Outstanding Room-Temperature Hydrogen Gas Detection by Plasma-Assisted and Graphene-Functionalized Core-Shell Assembly of SnO2 Nanoburflower.

Here, we have reported the synthesis of three-dimensional, mesoporous, nano-SnO2 cores encapsulated in nonstoichiometric SnO2 shells grown by chemical as well as physical synthesis procedures such as plasma-enhanced chemical vapor deposition, followed by functionalization with reduced graphene oxide (rGO) on the surface. The main motif to fabricate such morphology, i.e., core-shell assembly of burflower-like SnO2 nanobid is to distinguish gases quantitatively at reduced operating temperatures. Electrochemical results reveal that rGO anchored on SnO2 surface offers excellent gas detection performances at room temperature. It exhibits outstanding H2 selectivity through a wide range, from ∼10 ppm to 1 vol %, with very little cross-sensitivity against other similar types of reducing gases. Good recovery as well as prompt responses also added flair in its quality due to the highly mesoporous architecture. Without using any expensive dopant/catalyst/filler or any special class of surfactants, these unique SnO2 mesoporous nanostructures have exhibited exceptional gas sensing performances at room temperature and are thus helpful to fabricate sensing devices in most cost-effective and eco-friendly manner.

Introduction

To enlighten our surroundings, we have used different fuels like coal, petroleum, natural gas, etc. H2 is the latest in the succession of fuels with many socioeconomic advantages in its favor. But for practical use, it has high risk due to its combustible nature in aerial medium as it is a colorless, odorless, and tasteless gas, and not detectable by human sensorium. For this reason, deployment of H2 sensor is indispensable, where H2 is produced or used as fuel.

Significant developments have been made in the area of metal oxide-based semiconducting ceramic gas sensors since their invention. Currently, these sensors are extensively used for domestic, industrial, and environmental applications.[1−7] These are the simple type of gas sensors possessing superiority due to compact size, simple fabrication technique, durability, low cost, minimal power consumption, and simple electronic circuits.[1,2] In 1962, Taguchi was the first inventor to patent the capability of SnO2 in detecting low concentration of combustible and reducing gases by measuring their change in resistance on exposure to such gases.[4] Although SnO2-based thick-film gas sensors have been available commercially for a long time, their performances such as high sensitivity, preferential selectivity, response and recovery times, and durability still need further improvements especially in low- or room-temperature (RT) region. Hence, the development of SnO2-based gas sensors with optimum sensitivity at room temperature and selectivity toward a particular gas has been a major challenge in the field of semiconductor sensors in recent years. To tackle this problem, researchers worldwide have concentrated on the synthesis technique of SnO2 that can solve this delinquent during the grass route level without adding any expensive chemicals as catalyst/dopant/filler/modifier. Various morphologies of nano-SnO2 have been exploited, such as nanowires (NWs), hollow spheres, nanocrystals, etc., for selective or multiple detection of chemicals, where, in almost all cases, it has been kept in mind to enhance the effective surface area along with the other qualities, since sensitivity is primarily a surface-induced phenomenon. Wang et al.[8] exploited SnO2 NW-based sensors, which can detect hydrogen (H2) for a wide range of concentrations (10–1000 ppm), attributed to the undercoordinated atoms on the SnO2 NW surfaces. So, they focused on the material morphology to widen the detection concentration limit. In a comparative study, Brunet and his group[9] concluded that SnO2 NW sensors have higher detection efficiencies, i.e., more than 30 times than the SnO2 thin-film sensors toward CO, CH4, H2, CO2, SO2, and H2S, attributed to a lack of grain boundaries. Hence, they also focused on structural and morphological aspects to escalate the conventional sensing limit. Banarjee and her group[10] claimed about 91% sensitivity toward 1000 ppm n-butane with a recovery time of 15 s, using SnO2 nanoparticles (NPs) synthesized by the spontaneous gel autocombustion (Petchini) technique. By this way, they concentrated on the synthesis and fabrication procedure to overcome decade-long sensing drawbacks like optimum sensitivity with prompt recovery of their fabricated prototype.

With the recent advancement in nanotechnology, plasma-assisted nanofabrication has become an exciting new direction because plasma-based approaches can deliver unique predefined structures at the nanoscale level that cannot be achieved by other techniques in a more economical and environment-friendly manner. Over the last decade, there have been exiting breakthroughs in the utilization of plasma processes in the fabrication of a rich diversity of nanomaterials (nanowires, nanotubes, nanobelts, nanoparticles, etc.) and nanoengineered coatings. Some of these materials can simply not be derived by conventional means while many others have remarkable plasma-induced properties setting them apart from their conventional counterpart. In this regard, it is worth mentioning that plasma-enhanced chemical vapor deposition (PECVD) is an appealing technique to develop extremely controllable multifunctional oxide nanoarchitectures under relatively mild conditions, owing to the unique features and activation mechanisms of nonequilibrium plasma. Inspired by mother nature, the potential of plasma bombardment has been exploited in our study to synthesize tailor-made controlled structure of a naturally occurring flower (burflower), to increase the surface-to-volume ratio since sensitivity is a surface phenomenon.

In this work, SnO2 core–shell assembly has been synthesized using chemical (wet chemical synthesis) and physical (plasma bombardment technique) methods to fabricate solid-state gas sensors focusing on its response toward hydrogen gas in the ppm level to as high as volume percentage level. It is well known that this is a highly challenging task to fabricate semiconductor metal oxide (SMO)-based sensor devices (e.g., SnO2), which exhibit high sensitivity and prolonged stability at the same time and above all perform selective detection of a particular gas even at room temperature. Besides examining the interrelations between the material properties and the synthesis conditions, special focus is given to their emerging application for a practical need of room-temperature hydrogen gas detection for a wide range of concentrations. Here, we have proposed a novel H2 sensing material, which is basically a modified SnO2 pseudosphere consisting of several biddings on the surface of the SnO2 base substrates, which eventually appeared as “kadamba” flower (burflower).

Experimental Section

Materials and Methods

The synthesis technique of SnO2 nanoparticles includes chemical and physical routes, with the former involving the wet chemical technique followed by calcination for phase purification, and the latter involving deposition by e-beam and multiple plasma treatments (viz., H2 and Ar) with various operational parameters to fabricate nanoparticles with desired shape, size, and morphology.

Wet Chemical Synthesis

Calculated amount of powdered dihydrated stannous chloride (SnCl2·2H2O) is added to 250 mL of deionized (DI) water in a beaker to make the solution of 0.01 M in slightly acidic medium (in the presence of 2–3 mL of dil. HCl). Liquor ammonia (NH4OH, 30% (v/v)) is added to the solution dropwise until precipitation, which was confirmed by achieving pH 9.[11,12] The solution was then centrifuged at 4000 rpm to separate out the solid portion (precipitate) from the liquid. The solid part was collected and washed with DI water and acetone sequentially to remove impurities present, if any. Finally, it was collected in a Petri dish and then vacuum-dried at room temperature.

Pellet Formation

SnO2 nanoparticles synthesized by the wet chemical method were used to form pellets with anticipated diameter, which will be used as a target material in an electron beam deposition chamber. Granular poly(vinyl alcohol) (10 wt % (with respect to SnO2)) was used as binder in pellet formation to restrict cracking and breakage of pellets. SnO2 pellets of desired thickness and diameter were fabricated using high-pressure (1.5 Torr) hydraulic press (Polyhydron, Belgaum) under atmospheric temperature (∼27 °C). The pellets were placed in a programmable tube furnace (Nuskar, Kolkata) for calcination at 600 °C with 6 h soaking time to get phase pure material (SnO2) and for the removal of binder used.

E-Beam Deposition of SnO2 Nanoparticles

SnO2 pellets were used as a target material in an e-beam chamber (AUTO-500 HHV) having graphite crucible holder. A SnO2 thin film (25 nm) was deposited as a conformal layer on the surface of a polished Si wafer, where the substrate temperature was kept at 200 °C. The deposition was done in the presence of oxygen gas (purity, 99.999%) with a deposition rate of 3 Å/s, and the system vacuum was maintained at 1.04 × 10–4 mbar.

H2 Plasma in PECVD Unit

H2 plasma treatment was done in PECVD cluster tool (CT-100 HHV) on SnO2-coated polished Si wafer by maintaining the base vacuum of the system at 5 × 10–7 Torr. After achieving the equilibrium temperature of the PECVD cluster tool at 300 °C, H2 gas (purity, 99.999%) was purged into the chamber and the process pressure was maintained at 1 Torr. H2 gas flow rate of 100 sccm was controlled by a mass flow controller (MFC, Bronkhorst). Power density and plasma processing time were varied from 70 to 110 W/cm2 and 20 to 30 min to develop SnO2 nanoparticles of desired dimensions.

Bidding (Overgrowth) on SnO2 Nanoparticle Using Ar Plasma Treatment

Ar plasma treatment was performed by reactive ion etching (RIE) unit (ION ETCH 150-HHV), where the base vacuum pressure of the system was maintained at 1 Torr. The power of Ar plasma projected on the sample was 100 W with average deflection of 20 W inside the vacuum chamber. Argon gas (purity, 99.999%) flow was controlled at 100 sccm by MFC (Bronkhorst), and an inert atmosphere was maintained inside the chamber. The time of plasma treatment on the sample was varied from 10 to 20 min to grow the desired bidding (overgrowth) of SnO2 over SnO2 nanoparticles (self-cloning) by surface etching of the second one (Figure ).

Figure 1

Schematic showing the stepwise formation of SnO2 bidding over SnO2 nanoparticles along with graphene-coated nanocrystalline SnO2 bids.

Introduction of Reduced Graphene Oxide (rGO) on the Surface of SnO2

Graphene or reduced graphene oxide was synthesized through wet chemical bottom-up approach (modified Hummers’ method) with graphene oxide as the intermediate product, which was reduced by hydrazine hydrate to form amine-functionalized rGO, which may act as an n-type semiconductor[13] due to the presence of primary amine group (−NH2) at the edge of rGO as well as on the basal plane (very less amount) and form an inner armchair (zigzag) structure.[14] A thin rGO layer was developed on the surface of bided SnO2 by capillary action. It can be assumed that the highly ordered and repetitive patterned mesoporous tiny particles of SnO2 as well as the spaces between the particles behave like a capillary, which supports capillary action to cover the surface with a thin layer of graphene, which was suspended in isopropyl alcohol (IPA). Slightly increased temperature (∼45 °C), considerably low-dimensional capillary tube diameter (channel diameters, ca. 1–2 nm), and comparatively low surface tension of IPA (21.70 ± 0.05 mN/m at room temperature, while the same for water is 71.99 ± 0.05 mN/m) are responsible to support this action. Low density of the solvent also helps the phenomenon (IPA has a density of 0.790 g/mL at RT, whereas the same for water is 0.998 g/mL at the same temperature). However, it has been seen that coating continuity is more uniform in the lower portion of the substrate probably due to the action of gravity.

Sample(s) Nomenclature

Powdered SnO2 synthesized by the wet chemical (precipitation) route is named as “sample 1” (S1). This sample was used to form SnO2 tablet at a high pressure, which has been used as a target material to grow conformal thin film of thickness ∼25 nm deposited on a polished silicon wafer by e-beam deposition technique and named as “sample 2” (S2). Sample “S2” was used in PECVD cluster tool as a substrate under H2 plasma with different operating powers for 5 min. Samples in H2 plasma operated with 3, 5, and 7 W consumption power are named as “sample 3” (S3), “sample 4” (S4), and “sample 5” (S5), respectively. Sample S2 was used in PECVD cluster tool again as a substrate under an operating power of 5 W for different operating times. Sample S2 that underwent a 5 W H2 plasma treatment for 10, 15, and 20 min is named as “sample 6” (S6), “sample 7” (S7), and “sample 8” (S8), respectively. On the basis of particle’s shape, size, and morphology, sample “S7” has been chosen as the optimum H2 plasma-treated sample (through the reactive ion etching method). This sample 7 was then further treated with Ar plasma for different times under 1 Torr base pressure with 100 sccm Ar flow rate inside a vacuum chamber with operating power of 100 W. For the sake of simplicity and to reduce the total number of samples, 25, 50, 75, 100, and 125 W Ar plasma treatments were also performed on sample S7, among which, the 100 W treatment was chosen as the optimum one. Now, sample S7 treated in Ar plasma with 100 W for 10, 15, and 20 min operating time is named as “sample 9” (S9), “sample 10” (S10), and “sample 11” (S11), respectively. With preferred growth of bidding (overgrowth) on SnO2 surface, sample “S10” was chosen as the optimum one and is collected for graphene (rGO) coating treatment, and the coated sample is named as “sample 12” (S12).

Results

To confirm the phase purity, orientations of crystal lattice structures, and probable lattice plains of the synthesized materials, room-temperature X-ray diffraction study was performed in RIGAKU Ultima IV, and the corresponding peak positions along with the detail explanations are provided in the Supporting Information (SI). From the X-ray diffractograms, “lattice strain” was calculated using eq ,[15] where ∈ and η are the effective particle size and effective strain, respectively.The effective particle size (∈) can be estimated by plotting (β cos θ)/λ as Y axis against (sin θ/λ) as X axis (shown in SI), where λ = 1.541 Å (Cu Kα1 X-ray wavelength), “θ” is Bragg’s diffraction angle, and “β” is the angular full width at half-maximum (FWHM) intensity. The effective strain (η) may be induced in particles due to the prolonged plasma treatment with variable power. Degree of strain induced in the particle due to the size reduction has been estimated from the slope of the strain graphs (elaborated in SI).[16] Some physical properties of the synthesized SnO2 nanoparticles (NP) derived from X-ray diffractogram are shown in Table .

Table 1

Physical Properties of Synthesized SnO2 NP Derived from X-ray Diffractogram

sample no.	d-spacing value (Å)	crystallite size; DXRD (nm)	lattice parameter (Å)	c/a ratio	FWHM (2θ)	lattice strain (β cos θ)/λ)
S1	1.75	8.42	a = b = 4.525, c = 3.225	0.7127	0.45	0.0042
S2	1.72	1.50	a = b = 4.928, c = 2.997	0.6082	1.05	0.0082
S3	1.71	1.48	a = b = 4.814, c = 3.005	0.6242	1.01	0.0078
S4	1.72	1.49	a = b = 4.794, c = 3.125	0.6519	0.96	0.0066
S5	1.74	1.98	a = b = 4.778, c = 3.148	0.6589	0.81	0.0058
S6	1.75	2.62	a = b = 4.752, c = 3.172	0.6675	0.68	0.0052
S7	1.76	3.12	a = b = 4.737, c = 3.184	0.6722	0.56	0.0048
S8	1.77	4.08	a = b = 4.718, c = 3.206	0.6795	0.44	0.0040
S9	1.76	2.07	a = b = 4.721, c = 3.173	0.6721	0.81	0.0069
S10	1.771	2.76	a = b = 4.728, c = 3.178	0.6725	0.93	0.8100
S11	1.77	3.16	a = b = 4.734, c = 3.181	0.6719	0.90	0.0045
S12	1.77	2.76	a = b = 4.732, c = 3.183	0.6726	0.88	0.0080
JCPDS (77-0447)	1.76		a = b = 4.735, c = 3.185	0.6726

The interrelations between the material properties, structural orientations, and the synthesis conditions have further been corroborated using Fourier transform infrared (FTIR) spectroscopy, Raman spectroscopy, and X-ray photoelectron spectroscopy (XPS), as presented in the Supporting Information (SI).

Field Emission Scanning Electron Microscopy (FESEM) Studies

Figure depicts the FESEM images of samples S1–S12, along with the particle size distribution (PSD, where possible) with normalized Gaussian distribution function, showing the mean size of the particles. The morphologies of the synthesized nanoparticles were visualized on a field emission scanning electron microscope (SIGMA ZEISS) at 5 kV electron high tension using a high efficiency in-lens detector, having base pressure of the vacuum chamber in the range of 3 × 10–6 mbar (gun vacuum in the range of 1.95 × 10–9 mbar). However, in this technique, samples can potentially sublimate under high-vacuum conditions, thus increasing the risk of charging. To restrict this tendency, an aluminum (Al) foil tape was attached with the stage of each sample for discharging, if necessary. The micrographs behave as the footprint of step-by-step synthesis of the desired material in the form of nanobid composite. From Figure , it could be seen that S1 possesses nanoparticles having nonuniform diameter distribution in the range of ca. 20–200 nm. The shape of the particles includes rhombohedra-like, elongated spherical, quasi-spherical, as well as perfect spherical structures. The presence of such mixed shapes can probably be explained in terms of (i) interrupted growth of nanoparticles, (ii) overgrowth, (iii) incomplete lattice formation, (iv) introduction of defects (atomic voids, excess oxygen-ion incorporation, excess Sn-ion incorporation, interstitial defects, etc.), (v) agglomeration, (vi) recrystallization, and (vii) strain-induced formation of nanocrystal/particles. Such mixed-shape and mixed-phase SnO2 was used to generate uniform SnO2 pallet and was used as a target material in an electron beam deposition chamber to deposit SnO2 on a polished silicon wafer (S2). The detailed condition of deposition is discussed in the Materials and Methodssection. For the sample S2, thickness of the deposited thin film (conformal layer) of SnO2 was ∼25 nm, measured by the thickness profilometer (DektakXT; Bruker), and is presented as inset of S2 of Figure . To grow SnO2 nanoparticles from the conformal coating, hydrogen plasma treatment was performed in PECVD cluster tool (HHV CT-100), keeping the other conditions constant, viz., for all of the cases: (i) maintaining base pressure of the chamber at 1 Torr; (ii) keeping the temperature of the substrates at 300 °C; (iii) keeping the working distance of all at 7 cm, etc. Shape, size, morphologies, and distributions of the as-grown SnO2 nanoparticles on the polished silicon wafer substrate were studied for plasma treatment powers of 3, 5, and 7 W for 5 min and are displayed in Figure as S3, S4, and S5, respectively. The thickness of the film of samples S2, S3, and S5 was measured using a surface profilometer. The measured thickness of the film of sample S3 is ∼24 nm, which is almost identical to that of sample S2, supporting an early initiation of plasma-dependent nanoparticle growth. Decrement of film thickness can be a result of etching of SnO2 surface by hydrogen plasma at higher operating power. The SnO2 film thickness of the sample S5 is found to be roughly ∼38 nm, shown in the inset of the same figure. From the figure, it can also be found that the surface of the SnO2 is rough. This roughness of the surface is an effect of overetching of SnO2 due to prolonged plasma exposure. From the morphological point of view, it could be seen that for sample S3, nanoparticle formation started randomly. Higher operating power (5 W) was provided for 5 min to optimize the formation of nanoparticles from SnO2 thin film, and has been assigned as S4. From the micrograph of S4 sample, a continuous layer of SnO2 nanoparticles could be seen with uniform particle shape and size distribution (∼25 nm), estimated using Java-based ImageJ software (developed by National Institutes of Health) and presented as inset particle size distribution (PSD) histogram (with std. error of ±3 nm). But when moved on to sample S5, the particle deformed with uneven distributions. This is also evident from the surface profilometer study. From these micrograph studies, 5 W operating power of PECVD cluster tool was opted as the optimum power for nanoparticle growth formation.

Figure 2

FESEM microstructures of S1–S12 samples (with 200 nm bar) with corresponding particle size distributions histogram and surface profile study (whichever appropriate).

The operating plasma treatment time was changed further, keeping the power same (5 W), to realize further probable changes in the formation of SnO2 nanoparticle. In doing so, H2 plasma treatment time was chosen as 10, 15, and 20 min, and the corresponding micrographs are presented as S6, S7, and S8 in Figure . Gradual particle growth could be observed in S6, S7, and S8. For S6, a uniform spherical growth of SnO2 nanoparticle could be seen with a dense matrix having average particle diameter of ∼50 nm (evident from PSD graph, presented as inset in S6). The particle diameter increased to 100 and 150 nm for S7 and S8, respectively, with gradual increment of plasma treatment time from 15 to 20 min. Although the particle size is increased, the size, shape, and distribution of sample S8 deformed and the spherical shape became quasi-spherical with discontinued base, due to complete etching of the conformal base layer of SnO2 through H2 plasma treatment. For this reason, time progression study of the H2 plasma treatment was stopped at 20 min and no further study was performed with higher time of plasma treatment. The deformation tendency of sample S8 could be a result of exploiting higher power of hydrogen plasma (overplasma growth), which are not suitable for the size range of base particles. On the basis of these FESEM images, 15 min H2 plasma-treated sample (S7) was opted for further study. Ar plasma bombardment was especially performed to develop overgrowth (bidding) on the outer surface of SnO2 nanoparticles for the sake of surface area enhancement. S7 sample was further treated with 100 W Ar plasma for 10, 15, and 20 min time intervals, and subsequent micrographs are presented as S9, S10, and S11, respectively, along with their respective particle size distributions in Figure . The best shape was found in S10 sample with preferred growth of bidding (overgrowth) on SnO2 surface that appeared just like natural kadamba (burflower) with numerous tiny overgrowths, which would ultimately facilitate the sensing phenomenon. S12 represents the micrograph of n-type rGO-coated SnO2 NP in the same magnification range, where both the thickness of the rGO layer and the SnO2 particle distribution can be visualized.

Transmission Electron Microscopy (TEM) Studies

The particle size, morphology, and local crystallographic structures were studied by transmission electron microscopy (TEM) and high-resolution transmission electron microscopy (HRTEM), and the related micrographs are presented in Figure . The bright-field images help us to get an idea about the effective particle sizes of the different SnO2 nanoparticles. Figure b depicts the bright-field TEM image of wet chemically synthesized SnO2 nanoparticles (S1), which has been tinged (brownish) to enhance visualization. Figure a is its corresponding high-resolution (HRTEM) counterpart. The TEM image exhibits round particles of different sizes and shapes, and corroborates well with FESEM images. The HRTEM image clearly shows spacing between lattice fringes corresponding to the (200) plane orientation of SnO2 according to JCPDS card no. 77-0447. Figure c indicates the corresponding selected area electron diffraction (SAED) patterns with probable indexed planes of SnO2. Figure d,f depicts the micrographs of SnO2 nanoparticle prepared by H2 plasma treatment (S7). Here also a similar pattern of representation has been followed, i.e., Figure e is the bright-field TEM image, Figure d is the HRTEM image, and Figure f is the corresponding SAED crystallographic orientation of S7 sample, with the lowest crystallite size among the synthesized SnO2s. From this series of TEM images, the well-dispersed spherical particles having diameter in the order of 80–90 nm size can be seen with relatively less degree of agglomeration as the particles retain their boundary very precisely from one another.

Figure 3

Transmission electron micrography (TEM) images and their related images (like HR and demarcated SAED patterns) of wet chemically synthesized SnO2 nanoparticles (S1) illustrated as (a)–(c); H2 plasma-treated SnO2 nanoparticles (S7) illustrated as (d)–(f); Ar plasma-treated SnO2 bidding sample (S10) illustrated as (g)–(i); and rGO-coated bided sample (S12) illustrated as (j)–(l).

Figure g–i represents TEM images and their related images of Ar plasma-treated SnO2 bidding sample (S10). From Figure h, two parts of the image can be clearly identified, in which the denser and darker part has been assigned as “core” and lighter outer part is dispensed as “shell” originated probably by SnO2 nanoparticles formed by H2 plasma treatment and SnO2 bidding formed by Ar plasma treatment, respectively. The core part has a diameter of nearly 50 nm, and the shell part has a thickness of about 20 nm. This core–shell feature is also visible in its high-resolution counterpart as well. Therefore, these types of particles with average size of around 100 nm or lower can be thought of as a new type of core–shell structure that is yet to be reported by any group as this contains the same material (SnO2) as both core and shell components.

Generally, core–shell structures consist of two different compositions,[17−19] but here in this modified core–shell structure, it consists of materials with same compositions but with different stoichiometries and crystallographic orientations. With the gradual optimization of plasma packing from S7 to S10, the particle size increases although the degree of agglomeration decreases, as is evident from Figure e,h. Figure i confirms the SAED pattern of highly crystalline SnO2 with typical reflections of (110), (101), (111), (211) (220), and (310) crystal planes. The TEM image of rGO-coated sample (S12) with bids is demonstrated in Figure k, which clearly indicates the graphene layer coating on the outer surface of shell SnO2. In fact, this layer acts like a porous membrane and turns the particles as densely arranged black spheroid, which is well supported by the corresponding high-resolution HRTEM images. It is clear from the images that graphene is well dispersed and distributed throughout the SnO2 surface. In Figure j, a typical crystalline domain with interplanar spacings (d = 0.18 and 0.26 nm) corresponds to (211) and (101) reflections of SnO2 tetragonal rutile phase (JCPDS 77-0447), which confirms the presence of SnO2, and (002) reflection of carbon confirms the presence of finely dispersed rGO particles on the same matrix of SnO2. There is also a clear indication of a Moiré pattern in Figure j, which usually occurs due to the overlapping of crystal planes of either the same material or two different materials. The SAED pattern in Figure l taken from a representative area shows the presence of lattice fringes corresponding to (110), (101), (200), and (211) crystal planes of tetragonal SnO2. However, there is no indication of the presence of rGO in the SAED pattern probably because of its negligible amount compared to SnO2. Nevertheless, TEM images and the related images substantiate well with the results found from XRD studies.

Sensing Studies

Figure a,b and Tables –5 depict the sensing characteristics of SnO2 nanoparticles (NPs), SnO2 nanoparticles with bidding, and SnO2 nanoparticle-bidding coated with graphene (rGO) identified with sample numbers S1, S10, and S12, respectively. Figure a shows the dynamic response–recovery characteristics of S1, which exhibited n-type sensing characteristics with 5.27, 22.08, 48.82, 72.02, and 88.00% responses toward a wide range of H2 gas (viz., 10, 100, 1000, 10 000, and 100 000 ppm, respectively). Although it carried out satisfactory sensitivities only at and above 10 000 ppm H2 gas, it is not very much acceptable considering the practical applicability of a H2 gas sensor. On the contrary, the S10 sample exhibited excellent sensing appearances with almost full-scale response range with as high as 52.24, 68.78, 78.67, 91.0, and 99.89% sensitivities, respectively, for the same order of gradually increasing (of the order of 10-fold) five different H2 gas concentrations mentioned earlier. But when graphene (chemically synthesized, in the form of n-type reduced graphene oxide, or rGO) was introduced into the system as a thin surface layer, the sensing responses have been affected slightly. From Table , it could be seen that for the same concentration of same gas (H2), percentage of sensitivity increases drastically from S1 to S10, on the formation of tiny biddings on the SnO2 NP surfaces. The probable reason would be surface area enhancement due to overgrowth for bidding and consequent undulations leading to enhanced sensing response as sensing is primarily a surface phenomenon.

Figure 4

Combined dynamic response–recovery curves against five different concentrations of hydrogen gases (viz., 10, 100, 1000, 10 000, and 100 000 ppm; from left to right) at 250 °C operating temperate for the samples (a) S1 and (b) S10.

Table 2

Sensitivity against Different Concentrations of Hydrogen Gas at 250 °C Operating Temperature

material compositions	10 ppm H2	100 ppm H2	1000 ppm H2	10 000 ppm H2	100 000 ppm H2
SnO2 NP	5.27	22.08	48.82	72.02	88.00
SnO2 bidding	52.24	68.78	78.67	91.01	99.89
SnO2 bidding + rGO	37.77	51.12	63.39	75.50	90.00

Table 5

Cross-Sensitivities against Different n-Type Reducing Gases of Same Concentration at 250 °C Operating Temperature

material compositions	1000 ppm H2	1000 ppm methane	1000 ppm butane
SnO2 NP	48.82	22.44	14.02
SnO2 bidding	78.67	28.84	8.62
SnO2 bidding + rGO	63.39	25.56	5.66

However, sensitivity falls to some extent if rGO was also present on the surface. Probably, it formed a primary contact barrier between SnO2 NP and the analyte gases. But as the synthesized rGO was n-type in nature (evidence shown in the Supporting Information) and inherently possessed excellent room-temperature electrical conductivity[12] with low electrical noise due to its high degree of crystallinity (shown in the Supporting Information), sample S12 showed very good low temperature sensitivity even at room temperature (∼27 °C). To explore the effects of working temperature on sensing characteristics, all of the samples measured at different working temperatures and the measured sensing response data ranging from room temperature to 350 °C working temperature are shown in Table . The S1 sample exhibited sensitivities of 30.36, 72.02, 74.45, and 75.55% toward 1% H2 gas at 200, 250, 300, and 350 °C working temperatures, respectively. Below 200 °C, the sample (S1) was unable to show any sensitivity (resistance overload). Although SnO2 NP with bidding (sample S10) shows sensing response at 100 °C temperature also, the extent is quite poor (only ∼8%). From Table , it could be noted that only S12 sample could show room-temperature sensitivity and the values are considerably high enough (95 and 90.9% responses at room temperature and at 100 °C, respectively). Without rGO, SnO2 NP and/or SnO2 bidding did not show any room-temperature sensitivity; thus, it could be concluded that rGO is only responsible for low temperature sensing responses. However, sensitivity of S12 increases with increasing concentration of H2 gas (Table ). From Table , it could be claimed that all of the samples (S1 as well as S10 and S12) displayed fairly good selectivity toward hydrogen gas in comparison to the same concentration (1000 ppm) of some other n-type reducing gases like methane and butane. Some p-type oxidizing gases have also been reported in the literature[20−23] for similar kind of SnO2-based sensors, but the present study was limited to n-type reducing gases only for the sake of comparison with H2 (which is the main target gas in this study and is n-type reducing in nature). Above 250 °C operating temperature, SnO2 NP and SnO2 bidding responses increase only slightly, strictly speaking, sensitivity almost ceases. For this reason, sensing responses did not record above 350 °C working temperature. It should be mentioned here that at 250 °C working temperature, the base resistance of SnO2 NP (S1) was 180 MΩ and the same for SnO2 bidding (S10) and SnO2 bidding + rGO (S12) were 528 and 66 MΩ, respectively. The value for S10 sample was fairly explainable as with bidding/overgrowth formation, surface area increases, which in turn increases the extent of depletion region. Such enhanced depletion layer tried to encompass more ambient oxygen species, which has been reflected through higher base resistance or air resistance values compared to sample S1 comprising only SnO2 NP. On the contrary, when rGO was incorporated into the system, conductance of the system increased, as rGO is a good conductor at room temperature and it provides sufficient carrier electrons from its two-dimensional (2D) honeycomb lattice to reduce the resistance of the composite systems.

Table 3

Sensitivity at Different Operating Temperatures against 1% H2

material compositions	RT	100 °C	200 °C	250 °C	300 °C	350 °C
SnO2 NP			30.36	72.02	74.45	75.55
SnO2 bidding		8.22	68.82	91.01	93.34	94.68
SnO2 bidding + rGO	95.00	90.92	84.48	75.55	67.00	62.22

Table 4

Sensitivity against Different Concentrations of Hydrogen Gas at Room Temperature

material compositions	10 ppm H2	100 ppm H2	1000 ppm H2	10 000 ppm H2	100 000 ppm H2
SnO2 bidding + rGO	56.67	71.33	79.99	95.00	97.67

Figure a,b displays the combined dynamic response–recovery curves of sample S12 at different operating temperatures, e.g., Figure a at 250 °C operating temperature, and Figure b at room temperature. From the curves, it is evident that for both the cases, at room temperature and at elevated temperature (250 °C), S12 showed gradually improved sensitivity with increase of H2 gas concentration. However, room-temperature gas exposure exhibits better sensing responses than its elevated counterpart, which indicates that the sensor is best suited at room temperature. After reaching a saturation plateau, the dynamic curves switched from desorption mode to adsorption mode. Such steady states (plateau) indicate saturation of gas molecules at the depletion regions as well as the response time of the sensors.[24,25] From the curves, it could be seen that the responses are quick (within 20 s) even to a trace amount (0.001%, or 10 ppm) of H2 gas and the recovery is swift as the resistance output evenly comes back almost to its original level, suggesting highly reproducible and reversible response nature of the sensors against the analyte gas. However, the recovery time is not appreciably low probably because of slow oxidation of gas molecules on the surface of the sensors. It is noteworthy that a number of experiments have been carried out to measure the sensitivity as a function of operating temperature and time. In all of the cases, sensitivity of the sensor elements showed approximately constant values, indicating the repeatability and reproducibility nature of the sensors. To explore the reproducibility of the sensors and to examine their shelf-life, the following data have been acquired and are presented as Figure .

Figure 5

Combined dynamic response–recovery curves toward five different concentrations of hydrogen gases (viz., 10, 100, 1000, 10 000, and 100 000 ppm, from left to right) for the same sample (S12) at different operating temperatures, e.g., (a) 250 °C and (b) room temperature.

Figure 6

Dynamic response–recovery characteristics of S12 sample toward 100 000 ppm (10%) hydrogen gas measured at room temperature for a longer period of time (∼10 min).

The reproducible nature of S12 sensor sample was confirmed by its dynamic response–recovery features portrayed in Figure . It is evident that the S12 sample exhibited almost a similar kind of sensing rejoinders for the same gas exposure time (of ∼20 s) at room temperature for five consecutive cycles at a stretch, without showing any fatigue changes in its adsorption and/or desorption segment, leading to the evidence of its fast response and instant recovery of the sample with no drift.

Discussion

Working Principle

To understand the influence of surface properties on sensing performances, the mechanism of surface chemistry and lattice structure with stoichiometric defect of the synthesized components involved are to be understood primarily. Since gas sensing characteristics are primarily surface phenomena, the focus has been shifted to increase the surface area so that the surface-to-volume ratio can be enhanced. Inspired by mother nature, attempts to synthesize structures like naturally occurring burflower (kadamba), which has several spikelike bids (overgrowth), in turn would help increase the surface area of the synthesized material. The micrographs (FESEM and HRTEM) shown in Figure i portray the resemblance of the synthesized material with naturally occurring “burflower” and visualize the control over the synthesis parameters. The existence of a large number of bids on the SnO2 matrix has increased the percentage response to about 61% compared to only SnO2 NP (78.67% sensitivity of SnO2 with bids against 1000 ppm H2, measured at 250 °C operating temperature, and the same for SnO2 NP was only 48.82%, measured at similar conditions). Figure i illustrates a schematic depicting the surface chemistry mechanism of rGO-coated SnO2 in the presence of analyte gases like H2. When the sensor is exposed to hydrogen gas, it is expected to either remain as H2 (ads) on the surface or get dissociated into two H atoms. These adsorbed H atoms will interact with the active oxygen species on surface, as shown in eqs and 3Thus, the adsorption of O2– and O– ions on the nanocrystalline SnO2 surface is the key factor for enhancing the receptor function of the sensor, which in turn controls the response of the sensor. The greater the ability of the surface to oxidize the target gas, the higher could be the response of the sensor.

Figure 7

(i) (a) Real image of naturally occurring burflower and (b) its schematic impression illustrating the resemblance of the synthesized material with their electron micrographs [(c) FESEM and (d) HRTEM]. (ii) Proposed schematic surface chemistry mechanism of rGO-coated SnO2 in the presence of analyte gases.

Synthesized rGO, being n-type in nature (established through hot probe test[24,26] and elaborated in SI), facilitate the desired molecular exchange(s) on the sensor surface since SnO2 is also n-type intrinsically, leading to fortify the sensing performances by catalyzing interfacial reactions (Figure ii). Moreover, rGO possesses excellent electrical conductivity at room temperature, and it has also been seen that extremely small change in the resistance of a graphene sheet caused by gas adsorption even down to the molecular level is detectable. For these reasons, the presence of graphene enables the prototype to detect gases at low temperature, even down to room temperature.

Electronic Band Structure

The overall working principle of the gas sensing mechanism of SnO2 NP bidding and the same SnO2 NP with rGO (composite) at room temperature as well as at elevated temperature is illustrated in Figure . It is assumed that the total conductivity (σTotal) of a crystalline semiconductor is the sum of electronic (σe and σh) and ionic conductivities (σion), where subscripts “e” and “h” represent electrons and holes, respectively. Metal oxide semiconductors (MOSs) like SnO2 usually best operated at temperatures between 200 and 450 °C.[14,27,28] In this operating temperature range, the contribution from ionic conduction is negligibly low. So, the total conductivity of SnO2 can be assumed according to eq as follows[28]The resistance of the homogeneous bulk material (Rb) with bulk conductivity σbulk, mobility μ, length l, and cross section A can be calculated according to eq as follows[28]where the bulk conductivity (σbulk) isHere, “b” and “d” are the length and breadth of the material, respectively; “n” and “q” are the magnitudes of the free electron and hole (charge carrier) concentrations (i.e., the number of electrons or holes per cubic meter) for an intrinsic semiconductor like SnO2, respectively; and “e” is the charge of an electron (1.6 × 10–19 C). The n-type behavior of SnO2 associated with the oxygen deficiency in the bulk could be understood from Figure a. The donors are singly negative and doubly ionized oxygen vacancies (like O2– or O=) with donor levels DL1 and DL2 located around 0.03 and 0.15 eV below the conduction band edge.[25] In the case of SnO2, the extrinsic donors are multistep donors.[29] Therefore, donor and acceptor energy level concentrations and the operating temperatures determine the bulk conductivity of SnO2. In the typical temperature range for sensor operation (200–450 °C), the donors can be considered completely ionized and adsorbed with the semiconductor surface predominantly through two basic approaches, namely, physisorption and chemisorption. “Physisorption” is the process of adsorption with the least possible interaction with no charge transfer. “Chemisorption” is based on stronger forces and hence is connected with an electron transfer between the adsorbent and adsorbate and ultimately provides a band gap (Eg) of 3.6 eV for SnO2 (at standard condition). Hence, physisorption is associated with a neutral state where gas molecules condense on the metal oxide surface at low temperature while chemisorption takes place at temperatures higher than 150 °C, where the electron exchange takes place between adsorbed species and conduction band of SnO2 surface.[30] Ionosorption is basically “delocalized chemisorption” as the charge is transferred from/to the conduction band. In the presence of graphene (or reduced graphene oxide, rGO), however, these phenomena happen at much lower temperature, even at room temperature due to unique and excellent electrical properties of graphene at such temperature range. rGO possesses very high electron mobility at room temperature. Moreover, devices based on graphene are expected to have relatively low Johnson noise because of their high conductance and low crystal defect density. Although graphene can independently sense gas molecules in a quantitative manner,[12] here it has been incorporated only at the top of the samples as a surface layer to improve some SMO-based sensor drawbacks (like reduces operating temperature, enhances durability, etc.). rGO preserves the layer structures of the parent graphite, but the layers are buckled and the interlayer spacing is about 2 times larger (∼0.7 nm) than that of graphite.[31] Besides bridging oxygen atoms (oxygen epoxide groups), other functional groups are also found as carbonyl (C=O), hydroxyl (−OH), phenol (−C6H5), ammine (−NH4), etc.[12,26,32] It is actually an optically transparent multilayered film, which is impermeable under dry conditions. But when exposed to ambient moisture, they allow passage of molecules below a certain size. The films consist of millions of randomly stacked flakes, leaving nanosized capillaries between them, where ambient oxygen species may take shelter and develop a conduction band edge. When a large number of gas molecules were absorbed on the graphene surface, a transverse intermolecular force between the gas molecules would result, which is responsible for some changes to the microscopic corrugations of the adjacent sheets of graphene layers. Since the synthesized graphene is also n-type in nature, a significant decrease in the intersheet distance occurred due to intersheet electron tunneling effect (Figure b), which resulted in reduction of resistance, evident from the measured values of the pure and composite samples of SnO2 (at 250 °C operating temperature, base resistances of SnO2 NP, SnO2 bidding, and SnO2 bidding with rGO are 180, 528, and 66 MΩ, respectively). Such lowering of resistance also decreases the activation energy of sorption (physisorption and/or chemisorption) and hence the electrons of the conduction band are trapped creating a pretty good depletion layer (layer depth is denoted by the Debye length LD) leading to a decent and significant sensing output. In addition, due to having two-dimensional planner layers in it, graphene composite also increases the overall surface area, which ultimately facilitates the sensing phenomenon. However, after graphene incorporation, percent response was slightly affected due to the presence of 2D carbon layer in between semiconductor (SnO2) and the analyte gas(es), although it improves shelf-life of the sensor samples by developing a protective layer against corrosive environmental pollutants and surface contaminants.[33−35]Figure c illustrates that negative charges are built at the surface of SnO2 NP bidding due to sufficient adsorption of ambient oxygen species, which restricts further charge transfer across the surface. A band bending at equilibrium will influence the material resistance. Here, conducting electrons are represented by “e–”, and “+” represents the donor sites. During interaction, reducing gases release electrons into the surface of sensing materials and thus decrease the sensor resistance. If the surface-adsorbed species (Sad) possesses a dipole moment, the electron affinity (χ) gets changed.[36] Electron affinity and band bending influence the work function Φ of the sensing layer.[37−40] Here, it should be mentioned that the changes of the chemical potential have been considered to be negligible in the adsorption of the studied oxygen species. The H2 gas sensing mechanism in the case of an n-type semiconducting oxide like SnO2 involves two main surface chemical reactions. In the first reaction, the atmospheric oxygen gets ionosorbed on the metal oxide (SnO2) surface. This negative surface charge leads to an upward band bending of the conduction and valence bands resulting in an electron-depleted region.[41] This band bending produces an effective surface potential barrier. The height and depth of the band bending depend on the overall surface charge present from the beginning, which is determined by the amount and type of adsorbed oxygen, as can be seen in Figure c. The ionosorbed species act as electron acceptors due to their relative energetic position with respect to Fermi level (EF) as well as electron affinity (Figure a). Reactions of such oxygen species with analyte gases (reducing in nature) decreases the band tailing and ultimately restores it in the opposite direction with improved conductivity and less band bending. This change in resistance/conductance is measured through a source meter/multimeter and quantified as a percentage response or sensitivity with the conjecture that magnitude of the changes was proportional to the concentrations of analyte gases, which is believed to be the dominant sensing mechanism of surface-conductive gas sensors like SnO2. However, once the reducing analyte (H2) is removed, the surface is reoxidized by surrounding oxygen species, leading to the increase of resistance and return almost to its original level (reflected by the graphs shown in Figures , 6, and 7), indicating its highly reversible and reproducible nature. The reversible nature of these sensors also get fortified by thermodynamically stable irreversible reaction of the analyte gases (H2 in this case) to produce H2O molecules (shown in Figure c), which means environmentally favorable condition sustained with respect to positive entropy (ΔS) and negative enthalpy (ΔH).

Figure 8

(a) Schematic band diagram of the SnO2 bulk and two vacancy donor levels (DL1 and DL2). (b) Schematic diagram of structural band bending after chemisorptions of charged species (ionosorption of oxygen), where EV, EF, and EC designate energy levels of the valence band, Fermi level, and conduction band, respectively. (c) Schematic exhibiting hydrogen gas sensing mechanism by SnO2 nanoparticle overgrowth (bids).

It can be said that metal oxide semiconductors like SnO2 are normally high-band-gap oxides that have insulating properties, from which the semiconducting behavior originates due to deviation of stoichiometry, as depicted in eq using the Krögen–Vink notation[42]where subscript “o” defines the oxygen lattice position, “Vo” denotes oxygen (anion) vacancy, OoX represents a normal anion in an oxide with zero effective charge, e– is the charge of an electron, and “n” is the number of electrons (integer). Oxygen vacancies are generally singly (n = 1) or doubly (n = 2) ionized, depending on the temperature. This Krögen–Vink notation is a set of conventions that are used to describe electric charge and lattice position for point defect species to indicate various defect reactions. Therefore, bulk oxidation and reduction originates from surface lattice variations when it reacts with ambient oxygen. Here, the nonstoichiometry arises basically due to the oxygen vacancy, since oxygen atoms take positions at the interstitial places.

When SnO2 in the form of sample “S1” with SnO2−δ is coupled with highly nonstoichiometric sample S10 with SnO2−Δ, it forms an n–n-type homojunction.[43] In the same way, SnO2−Δ also forms an n–n-type heterojunction with rGO (which is deliberately synthesized as n-type)-coated SnO2 (sample “S12”). In such a junction, electron transfer may occur from semiconductors with low work function to semiconductors with high work function, until the Fermi levels (Ef) equalize. This creates an electron-depleted layer at the interface, which bends the energy band to some extent. The same logic is applicable for SnO2−Δ/rGO heterojunction, where the band tailing takes place to a substantial amount. The enhanced sensing performance of the composite of SnO2−δ/SnO2−Δ/rGO is attributed to the combined effect of the formation of depleted layer at the surface/interface of individual core–shell structure as well as the in-between graphene and SnO2 grains. Figure illuminates this idea in the form of a schematic presentation of the possible band bending by forming an “n–n+–n++”-type homo/heterojunction due to defect density or strain generated at the interface region. The formation of two depleted layers, one at the homojunction of SnO2−δ/SnO2−Δ core–shell interface and the other at the heterointerface of SnO2−Δ/rGO junction by adsorption of ambient oxygen species, promotes higher oxygen adsorption on the sensor surface to a greater extent, which might provide higher reaction sites.[44,45] Incorporation of rGO is responsible to reduce the base resistance (or resistance in air) of the sensor prototype, which may be considered as one of the major advantages of such combined heterojunctional composite material under investigation. For this reason, SnO2/rGO composite gas sensors can be measured at room temperature, which is rather impossible for only SnO2 sensors.

Figure 9

Schematic presentation showing defect-mediated homojunctional band bending at SnO2/SnO2−δ interface of core–shell nanoparticles of SnO2 and at the heteojunctional interface of SnO2/rGO (n-type) before (a) and after (b) gas sorption.

It can be seen from Figure that before any gaseous interaction at the homojunction of SnO2−δ/SnO2−Δ core–shell interface, a defect-mediated band bending takes place due to the presence of electron cloud, which arises mainly due to the large difference of electrons in the core (SnO2) and shell (SnO2−δ) materials (although the compositions of both the materials are same). The same assumption is also true for SnO2−Δ/rGO heterointerface. The presence of higher numbers of electrons shifts the Fermi level (Ef) nearer to the conduction band. After gas sorption, H2 molecules interact with the open surfaces of the nanocomposite and electronic exchanges take place, which reduce the density of the electronic cloud and minimize the barrier height at SnO2−δ/SnO2−Δ/rGO interfaces and channelize the electron flow toward the lower-energy states. The flow of electrons also minimizes band tailing, and for this reason, for n-type materials, resistance decreases after the gas exposure.

Plasma Effects

Plasma is an energy-rich hot ionized gas, consisting of approximately equal numbers of positively charged ions and negatively charged electrons. Plasma technology is versatile and capable of providing a large variety of processes. Among the plasma-assisted methods, three main approaches are especially promising for mild processing of nanomaterials: (i) plasma-enhanced chemical vapor deposition (PECVD), where gaseous molecular precursors are introduced into the plasma, whose role is either to activate the sole precursor or to avoid/enhance peculiar growth effects on the substrate surface. The potential utilization of metal catalyst nanoparticles can also be effective in this domain.[46,47] (ii) Sputtering, involving plasma bombardment of an externally biased target and the consequent ejection of neutral and charged species, enabling a high versatility in the preparation of supported nanoparticles or nanocomposites.[48,49] (iii) Plasma treatment of solid materials, based on the exploitation of bombardment leading to etching effects to modify the system surface, producing the tailor-made desired nanostructures.[50,51] However, there are some unresolved aspects in these mechanisms as the underlying elementary processes are yet to be known.

Here, among the three processes, the PECVD technique has been adopted for our sample preparations because PECVD is a state-of-the-art thin-film deposition technique that is widely known for semiconductor processing. In PECVD, the process material is delivered in the form of a special precursor gas that is broken down into a glow discharge (plasma), which transforms the gas mixture into reactive radicals/ions and other highly excited species interacting with the substrate. Depending on the nature of these interactions, either etching or deposition process occurs at the substrate. The desirable properties of this PECVD films are good adhesion, low pinhole density, good step coverage, and uniformity.[51] Moreover, the synthesized layer thickness can also be controlled in this technique. In general, using the PECVD method, the deposition, ablation, and surface modification processes are always concurrent and the predominance of one over the others has been attained in this work by a judicious choice of the experimental parameters, including pressure, power density, flow rates, reflected power, plasma composition, and their alterations. In this contribution, each parameter has been changed on an empirical basis to get the optimized conformal layer thickness with high-end purity.

Crystallite Size Effects

Sensitivity of thick- or thin-film gas sensor depends primarily on exposed surface area and hence on the particle size, shape, and their distribution and porosity of the film. All of these properties are directly related to the method of preparation of the samples. Thus, preparation techniques that can produce controlled size and shaped particles (like physically synthesized bottom-up approach as in the PECVD deposition method or top-down approach as in RI etching techniques) are of paramount importance in gas sensor material development, since wet chemical procedure has least control overgrowth dynamics. The gas response behavior of a semiconductor ceramic material mainly depends on grain boundary contacts, neck contacts, as well as inter- and intragrain contacts.[52] In most of the cases, they possess a wide range of mean crystallite size (D). Now if it is assumed that sensor elements are one-dimensional chain of SnO2 crystallites that are connected by substantial number of necks and less number of grain boundary contacts, then three possible situations may arise: (i) When D ≫ 2ts (where “t” is the thickness of the space charge layer or electron-depleted region), the electron conduction takes place in channels through necks. The grain boundary contacts share most of the electrical resistance of the chain and therefore the gas response is almost independent of D. (ii) In the cases where D becomes closer to 2ts, the necks become the most resistive part in the chain and control the gas response by neck control mechanism. (iii) When D < 2ts, each constituent of grain is depleted for conduction of electrons as a whole. Under this circumstance, grains share the dominant part of the resistance of the chain and control the gas response by grain control mechanism. The response in this region strongly depends on D, and therefore, the increase of gas response takes place with decrease in D.[52] When the grain or particle size gets reduced to nanometric dimensions, a dramatic improvement in the gas sensing properties takes place since a large fraction of the atoms are present at the surface or at the interface regions with structure and properties different from those of the bulk. The prominent effect of “nano” size, however, is associated with the thickness of the electron-depleted surface layer, which is defined as the Debye length “LD”. The Debye length LD for a semiconducting material could be calculated as[2]where “kB” is the Boltzmann constant; “ε” is the dielectric constant = εo × εr, where “εr” is the relative permittivity = ε/εo; “εo” is the permittivity of free space; “T” is the operating temperature in Kelvin scale; “e” is the electron charge (1.6 × 10–19 C); and “nd” is the carrier concentration.[53,54] It has been estimated by the researchers that LD for SnO2 is 3 nm with ε = 13.5, εo = 8.85 × 10–12 F/m, and nd = 3.6 × 1024 m–3.[3,8,25,53] Therefore, when the SnO2 particle size is reduced to a size that is comparable to or lower than 2LD, i.e., 6 nm, the whole crystallite will be fully depleted of electrons, which in turn causes the gas response of the element to change dramatically with D. Since LD for S10 sample is found to be 2.76 nm (overgrowth diameter), which is much lower than the critical size at which SnO2 could exhibit the “size-related nanoeffect”, its gas sensing response pitched to its highest level.

Conclusions

Three-dimensional, mesoporous, core–shell SnO2 nanobids have been synthesized by wet chemical as well as physical procedures, including plasma-enhanced chemical vapor deposition and reactive ion etching method to fabricate solid-state gas sensors. Experimental parameters have been altered judiciously to achieve tailor-made nanostructures, resulting in spectacular burflower-like structures originating due to screwlike overgrowth of SnO2. Such overgrowth or bidding can readily accommodate H2 molecules within its preferential structure with ready to adsorb and/or desorb properties and thus responsible for hydrogen selectivity for a wide range of gas concentrations (from 10 ppm to 1%) with very little cross-sensitivity against other similar types of gases. The synthesized core–shell assembly of the same material (SnO2) has been functionalized by introducing reduced graphene oxide (rGO) on its surface using modified thin-layer chromatography (capillary action) technique, aiming to be operative at a lower range of temperature, including room temperature. Hence, the synthesized SnO2 is a novel structure useful for selective H2 detection at room temperature for a wide range of gas concentrations and may be beneficial for industries like petroleum and chemical, coolant, welding, refining, metallurgical, glass, electronics, and so on. Besides using as reactant and reductant, H2 is also used as a green fuel, and development of such hydrogen-selective sensors operable at room temperature not only reduces the risk of its use, but also makes it user-friendly for common people and enhances socioeconomic advantages in its favor.