Demand for high performance and multi functional electronic devices is growing steadily; however, this places a huge demand on microelectronics and semiconductor manufacturing industries to develop innovative technologies to meet this requirement. Microelectronics packaging is an interconnection technology that plays a key role in meeting this demand.
Lasers help in developing advanced packaging technologies in an economical way. The DPSS q- switched nanosecond pulsed laser is one such device used for scribing of alumina ceramic (Al2O3) substrates employed in the development of microelectronics packages. Following laser scribing, the substrate comprising different electronic components is cleaved separately to make individual microelectronic device packages. The entire procedure is called as “scribe and break” process.
Thanks to rapid development in laser technology, the pulse repetition frequency (PRF) of DPSS nanosecond ultraviolet (UV) lasers and the energy available per pulse have increased significantly. However, on alumina ceramic scribing using a 355nm wavelength laser, only minimal data is available regarding comparative statistics of different methods used for improving process efficiency.
This article describes the impact of short pulse width DPSS q-switched 355nm laser fluence on scribing depth for alumina ceramic material. It also examines three different techniques for effective application of the PRF and high energy available from the laser source to improve material removal rate and maintain good scribe quality.
Approximately 180µm thick blank alumina ceramic substrate was used as the sample material. Laser processing was carried out at room temperature without any process assist gas.
For the experiment, a Spectra-Physics® Pulseo® 355-20 DPSS q-switched laser with high peak power and short pulse width was utilized. Table 1 shows a number of laser specifications.
Table 1. Specifications of the Spectra Physics laser system used in the ceramic scribing experiments
||>20 W at 100 kHz
Figure 1. Experimental optical setup for ceramic scribing.
Figure 1 shows a schematic diagram of the optical setup utilized for the experiments. It includes a scanning galvanometer system with 100mm focal length telecentric f-theta lens equipped with Newport’s linear motorized XYZ positioning stage systems. The laser beam was increased by a factor of 1.5X to produce ~10µm (1/e2) focal spot size that was calculated hypothetically. For the beam-splitting experiments, a diffractive optical element (DOE) was integrated into the optical system, and for the elliptical beam shaping tests, a couple of +/- 300mm focal length cylindrical lenses was placed into the system.
For the experiment, the ceramic plates were fastened to the flat side of the XY motion stages and the mounting surface was adjusted to sustain the focal plane along the ceramic scanning area. In order to produce laser scribes, the laser beam was scanned at different speeds across the ceramic plate material by means of the galvo scanner. A single scan was utilized for all the experiments.
Precision linear motorized Z positioning stages, a part of Newport’s IMS series, was then employed for precise beam focusing. In order to ascertain the location of the focal plane, test scribes were produced at different focal positions along the optical axis by shifting Z stage in approximately 100 micron increments. Next, the focal plane was defined relating to the deepest and narrowest scribe acquired for a specified set of laser parameters.
The material was cleaved and the resulting cross-section was examined under an optical microscope to determine the cut depths of the scribes. In order to cleave the ceramic plate, cleaving scribes were created on the opposite side of the sample.
The samples were then cross-sectioned at multiple locations along the scribes, and the average of various depth measurements was utilized to achieve a single data point. In this method, the ambiguity of the depth measurement is expected to be in the range of 2 to 4 microns.
Results and Discussion
In any laser material processing application, fluence management plays a critical role. At the stage of optimal fluence, the laser pulse energy is effectively utilized in material removal with minimum heat affected zone (HAZ).
Circular Single-Beam Processing
The aim of this analysis is to establish a baseline result for scribing alumina ceramic through a circular focused Gaussian beam. To this end, a range of scribe depths data at different laser fluence and scribe speeds was produced at the 100kHz laser PRF. Figure 2 shows the scribing speed data for producing a scribe depth of 30µm as a virtue of fluence.
Figure 2. Plot of scribe speed as a function of laser flounce at 100kHz PRF for scribing 30µm deep scribe.
The data given in Figure 2 agrees with a logarithmic trend line with R2 close to 1. For the PRF of 100kHz, small increase in fluence value initially results in instant increase in scribe speed, but beyond the fluence value of ~100J/cm2 the extent of increase in the scribe speed with increasing fluence starts to slow down and ultimately flattens out.
The data demonstrates that fluence of 80J/cm2 produces a scribe speed of about 27mm/s, while the fluence of 240J/cm2 creates a scribe speed of about 54mm/s. This shows that three times increase in the fluence results in just two-fold increase in the scribe speed.
The data thus obtained confirms that at higher fluence, all of the laser energy is not completely used for removing material and may be possibly diffused into the material as heat. This is fairly clear from the microscope images displayed in Figure 3 (a) and (b).
Figure 3. Top-view microscope picture showing quality of 30µm deep laser scribe at 100kHz PRF (a) 80J/cm2 (b) >400 J/cm2.
During material processing, the laser can be operated at higher PRF. This is one way to work in the low fluence regime without affecting the throughput. For a standard DPSS Q-switched laser, output power reduces with increase in the PRF. Therefore, the fluence available at higher repetition frequencies is lower, but faster scribe speeds can be obtained in view of higher PRF.
Elliptical Beam Processing
Elliptically-shaped beam is one of the beam shaping methods that was studied in the past to enhance laser process efficiency. In order to obtain an elliptically-shaped beam, the beam spot is stretched along the “major axis” and the beam focal spot size is sustained along the “minor axis”. A scribe is produced by scanning the beam along the major axis.
In an elliptical-shaped beam, the high energy available from the laser is spatially spread to obtain the preferred fluence at the surface. Figure 4 shows a top-view image of 30µm deep laser scribe produced using elliptical beam at a speed of 200 mm/s.
Figure 4. Top-view microscope picture of 30µm deep laser scribe generated using elliptical beam at speed of 200mm/s.
Circular Split-Beam Processing
Circular split-beam processing is a laser beam splitting technique used to enhance the process efficiency. The aim of using this method is to divide the high energy available from the laser into N different beamlets. In an ideal set up, each beamlet will exhibit the optimal fluence value for acquiring excellent process efficiency.
Figure 5. Top-view microscope picture of 30µm deep scribes generated using 1:7 split beam at effective speed of 175mm/s.
Beam splitting is realized by introducing a DOE within the optical beam path before the galvo scanner aperture, as illustrated in Figure 1. Besides higher effective scribing speed, the features obtained through this method have clean quality scribes with no visible HAZ with a scribe width of approximately 15µm (Figure 5).
This article has effectively shown three different fluence management techniques, such as circular single-beam processing, circular split-beam processing and elliptical beam processing, which can be utilized to leverage the high PRF and high energy available from current generation of lasers. By means of higher PRF, split-beam and shaped-beam processing method, the laser energy from the Pulseo 355-20 laser was customized to obtain 30µm deep scribes in alumina ceramic at highest possible scribing speed and also to maintain excellent scribe quality.
On the whole, results demonstrate a significant advantage with respect to processing efficiency and scribe quality that can be achieved by improving the applied laser fluence. Such methods can be effectively used to control the high PRF and high energy offered by existing Q-switched DPSS lasers. High energy and high PRF control can benefit material processing applications.
This information has been sourced, reviewed and adapted from materials provided by Spectra-Physics.
For more information on this source, please visit Spectra-Physics.