Γ point of the first Brillouin zone, the following optical phonons are predicted to be existing for different Ti3C2Tx MXenes: Γoptical (Ti3C2O2) = 6Eg + 3A1g, Γoptical (Ti3C2F2, Ti3C2 (OH)2) = 8Eg + 4A1g [23]. The atom vibrations of different Raman active modes of Ti3C2F2 and Ti3C2(OH)2 are illustrated schematically in Table 2. Their frequencies were theoretically calculated by Hu et al. [23] and are listed in Table 2. For Ti3C2(OH)2, there are four out-of-plane modes (A1g: 218, 514, 684, and 3734 cm−1) and four in-plane modes (Eg: 138, 278, 437, and 622 cm−1). For Ti3C2F2, there are three A1g modes (190, 465, and 694 cm− 1) and three Eg modes (128, 231, and 612 cm− 1) [23].

Fig. 3
figure 3

a Raman spectra of Ti3C2Tx flakes at different compression pressures; b Raman spectra obtained at different decompression pressures. Note that the unit of pressures in a and b is GPa

Full size image
Table 2 Assignment of typical Raman peaks in this work and reference. The theoretically calculated phonon energies of Ti3C2(OH)2 and Ti3C2F2 are listed for comparison. The vibration directions of atoms in a unit cell are also shown schematically. The unit of Raman shift is cm−1
Full size table

However, it can be seen the calculated phonon frequencies of pure Ti3C2F2 or Ti3C2(OH)2 cannot fit the experimental Raman spectra of Ti3C2Tx in Fig. 3. Because the surface of Ti3C2Tx is usually attached by a different type of chemical groups, a full interpretation of the experimental Raman spectra in Fig. 3 needs to consider the hybridized vibration modes of −F and −OH [23]. In a previous research [23], the Raman bands at ~ 200, ~ 500, and ~ 700 cm−1 were assigned to ω2, ω6, and ω3, respectively. Following this instruction, the prominent Raman bands at 205.6, 490.2, and 702.5 cm−1 of Fig. 3a, b can be assigned to ω2, ω6, and ω3 modes, respectively. Interestingly, these modes are all out-of-plane modes. Other Raman modes are difficult to be isolated from their neighboring modes due to their low intensities. To obtain robust data by eliminating the uncertainties, only these three modes are considered in the following calculations and analyses.

In Fig. 3a, b, it can also be seen that the relative intensities of these in-plane modes increase with the increasing compressive pressures (Fig. 3a). When the compressive pressure is ≥ 12.6 GPa, a new peak at ~ 600 cm−1 (ω4) emerges and becomes the prominent peak. In the decompression process, the intensities of this ω4 mode decrease significantly. The Raman spectra obtained at 0 GPa of decompression pressure contain almost all the in-plane and out-of-plane phonon modes. Such emergence of in-plane modes at high compressive pressure might be related with flake fracture or orientation rotation-induced polarization. Research about this effect is still ongoing and will be reported in the future.

With increasing pressure from 0.8 GPa to 25.6 GPa, ω2, ω6, and ω3 all show monotonic increasing blueshifts (Fig. 4a–d), which are similar to the pressure-dependent blueshifts of graphite [28] and MoS2 [31]. At 25.6 GPa, these three modes’ blueshifts are 66.7, 85.1, and 60 cm−1, respectively. Such pressure-dependent blueshifts are much larger than those of MoS2 [31]. To quantify the Raman shift vs pressure, the Raman shift plots in Fig. 4a, b, d were fitted using the following equation [28]:

$$ \omega (P)/{\omega}_0={\left[\left({\delta}_0/{\delta}^{\prime}\right)P+1\right]}^{\delta^{\prime }} $$
(2)
Fig. 4
figure 4

Raman shifts of different phonon modes as a function of different compressive pressures (solid sphere) and decompressive pressures (opened circles): a 210 cm−1, b 504 cm−1, c 620 cm−1, and d 711 cm−1. Solid lines are the fitting results using the equation \( \omega (P)/{\omega}_0={\left[\left({\delta}_0/{\delta}^{\prime}\right)P+1\right]}^{\delta^{\prime }} \)

Full size image

where δ0 and δ′ are the logarithmic pressure derivative(dlnω/dP)P = 0 and the pressure derivative of dlnω/dP, respectively. The fitted results are plotted as solid lines in Fig. 4a, b, d. Due to the high uncertainties at the low-pressure region, the Raman mode at 620 cm−1 was not fitted. It can be seen in Fig. 4a, b, d that the fitted curves are consistent with the experimental results, indicating the high accuracy of the fitting process. The obtained parameters of δ0 and δ′ are listed in Table 3.

Table 3 Calculated pressure-dependent parameters of Ti3C2Tx. \( \overline{\upgamma} \) refers to the averaged Grüneisen parameters that were calculated using equation γ =  − dlnω/(3dlnr)
Full size table

For anisotropic 2D materials with an atomic thickness, such as MXene and graphene, two independent components of the Grüneisen tensors are usually associated with the strains that are parallel and perpendicular to the c axis. For simplicity, we adopted the scaling relationship proposed by Zallen et al. [38], which has been used to fit the pressure-dependent Raman shift of graphite by Hanfland et al. [28]

$$ \omega (P)/{\omega}_0={\left[r(P)/{r}_0\right]}^{3\gamma } $$
(3)

where r refers to the in-plane and out-of-plane lattice constants for intralayer and interlayer modes, respectively. γ is equivalent to the Grüneisen parameter that was defined in other studies [39, 40].

As only the out-of-plane modes are observed in the compression process, c lattice parameter as a function of hydrostatic pressure is adequate for calculation. We adopted the (002) plane’s space distance data of 0 to 26.7 GPa in Fig. 2b for the calculation of Grüneisen parameter γ. The averaged \( \overline{\upgamma} \) up to 26.7 GPa for ω2, ω6, and ω3 were calculated to be 1.08, 1.16, and 0.29, respectively (Table 2). Similar to the graphite, the smaller \( \overline{\upgamma} \) of ω3 compared with the other two modes indicates a smaller change in force constants is involved in the rigid-layer motion [28]. To the best of our knowledge, the Grüneisen parameters of Ti3C2Tx have not been reported yet. However, we can still compare our data with other 2D materials. Zha et al. reported Grüneisen parameters as 4–5 for acoustic phonon modes of Ti2CO2 [41]. Because the Grüneisen parameters of high-frequency optical modes are usually one or two orders lower than those of the low-frequency modes [28], the Grüneisen parameters for optical phonons of Ti2CO2 can be estimated to be 0.05–0.5, which are similar to those of our values for Ti3C2Tx. Recently, Peng et al. [42] reported the Grüneisen parameters at room temperature as 1.22, 1.20, and 1.15 for MoS2, MoSe2, and WS2, respectively, which are larger than our results. Our results are also smaller than those of graphene (1.99 for E2g mode) [40] and graphite (1.06 for E2g mode) [28]. This finding indicates that Ti3C2Tx MXene has the weakest bonding anharmonicity among these ultrathin 2D materials [42].

Conclusions

In conclusion, we measured lattice deformation and phonon response of Ti3C2Tx thin flakes at different hydrostatic pressures up to 26.7 GPa. No phase transformation has been observed below a pressure of 26.7 GPa. All the phonon modes show a positive frequency shift with increasing pressures. The positive Grüneisen parameters of three out-of-plane phonons are calculated to be 1.08, 1.16, and 0.29. Our results increase understanding of the mechanical and thermal properties of Ti3C2Tx at high pressures.

Methods

Ti3C2Tx powder was prepared by a method reported by reference [43]. Briefly, Ti3AlC2 powder (Forsman, 10 g) was etched by HF solution (160 ml) at room temperature for 5 h. The obtained Ti3C2Tx powder was dispersed into DI water and ultrasonically exfoliated at a power of 700 W. The resulted solution was separated after being stored by 24 h. The obtained upper layer solution was used for further Raman, atomic force microscope (AFM), and scanning electron microscope (SEM) analyses. X-ray diffraction (XRD) spectra at ambient pressure were measured using an X-ray diffractometer (Rigaku, MiniFlex600). SEM images were obtained using a scanning electron microscope (Hitachi, Su1510). In situ high-pressure XRD measurements were performed at the Shanghai Synchrotron Radiation Facility by a gasket high-pressure diamond anvil cell (DAC) at room temperature. To produce a hydrostatic environment around the sample, we used methanol/ethanol/water (16:3:1) as a pressure transmitting medium. The pressure was determined by the pressure-dependent spectral shift of the sharp ruby fluorescence R1 line. The sample was placed in a stainless steel gasket hole (100 μm in diameter) with a diamond culet (400 μm in diameter). High-pressure Raman scattering measurements were performed using a Renishaw inVia Raman spectrometer with an excitation wavelength of 532 nm. The topographical measurements were conducted on an AFM instrument (Bruker, Innova).

The fitting of diffraction peak positions and Raman peak shifts was conducted on OriginPro package. A user-defined function, y = (A1 × x + 1)A2, was used by setting A1 and A2 as the fitting parameters. Proper fitting can be obtained easily by a simplex method.