The peak temperature results for all samples are shown in Fig. 1 (a), for the surrounding components located on the top PCBA side. Figure 1 (b) shows the peak temperature results for the mirror component locations located on the bottom side of the PCBA. The variability charts indicate that there was an interaction between the type of heat shield placement locations and the peak temperature for both the centre and corner of the BGA for the top and bottom PCBA sides during the rework process. The highest peak temperature was detected for sample W, which was reworked without a heat shield. For samples X, Y, and Z, reworking by applying heat shield(s) caused the centre and corner temperatures on surrounding components to be significantly lower than that of sample W.
The range of corner peak temperature for both top and bottom PCB sides in all samples was smaller than the range of centre peak temperature. This observation agrees with that of Sommerer et al.13, where the position of the TC wires from the heat source location was correlated with the amount of heat absorption. Weng and Martin14 also reported that TC temperature readings varied depending on the heat shield locations from the heat source. Samples X and Z had less variation in the centre and corner peak temperatures for both the top and bottom PCBA sides, but sample Z had a lower peak temperature range. Sample Y had a lower minimum peak temperature than samples X and Z, but the variability in the peak temperature range in sample Y was higher with the highest peak temperature, making it ineffective in controlling the heat dissipation. The mean for the peak temperatures indicated by the triangular shapes showed a downward trend, indicating a temperature reduction by using the heat shield during the rework process.
Validation by infrared thermography confirmed the heat dissipation of the components in the PCBA samples during rework15. Warmer temperatures, where more heat and infrared radiation are emitted, are indicated by brighter colours (red, orange, and yellow), while cooler temperatures are indicated by purple and dark blue or black where less heat and infrared radiation are emitted16. Here below shows pictures of the temperature distribution without heat shield application and Fig. 2 and the heat shield application in Fig. 3. Photographs of the BGA components on the top PCBA side during the rework process were taken with a regular camera as a reference for the infrared thermography images.
The bright yellow colour was the same as the heat source which came from the hot air nozzle can be seen on the side of the adjacent components of the rework location for sample W in Fig. 2 (b), indicating a surface temperature of 293.1°C. This was due to the rapid temperature increase along the side of the adjacent BGA components' surface during the rework process. The bright yellow colour from the heat source indicated the surface temperature of 333.9°C according to the heat scale was properly contained in sample Z, as shown in Fig. 3 (b). The adjacent components of the rework location were dark orange indicating that the temperature was lower than the heat source. With the heat shield application during the rework process, the temperature of the adjacent BGA components of the rework location decreased, as did the active heat-spreading area.
Figure 4 shows the number of BGA solder joints affected by dye penetration for both top and bottom PBCA sides in all samples. The quantity of solder joints affected in samples Y and Z was reduced from that of sample W, which showed that the heat shields enabled the reduction of thermal damage on the solder joints. Sample X, using an individual heat shield located on each adjacent component of the rework location, has the highest number of solder joints affected by the dye penetration, despite the temperature of the adjacent components of the rework location being much lower than sample W. This result has deviated from the expected outcome of the heat shield placement location.
The inner heat shield wall generates radiation heat from the hot air convection heat interacting with the outside heat shield wall. The interaction created transient heat transfer from the outside to the inside of the heat shield wall thereby transferring conductive heat toward the PCBA surface and BGA solder joints, as shown in Fig. 5. This aligns with the findings of Stein et al.17 on the temperature distribution conditions within the heat shield. Kong et al.18 reported that thermal fatigue failure of solder joints can occur in a lower temperature variation range.
The severity of the dye penetration percentage and its correlation with the centre and corner temperatures of each adjacent component of the rework location are shown in Figs. 6 (a) and (b), respectively. Sample W had the most severe dye penetration at 76–100% and mainly occurred at the bottom side. This was because no heat shield was applied during rework. Samples X and Y had the same dye penetration percentages at the bottom and top, respectively. In addition, 51–75% dye penetration was observed on the bottom side of sample X. The dye penetration percentage of sample Z was less than 50% on the bottom side. Dye penetration of 51% and above occurred when the centre temperature of the adjacent components of the rework location exceeded 195°C. Similar results were observed when the corner temperature of the adjacent BGA component exceeded 210°C.
In this study, the temperature of the adjacent components of the rework location during the rework process can be reduced by addressing the heat shield placement locations. Infrared thermographic images were validated using the TC wires temperature reading to integrate the heat distribution image of the surface temperature of the BGA components and the actual peak temperature of the solder joint array. The sample Z heat shield placement location had the most effective heat reduction on the adjacent components of the rework location, where the mean peak temperatures on the top PCBA side for the BGA component centre and corner were reduced by 6.70% and 6.85%, respectively.
For the bottom PCBA side, the heat reduction for the BGA component's mean peak temperature was 7.58% at the centre and 8.18% at the corner. Dye penetration of more than 50% due to the solder joint crack can be prevented as long as the adjacent components of the rework location temperature can be maintained below 195°C and 210°C for the centre and corner of the BGA component, respectively, during the rework process. This finding is in line with Chen et al.19 that lowering the temperature exposure to the solder joints will reduce the impact of reliability issues, such as the thickening of the intermetallic compound (IMC) layer that affects the shear strength of the solder joints.