Last updated : 27/Jan/2006
[Back]

Confirmation : the fitting method is changed. This time only Ti Kα1 mean is fitted. Other means are fixed by referential X-rays energy differences from Ti Kα1 mean.

Remark :
→ Ti Kβ fitting is not good. There is a little excess from the fit function. On the other hand Ni Kβ fitting is better than Ti one.

This difference is caused by the Cu Gaussians added in the fit function to conform an excess between Ni Kα and Kβ. Now let's have a think about this excess.

In the first cycle of E570 experiment, we regarded it as Cu X-ray from the Cu pipe for filling liquid helium. The Cu pipe was covered with pure Al foil to attenuate the X-ray. Because the Al foil's thickness is 0.3 mm, about 2 % of 8 keV X-ray can transmit this foil.

The fitted Cu X-ray intensity in the second cycle of E570 is estimated as about 2 % of Ni X-ray for SDD1 by comparing their heights (assumed same sigma and same background). This is too huge because SDD1 has no acceptance for the Cu pipe (SDDs were tilted 45 degrees to see the center of helium target cell) and the number of pion which hits Al foil is about one order smaller than Ni one. Even if assumed the cross section of K-shell ionization is same, this 2 % can't be realized.

I think it's better to take the excess as a response function of SDD which doesn't have any strange noise. As a semiconductor detector has some tail structure usually, SDD could have it.

It could be that the excess is caused by a pile-up of the signal of pion direct hit and X-ray. Analyzing the pion rate can remove it.