Shallow Angle Drilling of Inconel 718 Using a Helical Laser Drilling Technique

Laser drilling to generate holes is becoming a leading technique in many manufacturing processes. In particular, generating cooling holes in high-temperature superalloy components of an aero-engine is the largest application of laser drilling [1]. Reducing a drilling time for relevant aerospace applications of the laser drilling is a major task when considering that one jet engine generally needs over 1 × 10⁶ cooling holes [2].

There are conventional ways to produce holes in countless different materials. Those are punching, drilling by rotating twist drill bits, and drilling by electrochemical machining (ECM) and electrical-discharge machining (EDM). However, the punching or twist drilling turns out to be inefficient for superalloy material due to work-hardening characteristics of these alloys. The ECM and EDM are less dependent on the material properties of the material, but they are not free from the high cost of tools and long lead times [3]. For the last three decades, laser drilling which uses highly intensified radiation energy as a heat source has been considered as an attractive method for drilling of various materials. In this process, a target irradiated by a high intensity laser beam is quickly heated up to a melting or boiling temperature. Materials are then removed by vaporization and liquid expulsion due to high vapor pressure (or recoil pressure) and a hole is eventually produced [4,5].

To date, a number of studies for the laser drilling process have been reported. The relationship between process parameters and geometrical characteristics of the hole such as hole size, circularity of hole, and hole taper was tested. Ghoreishi et al. [6] investigated the effect of the focal position on hole geometry. According to their study, the effect of focal position on hole taper was most significant compared to other variables (peak power, pulse width, pulse frequency, and assist gas pressure). Positive focal position above workpiece surface produced holes with less taper (due to the enlargement of exit hole size) because a beam diameter for the positive focal position was larger at hole exit than that for a zero (on workpiece surface) or negative (below workpiece surface) focal position. Bandyopadhyay et al. [7] also investigated the effect of process parameters, mainly pulse energy, on hole taper. Higher pulse energy was found to lead to less taper since high pulse energy generally produced larger exit hole diameters.

Metallurgical characteristics such as recast, microcracks, delamination, and spatter deposition of laser-drilled holes were also investigated. Chien and Hou [8] studied the recast and microcracking in laser trepan drilling. The recast layer was generally thicker at entrance side than at any other point along the depth of a hole. Assist gas pressure, peak power, and focal position were the parameters that exerted the most significant influence on recast layer thickness. The increase in peak power and focal position on the working surface leads to a reduction in the recast layer thickness. A higher assist gas pressure generally tended to minimize the recast layer thickness. Li et al. [9] presented a novel high peak power pulse burst laser for drilling. They used a Q-switch pulse laser operated at a 400 kW peak power and a 5–50 kHz repetition rate burst for a percussion drilling experiment of Cu and Ni-alloys. A hole diameter of 200 $μ$m was obtained. A recast layer was about 10 $μ$m, and there was no microcracking observed with scanning electron microscope (SEM). Corcoran et al. [10] studied about laser drilling of multilayer aerospace material systems. The multilayer system used in the experiment was a René 80 substrate coated with a thermal barrier coating (TBC). High pulse energy reduced the level of microcracking in laser-drilled holes. Low pulse energy, however, reduced the level of adherent remelt material remaining on the edge of holes. Also, shorter pulse widths reduced the severity of microcracking and delamination between the TBC and bond coat (BC) significantly. Girardot et al. [11] investigated delamination in the laser drilling of multilayer material too. The laser drilling experiment was carried out by varying the beam angles of incidence. Delamination appeared at inclined drilling in the TBC at the BC/TBC interface at the leading edge. It was revealed that the pressure due to both the molten metal and vapor applying on the inclined leading ceramic wall of the hole was mainly responsible for delamination. Low et al. [12] studied the spatter deposition in laser drilling. It was found that short pulse width, lower peak power, and higher pulse frequency generally resulted in a reduction of spatter deposition. Low et al. [13] also reported that the spatter bonding strength was found to decrease with the use of oxygen and oxygen-containing assist gases. Yan et al. [14] carried out a numerical and experimental study on the CO₂ laser percussion drilling of thick section (4.4 mm thickness) alumina. It was found that the size and temperature of melt front significantly affected spatter deposition. The characteristics of melt front were mostly influenced by laser peak power, pulse repetition rate, and pulse duty cycle together.

Some researchers have employed a design-of-experiment (DOE) during laser drilling experiments to systematically investigate the effect of process parameters on laser drilling performance. Corcoran et al. [10] used the Taguchi method containing an L18 (2² × 3³) orthogonal array design. The effects of percussion laser drilling on material interface, bond strength, remelt layer, and microcracking were investigated, and optimal process parameters were determined. Tam et al. [15] also reported the use of the Taguchi method for optimization of the laser deep hole (25 mm) drilling of Inconel 718. The optimal process condition for a minimum drilling time of 31.51 s was a pulse energy of 30 J, a treble pulse shape with pulse duration of 1.8 ms, a focal position of 0.0 mm into the material, and an oxygen pressure of 0.35 MPa.

The researches above have made considerable contributions to the increase in the knowledge of the laser drilling process. However, most studies have focused on either percussion drilling or trepanning drilling and holes are primarily drilled at a right angle to specimen surface. In this research, a helical drilling technique is used to effectively produce holes with a diameter which is 1 order of magnitude larger than that of the focused beam spot. Holes with a shallow angle, 30 deg, to workpiece surface are drilled in Ni superalloy material to mimic cooling holes obliquely drilled in the turbine blades of the aero-engine. Inconel 718, which is one of the widely employed Ni superalloys in the aero-engine industry, is used as a target material. This study shows the systematic investigation for the effect of process parameters on hole quality such as drilling time, hole size, hole taper, and recast layer thickness through the design of experiment. From the part of the DOE results, a regression model to predict a correlation between experimental factors and hole quality was also developed. In the second phase of the study, an extra effort to improve drilling performance within the acceptable drilling time for industry application is made as well.

The laser used for drilling experiment was a high-average power diode-pumped solid-state Nd:YAG laser (DPSSL, model DP-11), manufactured by Northrop Grumman, (formerly known as TRW), Space & Electronics, Redondo Beach, CA. In this study, the laser was operated at 1064 nm and two different duty cycles: 7.5% (500 Hz pulse repetition rate and 150 $μ$ s pulse duration) and 10% (500 Hz pulse repetition rate and 200 $μ$ s pulse duration). In order to enhance drilling performance, the laser is modulated so that one pulse can have 17 (for 7.5% duty cycle) and 20 (for 10% duty cycle) small spikes with a high peak power. The full width half maximum of each spike was ∼200 ns. Pulse formats used in this study are shown in Fig. 1. A laser beam was focused through a lens with a 100 mm focal length, giving a spot size of approximately 30 $μ$m diameter. The laser was incorporated into an ANORD AD four-axis computerized numerical control (CNC) machine for precise control of the relative position between the laser beam and workpiece. The real images of laser, beam delivery optics, and CNC stage are shown in Fig. 2.

An assist gas jet was introduced co-axially to the laser beam through a conical nozzle with the orifice of ∼3.8 mm diameter. Ar and O₂ were employed as the assist gases in this study. Gas pressure shown in this study was the one measured at the outlet port of the gas cylinder and not in front of the laser nozzle. In this study, holes angled at 30 deg to a workpiece surface were drilled to mimic cooling holes obliquely drilled in turbine blades. In order to drill a hole angled at 30 deg to a surface, Inconel 718 plates of 3 mm thick, 5 mm wide, and 25 mm long were positioned on the wedge slanted by 60 deg with respect to a horizontal plane; therefore, the actual depth to be laser-drilled was 6 mm due to the relative angle (30 deg) between the laser beam and workpiece surface. A schematic diagram to describe processing area is shown in Fig. 3.

In the first stage of the study, a set of experiments designed using a Taguchi L16 orthogonal array were conducted to investigate the relationship between the process parameters and hole quality. In these experiments, a 7.5% duty cycle laser setting and a helical drilling method were used. The helical drilling method is applied to efficiently generate holes with a diameter which is around ten times larger than that of the focused beam spot. The helical drilling used is similar to a laser trepanning, but a key difference is that the laser beam path for helical drilling technique is composed of circles with different sizes as illustrated in Fig. 4. The laser beam first moves along circle 6, the largest one (radius 230 $μ$m), and then it follows circle 5 (radius 200 $μ$m), circle 4 (radius 170 $μ$m), circle 3 (radius 140 $μ$m), circle 2 (radius 110 $μ$m), and finally circle 1 (radius 80 $μ$m). Once the laser beam completes one sequence of the beam path (from circles 6 to 1), the laser beam path is repeated until it meets the experimental condition given. The number of drilling passes for helical drilling is the number of passes multiplied by the number of sequences. For example, the number of passes: 30 = 6 (passes per sequence) × 5 sequences. The focal position condition (shown in Table 1) defined in this study is based on the center of the hole that is to say the center of the drilling path (i.e., 0 mm focal position means beam focus at the workpiece surface, and this position is set as the center of the hole).

In these experiments, we tested five parameters and four levels as summarized in Table 1. After some initial drilling tests to set process window, the experimental conditions shown in Table 2 were finally selected for DOE study. In the second stage of the study, trials to enhance drilling performance were carried out based on the conditions which gave a hole of high quality from the DOE study. An O₂ assist gas, moving focal position during drilling, and high laser duty cycle were additionally tired to improve the drilling performance more. The drilling performance was evaluated in terms of exit hole size and drilling time.

After laser drilling experiments were completed, the geometrical and metallurgical characteristics of the laser-drilled holes were examined to investigate the effects of process parameters on hole quality and evaluate drilling performance. A drilling time, exit hole diameter, hole taper, and recast layer thickness were considered as the hole quality factors in this study.

For each set of parameters, the drilling time was the time taken to complete drilling process and it was measured using a stopwatch. A hole size was measured using an optical microscope (Nikon OPTIPHOT) and software (Adobe Photoshop CS2) for image processing. The schematic representation of the entry and exit hole is shown in Fig. 5.

The procedure measuring an entry and exit hole diameter is as follows. The image files of the entry and exit hole are first created by the software incorporated into the optical microscope. The areas of the hole at entry and exit surfaces are calculated using the image files and software for image processing. Once the areas of the hole at the entry and exit surfaces are obtained, they are projected onto a plane which is perpendicular to the drilling direction.

It is assumed that these projected areas at an entry and exit side are those of the perfect circles. Then, the hole areas projected are used to calculate the diameters of the entry and exit hole using the following equation, projected hole area = (⁠$π$d²)/4, where d is the average diameter of the hole.

where $θ$ is the taper angle, d_entry is the entry hole diameter, d_exit is the exit hole diameter, and L is the length of the hole. After geometrical measurements were completed, laser-drilled holes were sectioned using a precision diamond saw and ground to the center of the hole. During grinding process, the cross section of the hole was carefully checked to confirm that the cross section was at the middle of the entry and exit hole.

Polished samples were chemically treated with a chemical reagent composed of hydrochloric acid (HCl), nitric acid (HNO₃), and glycerol for several minutes in order to measure a recast thickness. The composition of each chemical was 50%, 16.7%, and 33.3% by volume, respectively. The procedure to measure the recast thickness is similar to that of the entry and exit hole diameter measured. An image file of hole cross section is created by the software incorporated into an optical microscope. Then, an area of recast region is calculated by the image-processing technique as before. It is assumed that the cross section of the recast has a rectangular shape. Then, the area calculated is the multiplication of the short side and long side of the rectangular, and the length of the short side of the rectangular (the long side of the rectangular is a hole length) is considered as the average thickness of the recast. Since a recast layer was not generally uniform for whole cross section, the direct measurement of the recast layer thickness at the random locations could give incorrect information; therefore, the whole area of the recast region was considered to quantify the thickness of the recast layer in this study. The elemental composition at the recast layer was analyzed using an energy dispersive spectrometer (EDS) attached on the scanning electron microscope (SEM, Phillips XL 30FEG).

After 16 experiments designed by the Taguchi method were completed, laser-drilled holes were characterized to investigate the effect of the process parameters on hole quality. In the 16 experiments, measured drilling time, entry hole diameter, exit hole diameter, hole taper, and recast thickness were in the range of 5.96–27.80 s, 0.544–0.798 mm, 0–0.243 mm, 1.92–3.35 deg, and 0.112–0.301 mm, respectively.

Once all the responses (real values) of the drilling experiment had been obtained (drilling time, exit hole diameter, hole taper, and recast thickness), they were normalized to values between 1 and 10 with 10 representing the highest quality. For example, the highest hole quality index 10 was assigned to the sample with the shortest drilling time, and the drilling time of other samples was normalized based on this value. Likewise, the highest quality index 10 was assigned to the sample with the largest exit hole size, and the exit hole size of other samples was normalized based on this value; therefore, each hole had an individual hole quality index (IHQI) according to four criteria (drilling time, exit hole size, hole taper, and recast thickness). Each ranges from 1 to 10 (1 = worst quality and 10 = best quality). These IHQIs were then averaged to obtain a total hole quality index (THQI) for each hole.

In this equation, n represents the total number of tests in a trial (n = 1 in this study) and y_i represents the response (IHQI and THQI). Figure 6 shows the effects of process parameters on the IHQIs. The highest SNR of each parameter gives the optimum condition to get the best IHQI. The most significant findings from the experiments are summarized below.

The clear tendency of the SNR is merely seen for drilling speed and number of passes since a drilling time is only affected by speed and number of passes. The results indicate that high speed and small number of passes reduce the drilling time.

It is revealed that high laser power and low speed are the two main factors to increase exit hole size. This is due to the absorption of more laser energy by a material at high power, which results in further removal of the material. Also, a deeper penetration of the laser beam at lower beam scanning speed contributes to an increase in exit hole size. Locating a beam focal position on workpiece surface (i.e., focal position is 0 mm) was also of help to enlarge the exit hole size. Setting the beam focal position on the workpiece surface yields the smallest beam spot size and the maximum energy intensity of the laser beam which is favorable for removal of the material at the initial stage of the process. In addition, a beam diameter becomes wider as the laser beam passes through its focal position due to the divergence of the beam. The largest beam diameter at the bottom surface of the workpiece is obtained when the laser beam is focused on the top surface of the workpiece (i.e., focal position is 0 mm). This expands exit hole at the later phase of the process; thus, using a focal position = 0 mm was useful to get larger exit holes. As the number of drilling passes increased, which was accompanied by the increase of drilling time, exit holes became larger. The results from experiments failed to show that using Ar as an assist gas was advantageous to enlarge exit holes under any of the experimental conditions included in this study. The Ar gas just played a role as a cooling source, which reduced drilling performance by cooling down the workpiece.

In general, an entry hole is greater than an exit hole in laser drilling for two reasons. First, the entry part of a hole tends to be exposed to a laser beam to a greater degree than the exit hole. Second, the erosion of the entry side by molten materials occurs due to recoil pressure. These are responsible for the presence of the hole taper in the laser-drilled holes. Entry holes become smaller in size as a beam focal position gets closer to the surface of the workpiece due to a smaller beam spot size on the surface. (Intuitively, the largest entry hole will be obtained when the beam focal position is −3 mm and the smallest entry hole size is obtained when the beam focal position is 0 mm); thus, as the focal position gets closer to the surface of the workpiece, the hole taper will become smaller because of both a decreased entry hole size and increased exit hole size. As illustrated in Fig. 6(b), exit holes increase in size under low speed and high laser power, which, in turn, results in the decrease in hole taper. In this experiment, the size of exit hole clearly tends to increase with the number of drilling passes. This also contributes to have the smallest taper angle under the largest number of drilling passes. It was observed that using an Ar assist gas lessened the hole taper. Of importance, this was not as a result of the increase in exit hole size (see Fig. 6(b)). Instead, it was primarily due to the decrease in entry hole size. It could be suggested that using the Ar as an assist gas did not contribute to hole expansion and it served more as a source to cool down the workpiece.

From the analysis of the signal-to-noise ratios (see Table 3), it was found that the beam focal position was the most influential parameter to determine the hole taper. The second important parameter was laser power, and the effect of assist gas and speed on the hole taper was turned out to be similar. The hole taper hardly varied with the number of drilling passes in this study.

A recast layer is formed within a hole if molten material is not blown away by explosion or an assist gas. It is also created by redeposited materials between laser pulses.

As shown in Fig. 7, a recast layer thickness varies in depth. For all the samples, the recast layer at an entry side is generally the thickest and it becomes thinner as a hole gets closer to an exit side. The entry side is generally more exposed to a laser beam during processing (more chance to melt hole wall at the entry side) and the redeposition of removed materials starts from it (accumulation of redeposited materials at the entry side is larger than that at other locations). These are why the recast layer decreases with increasing hole depth.

Figure 8 shows the high-magnification images of the recast layer. The recast layer is typically composed of several thin layers as shown in Fig. 8. The thickness of each layer is approximately 5–10 $μ$m. These thin layers are attributed to a by-product of the remelting and resolidification of the hole wall and the redeposition of removed materials during laser drilling.

Recast layers are primarily composed of fine columnar dendrites, which differ from the equiaxed grain structures shown in parent material. As shown in Fig. 8, the growth direction of the columnar dendrites is nearly perpendicular to the hole wall. This indicated that cooling mostly occurred via the hole wall during solidification process. Elemental composition of the recast was almost identical to that of the parent material. This occurred because an inert Ar assist gas suppressed oxidation. Table 4 shows elemental composition data of the recast and parent material.

Figure 6(d) shows that high laser power above 65 W is favorable to reduce the recast layer. In nanosecond laser drilling process, both expulsion of the melt and vaporization are generally two important material removal mechanisms. In this experiment, increased laser power enhanced the contribution of vaporization to the material removal. This reduced the recast by diminishing molten material formation. Low speed was shown to increase the recast in this study. At low speed, a laser beam could go down further in a workpiece. Under this condition, the hole of a relatively large depth-to-width ratio (deep and narrow hole) is likely to be formed. This geometrical characteristic of the hole could restrict the ejection of molten debris from the hole. It would contribute to the redeposit of molten materials on the hole wall during processing. An Ar assist gas also seemed to reduce the recast; however, its influence was not substantial under the range of gas pressure tested in this study.

During the helical motion, the actual size of the beam spot at the laser–material interaction zone keeps changing since the height of the interaction zone varies due to the tilted workpiece. The variation in the beam spot size also leads to continuing change of the beam intensity during the process. This may produce a nonuniform recast layer along the circumference of the hole because material removal mechanism is generally different for high intensity beam (vaporization-dominant material removal and production of the thinner recast) and low intensity beam (liquid expulsion dominant material removal and production of the thicker recast). However, an effort to verify the explanation above was not made in this study.

THQI was calculated using IHQIs, and the optimum set of process parameters to improve hole quality was obtained based on the THQI. The effects of process parameters on the THQI are summarized in Table 5. It was found that laser power had the most significant influence on the THQI, and the focal position had the second greatest influence factor. The effects of the other three variables were small (see rank in Table 5). The optimum set of parameters, that is, the set of parameters with the highest SNR, and its impact on the total hole quality are presented in Table 6.

where A is the laser power, B is the focal position, C is the speed, D is the assist gas pressure, and E is the number of passes. The model above has 93.42% of R². The R² is the regression coefficient for the model, which indicates that the fit of experimental data is satisfactory. R² = 93.42% indicates that the model is able to predict the response with high accuracy. The effect of all the main factors is included in the model. However, parts of their square and interaction terms between the main factors are not considered to enhance the accuracy of the model. Table 7 shows the comparison of actual value (result obtained in this study) and forecasting value of the THQI from the model. The error between two groups is 0.684–8.897%.

In the second stage of the study, trials to improve drilling performance more were carried out based on the conditions which gave a hole of high quality from the previous DOE study. An O₂ assist gas, moving focal position during drilling, and high laser duty cycle were additionally tired to improve the drilling performance. For the industry application of laser drilling, the economical efficiency of the process is important. The reduction of processing time is indispensable for this reason. In this experiment, the drilling time to generate one hole was restricted to less than 5 s. (This is the acceptable drilling time, for industry application, suggested by GE Aviation that financially and technically supported this study.) The drilling performance was evaluated in terms of exit hole size. A better drilling performance was defined as producing a lager exit hole.

The effect of the O₂ assist gas on the drilling performance was tested. Experiments with an Ar assist gas were also conducted for comparison. The laser duty cycle setting (7.5%) used in previous DOE study was adopted again for this experiment. Each experiment was repeated two times. A drilling time for all the experiments was determined to be ∼4.97 s to meet the requirement (less than 5 s drilling time) above by fixing drilling speed (8 mm/s) and number of passes (30). The average and standard deviation error are shown in Fig. 9.

As shown in Fig. 9, employing the Ar assist gas was not useful for enlargement of the exit hole as before. However, using O₂ as an assist gas was a great help for enhancing drilling performance at gas pressure of 50 kPa. The largest exit hole, ∼0.24 mm, was obtained at 70 W, 0 mm focal position, 50 kPa O₂ assist gas pressure, 8 mm/s speed, and 30 passes.

Using O₂ as the assist gas promotes the oxidation at exposed area. This oxidation process delivers more energy to a workpiece by exothermic reaction, which leads to better drilling efficiency. The enhancement of drilling performance by using O₂ assist gas was also reported [16]. In this experiment, 50 kPa O₂ pressure produced the best drilling performance; however, the increase of the O₂ pressure more than 50 kPa resulted in the loss of the drilling efficiency. This may be due to the rapid convective cooling of the workpiece by forced convection from the fast gas flow. The wt.% of O₂ in recast layer was measured by an energy dispersive spectrometer (EDS). The average and standard deviation error of the measurements are shown in Fig. 10. There was a significant increase of wt.% in the recast layer when O₂ was used as an assist gas. Due to the effect of the oxygen in the air, more oxygen was detected in the recast layer in the case of no gas when compared to Ar assist gas. Among three main elements (Ni, Cr, and Fe) of the Inconel 718, the amount of Ni in the recast layer was significantly reduced as compared to the amount of Ni in parent material for O₂ assist gas. However, the relatively small reduction of Fe and the small increase of Cr were observed in the recast layer as shown in the inset of Fig. 10. This implied that the Cr was preferentially oxidized compared to other elements. Once chromium oxide is formed, it tends to prevent the inflow of more oxygen into a material; therefore, the recast layer produced with the oxygen assist gas could be an effective shield against oxidation. The wt.% of O₂ did not vary much under the range of O₂ pressure tested in this study, which, in turn, indicated that recasts were fully oxidized at 50 kPa O₂ pressure already. This indicated that extra O₂ gas beyond that had no contribution to oxidation process. It only caused rapid convective cooling of the workpiece, which lowered the drilling performance.

The enhancement of the drilling performance by high laser duty cycle was also tried. For this experiment, a 7.5% (500 Hz pulse repetition rate and 150 $μ$ s pulse duration) and 10% (500 Hz pulse repetition rate and 200 $μ$ s pulse duration) duty cycle settings were used by adjusting the duration of the pulse. The drilling performance was compared for four different experimental conditions as shown in Fig. 11. In particular, the effect of the moving focal position on the drilling performance was also considered. In this case, a focal position was moved up to 3 mm into a workpiece during drilling process. The main goal of this method is to maintain high laser beam intensity at the interaction region of the laser–workpiece. Each experiment was repeated two times. An average value and standard deviation error are shown in Fig. 11. Around 4.97 s drilling time was used again for all the experiments as before (8 mm/s drilling speed and 30 drilling passes).

In general, the drilling performance for all the cases was enhanced using a 10% duty cycle laser setting, as shown in Fig. 11. In the case of a 7.5% duty cycle laser setting, the drilling performance was enhanced using the moving focal position method as shown in Fig. 11(a). As explained above, this is attributable to keeping high laser beam intensity at the interaction region of the laser–workpiece. Using both O₂ gas (50 kPa) and moving the focal position gave the best performance throughout the four different experiments. Moving a focal position was not useful to improve drilling performance in the case of a 10% duty cycle laser setting. Under the 10% duty cycle, laser power itself may be high enough. As a result, a rate of increase of the drilling depth would be higher than a rate of variation of the focal position. This may relatively decrease the beam intensity more at the interaction region of the laser–workpiece compared to 7.5% duty cycle. Figure 12 shows the cross section of the hole produced at a 10% duty cycle laser setting (90 W laser power, 0 mm focal position, 50 kPa O₂ assist gas pressure, 8 mm/s speed, and 30 passes). High-magnification SEM images in Fig. 12 show that the recast is composed of two types of layers. The inner part is a nonoxide layer which has the similar microstructure and elemental composition to those shown in Fig. 8 and Table 4. The outer part of the recast region is an oxide layer with an increased Cr content relative to other main elements (Ni and Fe). Elemental composition data of the oxide layer are shown in Table 8.

Cooling holes angled at 30 deg to workpiece surface were successfully drilled in Inconel 718 samples under an acceptable drilling time (∼5 s) for industry application. A helical drilling technique was used to effectively generate holes with a diameter which was around ten times larger than that of the focused beam spot (30 $μ$m).

The DOE using the Taguchi method was employed to investigate the effect of the process parameters on the hole quality. Higher laser power, lower speed, and closer focal position to the workpiece surface contributed to the further removal of material by the absorption of more laser energy and larger beam intensity. This resulted in a larger exit hole diameter and less hole taper. Using Ar as an assist gas was not useful to enlarge an exit hole size. A recast layer was generally thicker at an entry side of the hole than at any other part along the depth of the hole because of more chance to melt an entrance hole wall by a laser beam and larger accumulation of redeposits at an entry area. The increase in laser power reduced the thickness of the recast layer due to the contribution of vaporization to material removal at the larger power. A regression model which can estimate a correlation between experimental factors and THQI was developed. The error between actual values and forecasting values from the model was in the range of 0.684–8.897%. In the second stage of the study, trials to improve drilling performance were carried out based on the conditions which gave a hole of high quality from the DOE study. An O₂ assist gas, moving focal position during drilling, and high laser duty cycle were additionally tired to improve the drilling performance more. Using the O₂ assist gas of 50 kPa significantly enhanced the drilling performance owing to the delivery of more energy to a workpiece by exothermic reaction. However, the further increase of the O₂ gas caused rapid convective cooling of the workpiece, which lowered the drilling performance. The drilling performance was greatly improved as well using high laser duty cycle by providing more laser energy to the workpiece. The moving focal position was only beneficial to the drilling performance when the focusing of the laser beam was moderately kept on the interaction region of the laser–workpiece.

This work was supported by the National Science Foundation Industry-University Co-operative Research Center and GE Aviation under the NSF Subaward No. 0438917. The authors would like to thank Dr. Todd Rockstroh and Mr. Doug Scheidt for valuable input at discussion.