-
InSb IRFPAs主要由InSb芯片和硅读出电路(silicon readout integrated circuit,Silicon-ROIC)借助铟柱阵列倒焊互连混成[8],随后填入底充胶以提高铟柱的可靠性[12]。在建立InSb IRFPAs的有限元结构模型时,通常认为倒装焊之后的探测器处于零应力状态。考虑到在铟柱阵列和网状底充胶组成的中间层中,铟柱阵列与中间层的体积比通常小于20%,作为初步分析,可忽略铟柱阵列的影响。底充胶固化中引入的固化应力通常也比较小,可以忽略[13]。鉴于探测器结构的对称性,为降低计算成本,施加对称边界条件,建立起InSb IRFPAs的二维结构分析模型,如图1所示。结构模型建立过程中,采用自底向上的建模方法,三层结构均采用PLANE182单元。
Figure 1. Two-dimensional model of InSb IRFPAs (local model). (a) Without isolation trough; (b) With V-shaped isolation trough added; (c) Bottom of V-shaped isolation trough is connected with preset crack; (d) Preset crack is away from the bottom of V-shaped isolation trough with a distance of 1 μm
为提高填充因子,InSb 芯片大多采用台面结构形式,此时InSb芯片正面会刻蚀出光敏元隔离槽,即文中的V型隔离槽,填充底充胶后,隔离槽处会充满底充胶。为评估光敏元隔离槽对InSb芯片中累积热应力的影响,给出了不包含隔离槽结构的探测器模型,具体如图1(a)所示;加入V型隔离槽后探测器的结构模型如图1(b)所示,其中V型隔离槽开口宽度为10 μm,槽深4 μm,槽口坐标位于X=3.995 mm到X=4.005 mm的范围内。InSb晶体生长过程中会引入数量不等的位错线,在液氮冲击下,这些位错线可认为是裂纹成核的起源,伴随着热失配应力的增加,位错线会长大、扩展,最终形成穿通裂纹。为了描述隔离槽对InSb芯片中位错线的影响,在V型槽底部预置了不同长度的裂纹,其中预置裂纹下端点和V型槽底部重合的结构模型如图1(c)所示,预置裂纹下端点距V型槽底部为1 μm的结构模型如图1(d)所示。模型中预置裂纹的长度设定为a。
网格划分时,采用自由网格进行单元划分,并对隔离槽处进行加密处理。预置裂纹尖端具有奇异性,裂纹尖端周围的单元是二次奇异单元[14],因此在APDL代码中使用“KSCON”命令对裂纹尖端进行网格化分,第一行单元的半径是裂纹长度的1/10以下,裂纹尖端一圈的单元数量是15。在结构模型的左下角施加固定约束以防止结构移动。
-
为准确计算InSb IRFPAs中累积的热应变和热应力,材料模型均采用温度相关模型,其中,Silicon-ROIC视为各向同性线弹性材料;忽略底充胶固化过程的底充胶视为线弹性材料;InSb芯片视为各向异性线弹性材料,将InSb芯片的面内杨氏模量值设定为其体积弹性模量值,法线方向的杨氏模量设定为其体积弹性模量的30%,这归因于InSb芯片在背减薄过程中引入的损伤。具体数值如图2和表1所示[15-17],其中E为杨氏模量,v为泊松比。
α为底充胶的线膨胀系数,在77~300 K的温度范围内,可由公式(1)给出:
式中:T的单位为K。
Figure 2. Coefficients of thermal expansion (CTE) depending on temperature for InSb, underfill and Silicon-ROIC
Materials Elastic modulus E/GPa Poison's ratio v Temperature T/K Length/mm×height/mm InSb 409 (In-plane) 0.35 77-300 10×0.01 123 (Z-direction) Underfill 0.0002/α 0.30 77-300 10×0.01 Silicon-ROIC 163 0.28 77-300 12.8×0.3 Table 1. Mechanical parameters and specific sizes of InSb IRFPAs
Effects of isolation trough on cleavage of InSb chip in InSb detector
doi: 10.3788/IRLA20210599
- Received Date: 2021-08-24
- Rev Recd Date: 2021-10-28
- Publish Date: 2022-06-08
-
Key words:
- InSb detector /
- isolation trough /
- stress /
- energy release rate /
- chip cleavage
Abstract: Local cleavage of indium antimonide (InSb) chips restricts the improvement of the yield of InSb infrared focal plane detectors (IRFPAs) under cyclic liquid nitrogen shocking tests. Stress concentration effect may appear in isolation troughs surrounding mesa-junction photosensitive units, drives the dislocation line to nucleate and to propagate, ultimately to punch through InSb chips. In order to analyze quantitatively the influence of isolation troughs on the cleavage of the InSb chip, a structural model of InSb IRFPAs was established, and the in-plane normal stress distribution on the InSb front surface was obtained. Stress concentration phenomena appear on the bottom of V-shaped isolation trough added. Then, the assumed initial cracks with different lengths at the bottom of V-shaped isolation trough were put, here the preset initial cracks were employed to describe dislocation lines in InSb wafers, and were perpendicular to the InSb chips, and obtained the relationship between the energy release rates and the preset crack length. After analyzing these results, the in-plane stress concentration phenomena appears exactly at the bottom of V-shaped isolation trough, and originates from the added V-shaped isolation trough; the enlarged stress at the bottom of V-shaped isolation trough could drive the dislocation lines in the InSb chip to grow and to punch through the InSb chip, thus, the macro cleavage of InSb chip is created; once the preset cracks connect directly with the bottom of V-shaped isolation trough, cleavage of InSb chips is more likely to appear. All these conclusions provide a new perspective to understand the cleavage of InSb chips.