I'm doing a Monte Carlo simulation of a transmission type x-ray tube. A electron beam hits a thin tungsten target, causing X-rays to be emitted. These X-rays then transmit through the target and a beryllium window, hitting the scoring plane. The spectral distribution of these photons in the simulation look like the following:
So, both K-$\alpha$ and K-$\beta$ edges are accounted for. The general Bremsstrahlung curve seemed ok, but the attenuation was different than I expected. My knee-jerk reaction was that the simulation had failed in some way as the attenuation changes abruptly at k-$\beta_2$. A look at literature showed me that I was wrong and this is considered normal.
I understand that the attenuation coefficient should go up at the k-edges, as more photons will interact with the medium. What I don't understand is why this only happens at k-$\beta_2$? Why don't the photons interact with the three other k-edges, stepping the spectrum down at each of them? Looking at for example the NIST attenuation dataset for tungsten it's clear that there's only a single leap in attenuation for the four edges at these measurement accuracies. Why?