Author: omannion

I ran the same simulation as a few posts ago using SRIM but now I changed the angle of incidence of the alpha particle. I then looked at the energy loss plots for each of those simulations and recorded the energy loss per angstrom penetration for each individual substance. I then plotted the energy loss per angstrom penetration vs angle incident for each substance. This data clearly fits a sec(theta) equation which is plotted below in blue. The blue line is actually the equation y=E_(theta=zero)*sec(theta), where  E_(theta=zero) is the value of the energy loss per angstrom penetration for angle of incidence equal to zero.

Below is the graph for Magnesium fluoride which had an E_(theta=zero)=20

 

Below is the graph for Aluminium which had an E_(theta=zero)=16

 

Below is the graph for Copper which had an E_(theta=zero)= 38

 

The conclusion is that alpha particles lose more energy per angstrom penetrated  the closer the angle gets to being perpendicular to the surface. This is being caused by the fact that as the angle increases the distance these particles  must travel to penetrate the same distance is increased, when the distance is increased the number of particles in the substance the alpha particle interacts with increase, so it loses more energy along the way.

This is an example of the path of the particle, on the left is theta =15 degrees on the right theta =45 degrees. As you can see the distance traveled to get through the aluminum is larger for the 45 degree angle compared to the 15 degree angle.

 

One observation I could not explain was as the angle increased the “smoothness” of the energy loss plots went down substantially. Below on the left is a picture with theta=0 and on the right is theta=85.

SRIM plots of energy loss per angstrom vs depth in layer and plot of positing of particles in  vs depth in layer. I asked the author of the program if you can change the units of the axis but unfortunately you cannot.

(UPDATE) This is the energy loss plot with the magnesium fluoride first then the aluminum.

 

I have a lab today (3/31) at 1:25-4:25 so I may be able to explain this if the lab does not run late.

This is the data for the position data for the denoised 2991 run and the “noised” 2991. The positions are very similar except for in the x direction were there is a difference of almost 2 mm. The error for the denoised file was also about twice as larger which makes sense because of there being about half the events.

	Denoised		Noised
# Events	3216		6632
	Position (mm)	Error	Position(mm)	Error
Z	-24.51	0.048	-23.67	0.025
Y	9.021	0.057	9.05	0.033
X	260.77	0.2	258.96	0.101

This is the data for the +z and the -z. I understand that the +z and the -z should have roughly the same # of events, which they do. I do not understand though why the number of events in the +z and  the -z do not add up to the number of events when we just take the absolute value of z is > 187.  Also there was only a small difference in the +z vs the -z for the de-noised file. I think what I should do next is possibly run this again for many different runs and see is there a significant trend. For example is it the X position always larger in the -z cone?

	Denoised	Denoised	Differences	Noised	Noised	Differences
	“+z”	“-z”		“+z”	“-z”
# Events	3300	3256	44	6052	6019	33
	Position
Z	-24.54	-24.42	0.12	-26.34	-27.12	-0.78
Y	8.97	8.95	0.02	11.31	10.6	0.71
X	260.76	261.02	-0.26	269.62	269.57	0.05

This was the code for +z and similar for the -z just flipping the inequality. This checks if the z has a magnitude larger than 187, then it checks if that value is positive and gives back only that value.

if abs(clu.fZ) > 187:
if (clu.fZ) > 0:return []

I have checked if there was a difference when the code only selects positive z position and negative z positions. There was and the source positions found where

	AbsoluteValue(z)	(+z only)	(-z only)	Differences From +z and Abs(z)
# of Events	6632	6052	Fewer than 20 Events
X	258.966	269.627		10.661
Y	9.05981	11.3165		2.25669
Z	-23.6713	-26.3416		2.6703

There was only a slight difference.

I will meet with Sereres this week to see if there is something else I can do because the author of Numpy is taking a long time in-between email exchanges and the soerp package seems to not be going anywhere.

I was not able to make much progress last week because of two exams and another one today and a lab report due.

I did get to fix the Numbers problem with the ‘ad’ module. Now there seems to be a problem with the soerp code when I am running it on the noric computers which I never encountered on my home computer. When doing anything relating to soerp I get an error message, for instance this code which just assigns a value 4 and variance of 2 to x:

x=soerp.N(4,2)
Traceback (most recent call last):
File “<stdin>”, line 1, in <module>
File “/afs/slac.stanford.edu/u/xo/mannion9/pipeline/soerp/__init__.py”, line 872, in N
return uv(rv=ss.norm(loc=mu, scale=sigma), tag=tag)
File “/afs/slac.stanford.edu/u/xo/mannion9/pipeline/soerp/__init__.py”, line 722, in __init__
mn = rv.mean()
AttributeError: ‘rv_frozen’ object has no attribute ‘mean’

I am trying to find what this means and how to fix it but I have not had much luck finding information online.  I emailed the author.

After speaking with the author for a few weeks the soerp error analysis module and its required modules are finally fixed and I have them ready to go into the compton telescope code. I just need to now put it into the code and run it.

Also earlier this week I looked at the X and Y of the source position and calculated the radius of the detector tube. This analysis is posted in the data analysis section.

-Owen