Solder Paste and solder powder spattering is never caused by a single factor, but rather the result of the combined mismatch between the rheological properties of the solder paste itself, the geometric accuracy of the stencil aperture, squeegee downward pressure and rebound speed, the vacuum adsorption stability of the printer, and the time above liquidus during the solder paste reflow phase. When we discovered densely distributed tiny solder balls with a diameter of 15-40 μm next to the 0201 resistor pads on an automotive module production line in Dongguan, we first ruled out interfering factors such as reflow oven conveyor belt jitter and insufficient nitrogen purity, and instead focused on the dynamic behavior during the solder paste printing process.

The solder paste undergoes yield flow under the shear action of the squeegee. If the viscosity is too low or the thixotropic index is less than 0.65, it will be pulled into filaments due to insufficient cohesion at the moment of demolding, and break into mist-like solder powder. These solder powders are carried by the Flux vapor and spatter to adjacent pads or component bodies during the initial stage of reflow, and after cooling, they form tiny solder balls that cannot be cleaned. After we switched to Type 4 solder paste with a metal content of 90.5%, a viscosity of 720 Pa·s (@0.1 rad/s), and a thixotropic index of 0.78, the spatter volume decreased by 63%. At the same time, we reduced the stencil thickness from 100 μm to 80 μm, compensated the aperture area ratio at 1:1.05, and tightened the squeegee angle from 60° to 55°, so that the shear stress is more evenly transmitted to the central area of the pad.

The printing pressure must be dynamically matched with the stencil tension. When the tension is 35 N/cm², the pressure should be set at 5.8 kgf to ensure that the solder paste is completely filled without overflowing. We once failed to calibrate the pressure sensor, resulting in a local pressure of 7.3 kgf, which caused continuous solder balls on the side of the 0402 capacitor. This phenomenon disappeared after replacing it with a high-response piezoelectric sensor and performing zero drift verification every shift. At the same time, we compressed the post-printing residence time to ≤60 seconds to avoid premature volatilization of the solvent on the surface of the solder paste, which would cause edge collapse and aggravate subsequent spattering.

The time above liquidus during the solder paste reflow process must be strictly controlled between 65 and 90 seconds. Exceeding 95 seconds will not only aggravate the oxidation and agglomeration of the solder powder, but also cause the surface tension of the solder to become unbalanced and induce spattering. When we used a KIC thermocouple for actual measurement, we found that the actual heating slope in a certain temperature zone was 1.8°C/s higher than the set value, which caused some of the solder to melt prematurely at the end of the soaking zone, forming micron-sized droplets before entering the reflow peak. After recalibrating the PID parameters and correcting the heating slope to 1.2°C/s, the number of tiny solder balls was stably maintained below 170 ppm.

All improvement actions must be updated synchronously in the FMEA control plan. For example, add a rotational viscometer measurement step to the incoming inspection of solder paste batches, change the stencil cleaning frequency from every 4 hours to mandatory every 2.5 hours, and use XRF to detect whether the oxygen content in the furnace chamber is ≤100 ppm for the first piece of the reflow oven every day. If any item is not closed-loop, all previous adjustments will be invalid. We once missed the inspection of a batch of solder paste in a factory in Suzhou, where the particle size distribution D90 exceeded the standard to 28.3 μm, and the solder ball defect rebounded to 2900 ppm within three days. It was not completely controlled until all inventory was reworked and the upstream atomization process was traced.