Accelerated Material Testing

The only general rule of dumb is the Arrhenius Law. This works when the only variable is the temperature, and within temperature ranges which doesn't affect the behaviour of the components.

For any other types of degradation there are no rule of dumb.

To apply the Arrhenius Law, you need to know the constant which relates the reaction rate with temperature (as the Law is exponential, increasing the temperature you can get the same aging in less time)

Best regards

One factor (and I mean only one of the many possible factors) that we find useful is to VERY accurately weigh the material under test before and after exposure.

This will give some indication of potential absorbtion of the solvent/chemical you are handling, or at least surface reaction.

I would also suggest multiple samples to be examined at different time intervals. (Like 3 days, 30 days, 300 days) I am thinking that 72 hours of exposure for a pipe system that may be in place for years is rather short.

Other (less reliable) variables to consider is the material "hardness" before and after exposure (indicative of softening or embrittlement) "thickness" (indicating material absorbtion) and so on.

Also ask the material supplier for their opinion. I know the plastic companies publish extensive lists of "material compatabilities" and polyprop is relatively common and well understood.

Be careful that you have information relating the the GRADE and BRAND that you will be using. Often polyprop is "re-inforced" with various fillers ranging from fiberglass to talc which could also react with your solvents/fluids.

I am a materials chemist who runs reliability evaluations of materials.

1: Arrhenius beheavor is only one model, classic easy to calculate, but it is based upon a normal distribution. Nothing is normal in reliability evaluations.

2: Wiebell distributions is generally used for the calculation. This is not based upon a normal distribution.

3: Lognormal, and expendential distributions are also used.

First thing to do: Look at modern plastics, there is a chemical capadability section.

Now: Polypropylene (HDPP) High density PP should work well with the hydroxide and hypochlorite solutions.

Testing: 1: Weight gain is one metric.

Procedure: Cut samples into ~ 1"X1". Make 24 samples. Dry the samples 24 hours @ 50'C. Weight each sample to +/-1 mg. Place 8 samples into each of 3 round bottom flasks, attach a condenser (Make sure the ST24/40 flange will pass the samples through. If not cut smaller blocks. Into the flask add a known weight of the test solution. (Might be fun here as KOH will etch glass). Heat the solution to 100'C for 24 hours. Weigh each sample (rinsed and (pat dry and blow with N2). 24 hour 100'C soak is ~ equivelent to 500 hours @ 85'C. You are shooting for MTBF @ 20K hours. If you are going to do an arrhenius calc you need to run 3 temperatures to calculate.

I would suggest reading the literature from reliasoft.

Now, you can also contact the pipe molder and through them the OEM who will have this set of data for you.

I suspect you are runing a PTFE (Teflon) piping system. HDPP will respond to most chemistries like PTFE. HDPP is also a higher temperature material.

Question to the Material Chemist ( Guest) :

It is known that the rate of reaction increases as temperature increases...

Would increasing the concentrations of the reactants be also a way to simulate aging in an accelerated manner?

Concentration of reactants varies the rate of reaction, this is one of the reasons to increase at least one reactant concentration to displace the reaction (in reversible ones) to the right (reaction products) or the reason to try to extract products.

The problem, as with temperature is that has some limits, for example, you cannot foresee the behaviour of PP at 150F by heating it at 1500F. The same is with chemicals, one specific material can be immune to any chemical solution up to a specific concentration level.

Good info guys, thanks. Here's a little more info for you guys as I think some of the work required for testing might be overkill, you judge.

The chemical used is a foaming cleaner with a potassium hydroxide content of less than 12% and a sodium hypochlorite content of less than 2% per the MSDS. The system that the parts belong to is a mobile foaming unit that pumps the chemical in, which is pre-mixed with water, via an AODD or EDP pump. The chemical is pumped into a mixing chamber along with pressurized air and the output is a foam varying in cosistency based on the air pressure and turbulence induced.

We currently use either SS or Kynar fittings in these application. We are wanting to switch to PP for cost reasons.

I looked up the material information from the supplier - Eldon James, and based on their compatability chart, there is no issue to be concerned with. Also, they did not list any addatives to the material.

The only stresses induced on the part would be internal pressures ranging from 0-55psi. There are no mechanical stresses to consider such as bending.

Just An Engineer - I have done as you suggested and weighed the parts before and after testing(the test suggested to me). There appears to be virtually zero absorbtion. As far as the time intervals, it seems from what I read, that Arrhenius behavior model would indicate that the increased temp would accelerate these times with shorter testing. Although I have not calculated,(not sure I could even figure out the calc), at what temps and timeframes I need to test at, So I still do not know what 72 hours @ 120 degrees F equates to in the field. I also tested hardness and there was very little change, no greater than a common error percentage at least.

Guest - you stated that I could contact the OEM regarding the test you laid out. I do not think I have the tools to conduct such a test and couldn't justify the purchase for this one time use. What information exactly should I be requesting? Is there a specific name for the report?

Either way, after speaking with the engineer on the project, compatability is our biggest concern and the PP should be perfectly fine for the application.