Solution‑Derived ZnO Nanoshells for Flexible 3D Periodic Structures: A Low‑Cost, Non‑Vacuum Templating Approach

Background

3D periodic nanostructures have attracted intense interest for their exceptional optical and mechanical properties, with applications spanning photonic crystals (PhCs) [1,2,3], phononic crystals (PnCs) [4], battery electrodes [5,6], and microfluidic channels [7]. Templating—using self‑assembled colloidal spheres or photopolymers followed by infiltration and template removal—offers a straightforward route to such architectures [8–18]. However, vacuum‑based infiltration methods (ALD, CVD) require expensive equipment and can leave excess overlayers that must be removed by reactive ion etching [9,18,19]. Non‑vacuum alternatives (electrodeposition, sol‑gel) are attractive for their speed and cost but struggle to achieve uniform nanoshell thickness, especially when creating freestanding inverse structures [13,20,23].

ZnO stands out as a semiconductor with excellent optical transparency, electrical conductivity, and thermal stability [24], making it ideal for PhCs, sensors, and transparent electrodes [13,14,24]. Al‑doped ZnO has further shown promise in thermoelectric applications without the use of rare or toxic elements [25,26].

Artificially nanostructured materials are increasingly explored for thermal management, where hierarchical architectures with micro‑ and nanofeatures can reduce thermal conductivity and enhance thermoelectric performance [27]. Nanoshell‑based 3D structures—generated by templating—offer a high surface‑to‑volume ratio but typically rely on vacuum deposition, which is energy‑intensive and costly. A non‑vacuum, solution‑based approach that maintains structural control and reduces shrinkage would therefore represent a significant advancement.

Methods

Preparation of 3D Polymeric Template by PnP

A 0.16–0.19 mm thick cover glass was cleaned by 2 min oxygen plasma and used as the substrate. A bilayer film was then deposited to prevent delamination during development [28,32–34,37]. A negative‑tone SU‑8 photoresist (<2 µm) was first flood‑exposed for adhesion, followed by a 10 µm thick layer spin‑coated at 2000 rpm for 30 s. The substrate was soft‑baked at 95 °C for 10 min. A PDMS phase‑shift mask—featuring a square array of 600 nm periodic cylinders (480 nm diameter, 420 nm depth)—was placed in direct contact with the photoresist. UV irradiation at 355 nm (Nd:YAG laser, 300 mW) proceeded through the mask, and the sample was post‑exposed baked at 65 °C for 7 min. Unexposed regions were removed with PGMEA and rinsed with ethanol, yielding a 3D polymeric template [32–36].

Fabrication of Nanoshell-Based 3D Structure

Figure 1 illustrates the workflow. A 2.0 M ZnO precursor solution (SYM‑Zn20, Kojundo Chemical Lab.) served as the infiltration medium. The template was first coated with a few drops of precursor, then spin‑coated at 2000 rpm for 20 s to ensure uniform coverage. Vacuum degassing promoted penetration into the template’s voids, a technique routinely used in dye‑sensitized solar cell fabrication [38,39]. Pre‑baking was performed at 220 °C for 1 h in an O₂ atmosphere (14 L min⁻¹) to form a pre‑coated ZnO layer. This cycle was repeated multiple times to investigate the effect of infiltration cycles on structural quality.

Schematic diagram of the infiltration process using solution‑derived ZnO for nanoshell‑based 3D structures. a Preparation of template by PnP. b ZnO precursor infiltration and multiple pre‑bakes. c Post‑bake for template removal.

After infiltration, the template was removed by a post‑bake at 410 °C for 4 h in O₂, which simultaneously pyrolyzed the ZnO precursor and eliminated the polymeric scaffold [19]. This dual step is essential; a single high‑temperature bake (> 400 °C) without prior pre‑bake leads to structural collapse (Additional file 1: Figure S1a,b).

Characterization

Scanning electron microscopy (SEM; SU‑6600, Hitachi) revealed the morphology and uniformity of the 3D structures. Cross‑sectional images were used to quantify shrinkage by comparing dimensions between the ZnO inverse and the original SU‑8 template. Energy‑dispersive X‑ray spectrometry (EDX; 5.0 kV) confirmed template removal and determined the Zn:O ratio. UV‑Vis spectroscopy (V‑570, JASCO) measured reflectance spectra, while the transmission spectrum was used to extract the optical bandgap.

Results and Discussion

The periodicity along the out‑of‑plane axis (Talbot distance, Z_T) for the PnP template is calculated via Z_T = (λ₀/n_m)/(1 – √(1 – (λ₀/(n_mp))²)) [41] with λ₀ = 355 nm, n_m = 1.66, p = 600 nm. The measured Z_T was 29.2 % shorter than the theoretical value, consistent with known SU‑8 shrinkage during development [35,42]. Accounting for this factor is crucial for precise 3D design.

Figure 2 shows cross‑sectional SEM images of ZnO precursor/polymer structures after 1–6 infiltration cycles. Pre‑baked ZnO uniformly coats the template, acting as a protective layer that prevents collapse during the 220 °C pre‑bake. With six cycles, the template is completely filled, yielding a consistent filling factor from bottom to top—a key advantage over sol‑gel or electrodeposition methods, which often produce gradients or require high‑density infiltration [13,20,23].

Cross‑sectional SEM images of ZnO precursor/polymer 3D structures after 1–6 infiltration cycles (a–f).

Post‑baking at 410 °C for 4 h removes the polymeric template and converts the pre‑baked precursor into ZnO. Figure 3 reveals that structures with ≥ 4 cycles maintain the intended periodicity and exhibit minimal defects; 1–3 cycle samples show distortion and height reduction due to incomplete template removal and precursor shrinkage. The best results are obtained with six cycles, where the nanoshell thicknesses are < 85 nm, < 100 nm, and < 125 nm for 4, 5, and 6 cycles, respectively (Figure 4). This demonstrates controllable nanoshell thickness via cycle number—a feature difficult to achieve with ALD or sol‑gel processes.

Cross‑sectional SEM images of 3D inverse structures after post‑bake (a–f).

EDX analysis (Figure 5) shows a dramatic drop in carbon content from 47.8 % to 3.5 % after post‑bake, confirming effective template removal and precursor pyrolysis. The Zn:O ratio of 58.3:41.7 matches values reported for ZnO nanorods fabricated by non‑vacuum methods [45,46].

EDX results: (a) carbon content, (b) Zn:O ratio before and after post‑bake.

Histograms of structural dimensions (Figure 6) show that the height and in‑plane periodicity shrinkage for the six‑cycle ZnO inverse structure is only ~16 %, compared to 34 % reported for TiO₂ non‑vacuum templating [23]. The improvement is attributed to the protective ZnO layer that maintains the template’s geometry during high‑temperature processing.

2D structure height and in‑plane periodicity measurements; histograms of size distribution for template and inverse structure.

Reflectance spectra (Additional file 1: Figure S5) display peaks at 410 nm for the template and 450 nm for the ZnO structure, with a maximum reflectance of 62 %. The optical bandgap of the ZnO inverse structure, derived from the (αhν)² vs hν plot (Additional file 1: Figure S6), is 3.0 eV—consistent with ZnO nanorods produced by CBD [47].

Conclusions

We have demonstrated a scalable, non‑vacuum fabrication route that couples proximity field nanopatterning with a solution‑derived ZnO infiltration cycle to produce nanoshell‑based 3D periodic structures. The pre‑coated ZnO layer safeguards the polymeric template, enabling controllable nanoshell thickness, reduced height shrinkage (16 %), and high structural fidelity after post‑baking. EDX confirms effective template removal, while optical measurements validate the quality of the ZnO material. This cost‑effective, flexible process opens avenues for ZnO photonic crystals, sensors, and energy‑related devices.