Tracking Fluid Flow Patterns Around and Under the Wafer During Chemical Mechanical Planarization

 

Investigator:  Nicole Braun

 

Funding: Cabot Microelectronics and Intel

 

Background:

 

            Chemical Mechanical Planarization, also known as CMP, is the process used to prepare silicon wafers for deposition in the semiconductor industry.  A schematic of a typical CMP set up can be seen in Figure 1.  The silicon wafer and the platen are placed in contact.  The platen, which is covered with a polishing pad, rotates and is used to mechanically polish the wafer surface.  The wafer mimics the rotation characteristics of the platen. In addition, slurry is deposited between the wafer and pad and is used as lubrication.  The slurry is responsible for chemically weakening the wafer surface, making it susceptible for planarization.

 

Figure 1  A typical set used to planarize wafers for CMP. Image from www.agc.com

 

            CMP is an important area of research as the planarization process is a driving factor enabling the use of smaller semiconductor components.  Though CMP is thoroughly used in industry, it is little understood.  Described as an art, rather than a science, many researchers are attempting to understand the chemical and mechanical mechanisms present in hopes to improve the current state of the art. 

           

Research Method:

 

            The goal for my masters thesis is to measure the fluid patterns around and under the wafer during CMP polishing.  By more thoroughly understanding the fluid directions and paths, we hope to discover more about the removal processes present. 

            In order to measure the fluid patterns, Particle Image Velocimetry, or PIV, will be employed.  PIV is a method that allows for measurement of an entire flow field by adding tracer particles to the studied fluid.  A typical schematic can be seen in Figure 2.  A flow that contains seeding particles is illuminated, typically via a double pulsed NG/Yag laser.  A CCD camera detects signal given off by the seeding particles and captures their location in space.  In post processing, the images are then correlated and movement data, such as velocity, can be detected. 

Figure 2  A typical set up for PIV.