JEDEC's Generation of Wire Bond Pull Test Methods to Address Pulling of Copper Wire Bonds

Abstract

More than 50 years ago when the wire pull test method was initially added to Mil-Std 883, in Condition D of Method 2011, Bond Strength (Destructive Bond Pull Test), the test procedure and minimum pull force values were based on pull testing of mostly ultrasonic wedge bonded aluminum and gold wires of just a few different diameters. The minimum pull force values from that original data were extrapolated to cover a much wider range of wire diameters for both gold and aluminum wires. Since the release of this test method the electronics industry has manufactured copper ultrasonic wedge bonds, widely adopted copper thermosonic ball bonding roughly 15 years ago, and even developed a niche market for silver thermosonic ball bonding. The industry also developed specialty bonds such as security bonds, reverse bonds also called "stitch on ball", and even multi-loop wires and ribbons. In all that time neither the test procedure nor the minimum pull force values in Method 2011 were reviewed to determine their appropriateness for these new materials or new types of bonds, even though the industry widely referenced the test method for all of them and thus, by default, accepted its use for all of them. In late 2013, I led a working group within JEDEC's JC14.1 subcommittee, Reliability Test Methods for Packaged Devices, to update JEDEC JESD22-B116, Ball Bond Shear Test Method, to expand its scope to include the shearing of Cu ball bonds. It took the working group three years to address the necessary technical issues to ensure that the revised test method adequately addressed the shearing of copper ball bonds and propose minimum acceptable shear values. The working group produced a greatly improved document with drawings and images depicting the different shear fail modes of both gold and copper bonds and added several informative annexes to aid in the performing of the test method. By 2018 it was apparent that none of the most commonly referenced wire pull test methods in the electronics industry had made any significant progress in updating their documents to include Cu wire bonds. Therefore, the JC14.1 working group agreed to work jointly with the JC-13.7 Subcommittee, New Electronic Device Technology, to create a new, wire pull test method document under JC14.1 that would be a companion to the JESD22-B116. This new document will use Method 2011, Conditions C and D as its basis, but expand on its scope to cover copper wire bonds, both ultrasonic wedge and thermosonic ball bonds. The new test method will describe the process for a ball pull test and a stitch pull test that are referenced for copper bonds by AEC Q006, Qualification Requirement for Component Using Copper (Cu) Wire Interconnection. The test method will also provide guidance on how to perform pull testing on several different bond types used today including reverse bonds, multi-loop bonds, and stacked die. The working group plans to propose minimum pull values for copper wire bonds which JC14.1 will reference in JESD47, Stress-Test-Driven Qualification of Integrated Circuits. After the joint working group completes its work, which is targeted for some time in 2022, JC13.7 would then be able to use the output of this working group to update Method 2011 Conditions C & D. This paper will first briefly discuss the updates made to B116 to cover Cu wire bonds, but mainly focus on the work that has so far been completed by the joint working group, including a general outline of the proposed new document, JESD22-B120, Wire Bond Pull Test Methods.