Noble-Nanoparticle-Decorated Ti3C2T x MXenes for Highly Sensitive Volatile Organic Compound Detection.

Two-dimensional transition-metal carbides and nitrides (MXenes) have been regarded as promising sensing materials because of their high surface-to-volume ratios and outstanding electronic, optical, and mechanical properties with versatile transition-metal and surface chemistries. However, weak gas-molecule adsorption of MXenes poses a serious limitation to their sensitivity and selectivity, particularly for trace amounts of volatile organic compounds (VOCs) at room temperature. To deal with these issues, Au-decorated MXenes are synthesized by a facile solution mixing method for room-temperature sensing of a wide variety of oxygen-based and hydrocarbon-based VOCs. Dynamic sensing experiments reveal that optimal decoration of Au nanoparticles (NPs) on Ti3C2T x MXene significantly elevates the response and selectivity of the flexible sensors, especially in detecting formaldehyde. Au-Ti3C2T x gas sensors exhibited an extremely low limit of detection of 92 ppb for formaldehyde at room temperature. Au-Ti3C2T x provides reliable gas response, low noise level, ultrahigh signal-to-noise ratio, high selectivity, as well as parts per billion level of formaldehyde detection. The prominent mechanism for Au-Ti3C2T x in sensing formaldehyde is elucidated theoretically from density functional theory simulations. The results presented here strongly suggest that decorating noble-metal NPs on MXenes is a feasible strategy for the development of next-generation ultrasensitive sensors for Internet of Things.

Introduction

Volatile organic compounds (VOCs) exist as organic environmental emissions that pose adverse health effects to humans.[1] They also serve as biomarkers in breath analysis of early cancer and disease identification,[2] as well as indicators of crop health in agriculture.[3] Formaldehyde is one of the highly toxic VOCs in an indoor environment and is extremely harmful to human health even at a low concentration.[4] Thus, the monitoring of formaldehyde and other VOCs attracts great attention in the field of public safety control. To overcome the safety and health challenges by preventing industrial accidents caused by uncontrolled release or manipulation of VOCs, it is crucial to develop a reliable sensor that can detect formaldehyde at room temperature with high sensitivity, selectivity, reversibility, and reproducibility.

Recently, gas-sensing research presented various crucial advancements in materials including metal oxide semiconductors, transition-metal dichalcogenides, and carbon-based nanomaterials (e.g., graphene and carbon nanotubes).[5−7] However, these materials, even with a low-dimensional structure, inherently come with apparent limitations such as high optimal operation temperatures in metal oxide semiconductors, slow reversible recovery in transition-metal dichalcogenides, and cross-sensitivity issues in graphene.[8] Various methods have been proposed to improve the sensing performance of these materials by overcoming the existing problematic issues through advanced functionalization or heterojunction development with noble-metal nanoparticles (NPs; e.g., Au, Pd, and Pt),[9] metal oxides (ZnO, SnO2, and TiO2),[10] or conducting polymers.[11]

Two-dimensional (2D) transition-metal carbides and nitrides (MXenes) are recently emerging layered nanomaterials.[12] MXenes have the general structure of MXT, where M is an early transition metal, X is carbon or nitrogen, and T is a surface termination group, such as −OH, −O, or −F.[13] The process of synthesizing MXenes generally involves wet chemical etching of the Ti3AlC2 (MAX phase) precursor to remove Al layers, followed by an exfoliation treatment. Due to the distinct properties, such as high metallic conductivity, abundant active sites, good microstructural stability, and adjustable hydrophilicity, low-dimensional MXenes have been widely applied in the fields of energy storage,[14−16] electromagnetic interference shielding,[17] water purification,[18] photocatalysis,[19,20] and biosensing.[21,22] Selective etching converts the MAX precursor to Ti3C2 with various amounts and types of surface termination groups (e.g., −OH, −O, or −F), thus rendering MXenes theoretically suitable for selective gas-sensing applications.[23−25]

MXenes potentially pose as a more superior approach to gas sensing than other 2D materials.[26] However, the pristine MXene still suffers from low sensitivity, poor selectivity, and slow recovery and response, just as other pristine 2D materials.[27] Property advantages and room-temperature-sensing ability, however, have made MXenes increasingly popular candidates for gas-sensing applications in Internet of Things.

Recent studies on the modification of MXene surfaces with transition-metal dichalcogenides or metal oxide semiconductors have led to the creation of MXene-based hybrid structures for the enhancement of gas detection performance.[27−29] Moreover, Wang et al. proposed a nanocomposite composed of a Ti3C2T MXene and polyaniline NPs, showing excellent sensing performance due to the synergistic effect of nanocomposites and high catalytic/absorption capacity of Ti3C2T MXene.[30] However, only a few computational and synthesis studies on transition-metal- or noble-metal-functionalized MXenes have been proposed so far,[24,31] and more research on VOC sensing using MXene-based materials is needed. Gao et al. proposed density functional theory (DFT) simulation of MXenes functionalized with different noble-metal (e.g., Au, Co, and Ni) NPs, indicating that the substitution of surface groups of Ti3C2T by the noble-metal atoms is energy favorable, which potentially enhances the sensing capacity of noble-metal-MXene structures.[24] Zhao et al. examined biosensors based on Au–Pd-functionalized MXene nanocomposites and suggested that the nanocomposites possess superior electrical conductivity and a larger specific area, thus enhancing the electron-transfer mechanism in biosensing applications.[31] Noble-metal NPs not only offer resistance against environmental corrosion and oxidation but also provide catalytic effects, contributing to an enhancement of sensing reactions. In addition, noble-metal NPs serve as electron traps that prevent the rapid recombination of electrons and holes, thus further benefiting the electronic performance of sensing materials.[32] Therefore, functionalization of pristine MXene with noble-metal NPs provides a promising avenue toward sensing-enhancing mechanisms, charge carrier transportability, and gas selectivity. Compared to other NPs reported in the literature, Au NPs have shown higher promise for surface functionalization due to their catalytic properties and high resistance to environmental corrosion/oxidation.[33,34]

In this work, we have developed a facile solution mixing method to synthesize Au-decorated Ti3C2T hybrid materials and demonstrated their superior sensing performance toward the detection of formaldehyde, among other VOCs. The Au-decorated Ti3C2T sensor exhibits excellent sensitivity with an ultrahigh signal-to-noise ratio (SNR) and enhanced selectivity toward the detection of formaldehyde even at room temperature (21 °C). As a related issue, the response of the Au-decorated MXenes to oxygen-based (acetone, ethanol, and formaldehyde) and other hydrocarbon-based (toluene, hexane, and ethylene) VOCs is also examined, clarifying the effects of various VOCs on sensitivity and selectivity, along with the difference in sensing mechanisms. DFT simulations are performed to understand the interactions between two model gases (acetone and formaldehyde) on the Au-decorated MXene surface. The controlled surface functionalization of Au NPs not only significantly enhances the sensing performance of Ti3C2T but also greatly increases selectivity by distinguishing hydrocarbon-based VOCs from oxygen-based VOCs. The strategy presented here eliminates the property limitations of pristine MXene and thus provides a practical way for field-deployable sensing applications.

Experimental Section

Ti3C2T MXene Preparation and Au NP Decoration

Ti3AlC2 powder (2 g, particle size <40 μm, Carbon-Ukraine, Ltd.) was selectively etched to remove Al layers in a premixed acid solution of 9 M HCl (20 mL) and LiF (3.2 g) and stirred at 200 rpm for 24 h at room temperature. The mixture was washed through several centrifugation cycles with ultrapure water until the pH value of the supernatant reached approximately 6. The resulting Ti3C2T sediment was collected and rewashed with ultrapure water by vacuum filtration using a polyvinylidene fluoride membrane with 0.22 μm pore size and subsequently dried in vacuum at 80 °C for 24 h. Then, 100 mg of Ti3C2T powder was sonicated in 20 mL of ultrapure water with an ultrasonic bath (Branson, CPX2800H) for 1 h. The bath temperature was controlled at 4 °C to prevent material degradation by the release of thermal energy during sonication. The aqueous solutions of gold(III) chloride trihydrate (HAuCl4·3H2O, Sigma-Aldrich, 99.999%, trace metals basis) were then added to the Ti3C2T dispersion to functionalize Ti3C2T with Au NPs. Molar ratios of HAuCl4 to Ti3C2T were adjusted over a wide range of values from 0:1, 0.25:1, 0.5:1, and 1:1 for the preparation of Au-decorated Ti3C2T. The colloidal suspensions were collected for sensor fabrication.

Gas Sensor Fabrication

First, nanogold ink was purchased from UT Dots, Inc. for the printing of six pairs of gold interdigitated electrodes on a polyimide substrate using a commercial inkjet printer (Dimatix DMP-2850, Fujifilm). The total active electrode area was 8 × 8 mm. Then, sensing layers of Ti3C2T and Au–Ti3C2T were sprayed onto the interdigitated electrodes by using an airbrush (G-233, Master Airbrush) for 10 s and finally dried under vacuum. The spray conditions were achieved with an operating pressure of 80 psi, a nozzle of 0.5 mm size, an operating distance of 30 cm between the spray nozzle and the substrate, and a steady moving speed of 10 cm/s in all directions.

Characterization and Gas-Sensing Measurements

Surface morphologies of the Ti3C2T and Au-decorated Ti3C2T MXenes were examined by scanning electron microscopy (SEM; S-4800, Hitachi). The phase structures were investigated by X-ray diffraction (XRD; X’Pert Pro, Panalytical) operated at 45 kV and 40 mA using Cu Kα radiation. X-ray photoelectron spectroscopy (XPS; PHI 5000 VersaProbe, ULVAC-PHI) was conducted to examine the chemical components and bonding structures. To explore the sensing properties, all the gas-sensing measurements were tested at room temperature in a homemade gas-sensing system.[35] The fabricated sensors were loaded into a sealed Teflon chamber with a gas inlet and outlet. Then, the real-time electrical signals were acquired using a source measure unit (Keithley 2400) with a wireless sensor system (see below). The sensor response is defined using the following equation: ΔR/R0 = (R – R0)/R0, where R and R0 represent the real-time resistance of the sensors in the presence of VOC analytes and air, respectively. Mass flow controllers (5850E, Brooks Instruments) were used to maintain the flow rates of each VOC analyte at a total flow rate of 500 mL/min. The gas concentrations were calibrated with a commercial VOC detector (ToxiRAE Pro PID, Honeywell).

Wireless Sensor System

Electrical signals of the gas sensors were monitored by a wireless sensing platform.[35] The corresponding system block diagram is shown in Figure S1. The wireless sensing platform is powered by nRF52832 System-on-Chip from Nordic Semiconductor. The wireless sensing platform features an ultralow power 32-bit ARM Cortex-M4F processor with a built-in radio that operates in the 2.4 GHz ISM band with support for Bluetooth Low Energy (BLE). It is also equipped with a 512 kB of flash storage for storing data. An integrated humidity and temperature sensor (HDC 2010, Texas Instruments) provides high-accuracy measurements. The quad readout interface is implemented by TMUX1574, a CMOS switch offering a 2:1 single-pole double-throw switch configuration with four channels for selecting different signals from different inputs, which is connected to a Wheatstone bridge with a digital potentiometer (AD5241) controlled by the microprocessor via I2C bus. The potentiometer is adjusted automatically to match the input resistance of the sensors based on the analog-to-digital converter (ADC; NAU7802) reading. The system is powered by a two-stage power supply providing dedicated power for both digital and analog circuits. The first-stage power supply is enabled with a DC–DC converter (TPS62051) and provides power for the microcontroller and the digital circuit. The second-stage one is enabled with a power-distribution switch (TPS2051B) for powering the ADC and analog circuit. The system supports quad inputs simultaneously with minimal power consumption and an extremely high electronic resolution of 0.0008%.

DFT for Adsorption Energy Calculations

All the DFT calculations were implemented by using the Vienna ab initio software package.[36−38] The generalized gradient approximation with the parametrization of Perdew–Burke–Ernzerhof function was employed to describe the exchange-correlational energy term.[39] The projector-augmented wave method was used to express the electron–ion interactions of each atom.[40] The Grimme’s correction method was used to correct the effective van der Waals interactions.[41] The plane wave energy cutoff at 500 eV was conducted with a force threshold of 10–4 eV/Å for the ionic relaxation. To ensure the decoupling of consecutive slabs, the 4 × 4 × 1 supercell contained a vacuum thickness of 20 Å. The 5 × 5 × 1 Monkhorst–Pack k-point grids were used for geometric optimization of static energy and electronic structure calculations.[42] To evaluate the interaction between gas molecules on Au–Ti3C2T MXene, the adsorption energies (Ea) of the gas molecules on Au–Ti3C2T were calculated by the following equationwhere is the total energy of the gas molecule adsorbed on Au–Ti3C2T–Au, is the energy of the Au atom decorated on the Ti3C2T layer, and Egas is the energy of an isolated gas molecule.

Results and Discussion

Figure a illustrates the overall process flow of fabricating Au-decorated Ti3C2T MXene for VOC detection. First, Ti3AlC2 powders were selectively etched in an LiF/HCl aqueous solution to form the Ti3C2T MXene. For functionalization of Au NPs into Ti3C2T MXene, an aqueous solution of gold(III) chloride hydrate (HAuCl4·3H2O) was mixed with Ti3C2T at controlled molar concentrations. Au NPs were spontaneously formed on the Ti3C2T layers due to a redox reaction between Ti3C2T and noble-metal ions (Figure b). The work function of Ti3C2T is around 4.8 eV, which is well above the reduction potential of AuCl4– (+1.002 V vs standard hydrogen electrode, SHE).[33] Hence, the electrons spontaneously transfer from Ti3C2T to the Au ions, resulting in the formation of Au NPs. The chemical-sensing mechanism of the conductometric gas sensor is based on the electrical conductivity changes of sensing materials upon exposure to gas molecules with electron-donating or electron-withdrawing property.[43]

Figure 1

(a) Schematic preparation for the fabrication of Au-decorated Ti3C2T MXene and gas sensor device. (b) Energy diagram revealing the work function of Ti3C2T and the reduction potential of Au3+ (1.002 V vs SHE). SEM images of (c) Ti3AlC2, (d) Ti3C2T, and (e) Au-decorated Ti3C2T, along with (f) their XRD patterns.

SEM imaging was used to observe the successful synthesis of Ti3C2T from the selective etching of Ti3AlC2 and the even decoration of Au NPs on Ti3C2T surfaces. Figure c,d shows the SEM images of Ti3AlC2 and as-etched Ti3C2T, respectively, indicating a successful transition from a bulk Ti3AlC2 to an accordion-like Ti3C2T MXene with interlayer spaces that can serve as molecular sieving channels for hosting organic molecules and ions.[44−46] As shown in Figure e, the Au NPs with a diameter around 30 nm are uniformly decorated on Ti3C2T sheets. Size and distribution of surface-decorating noble-metal NPs significantly impact the performance of the gas sensors.[47,48] To study this issue, we fabricated samples of Ti3C2T MXenes decorated with Au NPs having average diameters approximately from 30 to 180 nm and compared their sensing performance. As shown in Figure S2a, the modification of size/distribution of Au NPs was made through varying the HAuCl4-to-Ti3C2T molar ratios. Figure S2b shows that the highest formaldehyde response is obtained from the Au-decorated Ti3C2T synthesized by an optimal HAuCl4/Ti3C2T molar ratio of 0.25:1. The Au-decorated Ti3C2T fabricated at a 1:1 HAuCl4 to Ti3C2T ratio demonstrates a significant decrease in sensor response with the increase of HAuCl4 molar concentration due to an excess amount of large Au NPs, which block adsorption sites of Ti3C2T, as also observed in a relevant SEM image (Figure S3).

XRD analysis was performed to examine the crystal structures and phases of Ti3AlC2, Ti3C2T, and Au–Ti3C2T (Figure f), showing the successful formation of Ti3C2T MXene after selective etching of Ti3AlC2 MAX phase, as well as decoration of Au NPs. Upon the etching treatment, the (002) peak broadens and downshifts from 9.5 to 7.0°, suggesting an increase of interplanar distance from 13.4 to 16.0 Å; the (104) peak at 39° originating from the Ti3AlC2 bulk greatly shrinks from the most prominent intensities to the background level, indicating the elimination of Al from Ti3AlC2T and the formation of 2D Ti3C2T MXene.[49] After surface functionalization, the Au–Ti3C2T sample exhibits the peaks at 38.3 and 44.3°, corresponding to the (111) and (200) planes of the Au NPs. The result confirms the successful functionalization of the Au NPs on Ti3C2T MXene.

The elemental composition and bonding configuration of Au-functionalized Ti3C2T were further investigated by XPS. The high-resolution spectra of Ti 2p, O 1s, C 1s, and Au 4f, taken from the Au–Ti3C2T sample, suggest the presence of Au NPs on Ti3C2T MXene after surface functionalization (Figure ). The Ti 2p spectrum in Figure a is deconvoluted into four peaks centered at 454.6, 455.6, 456.6, and 458.6 eV, corresponding to Ti–C, Ti2+, Ti3+, and Ti–O, respectively.[50] The spectrum is fitted with four doublets (Ti 2p3/2 and Ti 2p1/2) with an area ratio of 2:1 and a doublet separation of 5.7 eV, which are characteristics of Ti3C2T.[51] The O 1s spectrum in Figure b is deconvoluted into four peaks located at 529.7, 530.3, 531.3, and 533.1 eV, assigned, respectively, to Ti–O, C–Ti–O, C–Ti–(OH), and adsorbed H2O.[52] The C 1s spectrum in Figure c is fitted with four peaks centered at 281.6, 284.7, 285.1, and 288.1 eV, corresponding to C–Ti, C–C, CH/CO, and COO, respectively.[52] The Au 4f spectrum in Figure d is observed at 84.4 and 88.1 eV, which correspond to Au 4f7/2 and 4f5/2, respectively. This result also confirms the reduction of the Au precursor into metallic Au NPs and the successful decoration of Au NPs on Ti3C2T MXene.[33] The presence of fluorine groups on both pristine Ti3C2T and Au–Ti3C2T surfaces was also observed with a dominant peak around 685.1 eV in Figure S4, which corresponds to fluorine terminations. In general, the hydrophilic terminal groups are more favorable for gas adsorption than hydrophobic terminal groups and thus the fluorine terminations on the MXene surfaces are not desirable for the gas adsorption.[53]

Figure 2

High-resolution XPS spectra of (a) Ti 2p, (b) O 1s, (c) C 1s, and (d) Au 4f, showing the chemical composition and bonding configurations of Au-decorated Ti3C2T MXene.

The influence of Au NP decoration on sensing performance of Ti3C2T MXene, using individual Ti3C2T and graphene as references, was first examined by comparing its responses toward 20 ppm of formaldehyde (Figure a). First, the inkjet-printed graphene and pristine Ti3C2T show low responses of 0.85 and 1.5%, respectively. The graphene exhibits the lowest response due to its lack of inherent band gap, which reduces its sensitivity to gas analytes.[54] The pristine Ti3C2T, a layered 2D material, has a higher response since it not only provides larger specific surface area and more abundant active sites for gas reactions but also offers higher metallic conductivity for fast electron transport.[26] These properties and the higher response still overwhelmingly propose that Ti3C2T has an enhanced sensing capacity to formaldehyde over graphene. However, MXene materials do not have a significant increase in response potentially owing to their surface termination groups, which largely affect the properties of MXenes.[55] Decorating Au NPs on the surface of Ti3C2T achieves the highest response of 3.0% toward formaldehyde. This enhancement is attributed to the uniform decoration of Au NPs, which not only improves the electrical properties of MXenes but also provides a catalytic effect to the sensing reaction. The response and recovery time were defined as the time taken by the sensor to reach 90% of the total resistance change in the, respective, case of adsorption and desorption (Figure S5). The response and recovery times of Au–Ti3C2T sensors to 20 ppm of formaldehyde at room temperature were 3.0 and 6.7 min, respectively. The results demonstrate a fast response region and a slow response region, which are mainly a result of physical interactions and strong chemical interactions, respectively.[56] Moreover, a good electrical contact between sensing materials and electrodes is required for establishing good gas response. Thus, the electrical contacts of Au–Ti3C2T sensors are characterized by a current (I)–voltage (V) curve. Figure S6 plots the I–V characteristics of the sensing devices in ambient air at room temperature. The linear I–V curve indicates a good Ohmic contact of the sensors.

Figure 3

(a) Response of graphene, Ti3C2T, and Au-decorated Ti3C2T upon exposure to 20 ppm formaldehyde. (b) Response vs time toward formaldehyde (top) and acetone (bottom) for Au–Ti3C2T from 1 to 20 ppm and (c) the corresponding response vs concentration plots derived. (d) Five consecutive dynamic sensing (top) and long-term stability evaluation within 5 weeks (bottom) of Au–Ti3C2T toward 20 ppm formaldehyde.

Figure b shows the response versus time curves for two competing VOCs: formaldehyde and acetone over a concentration range from 1 to 20 ppm, and the response values versus concentrations derived is depicted in Figure c. Over the entire range of concentrations, the responses for both VOCs gradually increase with the increase of gas concentration, and Au–Ti3C2T displays a significantly higher response to formaldehyde than acetone. Au–Ti3C2T still has an excellent response (0.5%) toward formaldehyde, even suggesting its significantly low theoretical limit of detection (LOD). At higher concentrations of VOCs, the sensor shows a slight saturating behavior, which is attributed to the full occupancy of the adsorption sites on the surface.[54,57] The various levels of gas response to acetone and formaldehyde are attributed to the differences in size and geometry of the two molecules and are also related to the magnitude of the MXene-target molecule interfacial adsorption energy, as proposed later. Repeatability and long-term stability are crucial to developing practical state-of-the-art sensors. To evaluate their potential for long shelf life, sensors were stored in a desiccator at standard atmospheric pressure and removed every week to measure their response to 20 ppm formaldehyde. Figure d depicts cyclic curves of 20 ppm formaldehyde (top graph) and long-term stability of a Au–Ti3C2T sensor over 5 weeks (bottom graph). From the five-cycle curve, a consistent response of 3.0% is observed, showing repeatable and reliable sensing response toward formaldehyde detection over the 5 week time period. After 5 weeks of testing, the sensor maintains the original sensing characteristics and a standard deviation of less than 4.3% of average response values, indicating its high consistency and stability for long-term VOC sensing.

Sensitivity and selectivity of Ti3C2T and Au-decorated Ti3C2T sensors for formaldehyde (CH2O), acetone (C3H6O), and hexane (C6H14) were evaluated by the dynamic response plots represented in Figure a. The response plots in Figure b include an in-depth analysis of the two sensing films toward various oxygen-based and hydrocarbon-based VOCs. The responses to oxygen-based VOCs (formaldehyde, ethanol, and acetone) for both Ti3C2T and Au-decorated Ti3C2T are significantly higher than those to hydrocarbon-based VOCs (hexane, toluene, and ethylene). This is attributed to the fact that the interaction between MXene-type surfaces and hydrocarbon-based VOCs is minimal, where the electron clouds of their methyl groups induce merely dipole scattering, thereby reducing overall electron transfer and yielding a lower response.[58] The Au-decorated Ti3C2T exhibits extremely high sensitivity and selectivity toward all the three oxygen-based VOCs (ethanol, acetone, and formaldehyde). Notably, the response values of formaldehyde for pristine Ti3C2T and Au-decorated Ti3C2T sensors were 1.4 and 3.0%, respectively, which corresponds to an increase of 114% by the adequate decoration of Au NPs on Ti3C2T. The Au-decorated Ti3C2T induces the highest response to formaldehyde compared to other VOCs, suggesting the selective nature of the Au–Ti3C2T-based sensor toward formaldehyde. The reason for the selectivity can be ascribed to several factors such as the gas adsorption energy, reducibility of gases on the sensing material, and the distance of analytes from the material surface.[59,60] To unveil the mechanism of the observed enhancement of formaldehyde sensing by the decoration of noble-metal NPs on MXene, DFT simulations were used to investigate the interactions of gas analytes on Ti3C2T and Au–Ti3C2T as fully discussed later in Figure . Moisture interference is an important parameter that might affect the sensor performance. Thus, the Au–Ti3C2T sensors were also exposed to water vapors, with concentrations ranging from 200 to 2000 ppm (Figure S7). As observed, the conductometric responses to 200 ppm water vapors (1.28%) were significantly lower relative to those to 20 ppm formaldehyde (3.0%), indicating the selective adsorption of formaldehyde.

Figure 4

(a) Responses of Ti3C2T (top) and Au-decorated Ti3C2T (bottom) upon exposure to 20 ppm formaldehyde, acetone, and hexane, as well as (b) their selectivity toward 20 ppm of various VOCs indicated. (c) Maximal SNR values upon exposure to formaldehyde for graphene, Ti3C2T, and Au-decorated Ti3C2T. (d) State-of-the-art performance comparison of the LODs of room-temperature sensing toward formaldehyde based on various 2D materials, revealing the lowest LOD of the Au–Ti3C2T sensor.

Figure 5

DFT minimum energy configurations for (a) formaldehyde and (b) acetone molecules adsorbed on the Au–Ti3C2T surface. (c) Minimum binding energies of formaldehyde and acetone on Ti3C2T and Au–Ti3C2Tvs response values measured from sensing experiments.

One of the most critical requirements for practical state-of-the-art sensors is an extremely low noise level and a high SNR even at room temperature. As shown in Figure S8, the electrical noise levels for graphene, pristine Ti3C2T, and Au–Ti3C2T, determined by measuring the response fluctuations during air introduction, were 0.005, 0.01, and 0.005%, respectively. Graphene has a very small or nonexistent band gap and usually features a low signal in spite of a low baseline noise.[54] Based on the response values and electrical noise levels given above, the SNRs of the graphene, Ti3C2T, and Au–Ti3C2T toward 20 ppm of formaldehyde are displayed in Figure c. Graphene shows the lowest SNR of 25, pairing well with previous findings.[54] Ti3C2T yields a moderately low SNR of 149 due to its high baseline noise. Importantly, the baseline noise of Ti3C2T can be modulated through surface functionalization or alterations of termination groups.[61] Indeed, decorating Ti3C2T with Au NPs induces a significant reduction in baseline noise, establishing an ultrahigh SNR of 594. This finding indicates that functionalizing 2D MXene with noble-metal NPs is a promising strategy for the development of high-performance sensing devices.

To further investigate the LOD of the Au–Ti3C2T sensor, the achievable theoretical LOD for formaldehyde and acetone was calculated with an SNR of 3.[57] The relationship between gas response and the concentrations of formaldehyde and acetone with a linear fitting model is shown in Figure S9. The theoretical LODs of formaldehyde and acetone for Au-decorated Ti3C2T are 92.2 and 422.2 ppb, respectively. Au-decorated Ti3C2T exhibits superior sensitivity toward formaldehyde at room temperature and has the potential of detecting sub-100 ppb concentration. These estimations prove itself to be accurate as it not only sufficiently models the experimental data of both formaldehyde and acetone gases but also reflects low variations for low concentrations of VOCs. Figure d describes a reference comparison of room-temperature LODs for the state-of-the-art, advanced formaldehyde sensors. The Au–Ti3C2T MXene shows the lowest LOD among other pristine and functionalized materials, such as transition-metal dichalcogenides,[62−64] metal oxide semiconductors,[65−67] and carbon-based materials.[68−70] These results indicate that surface functionalization of noble-metal NPs is a promising approach to enhancing the sensing performance of MXenes for reliably detecting extremely low levels of various VOCs. Moreover, further improvement in the sensitivity and selectivity can be potentially achieved by different or dual-cation decoration, as well as by using more precise sensor fabrication methods, such as inkjet printing of such MXene-based composite inks, or atomic layer deposition.

Table S2 in the Supporting Information shows a selective summary of significant studies on the use of various 2D materials, including MXenes, for VOC detection. Among these reports, MXenes have not been previously employed toward the room-temperature detection of formaldehyde. The group’s recent work in MXene VOC sensors has also displayed its unique characteristics including room-temperature sensitivity to oxygen-based VOCs (ethanol and acetone), yet formaldehyde sensitivity on MXene has remained unexplored.[25,35,71] The previous works instead reported, MoS2, as well as graphene oxide hybrids, with either detection ability at higher temperatures than the work herein and/or lower sensitivity. Furthermore, Au-doped MXenes have not previously been reported for any VOC detection. Indeed, MXenes in various configurations have been reported for the detection of alcohols and acetone, albeit with lower performance in terms of sensitivity than that in this work. Thus, we can state with confidence that the Au–Ti3C2T sensors herein exhibit outstanding sensing performance toward formaldehyde, including high sensitivity, very low limit of detection (92 ppb), high durability, and room-temperature sensing. In addition, these Au–Ti3C2T sensors were fabricated on flexible substrates by a facile solution process. In contrast to other methods, such as hydrothermal and atomic layer deposition, the facile solution process methods used here have advantages of scalable and cost-effective production of wireless flexible VOC sensors. This finding brings a promising pathway for functionalizing 2D MXene with noble-metal NPs for room-temperature detection of VOC molecules.

DFT simulations were performed to investigate the mechanisms for (a) the enhancement of VOC sensing by the decoration of Au NPs on Ti3C2T and (b) Au–Ti3C2T being much more sensitive to formaldehyde than acetone. The most stable adsorption configurations of formaldehyde and acetone on Au–Ti3C2T are shown in Figure a,b, respectively. Formaldehyde has a shorter equilibrium distance at 1.40 Å, versus 1.69 Å for acetone, indicating a stronger interaction between formaldehyde and Au–Ti3C2T. Moreover, the calculated adsorption energies of the two gases on Ti3C2T and Au–Ti3C2T are presented in Table S2. We further plot the calculated binding energies against response values, as measured from dynamic sensing experiments, as shown in Figure c. The adsorption energies of formaldehyde for both Ti3C2T and Au–Ti3C2T are greater than those of acetone, indicating stronger sensing response of Ti3C2T-based sensors toward formaldehyde, which is consistent with the experimental results. Furthermore, the adsorption energy of formaldehyde on Au–Ti3C2T (−0.728 eV) is much larger than that on Ti3C2T (−0.615 eV), suggesting that Au–Ti3C2T is highly selective to formaldehyde sensing. As discussed, when compared to acetone, the formaldehyde-adsorbed Au-Ti3C2T system exhibits the smallest equilibrium distance and strongest binding energy, thus pointing at its superior formaldehyde adsorption and sensing properties.

Conclusions

We investigated the promise of a hybrid 2D material of Au NPs-Ti3C2T MXene to detect a variety of oxygen-based and hydrocarbon-based VOCs at room temperature using pristine Ti3C2T and graphene as controls. Au–Ti3C2T hybrid materials were successfully synthesized via a simple solution mixing method, which have a great potential for scalable fabrication. Ti3C2T after Au functionalization exhibits a significant improvement in sensing the VOCs in terms of sensitivity and selectivity, particularly for formaldehyde with an ultrahigh SNR and a low limit of detection below 100 ppb. The enhanced sensor performance is mainly due to the uniform decoration of Au NPs on Ti3C2T MXene, which not only improves the electrical properties of MXenes but also offers catalytic effect for sensing reaction. DFT simulations suggest that the adsorption of formaldehyde on Au–Ti3C2T is much stronger than that on pristine Ti3C2T, rendering a stronger interaction between formaldehyde and Au–Ti3C2T MXene for sensing enhancement. This study demonstrates that optimal Au functionalization of 2D MXene is an effective way for fabricating high-performance VOC-sensing devices.