2/4/2003

Given one Energy Dispersive Spectrum (EDS) peak width (principal peak),

for a Si(Li) detector

the widths of the other peaks are

                  sqrt[2500(Ei - Ep) + (2355 Sp)^2]

      Si   =      ---------------------------------

                              2.355

Si    standard deviation of peak i, keV

Sp    standard deviation of the principal peak, keV

Ei    energy of peak i, keV (I think).

Ep    energy of principal peak

from p. 386, eq. 7.19

Goldstein, J.I., et. al.

SCANNING ELECTRON MICROSCOPY AND X-RAY MICROANALYSIS, 2nd Ed.

Plenum, N.Y. 1992

Original work:

Fiori, C.E., Newbury, D.E., (1978), SEM 1:401.

------------------------------------------------------------------

2/4/2003    The above may have dimension problems.  Dale will check it out.

In the mean time, a simplified equation from the original article is:

FWHMx^2 = 2.5(Ex - Eref) + FWHMref^2

Where all units are eV.

Mn          usually principal peak        5895 eV     151 ev

Al          calculated from formula       1498 eV     107 ev

maclispix(45): (line-fwhm 1498 5895 151)

108.66692

close enough.

---------------------------------------------------------------------

2/28/2003   Trouble: 

Lispix(38): (xray-line-fwhm 1600 6392 96)

#c(0.0 52.573757)

Trouble point:   2.5(Ex - Eref) + FWHMref^2 =< 0

            or:   2.5(Ex - Er) + Wr^2 <= 0

                  2.5(Ex - Er) <= -Wr^2

                  2.5(Er - Ex) >= Wr^2

                  Edif >= Wr^2/2.5

                  4800 * 2.5 = 12000 - sqrt = 109.

Ex = 1600

Er = 6392

sqrt (2.5 (Er - Ex)) = 109.4

Lispix(47): (xray-line-fwhm  1600 6392 120)

49.193497

Lispix(48): (xray-line-fwhm  1600 6392 110)

10.954452

???

Ex - Eref >= FWHMref^2 / 2.5

Also, the Si peak range looks really broad - small side peak adds to the width.  That ain't right.