Mello D, Nacucchi M, Anastasi G, Sacchi M, Ricciari R, Burresi E, Tapfer L
STMicroelectronics, Stradale Primosole 50, Catania, Italy.
ENEA, Research Centre of Brindisi, SS 7 Appia km706, Brindisi, Italy.
Ultramicroscopy. 2018 Oct;193:143-150. doi: 10.1016/j.ultramic.2018.06.018. Epub 2018 Jun 30.
In fabrication of microelectronic devices two important steps are often recognized: i) all the processes performed on the wafer in order to build the active part of the devices and, ii) the assembly and packaging processes, typically performed on a chip, in order to fabricate interconnections between active part and exterior. The wafer back side is an active part of power devices and is normally coated with a stack of Ti-Ni-Au or Ti-Ni-Ag layers to ensure the best electrical contact with the frame on which the device is attached prior to the packaging. An important failure mechanism related to this particular process step is related to the diffusion of Ni to the surface of the stack that causes its oxidation on the back metal surface, inhibiting the correct connection to the metallic frame. Auger Electron Spectroscopy (AES) is a powerful analytical technique that can be used to detect this failure mechanism for its very high sensitivity in the characterization of surface layers. Unfortunately, its results are mainly qualitative. Quantitative extrapolations can be inaccurate using library Elemental Relative Sensitive Factor (ERSF) because they are mainly referred to a silicon substrate and could be not valid for a different matrix. A most accurate evaluation of the ERSF is based on the analysis, under identical experimental condition, of standard materials (with known concentration) that should be similar to the unknown sample and having the same matrix. However, the production of this kind of standard is not easy due to the mobility of Ni in Au and Ag. Another commonly used technique is the Energy Dispersive X-ray Spectrometry (EDS) which is less sensitive than the Auger and not sufficiently adequate for a quantitative analysis due to the limitation of the matrix correction methods. Recently, a new method to perform quantitative analysis by using Transmission Electron Microscopy (TEM) EDS was proposed, starting from bi-layers of pure elements. In this work we show how the use of TEM-EDS quantification of Ni in Ag could be a successful method for ERSF evaluation in order to overcome matrix effect in Auger quantification. For this purpose suitable foils of Ag/Al and Ni/Al were used. The validation of the method was performed on a sample with a tri-metal stack of Ti/Ni/Ag previously stimulated by means of a thermal budget to induce Ni migration on Ag surface. The quantitative analysis allowed us to use this characterized sample as AES standard for ERSF calculation.
在微电子器件制造过程中，通常认为有两个重要步骤：i）在晶圆上执行的所有工艺，以构建器件的有源部分；ii）组装和封装工艺，通常在芯片上进行，以制造有源部分与外部之间的互连。晶圆背面是功率器件的有源部分，通常涂覆有一层Ti-Ni-Au或Ti-Ni-Ag层，以确保在封装之前与器件所附着的框架实现最佳电接触。与这一特定工艺步骤相关的一个重要失效机制与Ni扩散到堆叠表面有关，这会导致背面金属表面氧化，从而抑制与金属框架的正确连接。俄歇电子能谱（AES）是一种强大的分析技术，因其在表面层表征方面具有非常高的灵敏度，可用于检测这种失效机制。不幸的是，其结果主要是定性的。使用库元素相对灵敏度因子（ERSF）进行定量外推可能不准确，因为它们主要参考硅基板，对于不同的基体可能无效。对ERSF进行最准确的评估是基于在相同实验条件下对标准材料（浓度已知）进行分析，这些标准材料应与未知样品相似且具有相同的基体。然而，由于Ni在Au和Ag中的迁移性，这种标准的生产并不容易。另一种常用技术是能量色散X射线光谱法（EDS），它比俄歇能谱法灵敏度低，并且由于基体校正方法的局限性，不足以进行定量分析。最近，有人提出了一种使用透射电子显微镜（TEM）EDS进行定量分析的新方法，该方法从纯元素的双层开始。在这项工作中，我们展示了如何使用TEM-EDS对Ag中Ni的定量分析作为评估ERSF的成功方法，以克服俄歇定量分析中的基体效应。为此，使用了合适的Ag/Al和Ni/Al箔。该方法的验证是在一个具有Ti/Ni/Ag三金属堆叠的样品上进行的，该样品先前通过热预算进行了刺激，以诱导Ni在Ag表面迁移。定量分析使我们能够将这个已表征的样品用作计算ERSF的AES标准。
