1. Introduction

Silicon is a highly significant semiconductor material due to its low-cost, nontoxicity, CMOS compatible, and high temperature stability [1]. The microstructured and hyperdoped silicon, also known as black silicon (bSi), fabricated by a femtosecond-pulsed (fs) laser, is a surface-modified silicon material. The microstructures enhance geometric light trapping through multiple reflections, while hyperdoped impurities like S [2,3], P [4], Ag [5], and Au [6] introduce intermediate states within the forbidden band of silicon extending light absorption into the near-infrared band. Therefore, it appears black to the naked eye, differing from the gray color of normal silicon, hence its name. Due to its ultra-low reflectivity and outstanding optical and electrical properties, bSi finds applications in solar cells [7,8], photodetection [9,10], etc.

Apart from the aforementioned applications, bSi exhibits potential as a gas sensing material due to its favorable surface microstructure for gas adsorption and unique photoelectric properties which are conducive to the conversion of surface information [11–15]. In our previous work, a simple bSi device with two co-planar electrodes was fabricated. Under an asymmetric light illumination between the electrodes, its ammonia sensing capability was demonstrated and its gas sensing properties were examined methodically [11]. Furthermore, the research showed that the adopted asymmetric illumination had the potential to significantly improve the gas responsivity compared to conventional gas-sensitive resistors. Subsequently, nitrogen dioxide (NO₂) gas sensing based on different ratios of sulfur and nitrogen co-hyperdoped bSi was also realized by asymmetric illumination [15]. All of these devices utilize photoelectrons or photocurrents generated by the lateral photovoltaic effect (LPE) to detect gases.

The scheme of the aforementioned gas sensing with asymmetric illumination is actually similar to the one used when the bSi device is regarded as position sensitive detector (PSD) to detect light. In this case, the asymmetric or non-uniform illumination originates from the light spot to be detected between the two co-planar electrodes. Through this illumination way, the PSD can measure the position, distance, and other related physical variables of light based on the LPE [16]. The principle of the LPE is that, when a semiconductor is under non-uniform illumination, charge carriers/electron-hole pairs are generated only in the illuminated region, creating a carrier concentration gradient between the illuminated and non-illuminated regions. This gradient allows for the diffusion of carriers towards the electrodes, thereby forming a photocurrent. In light detection, this photocurrent serves as a response signal to changes in light position, intensity, wavelength, and so forth. In gas sensing, the photocurrent acts as a response signal to variations in ambient gas. This means that the bSi device, when operated under the asymmetric illumination scheme, not only functions as a gas sensor but also as a photodetector. In other words, the photocurrent not only responds to changes in the gas environment, such as variations in gas type and concentration, but also to changes in the light environment, such as the intensity and wavelength of asymmetric irradiation light. Therefore, an intriguing question is how the light environment and gas environment individually affect the sensitivity of the device to gas and its photodetection performance.

In this study, three co-hyperdoped bSi devices with varying co-doping ratios of sulfur and nitrogen were fabricated. Under the asymmetric illumination scheme, the devices are considered for the first time as gas and light dual-function detectors, and their gas sensing and photo-detecting performance under different light and gas environments are investigated separately. The goal is to optimize the dual-function detectors’ performance in terms of gas sensing and photo-detecting. The results indicate that an appropriate light environment can significantly optimize the detection of NO₂ gas. Conversely, a specific NO₂ gas environment can greatly enhance the photo-detecting performance. In other words, for this light and gas dual-function detector, the appropriate light and gas environments have a mutually reinforcing effect on each other's detection performance. These findings not only provide valuable insights for optimizing gas sensor performance but also unveil an effective approach to enhance the photoelectric response of hyperdoped bSi.

2. Materials and methods

Co-hyperdoped bSi with sulfur and nitrogen was prepared on double-polished n-type Si (100) wafers measuring 5 mm × 5 mm, with a resistivity of 3-5 kΩ·cm and thickness of 250 µm. The fabrication utilized fs laser etching (515 nm, 1 kHz, 190 fs) in a mixed atmosphere of SF₆ and NF₃. The laser fluence was about 5.66 kJ/cm² with the laser average scanning speed of 0.7 mm/s. Three samples with different sulfur and nitrogen dopant ratios, labeled SN1, SN2, and SN3, were obtained when gas pressure (kPa) ratios of SF₆ to NF₃ were 56/14, 35/35, and 7/63, respectively. Previous research indicates that the mixed atmosphere led to the co-doping of sulfur and nitrogen, and doping concentrations in the surface layer are well above their solid solubility limit [17]. The surface morphology of the laser ablated regions of SN2 was visualized by scanning electron microscopy (SEM), as shown in Fig. 1(a). The SEM images of SN1 and SN3 samples are similar to that of SN2, details can be found in Ref. [18].

 

Fig. 1. (a) SEM images (tilt 45°) of SN2. (b) The device structure and its asymmetric illumination. (c) Schematic of the experimental setup to record photocurrent under different light and gas environments.

Download Full Size | PDF

After laser ablation, to reduce the contaminants and oxides on the surface of the samples, all samples were cleaned in acetone and diluted HF successively. Subsequently, two coplanar aluminum (Al) electrodes measuring 1 mm × 4 mm were thermally deposited on the bSi surface, with a thickness of 80 nm and an exact 2.5 mm spacing interval. A more detailed process can be found in earlier work [15]. The device structure is depicted in Fig. 1(b). An ohmic contact was established between the Al electrode and the bSi, enabling the study to concentrate on the surface properties of the bSi, devoid of interference from other factors. Additionally, the ohmic contact can effectively injects carriers into the conduction channel [19].

In this study, an asymmetric illumination method was employed, wherein a rectangular light spot, approximately 0.5 mm × 2 mm in size, was directed between the two electrodes of the device, resulting in a non-uniform light distribution on its surface, as shown in Fig. 1(b). According to the LPE, the device produces a lateral photocurrent as a result of the asymmetric illumination. The lateral photocurrent serves as a carrier for both gas signal detection and optical signal reading, measured by a Keysight precision source/measure unit (B2902A). The setup is depicted in Fig. 1(c), the design of which needs to take into account both the light environment and the gas environment. The constant light environment for the basic gas sensing tests was created by a 730 nm central wavelength LED for convenient observation of light position. And different light environments, characterized by varying wavelengths of illuminated light, were achieved by a 150 W bromine tungsten lamp (LSH-T150) coupled with an Omni-λ300 spectrometer from Zolix. To establish a particular gas environment, the device was immobilized in a 5-liter sealed gas chamber as shown in Fig. 1(c). Various gas environments, in this case with different NO₂ concentrations, were obtained by injecting different volumes of pure NO₂ (≥99.99%) into the gas chamber.

To ensure the reproducibility and reliability of the experiments, the entire experimental temperature was controlled at 25 ± 1 °C and relative humidity at 30 ± 2%. Additionally, to ensure stability, the devices underwent a one-month natural aging process before testing. First, we consider the devices as gas sensors. The transient response and repeatability of the lateral photocurrent to NO₂ were tested for all sensors in the constant light environment with the LED asymmetric illumination, providing a light intensity of approximately 2.8 mW/cm². In order to investigate the effect of different light environments on the gas sensing performance, besides transient response, we experimentally obtained the relative gas responsivity (R_es) of gas sensors illuminated with different light wavelengths. R_es is given by R_es = (I_gas - I_air)/ | I_air |×100%, where I_gas and I_air represent the lateral currents reflecting the measured gas signal and the background air signal, respectively.

When applying the devices for photodetection, our focus shifted from the light environment affecting gas sensing to the gas environment affecting photo-detecting. First, the lateral photoelectric properties of the detectors were investigated in a constant gas environment (i.e., air). Photocurrents were then measured under different gas environments, corresponding to different concentrations of NO₂, with the aim of investigating the effect of the gas environment on photo-detecting performance. To more intuitively reflect the impact of the gas environment on the photodetectors, we compared the photo-detecting performance in the air environment and a specific NO₂ concentration gas environment. This comparison included the on/off ratios, photoresponsivity (R_λ) to light of different wavelengths (photo response spectra), and the response to different light intensities. The on/off ratio is described by I_light/I_dark, where I_lgiht and I_dark are the lateral currents measured under illumination with a specific wavelength of light and in the darkness, separately. R_λ is define as R_λ=(I_light-I_dark)/P_in, where P_in is the optical power of the incident light. Based on the response (i.e., photocurrent) to different light intensities, the sensitivity of the photodetectors to light intensity was obtained from the first-order derivative of the photocurrent with respect to the light intensity.

3. Results and discussion

3.1 Basic gas sensing performance and enhanced gas response by modifying the light environment

First, the basic performance of the gas sensors was tested in the constant light environment. Repeatability is a crucial metric for characterizing the performance of a gas sensor and for ensuring the reliability of experimental results. The repeatability and transient response of the sensors was investigated by testing 20 ppm NO₂ four times under the same light environment illuminated by the 730 nm LED as shown in Fig. 2(a)-(c). According to data of Fig. 2(a)-(c), we can see that the response time, recovery time, and response value (photocurrent) are almost repeatable. The results demonstrate the sensors’ outstanding reversibility and repeatability in detecting NO₂.

 

Fig. 2. (a)-(c) Repeatability and transient response of SN1-SN3 in response to 20 ppm NO₂ gas. (d) Gas responsivity spectra of SN1-SN3 in different light environments. (e) The response of SN1 to 20 ppm NO₂ in two light environments: a 940 nm LED and a 730 nm LED, both with a light intensity of 1.5 mW/cm².

Download Full Size | PDF

After presenting the basic gas sensing performance, the following section will focus on the effect of the light environment on the gas response performance. The light environment was altered by adjusting the wavelength of the light that illuminated the bSi surface. Figure 2(d) shows the normalized gas responsivity spectrum for 300 ppm NO₂, which represents the gas responsivity in different light wavelength environments, spanning the range from 600 to 1200 nm. It should be pointed out that the optical power of the illuminating light has been uniformly normalized to make the gas response of the sensors comparable at various wavelengths. Thus, the data in this figure is not meaningful in numerical terms for comparison with other gas sensors. The most intriguing aspect of the graph is the significant variation in gas response among the three sensors as a function of light wavelength. Under the same optical power, all three sensors are enhanced to different extents under the optimal light environment. When illuminated with the 740 nm light used in the previous study [15], the normalized gas responsivity of SN1, SN2 and SN3 is approximately 52.8%/µW, 490%/µW, and 7.4%/µW respectively. While at the optimal point of operation, the gas responsivity improves to 457.5%/µW (at 1000 nm), 2885.2%/µW (at 1110 nm), and 1490.9%/µW (at 1190 nm), representing enhancement factors of about 9, 6, and 201 times, respectively. This indicates that the gas sensor's responsivity can be effectively enhanced by selecting specific light environment with appropriate wavelengths. Another interesting observation is the difference in the wavelengths corresponding to the peaks of the three curves. In the range of 600-1200 nm, the curves of SN1 and SN2 have a similar trend with the presence of two maxima points, and the corresponding wavelengths are 1110 nm and 1000 nm, respectively, whereas the gas responsivity of SN3 has only one extreme point at 1190 nm. This suggests a difference in the optimal light environment for gas response enhancement for different sensors. It should be noted that the SN1-SN3 response to above 1200 nm light is extremely weak, as will be explained later on. This means that both I_gas and I_air are quite minor. A high gas responsivity becomes irrelevant considering the non-negligible noise in small signals, so light environments above 1200 nm are not taken into consideration.

To visually show the influence of the light environment on the gas response, the SN1 sensor response to 20 ppm NO₂ was also tested in the two different light environments. Figure 2(e) depicts the response of SN1 to 20 ppm NO₂ in the light environments of a 730 nm LED and a 940 nm LED illumination. Both light sources had the same intensity of 1.5 mW/cm². As shown in Fig. 2(e), the response to 20 ppm NO₂ was significantly enhanced when illuminated with the 940 nm LED compared to the 730 nm LED. After exposure to NO₂, the change in photocurrent greatly increased from 163 nA to 510 nA. Our previous study systematically investigated the gas sensing performance of bSi co-hyperdoped with sulfur and nitrogen in different ratios [15]. The results show that changing the co-doping ratio of sulfur and nitrogen can effectively adjust the NO₂ gas sensing performance under an asymmetrical light illumination of a 730 nm LED. According to the results presented in Fig. 2(d) herein, it is clear that at the optimal wavelength, besides the significant enhancement of the responsivity of the three sensors, the differences between their gas sensing performances are also more significant. This indicates that when applying the 940 nm LED illumination, the sulfur and nitrogen co-doping ratios in bSi can more effectively regulate the gas sensing performance parameters of the sensors, to meet different application requirements.

3.2 Lateral photoelectric properties and enhanced photoelectric response by changing gas environment

The lateral photoelectric properties of the three photodetectors were first investigated in the ambient air environment. Figure 3(a)-(c) illustrates the current-voltage (IV) curves of the detectors under light and no light conditions, ranging from −1 mV to 1 mV in bias voltages. The detected light was produced by the 730 nm LED, with an intensity of 2.8 mW/cm². It is evident that all curves exhibit a linear trend, suggesting a reliable ohmic contact between the Al electrode and the bSi surface. Comparing the IV curves of the three devices under no light condition, there are differences in the numerical values of the slopes. The resistances of the three samples are 20.0 kΩ, 27.1 kΩ and 74.2 kΩ, respectively. These disparities can be attributed to varying levels of nitrogen doping in the bSi. Nitrogen doping has been found to have a significant reparative effect on crystal lattice disfigurement [18,20,21]. Therefore, as the proportion of nitrogen doping in bSi increases, the crystal lattice integrity can be significantly improved, resulting a resistance closer to that of crystalline silicon (∼70 kΩ). It is also evident from Fig. 3(a)-(c) that the lateral photocurrents at zero bias differ significantly among the three detectors. The contributions of sulfur and nitrogen doping to the lateral photocurrent appear to be opposite. SN1 with a larger sulfur contribution exhibits a more positive photocurrent (38 nA), while SN3 with a larger nitrogen share exhibits a more negative photocurrent (−160 nA). Furthermore, SN2 exhibits a particularly small photocurrent (20 nA). This is consistent with the previous findings [15]. A more detailed explanation will be provided in the section on the mechanism of gas effects on the lateral photocurrent.

 

Fig. 3. (a)-(c) Current-voltage characteristics and lateral photocurrents of SN1-SN3. (d) The transient photocurrent curve of the SN1 upon exposure to different NO₂ concentration environments.

Download Full Size | PDF

In addition to the impact of the sulfur and nitrogen dopant ratios, the lateral photocurrent is also influenced by changes in the placement of the rectangular light spot. Since SN3 has a larger absolute value of lateral photocurrent than SN1 and SN2, which is more convenient for measurement and observation, we investigated the dependence of SN3 photocurrent on light spot position. The data in Fig. 4 represent the lateral photocurrent of SN3 obtained by altering the illumination position. The location of the electrodes is colored in the figure with shading. The photocurrent curve exhibits two peaks, located near the positive and negative electrodes, respectively. Additionally, when the spot is in the middle of the two electrodes, the photocurrent is nearly zero due to the uniform lateral distance and the symmetry of the carrier diffusion distance. This phenomenon has been observed in other similar structures as well [22,23]. The dependence of the lateral photocurrent on the spot position demonstrates its capacity to convey positional information. It is worth noting that the detector is most sensitive to light intensity when it is illuminated near the electrodes, where the photocurrent is maximized. Therefore, unless specified otherwise, the measured lateral photocurrents in this study were obtained at the maximum position, where the spot illumination was positioned close to the positive electrode.

 

Fig. 4. Illumination position-dependent lateral photocurrent curve of SN3.

Download Full Size | PDF

In the air environment, the lateral photocurrents of the three detectors were suboptimal, only a few tens or hundreds of nA. However, changing the gas environment resulted in a breakthrough. Figure 3(d) displays the transient photocurrent curve of the SN1 detector upon exposure to different gas environments, with NO₂ concentrations gradually increasing from 50 ppm to 900 ppm. As depicted in Fig. 3(d), the introduction of only 50 ppm of NO₂ gas leads to a substantial rise in the photocurrent. This indicates that even a minor alteration in the gas environment can drastically enhance the photocurrent of the detector. Additionally, the photocurrent gradually rises with increasing NO₂ concentration, without reaching saturation even at 900 ppm. The large operating margin for adjusting the gas environment to boost detector photocurrent indicates that the gas environment could serve as a flexible enhancer of photodetector sensitivity.

To provide an intuitive assessment of the influence of the gas environment on the detector's performance, we compared the detector's behavior in both air and NO₂ gas environments. Figure 5 shows the photocurrent versus time plots of the detectors irradiated by 730 nm LED with a light intensity of 2.8 mW/cm² in air and 300 ppm NO₂ gas environments. The LED lamp was periodically switched on and off at 20 seconds intervals under the control of the Keysight B2902A, and the photocurrent was measured at intervals of 0.1 seconds. Figure 5 clearly illustrates that in both air and NO₂ gas environments, the photocurrent first increases rapidly, then remains stable, and finally decreases rapidly in a single light source switching cycle. This suggests that all three detectors exhibit exceptional rapid rise time and decay time. The variations in the photocurrent display a remarkable level of consistency over multiple cycles, indicating that the device possesses a very high degree of stability. The on/off ratios for SN1, SN2, and SN3 in air are 24.5, 20.3, and 290.9, respectively. After injecting 300 ppm NO₂, there are jumps in the on/off ratios to 1438.7, 1132.9, and 6517.2, respectively, with negligible changes in the dark currents. This means that a certain concentration of NO₂ can drastically improve the on/off ratio without sacrificing the otherwise favorable rise time, decay time, and dark current of the device in air. It is noteworthy that SN3 behaves with a negative photocurrent in air but a positive photocurrent in NO₂, which is also illustrated in the transient response to 20 ppm NO₂ in Fig. 2(c). This phenomenon implies that the effect of 20 ppm NO₂ on the detector is strong enough to reverse the positive and negative of its photocurrent. Predictably, this gas enhancement approach can be attempted irrespective of the initial photocurrent performance of the detector, provided that it responds appreciably to this gas.

 

Fig. 5. Comparison of time-resolved photocurrent of the SN1, SN2 and SN3 in response to light on/off at 20 seconds intervals in air and 300 ppm NO₂ gas environments.

Download Full Size | PDF

Next, spectral photo response measurements of the SN1-SN3 in air and NO₂ gas environments were further performed to investigate how NO₂ environment affects the photocurrent generated by light of various wavelengths. Figure 6(a),(b) shows the spectral response of the three detectors in air and in 300 ppm NO₂. As evidenced by Fig. 6(a),(b), the photoresponsivity of SN1-SN3 in air is quite weak, with SN2 exhibiting the feeblest photoresponsivity, and even the highest one observed in SN3 is merely about 1 mA/W. While in NO₂ gas environment, the photoresponsivity of all the three detectors is remarkably enhanced. In particular, the peak photoresponsivity of SN2 increases from 0.08 to 19.27, more than two orders of magnitude. A similar finding can be obtained in the on/off ratios shown in Fig. 5. The more interesting point is the effect of NO₂ on the peak spectral response position. The peak photo response values for SN1, SN2, and SN3 in air correspond to wavelengths of 1060 nm, 1030 nm, and 970 nm, respectively. This can be attributed to the decreasing sulfur doping content of the three detectors, as the incorporation of sulfur doping in bSi leads to an increase in NIR absorption [2]. When exposed to the NO₂ environment, the spectral response for SN1 features a blue-shifted with the peak wavelength decreasing to 1020 nm. In contrast, the SN3 exhibits a red-shifted to 1070 nm, and the peak wavelength of SN2 remains constant at 1030 nm. In conclusion, NO₂ not only enhances the photoresponsivity significantly but also causes a slight shift in the peak response wavelength. The combined effect of NO₂ environment on the photo-detecting performance of the three detectors is summarized in Table 1.

 

Fig. 6. Photoresponsivity spectra of SN1, SN2 and SN3 in (a) air and (b) 300 ppm NO₂ environments. Photocurrent as a function of light intensity of SN1, SN2 and SN3 in (c) air and (d) 300 ppm NO₂ environments. Sensitivity to the light intensity of SN1, SN2 and SN3 in (e) air and (f) 300 ppm NO₂ environments, data obtained from (c) and (d) respectively.

Download Full Size | PDF

Table 1. Photo-detecting performance of SN1, SN2 and SN3 in air and 300 ppm NO₂ environments.

View Table

The dependence of the photocurrent on the incident light intensity is also a critical factor in evaluating the photo-detecting performance. The photocurrent of the three detectors was tested for the different light intensities from 0 to 4.48 mW/cm² by adjusting the drive current of the illumination 730 nm LED. According to the observation presented in Fig. 6(c), the three curves closely resemble straight lines, indicating a nearly linear relationship between the photocurrent and light intensity in air environment. The inclusion of 300 ppm NO₂ increases the photocurrent of all three detectors by more than a factor of 10 (Fig. 6(d)). The data in Fig. 6(e) and Fig. 6(f) represents the first-order derivatives of Fig. 6(c) and Fig. 6(d), respectively, with respect to light intensity, indicating the detectors’ sensitivity to light intensity. Comparison of Fig. 6(e) and Fig. 6(f) indicates a significant improvement in the sensitivity of all three detectors in the presence of NO₂, particularly at low light intensities. Although the sensitivity decreases with light intensity, it remains higher in a NO₂ environment than in air.

4. Mechanism of light and gas effect on the lateral photocurrent

Whether the gas sensing performance is affected by the light environment or the photo-detecting performance is affected by the gas environment, both light and gas are essential factors. Therefore, investigating the mechanism of light and gas effects on lateral currents will provide a better explanation of the above experimental phenomena. Differences in the effects of various doping ratios also need to be considered to explain the different phenomena that occur in SN1-SN3 devices.

Before discussing the effects of light and gas environments, we initially focus on the principle of lateral photocurrent generation by asymmetric illumination. This is imperative as the lateral photocurrent is utilized in this work to achieve light and gas dual-function detection. The Dember effect [24] provides elucidation for how co-hyperdoped bSi generates lateral photocurrents under illumination in the positive electrode. As shown in Fig. 7, when light is directed to the positive electrode, a significant amount of non-equilibrium electron-hole pairs are produced in its vicinity. Obviously, the electron concentration (n_e) and the hole concentration (n_p) in the illuminated region are considerably larger compared to the dark region. Due to the difference in concentration, electrons and holes will diffuse towards the dark region and be collected by the electrode, resulting in a negative diffusion current I_pe and a positive diffusion current I_ph, respectively. The lateral photocurrent I equals the sum of the electron diffusion current and the hole diffusion current, expressed as I = I_pe + I_ph. Defined by the diffusion current, I_pe is directly proportional to the gradient of non-equilibrium electron concentration and the diffusion coefficient of electrons (D_e). Similarly, I_ph is directly proportional to the gradient of non-equilibrium hole concentration and the diffusion coefficient of holes (D_h). Considering the extremely low concentration of non-equilibrium carriers in the dark region, n_e and n_p can to some extent represent their concentration gradients between the dark and illuminated regions. We assume that n_e and n_p in the illuminated region are the same, and their concentration gradients are also equal. Here, the transport of electrons and holes along the longitudinal direction between the bSi and the substrate is ignored. In fact, no notable n+/n or p/n junctions were observed between the bSi and the n-type silicon substrate in the unannealed bSi [25], so such an approximation is reasonable.

 

Fig. 7. Schematic diagram of the black silicon sensor section under the combined effect of light and gas.

Download Full Size | PDF

With similar values for n_e and n_p, the lateral photocurrent is contingent upon D_e and D_h. Generally, electrons exhibit greater mobility than holes, and their diffusion coefficient is also higher. This conclusion certainly holds true for the silicon substrate used in this experiment. However, during the fabrication process of bSi hyperdoped with chalcogens, the intense interaction between the fs laser and the silicon leads to the formation of numerous defects that hinder electron transport [26,27]. Therefore, D_e is likely smaller than D_h, resulting in a positive photocurrent for SN1 in air. As mentioned earlier, the introduction of nitrogen doping to bSi enhances its crystallinity by repairing the crystal lattice defects. This explains that SN3 demonstrates a negative photocurrent similar to the pristine silicon substrate. When the sulfur and nitrogen doping ratio is near, D_e and D_h become similar, so the photocurrent for SN2 is quite small.

Based on the mechanism of lateral photocurrent generation, we can gain further insight into the impact of the NO₂ environment on photo-detecting performance. NO₂ is widely recognized as a potent oxidizing gas. Consequently, when NO₂ molecules adsorb on the surface of bSi, they act as electron acceptors, capturing some of the free electrons available on the bSi surface, as shown in Fig. 7. Once a substantial quantity of NO₂ is adsorbed on the bSi surface, the Fermi energy level of the surface becomes close to the valence band. This will be more conducive to hole diffusion rather than electrons. Therefore, D_h becomes dramatically larger than D_e. On the other hand, the adsorption of NO₂ reduces n_e, making n_p larger than n_e. As a consequence, SN1, SN2, and SN3 all displayed extremely large positive lateral photocurrents when subjected to 300 ppm NO₂. While in the dark condition, electrons and holes are equally distributed between the two electrodes and there is no concentration difference and no diffusion. Therefore, although NO₂ impacts D_e and D_h, it has little effect on the dark current, resulting in an incredibly low dark current, approximately on the order of ${10^{ - 10}}$A. In other words, NO₂ can effectively increase the photocurrent without a concurrent increase in the dark current. This accounts for the improved photoresponsivity and on/off ratio of NO₂.

Similarly, we can better comprehend how the light environment affects gas sensing performance based on the mechanism of lateral photocurrent generation. From the definition of gas responsivity, the gas response spectrum is the combined effect of the photo response spectrum (photocurrent) in air and in NO₂, which corresponds to I_air and I_gas, respectively. On one hand, selecting an appropriate light wavelength enables more efficient generation of electron-hole pairs, manifested as larger I_air and I_gas, which contributes to signal extraction. On the other hand, however, an excessively high I_air value signifies a correspondingly high background signal, leading to a potential adverse impact on the gas responsivity. Therefore, as shown in experimental result (Fig. 2(d)), although the gas responsivity has been significantly enhanced in a certain wavelength range, there are slight fluctuations in the specific numerical values. The different gas response behavior of SN1, SN2 and SN3 can be explained as follows: The variation in co-doping ratios results in disparate diffusion coefficients for electrons and holes, and certainly may also lead to differences in the ability to adsorb NO₂. Furthermore, doping-induced changes to the bSi impurity band result in inconsistent absorption and response of SN1, SN2, and SN3 to the same wavelength of light, leading to different numbers of electron-hole pairs being produced. All of these factors ultimately affect the variation pattern of I_air and I_gas with wavelength, resulting in different response spectrum, as shown in Fig. 2(d).

5. Conclusions

In summary, the performance of NO₂ gas and light dual-function detectors based on bSi co-hyperdoped with sulfur and nitrogen was investigated under various light and gas environments. The results obtained in this study demonstrated that the two environmental factors, light and gas, produce surprising mutually reinforcing effects on the NO₂ and light detection performance of the device. The gas sensor's responsivity to NO₂ can be significantly improved by about 5 to 200 times when illuminated by the optimal wavelength compared to the 730 nm illumination. This finding reveals the substantial modulation effect of the illumination wavelength on the performance of the gas sensor. The experiments extended the sample types to three samples prepared at different SF₆/NF₃ ratios of 56/14, 35/35, and 7/63. This broader selection of samples provides a more comprehensive reference for optimizing the gas sensors by changing light. The photoresponsivity of the photodetector is enhanced by a minimum of 15 times in a 300 ppm NO₂ gas environment compared to when exposed to air. For the given sample with inherently poor photoelectric properties prepared at SF₆/NF₃ ratios of 35/35, the gas intervention boosted the photoelectric response by over 200 times. This indicates that creating a NO₂ gas atmosphere is an effective approach to enhance the performance of the bSi photodetectors. Specifically, enhanced photo-response can be achieved by encapsulating a certain concentration of gas within the photodetector. Similarly, optimized performance can be achieved by encapsulating a light source of a specific wavelength in a gas sensor. The mutual enhancement effect of gas and light is instructive for the design of higher performance photoelectric and gas detectors employing materials that are both photo-sensitive and gas-sensitive.