Much less is known about neutron structure than that of the proton due to the
absence of free neutron targets. Neutron information is usually extracted from
data on nuclear targets such as deuterium, requiring corrections for nuclear
binding and nucleon off-shell effects. These corrections are model dependent
and have significant uncertainties, especially for large values of the Bjorken
scaling variable x. The Barely Off-shell Nucleon Structure (BONuS) experiment
at Jefferson Lab measured the inelastic electron deuteron scattering cross
section, tagging spectator protons in coincidence with the scattered electrons.
This method reduces nuclear binding uncertainties significantly and has allowed
for the first time a (nearly) model independent extraction of the neutron
structure function. A novel compact radial time projection chamber was built to
detect protons with momentum between 70 and 150 MeV/c. For the extraction of
the free neutron structure function $F_{2n}$, spectator protons at backward
angle and with momenta below 100 MeV/c were selected, ensuring that the
scattering took place on a nearly free neutron. The scattered electrons were
detected with Jefferson Lab's CLAS spectrometer. The extracted neutron
structure function $F_{2n}$ and its ratio to the deuteron structure function
$F_{2d}$ are presented in both the resonance and deep inelastic regions. The
dependence of the cross section on the spectator proton momentum and angle is
investigated, and tests of the spectator mechanism for different kinematics are
performed. Our data set can be used to study neutron resonance excitations,
test quark hadron duality in the neutron, develop more precise parametrizations
of structure functions, as well as investigate binding effects (including
possible mechanisms for the nuclear EMC effect) and provide a first glimpse of
the asymptotic behavior of d/u as x goes to 1.
