1 Introduction

Being the main source of contamination, ammonia gas (NH3) is normally emitted as of organic decomposition, motor vehicles and manufacturing wastes [1]. Nevertheless, NH3 is a harmful matter and could cause swelling for skin, and damaging for eye and lung. Thus, an NH3 sensor with a great response that requires selectivity, high stability and a low detection threshold is important and immediately desirable. To understand this, researchers have completed many of the works and elaborated a diversity of sensitive materials [2,3,4,5,6,7,8]. Conversely, most of them are essential to be worked at the elevated temperatures owing to their low electrical conductivity at the room temperature (RT), which is uncomplimentary for power saving, sensor assimilation and thermal stability [9]. So, it is essential to improve an inexpensive, sensitive, fast and dependable ammonia gas sensor, which is still a confrontation for the familiar materials for gas sensing. Metal oxide semiconductor (MOS) was established, presented the cost effectiveness and improved the sensing performance as a profound sample in the species of a continuous layer or cluster landmasses [10,11,12,13,14]. For example, tin dioxide (SnO2) sample was prepared with copper oxide (CuO) nanoparticles [11, 12, 14]. Kumar et al. [15] have coated SnO2 and SnO2/CuO bilayer films straight on porous anodic alumina, the gas response (defined by Ra/Rg) toward 100 ppm of NH3 gas was about 3.2 at 180 °C. Li et al. [16] have studied the gas sensor performance based on SnO2:CuO deposited on silicon substrate, they observed that the sensitivity toward 100 ppm of NH3 gas molecules was about 54% using the formula ((Rg − Ra)/Ra) × 100. Eom et al. [17] have synthesized SnO2:Cu2O nanocomposite on glass substrate, they found that the sensitivity (R = (ΔR/Rg) × 100%) towards 200 ppm of NH3 is about 4% at the room temperature. Bo et al. [18] have prepared SnO2:CuO on graphene sheet, they reported that the sensitivity (R = (ΔR/Rg) × 100%) is 16% towards 50 ppm of NH3 gas. The prospective of the proposed porous silicon (PSi) excessive was used as gas sensor elements at the room temperature because of their amazing feature; a high specific surface area. Also, it has been described that the existence of NH3, NO2, O2 gases molecules and other gases can alter the conductivity of PSi layer [19, 20]. Thus, the aim of the current research is to boost the performance of the gas sensing based on a porous silicon layer by inserting of SnO2:CuO nanoparticles with different contents of CuO nanoparticles within a macroporous silicon layer (macroPSi). The content of CuO nanostructures was adjusted for the promoted response merits towards the ammonia (NH3) gas molecules.

2 Experimental work

2.1 PSi materialization

The n-type of silicon wafers substrates with an orientation (100) and a resistivity of 10 Ω. cm were used for the PSi fabrication. A 6 min dip in 10% HF was used to remove the surface native oxide. The laser-assisted method based on the combination solution of the etching process of 1:1 of ethanol (99.999 purity) and 40% of hydrofluoric acid was used to form the PSi layer. The current density of about 12 mA/cm2 of the etching process was fixed for 15 min, and the used laser has a 640 nm wavelength with the intensity of 50 mW/cm2. The setup of experimental work (Fig. 1) composed of: ammeter, photon source, suitable homemade HF resistance etching cell consists of Teflon as a container for HF solutions and fitting aluminum as an electrode for conducting the current to the bottom silicon surface, and the platinum electrode (immersing in solution) utilized as cathode. All these requirements are considered as effective parameters in laser-assisted method. Following the etching method, the porous silicon layer was cleaned by deionized H2O and then dried for a few minutes in the atmosphere. The porosity of porous silicon was obtained by the following equation [21]:

Fig. 1
figure 1

Experimental set-up of laser-assisted etching process

Full size image
$$P=({M}_{1}-{M}_{2})/({M}_{1}-{M}_{3})$$
(1)

where M1, the mass of the PSi film before anodization route, M2, the mass of the PSi film after anodization technique, and M3, the mass of the sample after eliminating the PSi surface utilizing the KOH solution.

2.2 Elaboration of the nanocomposite films

Afterward, the etching step was completed, the spray pyrolysis technique was used to deposit the nanoparticles on the porous silicon layer. The experimental set-up of the spray pyrolysis method comprises a solution, which was sprayed on the surface of porous silicon at 400 °C, where the reaction of the solution led to the synthesized optimal nanoparticles. The solution that used in this method consists of 0.1 M CuCl2 (by BDH, from ENGLAND) and 0.1 M SnCl4·5H2O (by BDH, from ENGLAND), different samples of (bare PSi, pure SnO2, 50%SnO2:50%CuO, 90%SnO2:10%CuO, and 70%SnO2:30%CuO) were fabricated with a 50 ml of the solution for each deposition process. The ideal parameter conditions for the spray pyrolysis method were detected by employing a plastic atomizer with a nozzle having an outlet of about 0.9 mm, and the spraying time was 7 s with an interval time of 70 s under a constant pressure of 2.5 bar. The flow rate of the spray was about 2.2 ml/min, and the distance between the sample and the nozzle was fixed at 20 cm.

2.3 Gas sensing characterization

The morphological aspects of SnO2:CuO/PSi Nanocomposites were examined via field emission scanning electron microscopy (FE-SEM) (INSPECT-550) and atomic force microscope (CSP model AA3000 AFM supplied by Angstrom Company). The chemical elements were investigated by electron dispersive spectroscope (EDS) (INSPECT-550). The crystal structure of porous silicon was provided by the X-ray diffract meter (Shimadzu 6000) via CuKα (λ = 1.54056 Å) radiation. The NH3 sensing performance was done in a chamber system consisting of an adjustable heat plate, a digital program readout system, and a glass cavity. The ohmic contact made of aluminum was deposited on the surface of the sample by thermal evaporation method using a mask of grid shape with a thickness of (25 nm). Figure 2 displays the representation of the gas sensing set-up. Throughout the test, the NH3 gas was introduced into the chamber, and the variation in the electrical resistance was measured at a fixed voltage of 5 V. The relative humidity was 32%, and the chamber dimensions are 1 × 1 × 1 m3. Different cylinders of ammonia were used to obtain various concentrations.

Fig. 2
figure 2

Representation of the gas sensing set-up

Full size image

3 Results and discussion

3.1 Morphological features

The morphology and the cross-sectional images of the surface of macroPSi, SnO2/macroPSi, and SnO2:CuO/macroPSi are investigated by FE-SEM, Fig. 3. Figure 3a depicts that the structure of macroPSi consists of non-completed pores, irregular and semi-cubic shapes, this is because of the overlapping of pores among the specific non-completed pores, and the macropores are randomly distributed over the layer. This process of overlapping among the pores under the fabrication path takes place because etching is a three-dimensional route with various etching rates [22]. The pore diameters have a range of about (0.35–4.8) μm and the average diameter is 2.82 µm. The porosity of macroPSi layer is about 71.6%. For the SnO2/macroPSi nanocomposite structures revealed in Fig. 3b, the appearances of SnO2 nanostructures that deposited on the surface of macroPSi are semi-cubic with particle sizes in the range of (82–100 nm). The morphological structure of SnO2:CuO/macroPSi nanocomposites with different concentrations of CuO is shown in Fig. 3c, and e, from these figures, one can notice that the SnO2:CuO nanostructures have a semi-spherical structure and the SnO2 particles have a nano nail-like shape on the walls of the pores. Figure 3d evinces a novel morphology for the CuO Bucky particles inside the pores and a nano nail-like structure for SnO2. A great number of nanoparticles are aggregated on the walls of the macroPSi, and the others of CuO nanoparticles are inside the macroPSi, this is in compatibility with [23]. By increasing the CuO nanoparticles content from 10 to 30%, the sizes of the nanoparticles decreased from 95 to 37 nm with enhancing the surface homogeneity, and for 50% content they returned to increase to 45 nm. From these results, it can be concluded that by monitoring the content of the copper oxide nanoparticles, the surface area over macroPSi layer will be improved.

Fig. 3
figure 3figure 3

FE-SEM images with cross section of a macroPSi, b un-mixed SnO2, c 10%CuO:90%SnO2, d 30%CuO:70%SnO2, and e 50%CuO:50%SnO2

Full size image

The SnO2:CuO nanoparticles elements were investigated by the spectra of EDS, as demonstrated in Fig. 4a–d. The existence of SnO2:CuO nanoparticles was confirmed by the appearance of elements Si, Cu, Sn, and O without any impurities. Moreover, it can be observed that as the CuO nanoparticles content increased, the peak of it increased within the SnO2:CuO nanoparticles. The elemental maps for Si, O, Sn, and Cu is displayed in Fig. 4e–h. The spreading of the elements of O, Si, Sn, and Cu clearly distinguishes the SnO2 from CuO ones.

Fig. 4
figure 4

Electron dispersive spectra a un-mixed SnO2, b 10%CuO:90%SnO2, c 30%CuO:70%SnO2, d 50%CuO:50%SnO2, whereas e–h is the element map for 30%CuO:70%SnO2

Full size image

3.2 Atomic force microscopy analysis

The topographical characteristics of SnO2/macroPSi and SnO2:CuO/macroPSi nanoparticles are elucidated in Atomic Force Microscopy pictures in Fig. 5. The AFM pictures confirm that the surfaces of the samples have a great degree of homogeneity. The root mean square (RMS) and the roughness of the surface are increased by increasing the CuO nanoparticles content, as depicted in Table 1. From this table, the highest value for the average roughness is (54.5 nm) and the lowest value for average grain size is (40.26 nm), which correspond to the CuO nanoparticles content of 30%. The granularity size distribution of the fabricated nanoparticle size, as presented in Fig. 5, emphasizes that the surface properties of the nanoparticles are varying according to the controlled content of CuO nanoparticles. As the particles sizes decreased to the lowest possible values the physical system under measurement reach to the nano regime boundaries and hence the expected vacancies within the topographical patterns will increase and hence the roughness is increased.

Fig. 5
figure 5

Two-Dimension and 3-Dimension AFM images of a Pure SnO2, b 10%CuO:90%SnO2, c 30%CuO:70%SnO2, d 50%CuO:50%SnO2

Full size image
Table 1 Roughness parameters and average grain size
Full size table

3.3 XRD analysis

Figure 6 manifests the peaks of the SnO2/macroPSi and the SnO2:CuO/macroPSi nanostructures analyzed by XRD. In Fig. 6d, the diffraction peaks at 26.6, 33.8, 37.9, 51.7, 54.7, 61.8, 78.7, 83.7, and 90.8 correspond to the (110), (101), (200), (211), (220), (310), (321), (222) and (411) phases of tetragonal SnO2 structure (PDF# 41-1445), respectively. In Fig. 6b and c, the diffraction peaks at 35.4 and 38.7 correspond to the (002) and (200) phases of monoclinic CuO, respectively, which were reliable with (PDF# 48-1548). The patterns that diffracted from the XRD were recognized to the typical SnO2 and CuO peaks, and no peaks were detected for other materials. In Fig. 6, the SnO2 and CuO peaks were widened. This is due to the poorer crystallization ratings of the film, which is related to the defects that arise from the spray pyrolysis technique, also the addition of impurities decreased the crystallization ratings [24]. From the XRD patterns, it can be noted that the average nanocrystal size decreased with the increasing of CuO nanoparticles content within the SnO2:CuO nanoparticles, therefore, the surface area will develop by the modification of CuO nanoparticles with the SnO2 nanostructure. Table 2 lists the surface area and the average nanocrystal size. The highest specific surface area, the largest amount of gas response for gas sensing performance.

Fig. 6
figure 6

X-Ray Diffraction patterns of; a SnO2, b 10%CuO:90%SnO2, c 30%CuO:70%SnO2 and d 50%CuO:50%SnO2

Full size image
Table 2 Average values of the nanocrystal size and the specific surface area
Full size table

The average crystallite size D was calculated using the Scherrer formula given by Eq. 2 [25].

$$D=(0.94 \lambda)/[\beta_{({2\theta})} {\text{cos}}\theta]$$
(2)

where, the wavelength of XRD is λ, the full width at half maximum of diffracted peaks is β(2θ), and θ is the Bragg angle.

The specific surface area (S.S.A) of SnO2:CuO nanoparticles was obtained by Sauter formula [26]:

$${\text{S.S.A}}=6000/(D\ast \rho_{\text{d}})$$
(3)

where, ρ is the density of SnO2:CuO nanoparticles, the Sauter’s formula involves a 'shape factor' which is 6 for the sphere.

3.4 Electrical properties

Current–Voltage characteristics were investigated under the dark condition and the different ammonia gas concentrations (50, 100, and 150) ppm and temperatures (room temperature (RT), 50, 75, and 100) oC, as shown in Fig. 7. All samples demonstrated that the current increased with the increasing of gas concentration and temperature. All the samples showed an ohmic contact for the bare PSi and modified PSi, this is due to the double junction: PSi/n-Si and SnO2:CuO, and the barrier height of the first junction could be in opposite direction to the second junction. One of the most important behaviors for the gas sensor property is the ohmic contact; this is related to the affected sensitivity of the sensor device by contact resistance and could be highly changed when introduced to the NH3 gas molecules. The same behavior was found by [27] for Au modified PSi, and also [17] obtained the ohmic behavior for SnO2, SnO2:CuO, and SnO2:Cu2O. The elevated resistance for the CuO loaded with SnO2 proposes that the depleted electrons of p–n junction were favorably created from the tin dioxide layer and are more effective than the oxygen adsorption [28]. Also it should be mentioned that there are two types of defects for the CuO dopant that could affect the resistance of the film. The first is the substitution defect, in which the Cu2+ could substitute Sn4+ this will increase the oxygen vacancy, and hence decrease the free electrons and increase the resistance. The second is the interstitial defect, in which the copper ions could occupy the interstitial positions in Sn4+, this will lead to the decreasing of oxygen vacancy and hence increasing the free electrons and the resistance of the sensor. Table 3 lists the resistance obtained from the current–voltage characteristics.

Fig. 7
figure 7figure 7

Current–Voltage characteristics for a PSi, b SnO2, c 90% SnO2:10%CuO, d 70% SnO2:30%CuO, and e 50% SnO2:50%CuO

Full size image
Table 3 Resistance of PSi, SnO2, and SnO2:CuO deposited on the etched n-Si
Full size table

3.5 Gas sensing measurement

The sensing features of bare porous silicon, SnO2 and SnO2:CuO/macroPSi nanocomposites were recorded to clarify the consequence of altering the CuO nanoparticles on the fabricated sensor. The measurement was repeated three times to confirm the obtained results. The FE-SEM images revealed that the CuO nanoparticles and Bucky particles inside the pores of macroPSi as well as the high specific surface area enhanced the response, which might offer additional adsorption locations for gas molecules. Furthermore, the suitable pore dimensions of (0.35–4.8 µm) might store the NH3 molecules provisionally to interrelate with the sensing element entirely. Figures 8a–e demonstrates that the resistance of the bare PSi, SnO2/macroPSi, and SnO2–CuO/macroPSi nanocomposite sensors varies with the interval upon experience to several NH3 gas concentrations (50, 100, and 150) ppm at different temperatures (RT, 50, 75, and 100) °C. All the SnO2–CuO/macro-PSi nanocomposites sensors showed a good reversible response. The CuO content enhanced the response to ammonia gas. In particular, the SnO2–CuO/macroPSi nanocomposites sensors responded to ammonia gas at the RT. After the exposure to NH3 gas, the resistance decreased and reverted to the previous case upon the exposure to air which corresponds to the typical n-type semiconducting behavior [29]. Figure 8a evinces a good response of macroporous silicon at all investigated temperatures, fast response time and recovery time at the RT, and when the temperature reached to 50 °C, the response and recovery times increased and so on at 75 °C, but at 100 °C, they became lower. Figure 8b illustrates that the SnO2 nanostructures have developed the sensitivity. The response time and recovery time get faster than that in macro-PSi at the RT. When the CuO nanoparticles incorporated with 10% to 90% SnO2, the response to ammonia improved at the RT and 50 °C with very fast response and recovery times. In Fig. 8d, as the content of CuO increased to 30%, the ideal response and recovery times obtained, especially at the RT as shown in Fig. 9, this could be attributed to the morphology of the copper oxide Bucky nanoparticles (Fig. 4d) inside the pores of PSi, high specific surface area, and particle size that boosted the response to NH3 gas. This is in agreement with those that reported the effect of the morphology and particle size on the gas sensing performance [30, 31]. The response and recovery times become little higher when 50% CuO nanoparticles loaded to 50% SnO2, this is due to the low specific surface area as illustrated in Table 2, which shows the great grain size, consequently the perfect response is prominent with 30% CuO Bucky particles content. The sensitivity is the maximum at the RT (25 °C) and it reduces with the elevated temperatures. The detected drop in the sensitivity might be linked to the shrinking sample resistance in air which is initiated from the improvement in the phonon assisted tunneling that results in dropping resistance [32]. Meanwhile, the developed structure in this work comprises relatively an extra density of O vacancies, and the diverse variation in the resistance of sensor with the elevating temperature appears reasonable. The noise in the sensor device could be attributed to many reasons, such as the thermal motion of electrical charges, the amplifier in locally manufactured gas sensor system, and Brownian motion which caused by the irregular forces arise from the random collisions of molecules in the low size particles. The last one comes to be more important when the size of a structure reduces. The noise in chemical sensors (adsorption–desorption processes) is related to the Brownian motion [33]

Fig. 8
figure 8figure 8

Resistance change as a function of time when introduced to different ammonia concentrations at various temperatures of a macro PSi, b SnO2, c 10% CuO:90% SnO2, d 30% CuO:70% SnO2, and e 50% CuO:50% SnO2 nanocomposites sensors

Full size image
Fig. 9
figure 9

Response and Recovery times for 50 ppm NH3 gas of nanocomposites sensors 70% SnO2:30% CuO/macro-PSi at room temperature

Full size image

Figure 10 indicates that the sensitivity as a function of temperature at 150 ppm gas concentration of macro PSi is good, but it decreases as the temperature increases. Sensitivity gets enhanced for SnO2 nanostructure to 37%. When the CuO nanoparticles were incorporated with 10%, the sensitivity improved at the room temperature with the value of 47%. The excellent sensitivity obtained when 30% CuO Bucky particles was loaded, it showed 57% of sensitivity at room temperature. As the CuO nanoparticles content was increased to 50%, the sensitivity fell back to 31% at the room temperature. All samples exhibited a perfect response at the room temperature, and the temperature, in which the sensitivity displays a supreme rate, is named the best working temperature. Conversely, if the working temperature rises more than the room temperature, the response will drop as the extent of the adsorbed ammonia gas on the nanocomposite internal exposure surface drops. Whereas, the desorption rate becomes principal with the growing working temperature, leading to a drop in response. Table 4 lists the response time, recovery time, and the sensitivity of the ammonia gas, and Table 5 illustrates the performances of various categories of nanocomposites sensors. The sensitivity of the gas sensor was calculated by the following equation [34]:

Fig. 10
figure 10

The sensitivity as a function of temperature at 150 ppm of ammonia gas

Full size image
Table 4 Response time, recovery time, and sensitivity to 150 ppm of NH3 gas at the RT
Full size table
Table 5 Comparison of the sensing properties of NH3 gas sensor based on different categories of nanocomposites sensors
Full size table
$${\text{Sensitivity}}= \frac{{R}_{\text{a}}-{R}_{\text{g}}}{{R}_{\text{a}}}\times 100\%$$
(4)

where, Rg is the measured resistance at the presence of ammonia gas molecules, and Ra is the resistance at the absence of ammonia gas molecules.

4 Conclusion

Briefly, the inserting of SnO2:CuO nanoparticles with different CuO nanoparticles contents deposited over photo electrochemically etched macroporous silicon represents an efficient route for developing a gas sensor. The 30% CuO content manifested larger specific surface area and smaller particle size with higher roughness surface than pure SnO2 and other CuO contents, as obtained from the XRD and AFM. FE-SEM revealed a pore depth of 2.5 μm with various pore diameters in the range (0.35–4.8 μm), and the porosity of the macroPSi was 71.6%. The NH3 gas sensor based on SnO2:CuO/macroPSi nanocomposite elucidated an extra sensitive performance, 4 s response time and 55 s recovery time of 4 s, especially when 30% CuO was loaded to 70% SnO2, it evinced a 57% of sensitivity at the room temperature. The high sensitivity of the nanocomposite is credited to its level of crystallinity, additional oxygen vacancies, and heterojunction creation. Consequently, SnO2:CuO/macroPSi nanocomposites with the improved sensitivity have a high probability in the NH3 sensor presentations.