I'm just wondering (though they are at different sin thetas), if the mechanism that produces a spurious peak on my TAP crystal only on Ti, is related to the mechanism producing the spurious peak on your TAP only for Mg.
Maybe do a similar scan on Mg and see if you can reproduce the spurious peak I see near F Ka.
I ran a scan on synthetic MgO (20 kV, 300 nA) and only see:
2nd-order peaks of Mg: K-beta, satellite K lines, and K-alpha (attached MgO-20kV-300nA-TAP.jpg).
The range was 60,000 to 80,000 in units of 100,000sin(theta), with 1000 steps and dwell-time of 1 s per step.
I ran some scans on synthetic rutile (from MTI Corp.) at the same conditions over the same range, and just see:
higher-order peaks of Ti (attached TiO2-20kV-300nA-TAP.jpg).
I then ran a set of 10 accumulations on synthetic rutile from 69,000 to 73,000 with 200 steps and dwell time of 2 s per step, at 20 kV and 300 nA.
The attached results have been smoothed (3-point smoothing), and the file is: TiO2-20kV-300nA-TAP-detail-10accumul.jpg
Conclusions: I don't see any anomalous features around 71,000 in these TAP-crystal spectra of MgO or TiO2.
(All of this was in PeakSight software - I was in a bit of a rush).
Cheers, Andrew