downhole fractiontion fits method

eemixon

Long story short, I have some zircon grains for which it is NOT a fair assumption that they behave the same as my reference material, in terms of downhole fractionation. As a result, the DRS method of using the refmat fits is inappropriate. I'd like to instead do a downhole correction where a fit is identified for each sample (e.g. all grains from one sample) and applied that way. Is there a way to do this in the existing DRS? Thanks!

Bence

Hi eemixon ,

This sounds like it would require a python DRS, and is not possible (as far as I'm aware) with the current DRS.

Have you tried out the python U-Pb DRS? You would just need to add a few lines to that.

-Bence

Joe

eemixon

I can confirm that this isn't possible with the built-in U-Pb data reduction schemes and would require a custom python-based DRS. I'm not sure I agree with Bence that you'd only need to add a few lines to the python U-Pb example, but I'd say < 100 🙂

You can use data.compileDownhole(group, channel) to get a tuple (beam seconds, channel mean) for the specified group and channel. You'd then need to fit that and use the fit to correct the relevant parts of each signal. It is doable, but would require some python skill. If you're up to the challenge, I suggest you check out the python U-Pb example and the python course lesson on custom data reduction schemes.

All the best,

kai2

Hi,
If you fit a separate DHF model to each group, what do you gain? You still need a reference material (to compare your sample to) that has an identical downhole fractionation pattern, or not? Otherwise, how would you know what is DHF and what is zonation, for example?
What we usually do is normalize younger samples to younger stds; and older samples that have higher radiation damage (and ablate differently) to e.g., og1. That's two quick crunches just changing the primary reference, and works reasonably well.
Is there a more sophisticated method that transforms your primary DHF to different ablation characteristics, so that you can accurately correct og1 with 91500??

Cheers, kai