a novel nuclear recoil calibration in the lux detector using a d-d

LUA Novel Nuclear Recoil

Calibration in the LUX Detector Using a D-D Neutron Generator

James Verbus Brown University

April APS Meeting

April 12th, 2015

James Verbus - Brown University APS April Meeting - April 12th, 2015

mWIMP (GeV/c2)

WIM

P−nu

cleo

n cr

oss

sect

ion

(cm

2 )

5 6 7 8 9 10 1210−44

10−43

10−42

10−41

10−40

Low Mass WIMPs - Fully Excluded by LUX

2

>20x more sensitivity

CDMS II Si Favored

CoGeNT Favored

LUX (2013)-85 live days LUX +/-1σ expected sensitivity

XENON100(2012)-225 live days

CRESST Favored

CDMS II Ge

x

DAMA/LIBRA Favored


Conservative Light and Charge Yields Assumed for LUX 2014 PRL

3

• Modeled Using Noble Element Simulation Technique (NEST)• Szydagis et al., arxiv:1106.1613

• NEST based on canon of existing experimental data.• Artificial cutoff in light and charge yields assumed below 3 keVnr, to be conservative.• Includes predicted electric field quenching of light signal, to 77-82% of the zero field light yield

• Conservative threshold used in LUX 2014 PRL Dark Matter Result arXiv:1310.8214v2

Artificial 3 keVnr Ionization Cut-off

(same as with scintillation)

pertains to S1 pertains to S2

http://arxiv.org/abs/1310.8214v2

Samuel Chan, Carlos Faham for the LUX Collaboration

e-e-e-

e-e-

e-e-e-e-e-

e- e-e-e-

θ


Monochromatic2.5 MeV neutrons

e-e-e-e-e-

e- e-e-e-

θ

S2[1y’]



Δt

θ



Δt

top hit pattern: x-y localization

Δt : z’ separation

θ : energy calculation

θ




Neutron Conduit Installed in the LUX Water Tank

5


0 10 20 30 40 50

0

50

100

150

200

250

300

y’ Distance into LXe [cm]Drift

time[µs]

log10(cts / µs / cm)

−1

−0.5

0

0.5

1

1.5

2

2.5

3

3.5

4

Adelphi DD108 Neutron Generator Installed Outside LUX Water Tank

6

• Neutron generator/beam pipe assembly aligned 16 cm below liquid level in LUX active region to maximize usable single / double scatters

• Beam leveled to ~1 degree

• 107.2 live hours of neutron tube data used for analysis

PreliminaryLUX


0 10 20 30 40 50

0

50

100

150

200

250

300

y’ Distance into LXe [cm]Drift

time[µs]


−1

−0.5

0

0.5

1

1.5

2

2.5

3

3.5

4

Adelphi DD108 Neutron Generator Installed Outside LUX Water Tank

7

• This cut eliminates shine from passive materials and ensures 95% of neutrons in beam sample have energy within 4% of 2.45 MeV

• The mean energy of neutrons produced at 90o

by the DD108 was measured to be 2.45 ± 0.05 MeV at Brown University

PreliminaryLUX

Approximate fiducial for yield measurements


Beam Projection in Active Region (x’)

• Accepting all events along y’. The shine from scatters in passive detector materials is visible.

8

−20 −10 0 10 20

0

50

100

150

200

250

300

x’ perpendicular to neutron tube [cm]

Drift

time[µs]


−1

−0.5

0

0.5

1

1.5

2

2.5

3

3.5

4

−20 −10 0 10 20

0

50

100

150

200

250

300

x’ perpendicular to neutron tube [cm]

Drift

time[µs]


−1

−0.5

0

0.5

1

1.5

2

2.5

3

• After application of y’ > 15 cm beam purity cut. This cut eliminates shine from passive materials and ensures 95% of neutrons in beam sample have energy within 4% of 2.45 MeV.

PreliminaryLUX

PreliminaryLUX


Energy Measured from Scattering Angle [keVnra]100 101

IonizationSignal

[electrons]

100

101

102

Observed Ionization Signal • Event Selection Cuts

• Event Identification

• Select double scatters

• Determine vertex ordering via scattering geometry only

• Neutron Beam Energy Purity

• Enforced via position of scatters along beam line / depth into active LXe

• After geometry cut, 95% of neutrons have energy within 4% of 2.45 MeV

• Data Quality

• Ensure quiet detector conditions

• Ensure properly reconstructed events

• Cuts are flat for S2[1y’] (first scatter

along beam direction) in energy region of interest

Grey Points - Individual double scatter events

9

PreliminaryLUX

Double Scatter (S1, 2xS2s > 50 phe)


100 101100

101

102

Energy Measured from Scattering Angle [keVnra]

IonizationSignal[electrons] Sys. uncertainty (flat)

Sys. uncertainty (±1σ)

Sys. uncertainty due to pos. rec. energy bias correction

Ionization Signal Absolutely Measured below 1 keVnra in LUX

• Reconstruct number of electrons at interaction site by matching ionization signal model with observed event distribution using extended maximum-likelihood

• Red systematic error bar shows common scaling factor uncertainty. Dominated by uncertainty in electron extraction efficiency.

• Lowest event energy included for analysis is 0.3 keVnra.

Grey Points - Individual double scatter eventsMagenta Crosses - Error bars for individual event from best 10% from each bin Blue Crosses - Reconstructed number of electrons at interaction site accounting for threshold effects in signal analysisBlack Dashed Line - Szydagis et al. (NEST v1.0) Predicted Ionization Signal at 180 V/cm

10

Reconstructed Number of Electrons with Associated Statistical Uncertainty

Example Error Bars for Individual Events


PreliminaryLUX


100 101100

101

102


IonizationSignal[electrons] Sys. uncertainty (flat)



Ionization Signal Absolutely Measured below 1 keVnra in LUX

• Reconstruct number of electrons at interaction site by matching ionization signal model with observed event distribution using extended maximum-likelihood

• Red systematic error bar shows common scaling factor uncertainty. Dominated by uncertainty in electron extraction efficiency.

• Lowest event energy included for analysis is 0.3 keVnra.

Grey Points - Individual double scatter eventsMagenta Crosses - Error bars for individual event from best 10% from each bin Blue Crosses - Reconstructed number of electrons at interaction site accounting for threshold effects in signal analysisBlack Dashed Line - Szydagis et al. (NEST v1.0) Predicted Ionization Signal at 180 V/cm

10

Reconstructed Number of Electrons with Associated Statistical Uncertainty

Example Error Bars for Individual Events


LUX 2014 PRL Conservative Threshold Cut-Off

PreliminaryLUX



IonizationYield

[electrons/keV

nra]


Sys. uncertainty (flat)


100 101100

101

Ionization Yield Absolutely Measured below 1 keVnra in LUX

• Red error bars show systematic uncertainties

• (1𝝈) bar dominated by statistical uncertainty in electron extraction efficiency

• (flat) bar accounts for detector parameter uncertainties and variations during the run

• Pos. Rec. bias correction error bars compensate for modest Eddington bias due to position reconstruction uncertainties

Blue Crosses - LUX Measured Qy; 180 V/cm (absolute energy scale) Green Crosses - Manzur 2010; 1 kV/cm (absolute energy scale) Orange Crosses - Manzur 2010; 4 kV/cm (absolute energy scale)Purple Band - Z3 Horn Combined FSR/SSR; 3.6 kV/cm (energy scale from best fit MC)Teal Lines - Sorensen IDM 2010; 0.73 kV/cm (energy scale from best fit MC)Black Dashed Line - Szydagis et al. (NEST v1.0) Predicted Ionization Yield at 180 V/cm

11

LUX 2014 PRL ConservativeThreshold Cut-Off

Reconstructed Ionization Yield with Associated Statistical Uncertainty

PreliminaryLUX



0 500 1000 150010−1

100

101

102

103

S2sc (phe)

Counts

Measuring the Scintillation Yield • Use single scatters with

suitable selection criteria

• MC using measured LUX D-D charge yield to simulate expected single scatter energy spectrum with LUX threshold, purity, electron extraction, energy resolution effects applied

• Simulation uses JENDL-4.0 angular scattering cross-sections with isotope selection determined based upon natural abundance and total elastic cross-sections

• Ly measurement range is 0-900 phe S2sc using bins of 100 phe

• Simulation event distribution is normalized outside of Ly measurement range using 900 < S2sc < 1500 phe

12

Single Scatter (S1, 1xS2s > 50 phe)

PreliminaryLUX


0 500 1000 150010−1

100

101

102

103

S2sc (phe)

Counts







12


PreliminaryLUX

Select slice of S2bc to use for S1

comparison


0 500 1000 150010−1

100

101

102

103

S2sc (phe)

Counts







12


PreliminaryLUX

Select slice of S2bc to use for S1

comparison

Normalize simulation

spectrum using this region


S1c Spectrum from 400-500 S2sc

• Parameter α used to scale input energy to simulated S1 spectrum for fixed slice in S2sc (absolutely calibrated by LUX D-D Qy)

• For each fixed S2sc bin, determine percent deviation from model Ly determined by optimizing α via unbinned maximum likelihood comparison between data and simulation S1 spectra

• Both absolute number of events and spectrum shape incorporated into optimization

13

S1 photons at interaction site [photons]0 50 100 150 200 250

Cou

nts

0

5

10

15

20


Cou

nts

0

5

10

15

20

PreliminaryLUX


Cou

nts

0

5

10

15

20

PreliminaryLUX

PreliminaryLUX

α = 0.5 α = 0.92 α = 1.5



Sys. uncertainty 32mKr yield (±1σ)

Sys. uncertainty (flat)

Uncertainty from Qy (±1σ)

Enegy (keVnrS2)100 101 102

Ly[photon

s/keV

nrS

2]

100

101

Lyrelative

to32

mKr(32.1keV)

10-1

Leff Measured in LUX Using Absolute Energy Scale

14

LUX 2014 PRL ConservativeThreshold Cut-Off

PreliminaryLUX


Blue Crosses - LUX Measured Ly; reported at 180 V/cm (absolute energy scale) Green Crosses - Manzur 2010; 0 V/cm (absolute energy scale)Purple Band - Horn Combined Zeplin III FSR/SSR; 3.6 kV/cm, rescaled to 0 V/cm (energy scale from best fit MC)Orange Crosses - Plante 2011; 0 V/cm (absolute energy scale) Grey Crosses - Aprile 2009 (absolute energy scale)Black Dashed Line - Szydagis et al. (NEST) Predicted Scintillation Yield at 181 V/cm

• LUX Ly values reported at 180 V/cm

• X error bars representative of error on mean of population in bin

• Energy scale defined using LUX measured Qy

• Method can be extended below existing 1.2 keVnrS2 point

• 32mKr light yield at 32.1 keV measured to be

45.7 ± 3.13 photons/keV using same D-D beam fiducial


Conclusions

• Novel LUX absolute nuclear recoil calibration performed using mono-energetic D-D neutrons in-situ

• Clear confirmation of the response used in the first LUX WIMP search analysis with an order of magnitude improvement in calibration uncertainties

• The existing WIMP analysis only assumed a detector response at and above 3 keVnr

• D-D neutron calibration technique allows us to calibrate detector response in region well below this, and so further improve LUX sensitivity to low mass WIMPs

• Coming soon

• LUX paper on D-D results

• NEST fit to D-D light and charge yields

15


Extra slides

16


Spin-Independent Sensitivity

17

mWIMP (GeV/c2)

WIM

P−nu

cleo

n cr

oss

sect

ion

(cm

2 )

101 102 103

10−45

10−44

LUX (2013)-85 live daysLUX +/-1σ expected sensitivity



ZEPLIN IIICDMS II Ge

Edelweiss II


Projected LUX 300 day WIMP Search Run

• LUX 300 day run is underway

• Extending sensitivity by another factor 5

• Even though LUX sees no WIMP-like events in the current run, it is still quite possible to discover a signal when extending the reach

• LUX does not exclude LUX

• WIMPs remain our favored quarry

• LZ 20x increase in target mass

• If approved plans to be deployed in Davis Lab in 2016+

18

mWIMP (GeV/c2)

WIM

P−nu

cleo

n cr

oss

sect

ion

(cm

2 )

101 102 10310−46

10−45

10−44

x5

LUX (2013)-85 live days

LUX (2013)-300 live days

a novel nuclear recoil calibration in the lux detector using a d-d

Documents