Different rho diags in cross profile of 3D and RZ

[ancientflow]

Hi, I have another question. For the same physical model, I found that there are significant differences in the 3D and RZ diagnoses. The simulation model is as follows: gas is injected from a ring-shaped area at zMIN plane with a Maxwell distribution velocity. Generally, in the grid closest to the zmin plane, the density should be the highest, which is consistent with the RZ diagnosis results. However, in the 3D diagnosis results, the density of the grid closest to the zmin plane is lower than that of the layer above it, which is inconsistent with the RZ diagnosis results and physical laws. From previous issues, I learned that the diagnosis results are processed through interpolation, but I think interpolation should not lead to this result. Therefore, what is the cause of this phenomenon? Will it have an impact on other modules (such as DSMC, charge deposition, etc.)? The different values of density are due to the varying times of output diagnosis.The first image is the 3D diagnostic result. Furthermore, my WarpX code was downloaded in September 2025.

[ancientflow]

Furthermore, I uniformly injected gas into a cubic area and set do_not_gather = 0 and charge = 1. At this point, the density should be almost constant, merely due to statistical fluctuations. However, I found that the density on the six boundary surfaces was significantly lower than that inside, and the density on the 12 edges of the cube was lower than that on the boundary surfaces, while the density at the six vertices was even lower. Have any other people encountered this situation? I output the diagnostic data in the openpmd (bp5) format, using the software Paraview.

[RemiLehe]

Thanks for raising this issue. Could you post the input scripts that you were using for the 3D and RZ simulations?

[ancientflow]

Hi, I have found the reason. I used the PEC boundary condition for field and absorbing boundary condition for particles, so the charge density at the boundary was set to zero. After interpolation, the result above was obtained. The input scripts for the second comment is here.
3d_test.txt