Laser die transfer
Prof.dr.ir. J. Meijer
MSc N.S. Karlitskaya
Ir. H.J. Kettelarij
Dr.ir. G.R.B.E. Römer
April 2003 – April 2009
This project is completed. For information contact G.R.B.E. Römer.
chip assembly, laser propulsion, sacrificial layer process
A key feature of the semiconductor manufacturing over the past 30 years has been its ability to produce very small chips, the current state-of-the-art in production being around 0.25 mm. Conventionally assembly is based on manipulation of individual components by the pick-and-place robots. This technology is unsuitable for many microsystems due the difficulty in handling small (less than 1 millimeter) and thin (less than 100 micrometers) components and in scaling up these techniques for mass production. Thus, micro-assembly of miniature electronics demands processes which shall be able to handle accurately components with dimensions of several tens of micrometers with high speed of placing onto a broad range of substrates.
In our project the study of the new technique based on laser-induced releasing of the micro component from its carrier, transferring it towards the interconnecting substrate and landing is carried out. This process has an advantage of very high transfer speeds (up to 100 components per second comparing with maximum 8 components per second with the conventional equipment). Furthermore, laser assisted transfer being a contactless process, opens the route towards transferring ultra-thin and small components.
The objective is to develop a laser-based die transferring process and to create a prototype based on this process.
There are several requirements within this process: component placement accuracy should be very high (±35µm), the temperature of the components at the active side (where the electrical scheme is situated) should not exceed 700 K, the process should be faster than 10 components per second. The following critical issues correlate with these requirements: controlling of the releasing velocity and dynamics, management of heat transport within the die, mitigation of drive instabilities (this is a typical problem associated with a relatively low density gas pushing the higher density object).
In our work we performed experimental study of laser induced transfer and assembly of micro components. The structure, which is considered in experiments, is the sample from the semiconductor wafer, which consists of a carrier polymer tape and a diced silicon wafer. Material of the components was predetermined by domination of the silicon in the microelectronics industry, and this situation is likely to continue for the foreseeable future. Polymer tapes used as a carrier for components are standard tapes used in the semiconductor manufacturing industry. In general, the assembly process involves releasing the parts from the carrier tape surface, transporting them to the specific sites on the target surface, see Fig.1.
Two different approaches of the releasing and propulsion processes were investigated: ablative releasing and thermal releasing. The similarity in those processes is that the carrier tape is transparent for the laser irradiation; light absorption is taking place in the silicon component itself, see Fig.2.
Figure 1: Basic laser induced transfer process top view Figure 2: Schematic side view of the laser transfer release
The driving force for component releasing is the gas formation and consequently the pressure buildup between the glue layer of the carrier tape and the component due to temperature rise. The difference between these approaches is based on the specific optical properties of the silicon, view specific carrier tape characteristics and phenomena, which take place in the whole system. Initial experiments with highly intense IR laser pulses resulted in a high speed unstable releasing dynamics which clearly showed bad placement accuracy. This effect became more pronounced at higher fluency levels. The process takes places at high temperatures (near evaporation temperature of the silicon). The thermal damage of the irradiated die proved to have only “cosmetic” effect since the opposite side where the active scheme is situated does not overheat.
The lower velocities approach – thermal release using low intense IR laser irradiation helps to make the process more controllable and reduce the risk of mechanical damaging of the components. However, the achieved placement accuracy (50 μm) was not enough for standard assembly applications. Experimental results show that the placement accuracy is largely dependent on laser pulse duration. Furthermore, for shorter pulses better placement accuracy is obtained. Another factor responsible for smooth releasing is the spatial profile of the light spot. Nonlinear processes, which take place in the silicon under low intensity infrared laser irradiation, could be considered as a cause for problems with process stability.
A strongly improved component flying dynamics was achieved using a frequency doubled Nd:YAG laser, with low intensity pulses. The nonlinearity of absorption in this approach is decreased due to the much higher, hence localized absorption in silicon. The other factor responsible for improving the releasing behavior is the better quality of the laser spot. The placement accuracy, which is achieved using this approach, is 35 μm for 95% of the components and suitable for many assembly applications for example, for the assembly of RFID (radio-frequency identification) chips to antennas. The releasing part of the process is found to be interesting for the Direct Die Feeders applications and the prototype is being built by Assembléon B.V.
This project is sponsored by IOP – Precision Technology