Optically Activated 3D Thin‐Shell TiO2 for Super‐Sensitive Chemoresistive Responses: Toward Visible Light Activation

Abstract One of the well‐known strategies for achieving high‐performance light‐activated gas sensors is to design a nanostructure for effective surface responses with its geometric advances. However, no study has gone beyond the benefits of the large surface area and provided fundamental strategies to offer a rational structure for increasing their optical and chemical performances. Here, a new class of UV‐activated sensing nanoarchitecture made of highly periodic 3D TiO2, which facilitates 55 times enhanced light absorption by confining the incident light in the nanostructure, is prepared as an active gas channel. The key parameters, such as the total 3D TiO2 film and thin‐shell thicknesses, are precisely optimized by finite element analysis. Collectively, this fundamental design leads to ultrahigh chemoresistive response to NO2 with a theoretical detection limit of ≈200 ppt. The demonstration of high responses with visible light illumination proposes a future perspective for light‐activated gas sensors based on semiconducting oxides.


Experimental Section
Fabrication of a 3D polymeric template on an IDEs substrate: Interdigitated electrodes (IDEs) with 5 μm spacing were fabricated by depositing platinum (100 nm) and titanium (30 nm) on a SiO 2 /Si substrate using photolithography and e-beam evaporation techniques. [3,4] After cleaning the IDEs substrate by dipping and sonicating in acetone and ethanol in sequence, the substrate was dried and treated by air plasma (CUTEMP, Femtoscience) at a flow rate of 45 s.c.c.m., pressure of 40 mTorr, and power of 60 W for 120 s. Afterward, the edges of the electrodes were selectively concealed with commercial adhesives (3M tape) to avoid contamination during the following photoresist (PR) spin-casting step. First, a thin layer (~2 μm) of the PR (SU-8 2, Microchem) was spin-coated onto a plasma-treated substrate as an adhesion-promoting layer. After removing the pre-deposited tape, the adhesion layer was patterned into the open-window structure at the electrode pattern and hard-baked on a hot plate at 210 °C for 5 min. Then, a relatively thick layer (~10 μm) of the PR (SU-8 10, Microchem) was spin-coated on the adhesion layer again. The PR layer was carefully soft baked in 2 steps: 65 °C for 30 min and 95 °C for 30 min. Then, the substrate was induced to soft contact with a conformal phase mask designed to have a square array of holes (with a diameter of ~480 nm, a depth of ~400 nm, and a periodicity of ~600 nm). [17,20,24,25] After the 3D nanopatterning process using the proximity-field nanopatterning (PnP) technique, conventional lithographic procedures including a post-baking step with a 65 °C hot plate for 7 min, a developing step (SU-8 developer, Microchem), and a rinsing step with deionized water and a drying process were performed.
Material conversion into a 3D TiO 2 : A thin TiO 2 layer (0-100 nm) was conformally deposited on the pre-fabricated 3D template by atomic layer deposition (ALD) technique at 90 °C (Atomic-Classic, CN1 Co., Ltd). [24] Tetrakis-dimethyl-amido titanium (TDMAT) (UP Chemical) was used as a precursor and H 2 O was used as a reactant, respectively. The deposition cycles were controlled to modify the deposited TiO 2 thickness with the deposition rate of approximately 0.75 Å per cycle. Then, the 3D polymeric template was placed in a furnace at 500 °C to obtain the 3D TiO 2 hollow nanostructure by thermally decomposing the 3D polymeric template and annealing the TiO 2 in a single thermal step.
Characterization: The morphologies of the 3D TiO 2 samples were characterized using field-emission scanning electron microscopy (FESEM, S-4800, Hitachi) operating at an accelerating voltage of 5-10 kV. The structural features of the fabricated 3D TiO 2 were obtained by field emission transmission electron microscopy (FE-TEM) using a Tecnai F20 (FEI Company) at an accelerating voltage of 200 kV. The crystallographic analysis of the samples was performed by high-resolution powder X-ray diffraction (XRD) using SmartLab (RIGAKU). Elemental mapping measurements were collected via energy dispersive spectroscopy (EDS, Magellan 400, FEI). The absorbance measurements were performed using UV-VIS spectrophotometry (SolidSpec-3700, Shimadzu). The spectra of photoluminescence (PL) were collected by a fluoro-spectrophotometer using an F-7000 (Hitachi) with an excitation source for Xe lamp (150 W). The chemical bonding and coreelectron binding energy of the 3D nanostructured TiO 2 film samples were investigated by insitu X-ray photoelectron spectroscopy (XPS) using an Axis-Supra (Kratos) with automated monochromatic X-ray source. To compensate for the surface charging, we used a low energy electron flood gun (charge neutralizer).

Gas sensing measurement:
The gas-sensing measurement was conducted in a quartz tube with a gas injection line and a ventilation line. The breadboard with commercial LEDs (PL-UVA521CMK (Photron Co., Ltd), 513SBC3Z, 513UGC, 513UYC3 (Britestone)) was placed on top of a quartz tube with the help of plastic support. The gas injection was conducted using an automated mass flow controller repeating a sequence of dry air for 3000 s and the target gas (balanced with dry air, Sinjin Gases) for 500 s. The gas flow was maintained at 1000 s.c.c.m. throughout the whole measurements. The gas responses toward the target gas were precisely measured at a DC bias voltage of 5.0 V using a source meter (Keithley 2400).
Additional experimental details about the gas-sensing measurement can be found in our previous works. [31] Simulations: The optical modeling and calculations were performed by commercial finite differential time-domain (FDTD, Lumerical Solutions Ltd.). The PnP process with a square array phase mask (period: 600 nm) was simulated in the set of a 5-nm cubic mesh and refractive indices of each component (PDMS mask: 1.4, SU-8 photoresist: 1.66) with a 355 nm plane wave source assuming periodic boundaries. The resulting 3D interference image in photoresist was modeled to the 3D SU-8 polymer template by the level-set method to possess a fill factor of 50 %. [18] To consider the TiO 2 thin-shells on the template by ALD deposition, the thin-shell structures were generated to conformally cover the SU-8 template with the measured refractive index by ellipsometry. For the optical responses of the TiO 2 optical mazes, the electrical intensity field distributions of each model were calculated to visualize and further analyzed to select intensities only in the TiO 2 thin-shell regions.