Synthesis of TiO₂-(B) Nanobelts for Acetone Sensing

Abstract

Titanium dioxide nanobelts were prepared via the alkali-hydrothermal method for application in chemical gas sensing. The formation process of TiO₂-(B) nanobelts and their sensing properties were investigated in detail. FE-SEM was used to study the surface of the obtained structures. The TEM and XRD analyses show that the prepared TiO₂ nanobelts are in the monoclinic phase. Furthermore, TEM shows the formation of porous-like morphology due to crystal defects in the TiO₂-(B) nanobelts. The gas-sensing performance of the structure toward various concentrations of hydrogen, ethanol, acetone, nitrogen dioxide, and methane gases was studied at a temperature range between 100 and 500 °C. The fabricated sensor shows a high response toward acetone at a relatively low working temperature (150 °C), which is important for the development of low-power-consumption functional devices. Moreover, the obtained results indicate that monoclinic TiO₂-B is a promising material for applications in chemo-resistive gas detectors.

1. Introduction

Metal oxide-based chemical gas sensors are now regarded as excellent candidates for environmental monitoring due to their good sensitivity, fast response, and recovery times toward different analytes [1,2,3,4,5,6,7]. Moreover, research studies are progressing to employing metal oxides at their optimum gas-sensing performances and develop low-cost, portable, and stable chemo-resistive sensors. Among the existing technologies of metal oxides, thick films, thin films, nanocrystals, and one-dimensional (1D) structures are widely investigated owing to their superior surface-to-volume ratio, stability, and low production cost, together with excellent analyte-sensing performances [8,9,10,11,12]. Furthermore, 1D SnO₂, ZnO₂, and TiO₂ materials are widely studied in the field of chemo-resistive sensors, accounting for over 78% (SnO₂ > 34%, ZnO₂ > 31%, and TiO₂ > 12%) of all reported 1D metal oxide chemical gas sensors, to the best of the authors’ knowledge [13,14].

TiO₂ is one of the interesting transition-metal oxides due to its biocompatibility, stability, abundance, and low cost of production [15,16,17,18,19,20,21]. The crystallite phases of TiO₂, i.e., anatase, rutile, brookite, and monoclinic, are shown in Figure 1a–d [22]. The chains of TiO₆ octahedra are the building blocks that make up the tetragonal lattice structures of TiO₂. In an anatase structure, zigzag chains of TiO₆ octahedra are linked to each other with four edge-shared bonding (faces sharing, Figure 1a), and, in the rutile phase, two opposite edges of each octahedron link at a corner of an oxygen atom, forming linear chains of octahedra at each corner (Figure 1b). Consequently, rutile is the most stable crystal structure in the bulk material. However, brookite structures (Figure 1c) are also commonly available based on specific growth techniques and mechanisms [23]. Thus, investigating novel approaches for the synthesis of TiO₂ nanostructures including simple fabrication technologies is still demanding. Hence, several growth techniques, for instance, atomic layer deposition [24], electrochemical anodization [25], hydrothermal method [26], chemical vapor deposition (CVD) [27], and electrospinning [28], have been employed to fabricate low-dimensional TiO₂ structures for gas-sensing applications [29,30].

Besides these three crystal phases, another crystal phase, TiO₂-(B) (monoclinic, Figure 1d), is also available based on synthesis parameters. TiO₂-(B) possesses a layer structure, resulting in low density, high specific capacity, and more open structural frameworks compared to other phases [30,31]. The various TiO₂ crystal phases react with interacting gases in distinctive ways, as is well known [18]. Due to their molecular structure, bond length, and reactive groups, crystal phases are sensitive to certain gases [32]. For instance, rutile TiO₂ nanorods are sensitive to isopropanol [33] and anatase TiO₂ nanowires are sensitive to NO₂ [34]. As a result, it is crucial to research the various TiO₂ crystal phases for gas sensing. However, monoclinic TiO₂ is very rarely reported in the literature because of synthesis difficulties. Pioneering work on TiO₂-(B) was reported by Marchand et al. in 1980, which was prepared using the hydrothermal method, one of the simplest, low-cost, and high-yield processes [35]. Subsequently, several interesting research works have been reported on the preparation of TiO₂-(B), showing high catalytic and electrochemical properties [36,37,38,39].

Moreover, the hydrothermal method is a well-known process for preparing metal oxide nanomaterials with various morphologies by simply varying the pH value of the solution and the growth temperature. However, some highly toxic chemicals are still involved in these growth methods to achieve high-quality nanostructures. Therefore, we investigated the potential of replacing highly toxic compounds with low- or moderately toxic chemicals to grow TiO₂-(B) nanobelts via low-cost hydrothermal techniques. Herein, harmful hydrochloric acid (HCl), which is usually used in the protonation phase, was replaced by acetic acid (CH₃COOH) for the synthesis of TiO₂-(B). Furthermore, the morphological, structural, and compositional characteristics of the material were investigated. Additionally, the gas-sensing properties of TiO₂-(B) were determined in order to employ the prepared material in low-power-consumption chemiresistive gas sensors.

2. Materials and Methods

2.1. Growth of TiO₂-B Nanobelts

The alkali-hydrothermal process was employed for the synthesis of TiO₂ nanobelts. First, sodium titanate hydrated (Na₂Ti₃O₇.mH₂O) was synthesized using the hydrothermal method. A quantity of 1 g of titanium dioxide (TiO₂, powder, 21 nm primary particle size, ≥99.5%, Aldrich) was mixed with 70 mL of 10.0 mol dm⁻³ sodium hydroxide (NaOH, 98%, Loba Chemie, Mumbai, India) aqueous solution and mechanically stirred for 30 min, followed by 15 min of ultrasonication. The stirring and sonication processes were repeated six times following each other. Later, the obtained suspension was transferred into a 100 mL Teflon-lined stainless-steel autoclave and thermally treated for 48 h at five different temperatures: 120 °C, 135 °C, 150 °C, 175 °C, and 200 °C. After that, the material was cooled down to room temperature. The resultant white slurry (Na₂Ti₃O₇) was washed thoroughly with deionized water, followed by a filtration process until the pH of the washing solution reached the value of 7. Subsequently, the wet slurry was immersed in 1 mol L⁻³ acetic acid (99.5%, DAEJUNG, Sihung, Republic of Korea) aqueous solution for 24 h to prepare the protonated titanate form (H₂Ti₃O₇). The prepared H₂Ti₃O₇ was washed thoroughly with distilled water and underwent filtration until the washing solution became pH neutral. Later, the obtained H₂Ti₃O₇ was dried at 80 °C for 24 h, and then calcinated at 500 °C for 3 h.

The basic chemical routing of the TiO₂ nanobelt growth can be written as follows [40]:

The reaction starts with the dissolving of TiO₂, in the presence of NaOH, as

3TiO₂ + 2NaOH → Na₂Ti₃O₇ + H₂O

(1)

During the CH₃COOH washing, the Na₂Ti₃O₇ nanobelt is known to convert as follows due to the ion exchange:

Na₂Ti₃O₇ + 2CH₃COOH → H₂Ti₃O₇ + 2CH₃COO⁻Na⁺

(2)

In the calcination process, H₂Ti₃O₇ will convert to TiO₂ as follows:

H₂Ti₃O₇ → 3TiO₂ + H₂O

(3)

2.2. Characterization

The morphological investigations were carried out using a field-emission scanning electron microscope (FE-SEM) TESCAN (Brno, Czech Republic) (MIRA-3) at a voltage of 5 kV. Further, the obtained nanostructures were investigated by means of a JEOL JEM ARM 200F analytical transmission electron microscope (TEM) operated at 200 kV, equipped with an EDS detector to acquire Energy-Dispersive X-ray (EDS) spectra or maps for elemental investigation. X-ray diffraction spectroscopy (XRD) was performed in the range of 20–80 degrees using an Empyrean diffractometer (PANalytical, Almelo, The Netherlands) equipped with a Cu-kα₁ (λ = 1.5406 Å) tube operating at 40 kV to 40 mA. Raman spectra were measured using a fiber-coupled confocal optical microscope (HORIBA, XploRA Nano) at 100× magnification. The spectra were recorded in the wavelength range of 200–1200 cm⁻¹ using a red laser source (638 nm).

2.3. Fabrication of Sensors

The drop-casting method was used to fabricate the conductometric sensing devices and several sensors were fabricated by varying the number of droplets, as shown in

Table 1. Nomenclature of the fabricated conductometric chemical sensors.

Number of Drops	Nomenclature
2	TiO2-2n
4	TiO2-4n
6	TiO2-6n
8	TiO2-8n

. The process can be explained in brief as follows: Alumina substrates (2 mm × 2 mm, Kyocera, Kanazawa, Japan, 99.9%) were ultrasonically cleaned for 20 min in an acetone bath, followed by drying with synthetic air. Next, TiW (10/90 wt.%) adhesion layer (~90 nm) was sputtered on top of the alumina substrates via DC magnetron sputtering at 75 W argon plasma, 7 (standard cubic centimeters per minute) argon flow, 5.0 mTorr pressure, and 300 °C. Then, Pt interdigitated electrodes (IDE, 1 μm) were deposited on the adhesion layer using the same DC magnetron sputtering conditions [41]. Additionally, a platinum heater was fabricated on the opponent side of the alumina substrates to investigate the performance of the sensor at different operating temperatures (100–500 °C). Meanwhile, the Al₂O₃ substrates with IDE and heater were soldered to the TO-39 package with gold wires.

Later, the dispersed TiO₂ nanobelts (5 mg) in absolute ethanol (5 mL) were drop-casted on the top of the soldered device at room temperature. Figure 2 shows the schematic of the prepared sensors.

2.4. Gas Testing Measurements

All the fabricated sensors were thermally stabilized at 400 °C for 48 h before mounting into a gas-testing chamber. The conductometric response of the fabricated sensors was tested in a stainless-steel gas chamber having a volume of 1 L, which was mounted inside a climatic chamber (Angelantoni, Perugia, Italy, model MTC 120). The temperature was set to 20 °C inside the climatic chamber, and 40% of relative humidity (RH) was maintained inside the gas chamber. To create humid air, dry air flowed through a Drechsel bottle operating at a water bath of 25 °C. Subsequently, the sensors were mounted in the gas chamber to investigate their response toward H₂, C₂H₅OH, C₃H₆O, NO₂, and CH₄ with a fixed voltage of 1 V. The gas-sensing dynamic was recorded for the period of 140 min for each concentration (30 min injection followed by 45 min recovery). The response of sensors is defined as ∆G/G (S) = (G_g − G_a)/G_a or (G_a − G_g)/G_g for reducing and oxidizing gases, respectively, where G_a is the conductance of the sensor in synthetic air while G_g is the conductance of the sensor in the presence of analyte gas. The gas-sensing measurements were carried out at a gas flow rate of 200 sccm.

3. Results

To study the formation and crystal structure of the prepared TiO₂, X-ray diffraction (XRD) analysis was employed. Figure 3a clearly shows the formation of intermediate H₂Ti₃O₇ and residual sodium titanate (JCPDS 41-0192, JCPDS 98-008-2059). However, these H₂Ti₃O₇ and residual sodium titanate have been completely demolished upon washing and annealing (Figure 3b). Further, H₂Ti₃O₇ is the most important intermediate state, which results in the formation of nanobelt-like morphology in the final product ascribed to the dissolution–recrystallization process [42]. After OH⁻ ions in the NaOH solution have gradually diffused into the initial 3D anatase TiO₂, the Ti–O–Ti bonds are dissolved, leading to the exfoliation of single-layered sheets (nanosheets) composed of TiO₆ octahedra. This dissolution is the main root for the presence of characteristic stretching vibrational peaks in the Raman spectrum at 280 cm⁻¹ (Ti–O–Na), 372 cm⁻¹ (Ti–O–Ti), and 670.8 cm⁻¹ (TiO₆) [38,39], as shown in Figure 3c.

Meanwhile, Na⁺ ions intercalate in the interlayer space between TiO₆ sheets to neutralize the negative charge of the formed TiO₆ sheets, as shown in Figure 4a. This step controls the Na/Ti ratio of the final titanate product and forms intermediate sodium titanate (Na₂Ti₃O₇), as described in Equation (1) [28]. However, such dissolution and exfoliation processes are relatively easier for small anatase precursors yielding titanate sheet units. These sodium titanate sheets copy the epitaxial crystal growth along the c-axis, resulting in titanate sheet units growing into a sheet-like structure (Figure 5a) [43]. Next, Na₂Ti₃O₇ nanosheets are transformed into H₂Ti₃O₇ nanobelts when the ion-exchange process is employed (Equation (2) and Figure 4b). Accordingly, the formation of nanobelt-like structure is ascribed to the splitting of nanosheets to release the excess strong stress upon the replacement of Na⁺ by larger H₃O⁺ cations when forming H₂Ti₃O₇ [44]. Finally, TiO₂-B nanobelts are obtained by annealing the H₂Ti₃O₇ at 500 °C, as presented in Equation (3) and Figure 4c.

The diffraction peaks in Figure 3b at the two-theta values of 24.98°, 28.59°, 29.79°, 33.41°, 39.47°, 43.48°, 44.61°, 48.66°, 52.89°, 58.47°, 62.38°, 67.27°, and 76.69° are, respectively, assigned to the reflections of the (110), (002), (40-1), (31-1), (31-2), (003), (60-2), (020), (113), (71-1), (31--), (02-3), and (712) planes of the monoclinic phase (JCPDs: 98-004-7691, space group C12/m1) TiO₂-B.

Figure S1a–f show FE-SEM images of the intermediate H₂Ti₃O₇ structures, which were synthesized by dissolving anatase TiO₂ powder (particle diameter 20–25 nm, P21, Figure S1a) with NaOH solution under hydrothermal conditions and using the ion-exchange process with acetic acid. Figure S1b shows that the hydrothermal treatment at 120 °C results in the transformation of anatase particles into a compact and thick flake-like morphology. Further, at 135 °C, anatase particles were transferred into large aggregated particles (Figure S1c). At 150 °C, these P21 particles aggregated and formed a compact film-like morphology (Figure S1d). However, P21 particles were completely transformed into belt-like shapes (Figure S1e–f) when the hydrothermal treatment was employed at 175 °C and 200 °C. Furthermore, the obtained nanobelts at 200 °C showed a compact and low aspect ratio. Accordingly, further investigation was carried out on the sample that was thermally treated at 200 °C.

The conventional TEM (CTEM) image in Figure 6a shows a one-dimensional nanobelt structure in the obtained intermediate H₂Ti₃O₇ to a very long extent in the range of ten micrometers. Figure 6b shows a CTEM image of TiO₂ nanobelts, which were annealed at 500 °C. The HRTEM image (Figure 6c), with its insert clearly showing (101) planes that are characteristic of anatase structure, demonstrates that the nanobelts are well crystallized. Figure 6d shows the selected-area electron diffraction (SAED) of such a nanobelt, which confirms the presence of (101), (200), and (105) crystallographic planes of TiO₂-(B) (CIF number 9009086) corresponding to the lattice parameters a = b = 3.785 Å, c = 9.514 Å, and α = β = γ = 90°. Also, Figure 6d clearly shows the formation of nanobelts in the direction of the c-axis. Furthermore, Figure 6e,f show the presence of a “porous-like” morphology in some regions of the prepared TiO₂-(B) due to the large density of structural defects in specific areas along the nanobelts.

Energy-Dispersive X-ray (EDS) spectra on the TEM grid were carried out to study the elemental distributions on the surface of TiO₂-(B) nanobelts. The EDS maps (Figure 7a–c) indicate the presence of constituent elements (Ti and O) on the surface of TiO₂-(B) nanobelts. Remarkably, Figure 7a–d also show some overlapping nanobelts, which can significantly enhance the response of the sensor when employed in gas sensing [45].

Gas-Sensing Performance

Figure 8 shows the conductance-temperature behavior of the fabricated TiO₂-(B) sensors. The conductance was found to increase with increasing operating temperature of the sensors, which is ascribed to the semiconducting characteristics of the prepared TiO₂-(B) material. However, the sensors fabricated with eight drops (TiO₂-8n) show the highest conductivity at the operating temperature of 400 °C.

Preliminary gas-sensing studies were carried out at different working temperatures. The highest conductivity was observed at 400 °C. Interestingly, the sensors (TiO₂-8n) demonstrated the highest response toward acetone compared to other TiO₂ sensors. Therefore, we further analyzed the sensing properties of the TiO₂-8n sensor.

We investigated the gas-sensing properties of TiO₂-8n devices toward hydrogen (H₂), methane (CH₄), nitrogen dioxide (NO₂), ethanol (C₂H₅OH), and acetone (C₃H₆O) at the temperature range of 100–500 °C (step, 50 °C). The electrical conductance of the sensors increases to a maximum value once reducing gases are injected and returns to their initial value when the gas flow is not present in the chamber. This is a typical n-type semiconducting behavior, which is in agreement with the results reported in the literature [46,47]. Usually, working temperature affects the gas interaction with the sensor up to a certain catalytic temperature and then hinders the interaction, thereby modulating the response value. Therefore, metal oxide gas sensors have different response values depending on their working temperature. Thus, investigating the optimum working temperature is significantly important with respect to the performance of the sensors.

The response values of the TiO₂-8n sensor toward 10 ppm of C₃H₆O at different working temperatures are shown in Figure 9. The sensor exhibited the highest response at the working temperature of 150 °C. Figure 10 shows the dynamic response plot of the TiO₂-8n sensor to different concentrations of C₃H₆O at the operating temperature of 150 °C. TiO₂-based gas sensors mostly operate at elevated working temperatures (>200 °C). So, the relatively low working temperature of the fabricated sensors can be due to the high catalytic property of TiO₂-(B), which is also evidenced in the conductance variation of the sensors [48]. Also, porous regions in the nanobelts encourage the oxygen adsorption/desorption mechanism [44]. Accordingly, the porous-like region of the prepared nanobelts can potentially be one of the reasons for the high response at low working temperatures [49].

In general, the gas-sensing mechanism of TiO₂-(B) nanobelts is acerbated by surface reactions and induced carrier charge transfer due to the adsorption of analyte gas molecules. Furthermore, the sensing mechanism is based on two main interactions known as oxygen adsorption and C₃H₆O interaction on the surface of TiO₂ nanobelts. The sensing mechanism is related to the adsorption of oxygen onto the surface of TiO₂-(B) nanobelts. Usually, the adsorption of oxygen molecules occurs on the exposed surface of the MOX surface [50]. Consequently, electron transfers from the conduction band of TiO₂ to oxygen molecules result in the formation of oxygen species at the surface. Typically, three types of oxygen species (${\mathrm{O}}_2^{-}$, O⁻, O²⁻) are formed depending on the temperature. The formation of oxygen species is shown in Equations (4)–(7). Furthermore, the extraction of electrons by oxygen molecules leads to an increase in the depletion layer width. Accordingly, resistance increases, as shown in Figure 11a [35].

O_2(gas) ↔ O_2(ads)

(4)

O_2(ads)+ e⁻ ↔ O₂⁻_(ads)

(5)

$${{\mathrm{O}}_2^{-}}_{\left(\mathrm{ads}\right)}+{\mathrm{e}}^{-}\leftrightarrow {2{\mathrm{O}}^{-}}_{\left(\mathrm{ads}\right)}$$

(6)

O⁻_(ads) + e⁻ ↔ O²⁻_(ads)

(7)

Once the sensor is exposed to acetone, it interacts with the adsorbed oxygen molecular species. The interaction with O₂⁻ is interesting to discuss here because the optimum working temperature is 150 °C. This interaction leads to the decomposition of C₃H₆O into carbon dioxide (CO₂) and water (H₂O), as shown in Equations (8) and (9) [50]. Thus, the released electrons (e⁻) reduce the depletion layer, hence lowering the resistance of the sensor, as shown in Figure 8b.

CH₃COCH_3(gas) ↔ CH₃COCH_3(ads)

(8)

CH₃COCH_3(ads) + 8O⁻_(ads) → 3CO_2(gas) + 3H₂O_(gas) + 8e⁻

(9)

The sensing characteristics of the structures indicate that their response is enhanced by increasing the concentration of C₃H₆O (Figure 12). Conversely, a higher gas concentration results in a higher number of surface interactions, thus resulting in a higher response [51]. The porosity of the nanostructures is one of the possible reasons for the not-too-slow gas adsorption and diffusion within the sensor, resulting in the estimated response and recovery times of 348 s and 600 s at 40 RH% toward 10 ppm C₃H₆O. The estimated response and recovery times are shown in

Table 2. Sensors’ performance to different concentrations of C₃H₆O.

C3H6O (ppm)	40 RH%
	Response	Response	Response Time (S)	Recovery Time (S)
10	1.3	6.5	378	600
25	2.1	9.6	348	960
50	4.4	12.7	324	1320

.

A gas-sensing material with a lower detection limit may have a good response to a smaller quantity of targeted gas, which ultimately helps to prevent serious accidents in a polluted environment. Hence, the power fitting (y = 6.4492x^0.6046) between response and concentration is a significantly important characteristic (Figure 13a), in which the gradient describes the sensitivity and the intercepts show the detection limit. As shown in Figure 13b, the fabricated sensors have a relatively better response with a detection limit of ~0.05 ppm (∆G/G = 1 on power fitting). Concerning the practical application of the sensors, reproducibility and stability in response are also considered essential parameters.

Hence, the reproducibility and stability of TiO₂-(B) nanobelts toward C₂H₆O were studied and reported herein. Figure S2 shows the reproducibility of the sensor toward three consecutive cycles of 50 ppm C₃H₆O at 150 °C and 40 RH%. Almost the same response value is achieved upon successive injections of C₃H₆O. Figure S3 shows the stability plot for two weeks toward 50 ppm C₃H₆O at 150 °C in 40 RH%. Figure S3 clearly shows the excellent performances of the sensors’ response during the tested period. Figure S4 displays the response values of the TiO₂-8 sensor toward different gases at different working temperatures. Furthermore, Figure 13b depicts the selectivity of the TiO₂-8 sensor toward the 10 ppm C₃H₆O, 10 ppm C₂H₅OH, 100 ppm H₂, 100 ppm CH₄, and 1 ppm NO₂ at the operating temperature of 150 °C.

Thus, the observed C₃H₆O-sensing performances, for instance, response, selectivity, response/recovery times, reproducibility, and stability, of the TiO₂-B nanobelts make them a potential candidate for selective sensing of C₃H₆O in a variety of applications when comparing the literature survey shown in

Table 3. Comparison of gas-sensing parameters of TiO₂-B nanobelt sensor with other reported TiO₂-based C₃H₆O sensors.

Material	Synthesis Route	Working Tempt. (°C)	Response (Ra/Rg)	C3H6O (ppm)	Res. Time (Tres) (s)	Rec. Time (Trec) (s)	The Lowest Detection Limit (ppm)	Ref.
TiO2 porous NPs	Hydrothermal	275	13.9	100	11	14	-	[52]
TiO2 NPs	Matrix-assisted pulsed laser deposition	400	6	100	240	-	20	[53]
Nanoporous TiO2	Hydrothermal	370	25.97	500	13	8	20	[54]
Ag-TiO2 nanobelts	Hydrothermal	260	28.25	500	6	8	0.8	[55]
TiO2 nanorods	Electrospun	500	13	300	12	6	-	[56]
TiO2-B nanorods	Hydrothermal	320	2.3	100	3	180	-	[57]
TiO2-B nanobelts	Hydrothermal	150	12.7	50	324	1320	0.7	This work

.

4. Conclusions

Monoclinic TiO₂ nanobelts were successfully prepared using the hydrothermal method. These studies indicated that the employment of acetic acid was successful in the protonation phase of TiO₂-(B). The fabricated TiO₂ gas sensors showed good gas-sensing performance. The highest response (∆G/G) was 6.5 toward 10 ppm of acetone at the working temperature of 150 °C. Moreover, the sensors demonstrated good selectivity toward acetone compared to ethanol, hydrogen, methane, and nitrogen dioxide. Thus, the fabricated monoclinic TiO₂-B nanobelts are promising candidates for developing functional gas-sensing devices that work at relatively low power consumption. Additionally, the potential of lower detection limit and high selectivity toward acetone could make it interesting to employ the synthesized monoclinic TiO₂ nanobelts in chemiresistive gas sensor applications.