Overview

Here I use neutron single events to measure the distribution of theta_abs, or the angle between an incident photon and a resultaning photo-neutron. The distribution of uncorrelated neutrons from the SF of californium 252 is used to "divide out" the effects of detector geometry, efficiency, drifts, ect. For D2O, the result is compared to an MCNP simulation which was built to model as many aspects of the experiment as possible.

Simulation

I performed an MCNP simulation with a D2O target (axis length = 2"; dia. = 0.75") subject to a 10.5 MeV end-point bremsstrahlung beam. A mock-up of the entire neutron detector array is included in the simulation. Detector physics is modeled by applying a detection threshold in terms of light output (MeVee), which is equal to the typical MeVee produced by 0.5 MeV neutrons within the scintillator.

A Cf252 source was also simulated, allowing me to apply the exact same analysis technique to simulation and experimental data .

The plot below shows the relative distribution of neutron direction cosines w.r.t. the incident photon beam. These neutrons are not effected by scattering or detector geometry, since the direction cosine are taken right as neutrons are created.

Analysis

Ten detectors are used in the measurement, 5 beam left and 5 beam right. For each detector on beam left, there is a corresponding detector on beam right which covers the same range in theta_abs. Each left-right pair of detectors produces a peak in the theta_abs distribution, seen below.

During the analysis, each peak is condensed into a single point whose value on the y-axis is the mean of the peak, and is the weighted average, or "center of mass", along the x-axis.

Below is a step by step outline of the of analysis procedure. All rates are in terms of event per pulse.

To correct the target distribution for dead-time, multiply the rate in each detector by [math]\frac{1}{1-R_{G}}[/math], where [math]R_{G}[/math] is the gamma rate in the detector. This assumes the gamma rate is constant, which makes for a reasonable approximation since the gamma rate fluctuations are small. This assumption was tested by calculating the average (weighted by counts) of dead-time corrections measured over 10 minute intervals. The result differed by <1%.
Subtract from the target distribution, a cosmics-subtracted, dead-time corrected Al distribution. This will eliminate noise that results strictly from the beam.
1. scale the Al distribution in each detector individually so that its gamma rate is equal to that of the target distribution.
2. Scale the cosmics distribution with the same scaling factors.
3. Correct Al for dead-time in the same manor described in step 1.
4. Subtract the scaled cosmics distribution from the scaled and dead-time corrected Al distribution. The resulting distribution should now contain only beam-related noise.
Subtract cosmics distribution from deat-time corrected target distribution.
Divide by Cf252 distribution to get the relative rate over theta_abs, in each detector.

Note: Cosmic noise from Cf252 distribution was less than 0.1% effect.

non-normalized D2O and Cf252 distributions

Thet_abs distribution of neutron singles from D2O. No corrections applied.

Cf252 neutron distribution.

Normalized distributions

D2O dist. normalized to Cf252. No corrections applied to D2O distribution.

Normalized D2O distribution after applying dead-time-correction. Correction for beam right: 1.816, 1.068, 1.011, 1.01, 1.007. Correction for beam left: 1.592, 1.079, 1.015, 1.01, 1.012