Facile Fabrication of Oxygen‐Enriched MXene‐Based Sensor and Their Ammonia Gas‐Sensing Enhancement

Various sensing materials have been demonstrated to increase the precision of sensing technology. Nevertheless, this complicates the fabrication process for materials integration to obtain devices that can simultaneously accommodate various gas detectors, like electronic nose. The study here focuses on exploring the sensing response of different functionalization of specific sensing materials to provide an alternative way to achieve selective response to multiple gases. Triethoxysilylpropyl succinic anhydride silane (TESPSA) was introduced on 2D material MXene‐Ti3C2Tx to form carboxylic acid terminated MXene (COOH‐Ti3C2Tx) and alternately coated with polyaniline (COOH‐Ti3C2Tx/PANI). This modification doubled up the gas binding sites and improved the binding strength of the Ti3C2Tx surface to NH3 gas molecules. The 5CC‐COOH‐Ti3C2Tx/PANI sensor prepared from five coating cycles showed the highest sensitivity (214.70 %) with fast gas response rate at 80 ppm NH3 (1.75 % s‐1). Therefore, the different signal responses from specific functionalization of the same sensing material functionalization will allow the possible sensor array fabrication to achieve fingerprint‐like sensing map recognition in the presence of mixed gases.


Introduction
Toxic gases, such as NH 3 , NO x , CO x , H 2 S, and SO 2 , are considered secondary pollutants harmful to human health and the environment. Among them, ammonia (NH 3 ) further reacts to generate other substances, e.g., ammonium sulfate and ammonium nitrate, resulting in the formation of particulate matter (PM 2.5 ). [1,2] Higher NH 3 dose in the air causes severe health impacts, including irritation to the skin, eyes, and nose; headaches; dyspnea; or more severe lung and brain damage. [3][4][5] drawbacks, including the low sensitivity in detecting specific gas from gas mixtures, challenge of multiple metal oxides integration, and requiring higher temperatures during operation. [29,30] Therefore, specially fabrication process, such as direct laser writing, was established to enable the multiple metal oxide integration for mix gas sensing. Conductive polymers are also utilized for NH 3 gas sensors due to good electrical conductivity, low-cost fabrication, high stability, eco-friendly, flexibility, and compatibility with other materials. [18] Among conductive polymers, PANI shows impressive sensing performance with NH 3 gas due to its small dimensions, tunable structure, and excellent electronic properties, improving the sensitivity of gas sensors. PANI-based sensor performed a high efficiency with the gas response of 26% changes in electrical resistance to 100 ppm NH 3 gas at low temperature (0 °C). [31] Moreover, PANI nanofibers owing to their larger surface area and interconnected structures, showed an improved sensing response of 62% at 100 ppm NH 3 . [17] Nevertheless, some drawbacks of PANI need to be addressed, such as poor mechanical strength, long response/recovery time, and relatively low selectivity with a particular gas in the gas mixture. [32] 2D materials are well-known to have larger surface areas, tunable electronic properties, facile surface modification, good compatibility, and robust mechanical properties. As one of the promising 2D materials, MXene, has gained massive attention in gas sensing applications. MXene is a part of the group of transition metal carbides and nitrides with the formula of M n+1 X n T x (n = 1-3). The most popular and easily synthesized MXene is Ti 3 C 2 T x , reported in 2011. [33] Ti 3 C 2 T x features functional groups (T x = O, OH, F), which determine its unique properties, including hydrophilic surfaces, high surface reactivity, and excellent electronic properties. Due to the diverse surface groups and metal layers, Ti 3 C 2 T x displays exceptional electrical conductivity and p-type semiconducting behavior with a tunable energy bandgap (0.92-1.75 eV). [34][35][36] These superior properties make Ti 3 C 2 T x a reliable candidate for gas sensing applications, especially for NH 3 detection due to the active surface (with O, OH groups) of MXene that can promote the interaction with NH 3 gas, showing the effective change of electrical resistance. [37] Lee and co-workers used TiC 2 T x as sensing layers for toxic gas sensors at room temperature and observed that Ti 3 C 2 T x responded highly sensitively to NH 3 gas with 21% of gas response at 100 ppm. [38] Besides, the gas response of Ti 3 C 2 T x at 500 ppm of NH 3 was 6.13%, four times higher than ethanol (1.5%) reported by Wu et al. [24] Hydroxyl and oxygen terminals play a key role in the chemical adsorption of NH 3 . Therefore, Yang and co-workers synthesized alkalized Ti 3 C 2 T x using sodium hydroxide, which improved the NH 3 gas response to 28.87% at 100 ppm. [39] Despite its excellent properties for NH 3 gas sensing application, sensing materials made purely from MXene film still have limitations regarding the easy restacking and agglomerating of MXene particles, causing the loss of active sites also low flexibility due to its particle-like rigid nature. [40] To address this shortcoming, the MXene layers were doped with extra components such as polymers, carbon materials, and other inorganic materials to increase interlayer spacing. [41] With the advancement of sensor technology, people are looking for more precise and accurate sensing techniques.
Electronic nose is currently an accepted platform to achieve high detection accuracy that uses a sensor array combined with an artificial intelligence algorithm. Such a sensor array does require many sensors, showing corresponding responses towards target gases. [40] Different sensing materials may need to be introduced into the sensor array to increase detection accuracy, which challenges the fabrication. Therefore, changing the functionalization of the sensing materials will provide an alternative way to achieve a selective response to multiple gases. This will allow the possibility of building up the sensor map or sensing fingerprint with the help of the artificial intelligence algorithm to resolve the problem of cross-sensitivity, especially in the presence of mixed gases. Therefore, the study here aims to verify the gas response from different surface functionalization under the same system control.
Strong interaction between sensing material and target gas, such as covalent and coordination bonds, is a key to increasing sensitivity and selectivity toward specific gas. Therefore, active terminations onto the Ti 3 C 2 T x surface are significantly important in sensing certain gases. The active hydroxyl terminal group (OH) of Ti 3 C 2 T x can be functionalized with silane molecules to obtain particular terminations for further application. [42] Appropriate terminations are prospected to improve gas sensors' sensing response and sensitivity. Various types of functional groups can be formed onto the surface of Ti 3 C 2 T x , such as amine, aldehyde, fluoride, and carboxyl, by silanization technique. For instance, amino silane and fluoride silane were assembled onto the Ti 3 C 2 T x surface by a silanization process in which hydroxyl terminal groups react with alkyl silane groups to create a covalent TiOSi bond and terminations of silane molecules later are seen as new Ti 3 C 2 T x terminations. [43,44] Besides, carboxyl functionalization onto certain sensing materials was demonstrated that improves the gas response with ammonia gas in gas detection. For example, the NH 3 sensor based on COOH-functionalized carbon nanotubes (CNTs) film increased sensing performance to double times higher than that of pristine CNTs. [45] Moreover, a study by Barkov et al. demonstrated that the efficiency of carboxylated graphene nanoribbons in ammonia detection improved almost twofold gas response since ammonia is adsorbed preferably by carboxyl groups. [46] Wang and co-workers synthesized carboxyl functionalized carbon nanocoil for NH 3 sensors and reported an improved sensing response more than twice compared with pristine carbon nanocoil. [47] Therefore, it is noticeably confirmed that the chemical interaction of ammonia gas molecules with carboxyl groups has the potential to detect ammonia gas.
In addition, due to the diversity of sensing materials and the requirement to enhance NH 3 detection, hybrid composites are synthesized and applied for NH 3 gas sensors. Apart from excellent NH 3 gas sensing capability, Ti 3 C 2 T x with hydrophilic properties and its active functional groups (O, OH) can interact and be compatible with various materials giving the composites for NH 3 sensing. For example, Ti 3 C 2 T x is hybridized with other materials to produce Ti 3 C 2 T x nanocomposites with improved sensing performance, such as Ti 3 C 2 T x /SnO 2 (40% at 50 ppm), [48] Ti 3 C 2 T x /WO 3 (22.3% at 1 ppm), [49] Ti 3 C 2 T x /TiO 2 (40.6% at 30 ppm), [26,50] Ti 3 C 2 T x /SnO (7.8% at 200 ppm), [51] Ti 3 C 2 T x / In 2 O 3 (63.8% at 30 ppm), [52] Ti 3 C 2 T x /GO (6.77% at 100 ppm), [53] Ti 3 C 2 T x -TiO 2 /MoS 2 (147% at 500 ppm), [54] Ti 3 C 2 T x /CuO-MOF, [55] www.advmatinterfaces.de Ti 3 C 2 T x /PEDOT:PSS (36.6% at 100 ppm), [56] and PANI/Ti 3 C 2 T x (400% at 50 ppm). [57] The combination of Ti 3 C 2 T x and PANI exhibited outstanding NH 3 gas response compared to other composites. This indicates the compatibility between the two materials due to their electrostatic interaction and homogeneousness of being the same p-type semiconduction behavior.
In this study, taking advantage of the facile surface functionalization of Ti 3 C 2 T x. , as-synthesized Ti 3 C 2 T x is alkalized with sodium hydroxide. This is, then, functionalized with triethoxysilylpropyl succinic anhydride silane (TESPSA) using a self-assembly technique to form carboxylic acid terminals onto Ti 3 C 2 T x surfaces (COOH-Ti 3 C 2 T x ) (Figure 1). The COOH-Ti 3 C 2 T x has further reacted with the as-synthesized PANI by alternately dip-coating method onto the alumina (Al 2 O 3 ) supported graphene electrode with various dipping cycles. PANI act as the glue that can assist the assembly of Ti 3 C 2 T x flakes, whose layer thickness can be increased by adjusting the alternately dipping cycle. The different coating cycles of the COOH-Ti 3 C 2 T x /PANI-based sensors are utilized to detect NH 3 gas with concentrations ranging from 20 to 80 ppm. Given the superior advantages of the alternately coated Ti 3 C 2 T x /PANI sensing layers and the increasing available gas adsorption sites from the carboxylic acid groups, high sensitivity and gas response with different NH 3 doses are observed. This difference in responsive signal corresponds to the available surface functionalization of the MXene surface. Moreover, the oxygen-contained functional groups such as hydroxyl and carboxyl are very sensitive to the variation in the environment, such as humidity and temperature. Therefore, our sensors are tested with different humidity to decouple the sensing mechanism. This can help to explain and clarify the sensing mechanism in detail to avoid the effect of cross-sensitivity and improve the accurate performance of the signal process that enable the possible utilization as multifunctional sensing materials to realize multimodal sensor. [58,59] The sensing performance and humidity impact on the NH 3 sensors is investigated at two different relative humidity (RH = 22% and 55%).

Materials
Titanium aluminum carbide (Ti 3 AlC 2 , MAX phase) 98.2% purity (Luoyang Tongrun Info Technology Co., China), lithium fluoride (LiF) 97% purity (Acros Organics), and hydrochloric acid (HCl) 37% (Carlo Erba) were used to synthesize titaniumbased MXene-Ti 3 C 2 T x . Sodium hydroxide (NaOH) anhydrous Figure 1. Schematic of COOH-Ti 3 C 2 T x functionalization and alternate dip-coating process with different cycles Ti 3 C 2 T x with PANI via chemical bonding formation. Route 1 represents the alternate coating process of pristine Ti 3 C 2 T x and PANI; Route 2 describes the alternate coating of COOH-Ti 3 C 2 T x and PANI.

Synthesis of Ti 3 C 2 T x and Surface Modification Procedure (COOH-Ti 3 C 2 T x )
Titanium-based MXene Ti 3 C 2 T x was synthesized from Ti 3 AlC 2 phases using a minimally intensive layer delamination (MILD) method, which used a mixed solution of lithium fluoride (LiF) and hydrochloric acid (HCl, 6 m) as an in situ etchant. To obtain higher-quality etched and delaminated Ti 3 C 2 T x , the synthesis process was conducted with two etching cycles (24 h cycle −1 ). At the first etching cycle, Ti 3 AlC 2 was etched in LiF/HCl at room temperature based on a mass ratio of Ti 3 AlC 2 :LiF = 1:1 to collect first-cycle synthesized Ti 3 C 2 T x (1st-Ti 3 C 2 T x ). The 1st-Ti 3 C 2 T x was then exfoliated in LiF/HCl etchant with the mass ratio of 1st-Ti 3 C 2 T x :LiF = 1:1. The reaction temperature was adjusted to 45 °C for 3 h at the beginning stage of the second etching cycle while keeping at room temperature for the remaining period. The product solution was then alkalized with 1 m NaOH solution for 1 h at ambient temperature, washed, and collected the sediments for drying at 60 °C for 24 h to obtain the as-synthesized Ti 3 C 2 T x particles for further modification. [60] The carboxyl functionalization onto the Ti 3 C 2 T x surface was carried out using a molecular self-assembly method. TESPSA (98% of purity) is used as a molecular source for silanization onto Ti 3 C 2 T x surface. 0.5 g of the as-synthesized Ti 3 C 2 T x was added in 5% TESPSA dissolved in 20 mL of toluene. The selfassembly process was conducted at room temperature for 24 h and stirred under a nitrogen atmosphere. The product mixture was washed using fresh toluene, ethanol, and distilled water to remove the loosely bounded TESPSA. The carboxyl-terminated Ti 3 C 2 T x (COOH-Ti 3 C 2 T x ) was finally collected after drying in a vacuum oven at 45 °C for 48 h.

Synthesis of Polyaniline
PANI was produced by oxidative polymerization, in which aniline monomers are oxidized in an acidic medium by APS as an oxidizing agent. Aniline (5.8 mL) and nitric acid (8.5 mL) were mixed into 400 mL of distilled water and stirred for 5 min. The APS oxidizing solution was prepared by adding 14.28 g APS in 80 mL distilled water and stirring to a uniform solution. The APS solution was then added to the above monomer liquid mixture and evenly stirred for polymerization until the color of the product solution changed into dark green after 1 h. Finally, the product was filtrated and washed with excess distilled water to remove the residues by vacuum filter and then dried at 45 °C in a vacuum oven for 48 h. The PANI nanoparticles were collected for further application.

The Alternately Dip-Coating Process of COOH-Ti 3 C 2 T x -PANI Sensors
The chemiresistive sensors (15 × 15 mm) with graphene electrodes screen-printed onto aluminum oxide (Al 2 O 3 ) substrate were used for gas sensing in the present work. The sensing material layers, including COOH-Ti 3 C 2 T x and PANI materials, were alternately dip-coated onto the sensor surface by the selfassembly method, as shown in Figure 1. In brief, the COOH-Ti 3 C 2 T x and PANI solutions with the same concentrations of 10 mg mL −1 were prepared by dispersing two types of particles into distilled water under sonication for 5 min. The sensing device was hydroxylated by oxygen plasma for 1 min before dipping into the sensing material solutions. After oxygen plasma, the sensor was rinsed with distilled water and dipped in the PANI solution for 30 min. The sensing device was dried under nitrogen, then rinsed with distilled water to remove unbounded PANI before immersing it in the COOH-Ti 3 C 2 T x solution. Likewise, after modification with PANI, the sensors were incubated in the COOH-Ti 3 C 2 T x solution for 30 min, dried, and rinsed to complete one coating cycle of COOH-Ti 3 C 2 T x / PANI (1CC-COOH-Ti 3 C 2 T x /PANI). By repeating the alternately dipping processes in PANI and Ti 3 C 2 T x solutions, sensors with three and five coating cycles of COOH-Ti 3 C 2 T x /PANI (3CC-COOH-Ti 3 C 2 T x /PANI and 5CC-COOH-Ti 3 C 2 T x /PANI) were well-designed, as shown in Figure 1. Similarly, the 1CC-, 3CC-,5CC-Ti 3 C 2 T x /PANI sensors were fabricated with the same process and used for further gas detection to compare with 1CC-, 3CC-,5CC-Ti 3 C 2 T x /PANI sensors.

Characterization and Measurement
Scanning electron microscope (SEM) with tungsten heated cathode (VEGA3, TESCAN, Czech Republic) was used for energy-dispersive X-ray spectroscopy (EDX) mapping. A field emission scanning electron microscope (FESEM) (JEOL JSM7800F, Japan) with higher magnification was used for the morphological observation. XPS spectra were obtained using a Thermo Fisher Scientific theta probe equipped with a monochromatic Al Kα X-ray source and operated at 1486.6 eV. The spectra were referenced to the main C 1s peak at 284.5 eV. The data were collected from a surface area of 100 µm × 300 µm with a pass energy of 224 eV, with step energy of 0.8 eV for survey scans and 0.4 eV for high-resolution scans. For quantitative analysis, the sensitivity factors used to correct the number of counts under each peak were: C 1s, 1 During the gas sensing process, the sensors were placed into a closed chamber with a gas inlet and outlet under dry nitrogen. Ammonia (NH 3 ) gas was introduced into the sensing www.advmatinterfaces.de chamber, and the resistance of the sample was measured by a Keithley 2450 source meter. The NH 3 gas concentration was adjusted by diluting NH 3 standard gas (100 ppm) with dry air to obtain 20, 50, and 80 ppm using a mass flow controller. For the humidity adjustment, different saturated salt solutions of LiCl, Mg(NO 3 ) 2 , NaCl, and K 2 SO 4 were used to create separated RHs of 11%, 55%, 75%, and 97% in airtight containers, respectively.
The gas-detecting process was performed as follows. The pure nitrogen was introduced inside the chamber for 60 s in a dry air condition. Subsequently, a controlled amount of ammonia gas (20, 50, and 80 ppm) was pumped into the chamber to interact with the sensing material layer for 180 s. The resistance values of samples during the gas adsorption were measured by Keithley 2450. Nitrogen gas was then blown into the chamber for 180 s to purge the NH 3 gas from the sensor platform. This testing procedure was taken place in several cycles to verify the consistency of the sensor. The total gas testing time is about 1140 s. The gas sensing experiments were conducted at a room temperature of 21 °C. Different humidity conditions were also carried out at RH ≈ 23 and 55% to investigate the dependence of the gas response on the humidity factor.
The sensing response is determined as follows where R g and R a are the resistance values of the sensor upon exposure to NH 3 gas and dry air, respectively, the sensitivity of the sensor is featured as the slope of the gas response curve versus gas concentration.

Material Characterization
The Ti 3 C 2 T x was synthesized by a simple and effective method that enables a high content of OH/O functional groups on the Ti 3 C 2 T x surface via alkalization with sodium hydroxide. [60,61] It is demonstrated that O and OH functional groups are interconvertible in an acidic medium. [62] Therefore, in mildly acidic toluene (pH = 5-6), the oxide groups were converted into hydroxyl terminations and bonded with silane groups of TESPSA via the TiOSi covalent bond. In the rinsing process by excess distilled water, the succinic anhydride groups of silane molecules hydrolyzed readily into succinic acid groups, [63,64] as shown in Figure 1. During the dip-coating process, the COOH and OH-terminals of Ti 3 C 2 T x interacted with amine groups on PANI by hydrogen bonding. [65] The interaction of COOH with protonated amine groups of PANI is expected to be stronger than those of the OH or O groups of Ti 3 C 2 T x with PANI's amine groups due to the higher acidic properties of the COOH terminals. Besides, the available unbound carboxyl group on Ti 3 C 2 T x can further provide stronger adsorption for NH 3 gas molecules. The coordinating interaction is expected to increase electron transfer during the gas sensing process because this bond is formed by the donation of a lone electron pair from one atom (N of NH 3 ) to another atom (H of COOH). [66] The chemical functionalization of the Ti 3 C 2 T x surface was characterized by XPS, as shown in Figure 2.
The XPS spectra showed signals of Ti, C, O, N, F, and Si elements on COOH-Ti 3 C 2 T x /PANI samples in the survey region of 0-1200 eV ( Figure S1, Supporting Information). The fitting results of Ti2p, C1s, and N1s spectra are shown in Figure 2A-I. For the Ti2p spectrum, five doublets under Ti2p3/2 and Ti2p1/2 regions were defined, including titanium carbide (TiC), Ti 2+, Ti 3+ , Ti 4+, and TiF 3 , where the TiC, Ti 2+ , Ti 3+ peaks contributed for Ti 3 C 2 T x structure situated at 454, 455.1, and 456.5-456.6 eV, respectively, in the Ti2p2/3 region. The Ti 4+ peak at around 457.8-458 eV displays for TiO 2 or TiO 2−x F 2x resulted from the Ti 3 C 2 T x oxidation. Moreover, the peak centered at 459.0-459.1 eV, attributed to the TiF 3 impurity. Similar to the Ti2p2/3 region, the peaks of TiC, Ti 2+ , Ti 3+ , TiO 2 /TiO 2−x F 2x , and TiF 3 in the Ti2p1/2 are located at 460.1, 461.1, 462, 463.2, and 464.8 9 eV. [60] The area ratios for the two spin-orbit peaks align well with the 1:2 ratio. Because only the Ti 3 C 2 T x surface is modified, the Ti2p spectrum did not show much difference in all three samples of Ti 3 C 2 T x (Figure 2A), COOH-Ti 3 C 2 T x ( Figure 2D), and COOH-Ti 3 C 2 T x /PANI ( Figure 2G).
The chemical surface modifications from Ti 3 C 2 T x to COOH-Ti 3 C 2 T x were confirmed through the changes of peaks in the C1s spectrum. Figure 2B shows the C1s spectrum of Ti 3 C 2 T x before surface modification. Four peaks are positioned at 281.9, 282.9, 284.6, and 286.0 eV, labeled as CTi, CSi, CC, and CO bonds. The CTi represents Ti 3 C 2 T x which is deposited onto a silicon substrate. The C1s spectrum of the COOH-Ti 3 C 2 T x sample ( Figure 2E) appeared with two more peaks of CO and OCO at 287.7 and 289.6 eV. The OCO and CO are presented to the carboxyl group (COOH) and unhydrolyzed succinic anhydride terminals. These two peaks confirmed the success of Ti 3 C 2 T x surface modification by TESPSA molecules to obtain carboxyl terminations. [67] Carboxyl terminals and a new peak of CN at 285.7 eV on PANI were also found in C1s of COOH-Ti 3 C 2 T x /PANI, as shown in Figure 2H. This demonstrates the presence of available carboxyl functional groups on the Ti 3 C 2 T x surface, which are expected to participate in ammonia absorption. Moreover, XPS results also detected the signal of N atoms in COOH-Ti 3 C 2 T x /PANI. The fitting results of the N1s spectrum are shown in Figure 2I, where two peaks at 398.4 and 400.4 eV were labeled to imine (N) and polyaniline (NH) components. In addition, no signal of the N element was observed in the XPS spectra of Ti 3 C 2 T x and COOH-Ti 3 C 2 T x ( Figure 2C,F). The existence of PANI was confirmed by these two peaks, which demonstrated the success of the alternated dip-coating procedure between COOH-Ti 3 C 2 T x and PANI. Moreover, the ratio of COOH groups over OH groups based on XPS fitting results of the O1s spectrum was estimated roughly two times according to the atomic ratios (Figure S1B-D and Table S1, Supporting Information).
For device fabrication, the alternate dip-coating procedure of MXene and PANI layers onto the sensor electrode is described in Figure 3A. The sensing layers were fabricated using Ti 3 C 2 T x and COOH-Ti 3 C 2 T x individually and were alternately coated with PANI. Moreover, coating COOH-Ti 3 C 2 T x /PANI with one, three, and five coating cycles (1CC-,3CC-,5CC-COOH-Ti 3 C 2 T x / PANI) were designed and used to detect NH 3 gases. The as-synthesized PANI's appearance is flat and thin fiber-like, as shown in Figure 3B. The average width of PANI fibers is estimated at around 100 nm, much smaller than the COOH-Ti 3 C 2 T x size www.advmatinterfaces.de (around 1-10 µm). Therefore, PANI particles or nanofibers (mentioned in the synthesis part) can partially fill the gaps between Ti 3 C 2 T x particles and prevent Ti 3 C 2 T x aggregation. This can ensure the available sensing area from both sides of MXene. Also, the desired amount of sensing materials can be controlled by this alternative self-assembly process. The morphology of COOH-Ti 3 C 2 T x in Figure 3C,D shows the loosely stacked Ti 3 C 2 T x structure and the wider d-space ranging from around a few to hundred nanometers ( Figure S2A, Supporting Information). Furthermore, the Ti 3 C 2 T x layers partially filled by PANI were obtained, as shown in Figure 3C and Figure S2B (Supporting Information). It is noticed that the Ti 3 C 2 T x particle was covered and surrounded by PANI. The 5CC-COOH-Ti 3 C 2 T x /PANI multilayer film was estimated at around 10.58 µm ( Figure S2C, Supporting Information). It is noted that the layer morphology is not uniformly increased along the coating cycle due to the particle-like nature of MXene and PANI. However, the average thickness of the layer structure increased, allowing further evaluation of their sensing performance. The EDX mapping of the COOH-Ti 3 C 2 T x particle was performed and detected the distribution of Ti, C, and Si elements within the Ti 3 C 2 T x particle, as displayed in Figure 3E-H. The clear Si signal proved the successful functionalization of TESPSA molecules on the Ti 3 C 2 T x surface ( Figure 3F).
The NH 3 sensing performances of the sensors based on Ti 3 C 2 T x /PANI and COOH-Ti 3 C 2 T x /PANI with various coating cycles were investigated at 21 °C and 23% of relative humidity conditions, as shown in Figure 4. During the gas sensing process, the resistances of all sensors increased and roughly reached saturation when exposed to NH 3 gas and then gradually decreased to their original values when introduced to the dry air. This confirms the p-type semiconducting behavior of all sensing multilayer films. The gas response could be calculated by (R g −R a )/R a , in which R g and R a are denoted to the resistance www.advmatinterfaces.de value of gas sensors in the NH 3 gas and dry air, respectively. The real-time gas response curves of 1CC-,3CC-,5CC-COOH-Ti 3 C 2 T x /PANI and 1CC-,3CC-,5CC-Ti 3 C 2 T x /PANI sensors to 20-80 ppm of NH 3 gas are exhibited in Figure 4A,B. The NH 3 sensing response of all sensors rises with the increase of NH 3 concentration ranging from 20 to 80 ppm. It is worth noting that the more coating cycles of sensing materials, the higher the gas sensing response values. In both types of COOH-Ti 3 C 2 T x /PANI and Ti 3 C 2 T x /PANI, the highest gas response, i.e., faster response and better saturation at the higher concentration of NH 3 , is observed in the sensors with five coating cycles. This is likely because the sensing mechanism depends on combining the electrical charge transfer properties of Ti 3 C 2 T x and PANI materials. Both Ti 3 C 2 T x and PANI have p-type semiconducting behavior, which is featured by the majority carriers of holes charge (positive charge carriers). The NH 3 molecules are adsorbed directly onto Ti 3 C 2 T x surface, inducing charge transfer and changing the electrical conductivity. Moreover, more coating cycles lead to higher content of the materials, which improves not only the conductivity of the sensing layer but also the number of active sites for adsorbing more gas molecules. The charge carrier transfer positively correlates to gas adsorption and the conductivity of semiconducting sensing materials. [68] The sensing response of multilayer films gradually improved as the coating cycles of Ti 3 C 2 T x /PANI increased. Therefore, the content of active sites of the 5CC-COOH-Ti 3 C 2 T x /PANI was the highest, allowing the enhanced adsorption and free transfer of charge carriers for the sensing performance enhancement. Figure 4C shows the real-time gas response of 5CC-COOH-Ti 3 C 2 T x /PANI and 5CC-Ti 3 C 2 T x /PANI. It can be seen that the gas sensing performance of the sensor using COOH-Ti 3 C 2 T x is much higher than with the original Ti 3 C 2 T x . The gas response value of the 5CC-COOH-Ti 3 C 2 T x /PANI sensor at 80 ppm NH 3 is calculated to be around 214.70%, which is 2.77 times higher than that of the sensor based on 5CC-Ti 3 C 2 T x /PANI film (≈77.62%). It is especially noted that the saturated state of COOH-Ti 3 C 2 T x /PANI film is greater than that of Ti 3 C 2 T x /PANI film. The response time of both sensors based on COOH-Ti 3 C 2 T x /PANI and Ti 3 C 2 T x /PANI showed a faster response at higher NH 3 concentrations. Due to the difference in adsorption capacity between COOH-Ti 3 C 2 T x and pristine Ti 3 C 2 T x , the response rate is used to investigate the sensing performance of Figure 3. A) Schematic drawing of sensor fabrication procedure with different coating cycles of COOH-Ti 3 C 2 T x /PANI. SEM images of B) PANI, C) COOH-Ti 3 C 3 T x , and D) 5CC-COOH-Ti 3 C 2 T x /PANI sample; elment mappings of a Ti 3 C 2 T x particle with E) Ti, Si, C elements, F) Si element, G) C element, and H) Ti element in 5CC-COOH-Ti 3 C 2 T x stacking structure.

www.advmatinterfaces.de
sensors and defined as the increase of gas response in response time's interval (% s −1 ). The response rate of the 5CC-COOH-Ti 3 C 2 T x /PANI sensor at 80 ppm NH 3 is around 1.75% s −1 and faster than the pristine one with 0.79% s −1 (Table S3 and Figure S4, Supporting Information). Furthermore, the observed recovery time for 5CC-COOH-Ti 3 C 2 T x /PANI sensor is quite comparable with that of the 5CC-Ti 3 C 2 T x /PANI sensor ( Figure S5, Supporting Information). The sensing response versus NH 3 concentration plots of all sensors is illustrated in Figure 4D. As seen in the figure, the gas sensing performance of 5CC-COOH-Ti 3 C2T x /PANI sensor appears to be linearly increasing along with the NH 3 concentrations. A much higher slope for 5CC-COOH-Ti 3 C 2 T x /PANI sensor confirmed the higher sensitivity of the sensors using COOH-Ti 3 C 2 T x compared to those prepared from pristine Ti 3 C 2 T x . According to the linear graph of 5CC-COOH Ti 3 C 2 T x /PANI and Ti 3 C 2 T x /PANI sensors in Figure 4D, the gas response values at 1 ppm NH 3 are estimated to be roughly 59%. And the limit of detection is calculated as 539 ppb based on the signal-to-noise ratio (SNR > 3) (see Table S4, Supporting Information).
PANI and COOH-Ti 3 C 2 T x are interacted based on the hydrogen bonding between the NH of PANI chains and the carboxyl groups of COOH-Ti 3 C 2 T x . Similarly, the pristine Ti 3 C 2 T x interacted with PANI by hydrogen bonding between hydroxyl and nitrogen groups of PANI. The acidity of COOH groups used as dopants can slightly increase the conductivity of PANI. Due to the interaction of COOH and NH groups, the electrical conductivity of COOH-Ti 3 C 2 T x /PANI multilayers is enhanced as a contributing factor for gas sensing performance improvement. [65,69,70] Moreover, the NH 3 detection by Figure 4. The real-time gas response-recovery curves of ammonia sensors at RH = 23% based on A) 1CC-,3CC-,5CC-COOH Ti 3 C 2 T x /PANI and B) 1CC-,3CC-,5CC-Ti 3 C 2 T x /PANI films. C) The comparison of NH 3 gas response between 5CC-Ti 3 C 2 T x /PANI and 5CC-COOH-Ti 3 C 2 T x /PANI sensors on real-time response-recovery curves. D) The sensing response versus NH 3 concentration curves of all sensing films at 20, 50, and 80 ppm of NH 3 concentrations.

www.advmatinterfaces.de
the sensor based on a pure Ti 3 C 2 T x MXene film without PANI alternated coating was carried out to investigate the sensing efficiency ( Figure S3, Supporting Information). The sensing performance of pure Ti 3 C 2 T x film with ≈25 µm thickness was measured at only 14.03% of sensing response at 100 ppm NH 3 . However, the average resistance of this sensor in dry air was observed to be around 172.55 ± 1 Ω, which is significantly lower compared with the sensors based on Ti 3 C 2 T x /PANI films (Table S2, Supporting Information). Furthermore, the resistance change after ammonia adsorption did not alter much. This could be because of the loss of active binding sites of Ti 3 C 2 T x due to restacking and agglomerating Ti 3 C 2 T x particles.
SEM images showed that the Ti 3 C 2 T x layers were not continuous and had gaps between the particles, which gave PANI underneath a chance to be exposed to NH 3 gas molecules. The sensing mechanism of the sensors based on COOH-Ti 3 C 2 T x / PANI films reflected by the resistance change may be associated with various interactions between different components. This can be attributed to the NH 3 adsorption mechanism both on Ti 3 C 2 T x and PANI. Therefore, the experiments were designed to demonstrate the significant contribution of the carboxylic acid functionalized surfaces. The adsorption of NH 3 gas onto the carboxyl-modified Ti 3 C 2 T x is promoted based on the protonation/deprotonation process. When carboxyl interacted with NH 3 , a lone electron pair from the nitrogen atom of ammonia was donated to the hydrogen atom of carboxyl groups and then formed an ammonium ion (NH 4 + ), as described in the equation below [Equation (2)]. [66] This is also termed a proton donating process due to the proton (H + ) transferring from the carboxyl group to ammonia via coordination bonding formation. The loss of protons (deprotonation) of carboxyl groups resulted in the increase of electron density showing the electron-rich state with a carboxylate ion (COO − ) formation. This can be further reacted with NH 4 + to form ammonium carboxylate salt by a reversible coordinate bond, as shown in Figure 5A (1). Due to the increase of electron density in carboxylate functionalization, the hole carriers in Ti 3 C 2 T x recombined with electrons and consequently decreased its conductivity, leading to the increase of resistance of Ti 3 C 2 T x under NH 3 gas exposure. The formation of ammonium carboxylate salt is favorable but is nevertheless reversible. Therefore, when flushed with air, it withdraws H + from NH 4 + and restores the initial state. The resistance of COOH-Ti 3 C 2 T x /PANI sensors then decreases and the sensing is recovered. (2) The interaction of COOH-Ti 3 C 2 T x with NH 3 is the coordination bonding, which is stronger than the hydrogen bond of the original O and OH terminations of Ti 3 C 2 T x with NH 3 . [37] Hence, it can improve the charge carrier transport of COOH-Ti 3 C 2 T x with NH 3 to obtain high sensitivity and enhance sensor selectivity. [71] In addition, as TESPSA self-assembled onto the Ti 3 C 2 T x surface, a hydroxyl group can generate two carboxyl groups which can double up the available adsorption sites over the Ti 3 C 2 T x surface. It can increase the amount of NH 3 molecules that can adsorb onto the Ti 3 C 2 T x surface, leading to a relatively higher resistance change in the COOH-Ti 3 C 2 T x /PANI sensors compared with the pristine Ti 3 C 2 T x /PANI sensors. The higher concentration of NH 3 , the more protons were delivered to NH 3 , and the resistance was increased considerably. Furthermore, one can notice that the percentage of gas response, saturation level, and response time of the COOH-Ti 3 C 2 T x /PANI sensors significantly outperform other sensors made of Ti 3 C 2 T x /PANI. In addition, due to the dip-coating process, in the layer of COOH-Ti 3 C 2 T x , the small gaps between Ti 3 C 2 T x particles still existed ( Figure S2C, Supporting Information). Therefore, a minor amount of PANI can be exposed to NH 3 molecules which also played a role in the increase of the gas adsorption led to the enhancement of gas response. The sensing mechanism of PANI with NH 3 is described in Figure 5A (2), which is based on the protonation/deprotonation mechanism. The PANI is initially protonated during the acidic synthesis condition to form an actively favorable state (N + H). According to the gas performance of these two different sensor types (COOH-Ti 3 C 2 T x /PANI and Ti 3 C 2 T x /PANI), one can see that the detection process is strongly dependent on the types of surface functional groups as well as the coating cycles of sensing materials. The surface modification with TESPSA not only doubles up the available binding sites but also provides a more active binding group of carboxyl compared to the pristine MXene surface of O and OH groups.
The influence of humidity on sensing behaviors of 5CC-COOH-Ti 3 C 2 T x /PANI and 5CC-Ti 3 C 2 T x /PANI sensors is observed and shown in Figure 5B. The NH 3 gas sensing process of both sensors was conducted at RH = 55% of humidity condition with NH 3 concentrations ranging from 20 to 80 ppm. At this high humidity value of 55%, both sensors showed a gradual decrease in sensing responses at all NH 3 concentrations. But the sensor made from COOH-Ti 3 C 2 T x /PANI still outperformed compared to Ti 3 C 2 T x /PANI sensor at higher humidity. The gas response values are calculated as 53.74% and 11.20%, respectively, corresponding to 5CC-COOH-Ti 3 C 2 T x /PANI and 5CC-Ti 3 C 2 T x /PANI sensors at 80 ppm of NH 3 gas ( Figure 5B). It demonstrates that the impact of humidity on the Ti 3 C 2 T x /PANI is more severe than on COOH-Ti 3 C 2 T x /PANI. However, compared to the gas performance in dry air (RH = 23%), the gas response values at RH = 55% are significantly lower. When RH was increased from 23% to 55%, NH 3 sensing responses of both 5CC-COOH-Ti 3 C 2 T x /PANI and 5CC-Ti 3 C 2 T x /PANI sensors at 80 ppm NH 3 dropped roughly 74.97% and 85.57% (from 214.70% and 77.62% to 53.74% and 11.20%, respectively) as shown in Figures 4B and 5C. The results indicate the importance of maintaining the consistency of humidity during the evaluation of the sensing performance of the sensors.
In humid conditions, the water molecules at higher density in the sensing chamber participate in the sensing process with NH 3 gas. The resistance changes of 5CC-COOH-Ti 3 C 2 T x / PANI can be explained through the adsorption of NH 3 or water molecules and their corresponding conductivity change. The water molecules were adsorbed on the surface of COOH-Ti 3 C 2 T x and PANI, forming the layers of water. Each adsorbed water molecule is ionized to a hydronium ion (H 3 O + ) by capturing protons from COOH groups. When the humidity is high, the H 3 O + ions are then intercalated and diffused between COOH-Ti 3 C 2 T x layers to generate a hydrogen-bond network between COOH-Ti 3 C 2 T x sheets by Grotthuss mechanism (H 2 O + H 3 O + → H 3 O + + H 2 O), leading to the increase of ionic www.advmatinterfaces.de Figure 5. A) The sensing mechanism of COOH-Ti 3 C 2 T x and PANI materials toward NH 3 . B) The real-time sensing performance of NH 3 sensors at RH = 55%, and C) its comparison at RH = 23%. D) The resistance changes of 5CC-COOH-Ti 3 C 2 T x /PANI sensor at different RH conditions without NH 3 gas exposure.

www.advmatinterfaces.de
conductivity of COOH-Ti 3 C 2 T x and the decrease in the resistance. [72,73] Similarly, the exposure of PANI to water molecules enhanced the charge carrier transfer between PANI chains which consequently decreased the resistance of PANI. This is because of the proton transfer between polymer and water and the delocalization of the dopant ions due to the distribution of water molecules. [74] Therefore, the resistance of COOH-Ti 3 C 2 T x /PANI sensors decreased with the increase in humidity. A detailed investigation was carried out to monitor the sensor behavior under different humidity without introducing NH 3 gas. The resistance of 5CC-COOH-Ti 3 C 2 T x /PANI sensors at different humidity levels was extracted and plotted in Figure 5D. The resistance of the sensor decreased by the increased humidity level, reflecting the water molecule adsorption onto the surface of 5CC-COOH-Ti 3 C 2 T x /PANI films.
While sensing ammonia at high humidity conditions, the NH 3 molecules react with the water molecule layer, and the protons from H 3 O + are transferred to NH 3 , which is much easier than capturing protons from COOH-Ti 3 C 2 T x and PANI. Moreover, the excessive amount of H 2 O molecules adsorbed on the sensing film prevented the adsorption of NH 3 molecules and hindered the delivery of protons to NH 3 . This leads to a deterioration of the NH 3 sensing response of the sensors at high humidity conditions. Although the sensor performance of COOH-Ti 3 C 2 T x is adversely impacted by the humidity, it is worth noting that this still outperforms other sensors prepared from pristine Ti 3 C 2 T x . Similarly, the pristine Ti 3 C 2 T x can adsorb water molecules onto its hydrophilic active sites on the surfaces or within its interlayers by hydrogen bonds, which form layers of H 3 O + molecules. The layer of water molecules onto the surface of MXene prevents the interactions of the NH 3 molecules with O/OH groups of Ti 3 C 2 T x and hinders the charge transfer among them, resulting in the decrease of sensing response. Moreover, as the number of active groups (O/OH) in pristine Ti 3 C 2 T x is lower than that of COOH-Ti 3 C 2 T x, the overall available active sites for NH 3 adsorption are much lower under higher humidity. Hence, the sensing performance of the pristine Ti 3 C 2 T x is significantly reduced compared to the sensor of COOH-Ti 3 C 2 T.

Conclusions
In this study, the surface functional groups of Ti 3 C 2 T x are modified by self-assembled TESPSA to obtain carboxylic acid terminals. Ammonia sensors based on dip-coating of COOH-Ti 3 C 2 T x /PANI at multiple cycles were fabricated to detect ammonia gas. The sensing performance of the 5CC-COOH-Ti 3 C 2 T x /PANI sensor was observed to reach the highest sensing response of 214.70%, which is 2.77 times higher than that of the 5CC-Ti 3 C 2 T x /PANI sensor. This proves that the high sensing efficiency of COOH-Ti 3 C 2 T x combined with PANI, as a result of increased gas adsorption on the surface and efficient charge carrier transport between COOH-Ti 3 C 2 T x /PANI with NH 3 gas, is suitable for NH 3 gas sensors. Moreover, the result from a dip-coating process with varied cycles of COOH-Ti 3 C 2 T x demonstrated a linear increase in sensing material content and better NH 3 detection. The response rate of 5CC-COOH/Ti 3 C 2 T x is more than two times faster than that of the sensors used Ti 3 C 2 T x /PANI hybrid films. However, the sensing performance of the 5CC-COOH-Ti 3 C 2 T x /PANI sensor decreased with the humidity in the sensing chamber. This can be further optimized to obtain better humidity tolerance sensors, such as removing the humidity effect at elevated temperatures using from self-heating of MXene. [59,75] In the future, with active chemical properties of carboxylic terminals, COOH-Ti 3 C 2 T x is expected to improve the selectivity of the gas sensor with a target gas (NH 3 ) in a mixture gas by selective chemical interaction. Therefore, the different signal responses from specific functionalization of the same sensing material will allow the possible sensor array fabrication to achieve fingerprint-like sensing map recognition, especially in the presence of mixed gases with the help of the artificial intelligence algorithm. Herein, this NH 3 gas sensor can be used to monitor the indoor environment or integrate with polymer substrate to detect meat spoilage that can provide accurate sensing response. [76] Furthermore, our sensor can be fabricated in nano-or microscales due to the usage of nanomaterials such as 2D materials (MXene, graphene), conductive polymer (PANI), and organic molecule (TESPSA). This nanomaterials-based sensor can be fabricated on an arbitrary shape with some advanced fabrication techniques, such as aerosol jet printing, adaptive 3D printing, or intense pulsed light-induced mass transfer. [77] Moreover, MXene materials was investigated to have piezoresistive property, [78] which can be integrated with other piezoelectric materials to realize the stretchable and selfpower device with enhanced sensitivity. Hence, it can apply to wearable sensors [30] or self-powered flexible (standalone) sensors by using flexible and self-powder platforms (triboelectric nanogenerator) such as PI/graphene, PVA/PET/Ag, PLA/Ag, and latex/PTFE. [79][80][81]

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.