Below we show three pictures of the Qweak Geant moller simulation with a Mini-torus Bfil factor of .95 up, .95 down, and .95 down without the beam pipe.
These pictures lead one to believe that electrons produced by the moller electrons slamming into the beam pipe are making their way through the secondary collimator. Below we show the energy versus radial distance of the moller electrons incident on detector 2 (just before the secondary collimator) with Mini-torus configurations as described above. There are certainly more electrons passing through the collimator when the mollers are bent down with the beam pipe than without the beam pipe.
When one calculates the rate of electrons (from the moller process in the target) with a Mini-torus Bfil Factor = .95 down but with and without the beam pipe one obtains the following plots:
This figure shows the total photon rates at each detector, as well as the rates due to each process, for three energy ranges. Note that at each detector the highest rate seems to be from high energy Moller electrons.

The following figure is a log-log plot of photon rate/MeV vs. photon energy at the GEM, middle and back chamber locations. Note that most of the rate at the GEM location seems to come from high energy photons.

To see separate GEM and Chambers log plots:

See details of the above plots
However, after making these plots I am not sure this difference in rate indicates that most of
the electrons entering det 4 are produced when the moller electrons (from the
target) hit the beam pipe.