Investigation of radioactivity-induced backgrounds in EXO-200

The search for neutrinoless double-beta decay $\left(0\nu \beta \beta \right)$ requires extremely low background and a good understanding of their sources and their influence on the rate in the region of parameter space relevant to the $0\nu \beta \beta$ signal. We report on studies of various $\beta$ and $\gamma$ backgrounds in the liquid-xenon-based EXO-$2000\nu \beta \beta$ experiment. With this work we try to better understand the location and strength of specific background sources and compare the conclusions to radioassay results taken before and during detector construction. Finally, we discuss the implications of these studies for EXO-200 as well as for the next-generation, tonne-scale nEXO detector.

Figure 1

Cutaway engineering drawing of the EXO-200 detector showing the relationship between the TPC, HFE, Cu cryostats, and Pb shield. The TPC, mounted on the wall of the inner cryostat with six copper supports, is surrounded by HFE cryogenic fluid. A vacuum layer lies between the inner and outer cryostats, and an air layer (‘Air gap') between the outer cryostat and the Pb shield. The clean room and surrounding active muon veto detectors are omitted for clarity.

Figure 2

Cutaway view of the EXO-200 Cu TPC. “Wiring system” refers to the flat polyimide cables behind the APD platter. Additional flat polyimide cables are within the Cu TPC legs (“cable duct” in figure). Acrylic spacers, insulators, and field cage resistors are components of the as shown “Field shaping rings” system, and are all external to the Teflon reflectors (“VUV (vacuum ultraviolet) Teflon reflector tiles”). A Cu tube and support system used for deployment of calibration sources outside the Cu vessel are not shown.

Figure 3

Comparison of MS PDFs for ${}^{232}\mathrm{Th}$ in the Cu TPC vessel and a ${}^{228}\mathrm{Th}$ source located just outside the TPC at the cathode. The PDFs are normalized to their respective 2615-keV peaks. The difference in spectral shapes arises from the energy-dependent attenuation length of $\gamma \text{s}$ and because some features are created by the coincident detection of two or more $\gamma \text{s}$. The dotted (red) vertical lines indicate the energy regions over which counts are integrated to form the ratio between the 2615-keV $\gamma$ peak and the summation peak at 3197 keV(see text).

Figure 4

In the ${}^{232}\mathrm{Th}$ spectrum, ratios of the summation peak at 3197 keV and the 2615-keV $\gamma$ peak for low-background and ${}^{228}\mathrm{Th}$ source data, and from simulation of ${}^{232}\mathrm{Th}$ point sources and ${}^{232}\mathrm{Th}$ distributed in particular detector components (see legend). Note, the following two pairs of points overlap: (1) HFE and source tube, and (2) TPC vessel + HFE + source tube and a point source at $x\sim 220$ mm. The dashed line is an empirical fit to the simulated point sources (hollow black squares) with the form $e^{a+bx}$, where $a$ and $b$ are fitted constants. Results from source data (blue filled circles) agree well with it. The fit is used to interpolate the effective distance from the center of the detector for all non-point-like simulated geometries. Statistical error bars on the points are all smaller than the size of the symbols. The shaded region is from the measured value in data, where the range is due to statistical error. This region agrees best with the combination of PDFs from ${}^{232}\mathrm{Th}$ in the TPC vessel, in the Cu source tube, and distributed in the HFE, indicating the presence of a ${}^{232}\mathrm{Th}$ background source external to the TPC vessel.

Figure 5

Behavior of ${}^{60}\mathrm{Co}$ fitted decay rate in the TPC vessel vs time. The rate shown is counts integrated above 980 keV. (The full range used by the fit is 980–9800 keV, though the ${}^{60}\mathrm{Co}$ PDF only extends to $\sim 2.5$ MeV). The vertical error bars on the points are those returned by the fit to the low-background data using the migrad function of minuit [15, 16]. Horizontal error bars show the size of the time bins. Two fits to the data are shown, including a constant and an exponential fit. The exponential fit produces a half-life of $1500\pm 1100$ days, consistent with the half-life of ${}^{60}\mathrm{Co}$. The ${\chi }^2/\text{NDF}$ is 0.5/2 (2.35/3) for the exponential (constant) fit.

Figure 6

${}^{137}\mathrm{Xe}$ decays vs time following a muon passing through the TPC. The horizontal error bars indicate the width of the time bins used. The data were fit to two functions: (1) exponential decay + constant function, with the decay fixed to the known ${}^{137}\mathrm{Xe}$ half-life, and (2) a constant. The ${\chi }^2$/NDF for 1 (2) is 0.84/2 (4.5/3), indicating a slight preference for 1.

Figure 7

${}^{222}\mathrm{Rn}$ decay rate in the LXe over run 2, measured by tracking the counts of the 5.5-MeV decay $\alpha$. Spikes in the count rate occur when the recirculation system is upset (typically by power outages), resulting in the feed of some xenon from the bottle rack. Such xenon contains a higher amount of ${}^{222}\mathrm{Rn}$ (presumably from components in the feed plumbing system) which then decays in the detector with the expected half-life. These periods of increased ${}^{222}\mathrm{Rn}$ activity are removed from the low-background data set.