Energy-Coupling Mechanisms Revealed through Simultaneous Keyhole Depth and Absorptance Measurements during Laser-Metal Processing

The interaction between high-irradiance light and molten metal is the complex multiphysics phenomenon that underpins industrial processes such as laser-based additive manufacturing, welding, and cutting. One aspect that requires careful attention is the formation and evolution of vapor depressions, or keyholes, within the molten metal. The dynamic behavior of these depressions can dramatically change the number of laser-beam reflections and is therefore intrinsically linked to the instantaneous energy coupled into the system. Despite its importance, there is a severe lack of direct in situ, experimental evidence of this relationship, which creates challenges for those who aim to model or control laser-based manufacturing processes. In this work, we combine two simultaneous state-of-the-art real-time measurement techniques (inline coherent imaging and integrating-sphere radiometry) to confirm and explore the definite positive correlation between the highly dynamic vapor-depression geometry and laser energy absorptance. For irradiances resulting in vapor-depression formation ($\ge 0.49{\mathrm{MW}/\mathrm{cm}}^2$), we observe excellent correlation (0.86) between the instantaneous depth (down to 800 $\mu \mathrm{m}$) and the absorptance (up to 0.92), directly demonstrating their interdependence. In the transition mode, an important regime for additive manufacturing, we observe temporary vapor-depression formation with concomitant changes in absorptance from 0.34 to 0.53. At higher irradiances, we detect stepwise increases in the absorbed laser power with a smoothly increasing keyhole depth, which is a real-time experimental observation of the effect of multiple reflections during laser-metal processing. The value of simultaneous depth and absorption measurements for predictive model validation is presented using ray-tracing simulations, which also confirm the absorption enhancement via incremental increases in the reflection count. This work provides insight into the underlying physics of laser-based metal manufacturing that is useful toward deterministic modeling and real-time process control.

Figure 1

The experimental setup, consisting of the inline coherent imaging system, the ISR system, and the processing laser: SLD, superluminescent diode; PC, polarization control; FC, fiber collimator; M, mirror; ND, neutral density.

Figure 2

The time-resolved depth and absorptance for incident irradiances ranging from 0.35 to 0.52 ${\mathrm{MW}/\mathrm{cm}}^2$ exhibiting behavior consistent with (a) the conduction mode, (b) the keyhole mode, and (c–e) the transition mode. The depth measurements are plotted against the left axis in black. The absorptance measurements are plotted against the right axis in red. The gray fill under the depth data is meant as a visual aid.

Figure 3

(a) The time-resolved depth and absorptance from the keyhole mode (incident irradiance of 0.92 ${\mathrm{MW}/\mathrm{cm}}^2$). (b) Vapor-depression depth measurements during the rapid-growth period for four keyhole-mode trials of differing irradiances. (c) Drilling rates as a function of the average absorbed irradiance. The blue circles are drilling rates calculated during the rapid-growth period. The green squares are drilling rates calculated for the period of steady-state absorptance. The black dashed lines are linear fits to the data according to Eq. (5).

Figure 4

(a) The absorptance and keyhole depth measured during laser irradiance of 0.92 ${\mathrm{MW}/\mathrm{cm}}^2$, compared to ray-tracing simulations. (b) The calculated absorbed laser power distribution on the keyhole wall and laser absorptance for depths of 238 $\mu \mathrm{m}$ (cases 1 and 3) and 476 $\mu \mathrm{m}$ (cases 2 and 4). (c) X-ray imaging shows a cone-shaped keyhole (top) and a SH-shaped keyhole (bottom) at different times during stationary laser illumination. From Ref. [17]. Reprinted with permission from AAAS.

Figure 5

(a) The absorptance and average reflection counts of rays for keyhole depths in and near the absorptance plateau for the cone-shaped simulation. The error bars represent the standard deviation of the average reflection count. (b) Schematic diagrams of ray trajectories for cone-shaped keyholes with depth of 119 $\mu \mathrm{m}$ (blue) and 179 $\mu \mathrm{m}$ (green). (c) The percentage of rays with reflection count $j$ for the six keyhole depths marked in (a).

Figure 6

(a) Spearman’s rank correlation coefficient (${\rho }_s$) for the depth and absorptance as a function of the incident irradiance. The colored points correspond to the like-colored plot below. The range of the error bars associated with these values is smaller than the size of the data markers. (b–d) Plots of the absorptance versus the depth for the conduction mode ($0.23{\mathrm{MW}/\mathrm{cm}}^2$), the transition mode (0.46 ${\mathrm{MW}/\mathrm{cm}}^2$), and the keyhole mode ($0.92{\mathrm{MW}/\mathrm{cm}}^2$), respectively. Note that the scales differ between each plot.

Figure 7

The coupling efficiency (red, right ordinate) and average depth (black, left ordinate) plotted as a function of the incident irradiance. The error bars on the coupling efficiency are representative of a combination of systematic uncertainty and experimental repeatability. The standard error on average depth resulted in error bars smaller than the marker size.

Figure 8

The frequency-resolved depth (bottom) and absorptance (top) from the keyhole regime. FFTs are performed on data from the last 8 ms of laser illumination. The first 1 kHz is excluded from the plots, as this region is dominated by the dc term and by low-frequency contributions associated with initial vapor-depression formation.