Get Started

Process Functional Modeling in Microchip Manufacturing: A Case Study on Innovative Problem Solving

April 2, 2023

This case study explores the usage of Process Functional Modeling (PFM) as a problem-solving tool in microchip manufacturing. PFM is an innovative way of thinking that involves breaking down a process into its components on multiple levels and generating innovative solutions. The tool enables us to move from general to specific, from process to operations, operations to components, and components to functions. In this article, we provide a real-life example of how a team applied PFM to the microchip manufacturing process. We want to demonstrate its effectiveness in creating a model of failure in the context of the whole process. This information is accessible to all, regardless of prior knowledge of microelectronics.

For general information on PFM, read our related articles Process Functional Modeling with PRIZ Innovation Hub and  Functional Modelling with PRIZ Innovation Platform, or watch our Webinar: The Art of Functional Modelling with PRIZ Platform.

PFM | wafer

Process background

The process is related to microelectronics – microchip manufacturing. The purpose of it is to create a SiO2 layer on the surface of a Si wafer.

Vertical furnace to heat the wafers in the Q2 atmosphere and perform oxidation on the wafer surface.

The oxidation occurs on the front and the back side of the wafer: Si (wafer) + O2 (gas within the furnace) –> SiO2 (thin film layer on the wafer surface).

The process is performed in batches of up to 150 wafers in every run.

Create a SiO2 thin layer with a certain thickness and low sigma – low standard deviation of the thickness between the wafers and within the wafer

Wafers from the lower zone have higher thickness and significantly higher within wafer sigma (standard deviation of the thickness within the wafer)

Problem Statement

As the platform suggests, we built the problem statement using the three steps as shown below.

Problem statement

The final problem statement is as follows:

The process conditions are different in the different zones of the furnace. Wafers in the low zone present worse performance.

The Process

Before delving deeper into the subject, let’s review the specifics of this oxidation process itself. The process is performed inside a vertical furnace, where the furnace temperature is maintained by three different heaters in three zones. The temperature in these zones is measured using various thermocouples.

The process consists of several distinct steps:

  • The wafers are loaded from the FOUP (a special plastic box for storing and transporting wafers) into a boat, which can hold up to 150 wafers in one run. The furnace is kept open at 300C, and N2 gas is supplied to both the furnace and the exhaust to prevent outside air from penetrating.
  • The boat, loaded with wafers, is moved up and inserted into the furnace.
  • The furnace, along with the boat and wafers, is left to dwell at 300C until it reaches a stable temperature.
  • Once stable, the temperature is increased to the oxidation process temperature, and O2 gas is supplied to the furnace to provide oxidation.
  • After the oxidation process is completed, the furnace is cooled down to 300C.
  • The boat is then moved down and removed from the furnace.
  • The wafers are unloaded from the boat into the FOUP.
  • The thickness of the SiO2 layer on the wafers is measured using the ellipsometric method.

The entire process can be broken down into the following operations:

  1. Pre-loading – loading wafers onto the boat
  2. Loading – inserting the boat with the wafers into the furnace
  3. Stabilization – dwelling and increasing the temperature up to the target
  4. Oxidation – supplying O2 to the furnace
  5. Unloading – removing the boat from the furnace
  6. Boat cooling & wafers unloading – unloading the wafers from the boat to the FOUP
  7. SiO2 thickness measurement

After defining the operation types, our process now looks like the following:

typed operations within the process

Functional Models

Following the process of PFM, the next phase in the analysis involves constructing a Functional Model for each operation in the process. This is particularly important to build PFM for every operation when there isn’t a clear culprit for the problem at hand.

An example of the second operation (LOADING) is shown below.

SFM of loading operation in the process.

After building the functional model for each operation in the process, the platform calculates its respective functional and problematic levels. The effectiveness of each operation is determined by the ratio of its functional rank to its problematic rank.


It is evident that the LOADING and UNLOADING operations appear problematic. Additionally, the LOADING operation has the lowest effectiveness ratio. This result is significant as it defines the direction of our thinking:

The problem arises during the movement of wafers within the furnace.

Using the concept of wafer movement, let us explore potential models for failures of the bottom zone wafers’ thickness sigma (i.e., the standard deviation of the thickness).

Model 1. The loading of the boat with wafers causes overheating at the bottom zone of the furnace

As the wafers are loaded into the furnace at 300C, the cold wafers cool down the bottom thermocouple (TC). This causes more power to be added to the bottom heaters, leading to strong overheating of the bottom zone.

Model 2. The Loading of the boat with wafers causes the presence of air residue in the bottom zone of the wafer.

While loading the wafers, the flow of N2 primarily removes the air from the furnace, but some amount may remain, which could collect in the bottom zone of the furnace. As a result, the wafers in the bottom zone will start oxidizing earlier and receive more O2, leading to an uneven SiO2 thickness distribution.

Model 3. The unloading of the wafers from the furnace will result in additional uncontrolled oxidation on the bottom zone wafers

During the unloading process, the wafers at the bottom of the boat are exposed to the atmosphere first, which results in them having a higher temperature compared to the rest. Additionally, due to their proximity to the massive bottom flange, they retain their temperature for a longer period.

Possible solutions

There are most likely more than just these two possible options to solve the problem. We will leave that to you to think about and propose in the comments. These are the two things we ended up implementing in the process.

  1. Disconnect the temperature control during the loading and unloading of the boat with wafers
  2. Use Ar instead of N2 during the loading and unloading

That is it. You can find the fully detailed project published under PRIZ Hub at SiO2 thin film creation in Diffusion furnace.

Leave A Comment

Read also