Baseline drift effect on the performance of neutron and gamma ray discrimination using frequency gradient analysis

Guofu Liu,Xiaoliang Luo,Jun Yang,Cunbao Lin,Qingqing Hu,Jinxian Peng

DOI: https://doi.org/10.1088/1674-1137/37/6/066201

2013-05-08

Abstract:Frequency gradient analysis (FGA) effectively discriminates neutrons and gamma rays by examining the frequency-domain features of the photomultiplier tube anode signal. This approach is insensitive to noise but is inevitably affected by the baseline drift, similar to other pulse shape discrimination methods. The baseline drift effect is attributed to the factors such as power line fluctuation, dark current, noise disturbances, hum, and pulse tail in front-end electronics. This effect needs to be elucidated and quantified before the baseline shift can be estimated and removed from the captured signal. Therefore, the effect of baseline shift on the discrimination performance of neutrons and gamma rays with organic scintillation detectors using FGA is investigated in this paper. The relationship between the baseline shift and discrimination parameters of FGA is derived and verified by an experimental system consisting of an americium-beryllium source, a BC501A liquid scintillator detector, and a 5 GSPS 8-bit oscilloscope. Both theoretical and experimental results show that the estimation of the baseline shift is necessary, and the removal of baseline drift from the pulse shapes can improve the discrimination performance of FGA.

Instrumentation and Detectors,Nuclear Experiment

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: **The influence of baseline drift on the discrimination performance between neutrons and gamma rays**. Specifically, the paper studied how baseline drift affects the discrimination effect between neutrons and gamma rays when using the Frequency Gradient Analysis (FGA) method, and proposed a data pre - processing method to reduce this influence.

### Problem Background

Frequency Gradient Analysis (FGA) is a method to effectively discriminate between neutrons and gamma rays by analyzing the frequency - domain characteristics of the anode signal of the photomultiplier tube. Although FGA is insensitive to noise, it is still affected by baseline drift, similar to other Pulse Shape Discrimination (PSD) methods. Baseline drift is caused by factors such as power supply fluctuations, dark current, noise interference, hum, and pulse tails in the front - end electronic devices. These factors will lead to baseline drift, thereby affecting the discrimination performance of FGA.

### Research Purposes

In order to improve the discrimination performance of FGA, researchers need to:

1. **Quantify the influence of baseline drift on FGA discrimination parameters**.
2. **Estimate and remove baseline drift** to improve the discrimination effect.

### Main Research Contents

1. **Theoretical Analysis**: Derive and verify the relationship between baseline drift and FGA discrimination parameters.
2. **Experimental Verification**: Verify the theoretical results through an experimental system (including an americium - beryllium source, a BC501A liquid scintillation detector, and a 5 GSample/s 8 - bit oscilloscope).
3. **Error Analysis**: Analyze the influence of baseline drift estimation error on discrimination parameters.

### Experimental Methods

Experimental data were collected from a radiation measurement system, which includes:
- **Americium - Beryllium Neutron Source**: Generate a mixed radiation field.
- **BC501A Organic Liquid Scintillation Detector**: Used to detect neutrons and gamma rays.
- **EMI 9807B Photomultiplier Tube**: Convert optical signals into electrical signals.
- **Tektronix Digital Phosphor Oscilloscope**: Used to collect and process pulse signals.

### Key Formulas

1. **Moving - Average Filtering**:
\[
y_1(m)=\frac{1}{q + 1}\sum_{k = 0}^{q}b_ky_0(m - k)
\]
where \(b_k = 0.5\), and after simplification:
\[
y_1(m)=\frac{y_0(m)+y_0(m - 1)}{2}
\]

2. **Baseline Drift Estimation**:
\[
d_L=\frac{1}{L}\sum_{k = M - L+ 1}^{M}y_1(k)
\]

3. **Normalized Pulse**:
\[
y_3(m)=\frac{y_2(m)}{\max[y_2(m)]-d_L}
\]

4. **Fourier Coefficient Calculation**:
\[
X(0)=\frac{1}{N}\sum_{n = 0}^{N - 1}x(n)
\]
\[
X(1)=\frac{1}{N}\sum_{n = 0}^{N - 1}\left[x(n)\cos\left(\frac{2\pi n}{N}\right)-ix(n)\sin\left(\frac{2\pi n}{N}\right)\right]
\]

5. **Discrimination Parameter Calculation**:
\[
k = X(0)-|X(1)|
\]

### Conclusions

Through theoretical analysis and experimental verification, the research found that:
- More accurate estimation of baseline drift can reduce the error of discrimination parameters, thereby improving the discrimination performance between neutrons and gamma rays.
- Increasing the data length \(L\) used for estimating baseline drift can reduce the estimation error and further improve the discrimination effect.

Future work will involve the research and verification of more effective baseline drift estimation algorithms to further optimize the performance of the FGA method.