英语翻译Determination of die thicknessFigure 5 shows the different die thickness design ofTFTBGA package,and the maximum die thickness of novoid happening issue is tried to be found as the suggestionof wafer grinding.Because of the large thicknes

英语翻译Determination of die thicknessFigure 5 shows the different die thickness design ofTFTBGA package,and the maximum die thickness of novoid happening issue is tried to be found as the suggestionof wafer grinding.Because of the large thicknes
英语翻译
Determination of die thickness
Figure 5 shows the different die thickness design of
TFTBGA package,and the maximum die thickness of no
void happening issue is tried to be found as the suggestion
of wafer grinding.Because of the large thickness
difference,the compound melt will run faster around the
die attached region than the die top.The air trp then might
be formed as the melt flows back.After mold flow simulation,the air traps occurrence
possibility and locations are easily to be predicted.For the
design of 13 mil die thickness,there is no air trap
happened even the melt flows are little unbalanced (shown
by Figure 6).But for the 18 mil die thickness design
shown as Figure 7,the air traps happened at the last filled
areas of die top regions due to the uneven melt fronts.The
study tell us the die thickness should be as small as
possible to introduce the balanced flow pattern and reduce
the void occurrence.A further study to predict the die thickness limitation
of no void happened issue is also made for the TlTBGA
package,and the 16 mil die thickness is found to be the
maximum value of grinded wafer (refer to Figure 8).The
thicker the die thickness is,the more risk it will take to
form the air trap.
Solutions of heat spreader thickness
The thickness of the heat spreader also affects the
flow pattern and will introduce the air trap due to the
unbalanced mold flow.A LQFP 208L DPH (die pad heat
spreader) package is studied to simulate the melt front
advancement.Figure 9 shows the cross section of the
LQFP 208L package,and the analysis results of 6 and 10mil DPH thickness designs are compared to find a better
solution.Due to the thickness difference of top and buttom
cavities,the melt flows faster on the top side and reaches
the buttom cavity through the pad openings.For the 6 mil
DPH design,the unbalanced flow is not so serious to form
the air trap (see Figure lo),so it is acceptable.But for the
10 mil DPH case,the air trap forms under the heat
spreader because of the flow back effect of melt front
(Figure 11).A short shot sample is also made to verify the
simulation result.Figure 12 shows the air trap positions of
experimental sample and mold flow predition result are
almost the same.The accuracy of mold flow analysis isapproved again to find the possible problems before real
mass production here.

英语翻译Determination of die thicknessFigure 5 shows the different die thickness design ofTFTBGA package,and the maximum die thickness of novoid happening issue is tried to be found as the suggestionof wafer grinding.Because of the large thicknes
冲模厚度测定
图五显示的是TFTBGA组合不同的冲模厚度设计以及在微量晶圆打磨的情况下试图建立最大无隙冲模厚度的情形.鉴于厚度的极大差别,冲模连接部分复合溶解物的反应速度将会大于上模部分.因此,当溶解物倒流,可能会形成气阱.经过铸模流量分析后,出现气阱现象的几率和位置可以被轻易地预测出来.对于13密耳的冲模厚度设计,即使熔体流动有一些不平衡（见图6）,也不会发生任何气阱现象.但是对于图7中18密耳的冲模厚度设计,由于不均匀的熔体峰值,气阱将会出现在上模中的最后充填区.这一研究告诉我们,冲模厚度的数值应尽可能的小些,以此来引入平衡流动形态,减少空隙.另一项为TITBGA组合服务的深度研究是针对如何预测无隙冲模厚度极限的,最后结论表明16密耳的冲模厚度对于晶圆打磨来说已经是最大值了（详见图8）.当厚度越厚,出现气阱的风险越大.
散热器厚度解决方案
散热器厚度同样影响着流动形态,并且也会由于熔体流动的不平衡而造成气阱现象.LQFP 208L DPH（冲模垫散热器）组合被用于研究模拟熔体锋面提升.图9是LQFP 208L组合的横截面,以及6密耳DPH和10密耳DPH厚度的设计对比.鉴于底部和顶部插线孔的厚度差异,顶端的熔体流动速度比较快,并且在冲模垫打开时达到底部插线孔.对于6密耳的DPH设计而言,不平衡的熔体流动并不会严重到形成气阱（见图10）,所以是可以接受的厚度.但对于10密耳厚度的DPH,由于熔体倒流影响到溶解峰值（图11）,气阱会在散热器下方形成.我们又使用了一个欠注样本来检验模拟结果.图12表明,实验样本的气阱位置和熔体流动预测结果基本相同.熔体流动分析于产品大量生产前找出潜在问题的准确性再一次得到了证明.