Strategies and challenges for improving the performance of two-dimensional materials-based gas sensors

ABSTRACT

Two-dimensional (2D) materials emerged in the last decade have been considered as an attractive class of gas sensing materials due to their large surface-to-volume ratio, high surface reaction activity and good gas adsorption performance provided by the atomically thickness and flat surface. The development of high-performance micro-nano gas sensors places higher demands on sensing materials, thus 2D materials must satisfy good sensing performance requirements such as excellent gas selectivity, reproducibility, long-term stability, and fast response/recovery time to achieve high sensitivity detection at low concentrations. Herein, we overview the recent literature on the sensor structure, sensing mechanisms and improved gas sensing performance of 2D materials-based gas sensors. In particular, we discuss how the sensing properties of 2D materials can be modulated through defect and doping engineering, modification with noble metal, constructing heterostructure, as well as external energy assistance, e.g. by light illumination. Besides reviewing the achievements of gas sensors based on 2D materials, we provide an outlook on their future developments. Understanding the relationship between structure and sensing mechanisms of 2D materials-based sensors provides vast opportunities for designing advanced gas sensors, making it promising for a wide range of practical applications.

KEYWORDS:

• Gas sensor
• two-dimensional materials
• strategies
• sensing performances

1. Introduction

Owing to its ability to sensitively detect molecules at ultra-low concentrations, gas sensors have attracted huge interests of researchers and been widely used in many fields such as environmental monitoring, hazardous gas detection, industrial production, and medical diagnosis [1–5]. The sensing material, as the core of the sensor, determines the performance of the device. Among various sensing materials, semiconducting metal oxides are most common in gas sensors due to their facile preparation and low manufacturing cost [6–9]. However, gas sensors based on semiconducting metal oxides suffer high operating temperature (usually 200°C to 500°C), long recovery time and poor selectivity, which limit their practical application [10,11]. Therefore, it is highly desirable to develop novel gas sensors with high selectivity and fast response at room temperature.

Two-dimensional (2D) materials have tremendous advantages in gas sensing. Firstly, the inherent atomically thickness of 2D materials leads to ultra-high surface-to-volume ratio, which is beneficial for gas sensors to exhibit excellent performances such as high sensitivity, low detection limits, and fast response/recovery rates [12–14]. Secondly, different from conventional semiconducting sensing materials that rely on chemical reactions with gas molecules to achieve gas sensing, most gas sensors based on 2D materials relies on direct charge transfer, making gas sensing workable at room temperature conditions [15,16]. Besides, 2D materials exhibit excellent compatibility in ultra-thin silicon channel technology and have been considered as promising candidates for the manufacturing and integration of gas sensor devices [17,18]. What’s more, the properties of 2D materials like morphology, structure, bandgap and carrier mobility can be easily modulated by defect engineering and doping [19,20], which provides new approaches to improving gas sensing performances and achieving more practical applications. For instance, p-type and n-type 2D materials exhibit different sensitivities to gases owing to the various band structures. n-type 2D semiconductors are sensitive to electron donors such as O₂ and NH₃, while p-type 2D semiconductors are sensitive to hole donors such as H₂S and NO₂, which is favorable for designing highly sensitive gas sensors [21,22].

In the last decades, 2D materials have been rapidly developed with the increasing variety, including classical graphene [23], borophene [4], transition metal dichalcogenides (TMDs) [24], black phosphorus (BP) [25], boron nitride (BN) [26], and emerging MOFs [27,28] and MXenes [29], providing a rich material selection for gas sensing manufacturing, which can be proved through the number of published papers. A comprehensive literature search was conducted in Web of Science since 2000, using the topic ‘2D materials for gas sensors’. As shown in Figure 1, the number of articles has been rapidly increasing since 2010 and has continued to grow exponentially especially in recent years from approximately 11 articles published in 2010 to 300 articles in 2022, showing the great prospects of 2D materials in gas sensing field. Moreover, with the development of nanotechnology, methods for large-scale preparation of 2D materials are becoming more mature, allowing precise control of thickness and number of atomic layers, such as exfoliation method, chemical vapor deposition (CVD), physical vapor deposition (PVD), etc., accelerating the practical application of 2D material gas sensors [30–32]. Although 2D materials have made great progress in gas sensing, there are still some limitations hindering their application. For example, the huge specific surface and ultra-high reactivity easily lead to the indifferent adsorption when exposing different gas atmosphere, resulting in poor selectivity [33]. Besides, gas molecules are hard to desorption from the surface of 2D material without heating or illumination, leading to long recovery time at room temperature [34]. As a result, the selectivity, room-temperature recovery rate, and the long-term stability of fabricated gas sensors based on 2D materials still need to be further improved.

Figure 1. Records of the number of publications on the topic of 2D materials for gas sensors. (internet search from web of Science on October 31, 2023).

In this review, we first summarize the structure of gas sensors based on 2D materials. Secondly, the sensing mechanisms of gas sensors are detailed. Subsequently, the recent progress of strategies to improve the performance of 2D materials-based gas sensors is systematically introduced (Figure 2), including defecting and doping engineering, modification with high work function noble metal, constructing heterostructure and external energy assistance. Lastly, the perspectives of 2D materials-based gas sensors are given in terms of the advantages and disadvantages of gas sensors. We believe that the comprehensive study of the sensing mechanisms and improvement measures for sensing performance of 2D materials-based gas sensing will provide new ideas for the development of 2D materials and be valuable to 2D material-based gas sensing fields.

Figure 2. Recent progress of strategies to improve performance of 2D materials-based gas sensors.

2. Sensor structure and testing method

Semiconductor gas sensors can be divided into two major categories based on their sensing mechanism: resistive and non-resistive sensors [1,35]. Concretely, as for resistive gas sensors, gas molecules interact with the device and change the resistance of the sensor to realize gas detection. Instead, non-resistive gas sensors utilize the physical and chemical reactions between the gas molecule and the sensing material to change other electrical parameters, such as the threshold voltage of a field effect transistor and the output characteristics of a diode, to detect the gas molecule. Among them, resistive gas sensors, with easy detection mechanism, have been extensively studied and applied in reported 2D materials [36–38]. This section provides a brief depiction concerning resistive gas sensors with different structures and the related testing method.

2.1 Field-effect transistors (FETs)

The structural schematic of FET gas sensors based on 2D materials is illustrated in Figure 3(a). The FET typically consists of a single 2D nanosheet deposited between the source and drain electrodes to form the sensing channel. The gate electrode is located on the other side of the substrate and is separated from the channel by a thin dielectric layer. Currently, the methods for preparing 2D sensitive materials are developing rapidly, including common methods such as mechanical exfoliation, chemical vapor deposition, atomic layer deposition, hydrothermal method, and solvent thermal method, which can successfully control the size and thickness of the materials as desired [31,40]. When a FET gas sensor interacts with gas molecules, the adsorption or reaction on the surface results in the movement of carriers, leading to alterations in electrical conductivity, volt-ampere characteristics, or surface potential. Through collecting and analyzing the change of resistance value, information such as gas composition and concentration can be obtained.

Figure 3. Schematic of the (a) FET and (b) heterojunction gas sensor based on 2D materials. (c) schematic diagram of the gas sensor test platform [39].

2.2 Heterojunction sensors

Heterojunction sensors are a novel class of gas sensors that consist of two different semiconducting sensing materials, and the structure is illustrated in Figure 3(b). According to the conductivity type of semiconductors, heterojunctions can be classified into p-p type, p-n type, and n-n type. After two semiconductor materials contacting with each other, the Fermi level at the interface will balance to the same energy state. This process is accompanied by the transfer of charge carriers and the formation of a depletion or accumulation layer at the heterojunction. Different from traditional semiconductor materials, the surface of 2D materials has no dangling bonds, therefore, the formed heterojunction has a high-quality interface. Moreover, the carrier concentration and band alignment of 2D materials can be effectively modulated through electrical or optical means, providing new possibilities for expanding the application of heterojunction structures. A large number of studies have shown that constructing heterojunctions can effectively enhance gas sensing performance [41,42], which will be discussed in detail in the fourth section.

2.3 Testing method

In a typical gas sensing test process, the 2D material gas sensor is placed in a closed chamber and a certain test voltage is applied to the device to obtain the resistance value during the gas sensing. Then, the target gas is passed into the closed chamber. At this time, the resistance of the gas sensor will increase or decrease significantly due to the interaction between the 2D sensing material and the target gas. When the resistance value is stable, the target gas is discharged from the chamber, so that the entire chamber environment and the device resistance are restored to the initial state. During the whole process, a data acquisition system such as a computer is used to record the change of the resistance of the gas sensor in real time to obtain the corresponding sensing information, the corresponding test platform is shown in Figure 3(c).

3. Sensing mechanism

3.1 Charge transfer

Charge transfer is the main mechanism of gas sensing in 2D materials. When the gas sensor is exposed to the target gas, charge transfer will occur between the sensing material and the gas molecule, as shown in Figure 4(a) [43]. The direction of charge transfer and the amount of charge transferred are differentiated for different kinds of gases, resulting in distinct changes of the sensor to achieve gas identification and classification. In general, oxidizing gases (such as NO₂, O₂, Cl₂, etc.) tend to take electrons from the surface of 2D materials [45–47], while reducing gases (such as CO, H₂, NH₃, etc.) provide electrons to sensitive materials [48–50]. When the sensor is exposed to the target atmosphere, the target gas molecules adsorb on the surface of the sensitive material with charge transfer, changing the resistance value of the sensor. On the contrary, when the target gas is removed, the target gas molecules already absorbed are desorbed from the surface of the material, so the resistance of the sensor is restored to the original state.

Figure 4. Schematic illustration of gas sensing mechanisms. (a) surface charge transfer model [43]. (b) Schottky barrier modulation [44]. (c) and (d) surface adsorption of oxygen ions

For p-type two-dimensional layered materials, when they are in contact with oxidizing gas molecules, the oxidizing gas tends to acquire electrons from the surface, resulting in an increase in the concentration of holes. As holes are the majority carrier for p-type semiconductor, an increase in conductivity can be observed. Conversely, when they are exposed to reducing gas molecules, their conductivity decreases. For n-type two-dimensional layered materials, the opposite change of conductivity in the device occurs. Typically, the surface charge transfer can be proven through computational simulations based on first-principles calculations using density functional theory (DFT) [51,52].

3.2 Schottky barrier modulation

Another mainstream explanation of the gas sensing mechanism of two-dimensional layered materials is the Schottky barrier modulation [44,53]. It means that the gas molecules adsorbed on the surface of two-dimensional materials will induce a built-in electric potential and modulate the Schottky barriers between 2D material and the electrode, as shown in Figure 4(b). Taking n-type 2D materials as an example, when the surface electrons of the material are captured by the oxidizing gas such as NO₂, the Fermi level of the n-type sensing material tends to move towards the valence band, which increase the Schottky barrier and reduce the built-in electric potential, resulting in a decrease in the electrical conductivity and an increase in the resistance of the sensing material. Conversely, when n-type 2D materials adsorb reducing gases (such as NH₃) on their surface, electrons from gas molecules transfer to the material surface, causing the Fermi level of the material to move towards the conduction band, which reduce the Schottky barrier while increase the built-in electric potential, resulting in an increase in the electrical conductivity and a decrease in the resistance of the sensing material. As for p-type 2D materials where holes are majority carriers, the trend of the conductivity is opposite to that of n-type semiconductors.

3.3 Surface adsorption of oxygen ions

Surface adsorption of oxygen ions is also one of the gas sensing mechanisms for two-dimensional materials, commonly used in high-temperature or light-illuminated conditions for gas sensing [54,55]. Under light irradiation conditions, some electrons transition from the valence band to the conduction band, increasing the density of free charge carriers in the sensing material. At the same time, O₂ in the environment adsorbs onto the surface of the semiconductor material through physical or chemical adsorption, allowing electrons near the surface of the 2D material to be captured by O₂ in the conduction band, forming negatively charged oxygen ions such as O^2-, O⁻, and O₂⁻ as a result. When oxidizing/reducing gases appear, the oxygen ions interact with the gas molecules, causing changes in the electrical parameters of the device. As shown in Figure 4(d), the hole accumulation layer on the WS₂ surface increases under 940 nm light illumination compared with that in dark (Figure 4(c)). When the target gas NH₃ appears, more NH₄⁺ reacts with oxygen ions in the hole accumulation layer, releasing electrons and narrowing the width of the hole accumulation layer. Therefore, the photo-generated electron–hole pairs promote the generation and reaction of oxygen anions, and provide more reaction sites for gas adsorption, which greatly improve the sensitivity for NH₃ gas sensing.

4. Methods for improving gas sensing performance

High sensitivity, low limit of detection (LOD), rapid response/recovery time, good selectivity and repeatability are the general requirements for excellent gas sensing performance. Based on the gas sensing mechanisms, many strategies have been proposed to improve the performance of gas sensors. For example, creating defects in 2D materials would provide more active sites for gas molecules, thereby improving charge transfer between gas molecule and sensing material. The strategies can be roughly divided into five categories: defect engineering, doping engineering, modification with noble metal, heterostructure construction and light illumination, which will be discussed in detail in the following part. Besides, a comparison list about improved performances of the 2D materials-based gas sensors by different strategies is summarized in Table 1.

Table 1. Improved performances of the 2D materials-based gas sensors by different strategies.

Methods	Sensing Material	Target gas	Sensitivity	LOD	Response time (s)	Recovery time (s)	Ref
Defect engineering	Graphene	NO2	248%	100ppm	38	238	[56]
	In-SnSe2	SO2	72.45%	3.46ppb	226	1662	[57]
	SnOx/SnS	NO2	171%	0.005ppb	1800	36	[58]
	2D SnS2	NH3	4.2	500ppm	16	450	[59]
	MoS2 aerogel	NO2	~22.5%	28ppb	33	107	[60]
	WS2/G	NO2	23%	50ppb	110	168	[61]
Doping engineering	B-graphene	NO2	31.5	1ppb	-	-	[62]
	NO2-graphene	NH3	25.3%	200ppb	-	-	[63]
	W- MoS2	NO2	171%	20ppm	20	19	[64]
Modification with noble metal	Au-WS2NSS	CO	1.061	0.2ppm	~50	~400	[12]
	Cd-Co3O4	NO2	3.38	154ppb	46	620	[65]
	Pd-WSe2	CO	9.25%	5ppm	52	97	[66]
	MoS2-Pt	H2	10	100ppm	4	19	[67]
	Au/MoS2	NO2	220.4%	50ppm	-	-	[68]
	Pt-SWCNTs	NO2	63%	2ppm	1000	-	[69]
	Ag-rGO	NO2	8.6%	6.9ppb	75	89.5	[70]
	SnO2/Au	ethanol	131	-	13	19	[71]
	SnO2/Pd	ethanol	91	-	15	40	[71]
	Au-ZnO nanoflakes	n-butylamine	73.2	23.9ppm	23–57	8–76	[72]
	a-Fe2O3-Au	acetone	113	10ppm	7	7	[73]
Heterostructure construction	WSe2/WS2	NO2	322%	10ppm	387	728	[74]
	WSe2/WS2	NH3	−23%	10ppm	-	-	[74]
	MoS2/WSe2	NO2	−36%	50ppm	-	-	[74]
	MoS2/WSe2	NH3	165%	10ppm	205	263	[74]
	MoS2/SnO2	NO2	~30%	0.5ppm	408	162	[75]
	ZnO-rGO	NO2	119%	50ppb	75	132	[76]
	MoS2/PbS	NO2	22.9%	5ppm	30	235	[77]
Light illumination	WS2	NH3	3.4	10ppm	252	648	[54]
	InSe	NO2	402%	0.98ppb	233	350	[78]
	ReS2	NO2	9.07	50ppb	55	180	[79]
	MoS2	NO2	35.16%	5ppm	29	350	[80]
	MoTe2	NO2	1300%	252ppt	300	160	[81]
	MoTe2	acetone	55%	200ppb	-	-	[82]
	ReS2	acetone	104.8%	-	-	-	[83]
	g-C3N4/rGO	NO2	0.65	100ppb	138	318	[84]
	g-C3N4/rGO	SO2	0.0032	2ppm	140	130	[84]
	ZnO/g-C3H4	NO2	44.8	38ppb	142	190	[85]
	ZnO/SnSe2	CO	14.8%	200ppm	19	13	[28]

4.1 Defect engineering

During the preparation and synthesis of 2D material, various defects such as vacancies, impurity atoms, and adsorbates often occur due to deviation from stoichiometric ratio. However, these defects are typically considered the main active sites for gas adsorption on the surface of 2D material [56,86]. Therefore, appropriately introducing defects will improve the adsorption of gases, thereby enhancing the performance of gas sensing. Vacancies are a common defect in 2D materials, the presence of which could change the electronic, optical, and chemical activity of two-dimensional materials [87,88]. Moreover, owing to the presence of vacancies, gas molecules are very prone to adsorption at these vacancies, resulting in a corresponding increase in the adsorption of gas molecules on the surface of gas sensing materials [57,58]. Therefore, two-dimensional materials with rich vacancies commonly exhibit enhancements in sensitivity and response/recovery speed.

Qin et al. synthesized 2D SnS₂ with rich sulfur vacancies for NH₃ gas sensing [59]. Figure 5(a) demonstrated the preparation process, the prepared 2D SnS₂ was achieved through lithium intercalation, water sonication, vacuum-filtration assembly and chemical exfoliation. The comparation of XPS spectra of bulk and exfoliated SnS₂ obviously reveal that the sulfur vacancies greatly increased with the chemical exfoliation process, as the new peak at 161.2 eV of S 2p region in exfoliated SnS₂ indicated the sulfur vacancies (Figure 5(c,d)). Owing to the high activity and catalysis characteristics, the sulfur vacancies serve as the major gas adsorption centers. As a result, the 2D SnS₂ gas sensor exhibited excellent performances for NH₃ as shown in Figure 5(b,e). The response achieved ~ 2.3 at 100 ppm and the response time only 16 s at 500 ppm. Long et al. prepared a NO₂ gas sensor based on three-dimensional molybdenum disulfide aerogel synthesized using thermal decomposition technology and a two-step vulcanization process [60]. Due to the annealing treatment during the sulfidation process, more sulfur vacancies can be produced in the molybdenum disulfide aerogel. Therefore, the molybdenum disulfide aerogel not only exhibits high sensitivity to NO₂ gas at room temperature, but also achieves fast response and recovery at 200°C. Ma et al. fabricated gas sensors on the WS₂/G heterostructure through the defect engineering controlled by Ga⁺ ion irradiation [61]. The Ga⁺ ion irradiation transformed partial region of WS₂/G to WS_2–0.2/G, changing the electronic properties of the heterostructure like the work function. As a result, the defect-engineered heterostructure showed an excellent gas-sensing response to NO₂ gas with the limit of detection (LOD) of 50 ppb and a fast response time of 110 s.

Figure 5. (a) illustration of the preparation process of 2D SnS₂. (b) response-recovery curves of 2D SnS₂ for NH₃ sensing in the concentration range from 20 to 800 ppm. (c)-(d) XPS spectrum of the bulk and exfoliated SnS₂. (e) the response as a function of NH₃ concentration from 20 to 800 ppm [59].

4.2 Doping engineering

Doping engineering has been successfully used to regulate the carrier concentrations and carrier type. Effectively regulating the electrical characteristics of gas sensors is an approach to achieve excellent sensing performance, such as sensitivity, response time and selectivity. Proper doping treatment of materials such as ion implantation [89], surface charge transfer through organic or inorganic species [90], and alloys and compounds synthesized by chemical vapor deposition (CVD) [91] may improve electrical properties and adsorption capacity.

In order to experimentally confirm that the gas-sensing performance of graphene could be effectively enhanced by doping with dopants, Lv et al. synthesized the p-type boron-doped graphene (BG) using a liquid precursor with a unique croissant-like feature in lattice [62]. More interestingly, the BG exhibited enhanced sensing capabilities to toxic gas such as NO₂ (Figure 6(a)) and NH₃ (Figure 6(b)). Compared with pristine graphene, the sensitivity of BG for NO₂ and NH₃ were enhanced by 27 and 105 times, respectively. Similarly, Zanjani et al. demonstrated the enhanced sensitivity of p-doped graphene using NO₂ molecules doping [63]. The increased hole concentration of p-doped graphene increased the binding affinity and charge transfer amount for electron-donating NH₃ molecules. Thus, the detection limit of NO₂-doped graphene gas sensor was improved by 7 times from ~ 1400 ppb to ~ 200 ppb, and the sensitivity was enhanced by one order of magnitude for NH₃ detection, as shown in Figure 6(c,d).

Figure 6. (a) sensor response of B-doped graphene to NO₂. (b) sensor response of B-doped graphene to NH₃ [62]. (c) sensitivity of un-doped graphene and NO₂-doped graphene sensors to NH₃ gas. (d) sensitivity as a function of NH₃ concentration [63].

Doping with foreign atoms (heteroatoms) (nitrogen, sulfur, boron, silicon etc.) can adjust the defect concentration in materials. Liu et al. reported the doped MoS₂ gas sensors with different proportions of metallic atoms [64]. They doped MoS₂ with metal atoms W at ratios of Mo: W = 1:0, 1:1, 1:2, and 1:3, and found that the gas sensor had the highest sensitivity (56.91%) and fastest response/recovery time (24 s and 19 s) when the ratio was Mo: W = 1:2. This excellent performance was mainly due to the suppression of excessive defects in MoS₂ by doping appropriate proportion of W atoms.

4.3 Modification with noble metal

The modification of gas sensing materials with noble metals including Au [12], Cd [65], Pd [66], Pt [67], and so on is a common approach to optimize gas sensing performance, mainly through two enhancement mechanisms. One is that the formation of the Schottky barrier between the noble metal and 2D materials modulates the electrical properties of the gas sensor, therefore improving the performance. Another mechanism is that after the target gas molecules or O₂ molecules in the air adsorb and accumulate on the surface of the precious metal material, the modified noble metal acts as new active sites, which increases the adsorption capacity of the gas sensing material, thereby improving the sensing performance of 2D material.

The modification with noble metal has been successfully implemented to improve the sensitivity of various sensing materials, such as carbon nanotubes, graphene, MoS₂ and so on [68,69,92,93]. Li et al. doped Ag nanoparticles with different concentrations on reduced graphene oxide films, which reduced the response time to NO₂ gas [70]. When the doping ratio of Ag nanoparticles was 28.4 wt%, the composite material had the best gas sensing performance, as the inherent catalytic activity of nanostructured Ag and the modification of Ag providing more active sites for graphene. Kim et al. decorated three kinds of noble metal nanoparticles on MoS₂ and the MoS₂ flakes exhibited drastic selectivity changes toward different gases [93]. After modifications with Pd, Au and Pt (Figure 7(a)), the Au-MoS₂ flake showed clear selectivity to C₂H₅OH, the Pd-MoS₂ sensors showed enhanced responses to H₂, and the Pt-MoS₂ exhibited both increased responses to NH₃ and H₂, as demonstrated in Figure 7(b). As a result, the sensor array consisting of the four MoS₂-based gas sensors was ability to identify various gas species, since the responses of the four gas sensors changed differently after decorating with noble metal. The obvious separated data points of NO₂, acetone, ethanol and H₂S gases in Figure 7(c) proved the drastic gas sensing selectivity. Suh et al. studied the effect of Pd and Au modification on the gas sensing performance of CVD-synthesized MoS₂ thin films, through detailed comparisons of the gas sensing performance of intrinsic MoS₂, Pd/MoS₂, and Au/MoS₂ [68]. Compared to intrinsic MoS₂ material, Au-modified Au/MoS₂ exhibited superior gas sensing response at different operating temperatures and achieved the highest response, approximately 220.4% at the optimal temperature of 70°C as shown in Figure 7(d), which is attributed to the changes in the electronic structure of the composite material after Au modification. In addition, compared with intrinsic MoS₂, Pd/MoS₂ showed a reduced gas response to NO₂ at different operating temperatures but a higher response to H₂ in Figure 7(e). The detailed mechanism was demonstrated in Figure 7(f) that Pd hydride will form at high temperatures due to the catalytic activity of Pd for H₂ gas. In addition, Kaneti et al. carried out a series experimental and theoretical studies on semiconductor metal oxide nanoflakes decorating with metal nanoparticles [72]. Compared with the pure ZnO nanoflakes, the Au nanoparticles-decorated ZnO nanoflakes show a greatly enhanced selectivity at low operating temperature and 4–6 times higher response to n-butylamine. Through characterization and DFT simulations, they attributed this feature to high catalytic activity of Au NPs and the additional electron depletion layer at the ZnO/Au interface, thus promoting a higher change in resistance. Overall, these findings will be helpful to design and modify gas sensors with noble metal for achieving higher performance.

Figure 7. (a) SEM images of p-MoS₂, Pd-MoS₂, Au-MoS₂, and Pt-MoS₂. (b) polar plot of response of p-MoS₂, Pd-MoS₂, Au-MoS₂, and Pt-MoS₂. (c) PCA result for 4 different types of gas in 4 different MoS₂-based arrays. (d) gas sensing response to 50 ppm NO₂ for MoS₂ (green curve), Pd/MoS₂ (red curve) and Au/MoS₂ (orange curve) at 70°C.(e) gas sensing response to C₂H₅OH, H₂, NH₃ and NO₂ as a function of temperature. (f) schematic diagram of formation mechanisms of Pd hydride [68].

4.4 Heterostructure construction

By constructing two materials and forming a heterojunction at the interface, the formation of the heterojunction can be used to enhance gas sensing performance. According to the conducting types of the materials, either n-n, p-p, or p-n type heterojunctions will be formed [94,95]. Furthermore, the difference in work function of materials at the heterojunction interface will drive carrier transfer, which lead to an improvement in the electronic structure of the heterostructure and thereby enhance gas sensing performance [42,74].

To address the issue of the instability of the original MoS₂ nanosheets caused by exposure to oxygen in the air, Cui et al. proposed a solution based on constructing heterostructure to enhance the stability of MoS₂ nanosheets [75]. They prepared a SnO₂/MoS₂ gas sensor for NO₂ detecting using the solution method, and the heterostructure sensor showed high sensitivity, excellent selectivity, and good reproducibility to NO₂ in dry air. At the same time, due to the larger work function of SnO₂ compared to MoS₂, electrons transferred from MoS₂ to SnO₂, greatly suppressing the reactivity of MoS₂ with oxygen and enhancing the stability of the sensor. Xia et al. improved the gas sensing performance of NO₂ at room temperature by forming a heterostructure with ZnO and graphene oxide [76]. The heterostructure exhibited higher sensitivity up to 119% (Figure 8(a)) and quicker response time of 75 s to 1 ppm of NO₂ (Figure 8(b)). The improvement was due to the rapidly transferring electrons of graphene in the heterostructure as illustrated in Figure 8(c), accelerating charge transfer between gas molecules and the heterostructure.

Figure 8. (a) dynamic response curves of ZnO/rGO heterostructure to 1.0–10.0 ppm of NO₂ at room temperature. (b) dynamic response curves of rGO, ZNR-3, and ZNR-3/rGO to 1 ppm of NO₂. (c) schematic illustration of the energy band structures of ZnO-rGO junction and electron transfer in the heterostructure [76]. (d) dynamic response−recovery curves of MoS₂ and MoS₂/PbS gas sensors at different NO₂ concentrations. The inset shows typical I−V curves of MoS₂ and MoS₂/PbS gas sensors in air and 100 ppm NO₂. (e) schematic diagrams of the gas-sensing mechanism for MoS₂/PbS gas sensors [77].

In addition to forming heterostructures with metal oxides, combining with other two-dimensional materials can also improve gas sensing performance. Xin et al. deposited PbS quantum dots on MoS₂ to construct the MoS₂/PbS heterostructure [77]. Due to the difference in work function values between MoS₂ and PbS, electrons accumulate on the PbS, resulting in a sufficient number of electrons on its surface (Figure 8(e)), which is beneficial for the material to adsorb more NO₂ molecules. Therefore, the sensor based on the MoS₂/PbS heterostructure exhibits high sensitivity to NO₂ molecules. In addition, the small-sized PbS material can form many microdomains on the MoS₂ with large specific surface area, providing sufficient electron transfer channels inside the heterostructure. Compared to the gas sensor prepared from the intrinsic MoS₂, the MoS₂/PbS gas sensor exhibits a 50-fold increase in sensitivity to NO₂ at 100 ppm as shown in Figure 8(d).

4.5 Light illumination

External energy such as light illumination is also an effective approach to enhance sensing sensitivity and reduce recovery time. This is mainly based on two reasons [54,78,79]: on the one hand, due to the inherent light absorption ability of 2D materials, when light with appropriate wavelength irradiate, electron–hole pairs will be generated in the two-dimensional material. These photo-generated carriers can not only change the carrier concentration, which is beneficial for improving the conductivity, but also participate in the reaction with adsorbed gas molecules, enhancing the interaction between the material and gas molecules. In addition, light illumination may reduce the activation energy of adsorption/desorption reactions on the material surface, accelerate the reaction rate of gas molecule, and achieve rapid response and recovery.

Kumar et al. reported a fast and reversible NO₂ gas sensor based on MoS₂, the significant improvement in sensitivity (∼30%) and response time (∼88%) were achieved under ultraviolet light excitation (Figure 9(a-c)). Wu et al. fabricated a p-type MoTe₂ gas sensor through rapid annealing for NO₂ gas sensing (Figure 9(d)), which exhibited greatly enhanced sensitivity and recovery rate under ultraviolet illumination [81]. Under 254 nm illumination, the sensitivity of MoTe₂ gas sensor was enhanced by 1 order of magnitude (1300% to NO₂ at 1 ppm) as compared to that in the dark condition and achieved a remarkable low detection limit of 252 ppt (Figure 9(e)). As shown in Figure 9(f), MoTe₂ exhibited a strong optical absorption around the 254 nm, which was corresponding to π-electron plasmon excitation. The photoexcited plasmons may accelerate the reactions between gas molecule and sensing channel by attaching to the adsorbed molecules on the material and promote desorption. Therefore, the MoTe₂ sensor recovered faster and exhibited fully reversible under UV illumination than in darkness.

Figure 9. (a) schematic illustration of MoS₂ gas sensor under UV illustration. (b) cyclic test to 100 ppm of NO₂ under UV light. (c) response time calculation to 100 ppm of NO₂ under UV [80]. (d) schematic diagram of MoTe₂ gas sensor for NO₂. (e) comparation of sensing curves of NO₂ in the dark and under UV light at 20 ppb. (f) optical absorbance spectrum of MoTe₂ flake on a SiO₂ substrate. Inset: diagram of UV-induced electron−hole pairs of MoTe₂ [81].

Besides improving sensitivity and response/recovery rates, ultraviolet light illumination can also enhance the selectivity of 2D materials-based gas sensors [83–85]. Wu et al. reported the specific and highly sensitive detection of ketone compounds based on MoTe₂ gas sensors under ultraviolet illumination [82]. The device exhibited an opposite sensing response to ketone compounds with and without ultraviolet illumination, whereas the responses to other types of volatile organic compounds like ethanol, IPA and methanol remain in the same direction regardless of the ultraviolet activation (Figure 10(a,b)). It is found that the polarity of ketone molecules undergoes a reversal under ultraviolet light irradiation, thus causing the opposite sensing response, enabling specific recognition of ketone. In addition, the sensing response enhanced by roughly 1.5–2 orders of magnitude under ultraviolet illumination comparing with that in darkness (Figure 10(c)), owing to the desorption of impurity molecules by ultraviolet. Zulkefli et al demonstrated that the selectivity of volatile organic compounds (VOCs) based on ReS₂ can be enhanced through light illumination and gate-bias (Figure 10(d,e)) [83]. The sensing characteristics of each gas species were different under light illumination; thus, the device successfully distinguished the gas concentration in a mixture of VOCs. This discovery confirms that the 2D materials-based gas sensors are promising candidates as sensors in important applications such as human breath analysis.

Figure 10. (a) sensing response of the MoTe₂ FET toward seven different VOC gases at 100 ppm in darkness. (b) repeated measurements in (a) under UV light. (c) dynamic response of the MoTe₂ FET at room temperature toward acetone from 100 to 2000 ppm [82]. (d) schematic illustration of the gas-sensing measurement setup based on ReS₂-FET. (e) selective detection of acetone vapor in a mixed gas [83].

5. Conclusion and outlook

In this article, we have summarized the latest developments in two-dimensional material-based gas sensors, including gas sensor structures, gas sensing mechanisms (charge transfer, the Schottky barrier modulation and surface adsorption of oxygen ions), and effective strategies for enhancing the gas sensing performances. 2D materials provide unique capabilities such as huge surface-to-volume ratio and numerous active sites for gas sensing beyond what can be easily achieved by conventional semiconductors. Significant advantages including single gas molecule detection have been achieved in 2D materials-based gas sensors. However, since gas sensing processes rely on the interactions between gas molecules and materials, a profound understanding of the related sensing processes including structural design and optimization, sensing mechanism analysis, information acquisition and intelligent processing is crucial to tailor gas sensors with the desired properties for different applications while maximizing the benefits which 2D materials can offer.

One of the challenges and largely unexplored topics of 2D materials-based gas sensors is the gas sensing mechanisms. Understanding the sensing mechanism in depth will give new approaches which can offer higher sensitivity, better selectivity, and faster response/recovery rates. The most used gas sensing explanation is direct charge transfer between gas molecules and 2D materials. However, the gas sensing mechanism is relatively complex, involving processes such as gas molecule diffusion, physical/chemical adsorption and desorption, carrier transport and recombination. Besides, the gas sensing mechanism is mostly characterized by testing the performance of gas sensors and obtaining qualitative descriptions of the shallow mechanism currently, lacking in-depth discussions of the sensing mechanism. Although reasonable models have been established to elucidate the sensing mechanism, the design of novel comparative experiments and detailed theoretical calculations and simulations are still needed to verify and refine the gas sensing mechanism.

Another obstacle to overcome in 2D materials-based gas sensors is that the practical application ability of sensors needs to be further verified. Despite the significant progresses have been made in the field of 2D material gas sensing, most current research of two-dimensional materials has remained at the laboratory stage, using synthetic air/nitrogen or laboratory-generated gases as a blank background and large-scale instruments such as semiconductor analyzer to facilitate the exploration of gas sensing properties. However, the gas background in practical application environments is quite complex. For example, human breath samples contain not only various volatile organic compounds that need to be detected, but also many other substances such as water molecules, which may interfere with the sensitivity and accuracy of the sensors. Besides, the long-term stability and environmental impact of sensors cannot be ignored in practical use. The temperature, humidity, or other gas molecules of the environment may impede the sensing performances for most gas sensors, resulting in poor accuracy. Gas sensing tests of actual samples rather than gas mixtures synthesized in the laboratory are needed to further improve the anti-interference ability of 2D materials-based gas sensors.

To achieve commercialization and reproducibility of gas sensors, optimization of the manufacturing process is also one of the future research directions, which involves material preparation, electrode processing and integration with CMOS circuits. Considering that thickness variations could lead to differences in sensing performance, the fabrication methods of 2D materials should be adjusted to control the thickness precisely, such as chemical vapor deposition, inkjet printing, mechanical exfoliation, etc. In terms of electrode processing, it may be possible to introduce the new mechanical electrode transfer technology that does not require a high-temperature heating process, avoiding interface damage between the sensing material and the metal. However, further refinement is still needed to achieve wafer-scale transfer and high-precision alignment. Owing to the compatibility between 2D material device fabrication and CMOS technology, gas sensors based on 2D materials can be easily integrated with CMOS circuits. More efforts should be devoted to the development of flexible, wearable, and miniaturized gas sensor devices.

The rapid development of gas sensors is highly anticipated to represent solutions to the challenges our society is currently facing. 2D materials has proven to be an exciting and promising materials for physicists, material scientists and engineers. With continuous efforts devoted to this research field, these materials will be implemented in practice soon.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

The work was supported by the National Natural Science Foundation of China under Grant [number 62304151, 62204170, 52107203]; the Natural Science Foundation of Tianjin under Grant [number 22JCQNJC01010]; and the Open Project of State Key Laboratory of Transducer Technology under Grant [number SKT2208].