[NewCandle] Nick's new pics

[John Winterflood]

Hi Guys,

Thanks for the occasional excitement!  But I am sorry to say that it 
looks like this anomaly might be inexperience with EDS and SEM.  I hope 
I am wrong but it does seem that your Ag peak is uncomfortably close to 
being 2xAl peak.  Also your Al peak is so high it goes way off the plot, 
at which point it becomes natural that you should start to see the 
"harmonics" due to counting of the occasional but unavoidable double hits.

For those who are not familiar with EDS, I will describe very briefly 
how the process works (I know very little about the business, but have 
attended a course and dabbled with one a few times).  The high energy 
(usually 5-40kev) electrons, knock electrons which are in the elements 
of the sample out of their orbits and when an electron spirals back in 
to take its place, it emits an X-ray photon characteristic of the energy 
level of the orbit and thus the element.  These photons plough into the 
semiconducting detector crystal releasing electron-hole pairs (I assume) 
which are picked up as a short pulse by a very sensitive charge 
amplifier.  The energy of the photon is indicated by the size (ie number 
of electrons) of the pulse and this size indicates which bin on the 
spectrum this pulse should be accumulated into.  Well of course if you 
get two photons hitting the detector in a time that is too short for the 
pulse discriminator to discriminate, then they will be counted as a 
single double-size (ie double energy) pulse.  So two aluminum pulses (of 
1487ev) arriving within a very short time of each other will look like a 
single double energy (2974ev) pulse which may be indistinguishable from 
a silver pulse (2984ev).

There are several ways to discover if this is the problem or not.  The 
simplest is to halve the beam current and count for double the length of 
time.  Since the length of each pulse remains the same, but you have 
doubled the average time between pulses, you expect the number of pulse 
overlaps to go down by a factor of two.  Similarly you can do the 
opposite - double the beam current and count for half the time.  In 
these cases you expect to get much the same reading for the aluminum 
peak, but a *pseudo* silver peak that changes by factor of 2 each time!

Another method is to change the electron energy (ie change the KV 
setting) in an attempt to excite another (usually higher) orbital energy 
in the silver atoms that is not so close to being double the energy of a 
peak in the aluminum.  Silver has some lines above 20kev whereas 
aluminum has none - this would differentiate them immediately.

Actually quick look at some charts on the web suggest that silver should 
have a reasonably bright (56%) line at 3151 in addition to the (100%) 
line at 2984 that you are probably mis-identifying.  There is no 
indication in these plots that your silver has such a double peak.  So 
sorry to say, I think your Ag line is an artifact of the EDS method of 
pulse discrimination!

J.

Keith Nagel wrote:
> http://www.kpnconsulting.com/newcandle/download/nick2.htm
>
> Here are the new SEM and EDS pics. Enjoy!
>
> K.