Investigation of Overflow-Water-Assisted Femtosecond Laser-Induced Plasma Modulation of Microchannel Morphology

Abstract

Laser-induced plasma micromachining (LIPMM) process is an effective approach to create microfeatures with high aspect ratio (AR) and reduced heat affected zone (HAZ). Therefore, LIPMM plays a crucial role in improving the morphology of microchannels. In this study, microchannels were fabricated using a femtosecond laser with two distinct sets of process parameters under three different processing methods: overflow-water-assisted laser-induced plasma micromachining (OF-LIPMM), laser direct writing (LDW), and static water laser-induced plasma micromachining (S-LIPMM). Furthermore, single-factor experiments were conducted to systematically analyze the effects of four parameters, namely single-pulse energy, scanning speed, scanning times, and frequency, on the HAZ, AR, and material removal rate (MRR) of the microchannels. Finally, the optimized parameters determined from the single-factor experiments were applied for large-scale grid fabrication on a surface. The experimental results revealed that OF-LIPMM enables the creation of two different kinds of microchannel surfaces: one microchannel was fabricated with a higher AR of 3:1 and a larger HAZ, while another microchannel was created with a lower AR of 1:1 and a reduced HAZ. Moreover, the parameters investigated in the single-factor experiments can be applied to large-scale processing. The results also indicate that variations of the scanning speed, frequency, and single-pulse energy have similar effects on the machining characteristics of the three processing methods. The findings enable the generation of microchannels with favorable morphological characteristics and have significant implications for the large-scale production of both types of microchannels.

1. Introduction

With the extensive utilization of microstructures in microfluidics, microelectromechanical systems (MEMS), and other microsystems, the ability to fabricate high-quality microstructures on diverse material surfaces has become of utmost importance. Research has revealed the crucial roles of microchannels with high aspect ratios (ARs) and minimized heat affected zones (HAZs) in achieving optimal performance in optical devices, biomedical instruments, micromechanical systems, nanofabrication, and high-precision electronic chips [1,2,3,4]. Moreover, microchannels with high ARs play a vital role in providing expanded operational space, reducing power consumption, and enabling fluid control. Additionally, microchannels with limited HAZs ensure the preservation of material characteristics and offer substantial advantages in achieving precise machining, processing micro features, and optimizing surface quality. In summary, microchannels serve as fundamental units in microfluidic devices and find extensive applications in the biomedical, chemical analysis, energy transmission, and various other domains. However, conventional microchannel fabrication methods are compromised by limited precision, high costs, and sluggish processing speeds. Therefore, the pursuit of a novel and efficient microchannel fabrication approach holds great importance.

Currently, various methods are employed for fabricating microchannel structures, including precision mechanical machining [5,6,7], microelectrical discharge machining, microelectrochemical machining, photolithography, ion beam machining, and laser machining [8,9]. For laser processing, short-pulse and ultrashort-pulse lasers have garnered significant recognition as a microfabrication technique [10]. The capability of ultrashort-pulse lasers to minimize damage and accomplish precise processing has been extensively investigated for numerous years [11,12,13,14], primarily attributed to their capacity to generate high-quality, high-precision microfeatures with high peak power densities and ultrashort interaction times. Furthermore, due to the ultrashort pulse duration, high processing precision, wide material processing range, and minimal thermal effects, femtosecond lasers have emerged as a promising tool for micro/nanomanufacturing in recent decades [15]. Rizvi et al. [16] summarized the progress of femtosecond laser micromachining of metal, glass, diamond, porcelain, various polymers, etc., and confirmed that femtosecond machining is an ideal micromachining tool from all aspects. However, the relatively high equipment cost and low material removal rate (MRR) impose restrictions on their application. Moreover, when processing brittle or transparent materials, the low laser absorption coefficient significantly reduces the processing efficiency.

In recent years, pulsed laser ablation in liquid has emerged as a promising technique for improving the processing performance of femtosecond lasers. In contrast to laser ablation in atmospheric conditions, the presence of a liquid medium offers several advantages for laser machining. When a femtosecond laser is employed with sufficient laser energy to induce plasma, plasma is simultaneously formed on the material surface and in the surrounding liquid. During the laser ablation process, the presence of the liquid layer could restrict the expansion of plasma, thereby prolonging the duration of the plasma and ultimately increasing the MRR. Based on this concept, Pallav and Ehmann [17] proposed laser-induced plasma micromachining (LIPMM), a high-processing-efficiency, high-MRR machining technique that eliminates the need for complex external tools. LIPMM enhances laser absorption efficiency and shows great potential for multimaterial processing. The method involves focusing a laser beam onto the surface of the workpiece immersed in a liquid, causing the decomposition of the dielectric material, and generating a plasma flow. The laser-induced plasma reaches extremely high temperatures, showing a Fermi–Dirac distribution law [18], exceeding the melting and vaporization thresholds of the workpiece material, enabling large-scale material removal when it impinges on the workpiece surface [19]. However, traditional LIPMM commonly utilizes static water laser-induced plasma micromachining (S-LIPMM), which poses several limitations. Firstly, static water is ineffective in efficiently removing the debris and bubbles generated in the machining area [20], resulting in compromised machining quality. Additionally, the presence of static water often leads to significant oscillations during the machining process, further affecting the stability and precision of the process, thereby compromising the achievement of larger AR and MRR.

To achieve microchannels with greater ARs, higher MRRs, larger and more stable processing scales, and improved surface morphology, researchers have investigated diverse approaches in the field of LIPMM. For instance, Malhotra et al. [21] manipulated the laser beam to create linear plasmas, which resulted in higher MRR. Furthermore, they developed magnetically controlled LIPMM to further enhance processing efficiency. Wolff and Saxena [22], Saxena et al. [23], and Tang et al. [24] also employed external magnetic fields to confine the created plasma, restricting its diffusion and consequently achieving smoother surfaces with reduced heat defects and increased ARs of the microchannels. Saxena [25] employed a saltwater medium instead of distilled water during LIPMM. By optimizing the composition of the liquid layer using saltwater, an MRR of about 97% was observed. However, all these studies either involved complex equipment or had limited processing ranges, restricting their practical application in engineering scenarios. While various methods, including salt- or alcohol-assisted techniques [26], have shown some effectiveness in LIPMM, flowing water stands out as a particularly advantageous auxiliary liquid. It offers several key benefits, including cost-effectiveness, environmental friendliness, and efficient heat absorption. These advantages make flowing water a preferred choice as an auxiliary liquid in LIPMM processes. Moreover, it offers the advantage of a simple setup and easy plasma generation. Therefore, flowing-water-assisted laser-induced plasma micromachining (F-LIPMM) techniques emerge as the most easily accessible approach for industrial processing.

Based on the aforementioned discussion, it is evident that the F-LIPMM technique, which introduces a flowing water layer on top of S-LIPMM, indicates a commonly used and highly effective auxiliary process. Huang et al. [27] incorporated a water nozzle near the clamped workpiece surface to create a tangential water-assisted setup, generating a flowing water layer with a thickness of approximately 1 mm. This approach successfully produced low HAZ microchannels with a 1:1 AR. Wang et al. [28] developed a coaxial water-assisted setup utilizing an optical window and a cylindrical water chamber, allowing laser transmission through a high-speed water column to the workpiece surface, resulting in microchannels with no protrusions and minimal HAZs. Wang et al. [29] employed an immersion water-jet-assisted method, where the laser was focused on the workpiece surface through a lens, and a high-speed water jet perpendicular to the submerged workpiece surface impacted the machining area. This technique produced microconical holes with a cone angle 376% smaller than those created by laser micromachining in air. However, the aforementioned flowing-water-assisted devices face challenges, such as significant water flow fluctuations, substantial losses due to water layer instability, limitations in attaining significant ARs, substantial MRRs, and expansive processing scales. In comparison to the conventional F-LIPMM that employs a side-axis auxiliary nozzle to generate high-velocity water jets, the OF-LIPMM utilizes gentle surface water assistance by introducing water from the bottom and expelling it around the periphery. This approach demonstrates significantly improved processing outcomes.

Utilizing overflow water as an auxiliary material in laser-induced plasma processing offers numerous advantages. These include improving processing quality, expanding processing scale, enhancing AR and MRR during processing, and providing environmental and economic benefits. In addition, it can be applied to plasma to modulate microchannel morphology. Taking all these factors into consideration, it is evident why OF-LIPMM is chosen for microchannel fabrication in this study. By comprehensively adjusting the process parameters in OF-LIPMM, this research successfully achieved microchannels with a high AR and a low HAZ. The effects of different process parameter variations on microchannel processing efficiency and surface morphology were evaluated through an analysis of AR, MRR, and HAZ. The comparison of OF-LIPMM with channels created by laser direct writing (LDW) and S-LIPMM processes demonstrates the role of OF-LIPMM in improving microchannel surface morphology and processing efficiency. Moreover, the trends of HAZ, AR, and MRR variations remained consistent among the four process parameters. The findings of this study extend the application of the overflow-water-assisted method, and thereby lay a foundation for further advancements in the field of micromanufacturing.

2. Experiment Details

2.1. Experimental Setup

The schematic diagram of the OF-LIPMM experimental setup used in this study is illustrated in Figure 1. It consists of several key components, namely a femtosecond laser, an optical system, a water-assisted system, and a displacement platform. The femtosecond laser employed is a FemtoYL-20 model, a semiconductor-pumped solid-state laser (YSL Photonics Co., Ltd., Wuhan, China) with a wavelength of 1030 nm and a pulse width of 400 fs. The optical system, on the other hand, is primarily composed of a beam expander, an aperture, mirrors, and a scanning mirror system. As the laser beam passes through the aperture, it undergoes expansion and is directed towards the beam expander. The expanded beam then traverses through two mirrors before reaching the scanning mirror system located at the end of the optical path. The scanning mirror system incorporates an F-theta lens at its base, which enables the laser beam to converge and form a high-quality Gaussian beam with a diameter of 35 μm precisely at the focal point on the surface of the workpiece. To ensure precise control over the laser processing, parameters such as scanning speed, frequency, and scanning times need to be set using our proprietary computer control software. This software has been specifically developed for in-house laboratory use. Furthermore, the water-assisted system is installed on the displacement platform and comprises an acrylic container, U-shaped pillars, a pump, and pipes. These components work in synergy to facilitate the delivery and circulation of water during the laser processing.

In the course of conducting OF-LIPMM experiments, as illustrated in Figure 2c, the workpiece was securely fastened to the U-shaped bracket. Water was pumped into the bottom orifice, forming a stable overflow water layer on the front surface of the workpiece, with a velocity of approximately 0.5 m/s. Subsequently, the water overflowed from the surroundings and entered the water tank, and it was then recirculated using a water pump. Within this circulation, the water flow facilitated by the pump ensured the removal of generated bubbles and debris during the machining process. Employing a bottom inlet and edge outlet configuration enabled a smoother flow of water across the workpiece surface. As a point of contrast, in S-LIPMM experiments, as depicted in Figure 2b, the bottom orifice was blocked to create a static water condition on the workpiece surface. For LDW experiments, as shown in Figure 2a, it was necessary to evacuate the water from the container to allow direct laser–air interaction for LDW. It should be noted that experiments conducted in air were performed under standard atmospheric conditions without the use of any auxiliary gases.

2.2. Experiment Design

2.2.1. Comparison of Different Methods for Achieving High AR Microchannels

In order to achieve high AR microchannels, we conducted experiments using 0.7 mm 304 stainless steel as the test material, employing the LDW, S-LIPMM, and OF-LIPMM techniques. The experimental parameters are presented in

Table 1. Parameters used for the high AR microchannels.

Medium	Air, Static Water Layer, Overflow Water Layer
Scanning speed (mm/s)	0.1, 0.5, 1, 1.5
Single-pulse energy (μJ)	10, 12, 14, 16, 18, 20, 22
Frequency (kHz)	70, 80, 90, 100, 110, 120, 130
Scanning times	6, 8, 10, 12, 14

. Single-factor experiments were performed on four experimental parameters (frequency, single-pulse energy, scanning speed, scanning times) that affect the HAZ, AR, and MRR. A comparison was made among different operating conditions, with particular focus on the variations observed in OF-LIPMM compared to the other two common conditions. Considering the thicker water film, a higher overlap rate of the laser beam was used due to the higher energy loss caused by refraction within the water layer. The calculation formula for the overlap rate is given by Equation (1). The experimental setup included a lower velocity range of 0.1–1 mm/s and a larger frequency range of 70–130 kHz. Regarding the single-pulse energy, a wider range of 10–22 µJ was chosen to capture more pronounced trends. The number of scanning times was set between 6 and 14 to achieve greater depth while ensuring reasonable processing time.

$$S_{over}=\left(1-\frac{v_{scan}}{f_{rep\times }{d}_f}\right)\times 100\%$$

(1)

where S_over represents the laser beam overlap rate, v_scan corresponds to the scanning speed, f_rep denotes the laser frequency, and d_f represents the laser beam radius.

2.2.2. Comparative Experiments on Microchannel Acquisition with a Low HAZ

To obtain microchannels with a low HAZ, we chose a smaller overlap rate to minimize the generation of ablation debris and cavitation bubbles. Consequently, a relatively higher scanning speed of 0.1–4 mm/s and a lower overall frequency of 10–35 kHz were selected. The energy values, as shown in

Table 2. Parameters used for the low HAZ microchannels.

Medium	Air, Static Water Layer, Overflow Water Layer
Scanning speed (mm/s)	0.1, 0.5, 1, 1.5, 2, 3, 4
Single-pulse energy (μJ)	9, 11, 13, 15, 17, 19
Frequency (kHz)	10, 15, 20, 25, 30, 35
Scanning times	11, 13, 15, 17, 19, 21

, are slightly lower than those in Table 1. This adjustment aims to reduce the overall impact of single-pulse energy on material ablation and achieve a lower HAZ. In terms of the number of machining pulses, the increased scanning speed allows for a greater number of pulses to compensate for any deficiencies in energy, speed, and frequency, thus achieving a greater channel depth. Furthermore, a horizontal comparison with traditional S-LPMM and LDW was conducted to demonstrate the superior performance of the improved OF-LIPMM in meeting the desired metrics, particularly the low HAZ.

2.3. Characterization and Measurement Methods

After the completion of the machining process, the samples were first subjected to ultrasonic cleaning using anhydrous ethanol, followed by drying with oil-free compressed air. Subsequently, the bottom surface topography, microchannel cross-sectional profile, surface roughness, and milling depth were observed and measured using a 3D laser confocal scanning microscope (OLS-4100; Olympus, Tokyo, Japan). Additionally, the ablation depth and ablation width of the microchannels on all processed surfaces were recorded. In order to assess the efficiency of laser milling microchannels, this study introduces the HAZ, AR, and MRR as the primary evaluation indicators for microchannels.

The HAZ is the width of the discolored area around the microchannel measured from the workpiece surface considering the combined effects of cavitation bubble blow-up, laser pulse ablation, and deposition of untaken debris. The AR is the ratio of the obtained microchannel width to depth as shown in Equation (2), and the size of the AR can be judged to divide microchannels with high and low ARs. Due to the synergistic effects of material removal expansion in LIPMM, the radial propagation of plasma, mechanical fractures resulting from shock waves, and thermal vaporization due to high-temperature plasma contact, the MRR is chosen as the evaluative parameter. MRR is calculated by dividing the volume of removed material by the processing time, as shown in Equation (3). The magnitude of MRR serves as an indicator of laser processing efficiency.

$$\mathrm{A}\mathrm{s}\mathrm{p}\mathrm{e}\mathrm{c}\mathrm{t} \mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}=\frac{D}{W}$$

(2)

$$\mathrm{M}\mathrm{R}\mathrm{R}=\frac{V}{t}=\frac{A\times L}{t}$$

(3)

where D represents the measured depth of the microchannel after processing, and W represents the channel width in Equation (2). In Equation (3), A denotes the cross-sectional area of the processed microchannel, which is calculated using the contour profile of the microchannel. L represents the length of the scanning process, and t is the laser processing time. In the subsequent investigation, we conducted a detailed exploration of the variations in the HAZ, AR, and MRR of the microchannels after laser milling in response to changes in laser processing parameters.

3. Results and Discussion

3.1. Results of Microchannel Parameters with High AR

In the manufacturing process of high AR microchannels, a high overlap rate effectively increases the AR and MRR of the microchannels. However, due to the mismatch with water velocity, the bottom of the microchannels becomes rough, resulting in a large HAZ. The representative microchannels obtained from the experiments are shown in Figure 3. Figure 3a represents the three-dimensional view of the microchannel with the maximum AR, while its cross-sectional view is depicted in Figure 3b. Figure 3 illustrates the results obtained under the following parameters: single-pulse energy of 16 µJ, frequency of 100 kHz, scanning speed of 0.1 mm/s, and 10 scanning times. From Figure 3b, it can be observed that the microchannel has a width of approximately 110 µm and a depth of around 330 µm, resulting in an AR of three for the microchannel.

Figure 4a–c illustrate the three-dimensional confocal images obtained under three different methods with a single-pulse energy of 10 µJ, laser frequency of 100 kHz, a speed of 0.1 mm/s, and 10 scanning times. Figure 4d–f depict the two-dimensional and three-dimensional confocal images obtained at a single-pulse energy of 22 µJ, laser frequency of 100 kHz, a speed of 0.1 mm/s, and 10 scanning times. From Figure 4, it is evident that increasing the energy leads to a significant increase in both the depth and width of the microchannels for all three methods, resulting in an enlarged HAZ. A comparison between Figure 4a,d reveals that the microchannels processed with OF-LIPMM are more uniform and complete. However, in Figure 4b,e, the microchannels produced by S-LIPMM display intermittent breaks at the bottom. In Figure 4c,f, it can be observed that during LDW, the molten residue accumulates on one side, causing an uneven surface. Regarding the issue of water flow causing a larger HAZ compared to static water, we believe that lower flow rates can carry away some of the ablation debris. However, due to excessive overlap, high plasma energy, and the large amount of debris generated by instantaneous ablation, the debris produced exceeds what is carried away, resulting in a larger HAZ. However, in S-LIPMM, all the debris accumulates on the surface and cannot be carried away, causing a loss of energy directly on the surface without ablating the material. This phenomenon contributes to the observation that, in general, the HAZ of OF-LIPMM is larger than that of S-LIPMM in the final experimental results.

Figure 5a–c depict the results obtained with a single-pulse energy of 16 µJ, laser frequency of 100 kHz, scanning speed of 0.1 mm/s, and 10 scanning times, while Figure 5d–f show the results obtained with a speed of 1 mm/s while keeping the other processing parameters consistent with Figure 5a–c. By varying the scanning speed and adjusting the spot overlap, processing efficiency can be improved. Upon comparing Figure 5a,d, we observe that although the AR of the microchannels decreases, the microchannel morphology becomes more regular, and the HAZ reduces in size. This phenomenon can similarly be observed in Figure 5b,e. However, the processing results for S-LIPMM are notably inferior to those for OF-LIPMM, exhibiting the intermittent discontinuities often encountered in S-LIPMM. Conversely, the effects are less prominent in LDW, indicating that parameter variations have a significantly smaller impact on femtosecond laser direct writing than on the other two methods.

The phenomenon of larger HAZs in underwater processing compared to LDW can be attributed to three primary factors: (1) The instability of the water layer caused by the movement of the laser processing position affects the processing accuracy, causing slight deviations in the subsequent laser scanning from the previously processed microchannels, thereby manifesting a larger HAZ. (2) Errors arising from laser propagation: Firstly, during point-by-point scanning of a material layer, femtosecond lasers traverse three distinct mediums—air, water, and the processed material—each with varying refractive indices, causing aberrations. These aberrations lead to variations in the excitation intensity and size of the processing spot, affecting microfabrication precision. Secondly, due to the influence of focusing lens diffraction effects and laser beam waist, the intensity distribution of the processing spot exhibits an ellipsoidal shape in the X–Y plane (lateral) and an elliptical shape in the Y–Z plane (axial). This increases interlayer spacing along the axial direction and reduces the density of processing points within the same layer in the lateral direction, resulting in decreased processing resolution, impacting precision and surface quality, thereby contributing to a larger HAZ. (3) Accumulation of thermal effects: With a higher repetition frequency of laser pulses, the interval between pulses becomes shorter than the thermal relaxation time, leading to increasingly prominent thermal accumulation between pulses.

For LDW, when the femtosecond pulse laser interacts with the material, the material in the interaction area experiences a rapid temperature rise, causing vaporization and taking away the majority of the heat. As a result, the energy in the heated region does not have sufficient time to diffuse, thereby having minimal impact on the non-interaction area, and thus demonstrating the smallest HAZ phenomenon.

Figure 6a–c present microchannels fabricated with a single-pulse energy of 16 µJ, laser frequency of 100 kHz, scanning speed of 0.1 mm/s, and 10 scanning times, while Figure 6d–f show the results obtained by solely changing the laser frequency to 130 kHz while keeping the other parameters constant. Comparing the upper and lower rows of Figure 6, we observe that increasing the laser frequency leads to an enlargement of the HAZ. Comparing Figure 6a with Figure 6b,d with Figure 6e, we can see that similarly to the effect achieved by increasing the energy, the bottom width of the microchannels increases, but the processing at the bottom becomes more uniform. Additionally, the influence of different parameters on the processing results is similar across the three conditions, while having a minor impact on static water and air. Similar to the effects observed with changes in energy and scanning speed, S-LIPMM also exhibits non-uniform depths in the bottom of the microchannels. This phenomenon is attributed to the presence of debris and bubbles covering the water surface in the channel, resulting in significant energy loss during transmission and causing uneven depths in the microchannels.

The microchannels fabricated by OF-LIPMM exhibit remarkable uniformity and smoothness without any protrusions. This is attributed to the utilization of a flowing water layer, which effectively washes away the debris generated in the plasma processing region, preventing the formation of protrusions on one side of the microchannels. Additionally, the erosion and vaporization of bubbles prevent molten material from adhering to the sidewalls and inlet of the processed microchannels, thus avoiding the formation of recast layers.

3.1.1. Evaluation of HAZ Size Criteria

In the comparative experiments for achieving high AR microchannels, the variations of these four parameters exhibit consistent trends in influencing the HAZ under three different methods. Additionally, the width of the HAZ decreases sequentially in the order of OF-LIPMM, S-LIPMM, and LDW. Notably, OF-LIPMM consistently yields the widest HAZ. As illustrated in Figure 7a, as the single-pulse energy increases from 10 µJ to 22 µJ, the ablative energy and thermal influence intensify, resulting in an increased HAZ size. Furthermore, the width of the HAZ shows an increasing trend across all three methods. Additionally, it can be observed in the subsequent figure that OF-LIPMM demonstrates the most noticeable variations under different laser process parameters, whereas the other two methods remain comparatively stable. The authors posit that this phenomenon is primarily due to limitations in material removal resulting from surface cavitation bubbles that cannot be dislodged on the upper side of S-LIPMM and residual melting residues in both of the other methods.

3.1.2. Evaluation of AR Size Criteria

The variations of AR under the three different methods exhibit similar trends among the four laser process parameters, as visually depicted in Figure 8. When aiming for a higher AR, OF-LIPMM demonstrates superior performance, consistently surpassing S-LIPMM in AR regardless of which parameter is modified, as shown in Figure 8a–d. On the other hand, LDW yields the smallest AR. Specifically, the maximum AR achieved in OF-LIPMM can reach approximately 3.0, as depicted in Figure 8a. This is precisely where the conclusions drawn in this study become intriguing. The single-line scan can achieve an AR of 3.0 in OF-LIPMM, an experimental outcome previously unreported in the literature. When compared to the other two methods, OF-LIPMM exhibits an AR enhancement of three to four times, significantly enhancing microchannel performance.

3.1.3. Evaluation of MRR Size Criteria

In the comparative experimental study of obtaining microchannels with a large AR, the influence of laser process parameters on the MRR is shown in Figure 9. OF-LIPMM consistently demonstrates superior performance in terms of MRR, surpassing both S-LIPMM and LDW across all parameters. Furthermore, for the same parameter settings, the MRR of OF-LIPMM is 2–4 times greater than that of S-LIPMM, with a maximum MRR reaching 80 × 10⁴ µm³/s. In contrast, S-LIPMM and LDW exhibit relatively small variations within a limited range compared to OF-LIPMM, and different laser parameters have a significant impact on the MRR of OF-LIPMM. It is worth noting that, as shown in Figure 9a, a significant discontinuity occurs in OF-LIPMM between 18 µJ and 20 µJ. At energy levels above 18 µJ, the internal thermal energy within the plasma deposits on the workpiece surface through conduction during the interaction process. This causes an instantaneous rise in the local temperature around the plasma–material interaction point, leading to melting and vaporization of the workpiece material. At this stage, smaller ablation fragments dominate the energy transfer. However, beyond 20 µJ, a substantial amount of ablation fragments and cavitation bubbles are generated. Due to the relatively low water velocity of 0.5 m/s in this experiment, the timely removal of surface ablation debris is limited, resulting in a reduced MRR.

3.2. Results of Microchannel Parameters with Low HAZ

In the fabrication process of low HAZ microchannels, the lower overlap rate effectively matches the water velocity, significantly reducing the width of the HAZ and enhancing machining efficiency. The results indicate that there are no protrusions at the entrance of the machining zone, leading to improved geometric accuracy. The representative microchannels obtained from the experiments are shown in Figure 10. The three-dimensional representation of the microchannel fabricated with a single-pulse energy of 13 µJ, laser frequency of 20 kHz, scanning speed of 1 mm/s, and 15 scanning times is shown in Figure 10a, while the corresponding two-dimensional diagram is presented in Figure 10b. In Figure 10a, the HAZ is remarkably uniform, and the bottom section appears continuous, exemplifying a typical microchannel with low thermal impact. From Figure 10b, it can be observed that the microchannel has a width of 100 µm, a depth of 100 µm, and an AR of approximately 1.

When a femtosecond pulse laser interacts with a transparent water layer, nonlinear absorption occurs only within a very small region around the focal point of the incident light, greatly enhancing energy utilization efficiency. In an extremely short time, the temperature in the absorption region rises rapidly to a value far above the critical point, transforming the region into a plasma state. The plasma promptly expels the energy injected by the laser, and the absorption region subsequently cools down. When using the OF-LIPMM method to process the parameters outlined in Table 2, due to the uniform flow of the water layer on the workpiece surface, both the expelled heat and the machining ablation residue can be carried away. This characteristic enables the significant reduction in the thermal damage range during femtosecond laser micromachining, leading to a noticeable decrease in the HAZ during the machining process and, consequently, obtaining microchannels with distinct performances, as depicted in Figure 10.

Figure 11a–c depict microchannel patterns obtained under the following conditions: energy of 9 µJ, laser frequency of 20 kHz, scanning speed of 1 mm/s, and 15 scanning times. Figure 11d–f, on the other hand, represent microchannel patterns obtained with a single-pulse energy of 19 µJ, while the remaining processing parameters remain consistent with Figure 11a–c. Upon comparing Figure 11a–c with Figure 11d–f, it is evident that the latter (Figure 11d–f) exhibits an improvement in single-pulse energy and an increase in the HAZ, accompanied by an increased depth of microchannels. By comparing Figure 11a with Figure 11b,d with Figure 11e, it is noticeable that the HAZ In OF-LIPMM is smaller than that in S-LIPMM, indicating that microchannels with a smaller HAZ can be achieved under the parameters specified in Table 2. When comparing Figure 11a with Figure 11c,d with Figure 11f, it is observed that the depth and morphology of the microchannels produced by LDW are inferior to those of OF-LIPMM. However, the HAZ is smaller. It is evident from the three-dimensional images that there are prominent molten residues on one side that cannot be removed through cleaning. Comparing Figure 11 with Figure 7, it is apparent that adopting a lower overlap ratio effectively reduces the generation of ablation debris and provides sufficient time for the removal of bubbles and debris particles. This helps decrease the HAZ generated during the processing and stabilize the turbulent region before the next plasma discharge begins.

Figure 12a–c depict microchannel patterns obtained under the following conditions: a single-pulse energy of 13 µJ, laser frequency of 20 kHz, scanning speed of 1 mm/s, and 12 scanning times. Figure 12d–f, on the other hand, represent microchannel patterns obtained with a speed of 4 mm/s, while the remaining processing parameters remain consistent with Figure 12a–c. Firstly, comparing Figure 12a with Figure 12d, we observe that increasing the scanning speed can reduce the width of the HAZ in OF-LIPMM. Furthermore, by comparing Figure 12a,d with Figure 12b,c, we can see that the HAZ in microchannels produced by OF-LIPMM is smaller than that of S-LIPMM. However, when considering the LDW microchannels shown in Figure 12c,f, the results are similar to the energy variation observed in Figure 11. Although OF-LIPMM exhibits a larger HAZ, the bottom morphology of the microchannels is noticeably superior to that of S-LIPMM and LDW. Considering the higher scanning speed shown in Figure 11c, although it can effectively reduce the HAZ and improve the morphology, it also affects the AR.

Figure 13a–c present microchannel patterns obtained under the following conditions: a single-pulse energy of 13 µJ, laser frequency of 25 kHz, scanning speed of 1 mm/s, and 15 scanning times. Figure 13d–f, on the other hand, display microchannel patterns obtained with a laser frequency of 35 kHz, while the remaining processing parameters remain consistent with Figure 13a–c. Similar to the results observed in Figure 9 after varying the laser frequency, a comparison between Figure 13a,d reveals a significant impact of the laser frequency on the HAZ, wherein an increase in the laser frequency results in an enlargement of the HAZ. Comparing Figure 13a,b, as well as Figure 13d,e, we observe that the HAZ in microchannels produced by OF-LIPMM is smaller than that of S-LIPMM, but larger than that of LDW. However, the microchannels created by OF-LIPMM exhibit a more uniform bottom morphology and greater depth than those obtained under the other two conditions. Additionally, from Figure 13b,c,e,f, it is evident that there are non-uniform regions present at the bottom of these microchannels.

3.2.1. Evaluation of HAZ Size Criteria

In the comparative experiment aimed at achieving microchannels with low HAZs, the parameters listed in Table 2 were utilized. The influence of varying four process parameters on the microchannels under three different conditions is illustrated in Figure 14. By selecting different machining parameters, it is evident from the graph that the HAZ of the microchannels produced by OF-LIPMM is significantly smaller than those produced by S-LIPMM, resulting in a reduction in width of approximately 10–20 µm. It is noteworthy that there is not much reduction in the HAZ when the single-pulse energy and scanning speed are changed. This can be attributed to the overall reduction in the process parameters, which leads to a decrease in the generation of shockwaves induced by plasma machining and the formation of ablative debris and cavitation bubbles. These two factors do not play a dominant role in reducing the HAZ. However, due to the “cold machining” characteristic of femtosecond direct writing, which involves a short interaction time with the workpiece surface and the poor thermal conductivity of air, some ablative debris only adheres to the surface of the workpiece, without damaging the underlying material. This remaining debris can be washed away during subsequent ultrasonic cleaning, resulting in a smaller observed HAZ compared to the other two conditions.

3.2.2. Evaluation of AR Size Criteria

Figure 15 illustrates the influence of different process parameter variations on the AR under three different conditions. Comparing graphs Figure 15a–d, it is evident that the AR of microchannels achieved by OF-LIPMM is significantly higher than the other two conditions. Particularly in Figure 15a,b, the AR of OF-LIPMM shows little variation and remains stable at around one. S-LIPMM oscillates around 0.5, while LDW ranges from 0.1 to 0.4. The variations in the four parameters are similar among the three conditions. In Figure 15c, as the scanning speed increases, there is an initial increase followed by a decrease in the AR for all three conditions. This is because although the scanning speed increase leads to a decrease in overlap rate and reduces the width and depth of the microchannel, the AR may not necessarily decrease. Instead, it exhibits an initial increasing trend. In Figure 15d, as the frequency increases, the AR decreases for OF-LIPMM. This is due to the fact that while the width increases, the depth does not increase significantly, resulting in an initial increase followed by a decrease in the AR for all three conditions.

3.2.3. Evaluation of MRR Size Criteria

Figure 16 illustrates the influence of the four different process parameters on the MRR for three different processing conditions: OF-LIPMM, S-LIPMM, and LDW. Among them, OF-LIPMM exhibits the most significant variation in MRR in response to changes in the processing parameters compared to the other two conditions, surpassing S-LIPMM by more than double. The maximum achieved MRR reaches up to 30 × 10⁴ µm³/s. LDW, on the other hand, maintains a relatively stable MRR at around 2.5 × 10⁴ µm³/s, with minimal fluctuations resulting from variations in the four parameters. The trends observed in OF-LIPMM and S-LIPMM are quite similar. It is worth noting that as the scanning times increase, the MRR decreases in OF-LIPMM while remaining relatively stable in S-LIPMM. This can be attributed to the flow of the water layer, where with more scanning times, energy attenuation increases and the depth of the microchannels does not continue to increase. As time progresses, the MRR tends to decrease. In S-LIPMM, an increase in the number of scanning times effectively improves the phenomenon of laser focus on the workpiece surface, which may not have been achieved during the initial few scanning times due to the accumulation of ablation debris. Consequently, the depth continues to increase, stabilizing the MRR.

3.3. A Comprehensive Discussion of Two Microchannels

Apart from melting and vaporization, high laser fluence can also induce phenomena such as plasma generation, phase explosion, and bubble motion. These phenomena were observed to be evident in the femtosecond laser processing conducted in this study. Figure 17 illustrates the underlying principles of laser ablation on the material surface under three different conditions. Figure 17a depicts laser dry ablation on the material surface, while Figure 17b,c represent plasma-assisted microfabrication in static water and microfabrication in flowing water during OF-LIPMM, respectively. The impact of the liquid layer on surface morphology depends on the dynamics of laser ablation at the solid–liquid interface.

Due to the introduction of the water layer, a single-pulse energy of 13 µJ and a spot diameter of 30 µm were employed, with the energy density threshold reaching a breakdown limit, as discussed in [9]. As a result, the medium experienced a rise in temperature and vaporization, leading to the formation of plasma due to ionization of the vaporized particles induced by the laser. The thermal energy within the plasma exceeded the vaporization temperature of the material. When the plasma contacted the workpiece, material removal primarily occurred through melting and vaporization. Once the plasma was focused on the workpiece surface, it interacted with the workpiece material. During this interaction, the thermal energy within the plasma was deposited on the workpiece surface through conduction. This resulted in an instantaneous temperature increase in the localized region of the plasma–material interaction and its vicinity, leading to melting and vaporization. Additionally, a portion of the thermal energy was lost to the surrounding medium through convection and radiation.

Simultaneously, the mechanical energy generated by the plasma caused shock waves and cavitation, displacing the medium near the material and exerting pressure on the workpiece, preventing the movement of molten material [30]. When the pulse laser ended, the plasma collapsed, and the medium rushed back to fill the gap, causing a sudden decrease in pressure, resulting in the rapid expulsion of molten material from the workpiece surface. The processing debris was also carried away by the medium, leading to material removal.

In S-LIPMM, as depicted in Figure 17b, a static water layer was used, causing the generated bubbles and processing debris to float above the workpiece, which hindered the transmission of the laser underwater. Moreover, significant oscillations occurred in the static water layer during the processing, affecting the processing quality and efficiency. However, in OF-LIPMM, as depicted in Figure 17c, the thickness and stability of the water layer were ensured, which guaranteed stable laser intensity and plasma. Additionally, the directed flow of water effectively carried away the floating debris and bubbles above the processing area, reducing blockage and interference during the propagation of the laser underwater and maintaining stable laser intensity. The presence of the flowing water layer also facilitated the timely removal of heat, refreshing the surface water layer and reducing the width of the HAZ.

Two sets of parameters from Table 1 and Table 2 were employed in the experiments for microchannel fabrication. The main difference between these two sets of parameters was the reduction in the three parameters affecting the overlap rate in Table 2 compared to Table 1. With the water velocity kept constant in the experiments, Table 2 had lower spot overlap rate, smaller energy density, and most of the heat was carried away, resulting in a smaller HAZ. Simultaneously, the lower energy density reduced the energy absorption of the generated plasma, resulting in a lower AR. In contrast, Table 1 had relatively larger values for these parameters compared to Table 2, and the higher spot overlap rate led to a larger AR. However, the flowing water could not carry away as much heat, resulting in a larger HAZ.

3.4. Parametric Processing Grid Verification Experiment

MEMS, microfluidic chips, and other miniature components have garnered significant attention due to their advantages such as low power consumption, light weight, high precision, and fast response. These components include a plethora of microchannels with high ARs and low HAZs, which have been extensively utilized in fields such as aerospace, automotive instrumentation, microrobotics, and others. Figure 18 presents the validation experiments conducted to verify the applicability of the parameters outlined in Table 1 for large-scale grid fabrication. The grids in Figure 18a,b were fabricated under the following conditions: single-pulse energy of 16 µJ, 100 kHz laser frequency, 0.1 mm/s scanning speed, and 10 scanning times. Additionally, the grids in Figure 18c,d were fabricated under the following conditions: single-pulse energy of 13 µJ, 20 kHz laser frequency, 1 mm/s scanning speed, and 15 scanning times. Figure 18a–d provide three-dimensional representations of the grids, with Figure 18b being an enlarged view of Figure 18a, and Figure 18d being an enlarged view of Figure 18c. Figure 18e,f display two-dimensional contour plots of a cross-section of the grid. From Figure 18, it can be observed that the application of the OF-LIPMM method for grid fabrication on stainless steel surfaces yields smooth and well-defined surface morphology. In the sections of the grid with high ARs, such as those depicted in Figure 18a,b,f, the channel width measures approximately 120 µm, with a depth of 360 µm, resulting in an AR of three, which aligns with the results obtained when fabricating microchannels alone. However, Figure 18a,b reveal numerous uneven ablation marks surrounding the elevated portions of the grid, indicating a larger HAZ and substantial material removal in the vicinity. In the low HAZ grids illustrated in Figure 18c,d, the ablation marks surrounding the elevated sections are noticeably reduced compared to those in Figure 18a,b. The elevated surfaces exhibit a smoother profile, with less material removal in the surrounding areas. The channel width in the low AR grid is smaller than that in the high AR grid, as demonstrated in Figure 18f, where the microchannel width measures 100 µm, with a depth of 100 µm, resulting in an AR of one, which is consistent with the results obtained in the single-factor experiment for microchannels.

4. Conclusions

In this study, we have conducted a comprehensive investigation of the fabrication of surface microstructures using OF-LIPMM. By systematically manipulating the experimental parameters, we have successfully achieved two microchannels: one with a high AR, and the other one with a low HAZ. Furthermore, a single-factor experiment was conducted to investigate the influence of laser scanning speed, scanning times, frequency, and single-pulse energy on the performance indicators of microchannels. Eventually, we performed grid fabrication experiments to validate the effectiveness of our approach. The key findings of this study are outlined as follows:

• Microchannels with high ARs of approximately 3:1 and a large HAZ can be obtained using the parameters in Table 1. The grid validation experiment confirms that large-scale microchannel grids with high ARs can be achieved using these parameters, and the microchannels within the grids closely resemble those obtained in the single-factor experiments. By utilizing the parameters in Table 2, microchannels with a low HAZ and low ARs of approximately 1:1 can be produced. The grid validation experiment confirms that large-scale microchannel grids with low HAZs and excellent surface morphology can be obtained using these parameters, which is consistent with the results of the single-factor experiments.
• In comparison to S-LIPMM, OF-LIPMM generates a uniform and stable flowing-water layer in the machining area. This flowing-water layer enables large-scale processing with improved stability and efficiency. By comparing with LDW, the superiority of OF-LIPMM under the same parameters is confirmed. As a result, it was observed that OF-LIPMM significantly improves the overall machining quality compared to other methods such as S-LIPMM and LDW. Specifically, AR and MRR values achieved with OF-LIPMM were found to be obviously higher than those achieved with S-LIPMM and LDW.
• The machining depth of microchannels increases with the increase in single-pulse energy, frequency, and scanning times, and increases with the decrease in scanning speed. The trends of the four process parameters (single-pulse energy, scanning speed, scanning times, and frequency) are similar for the LDW, S-LIPMM, and OF-LIPMM methods. However, this paper has not extensively delved into microscale mechanisms, which could potentially impose limitations on future applications.