Magboltz attachment coefficients at low electric field strength

Garfield++
1

I wasn’t sure where to post for help with Magboltz, so I’m hoping the ROOT newbie forum will be gentle!

I am a PhD student working on the NEWS-G dark matter experiment at Queen’s University, Canada. I am studying the effect of electronegative contaminants (primarily oxygen) on our signals. NEWS-G has used Magboltz often during simulation work for our design of gaseous detector, giving reliable predictions. We find it valuable.

However, we find a discrepancy when oxygen is considered, with the predicted attachment coefficient from Magboltz being significantly higher than we seem to observe. Some experiments last week confirmed that when we see attachment affecting about 20% of our ionization electrons, Magboltz gives an attachment coefficient so high that almost all of them should be attached.

Our detector design means that we operate at unusually low electric field strengths, with a drift field of only ~1 V/cm and ~0.005 Td, and wonder whether this might be the central issue, with limited measurements of the oxygen attachment coefficient published by anyone in this regime. For instance, looking at Figure 5 left in the 2020 paper coauthored by Steve Biagi, Electron transport in gaseous detectors with a Python-based Monte Carlo simulation code, I notice that the lowest field strength data plotted (Jeon and Nakamura, 1998) is 3000 times larger, although I also notice that they measured down to 0.03 Td in Jeon and Nakamura, 1998. DOI: 10.1016/j.cpc.2020.107357.

I wondered whether anyone has any insights or suggestions that might throw some light on this apparent discrepancy. Thank you! (I tried emailing Prof Biagi, however the CERN email server rejected it.)

2

Dear @jondclarke

The addition of Oxygen leads to an increased electron attachment through the 3-body process. This process depends on the concentrations of the spectator gases. The default scaling for the cross section of this process in Magboltz is for 100% O2. You however have to scale this to the percentage of O2 or other gases that can be involved in the de-excitation. You basically have to adjust manually this parameter in the Magboltz.f file and recompile it:

C THREE BODY ATTACHMENT
C ***************************************************************
C ENTER HERE SCALING FACTOR FOR THREE BODY ATTACHMENT IN MIXTURES:
C FOR NORMAL SCALING T3B=1.0
T3B=1.0
C SCALING FACTOR NORMALLY PROPORTIONAL TO OXYGEN FRACTION
C IN RARE GAS MIXTURES
C
C***********************************************************

Recently the Pisa group did a lot of work and investigations into it, to set the attachment rates correct in the gas mixture they use for the MEG-II drift chamber. You can find more information in Fabrizio Cei’s presentation at the DRD1 Collaboration meeting of October 2025 [1]. They recently published a paper detailing their problem and their solution [2].

Kind regards
Piet

[1] https://events.camk.edu.pl/event/124/contributions/1198/attachments/815/2192/O2%20in%20Drif%20Chambers%20and%20Garfield.pdf
[2] The use of O2 in gas mixtures for drift chambers - IOPscience

1 Like
3

Excellent! Thank you, Piet. This was exactly the kind of missing ingredient I was hoping for.