Nan
Zhang‡
a,
Xiaohui
Ma‡
a,
Yanyang
Yin
a,
Yu
Chen
ab,
Chuannan
Li
a,
Jingzhi
Yin
*a and
Shengping
Ruan
*a
aState Key Laboratory on Integrated Optoelectronics and College of Electronic Science & Engineering, Jilin University, Changchun 130012, P. R. China. E-mail: yjz886666@163.com
bInstitute of Semiconductors, Chinese Academy of Sciences, Beijing 100083, PR China
First published on 3rd December 2018
In this paper, CuO–CdS composites were synthesized by a facile two-step solvothermal method. The composite materials were characterized by a series of systematical measures, such as X-ray diffraction, Fourier transform infrared spectroscopy, X-ray photoelectron spectroscopy, scanning electron microscopy and transmission electron microscopy. Through detailed experiments, it was observed that the shape of the as-prepared materials could be tailored by varying the CuO content in composites. The experimental results manifested that CdS decorated with 0.05 at% CuO presents superior performance including response/recovery speed, detection limit, optimum operating temperature, and sensitivity in comparison with pure CdS. The possible mechanisms of enhanced sensing properties were discussed in detail. At the interface of the two different material surfaces, the capacity of surface-adsorbed oxygen was elevated, which can be attributed to the formation of the p–n heterojunction between CdS and CuO. In conclusion, the functionalization of CuO with CdS may be a promising method to enhance gas-sensing properties.
Cadmium sulfide (CdS) is widely used as a material for light detection for its excellent light response. Its application in gas sensors is rare and its gas-sensing performance is not as good as expected due to the low chemical activity and poor electrical conductivity of the sulfide semiconductor. Therefore, discovering a suitable material that can be used to decorate the surface of CdS is necessary to achieve a better performance of gas sensing. Cupric oxide (CuO) is a typical p-type semiconductor with an energy gap of about 1.5 eV and has been widely used in catalysis and gas sensing.16,17 Volanti et al. studied the effects of different morphologies of CuO on gas sensing properties.18 In addition, Deng et al. introduced CuO into TiO2 nanofibers to improve the response of materials.19 According to these studies, it is possible that decorating CuO nanoparticles on the surface of CdS to form p–n junctions could effectively improve the gas sensing properties of CdS.
In this study, CdS nanowires were synthesized by a one-step solvothermal method, which could achieve a fast response/recovery time of less than 1 s as expected. To enhance the performance, CuO–CdS (p–n heterojunction) composites were synthesized by a solvothermal method. Different contents of CuO nanoparticles were grown on the surface of the CdS nanowires, forming p–n heterostructures. As a result, the prepared CuO–CdS composites showed superior gas sensing properties. The CdS composites doped with 0.05 at% CuO showed higher sensitivity, lower optimum operating temperature, and quicker response/recovery speed. Furthermore, the detection limit was decreased. In general, the excellent gas-sensing performance was highly associated with the p–n heterojunction structure.
The gas sensing properties were measured through a CGS-8 intelligent gas sensing analysis system (Beijing Elite Tech Co., Ltd, China). The response was defined as S = Ra/Rg. Ra represented the resistance of the sensor in air, and Rg was the sensor's resistance in the atmosphere of target gas. The response–recovery time was taken by the sensor changing 90% of the total resistance change.
The FTIR spectrum was recorded to analyze the chemical composition of the samples. The FTIR spectrum in the frequency range (400–4000 cm−1) of nanocrystalline CdS is depicted in Fig. 3. The peaks at 601 cm−1 and 405 cm−1 were induced by the stretching frequency of the Cd–S bond.21,22 The strong interaction of water with CdS was reflected by the peaks at 3368 cm−1 and 1629 cm−1, due to O–H stretching and O–H bending modes, respectively.23 The band at 1401 cm−1 came from aromatic C–C stretches.22 However, the FTIR spectrum did not show the presence of copper, which can be predicted because the content of trihydrate cupric nitrate added in the reaction was very small.
In order to further confirm the existence of the compound containing Cu, the XPS spectra of all samples have been provided in the ESI.† The measured atom ratios of Cu to Cd of the as-prepared samples according to XPS are shown in Table S1.† Taking into consideration the surface decoration of CuO and shallow detecting depth of XPS, it is reasonable that the atom ratio of Cu to Cd detected by XPS is higher than the theoretical one. In addition, it is worth mentioning that for a material with surface decoration, the surface content may be more important than the total content of the decorated component in the sample. Fig. 4 shows the sample of 0.05 at% CuO–CdS. The elements of Cd, S, Cu, and O can be observed clearly in Fig. 4(a–d), respectively. The peaks at 405.3 and 411.8 eV were indexed to Cd 3d5/2 and Cd 3d3/2 regions; the peaks at 162.8 eV and 161.4 eV can be assigned to S 2p regions; the peak of O 1s was at 532.1 eV; and Cu 2p spectra showed the peaks’ positions at binding energies of 936.5 eV and 954 eV. Because the gap energy between Cu 2p3/2 and Cu 2p1/2 was 19.6 eV, the test values were consistent with the theoretical values, so it was corresponding to CuO.24 Therefore, CuO–CdS composites were successfully synthesized.
Fig. 4 The XPS spectra of the 0.05 at% CuO–CdS: (a) Cd element, (b) S element, (c) Cu element and (d) O element. |
To observe the morphology of CuO–CdS composites, SEM and TEM of the samples were measured. Fig. 5 displays the SEM images of the as-synthesized samples. The pristine CdS nanowires are shown in Fig. 5(a and b), with 30–70 nm in diameter and several micrometres in length. As can be seen from Fig. 5(b), the surface of the nanowires was smooth and the diameter was about 65 nm. Fig. 5 displays the effect on the morphology of the materials owing to different contents of trihydrate cupric nitrate added in the reactions. It can be clearly observed that CuO nanoparticles grow on the surface of CdS nanowires. With different contents of trihydrate cupric nitrate, the amounts of CuO nanoparticles were different on the surface of CdS nanowires. There are many CuO nanoparticles in the sample of 0.5 at% CuO–CdS as shown in Fig. 5(c), while the sample of 0.05 at% CuO–CdS shown in Fig. 5(f) has fewer CuO nanoparticles. To observe the detailed morphology, TEM and high-resolution TEM (HRTEM) were necessary.
Fig. 5 SEM images of the as-prepared (a–b) pure CdS, (c) 0.5 at% CuO–CdS, (d) 0.1 at% CuO–CdS, (e) 0.05 at% CuO–CdS, and (f) 0.01 at% CuO–CdS samples. |
For characterizing the microscopic morphology of 0.05 at% CuO–CdS nanowires, TEM and HRTEM images are shown in Fig. 6. Fig. 6(a) depicts the TEM image of 0.05 at% CuO–CdS. The HRTEM image in Fig. 6(b), the partially enlarged picture of Fig. 6(a) at the interface of CdS nanowires and nanoparticles, further shows the detailed crystallographic structure of 0.05 at% CuO–CdS, which clearly revealed two distinct sets of lattice fringes. The measured interplanar distance of 0.336 nm in the CdS nanowire region was in good agreement with the d spacing of the (002) lattice planes of the hexagonal CdS crystal, while in the nanoparticle region, an interplanar distance of 0.215 nm was observed, and it complied with the lattice spacing of the (002) planes of CuO, confirming that CuO nanoparticles were successfully decorated on the surface of CdS nanowires. This result is consistent with that of XRD. Fig. 6c is the scanning TEM (STEM) image of a single nanowire, and the EDS elemental mapping images (Fig. 6d and e) confirm the spatial distribution of Cd and S elements in the CdS nanowires. As shown in Fig. 6(f), the signals of the Cu element were detected on the surface of CdS nanowires which also confirm that CuO nanoparticles were successfully decorated on the surface of CdS nanowires. In addition, the energy-dispersive X-ray spectroscopy (EDS) analysis (Fig. S1†) further showed that the 0.05 at% CuO–CdS nanowires consisted of CdS, Cu and S elements.
Fig. 6 (a) Low-resolution and (b) high-resolution TEM images of the as-prepared 0.05 at% CuO–CdS nanocomposites. (c–f) Scanning TEM (STEM) image and corresponding elemental mapping images. |
Fig. 7 Response of sensors based on CdS and CuO–CdS nanocomposites in 100 ppm ethanol at different operating temperatures. |
Fig. 8 presents the responses of all sensors varying with different concentrations of ethanol from 100 ppb to 500 ppm. Fig. 8(a) indicates that with increasing ethanol concentration, the responses of sensors increased rapidly, and then reached a stable (maximal) value. Besides, the samples of CuO–CdS composites were more sensitive to a low concentration of ethanol compared with pristine CdS nanowires; and to the same concentration of ethanol, CuO–CdS nanocomposites showed significantly higher responses. In the range of 100 ppb to 5 ppm, the pristine CdS nanowires had almost no gas-sensing performance; other samples presented linear performance, which was important for gas quantitative analysis (Fig. 8(b)). Therefore, the CuO–CdS nanocomposites exhibited excellent properties for the detection of a low concentration of ethanol gas.
Fig. 8 Curves of responses of the sensors based on pure CdS and CuO–CdS versus ethanol concentration at optimum operating temperature. |
The real-time detection is critical for a gas sensor. Therefore, the response–recovery time is also an important parameter. The response–recovery time properties of the samples are shown in Fig. 9. Fig. 9(a) shows the responses of the sensors based on 0.01 at% CuO–CdS, 0.05 at% CuO–CdS, 0.1 at% CuO–CdS, and 0.5 at% CuO–CdS in 100 ppb, 500 ppb, 1 ppm, 3 ppm, and 5 ppm of ethanol. The responses increased with the increase of ethanol concentration. Fig. 9(b–f) demonstrate the transient responses of all the sensors. The pure CdS-based sensor was tested at 230 °C in 10 ppm of ethanol, and other sensors were tested at 182 °C in 3 ppm ethanol. The results indicated that both the response time and recovery time of the pure CdS-based sensor were less than 1 s, which were short as predicted. The response times of the sensors based on 0.01 at% CuO–CdS, 0.05 at% CuO–CdS, 0.1 at% CuO–CdS, and 0.5 at% CuO–CdS were about 1 s, less than 1 s, 1 s, and 2 s; and, their recovery times were 3 s, 2 s, 1 s, and 9 s, respectively. These data suggest that all the sensors commonly have a fast response and recovery speed. However, compared with the sensor based on pure CdS, the response–recovery time of those based on CuO–CdS nanocompounds was a little longer. Possible mechanisms will be explained in the later part.
Selectivity is another important parameter for distinguishing different poisonous gases. The selectivity properties of the sensors based on 0.01 at% CuO–CdS, 0.05 at% CuO–CdS, 0.1 at% CuO–CdS, 0.5 at% CuO–CdS and pure CdS were tested at 182 °C and 230 °C, respectively. The response of the sensors in 100 ppm different gases including ammonia (NH3), acetylene (C2H2), acetone (C2H6O), ethanol (C2H5OH), xylene (C6H5C2H6), methylbenzene (CH3C6H5), and carbon monoxide (CO) has been tested (Fig. 10). The results showed that the sensors based on 0.05 at% CuO–CdS, 0.01 at% CuO–CdS, 0.1 at% CuO–CdS, 0.5 at% CuO–CdS and pure CdS possessed higher responses in ethanol than in any other gases, and the corresponding values were 120, 82, 61, 40 and 20, respectively. Consequently, the modification of the CdS nanowire with CuO nanoparticles significantly improved the selectivity towards ethanol. It may be determined by the following factors: on the one hand, CuO nanoparticles on the surface of CdS nanowires act as a catalytically active site for lowering the activation energy of the reaction between ethanol and chemisorbed oxygen species;25 on the other hand, according to the previous studies, CuO based sensors possess good selectivity in ethanol.17 In addition, the responses of CuO/CdS composite based sensors in 100 ppm ethanol, methanol and isopropanol as a function of temperature are shown in Fig. S6.† It can be clearly observed that the response of CuO/CdS based sensors in ethanol is 115, which is almost 3 times larger than in methanol (37.5) and 5 times larger than in isopropanol (22.3).
Long-term stability of the CdS sensor in 10 ppm ethanol and CuO–CdS based sensors in 3 ppm ethanol within 30 days are shown in Fig. 11(a). Responses of all the sensors gradually decrease and tend to be stable. For CuO–CdS based sensors, the changes in response values were small and they retained high sensitivity to ethanol within 30 days indicating their good stability. The response and recovery properties of the pure CdS based sensor in 10 ppm ethanol and CuO–CdS in 10 ppm ethanol and the CuO–CdS based sensor in 3 ppm ethanol were tested consecutively three times to investigate their repeatability, as shown in Fig. 11(b–f). During the test, as can be seen, the resistance values of the pure CdS and CuO–CdS based sensors in air change a little and also in the presence of ethanol, i.e. the response values are stable. Also, the response and recovery times of each sensor are very short. These indicate that the pure CdS and CuO–CdS based sensors have good repeatability.
The performances of the ethanol sensors in the previous literature are summarized in Table 1.17,26–29 So far, a large amount of semiconducting metal oxides have been utilized to detect ethanol, including CuO nanoleaves, ZnO/NiO composites, Cu–Fe2O3 microcubes, Fe2O3 nanoporous networks, and ZnO/SnO2 spheres. Our CuO/CdS based sensor has a high response of approximately 120 in 100 ppm ethanol at 180 °C. Thus, the equilibrium consideration, including the response, operating temperature and the response/recovery speed makes the 0.5 at% CuO/CdS nanowire based gas sensors competitive when compared with most ethanol sensors.
Sensing material | Ethanol | Temp. (°C) | Response | T res. (s) | T recov. (s) | Ref. |
---|---|---|---|---|---|---|
CuO nanoleaves | 10 ppm | 260 | 1.38 | 23 | 11.6 | 17 |
ZnO/NiO | 50 ppm | 400 | 37 | 2.1 | 4.1 | 25 |
Cu–Fe2O3 microcubes | 100 ppm | 225 | 20 | 26 | ||
Fe2O3 nanoporous network | 100 ppm | 200 | 35 | 4 | 179 | 27 |
ZnO/SnO2 spheres | 30 ppm | 210 | 35 | 6 | 120 | 28 |
Our work: | ||||||
CdS | 100 ppm | 180 | 20 | ∼1 | ∼1 | |
CuO/CdS | 100 ppm | 180 | 120 | ∼1 | 2 |
The CuO–CdS nanocomposites exhibited excellent properties owing to the formation of a p–n junction. As shown in Fig. 12(a), when the two types of materials came into contact with each other (n-type CdS and p-type CuO), the energy band bent at the interface due to different band gaps and work functions.31,32 Additionally, the formation of the p–n junction would increase the barrier height, making electron transition difficult. As a result, the resistances of CuO–CdS nanocompounds are larger than that of the pure CdS sample in air. When the sensing film was exposed to an ethanol atmosphere, ethanol would react with surface-adsorbed oxygen ions, making the depletion layer thicker. Thus, the formation of the p–n junction was the significant factor that enhanced the properties of the sensors. However, when the surface was decorated with excess CuO nanoparticles, the CuO nanoparticles cover the surface of pristine CdS, which leads to a decrease in the active area of CdS and weakens the gas response. In addition, the heavy coating will cause more electrons to be transferred between CuO nanoparticles rather than between p–n junctions, impairing the gas sensing properties.
The second reason is that more surface-absorbed oxygen ions would exist on the interface of CdS and CuO. As shown in Fig. 12, when CdS and CuO came into contact with each other, they would join together to form the p–n junction at the interface because of different band gaps and work functions. Therefore, the barrier can be formed such that the transition of electrons is prevented. As a result, electrons cannot jump to CdS, but accumulate on the surface of CuO. Then, more and more oxygen molecules are adsorbed at the interface, grabbing electrons from CuO and forming chemical-adsorbed oxygen ions. This process is beneficial for the reaction between ethanol and chemically-adsorbed oxygen ions on the surface of composites. When the sensor is exposed to the low concentration ethanol atmosphere, oxygen ions adsorbed at the interface could react with ethanol molecules more easily, compared with the pure one. Therefore, the CuO–CdS nanocompounds were superior for the detection of low concentration ethanol.
Furthermore, sensors based on the two components mixed together are more sensitive than those based on the single component alone, suggesting a synergistic effect between the two components.33 On the one hand, CdS catalyses the breakdown of ethanol extremely effectively. A combination of CdS and CuO would effectively dehydrogenate ethanol and subsequently catalyse the breakdown of ethanol. This explanation might not apply to all composite gas sensors. Only when the catalytic actions of the components complement each other, the performance of gas sensors will be enhanced. CuO can work as an efficient catalyst for oxidative dehydrogenation of ethanol within air. Besides, it can detect the low concentration of aldehydes. In the present study, when the low concentration of ethanol is explored, the sensors based on pure CdS nanowires show a tiny response, whereas the sensors based on the CuO–CdS nanocomposites catalyse the reaction (1) and induce reaction (2) to improve the responses. However, compared with reaction (1), reaction (2) is weaker in some degree. Taken together, the response–recovery time of some of the CuO–CdS samples is a little longer than that of pure CdS.
C2H5OH(gas) + O−(ads) → CH3CHO + H2O + e− | (1) |
CH3CHO(gas) + 5O−(ads) → 2CO2 + 3H2O + 5e− | (2) |
In general, there are many advantages that a p-type CuO grows on the surface of n-type CdS to achieve an excellent performance in ethanol detection. Therefore, CuO–CdS nanocomposites are promising in the field of gas sensors.
Footnotes |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c8qi00951a |
‡ These authors contributed equally to this work. |
This journal is © the Partner Organisations 2019 |