Logo

Typical Wire Bond Process Optimization

A wire bond process optimization is essential for bonding process stability. The process optimization defines a process parameter window for ball and wedge bond quality.

The following is a basic procedure for wire bond process optimization:

1. Define ball and wedge bond specifications to be use as target response in optimizing a given application.

  • Ball bond specification: Upper Specification Limit (USL), Lower Specification Limit (LSL) and average value of deformed ball height, ball diameter and shear strength. 
  • Wedge Bond Specification: Lower Specification Limit (LSL), average value and failure mode of stitch pull.


2. Free Air Ball Optimization – based on ball bond diameter and height requirement, compute for the equivalent free air ball volume in terms of free air ball diameter. This will be used as a target reference for optimizing EFO parameter. Consistency of the target free air ball is the main response to be considered.

Example:

FABØ3 = 1.5H2 (H-WD) + (CD3-H3) / 4tan0.5CA + 1.5MBD2 (BHT)

Where:

H = hole size = 30µm WD = wire Ø = 23
CD = chamfer Ø = 38µm CA = chamfer angle = 90°
MBD = mashed ball Ø = 46µm BHT = 9.5µm

FABØ = 36µm for 23µm wire Ø

FAB Parameters

EFO Current = 22.78mA
EFO Time = 1.0msec

Wire Diameter (µm)

23

Actual FAB (µm)

36.05

Average Std Deviation

0.33

SEM Pictures

SEM Pictures of a Free Air Ball (FAB)


3. Define initial working machine parameters to be used as a starting point for optimization. This is normally given by the wire bonder manufacturer or can be based on an existing similar application.

4. Ball Bond and Wedge Optimization – Responses Surface Methodology (RSM), using Central Composite Design (CCD), generate a Design of Experiment (DOE) run matrix using the available working parameters range.
Responses Surface Methodology (RSM),



5. Determination of Process Window – With the aid of a graphical 3D Contour Plot, define the optimized parameter window based on the target response specification.

Example: Below is a typical contour plot result of ball bond optimization using CCD with contour plot. Take note that the 2 parameters, bond force and bond power had been previously identified in the earlier part of the DOE as significant to ball bond response with interaction. Statistical software equipped with experimental design module is available to generate this task.
Typical contour plot result of ball bond optimization using CCD



6. Validation Run – Using the optimized parameters, perform another run to validate the long-term stability of the process. Acquiring a larger sample size, standard deviation and Cpk are the responses to be verified.

 
Copyright by SPT-Group http://www.smallprecisiontools.com