1. INTRODUCTION
Hydrogen sulfide (H2S) is a highly flammable and dangerous gas, known for its strong acidic nature. It
is commonly emitted from industrial processes such as oil refining and metal smelting,
as well as from crude oil and natural gas fields [1,2]. The corrosive nature of H2S leads to the degradation of metal parts in equipment [3]. Furthermore, exposure to low concentrations of H2S can result in nausea, headaches, dizziness, and irritation to the eyes, skin, throat,
and nose [4]. At higher concentrations, it can cause severe damage to the cardiovascular and
neuromodulatory systems [5]. H2S is also considered as a biomarker for various health conditions, including halitosis
[6,7]. Thus, the reliable detection of H2S gas is important from multiple perspectives. Various sensors have been developed
for the detection of H2S, including optical [8], surface acoustic wave [9], electrochemical [10], and resistive [11] gas sensors. Among these, resistive sensors, which are mostly based on semiconducting
metal oxides, are favored because of their low cost, high sensitivity and stability,
as well as fast dynamics [12-14].
SnO2, an n-type semiconductor (Eg= 3.7 eV [15)], is one of the materials most used for gas sensors because of its high electron
mobility, non-toxicity, availability, low synthesis costs, and good sensing properties
[16,17]. SnO2 has been used in the detection of various toxic gases, including H2S [18,19].
SnO2 powders or films can be produced by several methods, including chemical precipitation
[20], vapor– liquid–solid (VLS) [21], combustion [22], evaporation [23], pulsed laser [24], hydrothermal/solvothermal [25,26], and sol-gel [15] methods. While chemical methods produce fine powders with high purity, they are
time-consuming and need highly pure precursors. In addition, certain physical routes
require specialized and expensive facilities. In comparison, milling is among the
less expensive and simpler methods for preparing small-sized SnO2 powders.
High-energy ball milling is a mechanical process in which the continuous collisions
of milling balls cause the powders to deform and fracture, resulting in crushing and
mixing. The process can also reduce the activation energy of reactions, and decrease
the particle size to the nanometer scale, as well as enhance powder reactivity, and
induce chemical reactions at low temperatures. High-energy ball milling involves several
stages, including particle size reduction, local hotspot formation, lattice relaxation,
and structural decomposition [27]. Nanoparticles (NPs) can be produced using high-energy ball milling, significantly
enhancing the physicochemical properties of the milled materials. NPs offer a larger
surface area, thereby increasing the rate of chemical reactions and improving mechanical
strength. Furthermore, this method ensures uniform mixing of different powder materials,
leading to consistent properties in the final product [28]. Smaller particles lead to more contact points in the powders, resulting in higher
resistance modulation when exposed to gases. Kersen et al. [29] used ball milling to prepare SnO2 powders for H2S gas sensing. Later, this group leveraged mechanochemical synthesis to prepare contamination-free
SnO2 powders for H2S gas sensing [30]. Sapkota et al. [31], prepared ZnO particles using planetary ball-milling to detect H2 gas. CaO et al. [32], used ball milling to prepare a Fe2O3-ZrO2 composite for low temperature sensing of oxygen gas. Shin et al. [33], synthesized a WO3-In2O3 composite for CH4 gas sensing. Yadav et al. [34], synthesized TiO2 particles via a ball milling process for liquefied petroleum gas sensing. However,
despite these advances, only a limited amount of research has focused on the H2S gas-sensing features of SnO2 powders milled for varying durations.
Effective high-energy ball milling can minimize powder contamination, which is crucial
for applications in the electronics and biomedical fields. High-energy ball milling
can be scaled up to industrial production, facilitating the rapid introduction of
new materials into industry. Additionally, it produces high-quality NPs at relatively
low cost, offering substantial economic benefits. As a result, the technique is valuable
in materials science and engineering and is extensively employed to synthesize NPs,
microwave dielectric materials, and composites.
In this study, commercial SnO2 powders were subjected to high-energy milling for 30, 60, and 90 min. After milling,
the powders were characterized using advanced techniques to evaluate their morphology,
chemical composition, and crystallinity. The results indicated that milling for 60
min produced SnO2 powder with an optimal combination of crystallinity and a nearly round morphology.
Subsequently, a gas sensor was fabricated, and its H2S detection performance was assessed. Gas sensors fabricated from these powders demonstrated
the highest H2S sensing performance, which was related to the generation of fine particles and the
formation of double Schottky contacts among the SnO2 powders. Our findings indicated that SnO2 powders milled for 60 min exhibited superior sensing performance, which could be
attributed to the reduced particle size and increased surface defects, which enhance
the interaction between the gas molecules and sensor material.
Additionally, we comprehensively discuss the basic sensing mechanism of the as-fabricated
SnO2 powder, providing insights into the role of milling in improving the sensing properties
of SnO2 powders. The findings highlight the potential of milling as a simple yet effective
technique to improve the sensing capabilities of metal-oxide-based sensors, with applications
extending beyond SnO2 to other metal oxides.
3. RESULTS AND DISCUSSION
3.1. Characterization studies
Figures 2 (a)–(b) provide SEM views of commercial neat SnO2 powders without milling, and Figs 2 (c)–(d) show SnO2 powders after being milled for 60 min.
A comparison of the two samples revealed that while the overall morphology remained
largely unchanged after 60 min of milling, the number of particles with an initial
diameter of 400 nm significantly decreased. This outcome indicated that the milling
process reduced the particle size. In other words, simple milling for 60 min can reliably
reduce the size of commercial SnO2 powders (sub-nano level) to the nanoscale.
Figures 3 (a)–(d) show the high-angle annular dark field (HAADF) elemental mapping analysis results
of the SnO2 powders after milling for 60 min, clearly demonstrating the uniform distribution
of both Sn and O elements.
Figure 3 (e) presents the results of the HAADF chemical analysis of the SnO2 powders after milling for 60 min. Both Sn and O elements were detected, and their
amounts were 14.08 and 85.92 wt.%, corresponding to 54.87 and 45.13 at.%, respectively,
closely matching the chemical formula of SnO2. Figure 4 (a) depicts the XRD patterns of commercial SnO2 powers as well as those milled for different durations.
All of the samples exhibited the same peaks, which wellmatched JCPDS Card No. 41-1446
for SnO2 with a tetragonal rutile crystal structure. The absence of other peaks is indicative
of high sample purity and a clean milling procedure, free from unwanted contamination.
To further examine the peak positions, Figs 4 (b)–(c) show enlarged XRD patterns within the Bragg angle ranges of 26.2°–27° and 33.4°–34.4°,
respectively. In both cases, the peak positions were slightly shifted toward larger
Bragg angles, indicating a decrease in the distance of crystalline planes according
to Bragg’s law (d = λ/2sinθ). Specifically, the 2θ value of the peak corresponding to the (110) plane
shifted from 26.63° to 26.76°, and the 2θ value of the peak corresponding to the (101)
plane shifted from 33.91° to 34.05°. Bragg’s law was used to calculate the change
in interplanar distance, indicating that the interplanar distance for the (110) plane
decreased from 0.334 nm to 0.333 nm and that for the (101) plane decreased from 0.264
nm to 0.263 nm. This decrease in interplanar distance can be partially related to
the formation of oxygen vacancies. However, in milled samples, lattice strain can
also contribute to a shift in peaks [35], therefore, the entire amount of peak shifting cannot be related to the formation
of oxygen vacancies.
Figure 5 (a) presents a high-resolution TEM (HRTEM) image of a single SnO2 particle after milling for 60 min.
The particle exhibited a round shape with a diameter of approximately 40 nm. Figures 5 (b)–(e) present higher magnification HRTEM images, with parallel spacings between planes
of 0.243 nm and 0.328 nm, corresponding well to the XRD results for the (101) and
(110) crystalline planes of rutile tetragonal SnO2, respectively [36,37].
The extent of crystallinity can be established using fast Fourier transform (FFT)
patterns. Figures 6 (a) and (c) show the FFT patterns of the (110) and (101) planes of commercial SnO2, and SnO2 powders after milling for 60 min, respectively.
The sharp patterns in both cases demonstrate the high crystallinity of both analyzed
powders [38]. Figures 6 (b) and (d) display the corresponding image intensity contrast distributions in two vertical
directions. The results confirm the consistency of the interplanar distances calculated
using Bragg’s law from the XRD results, the visually observed interplanar distances
from HRTEM images, and distances determined through FFT. After milling, the SnO2 particles were fractured, which led to the formation of oxygen vacancies and a reduction
in interplanar distances.
Figures 7 (a)–(d) illustrate the O1s XPS core-level regions of commercial SnO2 powders and SnO2 powders after milling for 30, 60 and 90 min, respectively.
In all cases, O1s is fitted into three peaks located at 530.4, 531.6, and 532.6 eV,
corresponding to lattice oxygen, oxygen vacancy, and adsorbed oxygen species, respectively.
As both oxygen vacancies and adsorbed oxygen species are important for gas sensing,
the amounts of these oxygen species were also calculated (Fig 7). For commercial SnO2 powders, the areas of the peaks related to oxygen vacancies and adsorbed oxygen species
were 15.24% and 6.31%, respectively. For SnO2 powders milled for 30, 60 and 90 min, these values were 14.99% and 7.35%; 17.30%
and 5.65%; and 16.61% and 5.50%, respectively. Accordingly, milling for 60 min resulted
in the highest amounts of oxygen vacancies and adsorbed oxygen ions, which are expected
to be beneficial for H2S detection.
3.2. Gas sensing studies
Figures 8 (a)–(d) present the response curves of both commercial and milled (30, 60, and 90 min) SnO2 gas sensors at 300 °C toward 2–10 ppm H2S gas, respectively.
Figure 9 presents the corresponding calibration curves, and Table 1 compares the sensing results.
The response of commercial SnO2 powders to 2, 4, 6, 8, and 10 ppm H2S was 1.02, 1.40, 1.64, 2.10, and 2.46, respectively. However, after milling for 60
min, the response to these concentrations improved to 1.44, 1.84, 2.26, 2.75, and
3.01, respectively. This indicates that sufficient milling time enhances the sensing
performance. Notably, after 90 min of milling, the response to most H2S gas concentrations decreased. Additionally, the sensor fabricated from SnO2 powders milled for 30 min showed poorer performance compared with that milled for
60 min.
The selectivity of the optimal sensor was further explored by exposing it to 10 ppm
of various gases (C2H5OH, NH3, C6H6, and NO2) at 300 °C (Fig 10 (a)-(d)).
The sensor exhibited no meaningful response to these gases, demonstrating its selectivity
to H2S gas (Fig 10 (e)). Moreover, we studied the sensing behavior in the presence of 60% relative humidity
(RH) and 10 ppm H2S gas at 300 °C (Fig 11).
The sensor response decreased by more than 30% in the humid environment (60% RH).
This decrease is expected as water molecules occupy the available adsorption sites
on the sensor surface, reducing the available sites for adsorption of H2S gas. Consequently, the amount of adsorbed H2S gas is lower, resulting in a diminished response in a humid atmosphere [39].
3.3. Gas sensing mechanism
Initially, oxygen molecules are adsorbed on the sensor, and due to their high electrophilic
nature, they extract the electrons from the conduction band of the SnO2 [40]:
Hence, an electron depletion layer (EDL) forms on SnO2 in air. Upon exposure to H2S gas, it reacts with adsorbed oxygen species, releasing electrons to the surface
of the gas sensor. The potential reaction is [41]
This reaction narrows the EDL, resulting in a decrease in the sensor resistance. Additionally,
in the contact areas between the SnO2 NPs, double Schottky barriers form in air. The effect of these barriers on the resistance
of the sensor can be expressed as follows [42]:
where R represents the sensor resistance; Ra is the initial resistance; qv represents the barrier height; and k and T denote the Boltzmann constant and temperature, respectively. Upon exposure to H2S gas, the released electrons return to the sensor surface, reducing the height of
the double Schottky barriers, contributing to the sensing signal.
The selectivity to H2S gas is attributable to the working temperature, the higher reactivity of H2S relative to other gases, and the small bond energy of H–SH. The bond energy of 381
kJ/mol for H–SH in H2S can be easily broken to react with the oxygen ions [43,44].
The enhanced response of the sensor fabricated from SnO2 particles milled for 60 min compared with other gas sensors is attributable to both
smaller particle sizes and a higher total amount of adsorbed oxygen and oxygen vacancies
on these powders. From a particle size point of view, we know that increasing milling
time transfers more energy to the particles, and more grinding and wearing of particles
occurs, leading to the formation of finer particles. However, when the particles become
very fine their surface reactivity significantly increases, leading to the agglomeration
of particles and the formation of larger particles.
Therefore, it seems that in the present study, the optimal milling time is 60 min,
which results in the formation of finer SnO2 particles. However, by further increasing milling time to 90 min, the agglomeration
of fine particles is likely, contributing to the generation of larger particles. Smaller
particles increase the contact areas among the powders, leading to higher resistance
modulation for the sensor fabricated from SnO2 particles milled for 60 min.
Additionally, based on the total amount of adsorbed oxygen and oxygen vacancies, the
presence of more adsorbed oxygen species leads to greater sensing reactions with H2S gas, thereby producing a higher response. Oxygen vacancies serve as favorable sites
for the adsorption of oxygen gas at the sensing temperature [45, 46], which increases the amount of adsorbed oxygen ions and enhances the sensor response.
Therefore, it seems that for the optimal sensor, the presence of high amounts of adsorbed
oxygen species and oxygen vacancies is effective for sensing performance towards H2S gas.