Test Wafer, part 2

First of all, thanks to everybody who participated in the test wafer usage poll in the last post. I received some decent feedback and below are the results:

For 150 and 200mm FABs (based on 15 data entries)

The data show that test wafers play a massive role in the day to day FAB business. Most FABs have significant amount of test wafers as well as allocated carriers to test wafers.

For the 300mm FABs:


The general theme is the same, but there are 2 important remarks to make:

  • there were only 5 feedback votes – it seems in 300mm the overall competition might be tighter and the willingness to share data is limited – so the data above might be not a good reprensentation of the 300mm situation overall
  • there were 2 data sets indicating less than 10% of carriers are used for test wafers – this is interesting since the overall test wafer percentage in the FABs is still high.

    My guess is that in 300mm FABs with the high material transport automation capabilities there might be single wafer stockers in use to store test wafers outside of FOUPs.

Overall, the data set from the polls show a picture which I was expecting. Test Wafers are plenty in the FABs and they consume a lot of carriers as well as storage space. Here is a high level thought to get a feeling for real numbers:

Let’s assume a small/mid size wafer FAB has the following WIP related indictors:

  • 10,000 wafer starts per week
  • average 30 mask layers
  • 1.8 days per layer cycle time on product lots
  • average lot size 24 wafers for product lots

That will translate into an overall FAB production WIP of about 80,000 wafers, which will sit in about 3,400 carriers. Based on the poll data that would also mean that this FAB has an additional 50,000 … 90,000 test wafers sitting in 2,000 … 4,000 test wafer carriers.

I think this is mind boggling – at least compared to how much is typically talked and written about test wafers and test wafer management.

So how come that there are such massive amounts of test wafers in the FABs ? For sure it has to do with the impact of test wafers – especially when they are missing. ( see the part 1 of the test wafer post)

In my next post – which will be the 3rd and final part for the test wafer topic – I will discuss some of the common strategies and methods:

How to have control over these massive amounts of test wafers and test wafer carriers and how to make sure that equipment uptime is not (too often) impacted by missing test wafers.

2 thoughts on “Test Wafer, part 2”

  1. Some thoughts on this, based on my experience in a 300, 200 and 150mm fab. The much higher level of physical automation in 300mm Fabs allows the individual test wafers to be processed into a foup. A big enabler is the separation of lot and wafer box at 300mm. So that up to 25 individual lots can be handled in one foup. This was formally not possible with 150 and 200mm level of automation(SEMI std.) . Even if individual companies have found a workaround for this.
    In addition in elderly Fabs monitor wafers are not always part of the MES.

    At 300mm I have also seen many more inspections on production wafers that have replaced regular particle or layer thickness measurements with test wafers.

    Due to the higher OEE requirement for 300mm tools, much more has been done to minimize the time of test wafer processing.

    Thank you very much for bringing this up.

    Best, Alexander

    Like

    Reply

    1. Hello Alexander,
      thank you for reading my blog and writing your comment.
      I fully agree with your comments about 300mm automation capability enables better ways of dealing with test wafers. I will discuss some of these aspects in part3.

      I also think, I know quiet a few 200mm FAB managers, which kind of would argue, that 300mm has higher OEE requirements compared to 200mm. They all want for good reasons high OEE numbers – one way to get there is of course automation !

      Thanks again
      Thomas

      Like

      Reply

Leave a comment