**3. Results and Discussions**

Nanoparticles produced using various sparking metals were imaged using AFM, as shown in Figure 2a. Because the lateral (horizontal) sizes of the nanoparticles estimated by AFM show a broadening effect due to a tip shape convolution [15], the nanoparticle sizes were determined by measuring their vertical heights. Due to the random nature of the nanoparticle agglomeration, some particles attached to others and created a secondary particle deposition on the substrate. To determine the primary size of the particles, only primary particles that appeared in the images were used. It can be clearly seen, that a low energy level is needed to melt the metals (Zn, Ag, Au), resulting in a higher density of the particles in the area under observation. The reason for this is that these metals require a lower energy to melt the electrode, and in turn, to melt a nano-droplet. The energy applied to the tips (*Eapp*) is separated to the energy lost to the environment (*Eloss*) and the energy used to melt the metal tips, as described by Equation (1)

$$E\_{app} = E\_{loss} + m \left[ c\_p (T\_m - T\_r) + L\_f \right] \tag{1}$$

where *c<sup>p</sup>* is the specific heat capacity, m is the effective mass of the metal tip, *T<sup>m</sup>* is the melting point, *T<sup>r</sup>* is the room temperature and *L<sup>f</sup>* is the heat of fusion. As shown in Table 1, the melting point of aluminum (933 K) is lower than that of silver (1235 K) and gold (1337 K), however, there a reduced amount of sparked aluminum observed. This is due to aluminum having a remarkably high heat fusion of 0.90 Jg−1K −1 , whereas the high heat fusions of zinc, silver and gold are in the range of 0.13–0.39 Jg−1K −1 . Moreover, aluminum is always covered with a microscopic thin layer of Al2O<sup>3</sup> that is extremely hard and inert. Therefore, this oxide also reduces the creation of nanoparticles during the discharge. Figure 2b shows the height measurement of zinc nanoparticle using section analysis of Nanoscope III (Veeco, New York, NY, USA) software.

The relationship between the nanoparticle heights and the surface energy at melting point of the electrode metals is shown in Figure 3. It can be clearly seen, that the particle heights decreased as surface energy increased. Nevertheless, the heights of all primary particles were less than 4.5 nm, which corresponds with the transmission electron microscopy (TEM) images of sparked nanoparticles published elsewhere [10]. The relationship between the height of the particles and the surface energy is given by a linear formula H = 1.6915 γ + 4.4783, where H is the average particle height (nm) and γ is the

surface energy (J·m−<sup>2</sup> ). This result strongly confirms that nanoparticles on the substrates were nucleated by the molten metal electrodes. V 2183 1.855 0.49 0.31 908 19,652 1.08 Mo 2896 2.080 0.25 1.38 761 14,238 0.92

*Crystals* **2021**, *11*, x FOR PEER REVIEW 4 of 14

**cp (Jg−1K−<sup>1</sup>**

**)**

**k (Wcm−1K−<sup>1</sup>**

Zn 693 0.789 0.39 1.16 244 3,843 3.30 Al 933 0.871 0.9 2.37 830 22,569 3.15 Ag 1235 0.925 0.24 4.29 240 5,319 2.71 Au 1337 1.145 0.13 3.17 158 4,233 2.63 Ni 1718 1.796 0.44 0.91 623 13,228 2.02 Co 1811 1.881 0.42 1.00 599 13,944 1.50 Ti 1941 1.525 0.52 0.22 702 19,136 1.23

**)**

**Lf (Jg−<sup>1</sup> )**

**Lv (Jg−<sup>1</sup> )**

**Height (nm)**

**Table 1.** Metal properties and experimental data [16].

**LV (Jm−<sup>1</sup> )** 

**Metals.**

**T<sup>m</sup> (K)**

**Figure 2.** (**a**) Atomic force microscope (AFM) images of various metal nanoparticles prepared by single spark on glass substrates (**b**) Section analysis to measured height of zinc nanoparticles. **Figure 2.**(**a**) Atomic force microscope (AFM) images of various metal nanoparticles prepared by single spark on glass substrates (**b**) Section analysis to measured height of zinc nanoparticles.

**Table 1.** Metal properties and experimental data [16].


temperature.

**Figure 3.** Plot of the height of sparked nanoparticles as a function of surface energy at melting temperature.

**Figure 3.** Plot of the height of sparked nanoparticles as a function of surface energy at melting Figure 4a,d show sparked particle films of gold and cobalt, respectively, on white paper substrates. The film formed in two semicircle shapes with a gap of 1 mm between them. The semicircle on the right-hand side was formed under the cathode and is larger than the semicircle on the left that formed under the anode. This is caused by the process when, during the sparking, atmospheric gas breaks down and creates the positive ions and free electrons. The bombardment of the anode by the electrons and the cathode by the positive ions melts the electrode surfaces. The thermal energy expands, and a shock wave forms. The resulting high air pressure pushes the melted metal layer away from the elec-Figure 4a,d show sparked particle films of gold and cobalt, respectively, on white paper substrates. The film formed in two semicircle shapes with a gap of 1 mm between them. The semicircle on the right-hand side was formed under the cathode and is larger than the semicircle on the left that formed under the anode. This is caused by the process when, during the sparking, atmospheric gas breaks down and creates the positive ions and free electrons. The bombardment of the anode by the electrons and the cathode by the positive ions melts the electrode surfaces. The thermal energy expands, and a shock wave forms. The resulting high air pressure pushes the melted metal layer away from the electrode gap, and the molten metal therefore splashes away from the gap in a cone-shaped pattern, while the nanoparticles are nucleated. The cone shaped splashing produces the two semicircles on the substrate. The black color of the sparked cobalt film in Figure 4a indicates the presence of cobalt oxide. Furthermore, the sparked gold film is pink, which is a feature of gold nanoparticles [17].

trode gap, and the molten metal therefore splashes away from the gap in a cone-shaped pattern, while the nanoparticles are nucleated. The cone shaped splashing produces the two semicircles on the substrate. The black color of the sparked cobalt film in Figure 4a indicates the presence of cobalt oxide. Furthermore, the sparked gold film is pink, which is a feature of gold nanoparticles [17]. Figure 4b,c show the cobalt anode and cathode electrode tips, respectively, after the sparking process. The erosion of the tip results in the round shape at both tip ends. The results are not the same for the gold anode and cathode tips, as shown in Figure 4e and 4f, respectively. The sparked gold tips are cone shaped. The cone angle of the anode tip is less than that of the cathode tip. The reduced cone angle of the anode tip corresponds with Figure 4b,c show the cobalt anode and cathode electrode tips, respectively, after the sparking process. The erosion of the tip results in the round shape at both tip ends. The results are not the same for the gold anode and cathode tips, as shown in Figure 4e,f, respectively. The sparked gold tips are cone shaped. The cone angle of the anode tip is less than that of the cathode tip. The reduced cone angle of the anode tip corresponds with the smaller area of the particle film produced under the anode (Figure 4d). The shape of the anode is caused by the expansion of the plasma diameter at the anode surface [18]. In addition, the sides of the cobalt tips were covered by nanoporous films. It is believed that some of the metal evaporated as a gas phase and then deposited on the tip surfaces. The sides of the gold electrodes are different from those of the cobalt tips and no deposited nanoporous film was observed because the gold vapor could not oxidize in this setup.

the smaller area of the particle film produced under the anode (Figure 4d). The shape of the anode is caused by the expansion of the plasma diameter at the anode surface [18]. In

some of the metal evaporated as a gas phase and then deposited on the tip surfaces. The sides of the gold electrodes are different from those of the cobalt tips and no deposited nanoporous film was observed because the gold vapor could not oxidize in this setup.

5 mm

5 mm

**Figure 4.** (**a**,**d**) Nanoparticle films on papers which were placed under the sparking gap of cobalt **Figure 4.** (**a**,**d**) Nanoparticle films on papers which were placed under the sparking gap of cobalt and gold, respectively. The electrode after 1000 cycles of spark: (**b**) Co-anode (**c**) Co-cathode (**e**) Au-anode and (**f**) Au-cathode.

and gold, respectively. The electrode after 1000 cycles of spark: (**b**) Co-anode (**c**) Co-cathode (**e**) Au-anode and (**f**) Au-cathode. More studies are needed in order to better understand the formation of particle coating on the electrode wires. In this study, titanium wires were used as sparked electrodes. The anode and cathode titanium wires were sparked 1000 times. The sparked electrode wires are shown in Figure 5a,b for the anode and cathode, respectively. In the pictures, the nanostructure films are seen to almost fully cover the anode surface. On the other hand, most of the surface area of the cathode is covered by the freeze splashing of molten metal droplets. As previously explained, positive ions have a larger mass and transfer more energy to the cathode compared to the energy transferred by electrons to the anode. Therefore, the cathode shows more surface damage than the anode. The molten electrode splashed and then froze very quickly due to the rapid cooling of the room atmosphere. More studies are needed in order to better understand the formation of particle coating on the electrode wires. In this study, titanium wires were used as sparked electrodes. The anode and cathode titanium wires were sparked 1000 times. The sparked electrode wires are shown in Figure 5a,b for the anode and cathode, respectively. In the pictures, the nanostructure films are seen to almost fully cover the anode surface. On the other hand, most of the surface area of the cathode is covered by the freeze splashing of molten metal droplets. As previously explained, positive ions have a larger mass and transfer more energy to the cathode compared to the energy transferred by electrons to the anode. Therefore, the cathode shows more surface damage than the anode. The molten electrode splashed and then froze very quickly due to the rapid cooling of the room atmosphere. Furthermore, the nanoporous film coverage observed on the tip is rather patchy. In addition, during the sparking, the glass substrate was placed 2 mm under the sparking gap to collect the nanoparticles from the metal droplets (the result shown in Figure 7).

Furthermore, the nanoporous film coverage observed on the tip is rather patchy. In addition, during the sparking, the glass substrate was placed 2 mm under the sparking gap to collect the nanoparticles from the metal droplets (the result shown in Figure 7). Figure 5 shows high magnification SEM images displaying the tip surface located 200 m, 400 m and 600 m from the tip end (the positions of the tips are shown in Figure 4). Nanostructures grown on the anode tip (Figure 5a,c,e for the position of 200 m, 400 m and 600 m, respectively) are sponge-like structures composed of numerous irregularshaped and randomly-oriented grains. It is believed that, the nanostructures are the deposition of titanium vapor. The difference in the size and shape of the nanostructures is explained by the influence of different substrate temperature [19,20]. On the cathode tip (Figure 5b,d,f for the position of 200 m, 400 m and 600 m, respectively) no nanostructures can be observed, since the splashing of the melting titanium prevents the deposition of titanium vapor on the cathode tip surface. The results are the opposite for the cobalt, where the cathode tip is fully covered in cobalt oxide nanostructures, as shown in Figure Figure 5 shows high magnification SEM images displaying the tip surface located 200 µm, 400 µm and 600 µm from the tip end (the positions of the tips are shown in Figure 4). Nanostructures grown on the anode tip (Figure 5a,c,e for the position of 200 µm, 400 µm and 600 µm, respectively) are sponge-like structures composed of numerous irregular-shaped and randomly-oriented grains. It is believed that, the nanostructures are the deposition of titanium vapor. The difference in the size and shape of the nanostructures is explained by the influence of different substrate temperature [19,20]. On the cathode tip (Figure 5b,d,f for the position of 200 µm, 400 µm and 600 µm, respectively) no nanostructures can be observed, since the splashing of the melting titanium prevents the deposition of titanium vapor on the cathode tip surface. The results are the opposite for the cobalt, where the cathode tip is fully covered in cobalt oxide nanostructures, as shown in Figure 4c. Therefore, based on the results for gold, cobalt and titanium, it can be concluded that the nanoparticles were nucleated by the droplets of molten metals and deposited on the substrates under the sparking gap for all sparking metal electrodes. However, the nanostructure nucleation caused by the deposition of metal vapor on the sparking tip depends on the properties of each metal.

pends on the properties of each metal.

substrates under the sparking gap for all sparking metal electrodes. However, the nanostructure nucleation caused by the deposition of metal vapor on the sparking tip de-

**Figure 5.** (**a**) Anode-Ti (**b**) Cathode-Ti after 1000 cycles of spark and surface structures on the sparked tips at the position of 200 m (**c**,**d**), 400 m (**e**,**f**) and 600 m (**g**,**h**) from the tip ends for anode and cathode titanium tips, respectively. **Figure 5.** (**a**) Anode-Ti (**b**) Cathode-Ti after 1000 cycles of spark and surface structures on the sparked tips at the position of 200 µm (**c**,**d**), 400 µm (**e**,**f**) and 600 µm (**g**,**h**) from the tip ends for anode and cathode titanium tips, respectively.

Figure 8.

*Crystals* **2021**, *11*, x FOR PEER REVIEW 8 of 14

Figure 6a,b illustrate the SEM images and EDS line scan result of the sparked anode and cathode titanium tips, respectively. At the tip of the anode electrode, a large titanium ratio and a small amount of oxide cover were found. Given that there was no oxygen found at the aperture created by the focus ion beam (FIB), we can conclude that the oxygen was only on the surface. The oxygen content increased in the nanoporous film area due to the formation of titanium oxide in that area. As was the case with the anode, the oxygen content also increased in the cluster of nanoporous material on the cathode. Figure 6a,b illustrate the SEM images and EDS line scan result of the sparked anode and cathode titanium tips, respectively. At the tip of the anode electrode, a large titanium ratio and a small amount of oxide cover were found. Given that there was no oxygen found at the aperture created by the focus ion beam (FIB), we can conclude that the oxygen was only on the surface. The oxygen content increased in the nanoporous film area due to the formation of titanium oxide in that area. As was the case with the anode, the oxygen content also increased in the cluster of nanoporous material on the cathode.

**Figure 6.** SEM–EDX line scans of sparked titanium tip (**a**) anode (**b**) cathode. **Figure 6.** SEM–EDX line scans of sparked titanium tip (**a**) anode (**b**) cathode.

Figure 7 shows the SEM image of the sparked titanium particles that were deposited on the glass substrate, 5 mm under the sparking gap of titanium tips presented in Figure 5. The films were an agglomeration of nanoparticles with an approximate size of 20 nm. The size of the nanoparticles was almost ten times smaller than the size of the grain that occurred due to the vapor deposition on the tips electrodes. The crystallinity of the Figure 7 shows the SEM image of the sparked titanium particles that were deposited on the glass substrate, 5 mm under the sparking gap of titanium tips presented in Figure 5. The films were an agglomeration of nanoparticles with an approximate size of 20 nm. The size of the nanoparticles was almost ten times smaller than the size of the grain that occurred due to the vapor deposition on the tips electrodes. The crystallinity of the sparked films was initially amorphous and then gradually transitioned into the anatase phase of TiO<sup>2</sup> during the annealing at 400 ◦C for 1–5 h, as shown in Raman spectra in Figure 8.

**Figure 7.** Scanning electron microscopy (SEM) of sparked titanium nanoparticles. **Figure 7.** Scanning electron microscopy (SEM) of sparked titanium nanoparticles.

Unless wires are sparked inside of a magnetic field [21], in ambient conditions at atmospheric pressure, nanoparticles collected on the substrate during sparking discharge process will be amorphous. High energy will also break down gas molecules in the air between the sparking gap (even nitrogen) and will create different oxides, nitrides and carbonates [22]. Because of obtaining thermodynamically unstable products, annealing in controlled atmosphere will affect the crystallinity and phase of the sparked nanoparticles [23]. This is the reason why we see different phases in the Raman results of the sparked titanium.

To further explain the action of the sparking process in the air pressure atmosphere, the schematic illustrations of the nucleation of nanostructures were developed and are shown in Figure 9. When the voltage applied to the gaps is higher than the break down voltage, gas molecules in the sparking gap were ionized, and the electrons and ions produced from the neutral molecules migrated towards the anode and cathode, respectively. The bombardment of high energy electrons and ions melted and evaporated the metal tips. The expansion of the electron stream at the anode increased the bombardment area and therefore the sparked anode tips exhibited a low cone angle. For the cathode, positive ions have a larger mass and therefore transfer more energy to the cathode, resulting in a larger melting volume than the anodes. The nucleated molten nano-droplets were blown in the cone shape around the tip and deposited on the substrate which were placed under, next to, and over the gap. The reduced density and hemisphere area of the films under the tips of the anode are consequences of the reduced cone angle and lower molten volume. The growth of the nanoporous structures from the metal vapor on the sparking tips depends on the material and thermal properties and surface reactivity of the metal tips.

atures**.**

titanium.

**Figure 8.** Raman spectra of the sparked titanium nanoparticles films annealed at different temper-**Figure 8.** Raman spectra of the sparked titanium nanoparticles films annealed at different temperatures.

atmospheric pressure, nanoparticles collected on the substrate during sparking discharge process will be amorphous. High energy will also break down gas molecules in the air between the sparking gap (even nitrogen) and will create different oxides, nitrides and carbonates [22]. Because of obtaining thermodynamically unstable products, annealing in controlled atmosphere will affect the crystallinity and phase of the sparked nanoparticles [23]. This is the reason why we see different phases in the Raman results of the sparked

Unless wires are sparked inside of a magnetic field [21], in ambient conditions at

**Figure 9.** Schematic diagram of the nucleation mechanism of the nanostructure deposited by the sparking method.

**Figure 9.** Schematic diagram of the nucleation mechanism of the nanostructure deposited by the sparking method. There are a substantial number of applications involving the sparking discharge process in the field of aerosol science [24]. Even though we provide evidence in this research that most of the nanoparticles are made by the metal melt approach, the sparking dis-There are a substantial number of applications involving the sparking discharge process in the field of aerosol science [24]. Even though we provide evidence in this research that most of the nanoparticles are made by the metal melt approach, the sparking discharge process can still be used for aerosol production and fabric filtration calibration applications, which are now more important than ever in light of the current COVID-19 crisis [25].

To further explain the action of the sparking process in the air pressure atmosphere, the schematic illustrations of the nucleation of nanostructures were developed and are shown in Figure 9. When the voltage applied to the gaps is higher than the break down voltage, gas molecules in the sparking gap were ionized, and the electrons and ions produced from the neutral molecules migrated towards the anode and cathode, respectively. The bombardment of high energy electrons and ions melted and evaporated the metal tips. The expansion of the electron stream at the anode increased the bombardment area and therefore the sparked anode tips exhibited a low cone angle. For the cathode, positive ions have a larger mass and therefore transfer more energy to the cathode, resulting in a larger melting volume than the anodes. The nucleated molten nano-droplets were blown in the cone shape around the tip and deposited on the substrate which were placed under, next to, and over the gap. The reduced density and hemisphere area of the films under the tips of the anode are consequences of the reduced cone angle and lower molten volume. The growth of the nanoporous structures from the metal vapor on the sparking tips depends on the material and thermal properties and surface reactivity of the metal tips.

charge process can still be used for aerosol production and fabric filtration calibration applications, which are now more important than everin light of the current COVID-19 crisis [25]. In Figure 10, the results with the filtration barrier and without the filtration barrier are both presented. The experiment was conducted as previously described, [26], using aluminum wires, with the modification of one output, in which a commercially available surgical mask as a filter is attached onto the aerosol outlet, leading to a differential mobil-In Figure 10, the results with the filtration barrier and without the filtration barrier are both presented. The experiment was conducted as previously described, [26], using aluminum wires, with the modification of one output, in which a commercially available surgical mask as a filter is attached onto the aerosol outlet, leading to a differential mobility analyzer (DMA). The sparking discharge process suitability for the aerosol calibration was compared with the dispersing particle size standardsTM (from Polysciences, Inc. Nanobead NIST Traceable Particle Size Standards). The results indicate that filtration capacity of the sparking discharge is 99.6%; obtained using the following equation.

$$\mathbf{E\_f} = 100 - \frac{number\ of\ particles\ measured\ with\ DMAA\ with\ filter}{number\ of\ particles\ measured\ with\ DMAA\ without\ filter} \tag{2}$$

This was compared by using a 50 nm particle size standard and the results obtained are not significantly different. A 50 nm particle size was used because this was the size that was generated from the sparking aluminum wires with the machine described in our experiment.

Nanobead NIST Traceable Particle Size Standards). The results indicate that filtration ca-

This was compared by using a 50 nm particle size standard and the results obtained are not significantly different. A 50 nm particle size was used because this was the size that was generated from the sparking aluminum wires with the machine described in our

ℎ ℎ

ℎ ℎ (2)

pacity of the sparking discharge is 99.6%; obtained using the following equation.

Ef = 100 −

experiment.

(**b**)

**Figure 10.** Filtration experiment measured by the differential mobility analyzer. (**a**) The aluminum sparked wire without a surgical mask as a filter. The concentration of nanoparticles produced by sparking was 2.32 × 10<sup>7</sup> particles at 1 cm<sup>3</sup> at size of 44.5 nm (**b**) Results of differential mobility analyzer (DMA) from aluminum sparked wire with a filter from a commercially available surgical mask. Concentration of nanoparticle produced by sparking was 9.11 × 10<sup>4</sup> particles at 1 cm<sup>3</sup>at size of 47.8 nm. **Figure 10.** Filtration experiment measured by the differential mobility analyzer. (**a**) The aluminum sparked wire without a surgical mask as a filter. The concentration of nanoparticles produced by sparking was 2.32 <sup>×</sup> <sup>10</sup><sup>7</sup> particles at 1 cm<sup>3</sup> at size of 44.5 nm (**b**) Results of differential mobility analyzer (DMA) from aluminum sparked wire with a filter from a commercially available surgical mask. Concentration of nanoparticle produced by sparking was 9.11 <sup>×</sup> <sup>10</sup><sup>4</sup> particles at 1 cm<sup>3</sup> at size of 47.8 nm.

#### **4. Conclusions 4. Conclusions**

There are two approaches that are utilized in the generation of nanoparticles from a sparking discharge process. The first is an aerosol-based process [27], and the second is a metal melt process, that we used to modify the surface and create nanomaterials [28]. The sparking discharge studies by Kohut et al. [29] describe craters, undulated areas, and dendritic areas of nickel and copper electrodes (rode) in low numbers (one to three) as opposed to our study (100). Their results also support our approach to sparking discharge for nanoparticle generation trough metal melt, with only a small portion of the electrode content being aerosolized for nanoparticle synthesis. This confirms the production of nanoparticles by the sparking discharge methodology and spark discharge apparatus as mainly coming from the metal melt process.

The findings by Domaschke et al. [30] are significant because they support our exclusive use of wires as the electrode material. In their work regarding the sparking discharge process, a high capacitance of the apparatus is also expected in order to improve the energy efficiency of the sparking discharge process. They also measured the effect of the gas flow during sparking discharge on particle size and the concentration of the produced particles, which we did not deal with because we did not use the gas flow during the synthesis of nanoparticles. Our sparking discharge process was done in an ambient atmosphere and under normal pressure conditions, while metal wires were discharged with no air flow.

In summary, we have demonstrated the preparation of nanoparticles using the sparking process of various metal tips at an atmospheric air pressure. It was observed that particle sizes are related to the surface tension of molten metals. The film coated under the sparking gap was produced by molten metal droplets, whereas the nanostructure covered on the sparking tips was formed by the deposition of metal vapors. The model was created to explain the coating pattern, and this model provides a practical tool that can be used to design a sparking machine for large scale application (see Supplementary Information for examples). This sparking discharge machine can be used for the calibration of fabric filtration barriers as a replacement for dispersing particle size standards.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2073-435 2/11/2/140/s1.

**Author Contributions:** T.K. wrote, analyzed, and collected the data, S.R., wrote, collected and analyzed the data, and edited the manuscript E.K., collected the data, N.J., collected and analyzed the data, W.S. collected the data, W.T., collected the data. S.P. collected the data, A.P. collected the data, P.S. (Panupong Sanmuangmoon) collected the data, P.S. (Pisith Singjai) supervised the edit and acquire funding. All authors have read and agreed to the published version of the manuscript.

**Funding:** Postdoctoral Fellowship of Office of Research Administration, Chiang Mai University.

**Acknowledgments:** This research was supported by Chiang Mai. Stefan Ruˇcman would like to thank CMU Presidential Scholarship for Post-Doctoral Fellowship. Special thanks to Assoc. Panich Intra, Research Unit of Applied Electric Field in Engineering (RUEE), https://www.facebook.com/ ruee.rmutl/ Rajamangala University of Technology Lanna for helping us and allowing us to use the differential mobility analyzer (DMA).

**Conflicts of Interest:** The authors declare no conflict of interest.
