I had another new revelation this morning:
I think I have found the probable reason why Jones thinks he found elemental Al in the MEK-chip - and is honest about it!
To recap:
The MEK-Chip, presented in Fig. 12-18, was a red-gray chip that most likey was chipped off the WTC perimeter columns, as the red layer is almost certainly Tnemec primer, according to the XEDS-Spectrum of the entire chip in Fig. 14.
Fig. 16-18 are three SEM-EDS-Spectra from different, albeit unspecified spots on that chip, taken after it had been soaked in MEK. Strangely, while Fig 14 was recorded at 20 keV, as most other spectra in the paper, and
Fig 16, was, too, Jones did
Fig 17 at only 10 keV, and
Fig. 18 at 15 keV. No reason is given.
Of particular interes is Fig 17 (10 keV), since it shows mainly a large Al-peak, much smaller C- and O-peaks, and tiny traces of Mg, Fe and Si. A quantitative interpretation of that result would indeed show too many Al-atoms and too few O-atoms, such that a significant proportion of the Al would have to be elemental. There is no elemental Al in Tnemec, and it is pretty unlikely to find small particles of Al contaminating the surface of such a chip.
So I have already suggested that perhaps this spectrum somehow picks up signal mostly from the sample holder, which in such electron microscopes is often an aluminium stub. Either the electron beam was pointed at the sample holder outright, or this is somehow a "reflection" of stub-signal, reflecting from the "steep" (almost perpendicular to the holder surface) edge of the chip on the lower left as shown in Fig 15, the EDS-map of a part of that chip after soaking:
[qimg]http://i1088.photobucket.com/albums/i328/MikeAlfaromeo/ActiveThermiticMaterial/th_ActiveThermiticMaterial_Fig15_a-d.jpg[/qimg]
But I have wondered: Jones couldn't be that stupid!? He'd have to
see and
know somehow that his beam is pointed at an inopportune spot - he'd have to be aware of the deceptive nature, not?
But then it struck me:
Fig 17 was recorded at 10 keV
Fig 15 was recorded at 10 keV! (the EDS map)
And this is what I think happened here:
When you record an EDS-map, your SEM machine basically records many individual EDS-spectra from many spots - of which we can think of as "pixels". So if the SEM scans, say, 200 x 150 "pixels", it actually records 30,000 EDS-spectra for 30,000 spots - that goes into the raw data. From that raw data, software computes the colored map images.
As we can see in Fig. 15, the raw data included raw data, i.e. individual EDS-spectra, for spots outside of the chip, many of which are shown in purple as "contain Al". Fewer contain high O or C, fewer still high Fe, and
none high Si.
I think that Jones did not pick a "region of high aluminum concentration on the MEK-soaked red chip" to run a seperate EDS spectrum on when he obtained Fig. 17;
rather; I think he just dug into the raw data from the EDS mapping, grabbed the individual spectrum with the highest Al-count or -proportion, and that's what he presented. I think it is well possible that he never checked where on the EDS-map that spectrum was from, and never became aware he had picked one from outside the chip!
(ETA: Two minor questions open in my mind:
a) Why doesn't the Al-map show Al
everywhere outside of the chip? -> Perhaps the equipment, running in standard configuration, is smart enough to filter out
most of the signal that's not from specimens - but obviously it records
some of such signal - it's clearly there in the raw data, not very obtrusive, and available to be picked
b) Why 15 keV for Fig. 18? That spectrum also shows P and Cl, which otherwise don't appear anywhere else. I consider that an inconsequential quirk for the time being.)