Analyzing noisy spectral line shapes

Abstract

In many situations it is necessary to analyze spectral line shapes where the data contain significant amounts of noise or statistical fluctuations. Our visible spectroscopy measurements exploring ion diode physics on the PBFA II accelerator typically result in noisy spectra because the harsh environment limits the number of photons collected. The spectral line profiles include contributions from Doppler (Ti∼1–3 keV), Stark (ne∼1017 cm−3), and instrument broadening, as well as from Stark shifting (E∼3–10 MV/cm) and Zeeman splitting (B∼2–10 T). We extract a range of parameters (e.g., ion temperature from Doppler broadening) that fit the data by determining a range of fits that are consistent with the uncertainty due to the noise in the data. The range of fits is generated by a Monte Carlo technique. This method effectively distinguishes between actual spectral features and artifacts due to noise. It provides not only estimates of physical parameters, but also their uncertainties. We evaluated the technique over a range of signal-to-noise ratios and found that it works well for our application.