Transcript
Application Tech Note Soldering with High-Power Diode Lasers – Process Optimization Introduction The trend towards miniaturization in the electronics industry has led to an increase in the population density of circuit boards across a whole range of products—from cell phones to under-the-hood sensors for the automotive industry. In these circumstances wave soldering is not always possible, and this has led to a demand for selective soldering processes. The use of direct-diode lasers for laser soldering has, therefore significantly increased within the last two years. The majority of systems currently employed for laser soldering produces joints sequentially, as opposed to conventional soldering processes, such as wave soldering, which produces a large number of joints simultaneously. The average power of the laser required in most cases is relatively low. Most joint configurations require less than 10 watts to produce a single joint. However, there are some solder reflow applications where a number of pads are scanned simultaneously, and these require higher average power. Sequential soldering puts the process well within the power range of different types of diode laser devices. Power density requirements (watts/cm2) are also relatively low. Essentially a relatively slow (in laser processing terms) heating process, melting typically takes place in the 100s of milliseconds. In fact, if excess power density is employed, poor joint quality results due to spattering. A number of worldwide commercial manufacturers are now offering fully integrated diode laser soldering systems. These systems come equipped with X-Y positioning stages and/or robots to provide precise, contactless soldering.
Optimization of the soldering process A previous Application Note, “Soldering with High-Power Laser Diodes,” laid out the general approach for laser soldering using Coherent FAP™ systems. The purpose of this new note is to expand on this method and to provide additional information, both on how the soldering process can be optimized and how the defects associated with non-optimized joints can be identified.
Since the start of the electronics industry, leadcontaining solders have been used almost exclusively. The most widely used composition has been the low-melting point (183°C) eutectic lead-tin (63%Sn/37%Pb) composition. Lead-free soldering has been introduced as a response to proposed legislative restrictions on the use of lead in electronics and by marketing activities in the Far East and Europe. However, the majority of soldering applications continue to use this or very similar compositions. As board densities increase and pad size decreases, the requirement for precision in the soldering process has led to an increasing need for the accuracy that lasers provide. This improvement in accuracy has put increasingly demanding requirements on the positioning system. Two approaches are currently employed: Either positioning the workpiece beneath the stationary laser spot, or manipulating the fiber-coupled laser beam output. Both approaches require precise alignment. Typically, this has been achieved using a visible laser beam, usually a red laser diode, as a ‘pointer.’ There are limitations to this approach, because the red beam never entirely nor accurately delineates the real laser spot. However, this can be achieved if a CCD camera is employed to directly view the laser spot, and using accessories (now available) that allow co-axial viewing of the laser beam in real time. Because of the greatly reduced cost of such cameras, almost all precision soldering stations are now equipped with CCDs. All of the cameras supplied have some sensitivity in the infra-red range and, hence, our approach is to align using a remotely viewed (via a TV monitor) low-power diode laser beam. This has two benefits: First, it is far safer. Second, because the diameter of the laser spot does not increase in size with an increase in power, the exact area covered by the laser spot is seen. This is critical if very careful alignment is required, as is the case in many microsoldering situations.
dard 44-pin chip carrier with a 2.54 mm (1/10") pin pitch was soldered into a standard conformal-coated FR4 printed circuit board material. To achieve reproducible solder joints, 1.7 mm OD (0.07") solder pre-forms were used. The pre-form was selected over solder paste or solder wire feeding, as the use of the pre-forms provides more reproducible results, especially in a laboratory environment. In this case, 96.5%/3.5% tin/silver solder pre-forms were used, melting point 183°C, supplied by AlphaFry Technologies. These are coated with an RMA-type low-solids flux, making it a ‘no clean’ solder. The samples shown here were soldered in air using an 810 ±10 nm Coherent FAP™ System with a total average power capability of 30 watts. The energy was delivered via an 800 µm diameter delivery fiber and a 1:1 Optical Imaging Accessory (OIA), which had a working distance of 33 mm and produced an 800 µm spot. To ensure a valid optimization trial, soldering time was fixed at 0.8 sec.: Therefore, average power and solder composition were the primary variables studied. Average power was changed in increments of 2 watts over the range at which solder joints were produced. At each setting, a total of 13 pins were soldered. Results over the optimum range of 6 to 12 watts average power are reported here. To confirm and to expand on visual observations, metallographic samples were prepared for examination. Transverse diametral cross-sections of several joints from each parameter
Lead-tin Soldering –Over-heated Joints
Experimental details A single, standard solder joint configuration was used to demonstrate a generic experimental approach to minimizing heat input. A stan-
Figure 1. PbSn solder, 10 watts, 0.8 sec
Soldering with High-Power Diode Lasers – Process Optimization Lead-Tin Soldering –‘Dry’ Joints
Lead-tin Soldering –Over-heated Joints
Lead-Tin Soldering –Optimized Joints
Figure 2. PbSn solder, 10 watts, 0.8 sec
Figure 3. PbSn solder, 8 watts, 0.8 sec
Figure 5. PbSn solder, 6 watts, 0.8 sec
Figure 4. PbSn solder, 8 watts, 0.8 sec
Figure 6. PbSn solder, 6 watts, 0.8 sec
setting were prepared using conventional metallurgical techniques. Figure 1 shows how a slight excess of power (for a fixed solder time) produces a slightly overheated appearance to the joint, with some degradation of the flux apparent and a lessclean appearance. Figure 2 is a transverse section across the diameter of the pin (0.5 mm/0.02" diameter). The pin has been sectioned and prepared using a standard metallurgical technique. The section is then acid-etched to reveal the internal grain structure of the joint. Figure 2 confirms this slightly overheated microstructure with some porosity showing. The profile of the joint is also confirmed to be less than ideal. Some slumping of the solder has occurred and grain structure generally appears coarse. Figure 3 shows what appears from visual examination alone to be close to optimum parameters for this particular joint configuration. Figure 4 shows a porosity-free microstructure with good wicking of solder down into the board. All contacting surfaces appear wellwetted. Grain structure appears fine.
Figure 5 shows a noticeably dry joint. Figure 6 shows a cross-section of a dry joint (not as serious as Figure 5). Serious porosity and lack of wetting are both visible in this cross-section, although the profile of the solder joint is not as poor as that shown in Figure 5. This suggests that at this low power, process variability increases, as no such lack of consistency was seen at higher average powers. There is evidence from all these cross-sections, that in addition to the macro-scale defects, there are more subtle variations in the microstructure that would significantly affect the mechanical properties of the joint. These will be covered in a future Application Note.
Conclusions • High-quality solder joints can be produced using Coherent FAP™ Systems by a straightforward process optimization technique. • Laser power has a critical effect on solder joint quality. • Metallography has confirmed that initial visual observations of solder joint quality were correct. • Small but highly repeatable changes in laser power can produce controlled metallurgical changes in the microstructure.
Coherent, Inc. Produced by the Laser Application Center at: 5100 Patrick Henry Drive Santa Clara, CA 95054 Telephone: 877-434-6337 Fax: 408-764-4329 E-mail:
[email protected] Web: www.CoherentInc.com MC-045-02-1M0302
03/2002
