**3. Results and Discussion**

Suspensions comprising microdroplets of Galinstan and EGaIn were individually prepared as shown in Figure 1a. It exhibited a matt dark gray color due to diffuse reflections and surface oxide layers of the microdroplets. As schematically illustrated in Figure 1b, the suspension was drop-dispensed onto a flat PDMS substrate and then slowly dried at 30 ◦C for 24 h. As the slow drying process hindered the formation of structural defects induced by rapid evaporation of the solvent, the microdroplets were densely aggregated in thin-film form with high uniformity. For Galinstan, the size of each microdroplet was less than 5 μm, and rod-shaped particles were also observed between the rounded droplets. It is well known that rounded Galinstan microdroplets are surrounded by a thin layer of carbon and Ga2O3, of which the inner core is composed of Ga, In, and Sn [41,42]. As previously reported, the rod-shaped particles possibly consisted of Ga oxide monohydroxide ((GaO)OH) [43]. Note that Ga in Ga-based metal alloys can react with decomposed OH− in the presence of O2, leading to the crystallization of (GaO)OH as follows:

$$2\text{Ga} + 2\text{OH}^- + \text{O}\_2 \rightarrow 2(\text{GaO})\text{OH} \tag{1}$$

The amount of rod-shaped particles is significantly less than that of round particles, and it is expected to be further reduced at low-temperature conditions because the crystallization strongly depends on heat and reactive oxygen species originating from sonication [44,45].

The thin film comprising aggregated Galinstan microdroplets was not electrically conductive because each droplet was fully covered by a non-conductive Ga2O3 layer. Thus, another flat PDMS substrate for protecting Galinstan was brought into contact with the thin film, and an external pressure of 15 MPa was sequentially applied to the sample to collapse the surface Ga2O3 layers and connect Galinstan. With the collapse of the surface Ga2O3 layers, the Galinstan microdroplets were connected to achieve a continuous phase between the two PDMS substrates. After peeling off the upper PDMS substrate from the bottom PDMS substrate, thin Galinstan films were consequently formed on both PDMS substrates (see Figure 2a). The thickness of each glossy film was measured to be less than 1 μm, of which the surface partially cracked due to rapid surface oxidation during the peeling process.

**Figure 2.** (**a**) Photographs of (left) a dried suspension of Galinstan microdroplets in thin-film form on a flexible PDMS substrate, and (right) a conductive Galinstan film fabricated by compression and separation using PDMS. (**b**) SEM images of patterned Galinstan films, of which the minimum line width is ~20 μm. (**c**) Photographs of exemplary Galinstan films, patterned by laser ablation. (**d**) Optical microscopy image and its close-up SEM image of an exemplary Galinstan structure with complex pattern.

A fiber laser marking machine (λ ~ 1064 nm) was employed for the direct patterning of the Galinstan films, enabling the fabrication of accurate and desirable features with sub-100 μm resolution. Various exemplary features of the patterned Galinstan films are shown in Figure 2b–d. The smallest feature size of the Galinstan line was 20 μm. Laser ablation is a fast and precise method for patterning liquid metal electrodes, facilitating the fabrication of complex and hollow patterns. In addition, this light-based patterning method does not cause significant damage to the transparent substrates, such as glass and PDMS, which do not directly absorb the energy from a fiber laser. After completing the laser ablation process, partial buckling was observed on the PDMS substrates.

The electrical resistance and conductivity of the Galinstan film were measured corresponding to the structural deformation of the PDMS substrate. For the measurements of electrical properties, the patterned Galinstan electrode (80 μm × 5 mm) was used as shown in Figure 3. Its thickness was less than 1 μm. The electrical resistance and conductivity were initially measured as 48.3 Ω and ~1.3 × 10<sup>6</sup> S m<sup>−</sup>1, respectively. In response to the deformation ratio, the electrical resistance gradually increased to 73.8 Ω, corresponding to an electrical conductivity of ~8.5 × 10<sup>5</sup> S m<sup>−</sup>1. In comparison with pure Galinstan, in which the electrical conductivity was found to be ~3.5 × 10<sup>6</sup> S m<sup>−</sup>1, the relatively low electrical conductivity of the Galinstan films used in this work could be attributed to the partial cracks and insulating components remaining in the films, such as (GaO)OH and Ga2O3. However, since the majority of the film components were Galinstan, the fabricated films still possessed electrical conductivity high enough to be used as flexible electrodes. It should be noted that the electrical conductivity of the fabricated Galinstan film (~1.3 × 10<sup>6</sup> S m<sup>−</sup>1) is slightly lower than that of the thin EGaIn film (~2.2 × 10<sup>6</sup> S m<sup>−</sup>1) prepared using the same procedure [39]. It is possibly originating from the content of the insulating material in the suspension. As shown in Figure 1a, the content of the rod-shaped particles in the Galinstan suspension is significantly higher than in the EGaIn suspension under our experimental conditions, which may cause a decrease in overall electrical conductivity.

**Figure 3.** Electrical resistance and conductivity of a line-patterned Galinstan electrode (80 μm × 5 mm) corresponding to the structural deformation of the PDMS substrate.

One advantageous property of Galinstan is the liquid phase, maintaining its flexibility and stretchability, even below 0 ◦C. To compare with EGaIn, electrical resistances of the two materials were measured corresponding to lateral stretching (up to ~130%) of the PDMS substrates. For the measurements, the flat Galinstan and EGaIn electrodes were individually prepared on the PDMS substrates (15 × 25 × 1 mm3) and then stretched up to 130% at room temperature and −10 ◦C, respectively. Changes in the electrical resistances upon lateral stretching are shown in Figure 4a. At room temperature, the electrical resistances of both materials slightly increased with the stretching ratios, which is possibly originating from structural deformation [26]. At the temperature of −10 ◦C, the cracks were generated inside the EGaIn film upon the lateral stretches (see Figure 4b), leading to significant reduction in the film continuity. When the stretching ratio was above 110%, the electrical conductivity of the EGaIn film was thus not observed. However, differently from EGaIn, the Galinstan film was still electrically conductive even if the stretching ratio increased up to 130%. These results were caused by the difference in the melting points of the two materials (i.e., <sup>~</sup>−19 ◦C for Galinstan and ~16 ◦C for EGaIn). At −10 ◦C, the EGaIn film in the solid phase was significantly damaged, whereas the Galinstan film in the liquid phase showed excellent film continuity (see Figure 4c).

**Figure 4.** (**a**) Changes in electrical resistances of EGaIn and Galinstan films corresponding to lateral stretching at varied temperatures (i.e., room temperature (RT) and −10 ◦C). (**b**) SEM image of a cracked EGaIn film after lateral stretching below 0 ◦C. (**c**) SEM image of a stretched Galinstan film below 0 ◦C, exhibiting excellent film continuity.

In addition, the Ga2O3 layer, which was spontaneously formed on the Galinstan film, was neatly exfoliated using an elastomeric PDMS stamp for further investigation. It should be noted that inherently high adhesion between the thin oxide shell and PDMS was reported [42], and the elastomeric PDMS stamps enabled intimate contact with the oxide

surfaces. In this work, the mixing ratio of the PDMS precursor was further modified to enhance the adhesive properties, demonstrating excellent contact characteristics with a relatively rough Ga2O3 surface. The elastomeric PDMS stamp was placed on the surface of the Galinstan film without applying any external pressure and was then detached. In this process, the transparent Ga2O3 layer was successfully transferred onto the stamp. The transferred Ga2O3 film was slightly darker than the bare PDMS substrate because small Galinstan residues in the form of islands (<10 μm) remained on the substrate, as shown in Figure 5c. However, all Galinstan residues were entirely isolated from each other and wrapped with Ga2O3, resulting in the formation of a non-metallic film. The thickness of the exfoliated Ga2O3 film was measured as ~13 nm using AFM, as shown in Figure 5b. It is worth noting that the measured thickness of the Ga2O3 film in this work is thicker than that of the single surface oxide layer of Galinstan (i.e., ~3 nm) due to further oxidation during the exfoliation process. The surface roughness and embossed features of the fabricated Galinstan film, as shown in Figure 2d, could also affect the thickness of the Ga2O3 film.

**Figure 5.** (**a**) Schematic illustration of solar-blind photodetector. (**b**) AFM image of transparent Ga2O3 film, exfoliated using a PDMS stamp. The height profile along the white line is also indicated. (**c**) Optical microscopy image of a channel between two patterned Galinstan electrodes. The black areas correspond to Galinstan beneath the transparent Ga2O3 film. Photograph of the fabricated photodetector is also shown in the inset. (**d**) Output characteristics of the solar-blind photodetector under irradiations of visible light (with a halogen lamp; ranging from 350 to 900 nm) and UV light of 365 nm wavelength, respectively. (**e**) Output characteristics of the solar-blind photodetector under deep UV irradiation (254 nm wavelength). The on/off switching of each irradiation was manually performed at 30 s intervals, and the output characteristics were constantly measured at a sample bias voltage of 0.1 V.

To investigate the solar-blind photodetective properties of the exfoliated Ga2O3 film in consideration of its wide bandgap (~4.9 eV) [35–38], a channel between two patterned Galinstan electrodes was bridged using Ga2O3, as shown in Figure 5a. For this work, the conductive Galinstan film, prepared on a large area (50 mm × 50 mm), was patterned by laser ablation to form a channel. The transparent Ga2O3 film, individually prepared on the PDMS stamp, was placed between the channel to complete the device structure (see Figure 5b). The output current was constantly measured at a sample bias voltage of 0.1 V. Under visible light irradiation with a halogen lamp (ranging from 350 to 900 nm), the output current only increased by ~2.7% at a high light intensity of 30 mW cm<sup>−</sup><sup>2</sup> (see

Figure 5c), and no significant change in the output current was observed at lower light intensities. As the contribution of the short-wavelength region in the emission spectrum is not negligible at high intensity, a small increase in the output current could be detected at light intensities above 30 mW cm<sup>−</sup>2. To confirm this speculation, the output current was also measured under irradiation of UV light of a 365 nm wavelength, which was contained in the emission spectrum of the halogen lamp, and the output current in effect increased by ~14.9% at a low light intensity of 0.2 mW cm<sup>−</sup>2. Eventually, under deep UV irradiation (254 nm wavelength) with an extremely low light intensity of 0.1 mW cm<sup>−</sup>2, the output current sensitively increased by up to 15.1% (see Figure 5d). These results strongly sugges<sup>t</sup> that the combination of Galinstan and its surface oxide layers can be used for sensitive solar-blind photodetectors that possess remarkable advantages, such as low-cost and easy processability under ambient conditions, and flexibility.
