Dark Matter Direct Detection

with Liquid Xenon

Kaixuan Ni

University of California San Diego

1

Revealing the history of the universe with underground particle and nuclear research 2019

Tohoku University, Sendai, March 7-9, 2019

How to detect dark matter directly?

Via nuclear recoil (NR)

  • Spin-independent DM
  • Spin-dependent DM
  • Pion-coupling WIMPs
  • more than these: EFT approach
  • Self-interacting DM

Via electronic recoil (ER)

  • sub-GeV DM
  • Dark photons
  • Axion-like particles
  • SuperWIMPs
  • Axial-vector
  • Luminous DM

Or a mixture of ER & NR

  • inelastic DM
  • Magnetic inelastic DM
  • Mirror DM
  • Migdal/Bremsshtrahlung

2

XENON100, arXiv:1704.05804

NR

DM

ER

DM

Detection techniques and target materials

3

DM

NR &/or ER

A booming research field

4

this talk

What makes LXe the most favorable target?

Rich Physics Goals

  • Probe many DM models
    • SI & SD & EFT
    • Inelastic etc.
    • Heavy or sub-GeV
    • ALPs, dark photon
    • etc.
  • Neutrino astrophysics
    • Elastic scattering of solar neutrinos (pp)
    • CEvNS of B8 neutrinos
    • Supernova neutrinos
  • Neutrino physics
    • 0vbb with Xe-136
    • DEC with Xe-124

Mature Technology

  • Large target
    • online purification of the liquid/gas target
    • multi-ton target demonstrated
    • Next generation: 50~100 ton
  • Low background
    • Intrinsically pure and purifiable
    • self-shielding
    • 3D localization
    • ER/NR discrimination
  • Low threshold
    • keV threshold with both charge and light
    • O(10) eV threshold with charge only

5

Best DM

Target

Mature Technology

Rich Physics Goals

Reasonable Cost

How to build your LXe dark matter detectors?

Single phase

Two phase

6

XMASS: the largest single-phase liquid xenon detector for dark matter

  • Yasuhiro Kishimoto Talk

XENON1T: the largest two-phase liquid xenon detector ever built for dark matter

Rates for “standard” WIMP spin-independent interactions

Heavy WIMPs

Low-Mass WIMPs

7

LXe detectors push the frontier of DM detection

8

XENON1T

LUX

The evolution of dark matter detectors with LXe

9

~2000

Concept of using LXe for DM Detection

DAMA/LXe, ZEPLIN, XMASS, XENON

2007

First Results from Two-Phase Xe detectors

XENON10, ZEPLIN-II/III

2010

G1 Experiments

(0.1~1 Ton)

XENON100, LUX, PandaX-I/II, XMASS

2017

First results from the ton-scale detector

XENON1T

2020

G2 Experiments

(1~10 Ton, two-phase)

XENON1T/nT, PandaX-4T, LZ

2025

G3 Experiments

(10~100 T, two-phase)

DARWIN, PandaX-30T

(size not in scale)

XENONnT

Understanding the signals (S1 & S2)

10

Nuclear Recoil

Electronic Recoil

Heat

Excitation

Ionization

photon

electron

Recombination

(type, energy, field dependent)

S1

S2

g1 detector parameters g2

Years of effort to calibrate and understand LXe

  • External gamma rays: limitations
  • Gaseous sources are developed:
    • 129mXe, 131mXe, 127Xe, 83mKr, 37Ar
    • Tritium (CH3T), 14CH4, first in LUX
    • 220Rn, first in XENON1T
  • Nuclear recoils
    • Neutrons from AmBe or DD generator
    • High energy: DT?
    • Low energy: YBe?
    • Gaseous source??

11

DD

AmBe

Rn220

Kr83m

Calibrating LXe detectors: more accurate than ever

12

XENON1T, arXiv:1902.11297

LUX, arXiv:1712.05696

“Doke plot” to determine g1, g2 factors

Calibrating LXe detectors: more accurate than ever

13

Noble Element Simulation Technique (NEST) provides liquid xenon responses from global data fitting

NEST v2.0 is now available: http://nest.physics.ucdavis.edu

Using the S1 & S2 signals: ER/NR discrimination

14

ER vs NR

LUX as an example

arXiv:1712.05696

Most of LXe experiments show discrimination in the range of 99.5% to 99.9%. This is sufficient so far but it will become necessary to go above 99.9% for future experiment to suppress ER background events from solar neutrinos

99%

99.9%

Simulated results using NEST v2.0 (Zehong Zhao, UCSD)

Using the S1 & S2 signals: ER/NR discrimination

15

ER vs NR

LUX as an example

arXiv:1712.05696

Most of LXe experiments show discrimination in the range of 99.5% to 99.9%. This is sufficient so far but it will become necessary to go above 99.9% for future experiment to suppress ER background events from solar neutrinos

99.9%

Simulated results using NEST v2.0 (Zehong Zhao, UCSD)

Using the S1 & S2 signals: positions and fiducialization

16

XENON1T as an example

0.65 t

0.9 t

1.3 t

2 t

Combining ER/NR discrimination and fiducilization makes two-phase LXeTPC experiments very powerful in background rejection

arXiv:1805.12562, PRL

Background reduction over the years

17

LXe experiments reduce ER background significantly thanks to:

  • Low radioactive material selection
  • Purification of xenon gas
  • Powerful fiducilization

ER background rate before ER/NR discrimination

XENON1T, background rate evolution with online Kr-reduction (distillation)

ER background: lowest achieved by XENON1T, but dominated by Radon in the bulk LXe

18

XENON Preliminary

NR Background from neutrons

XENON1T, arXiv:1902.11297

Neutrons make multiple scattering in LXe. Multiple scatter neutrons are rejected in DM search, but can be used to estimate single scatter neutron background. Single NR background is a concern for the upcoming experiments.

19

Highlight of Recent Dark Matter Results

20

XENON1T: largest exposure & lowest background

  • Exposure: one tonne x year (Nov.22, 2016 ~ Feb.8, 2018)
  • Dominant ER background: 82 events/ton/yr/keVee
  • Best Spin-independent limit: 4.1 x 10-47 cm2 at 30 GeV/c2

21

arXiv:1805.12562

Phys. Rev. Lett. 121, 111302 (2018)

XENON1T: the best SD-neutron limits

  • Best WIMP Spin-Dependent (neutron) limits: 6.3 x 10-42 cm2 at 30 GeV/c2

22

arXiv:1902.03234, submitted to PRL

Solid line: 10 GeV/c2

Dashed line: 100 GeV/c2

Instead of coupling to one nucleon, the WIMP couples to a virtual pion between two nucleons

XENON1T: first results on WIMP-pion coupling

  • Best WIMP-pion limit: 6.4 x 10-46 cm2 at 30 GeV/c2

23

arXiv:1811.12482

Phys. Rev. Lett. 122, 071301 (2019)

For cross sections at 10-46 cm2

Instead of coupling to one nucleon, the WIMP couples to a virtual pion between two nucleons

PandaX-II: constraints on the SIDM with a light mediator

  • Exposure: 54 ton x day from 2016~2017 runs

24

arXiv:1802.06912

Phys. Rev. Lett. 121, 021304 (2018)

Sub-GeV dark matter scattering

  • NR from sub-GeV DM scattering: energy too low
  • DM-nucleus scattering accompanied by a Bremsstrahlung photon or “Migdal” electron: ER signal

25

  • M. Ibe et al., JHEP 02 (2018) 194
  • Dolan et al., PRL 121, 101801 (2018)

LUX, arXiv:1811.11241

ER signal for

1 GeV DM at 10-35 cm2

Dolan et al., Phys. Rev. Lett. 121, 101801 (2018)

XMASS constraints on dark/hidden photon and ALPs

arXiv:1807.08516 (PLB)

26

Annual Modulation Signal Search

  • Excluding the leptophic DM models favored by DAMA’s modulation signals
  • Demonstrate LXe detector’s long-term operational stability.

27

XMASS, 2.7 years data

XMASS, arXiv:1801.10096

PRD 97, 102006 (2018)

LUX, 20 months

LUX, arXiv:1807.07113

PRD 98, 062005 (2018)

XENON100, 4 years

XENON, arXiv:1701.00769

PRL 118, 101101 (2017)

The Near Future: PandaX-4T, XENONnT, LZ

28

PandaX-4T

  • A scale-up from PandaX-II at Jin-Ping Lab
    • 1.2 m diameter
    • 1.2 m drift length
    • 4-ton active LXe target
  • Schedule:
    • assembly/commission: 2019~2020
    • Science data taking: 2020~2022
  • Sensitivity reach:
    • SI interaction: 6 x 10-48 cm2

29

arXiv:1806.02229

1 mDRU = 1 event/keVee/ton/day

XENONnT (talk by Shigetaka Moriyama)

30

LZ

31

arXiv:1802.06039

Construction underway NOW at SURF

1000 live-days x 5.6 ton

Technical challenges to be solved in these (G2) experiments

  • Radon concentration in the bulk liquid xenon
    • Lowest achieved in XENON1T: 5~10 μBq/kg
    • Goal of the G2 experiments : 1~2 μBq/kg
    • Rn control, online distillation, charcoal adsorption
  • Neutron background (neutron veto needed)
    • LZ: liquid scintillator
    • XENONnT: Gd-doped water (see Poster by Ryuichi Ueno)
  • Long electron drift length (1.2~1.5 m)
    • Require >1 ms electron lifetime: fast/efficient purification
    • Need faster drift velocity to avoid too much diffusion: 30~100 kV on cathode
  • Large diameter (1.2~1.5 m) TPC
    • Electron emission rate from gate/cathode electrodes needs to be controlled
    • Signal uniformity

32

The lowest achieved Rn in XENON1T is ~5 uBq/kg

Dark Matter sensitivity reach in the next 5 years

WIMP Dark Matter Detection in five years?

33

LZ, arXiv:1802.06039

PandaX, arXiv:1806.02229

Benchmark point: 10-47 cm2 at 250 GeV/c2

XENON, arXiv:1512.07501

The G3 LXe Experiment

34

The case for a G3 LXe detector

  • As already demonstrated by past experiments, two-phase LXeTPC is an ideal choice for dark matter detection
  • But science reach of the LXeTPC is more than dark matter…
    • Neutrinoless double beta decay (Xe-136)
      • 100-ton natural xenon contains 9 ton Xe-136!
    • Neutrino Astrophysics
      • Electron scattering: pp, Be-7, etc.
      • Coherent scattering: B-8, DSN, atmospheric neutrinos
  • The call for a global effort to build the next generation (G3) LXe detector
    • LXe mass: at least 50 tonnes
    • Technical design & demonstration: 2020~2024
    • Construction: 2024~2025
    • Commissioning and Science data taking: 2025-2035

35

DARWIN: the G3 “Ultimate” dark matter detector

Baseline design:

  • 2.6 m x 2.6 m TPC
  • 40 ton active LXe target (total 50 ton)
  • ~10 m2 photo-sensor coverage (top/bottom)

36

https://darwin.physik.uzh.ch/

VUV-MPPC

JCAP 11, 017 (2016)

WIMP Spectroscopy with DARWIN

1 and 2 sigma credible regions of simulated WIMP signals for SI interactions at various WIMP masses and cross-sections for a 200 ton x year exposure in DARWIN

37

JCAP 11, 017 (2016)

DARWIN sensitivity to solar axion and ALPs

Solar axion

Galactic ALPs

38

JCAP 11, 017 (2016)

For solar axions, both flux and cross section depend on g_Ae, thus rate scales with 4th power of the coupling

For nonrelativistic galactic ALPs, the flux is independent from the coupling, thus the rate depends on g_Ae^2

Solar neutrino-electron scattering in DARWIN

as signal

→ 2850 neutrinos per year (89% pp)


→ achieve 1% statistical precision
 on pp-flux with 100 t x y

as background

ER rejection efficiencies ~99.98% at
 30% NR efficiency are required to
 reduce to sub-dominant level

Other physics channels:

  • CEvNS
  • 0vbb
  • etc.

JCAP 11, 017 (2016)

Science Channels for the G3 LXe Experiment

40

Credit: XENON1T, Purdue University

Planck mass ~ 10^19 GeV/c^2 (0.02 mg)

Summary

It’s a golden time to work on liquid xenon experiments!

  • Liquid Xe has become the most favorable target for dark matter detection; Ton-scale experiment is already probing many interesting DM models.
  • The upcoming G2 experiments (PandaX-4T, XENONnT, LZ) with unprecedented low background may give us a first glimpse of the nature of dark matter in 5 years.
  • The G3 LXe experiment at 50~100 tonnes scale, e.g. DARWIN, will be the ultimate dark matter detector and may reveal the history of universe in nuclear, particle and astro-physics in the next two decades.

41

LXe-DarkMatter-Review - Google Slides