• Journal of the Semiconductor & Display Technology
• /
• v.5 no.3 s.16
• /
• pp.5-10
• /
• 2006

This paper investigates cutting qualities after laser dicing and predicts the problems that can be generated by laser dicing. And through 3 point bending test, die strength is measured and the die strength after laser dicing is compared with the die strength after mechanical sawing. Laser dicing is chiefly considered as an alternative to overcome the defects of mechanical sawing such as chipping on the surface and crack on the back side. Laser micromachining is based on the thermal ablation and evaporation mechanism. As a result of laser dicing experiments, debris on the surface of wafer is observed. To eliminate the debris and protect the surface, an experiment is done using a water soluble coating material and ultrasonic. The consequence is that most of debris is removed. But there are some residues around the cutting line. Unlike mechanical sawing, chipping on the surface and crack on the back side is not observed. The cross section of cutting line by laser dicing is rough as compared with that by mechanical sawing. But micro crack can not be seen. Micro crack reduces die strength. To measure this, 3 point bending test is done. The die strength after laser dicing decreases to a half of the die strength after mechanical sawing. This means that die cracking during package assembly can occur.

• Proceedings of the KIEE Conference
• /
• 2006.04a
• /
• pp.327-329
• /
• 2006

In this paper, the cutting qualities by laser dicing and fracture strength of a silicon die is investigated. Laser micromachining is the non-contact process using thermal ablation and evaporation mechanisms. By these mechanisms, debris is generated and stick on the surface of wafer, which is the problem to apply laser dicing to semiconductor manufacture process. Unlike mechanical sawing using diamond blade, chipping on the surface and crack on the back side of wafer isn't made by laser dicing. Die strength by laser dicing is measured via the three-point bend test and is compared with the die strength by mechanical sawing. As a results, die strength by laser dicing shows a decrease of 50% in compared with die strength by mechanical sawing.

• Proceedings of the Korean Society Of Semiconductor Equipment Technology
• /
• 2006.05a
• /
• pp.251-254
• /
• 2006

In this paper, cutting qualifies and fracture strength of silicon dies by laser dicing are investigated. Laser micromachining is the non-contact process using thermal ablation and evaporation mechanisms. By these mechanisms, debris is generated and stick on the surface of wafer, which is the problem to apply laser dicing to semiconductor manufacture process. Unlike mechanical sawing using diamond blade, chipping on the surface and crack on the back side of wafer isn't made by laser dicing. Die strength by laser dicing is measured via the three-point bending test and is compared with the die strength by mechanical sawing. As a results, die strength by the laser dicing shows a decrease of 50% in compared with die strength by the mechanical sawing.

• Journal of Institute of Control, Robotics and Systems
• /
• v.8 no.5
• /
• pp.419-424
• /
• 2002

After the patterning and probe process of wafer have been achieved, the dicing processing is necessary to separate chips from a wafer. The dicing process cuts a semiconductor wafer to lengthwise and crosswise directions to make many chips. The existing general dicing method is the mechanical cutting using a narrow circular rotating blade impregnated diamond particles or laser cutting. Inferior goods can be made by the mechanical or laser cutting unless several parameters such as blade, wafer, cutting water and cutting conditions are properly set. Moreover, we can not apply these general dicing method to that of GaN wafer, because the GaN wafer is harder than general semiconductor wafers such as GaAs, GaAsP, AIGaAs and so forth. In order to overcome these problems, this paper describes a new wafer dicing method using fixed diamond scriber and precision servo mechanism.

• Journal of the Microelectronics and Packaging Society
• /
• v.24 no.1
• /
• pp.91-101
• /
• 2017

In this study, we conducted the preliminary research for high speed laser cutting of LED module. In particular, the feasibility of ultra-high speed laser cutting of 100 mm/s which exceeds the cutting speed of conventional dicing saw was examined. For this, copper/ceramic and silicone/ceramic hybrid substrates, which are the components of the LED module, were fabricated, and the surface morphology, surface roughness and flexural strength of the laser-cut samples were investigate and compared with the dicing-cut samples. To investigate optimal laser cutting conditions for hybrid substrates, the effects of various laser cutting conditions on cutting surface characteristics were studied using single ceramic and copper substrate. Optimal laser cutting conditions of the hybrid substrates were the use of Ar assist gas, high laser power and high assist gas pressure. Comparing the cutting surface of the hybrid substrates, the surface characteristics of the laser-cut samples are slightly inferior to those of the dicing-cut samples. The average surface roughness of the laser-cut samples was about $9{\mu}m$, and that of the dicing-cut samples was about $4{\mu}m$. However, considering very low cutting speed (3 mm/s) of the dicing saw, the surface morphology of the laser-cut sample was relatively uniform, and the surface roughness was not much different from that of the dicing-cut sample. The flexural strength of the laser-cut samples was equivalent to or slightly inferior to the flexural strength of dicing-cut samples. However, if the laser processing conditions are sufficiently optimized, the ultra-high speed laser cutting of the LED module will be possible.

• Journal of the Microelectronics and Packaging Society
• /
• v.21 no.4
• /
• pp.63-68
• /
• 2014

For a nanoscale Cu/low-k wafer, inter-layer dielectric (ILD) and metal layers peelings, cracks, chipping, and delamination are the most common dicing defects by traditional diamond blade saw process. Sidewall void in sawing street is one of the key factors to bring about cracks and chipping. The aim of this research is to evaluate laser grooving & mechanical sawing parameters to eliminate sidewall void and avoid top-side chipping as well as peeling. An ultra-fast pico-second (ps) laser is applied to groove/singulate the 28-nanometer node wafer with Cu/low-k dielectric. A series of comprehensive parametric study on the recipes of input laser power, repetition rate, grooving speed, defocus amount and street index has been conducted to improve the quality of dicing process. The effects of the laser kerf geometry, grooving edge quality and defects are evaluated by using scanning electron microscopy (SEM) and focused ion beam (FIB). Experimental results have shown that the laser grooving technique is capable to improve the quality and yield issues on Cu/low-k wafer dicing process.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2011.06a
• /
• pp.1519-1520
• /
• 2011

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2011.06a
• /
• pp.39-40
• /
• 2011

• Proceedings of the KIEE Conference
• /
• 2004.11c
• /
• pp.478-480
• /
• 2004

다이아몬드 톱날을 이용한 얇은 Si 웨이퍼의 기계적인 다이싱은 chipping, crack 등의 문제점을 발생시킨다. 또한 stacked die 나 multi-chip등과 같은 3D-WLP(wafer level package)에서 via를 생성하기 위해 현재 사용되는 화학적 etching은 공정속도가 느리고 제어가 힘들며, 공정이 복잡하다는 문제점을 가지고 있다. 이러한 문제점을 해결하기 위해 현재 연구되고 있는 분야가 레이저를 이용한 웨이퍼 다이싱 및 드릴링이다. 본 논문에서는 UV 레이저를 이용한 얇은 Si 웨이퍼 다이싱 및 드릴링 시스템에 대해 소개하고, 웨이퍼 다이싱 및 드릴링 실험결과를 바탕으로 적절한 레이저 및 공정 매개변수에 대해 설명한다.

• Transactions of the Korean Society of Machine Tool Engineers
• /
• v.17 no.3
• /
• pp.53-58
• /
• 2008

This study was conducted to evaluate the flatness of the porous type of dicing chuck. Two measurement systems for a vacuum chuck with a porous type of ceramic plate were prepared using a digital indicator and a laser interferometer. 6 inch of silicon and glass wafer were also used. Vacuum pressure from 100mmHg to 700mmHg by 100mmHg was increased. From experiments, chucked-wafer flatness was converged to the dicing chuck flatness itself even though the repeatability of contact method using indicator was unstable. Finally, the chuck flatness was estimated below $2{\mu}m$ with peak-to valley value.