Effect of Ni, Pd, and Pt Nanoparticle Dispersion on Thick Films of TiO2 Nanotubes for Hydrogen Sensing: TEM and XPS Studies.

Crystal structure, morphological features, and hydrogen-sensing properties of thick film sensors of TiO2 nanotubes (NTs) impregnated with nanoparticles of elements of Group 10, viz., nickel, palladium, and platinum, having average grain size of about 25, 20, and 20 nm, respectively, are presented. The sensitivity is observed to be higher for Pd/TiO2 NTs than for Pt/TiO2 NTs. Ni/TiO2 NTs exhibited very poor sensitivity. X-ray photoelectron spectroscopy (XPS) studies confirm reduction of the oxide layer of palladium nanoparticles, which, in turn, is responsible for the generation of Ti3+ ion in TiO2 NTs through hydrogen spillover. For Pt/TiO2 NTs, only reduction of the oxide layer over Pt nanoparticles takes place without any spillover effect. For Ni/TiO2 NTs, neither NiO nor TiO2 undergoes any reduction. Changes in the Fermi level difference of PdO and TiO2 along with Ti3+ generation synergistically operate for Pd/TiO2 NTs, whereas the difference in Fermi levels of PtO and TiO2 alone operates for Pt/TiO2 NTs during sensing.

Introduction

Commercially available hydrogen (H2) sensors operate on different principles: combustion, pellister type, semiconductor type, electrochemical, fiber optic, and field-effect transistor, each of which is designed for a selected purpose.[1−5] Among them, semiconductor oxide-based sensors are simple, robust, and economically cheap. Different types of semiconductor oxides have been surveyed to exploit their surface conductivity for sensing a specific analyte. SnO2, TiO2, WO3, ZnO, etc., are some of the semiconductor oxides[6−8] that have been demonstrated for sensing H2. The gas–solid interaction that leads to changes in electrical conductivity is exploited in this type of sensors. For sensing trace levels of gases, the probability of chemical interaction needs to be maximized, which is achieved either by porous structures or in nanostructured thick/thin films as they offer high specific surface area with higher probability for interaction along with a reduction activation energy owing to high surface free energy of the nanoparticles.[9] The developments in nanoscience trigger the immediate application in the field of chemical sensors for trace level detection (ppm and ppb) of a variety of harmful, pollutant gases for environmental and process control. The wide band gap semiconductors with variety of donors and acceptors to alter the carrier concentration are desirable for this purpose. Moreover, the base semiconductor oxides should not form any irreversible permanent damage upon interaction with the gaseous species and based on the above considerations, the following materials, viz., SnO2, In2O3, and TiO2, are shortlisted and among them TiO2 is chosen (Eg: 3.4 eV) for our study. Generally, this type of sensor is not selective yet the kinetics of the specific interaction of the analyte with TiO2 could be made selective by manipulating the temperature. Although, sensor studies on nanocrystalline TiO2 are already reported, most of them are centered on the general response to gases without any specificity to an analyte.

First of all, for sensing H2, the strong covalent bond in H2 with the bond dissociation energy of 436 kJ mol–1 needs to be broken, which requires a very high temperature.[10] Hence, the first step would be to use an appropriate catalyst to break this bond at as low a temperature as possible. Elements of group 10 in the periodic table, viz., Ni, Pd, and Pt, or their alloys have been known for their catalytic hydrogenation reactions.[11] Their partially filled “d” orbitals play an important role in the hydrogenation reactions. The literature shows that H2 gets chemisorbed on these metals and dissociates into nascent hydrogen (highly reactive), which subsequently spills over to the neighborhoods to reduce them.[12] To improve the efficiency, nanoparticles of Ni, Pd, and Pt have been chosen in addition to TiO2 nanotubes (NTs) for this purpose. Therefore, it is proposed to exploit the catalytic role of nanocrystalline Ni, Pd, and Pt for gas-sensing behavior of TiO2 NTs. For this purpose, TiO2 NTs were impregnated with nanoparticles (NPs) of Ni, Pd, and Pt, respectively. The fabrication of these group metals-loaded TiO2 sensors to detect H2 in an Ar ambience is needed to monitor ppm to the percentage level of H2 in Ar cover gas to identify steam generation leaks in sodium-cooled fast breeder reactors during low operation or during shutdown or during startup of the reactor.[13,14] We report the results of our investigation in this paper.

Results and Discussion

Structural Studies

X-ray diffraction (XRD) pattern of TiO2 NTs (Table ) confirms the anatase phase as most of the reflections are indexed in terms of JCPDS file No: 89-4921 (Figure ). The additional peaks correspond to characteristic reflections of the respective metals, as marked in Figure . XRD data from Table affirms the face-centered cubic (FCC) structure of Pt, Pd, and Ni NPs dispersed on TiO2 NTs [JCPDS Nos. 87-0646, 89-4897 and 04-0850], respectively.[15−17]

Table 1

XRD Data of TiO2 NTs

2θ	25.35	37.67	47.71	53.74	54.84	62.49	70.03	75.11
peaks	(101)	(004)	(200)	(105)	(211)	(204)	(220)	(215)

Figure 1

XRD patterns of TiO2 NTs and their composites.

Table 2

XRD Data of TiO2 NT Composites

2θ (Pt)	peaks (Pt)	2θ (Pd)	peaks (Pd)	2θ (Ni)	peaks (Ni)
39.73	(111)	39.67	(111)	45.41	(111)
46.28	(200)	46.28	(200)	51.82	(200)
67.80	(220)	68.61	(220)	77.49	(220)

Selected area electron diffraction (SAED) pattern discloses the polycrystalline nature of pristine TiO2 NTs (Figure a) as well as nanocomposites (Figure b,c). The ring patterns of Pt and Pd NPs in their respective SAED micrographs of the composites show that they are crystalline, as observed in Figure b,c. The SAED patterns of the nanocomposites reveal that the NPs are dispersed on the surface of TiO2 NTs only as there was no change in lattice parameters of TiO2.

Figure 2

SAED patterns of (a) TiO2 NTs, (b) Pt/TiO2 NTs, and (c) Pd/TiO2 NTs.

Vibrational and Thermal Studies

The Raman spectrum of the TiO2 NTs shows six characteristic Raman peaks at wavenumbers 146 (Eg), 199 (Eg), 397.5 (B1g), 510 (A1g), 515.1 (B1g), and 639.2 (Eg) cm–1 for TiO2 NTs[18,19] and all of them characteristic of the tetragonal anatase phase, are indexed to RRUFF id: R120064, as shown in Figure a. Further, the broadened peaks signify the nanoparticulate nature of the TiO2 NTs. On the other hand, the Fourier transform infrared (FT-IR) spectrum of TiO2 NTs shows three major peaks at 454, 1622, and 3382 cm–1 characteristic peaks of TiO6 octahedral, deformation mode of chemisorbed water, and stretching mode of hydroxyl group, respectively (Figure b). The bending and stretching vibration of chemisorbed moisture are observed at 1622 and 3382 cm–1.[20] Thermal history of TiO2 NTs at higher temperatures was studied using thermogravimetry–differential thermal analysis (TG–DTA) (Figure c). Endothermic events marked by gradual weight losses were noticed at 135 and 225 °C, which are attributed to loss of adsorbed moisture and chemisorbed hydroxyl groups, whereas the exothermic event at 632 °C without the weight loss is characteristic of phase transformation of anatase TiO2 to rutile TiO2.[7][7]

Figure 3

(a) Raman spectrum, (b) FT-IR spectrum, and (c) TG–DTA curves of TiO2 NTs.

Morphological Studies

The field emission scanning electron microscope (FESEM) micrographs of TiO2 NT powder and its composites obtained through the rapid breakdown anodization (RBA) technique are shown in Figure . Well-defined bundles of TiO2 NTs were formed and composed of uniform tubes of length of typically about 5 μm with wall thickness of around 20–25 nm (Figure a). The length indicates that the growth rate of the TiO2 NTs is about 20 μm h–1 since the present length is attained within the reaction time of 20 min in 0.3 M NaCl electrolyte. Notably, the literature reports that the bundles could be separated into individual TiO2 NTs.[21] On the other hand, TiO2 NT composites loaded with Pt (Figure b), Pd (Figure c), and Ni (Figure d) nanoparticles displayed a slightly different surface morphology. Each of the composites exhibit the existence of agglomerated metal NPs as spheres that are present over the bundles of TiO2 NTs. There is a significant reduction in the wall thickness of the TiO2 NTs to about 3–4 nm and the inner tubular diameter to about 15–20 nm, as seen from the transmission electron microscopy (TEM) image displayed in Figure a.[16] The diameter of the Pt and Pd nanoparticles was about 20 nm (Figure c,e). The dark and bright field images of pristine TiO2 NTs and nanocomposites clearly indicate that the metal nanoparticles are present at the periphery of the tube walls (Figure b,d,f).

Figure 4

FESEM images of (a) pristine TiO2 NTs, (b) Pt/TiO2 NTs, (c) Pd/TiO2 NTs, and (d) Ni/TiO2 NTs.

Figure 5

TEM micrographs; bright and dark field images corresponding to (a, b) TiO2 NTs, (c, d) Pt/TiO2 NTs, (e, f) Pd/TiO2 NTs.

H2 Sensor Studies

H2 sensing response of thick film sensors made of Pt/TiO2 NTs, Pd/TiO2 NTs, and Ni/TiO2 NTs at the operating temperature of 150 °C in Ar ambience are shown in Figure . All sensors demonstrated a stable baseline in the Ar atmosphere. While a specified concentration of H2 is introduced into the Ar stream, the resistance of the corresponding sample decreases systematically. The sensors retrace to the baseline once the H2 is withdrawn. The origin of the response is with basic semiconductor oxide TiO2 and is measured by the changes in resistance. Moreover, the addition of H2 in the ppm level does not change the metallic properties of the loaded metals significantly. The relative change in the Fermi level of metals with addition of H2 can also be a factor in the changes in resistance. Generally, H2 directly interacts with semiconductors like SnO2,[22] ZnO,[23] etc., leading to the changes in conductivity, whereas in this case, it does not interact with TiO2 at least in the chosen temperature range. But the addition of metals introduces the gas-sensing phenomenon. Among the three sensors, Pt/TiO2 NTs (Figure a) responded to H2 in the range of 500–9000 ppm with the response time of 2–3 s and recovery time of 2–3 s, whereas Pd/TiO2 NTs (Figure b) exhibited a higher response with a working range of 1000–9000 ppm having nearly same response and recovery times. On the hand, sensors made of Ni/TiO2 NTs (Figure c) showed a very poor response to H2. These sensors compared with the reported sensors with similar materials in the literature[24−32] indicate the effectiveness of the present sensors (Table ). The mechanism of H2 sensing by each of the samples is investigated using X-ray photoelectron spectroscopy (XPS).

Figure 6

Typical H2 sensing response of (a) Pt/TiO2 NTs, (b) Pd/TiO2 NTs, and (c) Ni/TiO2 NTs nanocomposites.

Table 3

Comparison of H2 Sensors Based on Semiconductor Oxides

semiconductor oxides or its composites	synthesis method	operating temperature (°C)	concentration of H2 gas (ppm)	response time (s)	refs
Au or Pt/SnO2	sol–gel annealing	150	10 000	few minutes	(24)
Pd/WO3	sol–gel annealing	20	1000	<100	(25)
Pt/WO3	RF magnetron sputtering	200	200	42	(26)
NiO	magnetron sputtering	400	5000	300	(27)
Pd/SnO2	reactive magnetron sputtering	200	1000	few minutes	(28)
NiO	pulsed laser deposition	125	30 000	600	(29)
NiO	hydrothermal	200	100	180	(30)
anatase TiO2	micro-arc oxidation	250	1000	45	(31)
nanoporous TiO2	thermal oxidation	500	500	10	(32)
Pt/TiO2 NTs	rapid breakdown anodization	150	1000	2–3	this work
Pd/TiO2 NTs	rapid breakdown anodization	150	1000	2–3	this work
Ni/TiO2 NTs	rapid breakdown anodization	150	1000	15	this work

Core–Shell XPS Studies on Pt/TiO2 NT, Pd/TiO2 NT, and Ni/TiO2 NT Nanocomposites

Elements of group 10 of the periodic table, viz., Ni, Pd, and Pt having electronic configurations 3d8, 4d8, and 5d8, respectively are known to be efficient hydrogenation catalysts[33,34] with catalytic activity proportional to their specific surface area. Nanoparticles of these elements with high specific surface area when dispersed in TiO2 NTs, are expected to enhance the gas-sensing properties of TiO2 toward H2. X-ray photoelectron spectroscopy, the surface-sensitive tool, is ideal to probe the mechanism of interaction of H2 with TiO2. XPS patterns were recorded before and after exposure to a specified concentration of 1000 ppm of H2 in argon (Ar).

Pt/TiO2 NT Composite before and after Exposure to H2

The XPS patterns of Pt dispersed TiO2 recorded for the Ti 2p level before and after exposure to H2 are shown in Figure . The patterns are nearly identically exhibiting characteristic satellites, and no significant changes could be seen after exposure to 1000 ppm H2. The Ti 2p3/2 pattern of Pt/TiO2 NTs before exposure to H2 shows features that can be fitted with two peaks. The less intense peak with the center of 456.8 eV and the high intense peak with the center of 458 eV are observed. The same features are also seen for the corresponding Ti 2p1/2 components. The binding energy values show that the peak at 456.8 eV corresponds to Ti3+ ion whereas the peak at 458 eV corresponds to Ti4+ ion.[35] The 3+ valence for Ti could be due to the oxygen nonstoichiometry in TiO2. The overall fit gives rise to a correlation coefficient of 0.99. After exposure to 1000 ppm of H2, the Ti 2p pattern remains identical without any changes in full width at half-maximum (FWHM), as shown in Figure b. On the basis of the curve fit, there is no significant change in the peak characteristic of Ti3+ ion, which implies that there is no reduction of TiO2 although hydrogen spillover normally operates in hydrogenation reactions.

Figure 7

Curve fitting of Ti 2p level of Pt/TiO2 NTs (a) in the presence of 1000 ppm of H2 and (b) absence of H2.

A drastic decrease in conductivity of the sensor during exposure to 1000 ppm of H2 shows that Pt should have a different role in sensing and therefore demands the investigation of the oxidation state of Pt. The XPS pattern of the Pt 4f level in Pt/TiO2 NTs before exposure shows a complex pattern, indicating at least two different oxidation states for Pt (Figure a).[36,37] The weak shoulder at the low binding energy of about 71.5 eV shows the fingerprints of Pt in an elemental form, whereas the intense peak at around 73.2 eV is characteristic of Pt2+. Peak fit was carried out for these two species, and the overall fit gives rise to a correlation coefficient of 0.99. The two species are elemental Pt and most likely PtO as chemisorption of oxygen can cause surface oxidation, giving rise to PtO, although there are no evidences from XRD to support this claim. On the other hand, the Pt 4f pattern recorded after exposure to H2 shows reversal in intensity of peaks under consideration, as seen from Figure b. That is, the peak intensity of the elemental Pt increased at the expense of the Pt2+ peak intensity. This shows that a fraction of the surface species in which Pt is in +2 oxidation state undergoes reduction upon exposure to H2 to form elemental Pt.[38] The reduction of surface PtO is responsible for the changes in conductivity during sensing. Hence, TiO2 remains more or less a passive medium in Pt/TiO2 NT composite and the hydrogen spillover by Pt does not operate during sensing. This leaves the only option that the relative changes in Fermi levels between PtO/TiO2 and Pt/TiO2 are responsible for the changes in conductivity of Pt/TiO2 NT composite during sensing. Incidentally, this reduction reaction (eq ) can occur at as low as 150 °C. The schematic diagram of the H2 sensing mechanism of Pt/TiO2 NTs is presented in Figure .

Figure 8

Curve fitting of the Pt 4f level of Pt/TiO2 NTs (a) in the presence of 1000 ppm of H2 and (b) absence of H2.

Figure 9

Schematic diagram of the probable H2 gas-sensing mechanism for Pt/TiO2 NTs.

Pd/TiO2 NTs Nanocomposite before and after Exposure to H2

XPS studies on Pd 3d and Ti 2p levels of Pd/TiO2 NT composites were recorded before and after exposure to 1000 ppm of H2. The Ti 2p level before exposure to 1000 ppm of H2 is shown in Figure a. A weak shoulder observed around 456 eV followed by an intense peak at 458 eV shows that the Ti 2p3/2 curve can be fitted with two peaks, as shown in the Figure a. The overall fit gives rise to a correlation coefficient of 0.99. The peaks at 456 and 458 eV are characteristic of Ti3+ and Ti4+ ions, respectively.[35] The Ti 2p pattern after exposure to H2 gas shows an increase in the intensity of the shoulder at 456 eV, which allows us to perform the curve fitting with two components, and the fitted curves are shown in Figure b. It is clear that after exposure to H2, the area under the peak of Ti3+ is higher than that of the virgin sample, which confirms the surface reduction of TiO2. This implies that TiO2 NTs undergo a reduction process upon exposure to 1000 ppm H2.

Figure 10

Curve fitting of the Ti 2p level of Pd/TiO2 NTs (a) in the presence of 1000 ppm of H2 and (b) absence of H2.

The XPS pattern of the Pd 3d level recorded before exposure is shown in Figure a. The pattern is fitted with a doublet having the peak position of 336.6 eV for Pd 3d5/2 with its component Pd 3d3/2 at 342 eV having a spin–orbit coupling of 5.4 eV.[39] The peak at 336 eV implies that Pd is in +2 valence state, whereas the XRD pattern provides the evidence for elemental Pd only and not for PdO. The only way this can be resolved is that the oxide layer should extend up to a depth of at least few tens of angstroms as the escape depth of XPS is only about 10–20 Å. The XPS pattern of the Pd 3d level after exposure to H2 looks complex showing additional features (Figure b). A new peak at 334 eV emerges, which corresponds to elemental Pd along with the characteristic peak of PdO. A curve fit constrained by the spin–orbit separation of 5.6 eV among the components and the intensity ratio of (2J + 1) was carried out, and the fitted curve is shown in Figure b. Upon exposure to 1000 ppm H2, partial reduction of PdO is observed. XPS studies show that it is likely that the outer PdO layer undergoes reduction first to form Pd (eq ), which subsequently transfers hydrogen to TiO2 through the spillover mechanism, causing surface reduction of the latter. Thus, in the case of Pd/TiO2 NTs, the spillover mechanism operates during sensing. The observed increase in conductivity is due to the generation of Ti3+, which acts as a donor for the conduction band of TiO2 NTs (eqs and 3) (Figure b) in addition to Fermi level changes of PdO/TiO2 to Pd/TiO2. Incidentally, the higher response exhibited by thick films of Pd/TiO2 sensor toward H2 than that of Pt/TiO2 indirectly confirms this result. The schematic diagram of the H2 sensing mechanism of Pd/TiO2 NTs is presented in Figure , and the proposed sensing reactions at 150 °C proceed through the following reactions (eqs –4).

Figure 11

Curve fitting of the Pd 3d level of Pd/TiO2 NTs (a) in the presence of 1000 ppm of H2 and (b) absence of H2.

Figure 12

Schematic diagram of the probable H2 gas-sensing mechanism for Pd/TiO2 NTs.

Ni/TiO2 NTs Nanocomposite before and after Exposure to 1000 ppm H2

The Ti 2p recorded before and after exposure to 1000 ppm H2 of Ni/TiO2 NT composite is shown in Figure a. The curve fit was carried using constraints of spin–orbit separation of about 6.0 eV and the intensity ratio of (2J + 1). Only a single doublet of Ti 2p3/2 and Ti 2p1/2 without any additional feature is observed. The position of the 2p3/2 peak shows that Ti is in a +4 valence state. After exposure, the pattern remains the same, indicating nearly all of the Ti is in the +4 valence state even after exposure.

Figure 13

Curve fitting of the Ti 2p level of Ni/TiO2 NTs (a) in the presence of 1000 ppm H2 and (b) absence of H2.

The Ni 2p pattern before exposure exhibits a complex pattern with different valence states for nickel, as shown in Figure a. However, the XRD pattern shows a face-centered cubic nickel pattern without any impurity lines for NiO, implying that the valence state higher than the elemental state should be arising from the oxidized surface of nickel, as attested by the characteristic satellites at 852 and 870 eV for Ni2+ and elemental Ni.[40,41] Curve fitting was carried out using spin–orbit coupling of 17.3 eV and the intensity ratio of (2J + 1) and is shown in Figure b. A strong line at 853 eV shows elemental Ni and another strong peak at 856 eV shows the Ni2+ in Ni/TiO2 NT composite before exposure.[42] However, after exposure to H2, a nearly identical pattern is observed, as seen from Figure b. Neither the peak shape nor FWHM undergo a change, implying that hydrogen interaction did not have any effect on Ni at this operating temperature. In other words, although nickel is used as a catalyst for hydrogenation reactions, the temperature is too low for nickel to interact with H2 (Figure ). Hence, the response toward H2 was very insignificant compared to that of Pt or Pd dispersed TiO2 NTs.

Figure 14

Curve fitting of Ni 2p level of Ni/TiO2 NTs (a) in the presence of 1000 ppm H2 and (b) absence of H2.

Figure 15

Schematic diagram of the probable H2 gas-sensing mechanism for Ni/TiO2 NTs.

Conclusions

Hydrogen-sensing properties of TiO2 NTs impregnated with nanoparticles of Ni, Pd, and Pt are presented. Among the three, the sensitivity is the highest for Pd/TiO2 NTs. The sensitivity of Pt/TiO2 NTs is reasonably good, whereas it is poor for Ni/TiO2 NTs. XPS studies show that the oxide surface of Pd reduces to elements and causes the generation of Ti3+ ion in TiO2 nanotubes through hydrogen spillover. In the case of Pt/TiO2 NTs, only reduction of the oxide layer of Pt nanoparticle takes place without any changes in TiO2 NTs. For Ni/TiO2 NTs, neither the reduction of oxide surface over Ni nor Ti4+ to Ti3+ in TiO2 occurs. Relative to Ni 3d or Pt 5d, the overlap of Pd 4d orbitals with the conduction band of TiO2 along with spillover effect is responsible for the high sensitivity for Pd.

Experimental Section

Materials and Reagents

The RBA technique was applied using a two-electrode system with Ti foil (99.7%, 1 × 2.5 × 2.5 mm3, Good Fellow) as the anode and Pt gauze (99.9%, 52 mesh, 2.5 × 2.5 cm2, Sigma-Aldrich) as the cathode. NaCl (0.3 M, Merck) was prepared using Millipore water and employed as the electrolyte. Chloroplatinic acid, H2PtCl6 (Sigma-Aldrich), palladium chloride, PdCl2 (Sigma-Aldrich), and NiCl2·6H2O (Sigma-Aldrich) were employed as precursors for Pt, Pd, and Ni, respectively. Sodium borohydride, NaBH4 (Merck) was used as the reducing agent.

Preparation of TiO2 NT Powder and Metal NPs/TiO2 NTs

The rapid breakdown anodization (RBA) process was conducted using degreased Ti foil and Pt gauze as the working and counter electrodes, respectively, in a two-electrode cell with 0.3 M NaCl of pH 4.4 as the electrolyte.[35] A potential of 20 V was applied for conduction of the RBA process. As the RBA process proceeded, TiO2 NTs fell into the electrolyte as bundles and were later collected in their powder form to dry overnight in an oven at 70 °C.[43] Thereafter, the dried sample was annealed at 450 °C for 3 h to achieve better crystallinity for the TiO2 NT powder.[3,44] The Pt/TiO2 NT, Pd/TiO2 NT, and Ni/TiO2 NT composites were prepared by the chemical reduction of H2PtCl6, PdCl2, and NiCl2·6H2O over TiO2 NTs using NaBH4 as the reducing agent, correspondingly.

Sensor Assembly and Measurement Conditions

About 1 g of Pt/TiO2 NT powder was mixed thoroughly with a calculated quantity of tetraethyl orthosilicate (TEOS) and an organic binder. The composition was allowed to cure for about 12 h, and thereafter it was screen-printed on alumina substrates coated with electrodes for the sensing process. The thick films were annealed at 425 °C for 10 h. The substrates were mounted on a specially designed testing chamber provided with an inlet and outlet valve for argon (Ar) containing a specified concentration of H2 gas, and the sensor studies were carried out at 150 °C. The sensor was tested in an Ar ambience, since the condition is required to monitor trace levels of H2 gas in Ar cover gas of the sodium-cooled fast breeder reactors at low operation, in shutdown monitoring, and at startup of the reactor.[13,14,45] The sensor samples displayed a steady baseline at around 90 MΩ in pristine Ar ambience, and at the introduction of H2 gas in Ar (H2/Ar), the baseline signal indicated increase in conductance. The changes in resistivity were measured using Agilent digital multimeter (34972A) interfaced to a computer. XPS studies were carried out on thick films before and after exposure to hydrogen to understand the mechanism of sensing. For this study, the virgin sensor (i.e., before the injection of H2/Ar) was subjected to XPS studies. Thereafter, the sensor after being purged with 1000 ppm H2/Ar at 150 °C was analyzed for the surface phenomenon using XPS. Both the XPS studies are compared to understand the changes with respect to injection of H2/Ar and are utilized to propose the plausible mechanism. Likewise, the studies were also performed on Pd/TiO2 NTs and Ni/TiO2 NTs to assess their response patterns and sensing mechanisms.[45]

Characterization of Materials

The Inel Equinox 2000 X-ray Diffractometer (XRD) of ThermoFisher Scientific was utilized to acquire the real-time crystal pattern of the sample using Cu Kα1 radiation of wavelength 1.5406 Å. Transmission electron microscopy (TEM, LIBRA 200 TEM, Carl Zeiss, Germany) was employed to investigate the crystal structures of modified Pt, Pd, and Ni and pristine TiO2 NTs. Raman spectroscopy was carried out using the Raman Microscope, inVia, U.K. FT-IR analysis was performed using Perkin-Elmer spectrum 2. Morphology of the samples was analyzed using FESEM (Carl Zeiss Auriga system, Germany) in the secondary emission mode. The thermal history of the samples was evaluated using TG–GTA using Perkin-Elmer FTA 6000. The chemical states of the elements Ni, Pd, Pt, Ti, and O, before and after the exposure to hydrogen gas of the prepared composites, were analyzed using the SPECS X-ray photoelectron spectrometer (XPS), Germany.