Enhanced Sensitivity of MoTe₂ Chemical Sensor through Light Illumination

Abstract

Two-dimensional (2D) transition metal dichalcogenides (TMDCs) semiconducting materials have recently attracted wide attention and been regarded as promising building blocks for chemical sensors due to their high surface-to-volume ratio. However, their low response hinders the realization of high-performance 2D TMDCs chemical sensors. Here, we demonstrate the improvement of sensing performance of molybdenum ditelluride (MoTe₂) gas sensor through continuous light illumination. The dependence of sensing performance on the energy of photons and light intensity is systematically studied. The response to NH₃ is dramatically enhanced by more than 25 times under 254 nm ultraviolet (UV) light illumination with intensity of 2.5 mW/cm². Moreover, a remarkable low detection limit of 3 ppb is achieved, which is improved by 80 times compared with that in dark condition. The results demonstrate that light illumination is a promising method to improve the sensitivity of 2D TMDCs chemical sensors.

1. Introduction

Detecting ultra-low concentration of toxic and harmful gases has become more and more important in many promising applications such as environmental monitoring, industrial manufacturing and diagnostic processes, etc. [1]. In particular, semiconductor chemical sensors have attracted considerable attention owing to their low cost, ease of integration and simple sensing mechanism based on the conductance change upon gas exposure [2]. Over the past few decades, metal oxide semiconductor sensors have delivered satisfactory sensitivity in many cases [3,4,5], however, they require very high temperature (generally above 200 °C) to activate the chemisorption of atmospheric oxygen on the metal oxide surface, which raises several thermal-related issues and restricts their applications in oxygen-free environments or hazardous environments containing flammable gas species. Therefore, highly-sensitive chemical sensors which can operate at room temperature are highly desired. Recently, semiconducting two-dimensional (2D) transition metal dichalcogenides (TMDCs) have been widely studied and considered as promising gas-sensing materials due to their high surface-to-volume ratio and favorable surface energy level for gas adsorption [6,7,8,9,10]. Besides, researchers have also developed various approaches, for instance, surface functionalization [11] and using vertically aligned structure [12] to further enhance the sensitivity of 2D TMDCs based chemical sensors. However, the results show limited improvement, which is usually less than 10 times. Illuminating sensors with light is another effective way to dramatically improve the performance of semiconductor chemical sensors [13,14,15]. However, few have systematically investigated the influence of light illumination on the performance of 2D TMDCs based sensors.

Molybdenum ditelluride (MoTe₂), a new addition to the class of 2D TMDCs, has a smaller band gap of ~1.0 eV compared with other TMDCs, which extends the photo detection range of TMDCs based phototransistor from the visible to near-infrared range [16]. Recent experimental investigations show that MoTe₂ photodetector has a high photoresponsivity and a fast photoresponse in broad spectral range [17,18]. In the view of growing attention on MoTe₂ for possible optoelectronic device applications, it is also important to investigate the influence of light illumination on the sensing performance of MoTe₂ based chemical sensors.

In this work, we report the improvement of gas-sensing properties of MoTe₂ chemical sensor through light illumination. The sensitivity of MoTe₂ sensor for NH₃ detection was enhanced by more than 25 times and the detection limit as low as 3 ppb was achieved under continuous ultraviolet (UV) light illumination. The influences of wavelength and intensity of light on the improvement of the sensing properties have also been systematically studied.

2. Materials and Methods

Few-layered MoTe₂ flakes were prepared by mechanical exfoliation of a MoTe₂ bulk crystal on n+ doped silicon substrate covered with 285 nm of thermally grown SiO₂. The tape residues after exfoliation were removed by dipping the substrate along with flakes in acetone for 10 min. The electrical contacts on MoTe₂ were fabricated by lift-off process using electron beam lithography and subsequent electron beam evaporation of 20/60 nm Ti/Au. Figure 1a shows the optical microscope image of a typical fabricated MoTe₂ field-effect transistor (FET) device. Thickness of the MoTe₂ flake is 3.4 nm, as confirmed by the atomic force microscopy (AFM, Bruker, Santa Barbara, CA, USA) measurement shown in Figure 1b. MoTe₂ has two stable crystal phases: semiconducting 2H (hexagonal) and metallic 1T′ (distorted octahedral) phases [19]. Density functional theory (DFT) calculations revealed that molecular adsorption favors semiconducting 2H phase other than 1T′ metallic phase [20]. So it is important to determine the exact crystal structure of the MoTe₂ flakes we used. Figure 1c shows the Raman spectra (532 nm excitation wavelength) for the exfoliated few-layered MoTe₂. The characteristic Raman-active modes of the out-of-plane A_1g mode at 170 cm⁻¹, prominent in-plane E¹_2g mode at 231 cm⁻¹ and the out-of-plane B¹_2g mode at 287 cm⁻¹ are clearly observed. These Raman features are signature peak positions of 2H phase MoTe₂ and in good agreement with the results reported previously [19]. The crystallographic structure of the MoTe₂ was further characterized by high-resolution transmission electron microscopy (TEM, JEOL, Tokyo, Japan). Figure 1d shows a typical TEM image of as-exfoliated MoTe₂ film. We could clearly observe a high level of crystallinity and hexagonal symmetry. X-ray photoelectron spectroscopy (XPS, PerkinElmer, Waltham, MA, USA) was utilized to determine the elemental composition and bonding types of the MoTe₂ flakes. The XPS spectrum displayed in Figure 1e reveals the presence of Mo and Te elements in the flakes. Peaks are observed at 228.4, 231.5, 573 and 583.4 eV, corresponding to the Mo 3d_5/2, Mo 3d_3/2, Te 3d_5/2 and Te 3d_3/2, respectively, as shown in Figure 1f. These features are consistent with previous results from 2H MoTe₂ [16].

3. Results and Discussion

We then investigate the effect of UV light illumination on the electrical performance of the MoTe₂ FET device in pure N₂ (99.9999%) environment. The wavelength of UV light source is 254 nm and the intensity is 2.5 mW/cm². Figure 2a shows the transfer characteristics of the MoTe₂ FET device in both linear and semi-log scales under dark condition. The device shows ambipolar behavior with a minimum current around V_gs = −18 V. The I_ds increases dramatically by 2 × 10³ when V_gs increases from −18 V to 30 V, indicating electron accumulation in the MoTe₂ channel. When V_gs decreases from −18 V to −30 V, the I_ds is dominated by holes, as evidenced by increasing I_ds. The observation is also consistent with previous studies [21]. After UV light illumination for 2 h, we note that the p-type behavior vanishes and the MoTe₂ FET device behaves as n-type transistor as evidenced in the I_ds–V_gs curves in Figure 2b. When V_gs increases from −30 to 30 V, I_ds increases monotonically by more than 3 orders of magnitude due to accumulation of electrons in the conduction channel. The vanished p-type behavior of the MoTe₂ FET under UV light illumination can be attributed to the removal of O₂ molecules from MoTe₂ flake. This can also be confirmed by the reversible electrical characteristics of one MoTe₂ FET (Figure S1). The oxygen molecules adsorbed on the surface of MoTe₂ flake may act as acceptors which induce p-type doping [22].

We then carried out systematic sensing tests of the MoTe₂ FET toward NH₃ in both dark and light illumination conditions to characterize its sensing performance at room temperature. Prior to the sensing experiments, the gas chamber and device were purged with N₂ for 2 h under UV light illumination. N₂ was used as the carrying gas to dilute the target gases to desired concentrations through flow rate adjustments. The gate electrode is grounded throughout all measurements.

Figure 3a presents the real time response of the MoTe₂ sensor toward exposure of NH₃ from 300 ppb to 30 ppm under both dark and light illumination conditions. The light intensity was kept at 2.5 mW/cm². The device conductance increases monotonically with increasing NH₃ concentrations for all conditions as NH₃ molecules adsorption increase the electron concentration in the n-type semiconducting channel. The response is defined as (G − G₀)/G₀ × 100%, where G₀ and G are the initial conductance and 5 min after NH₃ exposure, respectively. We note that the response of the MoTe₂ sensor is enhanced remarkably under light illumination from near-infrared to UV region. For example, the device shows responses ranging from ~4 to 25% under dark condition, whereas it presents larger responses ranging from ~100 to 790% under 254 nm UV light illumination at NH₃ concentrations of 300 ppb–30 ppm. Therefore, the response is enhanced by more than 25 times by UV light illumination. The response values obtained under UV light illumination are well above the results from 2D TMDCs chemical sensors for NH₃ detection. For comparison, the study in [9] and [23] demonstrated sensitivity of MoSe₂ and WS₂ sensors from a few to tens of ppm concentrations. It is noteworthy that for the MoS₂ sensor with a lower detection limit down to 300 ppb, the response is ~0.04% [24], which is 3 orders lower than the MoTe₂ sensor. By comparing the data points at different concentrations of NH₃ under light illumination with different wavelength, we also note that the sensitivity enhancement increases monotonically with reducing wavelength (increasing photon energy), as shown in Figure 3b. The sensitivity improvement is 2–6-fold in the near-infrared to red region, while it increases to more than 10 times in the UV region. The largest sensitivity enhancement is achieved under 254 nm UV light illumination. The enhanced sensitivity under light illumination can be attributed to the following reasons: The presence of defects in MoTe₂ during the crystal growth process can act as active sites. Density functional theory calculations revealed that the presence of defects strongly increases the binding energy of the molecular interaction on MoTe₂ surface [25,26]. For example, interaction at the defect sites, such as Te vacancies, yields much larger binding energy between MoTe₂ and O₂ (~166 meV) compared to the case of pure MoTe₂ (~36–64 meV) [25]. The O₂ molecules initially adsorbed on the surface of MoTe₂ may occupy the defect sites. Therefore, removal of O₂ molecules could significantly enhance the sensitivity. However, the thermal energy at room temperature could not desorb O₂ molecules [27,28]. In this respect, UV illumination can provide the energy for O₂ desorption through several mechanisms [29,30]. The study in Reference 29 suggested that photo induced desorption strongly depends on the wavelength of light. UV light illumination has the highest efficiency for “cleaning” the nanomaterial surface owing to electron plasmon excitation. Therefore, the best sensitivity was achieved under UV light illumination. By comparing the response time and recovery rate in dark and light illumination conditions (Figure S2), we also note that the response time is not changed too much while the recovery rate is slower under light illumination than that in dark. This can be attributed to the following reason: NH₃ molecules are mainly adsorbed on the pristine surface of MoTe₂ due to occupancy of active sites by O₂ molecules in dark condition. Under light illumination, the active sites are released and the binding energy between NH₃ and active sites are much higher than the pristine case. On the other hand, we also note that the recovery rate is faster in the UV region than in the near-infrared to red region. Under UV light illumination, the desorption rate of NH₃ from active sites is increased due to electron plasmon excitation [29] so that a dynamic equilibrium is established between adsorption and desorption processes.

Figure 3c plots the response as a function of NH₃ concentration in the dark and under light illumination. The response is linear at low concentrations and shows a quasi-saturation behavior as the concentration increases. The linear portion of the curve can be then used to estimate the detection limit, which is typically defined as the concentration level corresponding to a signal-to-noise ratio (SNR) of 3 [31]. Under 254 nm UV light illumination, the detection limit is estimated to be ~3 ppb by comparing the noise level (0.17%) of the sensor to the slope (196% ppm⁻¹) of the response at low concentration of NH₃. This number is 80 times superior to that in the dark, which is estimated to be ~243 ppb according to the slope of the response (Figure 3d). We also note the detection limit of our MoTe₂ sensor under 254 nm UV light illumination is the record number using 2D TMDCs so far [6,7,9,23,24,32,33] and competitive with other nano-sensors [34,35,36].

Figure 4a shows the dynamic response of the MoTe₂ sensor to different NH₃ concentrations (300 ppb–30 ppm) under UV light illumination (254 nm) with intensity ranging from 0.25 to 2.5 mW/cm². We note that the sensing properties of the MoTe₂ sensor behave similar but the responses are strongly influenced by the light intensity. For example, the response to 30 ppm NH₃ is ~290% at intensity of 0.25 mW/cm², while it increases up to ~790% at intensity of 2.5 mW/cm². Figure 4b presents the dependence of the response on the UV light intensity. The response increases rapidly when the intensity increases from 0.25 to 1 mW/cm² and then shows slow saturation trend as the intensity increases up to 2.5 mW/cm², the highest intensity we could experimentally achieve with the setup. Generally, the light intensity should have an optimal value for maximum sensitivity as the desorption rate will increase more than the adsorption rate if the intensity is too high [37]; however, we did not observe this phenomena due to the limited power density of our light source. The detection limit is also improved with higher light intensity. The slopes of the response plotted in Figure 4c are 196% ppm⁻¹ and 90.4% ppm⁻¹ at intensity of 2.5 and 0.25 mW/cm², respectively, from which we can obtain a detection limit of ~3 and 6 ppb, respectively. Therefore, the detection limit is improved by 2-fold as the intensity increases by 10 times.

4. Conclusions

In summary, we report a systematic study on the sensitivity enhancement of MoTe₂ gas sensor through light illumination. Compared with dark conditions, the sensitivity is dramatically improved by light illumination and monotonically dependent on the photon energy. With continuous 254 nm UV light illumination, the response to NH₃ gas is enhanced by more than 25 times and a detection limit as low as 3 ppb is achieved. It is also found that the enhanced sensitivity is dependent on the light intensity. The response increases from 290 to 790% at 30 ppm NH₃ when increasing the UV light intensity from 0.25 to 2.5 mW/cm². Our results prove a strong influence of light illumination on the sensitivity of MoTe₂ sensor and demonstrate the potential of 2D MoTe₂ as a promising candidate for ultrasensitive chemical sensing applications.