Design and realization of a vacuum compatible substrate transfer system

The goal of this project is to design, realize and test a manipulator for handling flat substrates, e.g. silicon wafers, in a vacuum environment. Key requirements are high cleanliness (both particle and molecular) and minimizing the amount of moving mass.

PhD Candidate: ir. Rick Baade
Supervisor: dr. ir. P.C.J.N. Rosielle
Promotor: prof. dr. ir. M. Steinbuch
Project Financing: VDL ETG / Impulse 2
Project Period: December 2015 – December 2019

Manufacturing of integrated circuits (IC) such as microchips and LED’s, among others, involves multiple process steps such as patterning (lithography), etching and layer deposition on flat substrates. Each of these process steps is performed on dedicated tools, often under vacuum conditions. One category of these tools are so called cluster tools (see figure 2). A cluster tool consists of a central transfer chamber surrounded by several process chambers and one or more load locks. The transfer chamber contains a robot. 

The functionality of the robot is to transport the substrates throughout the tool (process chambers and load lock). Critical requirements of the robots are: availability, motion speed, motion accuracy and cleanliness (particle and molecular cleanliness). These robots usually have three degrees of freedom in cylindrical coordinates, namely R ϕ and Z. Commercially available robots are mostly SCARA or double SCARA robots, as shown in figure 3. 

The moving mass of these robots in the order of 10 – 40 [kg]. High cleanliness robots typically show a large moving mass, since high cleanliness is achieved by shielding mechanical components using enclosures and seals. Commonly used substrates such as silicon wafers have a mass of approximately 100 [g]. This means that the effective payload of these robots is less than 1% of the total moving mass. 

The objective of this project is to design a high cleanliness in vacuum wafer manipulator while minimizing the moving mass, a target is set at a reduction with factor 5 to 10 compared to currently available robots.