Pouch Sealing and Peel Testing Process Improvement

We have a lower specification of 1.0lbf/in for pouch peel strength, and our sample size is 4 pouches per lot. Our process is not always stable and has inherent large variation, meaning that we do get values that range from around 1.0lbf/in to 10+ lbf/in. The Cpk that I calculated from the peel strength values is only 0.8.
Can someone give me advises on how to segregate the Pouch Sealing process variation from the overall of Pouch Seal + Peel strength testing? The strength of the seal cannot be measured without peeling it, and they are 2 separate processes with variations of its own coming from the equipment used, operators, methods, etc. How do I isolate the process variations from these 2 processes so that we can reduce the variation?

Thank you

biscotti said:

We have a lower specification of 1.0lbf/in for pouch peel strength, and our sample size is 4 pouches per lot. Our process is not always stable and has inherent large variation, meaning that we do get values that range from around 1.0lbf/in to 10+ lbf/in. The Cpk that I calculated from the peel strength values is only 0.8.
Can someone give me advises on how to segregate the Pouch Sealing process variation from the overall of Pouch Seal + Peel strength testing? The strength of the seal cannot be measured without peeling it, and they are 2 separate processes with variations of its own coming from the equipment used, operators, methods, etc. How do I isolate the process variations from these 2 processes so that we can reduce the variation?

Thank you

Click to expand...

Maybe I'm misunderstanding something, but if you have only a minimum limit, what are you concerned about?

biscotti said:

We have a lower specification of 1.0lbf/in for pouch peel strength, and our sample size is 4 pouches per lot. Our process is not always stable and has inherent large variation, meaning that we do get values that range from around 1.0lbf/in to 10+ lbf/in. The Cpk that I calculated from the peel strength values is only 0.8.
Can someone give me advises on how to segregate the Pouch Sealing process variation from the overall of Pouch Seal + Peel strength testing? The strength of the seal cannot be measured without peeling it, and they are 2 separate processes with variations of its own coming from the equipment used, operators, methods, etc. How do I isolate the process variations from these 2 processes so that we can reduce the variation?

Thank you

Click to expand...

Is the seal large enough to take multiple readings? If I understand you correctly you want to do an MSA on the seal strength measurement system but of course this is a destructive test. Depending on if you can take multiple measurements on the same seal (cut the pouch into several samples) determines how you design the study.

Jim Wynne said:

Maybe I'm misunderstanding something, but if you have only a minimum limit, what are you concerned about?

Click to expand...

It is obvious from the control chart and Cpk study that the variations are large (even though I can't be sure if they are mostly coming from Sealing or Peeling). Since the peel testing is done on a sampling basis, I cannot rule out the chances that the peel strength is less than the lower limit within the production lot.

I can cut the same pouch into multiple strips for measurement. However, that would be a MSA on the peeling process. If the MSA shows that the peel test is repeatable and reproducable, would that be sufficient to say that the peel test is stable and that the Sealing process is the source for special causes?

biscotti said:

It is obvious from the control chart and Cpk study that the variations are large (even though I can't be sure if they are mostly coming from Sealing or Peeling). Since the peel testing is done on a sampling basis, I cannot rule out the chances that the peel strength is less than the lower limit within the production lot.

I can cut the same pouch into multiple strips for measurement. However, that would be a MSA on the peeling process. If the MSA shows that the peel test is repeatable and reproducable, would that be sufficient to say that the peel test is stable and that the Sealing process is the source for special causes?

Click to expand...

Again, if I understand you correctly, you don't know whether or not there are any special causes. You might have a perfectly stable process. I agree that you should first qualify your measurement system, and for information on that I commend you to Miner's excellent blog post R&R for Non-Replicable Measurements.

For the time being at least, I recommend that you put CpK out of your mind. Once you know something about your measurment system, start collecting data and then just chart it and look at what it's doing. If you're not sure whether or not the measurements are dipping below the minimum, or are likely to, you probably don't have enough data.

biscotti said:

If the MSA shows that the peel test is repeatable and reproducable, would that be sufficient to say that the peel test is stable and that the Sealing process is the source for special causes?

Click to expand...

If you do your MSA correctly, you ca show the test method is stable and repeatable and reproducible, then the apparent variation is assumed to be process variation. Typically peel tests are repeatable if the fixturing is build correctly (no slipping of the sample). Therefore if you see very variable data, you usually conclude you need tighter control over your process.

Is this a Tyvek pouch (on one side) and a LDPE/PET or LDPE/Nylon film on the other side? In my experience with medical device Tyvek pouches tested per ASTM F88, where 1.0 lbf/in is the lower spec limit, the upper values should not typically exceed 3.0 lbf/in. Much above that and you start getting Tyvek delamination and the sealer heat setting can cause transparency which means Tyvek weakening - not good. The range of values you are getting is remarkably wide. It makes me think the heat sealer is not performing very consistently. Is it a bar sealer? impulse sealer? Does it pass Installation Qualification?

If I understand you correctly, you wish to validate the test method to ensure the variation in test results is from the sealing process, and not the test method. This is a very good thing to do and a requirement of ISO 11607-2. The distance between the clamp and the seal can impact the test result due to the elasticity of the material. Depending on the packaging material, humidity and temp can also impact the results. This is why ISO 11607-2 requires conditioning. Also, the seal can fail by several different modalities. If the modalities vary, the results will be very different. There are also different ways to support the material during the test. If this varies, the results may vary. Rate of separation is also important.

I suggest taking a look at ASTM F 88. It contains a section on a round-robin interlaboratory study conducted to validate the test. This validation does not perclude you from validating the test in your own laboratory. However, the methodology described can be used to validate your test.

Sorry to be so late to this party. I agree with Sam Lazzara -- 1.0 - 10.0 lb/inch is a much wider range than I have seen in working with a number of different companies.

Have you done a histogram of the seal strengths? You might find a bi-modal distribution, for example, instead of a normal distribution with wide spread. If you have two pouch sealers, perhaps one of the sealers is creating much stronger seals than the other sealer.

Doing R&R on the measurement system is good idea. Also, depending on the test-up, the sample may have a "tail" as the seal is pulled open. Some people test with the tail free, and some people clamp the tail. I have seen this make a big different in the peel strength readings.

Just wanted to offer a couple more ideas.

Hi I have a question regarding the same issue. We have chosen our sample size to be 15 per operator for the peel strength test. The same specification 1.0 lb/in and the upper limit is typically lower than 3 lb/in. We have 2 operators and the Cpk value is greater than 1 for both of them. So I understand that the process in under control. But how do we justify the number of samples that we took were right? The auditor is asking for a justification. I am not sure how I should do this. I looked at determining sample size. It was very complicated. Please help me with this.