1. Introduction
TiO
2, known for its affordability, stability, and abundance on Earth, exhibits excellent properties such as UV light absorption and water splitting capabilities [
1,
2,
3]. However, its application is limited by its large bandgap of 3–3.2 eV, which restricts its absorption to only UV spectrum [
4]. TiO
2 nanoparticles (NPs) have been utilized to reduce its bandgap and shift its color [
5]. Therefore, TiO
2 NPs have found wide-ranging applications in hydrogen generation, solar cells, and photocatalyst to degrade organic pollutants [
3,
4,
5,
6,
7]. Especially for electrochemical nitrate-to-ammonia conversion, Ti-based NPs have been widely acknowledged as the most promising catalysts [
8].
Among the various methods available for synthesizing TiO
2 NPs, pulsed laser ablation in liquid (PLAL) stands out as an effective and environmentally friendly approach. This technique offers numerous advantages, including easy production and extraction, simplified synthetic procedures without the need for multiple steps, reduced reaction time, absence of reducing agents, enhanced laboratory safety, and low toxicity [
9,
10,
11,
12]. Upon laser irradiation of the metal target, the radiation is absorbed by the target surface electrons, leading to instantaneous energy transfer to the lattice [
10]. The high power density results in explosive vaporization of the local surface, creating a plasma plume with a volume defined by the laser spot size and a height of approximately 100 nm within a few hundred picoseconds [
13]. The rapid volume expansion induces a shockwave both in the target and the liquid, causing a phase transition in the target. The interaction between plasma and the liquid vaporizes the liquid, initiating a cavitation bubble primarily composed of liquid vapor. This cavitation bubble encloses a plasma layer with a thin transient metal layer on both sides. Thus, each pulse generates a structure consisting of a bubble cap, a transient thin metal layer between the cap and the plasma, the plasma layer itself, and another transient thin metal layer between the plasma and the target surface. The NPs are mainly expulsed from the bursting of the cavitation bubble. The cavitation bubble grows and collapses typically within the time scale of hundred of microseconds, which becomes a limiting process in NP generation. Furthermore, reflection from the water/air as well as water/target interfaces, refraction, and self-focusing in water lead to a decrease in the effective energy deposited on the target surface compared to the nominal output power [
14]. Therefore, the production rate is low. The scattering of the cavitation bubble, the shielding effect of the plasma layer, and the absorption and re-ablation of the ablated particles reduce further the effective energy deposited on the target [
13]. Stirring a flow liquid has been found to assist in removing generated bubbles and particles compared with a static liquid, thereby improving production [
15]. Attempts to increase production through the application of electrical, magnetic, and thermal fields revealed that only a magnetic field proved effective by increasing the kinetic energy of charged electrons and the plasma density or accelerating plasma breakdown due to the stirring effect [
10,
16]. High laser-scanning speed has been reported to increase production dramatically as the interpulse distance becomes much larger than the bubble size [
14]. Under such conditions, the cavitation bubble is bypassed, and the bypassed bubble and the enclosed plasma layer cannot shield the following laser pulse [
14]. In summery, circumventing the bubble for the following pulse, accelerating the break down of the bubble, increasing the plasma density, removing the generated bubbles and particles have proved to be effective strategies to increase production rate of PLAL. For further scale up, remote control of PLAL has been demonstrated [
17] and the application of an electrical field has been found instrumental in confining the plasma plume as well as collecting particles [
18]. Apart from the self-focusing and filamentation effect [
13], the penetration depth per pulse for ultrashort laser such as fs and ps lasers is relatively shallow, two orders of magnitude lower than that of ns lasers [
19], resulting in lower ablation efficiency [
13]. In this regard, low-power, inexpensive, and compact ns lasers have shown promise in terms of ablation efficiency, with a factor of 8, which is higher than that of high-end ps lasers [
13]. Moreover, laser ablation in water enables the production of small NPs with a narrow size distribution compared to other liquids [
9], exhibiting a log-normal size distribution [
20].
Laser marker is cheap and facile to operate with high energy density and fast scanning speed, which is beneficial to bypass the cavitation bubble and achieve high production. In this work, a laser marker was adopted to ablate Ti target in distilled water to produce NPs. Meanwhile, sonication was introduced to laser NP production and its effect on NPs was investigated, which may accelerate the collapse of the bubble and is compatible with other production enhancement methods.
Ceria-doped Titania NPs have been utilized as a promising efficient drug delivery system [
21]. Ceria and Titania composites have also been active catalysts and photocatalysts for wide applications and have drawn wide attention [
22,
23,
24,
25,
26,
27,
28]. In this study, the doping of ceria in titania was attempted using PLAL, adding to the exploration of this composite material.
2. Materials and Methods
A measuring cylinder with a 1% volume precision was used to measure fixed volume ~500 mL distilled water into the supersonic bath tank. A 100 × 100 × 5 mm 99.995% Ti block was immersed in the water ~10 mm under the water surface and ablated by a pulsed optical fiber laser marker (YLP-M30, Qingdao, China) with wavelength at 1064 nm, 40 KHz repetition rate, pulse duration of 100 ns, and 30 W nominal outpower. The laser beam was focused onto the target using a scanning head equipped with a focusing lens, resulting in a focused beam size of approximately 50 μm in diameter. At nominal output power, the laser fluence was 38 J/cm
2. However, due to reflection, refraction, and non-linear effects in water, the effective fluence deposited on the target surface would be much reduced. The laser ablation was conducted in raster scan mode with the laser head move controlled by a computer. The laser ablation was taken with and without sonication for comparison. The sonication was carried out at a frequency of 40 KHz, matching the laser repetition rate. Laser fluence was also varied at different levels: 7.6 J/cm
2, 19 J/cm
2, 28.5 J/cm
2, and 38 J/cm
2. The scanning rate was 10 m/s with line spacing of 0.06 mm and scanning duration was fixed 0.5 h for all processing conditions. In addition to the pulsed optical fiber laser marker, a pulsed UV laser marker (XC-XMU5W, Qingdao, China) with a wavelength of 355 nm, a repetition rate of 40 KHz, and a nominal output power of 5 W was also employed. The raster scan settings were the same as those used for the optical fiber laser, except for the pulse duration, which was set at 1 microsecond. At such settings, the laser fluence was about 6 J/cm
2. After each round of TiO
2 NPs fabrication, the entire fluid was transferred into a standard jar for storage.
Figure 1 shows typical colloids synthesized with and without sonication in the storage jars. Sonication resulted in deep color which suggested high NPs concentration and less large precipitates at the bottom.
In order to dope ceria into Ti-based particles, ceria powders with a diameter of 10 μm were dispersed into the water prior to laser ablation. The mass of the powder was varied for comparative analysis, specifically at 0.025 g and 0.1 g measured by a micro-balance, corresponding to concentrations of 50 ppm and 200 ppm in the water medium. All the laser ablation parameters were the same for both.
The liquid sampling drop from the stored colloids was dispensed onto a Al substrate by a pipette followed by drying in an oven for measurement. Scanning electron microscope (SEM) images were taken in Sigma300 from CARL ZEISS (Jena, Germany) with a resolution of 1 nm for extracting the morphology of laser ablated NPs. Energy dispersive X-ray spectroscopy (EDS) mappings were also performed to extract composition distribution over the selected regions. EDS of selective points was also measured. Bruker D8 Advance was adopted to extract X-ray diffraction (XRD) spectra of the colloidal solutions. A laser scanning confocal microscope with the model number OLS5100 from Olympus Corporation (Tokyo, Japan) was employed to examine 3D profiles of the ablated target surface with and without sonication to obtain the ablated volumes for comparison of the production rates. The ablated volumes were extracted from the attached software. The UV-vis spectrophotometer (UV2600i by Shimadzu, Kyoto, Japan) was used to measure the transmission of the colloidal solutions in the wavelength range of 190 nm to 1100 nm. Nanoparticle size distributions were analyzed by means of a dynamic light scattering analyzer (Malvern Zetasizer Nano ZS, Malvern, UK).
3. Results and Discussion
SEM images of the laser-produced Ti particles at varying laser fluences with and without sonication are shown in
Figure 2. The images demonstrate that all particles were regularly spherical. At a low fluence of 7.6 J/cm
2, the particle size exhibited a wide distribution, with the majority of particles appearing as NPs. Particle size increased with laser fluence until the plateau fluence 28.5 J/cm
2 followed by a size decrease at 38 J/cm
2. This decrease in particle size at higher fluence can be attributed to the further breakdown of the ablated particles under the influence of high laser energy. Upon close examination at high magnification in
Figure 2e (at 7.6 J/cm
2), it can be observed that the laser fluence produced NPs with a majority well below 100 nm, and some even as small as approximately 20 nm.
Figure 3a presents size distribution of particles produced at 38 J/cm
2, which is consistent with
Figure 2d.
Figure 3a demonstrates the maximum abundance of particles occurs around 600 nm, a size range that aligned with ~600 nm sized particles shown in
Figure 2d.
Figure 3b,c show two additional DLS results for colloidal solutions processed at 7.6 J/cm
2 without and with sonication, respectively. Sonication generally produced smaller particles with narrower size distribution, corroborating the findings depicted in
Figure 2e,f. It is essential to acknowledge a slight discrepancy between the particle distribution data obtained from DLS spectra and SEM images. SEM provides insights into only a relatively small, localized portion of the particle distribution and, as a result, may sometimes contradict the overall particle distribution observed through DLS analysis. These disparities can be attributed to the inherent limitations of each technique, wherein SEM offers higher resolution imaging of specific small portion of particles, while DLS provides a more comprehensive characterization of the entire colloidal solution. It is important to note that only the spherical particles in
Figure 2e,f were Ti-based particles, which are highlighted with circles. The non-spherical particles were contaminants, which would be verified in
Figure 4.
Figure 2e took much attention due to its low concentration and sparse distribution of particles. Conversely,
Figure 2f was easier to capture as the concentration of particles was higher. The stirring effect of sonication increased the chances of laser interaction with the particles, leading to further breakdown and size reduction. Sonication also likely played a role in splitting bubbles and expulsed melts generated during the laser ablation process, contributing to particle size reduction [
10]. Furthermore, sonication might potentially expedite the collapse of the cavitation bubble and plasma plume [
10], enhancing the particle production rate. This effect arose from the operating frequency of sonication, which was nearly one order magnitude higher than that of cavitation bubble burst. Given that both the cavitation bubble and the plasma layer exist in a metastable state, a minor perturbation such as high-frequency sonication is sufficient to disrupt this state, resulting in the expulsion of NPs. This could explain why sonication led to a deep-colored colloid in
Figure 1, indicating a higher concentration of particles.
Figure 4 shows a SEM image of a different region of the same sample presented in
Figure 2f and EDS spectra of selected non-spherical particles.
Figure 4 suggests that the non-spherical particles were mostly salts (Na and Cl peaks) or organic contaminants (C peak), which might come from fingerprint residuals during sample preparation for measurement and airborne pollutants.
Figure 5 showcases confocal images of treated targets with and without sonication. The darker regions represent the ablated areas, while the lighter regions represent the untreated intact parts. The introduction of sonication resulted in deep, ablated channels and a pronounced contrast between the ablated and intact regions. The Z dimension values further indicate that sonication led to a larger ablation volume and a higher production rate. Specifically, the ablated volumes of the regions marked in
Figure 5c,d were measured to be 0.46 mm
3 and 0.26 mm
3, respectively. These findings align consistently with the observations presented in
Figure 1. It is important to note that the observed ablated channels in our study differed from the craters described in the literature. However, the objective of this measurement was consistent with the literature, which aimed to assess the production rate of nanoparticles [
20,
29].
Figure 6 shows transmission of reference water and colloidal solutions produced with and without sonication. In comparison to the reference water, the transmission of colloidal solutions decreased thanks to scattering, reflection, and absorption of the nanoparticles dispersed within them. Sonication resulted in a further reduction in transmission. This can be attributed to the enhanced production rate of nanoparticles facilitated by sonication, a phenomenon consistent with the findings presented in
Figure 1 and
Figure 5.
Figure 7 presents EDS mapping of the particles depicted in
Figure 2b and it indicates that the majority of the Ti-based NPs were largely oxidized, indicating the possible presence of TiO
2. This oxide was the reaction product between highly active ablated Ti and O in water. O and Ti signal did not exactly match the spherical particles because the signal accumulation time was not adequate.
Figure 8 depicts EDS analysis of a particle from a representative ablation and the Al substrate. The crossings labeled with numbers 67 and 68 represent the selected particle and the Al substrate, respectively, for EDS measurement. The C/O/Al atomic ratio for Al substrate was found to be 2.3:2.4:95.3. As for the selected particle, the C/O/Al/Ti atomic ratio was determined to be 1.7:61.3:1.6:35.4, where C was mainly attributed to airborne pollutants and may be bonded with O. EDS typically has a detection depth in the range of microns. Additionally, Al forms a very thin native protective oxide layer on its surface, usually at the level of a few nanometers [
30]. Due to this thin oxide layer, the majority of Al-bonded oxygen is primarily associated with surface-bound Al. EDS also has its detection resolution generally ~1% [
31]. Consequently, the atomic concentration of O bonded to Al and C was assumed to be relatively constant for both
Figure 8b,c, estimated to be approximately 2% based on measurements. As a result, the atomic ratio between O bonded with Ti and Ti was determined to be 59.3:35.4, equivalent to 1.68. This finding indicates insufficient oxidation of Ti, with approximately 16% of the metal remaining unoxidized. Additionally, the observation suggests a core–shell structure of the particle, featuring a TiO
2 shell encompassing a pure Ti core.
Figure 9 shows XRD spectra of colloidal solutions, which were fitted using Origin. It indicates the presence of rutile nano-crystal. According to Scherrer equation [
32]:
where D is the particle diameter,
K is constant,
is the wavelength of X-ray, 0.154 nm,
is full width half maximum (FWHM),
is the diffraction angle. The narrow, fitted red peaks observed in
Figure 9a,b provide valuable insights into the nano-crystal sizes of approximately 1 nm and 1.4 nm, respectively, for rutile. In contrast, the sharp Ti peak represented by the black line in
Figure 9b, as opposed to the fitted blue one, corresponds to a larger nano-crystal size of approximately 5 nm for Ti. These calculated sizes are notably smaller than the particles depicted in
Figure 2b,c.
This intriguing observation suggests that the particles’ outer shell is likely composed of amorphous Ti oxide, while the core contains a small amount of nano-rutile and nano-Ti. This structural arrangement may explain the significant difference in size between the crystalline structures observed in
Figure 9 and the larger particles shown in
Figure 2b,c. Furthermore, it is important to recognize that stress-induced diffraction peak broadening during the particle formation process may provide an alternative explanation for the observed phenomenon. The sudden changes in pressure and temperature experienced during particle formation can induce stress, leading to diffraction peak broadening, which is beyond the scope of application of Scherrer equation.
Figure 10 demonstrates that the implementation of the UV laser generally led to smaller particle size compared to its 1064 nm counterpart. Sonication proved, once again, to lead to a reduction in particle size and a narrow size distribution, as is evident from the comparison of the sonicated and non-sonicated samples. It is worth noting that sonication operated at the same frequency as the laser, which could potentially enhance the ablation of particles.
Figure 11 illustrates EDS mapping of particles depicted in
Figure 10b. It indicates that the majority of the nano-particles were predominantly oxidized, confirming the presence of TiO
2. It is worth noting that only the spherical particles corresponded to Ti-based particles, as highlighted in the mapping results.
Figure 12 presents EDS spectrum of a selected particle produced with 0.025 g ceria. It suggests the successful doping of ceria into titania particle. The Al peak was from the substrate, as EDS detection depth was in a few micrometers level.
Figure 13 presents EDS spectrum of a selected particle produced with 0.1 g ceria. It suggests that ceria became the majority and titania turned into the doping minority. It reveals that even in a very high diluted concentration 200 ppm in water, ceria was readily further broken down by the laser ablation, partially blocking the ablation of Ti target underneath.