second guessing is hard on one's sleep
May. 23rd, 2008 12:20 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
kareinaYesterday I had time on the microprobe. This is a task that I enjoy--it is fun to be able to look at the individual crystals in my sample in such close detail. In theory, I should have completed this part of my project before the end of last year. However, when I did my previous work I analyzed spots in lines across the garnets, because they are known to be zoned--with a different composition in the middle than at the edges, but we just probed random spots on the other minerals, since they aren't as likely to be zoned. This would normally be all well and fine, but the problems I've mentioned in other posts about how the dearth of calcium in my rocks means that I've got problems finding feldspar with a suitable composition for calculating the pressure of formation. It turns out that in other areas (like Scotland) which have similar metamorphic conditions to those in Tasmania feldspar often occurs as large, zoned, crystals. Some of my samples have large feldspars, and those samples, which had had just a few random spots per crystal analyzed, also show a range in composition. Therefore back to the probe for me, to put lines across some feldspars, to see if the change in composition is like the garnets, with a steady increase from core to rim of some elements, and a steady decrease in others, or if it is more random (less likely, given that minerals tend to grow from the centre out, and the changes in composition tend to arise because of changes in conditions while they are growing).
And, since I was putting samples back in, it was also time to see if I could solve some other problems I've been having with pressure-temperature calculations. You see, it works best if one takes compositions of minerals which are in equilibrium with one another when doing these calculations--and that is much easier for them to have achieved if they are quite near one another. Some of my early probe sessions wound up analyzing the different types of minerals on quite distant parts of the sample, so this was an opportunity to try again, only this time use wisdom gained from experience to select a better set of minerals.
Because some of the minerals I need are rather small in these samples, rather than pre-digitizing all of the points in advance, I decided just to mark the location of the large feldspars and do the actual point selection on the probe. Sure, I knew it would take longer that way, but I didn't really realize just how much longer.
By the time I'd looked at all of the marked feldspars (about four on each of two samples) and decided which three looked best, taken photos, printed them out, and set up the lines across them, set up lines across some large near-by garnets, and some individual spots on small, closer garnets and on micas (making notes on the photos recording all of those details), then went through each spot using the different view settings needed to adjust the focus/distance from the beam, and then went through all the points again using the "stage backlash" button (which moves the sample holder away from the point in a specific semi-circle, then brings it back to that spot from a specific direction) and re-saving the coordinate (this is because it is more accurate about returning to the location if it got to the location from the same direction it will approach it later, when it is running on automatic), it was already 22:00! (I started around 17:00) Then I looked at how long it thought it would take to analyse all of those spots, and realized that it wouldn't be done till mid-day the next day. Oops--can't be doing that--other people need the probe in the morning.
So then I spent another half an hour deleting one complete set of marked spots from each sample, bringing the estimated time for completion down to done at around 8:30 in the morning--a much more reasonable time! I told it to "run", and went on home, had a quick walk, some yoga, and crawled in to bed. However, I woke up often all night long, wondering if I'd picked the correct ones to delete, or if those spots would have given me better results than the ones I left in...
And, since I was putting samples back in, it was also time to see if I could solve some other problems I've been having with pressure-temperature calculations. You see, it works best if one takes compositions of minerals which are in equilibrium with one another when doing these calculations--and that is much easier for them to have achieved if they are quite near one another. Some of my early probe sessions wound up analyzing the different types of minerals on quite distant parts of the sample, so this was an opportunity to try again, only this time use wisdom gained from experience to select a better set of minerals.
Because some of the minerals I need are rather small in these samples, rather than pre-digitizing all of the points in advance, I decided just to mark the location of the large feldspars and do the actual point selection on the probe. Sure, I knew it would take longer that way, but I didn't really realize just how much longer.
By the time I'd looked at all of the marked feldspars (about four on each of two samples) and decided which three looked best, taken photos, printed them out, and set up the lines across them, set up lines across some large near-by garnets, and some individual spots on small, closer garnets and on micas (making notes on the photos recording all of those details), then went through each spot using the different view settings needed to adjust the focus/distance from the beam, and then went through all the points again using the "stage backlash" button (which moves the sample holder away from the point in a specific semi-circle, then brings it back to that spot from a specific direction) and re-saving the coordinate (this is because it is more accurate about returning to the location if it got to the location from the same direction it will approach it later, when it is running on automatic), it was already 22:00! (I started around 17:00) Then I looked at how long it thought it would take to analyse all of those spots, and realized that it wouldn't be done till mid-day the next day. Oops--can't be doing that--other people need the probe in the morning.
So then I spent another half an hour deleting one complete set of marked spots from each sample, bringing the estimated time for completion down to done at around 8:30 in the morning--a much more reasonable time! I told it to "run", and went on home, had a quick walk, some yoga, and crawled in to bed. However, I woke up often all night long, wondering if I'd picked the correct ones to delete, or if those spots would have given me better results than the ones I left in...
