Thursday, August 21, 2008

Accelerated Aging - Getting Bad Data Even Faster, Part 3

UV light is a very common culprit in degradation of plastics as these photons have enough energy to cause problems if they are absorbed by the polymer. E = h c/l so shorter wavelengths have higher energy. So the shorter the wavelength, the better, right?

Unfortunately this is also not correct for two reasons. First, the shorter wavelengths become increasingly rarer on the surface of the earth thanks to all the ozone up high. So going with lots of really short wavelength is unrealistic. But secondly, different plastics react differently to the spectra. A plot of degradation rate vs. wavelength will show a peak and it is seldom at the shortest wavelengths. This plot is known as the activation spectra (see Andrady's article for a good review) and is key to running good tests. Polypropylene for example, has a peak at about 350 nm, much higher than other plastics.

UPDATE: The formula for energy has been corrected. I was so excited to get the HTML right that I ignored the content.


Ψ*Ψ said...

Sounds useful! Too bad my lab doesn't have the setup for it.

David Eaton said...

Getting bad data even faster!

I nearly ripped a stitch when I read that. I have been struggling against similar ideas lately. Two things never seem to fail when a crisis arises at work: 1. The fastest data collection method wins, relevance be damned, and 2. Outliers will be given more weight than the bulk of the results.

I work with a lot of engineers. I love their abilities, their creativity, and their tenacity. But some of them have absolutely no instinct or training when it comes to doing real experiments.

Anonymous said...

I think you mean that E = hv rather than lv since lambda and frequency are inversely related.