Facile Solution Processing of Stable MXene Dispersions towards Conductive Composite Fibers

Abstract 2D transition metal carbides and nitrides called “MXene” are recent exciting additions to the 2D nanomaterials family. The high electrical conductivity, specific capacitance, and hydrophilic nature of MXenes rival many other 2D nanosheets and have made MXenes excellent candidates for diverse applications including energy storage, electromagnetic shielding, water purification, and photocatalysis. However, MXene nanosheets degrade relatively quickly in the presence of water and oxygen, imposing great processing challenges for various applications. Here, a facile solvent exchange (SE) processing route is introduced to produce nonoxidized and highly delaminated Ti3C2Tx MXene dispersions. A wide range of organic solvents including methanol, ethanol, isopropanol, butanol, acetone, dimethylformamide, dimethyl sulfoxide, chloroform, dichloromethane, toluene, and n‐hexane is used. Compared to known processing approaches, the SE approach is straightforward, sonication‐free, and highly versatile as multiple solvent transfers can be carried out in sequence to yield MXene in a wide range of solvents. Conductive MXene polymer composite fibers are achieved by using MXene processed via the solvent exchange (SE) approach, while the traditional redispersion approach has proven ineffective for fiber processing. This study offers a new processing route for the development of novel MXene‐based architectures, devices, and applications.


Drying of mlMXenes and re-dispersion in organic solvents
The aqueous mlMXene dispersion was filtered using a 0.22 µm PTFE filter membrane, dried and stored in a desiccator under vacuum until ready for use. This mlMXene powder was redispersed and delaminated in organic solvents (ethanol, IPA, acetone, DMF, DMSO, chloroform and toluene) by adding the mlMXene powder (2 g) into the solvent (400 mL). The mixture was stirred for 18 h at room temperature prior to bath sonication for 1 hr. The resulting dispersion was centrifuged at 500 rpm (38 g) for 1 h (Beckman J2-MC) and the supernatant containing the delaminated MXene was collected. These dispersions were referred to as "RD-dMXene".

Atomic force microscopy of dMXene
The atomic force microscopy (AFM) images of dMXene samples were obtained using Multimode 8-U AFM (Bruker) in a tapping mode (ScanAsyst) using a SCANASYST-AIR probe (resonant frequency 70 kHz and spring constant 0.4 N m -1 ). Samples were prepared by drop-casting a very dilute dMXene dispersion (~50 µg mL -1 ) on a mica substrate and then airdried.
The average size and ζ-potential of dMXene in dispersions were measured by Zetasizer (Malvern Nano-ZS) using disposable folded capillary cells (Malvern DTS 1070). Each test was repeated for at least three times and the dispersions were vigorously shaken before the measurements.
X-ray photoelectron spectroscopy (XPS) data were acquired on a Kratos AXIS Nova instrument (Kratos Analytical Ltd) equipped with a monochromated Al Kα source (hν = 1486.6 eV) operating at 150 W. The samples were immobilized for the analysis by pressing S4 them onto a double-sided carbon tape. Survey spectra were acquired at a pass energy of 160 eV and at an energy interval of 1 eV. The Ti 2p, C 1s, and O 1s spectra were acquired at a pass energy of 20 eV and an energy interval of 0.1 eV. The data were processed using the CasaXPS software program (Casa Software Ltd).

Validation of mlMXene synthesis
Ti 3 C 2 T x MXene synthesis was carried out using previously established LiF/HCl method [1] as described in detail in the Supplementary experimental section. Here we first validate the successful etching of the aluminum layer from MAX phase by scanning electron microscopy (SEM), X-ray diffraction (XRD), and XPS analyses. SEM images (Figure S1a,b) showed that the MAX phase was carved into separated sheets after etching suggesting the successful production of multi-layer MXene (mlMXene). This observation was confirmed by XPS and XRD results. XPS revealed that the characteristic Al 2p peak seen in MAX phase was absent in mlMXene ( Figure S2). X-ray diffraction (XRD) analysis showed that the prominent (104) diffraction pattern at 2θ ~ 39° for aluminum seen in MAX phase was no longer visible in mlMXene ( Figure S1e). Furthermore, all (00l) peaks shifted to lower 2θ after etching and all except the (002) diffraction pattern broadened and their intensities decreased. This (002) peak for mlMXene increased in intensity and sharpness, and shifted to 2θ ~6.5° from ~9.5° (for MAX phase). Based on this (002) diffraction, the interlayer spacing increased to 13.5 Å in mlMXene from 9.3 Å in MAX phase. These results are in accordance with literature reports. [1][2][3][4][5]

Oxidation study of dMXene in water
Relevant reports on Ti 3 C 2 T x MXene stability have proven that MXene oxidation can proceed under several conditions. Studies involving flash heating, [6,7] hydrothermal, [7,8] and solvothermal [9] treatments of multi-layered MXene revealed the formation of TiO 2 particles on nanosheet edges, which were found to be predominantly in the anatase phase. Similar conclusions were reached when Ti 3 C 2 T x MXene was stored in water (in the presence of oxygen). [10] An earlier report had indicated the presence of both rutile and anatase phases upon storage of MXene in water. [11] Contrasting these reports which were carried out for HFderived MXenes, here we show that Ti 3 C 2 T x MXenes obtained by the LiF/HCl method degrade primarily as rutile TiO 2 and this degradation occurs randomly on the sheet surface ( Figure S3). A gradual decrease in the UV absorbance was observed with increasing the storage time ( Figure S4c). Also, the wavelength at maximum absorbance (λ max ) shifted from ~278 nm (day 0) to ~368 nm (day 28) and the broad peak between 700 and 800 nm gradually disappeared. This result agrees well with and follows the trend reported in the previous work on the oxidation of the Ti 3 C 2 T x MXene synthesized by HF route. [10] Our electron microscopy studies suggested that the dMXene oxidation process began with TiO 2 nucleation, which grew over time to form TiO 2 clusters and finally coalesced to form large TiO 2 aggregates. This MXene [12] at ~200 cm -1 broadened on day 3. On day 7, a new sharp peak at ~160 cm -1 appeared, which could be attributed to the doubly degenerate (E g ) vibrations of Ti in TiO 2 anatase phase. [13,14] On day 28, the broad Raman signals for pristine (day 0) dMXene at ~373 cm -1 and ~581 cm -1 shifted to ~433 cm -1 and ~617 cm -1 , respectively. These signals could be attributed to the E g and in-plane stretching (A 1g ) vibrations of oxygen in TiO 2 rutile phase, respectively . [6,7,13,14] Also, a broad peak emerged at ~512 cm -1 , which could be assigned to the antisymmetric (A 1g ) bending vibrations of oxygen in TiO 2 anatase phase. [6,7,13,14] This is in agreement with theoretical studies where MXene surfaces in aqueous medium are shown to be saturated with oxygen that can dissociate and readily diffuse into MXene layers. [15,16] From these studies, it can be deduced that the oxidation process begins with oxygen diffusing into the MXene lattice, resembling the oxidation of titanium carbide (TiC) powder in water. [17] The dissolved oxygen can form strong covalent bonds with the Ti and C atoms of MXene. [15] However, oxidation of Ti occurs first before C because of the more negative Gibbs free energy of Ti oxidation than C. [18] This aqueous-based oxidation process is S10 slower than the oxidation of MXene in air by flash heating, [6] therefore we can expect that over time, the oxidation of C also takes place in regions where Ti oxidation has already occurred to form TiO 2 -rich regions and C-deficient regions on MXene surfaces as illustrated in Figure S3.
We also measured the ζ-potential and pH of the dMXene aqueous dispersions over the storage period. We noticed that the fresh dMXene dispersion was slightly acidic with pH ~5, although the supernatant of this dispersion after centrifugation was relatively neutral and had a pH of ~6 (Figure S5a). For the freshly prepared dMXene dispersion, we found that the ζpotential was -32 mV (at pH ~5), which further decreased to -35 mV when neutralized to pH ~7 by titrating with NaOH ( Figure S5b). These results are comparable with values reported by Ying et al. [19] where ζ-potential in the range of +5 to -20 mV in acidic conditions and around -20 to -40 mV in neutral pH (6 to 8) were reported. Alhabeb et al. [20] reported more negative ζ-potential (below -30 mV at pH 4 to 8 and -60 in a basic conditions). This might be due to the different MAX phase (Ti 3 SiC 2 ) used as the starting material. Monitoring the dMXene dispersion during the degradation process showed a slight decrease in the pH from ~5 to ~3.6 after 28 days (Figure S5a). Similarly the pH of the supernatant of the dMXene dispersion decreased from ~5.7 to ~3.6. We observed that the ζ-potential of the dMXene dispersion increased rapidly from -32 mV (day 0) to around -5 to -10 mV after 3 days ( Figure   S5b). Nevertheless, when the pH of the aliquot was adjusted to around 7 (by titrating with NaOH), the ζ-potential remained stable at around -30 mV throughout the 28 days period.
These observations suggest the release of acidic products and formation of neutrally charged species during the reaction. S11 Figure S5. The effect of storage on pH and ζ-potential of dMXene dispersion. (a) pH of dMXene dispersions and supernatants during the storage period. (b) ζ-potential of original and neutralized dMXene dispersions during the storage period.

Preparation of stable dMXene samples
The preparation of various dMXene dispersions are summarized in Figure 2 (main text).
Using the solvent exchange (SE) approach, we prepared dMXene dispersions in eleven organic solvents from the dMXene dispersion ( Figure S6). We also used the re-dispersion (RD) approach to obtain dMXene dispersions in seven organic solvents by processing the a b S12 mlMXene powder. We compared the above dMXene samples against the aqueous dMXene dispersion under continuous bubbling of argon (Ar-dMXene) for a period of 28 days. Our observations indicated that the SE route was more versatile than the RD route and could be used for a wide range of organic solvents. Most organic solvents (except toluene and hexane) worked for the SE route, while the RD approach only worked for ethanol, DMF, and DMSO.
SEM observations of dMXene samples on day 0 revealed that the flakes retained their sheetlike morphology in all of the three processing approaches (Figure S6). The SE approach also allowed for the sequentially transfer of dMXene nanosheets into other solvents ( Figure S7).

Comparison of UV-Vis absorbance of dMXene samples
In contrast to aqueous dMXene dispersions, the UV-Vis absorbance spectra of the SE-dMXene and RD-dMXene dispersions remained almost the same with no detectable sign of oxidation or degradation of MXene ( Figure S8). Figure S8. UV-Vis absorbance spectra for SE-and RD-dMXene dispersions over 28 days.
The absorbance spectra of the dispersions remained almost the same in all samples. S15

Size measurements of dMXene samples
Dynamic light scattering measurements ( Figure S9) showed an increase in the mean size from ~578 nm to ~2,610 nm for the dMXene dispersion stored in the ambient condition.
Nevertheless, the increase in particle size was less obvious for the Ar-dMXene dispersion.
Except for SE-dMXene in ethanol and IPA that showed slightly larger sizes in DLS measurements, all other dispersions on day 0 were of similar sizes to the original dMXene dispersion. Also, the sizes of SE-dMXenes and RD-dMXenes remained relatively unchanged over the storage period. Compared to the HF-derived Ti 3 C 2 T x MXene dispersions reported previously, [21] our dispersions contain MXene sheets with larger sizes. For instance, for dispersions in ethanol obtained by similar RD route, we measured a z-average of 2,864 nm, while a z-average in the range of 300-400 nm was observed for the HF-based MXene. This difference in particle size could be attributed to the differences in the synthesis approach.

XRD, Raman spectroscopy and XPS analyses of dMXene samples
The characterization of powder samples collected from dMXene dispersions obtained by the SE, RE, and Ar approaches on day 0 showed signatures of delaminated MXene i.e. (002) diffractions in the XRD spectra (Figure S10a), and the A 1g mode of MXene and the D and G bands vibrations in the Raman spectra ( Figure S10b). The shifts in the (002) diffraction to lower 2θ ( Figure S10a) were observed to depend on the solvent used and the preparation of the dispersions. The highest downshift was observed for SE-dMXene in DMSO with 2θ ~5.1° corresponding to a d-spacing of 17.2 Å, which was higher than that of dMXene in water (13.9 Å). This downshift in (002) diffraction and the corresponding increase in the inter-layer spacing has been previously attributed to the intercalation of solvents between MXene layers. [2,3,22] Table S1 and Table S2 summarize the results of peak fitting on high-resolution XPS spectra in the Ti 2p and C 1s regions respectively for different samples.
In the previous report by Zhang et al. [10] , the oxidation process of HF-etched Ti 3 C 2 T x MXene dispersion was studied. In this work, Ti 3 C 2 T x MXene was prepared using LiF/HCl as the etching agent. We observed slightly different oxidation process where the carbon content gradually decreased as the oxidation progressed. This is in contrast to the previous report where the carbon layer developed into disordered carbon upon MXene oxidation. A previous study found that the oxidation of MXene depends on the synthesis route. [23] In this study, we observed that the degradation of Ti 3 C 2 T x MXene into rutile TiO 2 occurs randomly on the sheet surface without noticeable preference to the edges. This indicates that the oxidation is likely to start from the exposed Ti-of MXene on the surface. The variation in synthesis methods could introduce different amount of defect on the Ti 3 C 2 T x MXene, which could result in a different oxidation behavior due to the differences in the exposed Ti-. However, further experiments and modelling are required to clearly elucidate the mechanism and kinetics of MXene oxidation. S18 Figure S11.