Hello,
I am a beginner with allpix2. Recently, I have been using allpix2 to create a somewhat complex detector device. I have uploaded both the 3D view and cross-sectional view of the device. While building this device model, I encountered a few issues and hope someone can help me with the answers:
In the detector model in allpix2, are the pixels predefined PIN devices? Do they already define the doping concentration?
Since the device model I want to create is divided into several regions with different geometric shapes (some are not boxes) and varying doping concentrations, where the insulator is SiO2, electrodes are aluminum, and the other doping layers are SiC, I initially divided it into 25 detectors to assemble this device. I would like to ask if this is reasonable. Will it affect the subsequent import of electric fields and doping concentrations from TCAD?
I also tried to use GDML to establish the shape of this device, but I found that I could only set it as a passive volume and could not set it as an active volume. I am unsure how I can realize such a complex device in allpix2.
Thank you very much!
I am not a big expert on Allpix2 but i can try to give my 2 cents on your questions.
I do not think they are predefined PIN devices, i should better read the documentation on that note, and i am quite sure they do not have a predefined doping concentration.
From my general understanding the pixels in allpix2 are kind of geometrical boxes without much predefined and hardwired parameters such as doping. If you want to simulate a specific detector design in which the pixels are squared and have a certain doping and composition, what you then do is to calculate (on an external software) the electric field and the weighting field from a more accurate representation of your device in which you can set doping, maybe more accurate geometrical distinctions of the p and n doped region the substrate and so on…
Once this is done you can import and map to the allpix detector geometry these fields and this will effectively make the detector simulated in allpix behave as you designed. The pixels in the detector of allpix will not have knowledge of the exact geometry of your picture here above, but they will “behave” as the pixel geometry you showed here thanks to the information contained in the electric and weighting field.
This then correlates to also the remaining questions you asked, but once again this is just from my general understanding of allpix2 concept, i might be wrong or have missed some important details, if so feel free to correct me.
@luca_tero almost said it all and correctly. I’d like to add that in most simulations we only define a block of silicon for Geant4 to determine for the interaction e.g. with charged particles. Of course you could be more precise by adding layers of SiO2 or even metal layers, but usually those are negligibly thin when it comes to the interaction with incoming particles. They do certainly make a difference when it comes to the electrical behaviour of the sensor, but this is something that, as @luca_tero mentioned, cannot be simulated in Allpix Squared, but you’d have to use e.g. TCAD for this or use approximations (which in the presence of a gain layer will not be accurate).
Then, the usual workflow is to use the electric field simulated via TCAD and project it onto the “block of silicon” defined via the model files, see examples here. This should give you the highest precision possible.
