[Gate-users] Help on simulating x-ray tube

[Paul Betteridge]

Dear Emmanuel,

> Please, I  am trying to simulate real x-ray tube for our CTscanner. I 
> have to try several times to make the electron source to hit the 
> anode(tungsten) to get photons but the electrons are not hitting the 
> anode for me to get photons. I do not know if what am doing is wrong or 
> am missing something. 

I can't tell you exactly what is wrong with your simulation, but I have 
used Gate for similar tasks before, so perhaps I can give some guidance 
that will let you sort out your own problem.

1) Generating x-rays by hitting a target with an electron beam is 
inefficient, both in the real world and in simulation. Of course, for 
some questions, it is the only thing you can do. But if you want to 
study what happens to the x-rays afterwards, you will want to generate 
them directly, perhaps using a spectrum you have determined with a 
simulation like that you are attempting.

2) Have you tried looking at effect of your simulation - I mean looking 
at the particle tracks using (after having set up some kind of 
visualisation) commands like these?

/vis/scene/add/trajectories

/tracking/storeTrajectory 1

# keep up to 100 events/tracks for display

/vis/scene/endOfEventAction accumulate 100

...

/gate/application/start


3) You have included a lot of detail in your simulation at this early 
stage. I would be inclined to simplify it to just the electron source 
and the anode.

4) One technique which I have found useful is the ability you have in 
simulation to make changes like setting components of the system to have 
the material "Vacuum". If you don't want to delete the elements of your 
model, you can set their material that way to remove their effect.

It may still be difficult to properly visualise the components though, 
so bear that in mind.

5) I'm also a little concerned about both the geometry of your electron 
source - I usually use a much simpler definition intended to send a beam 
of electrons along the z axis - and its placement relative to the 
target/anode. Try visualising the system, and look at it from different 
directions. Using the command:

/gate/source/electron_beam/visualize 1 yellow 10 mm

will help you see where the electrons start from. (In the version of 
Gate I am using, I have to place that after /gate/run/initialize; 
otherwise it crashes the program.)

Should the source be right next to the anode?

For your information, my definition of an electron source looks like this:

/control/verbose

/gate/source/addSource electron gps
/gate/source/electron/gps/particle e-

/gate/source/electron/gps/energytype Mono
/gate/source/electron/gps/energy 120 keV

/gate/source/electron/gps/type Point
/gate/source/electron/gps/centre 0 0 0 mm

/gate/source/electron/gps/ang/type iso
/gate/source/electron/gps/ang/mintheta 180 deg
/gate/source/electron/gps/ang/maxtheta 180 deg

6) I do recommend picking a good initial direction for your primary 
particles, and I think the z axis is a good choice - you can then place 
things along it easily. That works for me, because I am using a 
transmissive target usually, so you may need to make another choice. But 
it may help to try this arrangement. (It helps with visualisation too, 
because you can view along, or close to, each of the x, y, z axes.)

7) You obviously need to have appropriate physics enabled. 
Bremsstrahlung is necessary. ElectronIonisation adds the impressive 
detail of characteristic lines (probably without actually changing your 
results much.)

Hopefully one or more of these suggestions will be useful.

Best wishes,

Paul

-- 
Paul Betteridge
Adaptix Imaging