A series of studies reveal the critical factors.

Tech Tips The profiling basics for attachment of large ball grid array packages using industry-specific precision systems consist of bottom heaters (commonly referred to as preheaters), and a top heater (also referred to as a nozzle heater). The convection settings of temperature and airflow rate are controlled by profiles within system-specific software. When processing on a rework system, profiles are also created. The process is similar to that of a reflow oven; however, the board is not conveyed through the temperature zones but remains in place, and the system heaters are ramped instead.

From a system viewpoint, the process attaches a large BGA to a PCB composed of epoxy, fiberglass, and many layers of copper metal. The substrate will also have various components already attached. Each of these materials has specific thermal properties. Some components are large, some small.

As heated air is applied to the board bottom by the preheaters, conduction takes place within the materials. Copper is an extraordinary conductor of heat with a thermal conductivity of 401 W·m-1·K-1 (at 25°C). With large boards and high thermal mass, the input heat must be significant to heat the localized area of interest. For a Pb-free solder attach profile, a typical preheat will target 150°-180°C at the end-of-zone temperature. This temperature is read at the component/substrate interface, S. A consistent preheat is important because temperature variations in environment will be present, and elimination of this uncontrolled factor is best kept in check.

For large substrates, the preheater will apply most of the heat in order to minimize the temperature difference across a component. This is accomplished by having the preheater temperature higher than the nozzle temperature. This will be the case for the first two or three zones. This practice also prevents the drying or early activation of flux materials. Figure 1 shows an example of a four-zone ramp. Zone two should have a target end-of-zone temperature of 190°C. This is where flux is activated and specifications require a defined amount of time within this range. The flux materials used for Pb-free rework conditions are typically referred to as gel flux or tacky flux. These materials are stable to higher temperatures than those of traditional solvent-based flux materials.

Image

SAC 305 has a liquidus temperature of 217°C and is reflowed at peak temperatures of 235°C. Zones three and four are where the nozzle temperature is raised in conjunction with an increased airflow rate. For relatively large BGAs – 35-50 mm on a side – airflows will generally be high for zones three and four.

To summarize the sequence of events:


There are two values to consider for temperature differences. The first is across the component, and is measured from the center of the component to a corner. Faster profiles will generally create a larger temperature difference across the component. This is important because when the grid begins to reflow, the center may liquefy first, followed by the edges. For Pb-free components, a differential as low as 5-10°C is required. The second temperature gradient is measured from the component top to the substrate bottom. A value for this is directly measured from the placement of monitoring thermocouples on the top of the lid and at a secure attachment on the board bottom, centered below the component. A value of 10°C is generally required and is important for fine-pitch components.

The justification for using process parameters with a higher bottom heat relative to the nozzle temperature is to prevent component warping. Component packaging is considered safe below 250°C. However, with a higher preheat, the board is another limiting factor. Excessive heat will cause board warpage. In addition, components that are localized to the area directly heated by a chimney heater on the bottom side may also attain liquidus temperatures.

The EMPF has conducted profile experiments that varied the top and bottom heaters and measured temperatures at the chip top, board bottom, and target (internal thermocouple at solder ball area). The goal of the profile was to attain the liquidus temperature of the SAC 305 alloy (217°C) and then attain a peak that is 20°C higher.

From the data, several conclusions can be made to fine-tune the profile. The most interesting and useful observation was that the average temperature (between the top and bottom heaters) approximated the target temperature at the interface.

For example, when the top heat was 280°C and the bottom heat 200°C, the target thermocouple measurement at the BGA interface was 238°C, and the TC measurement at the top of the component was safe at 248°C. There were also many useful small empirical observations. For instance, the bottom board TC was on average 15°C lower than the average heat input. Based on these observations and the desire not to reflow the board bottom side, the target temperatures and average heat inputs were determined.

The time and duration of each stage is another consideration. The preheat stage depends on the mass of the substrate, the preheat temperature and airflow rates. When the TC response becomes asymptotic (the temperature is no longer increasing rapidly and has flattened off), the zone duration may be decreased.

Using thermocouples to monitor the temperature response of electronics being reworked, observations can be made such as end-of-zone temperatures and peak temperatures. Using these observations, modifications to the system heaters can be adjusted incrementally to improve the attachment process. These are key elements to proper temperature profiling.

The American Competitiveness Institute (aciusa.org) is a scientific research corporation dedicated to the advancement of electronics manufacturing processes and materials for the Department of Defense and industry. This column appears monthly.

Submit to FacebookSubmit to Google PlusSubmit to TwitterSubmit to LinkedInPrint Article
Don't have an account yet? Register Now!

Sign in to your account