The particle physics nature of dark matter (DM) is one of the most fundamental scientific questions nowadays. The leading candidates, weakly interacting massive particles (WIMPs), can be directly detected by looking for WIMP-nucleus scattering events in deep underground laboratories. Since the 1980s, physicists have improved the sensitivity of direct DM detection by about seven orders of magnitude. In the last decades, dual-phase xenon detectors exhibit their advantages in background rejection and scalability and lead the sensitivity in high mass WIMP direct searches. Experiments in XENON and LUX projects have been continuously pushing the exclusion limits of the elastic WIMP-nucleon scattering cross section into the parameter space predicted by various theoretical models.

The Particle and astrophysical Xenon (PandaX) project is a series of xenon- based ultra-low background experiments in the China Jinping Underground Laboratory (CJPL) targeting the unknown physics of DM. The first stage of the project, the PandaX-I experiment, with a 120 kg sensitive liquid xenon (LXe) target, performed the WIMP search in 2014 with a 54×80.1 kg-day exposure. PandaX-I reported a strong limit on the WIMP-nucleon cross section for a WIMP mass of less than 10 GeV/c2, strongly disfavoring all positive claims from other experiments.

The construction and installation of the second stage, PandaX-II experiment, with a half-ton scale LXe target, commenced after PandaX-I. In 2015, PandaX- II reported a WIMP search result with a 306×19.1 kg-day exposure from a short physics commissioning run with a notable 85Kr background. With 580 kg LXe in the sensitive region, PandaX-II was the largest running dual-phase xenon detector before the XENON1T detector in 2017. PandaX-II reported the most stringent limit on the WIMP-nucleon scattering cross section at 2.5×10−46 cm2 for the WIMP mass 40 GeV/c2 with a total exposure of 33 ton-day in 2016 and updated the limit to 8.6×10−47 in 2017 with a total exposure of 54 ton-day. In this dissertation, I will focus on the PandaX-II experiment, data analysis and its constraints on theoretical models.

After a distillation campaign for krypton removal in 2017, the PandaX-II experiment achieved a background level of 0.8×10−3 event/kg/day/keV which was the lowest among similar detectors at the time. Compared to other dual-phase xenon detectors, we drift electrons by applying bias voltages on the electrodes which producing a stronger uniform electric field at a strength about 400 V/cm. After running for more than three years, more than 97% of 110 3′′ photomultiplier tubes (PMTs) perform stably.

The analysis processes are continuously improved in various run periods in PandaX-II. The recorded waveforms were processed using a custom-developed software through several steps including hit finding and calculation, signal clustering, and so on to the final pairing analysis of scintillation and ionization signals. In this dissertation, I shall cover some topics in these steps as following. The amplification factors (gains) of the PMTs are calibrated using Light Emitting Diode (LED) light periodically for transforming the recorded waveform to the number of photon electrons (PE) and energy. An inefficiency raised from zero length encoding (ZLE), a data suppression firmware of the data acquisition system, is investigated in run periods with relatively low PMT gain. A data-driven algorithm is developed for the X-Y position reconstruction using the hit pattern of the proportional scintillation on the top PMT array. The mono-energetic events from xenon isotopes are studied to correct the non-uniformity of the detector response, and key parameters are extracted to reconstruct the deposited energy of events. The background analysis is critical for rare-event search experiments. I will present the study on the intrinsic electron recoil backgrounds from krypton, radon and xenon isotopes. The constraints of PandaX-II data on various theoretical models are investigated.