Per default, AtRIS uses a spherical phantom corresponding to an ICRU sphere (15 cm radius) but made of water.

We did not want to use Tissue Equivalent Material (TEM) since it express the radiation hazard to humans in specific. Although the use of non spherical phantoms could be possible, since to AtRIS they are nothing more but a 2D histogram, the question of normalization becomes much more difficult. Let us consider how AtRIS works:

1. Read in “icru_response.txt”
2. For each particle (primary and secondary) do the following when it crosses from one sensitive detector (atmospheric shell) to another:
   1. Identify particle, check if relevant to dose rate calculations.
   2. If so:
      1. Retrieve from “icru_response” the efficiency corresponding $\mathcal{I}_R$ to this particle and energy $E$
      2. in the absorbed dose matrix:
         1. determine the altitude and primary energy index
         2. increase the corresponding value by $E\cdot\mathcal{I}_R$
      3. check if the particle has a non-zero radiation weighting factor $W_R$. if so:
         1. in the equivalent dose matrix $\mathcal{I}_A$ increase at proper location the existing value by $E\cdot\mathcal{I}_R\cdot W_R$
   3. Finish simulation
   4. Perform normalization:
      1. Divide each column of matrices $\mathcal{I}_R$ and $\mathcal{I}_A$ with the number of primary particles $n_j$ generated in the corresponding bin
      2. Multiply with $\pi\cdot A_{PHANTOM}$ to account for the geometric factor of the phantom
      3. Divide with $A_{SDIID}$ to scale the areas properly
      4. Divide with the mass of the phantom $m=\rho\frac{4}{3}\pi r^3$, to get energy/mass
      5. When performing the output, force Geant4 to use Greys. Thereby the energy/mass unit conversion is performed automatically.

If one desires to change the phantom, one would need to:

• Provide a new icru_response.txt file and replace the original
• Provide the new mass of the phantom
• Provide the new cross-section surface of the phantom (for a sphere simply $\pi r^2$).

• Particles relevant for dose rate calculations are: $\gamma,e^-,e^+,\mu^-,\mu^+,p^+,p^-,n,\bar{n},\pi^+,\pi^-,\pi^0,K^+,K^-$. This can be changed.
• The internal energy binning for the phantom is fixed:

  python

  10**np.linspace(-6,7,651)

• The header specifies the PDG codes of the particles. The default is:

# 22 1000020030 1000020040 1000010020 1000010030 11 -11 2212 -2212 321 -321 211 -211 13 -13 2112 -2112

• If you wish to remove or add particles to the dose rate calculation, please modify the header.
• When the default phantom is changed, you need to configure using the custom macro commands the mass and the crosssectional surface of the phantom in order for the normalization to be performed properly

You can create your own phantoms. The steps involved are:

• Specify the phantom gdml file. Take a look at the examples.
• Perform a geant4 simulation for all the relevant particles. The script phantom.py does this.
• Plotting can be done with icru_phantom.py
• The script XXX creates a new icru_response.txt
• To use the new phantom, build AtRIS and THEN replace the icru_response file. Proceed to add the relevant information to the macro or to your execution script. AtRIS will not recognize automatically that you have changed your phantom.
• If you have a phantom that does not have a geometric factor of $A\cdot\pi$, pass a false value for A which scales this value so that your geometric factor is accounted for.