Effect of Ti₂CT_x MXene Oxidation on Its Gas-Sensitive Properties

Abstract

The oxidation process was studied for the synthesized low-layer Ti₂CT_x MXene deposited on a special Al₂O₃/Pt sensor substrate using in situ Raman spectroscopy. It is noted that on the ceramic parts of the substrate (Al₂O₃), the beginning of oxidation (appearance of anatase mod phase) is observed already at 316 °C, in comparison with platinum, for which the appearance of anatase is noted only at 372 °C. At the temperature 447 °C, the initial MXene film is completely oxidized to TiO₂. Using scanning electron microscopy and atomic force microscopy, the microstructure and dispersity of the obtained MXene film were studied. It was found that the obtained films exhibit chemoresistive responses to the detection of a wide group of gases, H₂, CO, NH₃, C₆H₆, C₃H₆O, CH₄, C₂H₅OH and O₂, at room temperature and RH = 50%. The highest sensitivity is observed for NH₃. The partial oxidation of the Ti₂CT_x MXene was shown to favorably affect the gas-sensitive properties.

1. Introduction

Two-dimensional titanium-containing carbide MXenes with the general formula Ti_n+1C_nT_x have attracted increased attention from the scientific community in recent years due to their unique properties—a large surface-to-volume ratio, high electrical conductivity, which can be metallic or semiconducting, and surface functionalization by various groups (primarily −F, −OH, −O), whose composition can be adjusted in some limits by choice of synthesis method. [1,2,3]. This combination of unique MXene properties opens up the potential for their use in various fields: in lithium/sodium-ion batteries and supercapacitors [4,5], fuel cells [6,7], photocalysis [8,9], as well as chemical gas sensing [10,11,12,13,14] etc.

In contrast to metal oxide semiconductors (MOS), the classical receptor materials for chemoresistive gas sensors, Ti_n+1C_nT_x MXenes often have metallic conductivity, their outer surface is completely covered by functional groups, and their morphology refers to 2D-nanomaterials. This combination makes this class of compounds attractive for chemoresistive gas sensors with a high signal-to-noise ratio (SNR) [15], including room temperature operation.

The detection of the MXenes mechanism differs significantly from MOS materials. A universal model describing the interaction of the analyte gas with ion-adsorbed oxygen and the formation of an electron-depleted layer (EDL) on the surface layer of the sensitive material can be used for MOS-gas sensors [16]. At present, the question about the mechanism of gas detection by MXenes is debatable. Nevertheless, theoretical and experimental models allowing the description of the appearance of the chemoresistive signal in the atmosphere of different gases have already been observed [17,18,19].

MXenes, due to their predisposition to form hydrogen bonds with functional groups, are extremely sensitive to moisture. The harmonic structure characteristic of MXenes multilayer, the metallic type of conductivity, and the fact that water molecules do not form covalent bonds with the surface of MXenes make them extremely sensitive to humidity and allow a reproducible response. In Zhang C. et al. [20], using FT-IR spectroscopy and the study of the wetting angle on the MXenes surface, the authors found that water molecules from the gas phase can be chemically sorbed on the surface and in the interlayer space of Ti₃C₂T_x MXene. The authors proposed a mechanism whereby the hydroxyl groups thus formed on the surface provide an electrostatic field and prevent charge transfer, which is explained by an increase in electrical resistance (positive response: p-type) with increasing humidity. The nanoscale interlayer space (~1.01 nm) of Ti₃C₂T_x, as well as the surface super hydrophilicity found by the authors, prevents capillary condensation of water and limits the movement of hydroxyl groups, which are characteristic of bulk nanomaterials.

Structural water plays an important role in reducing the overall interaction between the MXenes layers and creating additional space for the adsorption and diffusion of gases with affinity to functional groups. This feature is an advantage over other 2D-layered materials, such as graphene, in which strong Van-der-Waals interactions between two-dimensional sheets and small interlayer distances prevent the intercalation of gas molecules. Koh H-J. et al., in their article [21], in situ using XRD demonstrated that the interlayer space of MXenes contains water fragments (hydroxyl groups or water molecules themselves), which can be removed when the material is exposed to dry nitrogen. The authors showed that ethanol molecules can chemically bind with Ti₃C₂T_x MXene in its interlayer space. As a result of these interactions, the MXene swells, and its interlayer distance increases due to the steric effect. The swollen two-dimensional MXene decreases the number of electrons involved in charge transfer between the layers during ethanol detection, which leads to an increase in electrical resistance and allows the positive (p-type) chemoresistance response to be recorded.

The MXenes show the best sensitivity when detecting gases able to form hydrogen bonds. Majhi S.M. et al., in article [18], showed a positive (p-type) response when acetone is detected by the Ti₃C₂T_x MXene. Kim S.J. et al. [15] also showed that when used as a receptor material, Ti₃C₂T_x MXene demonstrates the greatest response to gases capable of forming a hydrogen bond (acetone, ethanol, propanal and ammonia) than observed for acid gases (NO₂, SO₂ and CO₂). The authors demonstrated the ability to detect 50–1000 ppb of acetone, ethanol and ammonia with a low SNR value. DFT calculations found that hydroxyl groups (−OH) made the largest contribution to acetone bonding to functional groups on the Ti₃C₂T_x MXene surface compared to oxygen (−O) and fluoride (−F) groups. It was found that for OH-groups, the bonding energy value is significantly more negative compared to −O and −F: −0.774 compared to −0.317 and −0.311 eV, respectively. The authors concluded that the terminal hydroxyl groups on the surface of MXenes play a key role in the detection of various gases to form hydrogen bonds [15].

In most of the papers presented in the literature, a positive chemoresistive response of the p-type is recorded when detecting various gases with carbide titanium-containing MXenes. Nevertheless, there are also studies in which the electrical resistance decreases when gases are injected (negative response: n-type). For example, Wu M. et al. [22] observed a decrease in electrical resistance when detecting ammonia by the Ti₃C₂T_x MXenes. The authors explain this by the strong functionalization of the surface by chloride groups (−Cl), which were formed on the surface of MXenes due to the peculiarities of the synthesis (etching MAX-phase in the Lewis acid ZnCl₂ melt, which is different from the classical liquid-phase methods). This phenomenon is not fully explained in this paper. The authors suggest that chloride groups can affect the MXene band structure, which leads to the reverse chemoresistive response.

Nanocomposites based on them, primarily based on titanium dioxide Ti_n+1C_nT_x/TiO₂, are no less interesting receptor materials for gas sensors as compared to individual Ti_n+1C_nT_x MXenes. Such nanocomposites can be obtained, in particular, by thermal oxidation in an oxygen-containing atmosphere of the initial Ti_n+1C_nT_x MXenes. Such Ti_n+1C_nT_x/TiO₂ nanocomposites can have improved gas-sensitive characteristics. In [23], the authors oxidized a Ti₃C₂T_x film deposited on a special multi-sensor chip in air at 350°C. The resulting nanocomposite demonstrated an increased semiconductor response to ethanol at relatively high temperatures (300–350 °C), and a decrease in the response and recovery time of the sensor was noted. In [24], the authors used oxygen plasma to oxidize MXene. The resulting nanocomposite showed an increased response to ethanol by several orders of magnitude. The formation of Ti_n+1C_nT_x/TiO₂ nanocomposites is also possible by holding MXene in an aqueous solution, specifically under the influence of light. In [25], the authors obtained Ti_n+1C_nT_x/TiO₂ by holding the individual components in an ethanol-water mixture. For the synthesized nanocomposite, they were able to increase the sensitivity, as well as lower the threshold of sensitivity to ammonia to 100 ppb. In [26], the authors in situ-oxidized the initial Ti₃C₂T_x MXene using hydrothermal synthesis. The resulting oxidized nanocomposite demonstrated increased sensitivity and selectivity to NO₂ detection. In [27], the authors used a “simple spray method” to spray an aqueous dispersion of TiO₂ onto the surface of Ti₃C₂T_x MXene. The resulting nanocomposite showed improved gas-sensitive characteristics when detecting low NH₃ concentrations at RT, as well as a more stable signal compared to the individual Ti₃C₂T_x MXene.

The present paper is devoted to the study of the oxidation process of the Ti₃C₂T_x MXene film using in situ Raman spectroscopy and to the study of the gas-sensitive chemoresistance properties of the obtained nanomaterials to a wide group of analyte gases.

2. Materials and Methods

2.1. Mxene Synthesis and Film Application

Reagents: powders of metallic titanium (99.9%, 0.5–100 µm, Ruskhim, Moscow, Russia), aluminum (99.2%, 30 µm, Ruskhim), graphite (MPG-8 grade, Ruskhim), potassium bromide KBr (99%, Ruskhim), sodium fluoride NaF (osc. Part 9-2, Reackhim, Moscow, Russia), hydrochloric acid HCl (>99%, Sigma Tech, Moscow, Russia), tetramethylammonium hydroxide solution (CH₃)₄N(OH) (TMAOH, 25%, aqueous solution, Technic, Saint-Denis, France).

The synthesis of Ti₂CT_x MXene was performed by selective etching of aluminum layers contained in the MAX-phase of Ti₂AlC. For this purpose, Ti₂AlC powder was added to a sodium fluoride solution in hydrochloric acid; the Ti₂AlC and Ti₂CT_x synthesis techniques used in this study are described in detail in [28,29].

Briefly, to obtain Ti₂AlC, powders of aluminum, titanium, graphite and potassium bromide were mixed in the ratio n(Ti):n(Al):n(C) = 2:1.2:0.8 and m(Ti+Al+C) = m(KBr), then co-milled, pressed into tablets and heat treated in a muffle furnace at 1000 °C in a protective melt KBr.

To obtain the multilayer Ti₂CT_x MXene, a 1 g sample of the MAX-phase Ti₂AlC mass was slowly added to 20 mL of 6 M HCL solution with 1 g of NaF. The system was then incubated under stirring and at 40 ± 5 °C for 24 h. The resulting powder was separated by centrifugation and repeatedly washed with distilled water until pH ~5–6 was reached. Delamination was performed using tetramethylammonium hydroxide (CH₃)₄N(OH) solution under ultrasonic influence.

The obtained Ti₂CT_x MXene sample was held in an aqueous dispersion at 4–6 °C for 7 days and then used to apply the receptor layer.

The Ti₂CT_x MXene film was deposited by drop casting method on the surface of a specialized Al₂O₃ substrate with platinum interdigital electrodes (from the front side) and a platinum microheater (from the back side). For this purpose, the dispersion system (carbide nanosheets and deionized water in a volume of 50 µL) was applied to the substrate surface in the area of counter-pin electrodes using an automatic dispenser. Then, a step drying in the range of 25–150 °C under reduced pressure was performed to remove the solvent.

2.2. Instrumentation

The microstructure and chemical composition of the film surface were studied by scanning (NVision 40 scanning electron microscope, Carl Zeiss (Oberkochen, Germany, secondary electron detector, accelerating voltage 1–10 kV) and transmission electron microscope (JEOL JEM-1011, Akishima, Japan), as well as X-ray spectral elemental microanalysis (energy-dispersive X-ray (EDX) spectrometer INCA X-MAX 80, Oxford Instruments (Oxford, UK), accelerating voltage 20 kV). The powder phase composition and films were studied by means of X-ray diffraction (XRD) on D8 Advance device (Bruker, Billerica, MA, USA,), CuKα = 1.5418 Å, Ni filter, E = 40 keV, I = 40 mA; the range of 2θ: 5–45°; resolution: 0.02°; signal accumulation time in the point: 0.3 s).

The obtained Ti₂CT_x MXene films were examined by atomic force microscopy (AFM). Surface topography and Kelvin-probe force microscopy (KFM) were studied on a Solver Pro-M scanning probe microscope (NT-MDT production, Zelenograd, Moscow, Russia) using ETALON HA-HR probes with W₂C-based conductive coating (tip curvature <35 nm). The microstructure of the obtained films was studied using SEM and AFM on the Al₂O₃ section between platinum microelectrodes.

Raman spectra were obtained with a Renishaw inVia Reflex Microscope system equipped with a Peltier-cooled CCD. The 532-nm lines of a Nd:YAG laser were used for excitation. The laser light was focused on the sample through a 50× objective to a spot size of ~2 μm. The power of the sample was <0.3 mW. The variable-temperature Raman scattering measurements were made using a THMS600 stage (Linkam Scientific Instruments Ltd., Redhill, UK). The in situ film heating rate was 5 °C/min. After conducting Raman in situ heating experiments on the samples, the temperature was recalibrated on a typical sample using a high-precision Testo 868 thermal imaging camera. Due to the fact that Raman spectroscopy can be destructive for some samples (especially those prone to oxidation, as in this case) due to local heating of the region from which the spectrum is taken after laser exposure, spectra were recorded in different areas of the film, which are nevertheless localized close to each other. This approach avoided the local overheating that could occur with repeated laser exposure, even at low power, which allowed more reliable data to be obtained. All spectra obtained were normalized with respect to the most intense spectrum.

The chemoresistive responses were obtained using a special laboratory setup, a detailed description of which can be found in our earlier papers [30,31]. The gas medium was created in a quartz cell using two Bronkhorst gas flow controllers with a maximum throughput of 100 and 200 mL/min. The quartz cell volume is 7 × 10⁻⁵ m³. The electrical resistance of the obtained oxide films was measured using a Fluke 8846A Digit Precision Multimeter, which has an upper detection limit of 1000 MΩ. The temperature of the sensor was monitored using a platinum microheater pre-calibrated with a high-precision Testo 868 thermal imaging camera. The measurements of gas-sensitive properties, among others, were carried out at room temperature. Before starting the gas-sensitive measurements, the film to be measured was incubated in a baseline gas atmosphere until a stable signal was established.

To measure the signal at different relative humidity (RH), we used a special unit with a bubbler flask; the RH of the gas mixture was controlled by a digital flow-through hygrometer “Excis”. The temperature value of the relative humidity was set and then measured at 20 °C.

All gas-sensing measurements were carried out at room temperature (RT) and 50% relative humidity. The response to H₂, CO, NH₃, benzene (C₆H₆), acetone (C₃H₆O), methane (CH₄), ethanol (C₂H₅OH) and oxygen (O₂) was calculated using the following ratio:

$${\mathrm{S}}_1=\frac{\left|{\mathrm{R}}_{\mathrm{BL}}-{\mathrm{R}}_{\mathrm{g}}\right|}{{\mathrm{R}}_{\mathrm{BL}}}\times 100\%$$

(1)

where R_BL is the baseline resistance (nitrogen (99.9999%) and was used as the baseline for oxygen detection and synthetic air for other gases, and R_g is the resistance at a given concentration of analyte gas.

The response to humidity was calculated using the following ratio:

$${\mathrm{S}}_2=\frac{\left|{\mathrm{R}}_{\mathrm{BL}}-{\mathrm{R}}_{\mathrm{RH}}\right|}{{\mathrm{R}}_{\mathrm{BL}}}\times 100\%$$

(2)

where R_BL is the resistance at 30% relative humidity, and R_RH is a given relative humidity.

3. Results and Discussion

3.1. Investigation of the Phase Composition and Microstructure of the Obtained Ti₂CT_x Mxene

As shown in Figure 1a, the obtained MXene sample contains no significant admixtures of the original Ti₂AlC, and the shift in the reflex position (002) toward lower angles to 2θ = 4.6° indicates the successful delamination of the multilayer aggregates formed by etching the MAX-phase with NaF-HCl. The shift of the reflex associated with the Ti₂CT_x phase for the coating on the specialized substrate toward higher angles to 2θ = 5.3° indicates that the reverse aggregation of the MXene sheets and the decrease in the interlayer distance during vacuum drying occurs during coating formation.

The TEM data indicate partial destruction of the MXene sheets (Figure 1b,c) while holding it in aqueous dispersion, as well as the appearance of loose nanoparticles on the sheet boundaries, probably related to oxidation products—titanium dioxide in different crystalline modification.

3.2. Raman Spectroscopy

In situ Raman spectra recording for Ti₂CT_x MXene films on an Al₂O₃ chip with Pt interdigital microelectrodes (Figure 2a) was performed both for platinum microelectrodes and for the Al₂O₃ area between them (Figure 2b). The results of which can be attributed to the inhomogeneity of the prepared film. It should be noted that modes on the Raman spectra characteristic of the MXene phase are more intensive on platinum.

3.2.1. MXene Raman Spectra for the Platinum Electrodes Area

A Ti₂CT_x MXene film is characterized by three normal modes, ω₁, ω₂ and ω₃, in Raman spectra, located at 270, 380 and 697 cm⁻¹, respectively (Figure 3a). Modes ω₁-ω₃ can be attributed both to vibrations of MAX-phase Ti₂AlC, remaining after synthesis, and to Ti₂CT_x MXene itself [32,33,34,35]. Analysis of the spectrum obtained at RT also reveals other modes: ω_R1 and ω_R2 modes are located at 447 and 821 cm⁻¹, which can be attributed to high convergence to E_g and B_2g modes characteristic of rutile TiO₂ [36,37]. For the rutile phase, the largest intensity should be observed for E_g and A_1g modes with maxima at ~448 and 613 cm⁻¹, respectively, while B_2g should be the least intensive of all first-order modes characteristic of rutile [38]. In this case, however, the ω_R2 (B_2g) mode is the most intensive of all rutile phase modes, which is highly irregular for rutile TiO₂ materials in the literature [38,39]. Such behavior allows us to suppose that the obtained MXene at RT does not include a separate rutile TiO₂ phase, with observed ω_R1 (E_g) and ω_R2 (B_2g) modes arising from Ti−O bonds, with energies close to analogous bonds in rutile, on the MXene surface. Moreover, the observed rutile species on the surface of Ti₂CT_x MXene likely formed as a result of keeping the synthesized sample in water suspension.

In the more long-wave region of the spectrum, ω_D and ω_G modes are located at 1349 and 1583 cm⁻¹, which are characteristic carbon D- and G-bands [32], which are common for many carbon systems with sp² hybridization of the C−C bond. Therefore, they can be attributed to carbon in MXene as well as to graphene layers resulting from the excessive etching of Ti₂AlC by hydrofluoric acid and the partial removal of titanium atoms beside aluminum atoms [40,41,42].

In addition to the main MXene modes, low-intensity modes ω_T1, ω_T2 and ω_T3 were found at 754, 935 and 1454 cm⁻¹, respectively. Modes ω_T1-ω_T3 are characteristic of tetramethylammonium hydroxide (TMAOH) [43], fragments of which had likely been captured in MXene interlayer space after delamination and washing. At 1042 cm⁻¹, the ω_N mode is also present, attributed to the nitro group (NO₃⁻) [44,45], which could have formed on the MXene surface after the gas-sensing properties studies in an atmosphere containing NO₂. The possibility of gaseous NO₂ transformation into an NO₃− group on the surface of the Ti₃C₂T_x/TiO₂ material has been considered in studies [46,47].

The described set of modes is retained by the sample under heating up to 109 °C. At higher temperatures, the spectra undergo significant changes. Starting with 147 °C, no ω_T1-ω_T3 and ω_N modes (characteristic of TMAOH and NO₃⁻ groups, respectively) appear in the spectra. This can be attributed to their desorption and/or decomposition with the following desorption of gaseous decomposition products from the MXene surface. Heating to ≥147 °C results in the overlap of ω₂ mode with the rutile ω_R1 band, yielding a new widened peak. Further heating leads to the transformation of this peak with maximum shifting to a lower wave number. At 410 °C, the maximum of this phase is located at 420 cm⁻¹, which is close to the B_1g mode of the titanium dioxide anatase phase [38]. This might indicate partial restructuring of the crystal lattice. Modes characteristic of Ti₂CT_x (ω₁ and ω₃), as well as of TiO₂ rutile (ω_R2), still remain on the film spectra, with their maxima not undergoing any changes.

The most intensive E_g mode at 156 cm⁻¹ (ω_A1), characteristic of anatase, starts to appear on spectra at 372 °C. The intensity of this mode rises with the temperature, which indicates that MXene is being oxidized to titanium dioxide. The A_1g mode at 608 cm⁻¹ (ω_R3), which is the most intensive mode of the rutile phase [38], also appears on the Raman spectra at temperatures ≥ 372 °C as a result of the oxidation process. The Raman spectrum at 447 °C significantly differs from those recorded for the MXene film at lower temperatures, having the following set of bands: ω_A1 (E_g) and ω_A2 (B_1g) at 156 and 396 cm⁻¹ from the anatase phase and ω_R3 (A_1g) and ω_R2 (B_2g) at 615 and 821 cm⁻¹ from the rutile phase. The obtained values are in good agreement with the literature [37,38].

D- and G-modes are retained up to the temperature of 447 °C. An increase in temperature gives rise to the systematic shift of the G-mode to lower wave numbers. Thus, at 410 °C, the G-mode maximum is situated at 1570 cm⁻¹. The shift of the maximum position can be related to temperature effects arising from in situ heating of the sample.

Thus, analysis of the recorded Raman spectra for MXene film on platinum allows us to infer that the oxidation of MXene begins at 372 °C. It is at this temperature that the Ti₂CT_x/TiO₂ composite with a large content of rutile and anatase phases of TiO₂ forms. At the temperature of 447 °C, the Ti₂CT_x/TiO₂ film fully oxidizes TiO₂, consisting of rutile and anatase.

3.2.2. Raman Spectra for the Al₂O₃ Area

Considering the features of sensing studies, it is the Raman spectra of MXene on a ceramic Al₂O₃ surface between platinum microelectrodes that are of the most interest since chemoresistive signals of the gas-sensitive film arise from this area of the sample. We have observed that Ti₂CT_x spectra recorded on local areas on the Al₂O₃ surface significantly differ from those recorded for MXene film areas situated on the surface of platinum electrodes. We were able to obtain intensive signals for Ti₂CT_x on the surface of the platinum electrodes related to vibrations of various phases. On the Al₂O₃ surface, however, most modes are of low intensity, and for the initial TI₂CT_x spectrum, large background luminescence was observed, which obstructs registering Raman scattering. For this reason, only the region up to 900 cm⁻¹ is shown in the graph with characteristic MXene modes and oxidation products.

On all spectra of the MXene film (Figure 3b), the characteristic set of modes ω_S1-ω_S5, typical for α-Al₂O₃ substrate, can be seen at 377, 417, 574, 644 and 749 cm⁻¹ [48]. At RT, only weak ω₁ mode of MXene at 276 cm⁻¹ and intense mode ω_R2 (B_2g) at 828 cm⁻¹ from rutile (these modes were the most intense for Raman spectra of MXene on the platinum surface—see Section 3.2.1) are observed. No other modes, characteristic of MXene, were found for the film on Al₂O₃‘s surface due to their low intensity and background luminescence. The described set of modes is retained up to 316 °C when the E_g (1) mode at 156 cm⁻¹ (ω_A1) from anatase, as well as E_g (3) at 634 cm⁻¹ (ω_A3), appear, with the latter mode not observed in spectra for the platinum surface. The obtained data are in good agreement with those in the literature [38]. It should be noted that the start of anatase formation owing to MXene oxidation already starts at 316 °C on the Al₂O₃ substrate, while for the platinum substrate, the same is only observed at 372 °C. Further increases in temperature to 410–447 °C result in full oxidation of MXene to TiO₂. Modes characteristic of anatase can be observed in this case on the Raman spectra: strong E_g (1) at 156 cm⁻¹ (ω_A1) and weaker B_1g(2) + A_1g at 510 (ω_A2) and E_g(3) at 634 cm⁻¹ (ω_A3), as well as B_2g at 828 cm⁻¹ (ω_R2) from the rutile phase [38]. It should be noted that on Al₂O₃‘s surface, the oxidation of Ti₂CT_x MXene can be seen to proceed with the formation of mostly anatase phase. Only the weak B_2g mode can be attributed to rutile, which was present from the start. In the case of the final Raman spectra for platinum substrate, ω_R1 (E_g) and ω_R3 (A_1g) of the rutile phase can be observed. These seeming differences in oxidation behavior of the Ti₂CT_x MXene-receptive layer can be connected to the effect of the substrate material used for MXene film deposition on the recorded Raman spectra.

For further studies of the gas-sensing properties of the prepared Ti₂CT_x/TiO₂ nanocomposite, we have chosen the film oxidized at the minimal temperature of 316 °C.

3.3. Microstructure

SEM micrographs for films on the special Al₂O₃/Pt substrate are given in Figure 4. The initial MXene film (Figure 4a,b) exhibits the folded wavelike microstructure: the surface is reminiscent of sea crests. The length of such crests can reach several microns. It can be seen from the micrograph that the film surface is rather developed, which is important for gas-sensing applications, which require a high specific surface area for better adsorption of gases from the atmosphere. Separate nanoparticles can also be seen on the micrographs, having a mostly spherical shape, with sizes of about 200 nm. Such particles might consist of impurities.

Heating up to 316 °C gives rise to significant microstructural changes in the MXene film (Figure 4c,d). It becomes more uniform, although the wavelike microstructure is retained. Spherical particles 1–3 μm in diameter start to appear, most likely being aggregates of TiO₂ in rutile crystal modification. The enlargement of TiO₂ particles can be attributed to the aggregation of smaller particles, as well as to overall phase transformation due to heat treatment of the films. Judging from Raman spectroscopy data given earlier, a nanocomposite is formed under such temperatures, consisting of the MXene phase and TiO₂.

Further increases in heat treatment temperature to 447 °C result in even more pronounced changes in the microstructure (Figure 4e,f). The wavelike structure fully gives way to a smoother surface, consisting mostly of spherical aggregates 1–3 μm in diameter. Smaller spherical nanoparticles, 20–50 nm in size, can be seen from micrographs with higher magnification (Figure 4f). Raman spectroscopy analysis suggests that under such temperatures, the TiO₂ film forms, consisting of a rutile and anatase phases mixture.

The microstructure of MXene films after their heat treatment at 316 and 447 °C was additionally studied using AFM. Topography scans are given in (Figure 5a,b), as well as the distribution of surface potential for the sample of MXene film oxidized at 316 °C. Topography results from AFM are in good agreement with SEM results (Figure 4c,d). Areas with spherical aggregates 2–4 μm in size, as well as narrow (width of ~100–250 nm) and elongated wavelike structures, roughly 100 nm in size, can be seen. Surface potential values for different film areas differ noticeably, with a work function changing varying from 4.57 to 5.00 eV, which is a consequence of film inhomogeneity. The mean work function value was 4.71 eV, which is rather close to the value for individual Ti₂CT_x MXene with a varyingly functionalized surface: 4.5–4.98 eV [49,50]. Thus, the obtained AFM results correlate well with Raman spectroscopy and SEM data and also confirm the formation of the Ti₂CT_x/TiO₂ nanocomposite.

The topography scans are given in (Figure 5c,d), as well as the distribution of the surface potential for the sample of the MXene film oxidized at 447 °C. The obtained film surface is relatively smooth, consisting of low (50–80 nm) spherical aggregates 2–4 μm in size, with small spherical agglomerates 200–400 nm in size situated on their edges and in between them. Under higher magnification and in the phase contrast regime, nanoparticles with a shape close to spherical, size of ~45–95 nm, were found, which is in agreement with the SEM results. The mean square roughness on an area of 100 μm² is just 20 nm. It can be seen from the map of the potential surface distribution obtained from KPFM (Figure 5d) that charge carriers are spread relatively uniformly on the film surface. The mean work function value for all scanned areas was 5.00 eV. This value is in good agreement with those reported in the literature for anatase [51]. For anatase, the work function lies in the range of 4.94–5.07 eV, while for rutile, it is ~4.80 eV. Thus, AFM also confirms the formation of TiO₂ with mostly anatase crystal structure.

3.4. Gas-Sensing Chemoresistive Properties

The gas-sensitive chemoresistive properties at room temperature were studied for the initial Ti₂CT_x MXene film as well as the Ti₂CT_x/TiO₂ film oxidized in situ in a Raman spectroscopy cell at 316 °C. A further increase in the operating temperature (including in an atmosphere of high humidity) led to a significant increase in electrical resistance (R > 1 GOhm), which made it impossible to measure the gas-sensitive properties.

In the first stage, the obtained Ti₂CT_x and Ti₂CT_x/TiO₂ films had their baseline resistances measured in an atmosphere of dry air and dry nitrogen. In both cases, the resistance of the films exceeded 1 GOm, which did not allow their gas-sensitive properties to be measured in these atmospheres. The Ti₂CT_x and Ti₂CT_x/TiO₂ films showed increased chemoresistance sensitivity to humidity: a significant decrease in the baseline was observed as the RH increased. Figure 6a shows the responses to humidity variations in the air atmosphere. As can be seen, when the RH was increased from 30% to 40%, 50%, 64% and 93%, the response value increased from 0% to 46%, 67%, 83% and 96% and to 50%, 76%, 87% and 91% for Ti₂CT_x and Ti₂CT_x/TiO₂, respectively. The initial Ti₂CT_x film showed a higher sensitivity to humidity at elevated RH values (93%). The high MXenes sensitivity to humidity is typical and is well described in the literature. For further gas-sensitive measurements, a gas atmosphere with constant humidity RH = 50% was chosen.

In the following step, the chemoresistive responses were determined for the detection of 100 ppm CO, NH₃, benzene (C₆H₆), acetone (C₃H₆O), ethanol (C₂H₅OH) and 1000 ppm methane (CH₄), H₂. The experimental responses (S1) to the above gases are shown in Figure 7a,b. The selectivity diagrams (Figure 7c,d) are plotted from the obtained response array. The bar chart (Figure 7c) shows the sign of the resulting response: a positive value corresponds to an increase in electrical resistance when the analyte gas is injected (p-response), and a negative value corresponds to a decrease (n-response). The radar chart (Figure 7d) shows the response values (S1 in %) for all analyzed gases.

For the initial Ti₂CT_x MXene film containing its oxidation products in water dispersion, the highest response (24%) was observed when NH₃ was detected. Notable responses were recorded for CO (15%), C₆H₆ (13%), C₂H₅OH (11.5%) and C₃H₆O (8.3%); the response for H₂ and CH₄ detection was 4%. The sensitivity to 0.4–10% oxygen was also additionally measured (Figure 6b). A gradual increase in the response (S1) from 4% to 33% was observed when the oxygen concentration was increased from 1% to 10%. The response to oxygen by MXenes is practically not described in the literature but was nevertheless recorded by us in our previous study [29]. When all the above gases were injected, regardless of their chemical nature, the resistance decreased, i.e., an n-response was observed.

A significant change in gas-sensitive properties was observed for the Ti₂CT_x/TiO₂ composite film obtained by oxidation in air at 316 °C. The highest response (61%) was recorded for NH₃ detection, as for the initial Ti₂CT_x film. Further, notable responses were obtained for the detection of C₂H₅OH (32%) and NO₂ (43%), and the response for all other gases did not exceed 7.6%. As can be clearly seen from the selectivity radar chart (Figure 7d), the oxidized Ti₂CT_x/TiO₂ film showed an increased response to all analyzed gases (except oxygen) compared to the Ti₂CT_x film. Especially noticeable is the sensitivity to NO₂ (the response increased by 10 times), as well as a completely absent response to oxygen. In contrast to the Ti₂CT_x film, for the air-oxidized Ti₂CT_x/TiO₂ film, a different type of response was observed for the detection of almost all gases (except for NH₃ and NO₂, to which the maximum response was observed). When detecting CO, C₆H₆, C₃H₆O, C₂H₅OH, CH₄, H₂ and O₂, a p-type response was obtained, and when detecting nitrogen-containing gases NH₃ and NO₂ an n-type response was obtained.

Figure 8 shows the response of 4–100 ppm NH₃ and the signal reproducibility when detecting 10 ppm NH₃ by the Ti₂CT_x/TiO₂ composite film. As can be seen (Figure 8a), with an increase in NH₃ concentration from 4 to 100 ppm, there is a gradual increase in the response (S1) from 16% to 61% with a noticeable drift of the baseline (which can be explained by the high interaction energy of the ammonia molecule and receptor material), and the response itself is well reproduced (Figure 8b)

During the consideration of the mechanisms of the detection of various gases by the obtained materials, it is necessary to take into account the fact that all measurements were carried out in an atmosphere of 50% RH. Therefore, it can be assumed that the MXene surface, in this case, contains a large amount of sorbed water, as well as hydroxyl groups (OH⁻), i.e., hydroxyl groups can be directly involved in reactions with various gases that precede the chemoresistive response.

In the present work, only the n-type response was observed for the initial Ti₂CT_x sample. When detecting the reducing gases (all gases in this work except NO₂ and O₂), this behavior is typical for n-type metal oxide semiconductors [52] but is not typical for Ti_n+1C_nT_x MXenes, for which a p-type response is most often observed [15,18,53]. Raman spectroscopy and TEM data revealed that the surface of the initial MXene sheets contains oxygen groups, and an impurity of the rutile TiO₂ phase was noted. Thus, in addition to the hydroxyl functional groups of MXene, as well as sorbed water molecules, TiO₂ particles may also be involved in the detection mechanism. The presence of the n-type response detected in our study is not fully understood at this time and requires further investigation.

After oxidation of the initial MXene film in air to Ti₂CT_x/TiO₂ composite, a change in the response character from n- to p-type is observed when detecting CO, H₂ and VOC’s. According to Raman spectroscopy data, it is found that a single n-type TiO₂ semiconductor phase is formed for this sample. The surface microstructure of the resulting composite also changes. As a rule, large-band-gap semiconductors, such as TiO₂, have poor electrical conductivity at room temperature. At room temperature, it is the MXene phase, which is characterized by p-type conductivity, that makes the main contribution to the gas sensitivity. The increase in sensing signals can be related to the fact that at the Ti₂CT_x/TiO₂ interface, a Schottky barrier appears, the height of which will change during the adsorption of different gases [23].

Table 1. Comparison of gas-sensitive characteristics of chemoresistive gas sensors based on Ti₃C₂T_x/TiO₂ and Ti₂CT_x/TiO₂ MXenes operating at room temperature that is presented in the literature.

No.	Year	Composition	Target Gas	Conc., ppm	Response	RH, %	Ref.
1	2019	Ti3C2Tx/TiO2	NH3	10 ppm	3%	60	[27]
2	2020	Ti2CTx/TiO2	NH3	10 ppm	1.9%	0	[25]
3	2020	Ti3C2Tx/TiO2	NO2	5 ppm	16.05%	0	[46]
4	2021	Ti3C2Tx/TiO2	C2H5OH	100 ppm	22.47%	0	[55]
5	2021	Ti3C2Tx/TiO2	hexanal	100 ppm	8.8%	0	[56]
6	2021	Ti3C2Tx/TiO2	NO2	100 ppm	4%	0	[26]
7	2021	Ti3C2Tx (N-doped)/TiO2	NH3	200 ppb	7.3%	0	[57]
8	2022	Ti3C2Tx/TiO2	C2H5OH	90 ppm	91 a.u.	0	[24]
9	2022	Ti2CTx/TiO2	NH3	4–100 ppm	16–61%	50	This work

provides information on the gas-sensitive properties of composites of MXenes Ti₃C₂T_x/TiO₂ and Ti₂CT_x/TiO₂ operating at room temperature for chemoresistive gas sensors reported in the literature. As can be seen, in the field of gas-sensing composites based on Ti₃C₂T_x, MXenes are mainly used, and the sensory properties of Ti₂CT_x MXenes and nanocomposite Ti₂CT_x/TiO₂ are practically not studied. For many of the composites presented in Table 1, it is ammonia that shows the greatest response. This corresponds to the data of theoretical calculations [54] for the maximum oxidized compound, Ti₂CO₂, which is associated with high adsorption energy and the calculated value of charge transfer. A comparison of the experimental data on the gas-sensitive properties shows that the materials obtained in this work have a sufficiently high response (to a wide range of NH₃ concentrations), significantly exceeding most of the literature analogs. The increased response can be attributed to the more flexible heat treatment conditions used in this work under the control of in situ Raman spectroscopy for the controlled formation of heterojunctions between Ti₂CT_x and TiO₂. In addition, the responses obtained in this work were recorded at relative humidity RH = 50% (closest to the real conditions), which has not been noted in other works for TiO₂-modified MXenes.

4. Conclusions

The in situ oxidation of the film of layered Ti₂CT_x MXene synthesized as a result of the exposure of the MAX-phase of Ti₂AlC to sodium fluoride solution in hydrochloric acid, delaminated by tetramethylammonium hydroxide and held in water dispersion, has been studied, in detail, using Raman spectroscopy. The phenomenon of substrate (Al₂O₃ ceramics or platinum) influence on the intensity of Ti₂CT_x MXene phase modes, as well as the modes formed during oxidation in the air atmosphere of rutile and anatase phases, has been noted. It is shown that the formation of a new oxidation product phase (TiO₂ with anatase structure) on the ceramic Al₂O₃ substrate can be fixed at a lower temperature (316 °C) than it is typical for the Ti₂CT_x layer on the platinum electrodes of the specialized substrate (372 °C). It was found that at 447 °C the full oxidation of the initial Ti₂CT_x MXene under the conditions of heating in an air atmosphere occurs.

The coating microstructure was studied using a complex of methods (SEM, TEM and AFM), and the work function was determined for both the Ti₂CT_x/TiO₂ composite obtained at 316 °C and the fully oxidized composition using Kelvin-probe force microscopy.

It was found that the obtained Ti₂CT_x and Ti₂CT_x/TiO₂ films exhibit chemoresistive responses when detecting a wide group of gases, H₂, CO, NH₃, C₆H₆, C₃H₆O, CH₄, C₂H₅OH and O₂, at room temperature and 50% relative humidity. It was shown that partial oxidation of MXene favors the gas-sensitive properties, and the most sensitivity was observed for NH₃.