Optimizations and Case Studies: Decapsulation of Hardened Epoxy SiC MOSFETs and Diodes via JIACO Microwave-Induced Plasma Etching

J Fail. Anal. and Preven. (2025) 25:1976–1985 https://doi.org/10.1007/s11668-025-02266-9 TOOLS AND TECHNIQUES Optimizations and Case Studies: Decapsulation of Hardened Epoxy SiC MOSFETs and Diodes via JIACO Microwave-Induced Plasma Etching Connor J. Sorrells . Kyle E. Nielsen . Bret A. Walters . Jiaqi Tang . Mark W. McKinnon Submitted: 18 February 2025 / in revised form: 20 May 2025 / Accepted: 26 June 2025 / Published online: 25 August 2025 Ó ASM International 2025 Abstract One of the foremost challenges in the field of SiC MOSFET and diode failure analysis is the effect of thermally modified mold compound on the decapsulation process. The extended total etch time that thermal modification imposes on the process of wet chemical decapsulation has created a niche for new techniques to fill. This paper focuses on use cases for the JIACO microwaveinduced plasma (MIP) etching system and how to best optimize the tool’s settings to facilitate time-efficient decapsulations. The words and data that follow aim to present what has been determined to be a successful alternative for the decapsulation of thermally modified Si and SiC power devices when wet etches prove to be ineffective. Keywords SiC
JIACO
MIP
Thermal modification
Decapsulation Introduction This article is an invited paper selected from presentations at the 50th International Symposium for Testing and Failure Analysis (ISTFA 2024), held October 28–November 1, 2024, in San Diego, California, USA. It has been expanded from the original presentation. The special issue was organized by Prof. Umberto Celano, Arizona State University, and Prof. Frank Altmann, Fraunhofer Institute for Microstructure of Materials and Systems, on behalf of the ASM International Electronic Device Failure Analysis Society. C. J. Sorrells (&)
K. E. Nielsen
B. A. Walters Microchip Technology Inc, 2355 W Chandler Blvd, Chandler, AZ 85224, USA e-mail: K. E. Nielsen e-mail: B. A. Walters e-mail: J. Tang
M. W. McKinnon JIACO Instruments, Feldmannweg 17, 2628 Delft, CT, The Netherlands J. Tang e-mail: M. W. McKinnon e-mail: 123 The arduous task of removing a relatively large volume of material during decapsulation is further complicated by the thermal modification of the encapsulant. All encapsulants have the potential to suffer from thermal modification whenever the package is subjected to extreme temperatures, for instance, during stress tests (temperature cycling, for example) or as a direct result of an EIPD (electrically induced physical damage) event. These sources of extreme, localized heat can cause the encapsulant to polymerize. Polymerized mold compound is highly resistant to decapsulation techniques that utilize wet chemistry, such as nitric acid or sulfuric acid [1]. The amount of time that thermal stress adds to the decapsulation process is difficult to quantify as it varies greatly depending on the severity of the thermal stress and the composition of the encapsulant. Prolonged periods of acid exposure present additional risks, as well—internal components such as wires and exposed source bonding pads can be over-etched by nitric or sulfuric acid, potentially leading to unintended decapsulation artifacts [2, 3]. J Fail. Anal. and Preven. (2025) 25:1976–1985 Proof of Thermal Modification’s Effects on Power Device Decapsulation Two of microchip’s own SiC devices packaged with the same mold compound underwent attempted decapsulation using the same procedure. One unit was an unstressed device, and the other unit exhibited drain to source leakage after temperature cycling. Both devices were subjected to laser ablation using the same settings and procedure to create cavities 1.91 mm in depth to prepare for wet etch decapsulation (Fig. 1a, b). First, 5 ml of 98% nitric acid was administered into the laser cavity of each device via pipette, while the device was heated to 150 °C on a hot plate. Each unit was rinsed with water and acetone before being placed in an ultrasonic bath. Next, each device’s laser cavity was inspected using an optical microscope to assess the effects of the nitric acid treatment. Very little material was removed from either device, but the acid had a greater effect on the good device as some of the wire bond wedges are now visible (Fig. 2a), while the failing device does not have any wire bond wedges visible (Fig. 2b). Next, each unit was placed in a separate beaker to hang suspended (pins pointing up) in a beaker of 96% sulfuric acid solution on a hotplate set to 285 °C. The beakers were filled with enough sulfuric acid to fully submerge the laser cavity. After 30 minutes of etching, a water/acetone rinse, and ultrasonic bath, the die of the unstressed device had been fully uncovered by the acid, except for one small clump of residue below the source wires (Fig. 3a). In contrast, the sulfuric acid treatment had very little effect on the thermally stressed device. The 30 minutes of exposure barely removed enough material to expose the wire bond 1977 wedges of half of the device’s wires, with no areas of the die being uncovered by the etchant (Fig. 3b). The thermally modified device could have been allowed to sit in the acid for longer, but as stated previously, prolonged etch time imposes notable risks of corrosion upon the integrity of wires, bonding paddles, and any exposed areas of the die’s source metal. These risks [2, 3] as well as the inherent health and safety concerns surrounding the use of strong oxidizing agents are completely mitigated by using MIP instead of conventional wet etches. SiC MOSFET Decapsulation via the JIACO Instruments MIP System The JIACO Instruments MIP (microwave-induced plasma) system accomplishes etching by using argon plasma as a carrier to transport atomic oxygen radicals to the sample surface. These radicals then react with the epoxy in the encapsulant, yielding water and carbon dioxide byproducts. The sample stage is submerged in the integrated ultrasonic cleaner to shake off the now-loosened SiO2 filler beads and then dried, which accounts for one full cycle [4]. MIP has been successfully employed in the decapsulation of silver wire [5] and copper wire [6] IC applications for quite some time now, but the JIACO had not yet been tested in SiC applications at the time. However, cases involving SiC devices that are impervious to wet chemistry have become increasingly prevalent over the past few years, and an alternative has become necessary. The exact same SiC unit from the first experiment was allowed to run in the JIACO system overnight using the tool’s ‘‘default’’ settings that are used for copper wire ICs (30W, 15 lm/s, and 29 sccm—parameters that Fig. 1 (a) Unstressed SiC device after laser ablation. (b) Thermally modified SiC device after laser ablation 123 1978 J Fail. Anal. and Preven. (2025) 25:1976–1985 Fig. 2 (a) Unstressed SiC device after 5 ml of nitric acid. (b) Thermally modified SiC device after 5 ml of nitric acid Fig. 3 (a) Unstr (...truncated)