Taking Inventory of the Universe:
The Mystery of Dark Matter
Stanley SeibertUniversity of PennsylvaniaLBNL, February 20, 2013
Today’s topic: A gravitational mystery...
...brought to you by precision astronomy
Abell 1703
Seven Decades of “Excess Gravitation”
Rotation Curves Gravitational Lensing
CMB Power Spectrum
Rotation Curves
Baryon Acoustic Oscillations
Cluster Collisions
Simulations of Structure FormationAnd many others!
Responses to the Unexpected
1. “Observational” error.
2. Interpretation or modeling error.
3. New or modified interaction between matter.
4. New material constituent. (usually a new particle)
“Precision astronomy is hard. Maybe you made a mistake.”
“Are you really sure you know where all the baryons are?”(Black holes, neutron stars, brown dwarfs, etc.)
Modified Newtonian Dynamics (MOND), TeVeS
Neutralinos, axions, Kaluza-Klein states, ...
Pick any combination of all 4!
The Dark Matter Hypothesis
Rotation Curves
A substantial fraction of the matter in the
universe is in a form that rarely (or never)
interacts with photons, rendering it invisible
(“dark”) to direct electromagnetic observation.
But, isn’t proposing a new form of matter based purely on astronomical
evidence preposterous?
Rotation Curves
radiation of a great thickness of hydrogen’.19 Indicating his uncertainty concerningthe origin of the line, he referred to ‘the D line of hydrogen (?)’.20
By that time Lockyer had teamed up with Edward Frankland, professor at theRoyalCollege ofChemistry inLondon andoneofBritain’smost distinguished chemistsin the Victorian era. Importantly, Frankland had considerable experience withchemical spectroscopy, an area of research which was new to Lockyer. The astronomerand the chemist collaborated in comparing solar lines with lines produced in thelaboratory of gases under varying pressure and temperature. One of the solar linesunder examination was D3, although this was only a minor part of their work. Theyconfirmed that ‘there is a line near D visible in the chromosphere to which there is nocorresponding Fraunhofer line’ and that the new D line was unlikely to belong to thehydrogen spectrum, still poorly understood at the time. It appears that Frankland waslesswilling thanLockyer to exclude theD3 line as being of hydrogenic origin. ‘I thinkweought not so easily to give up all efforts to get it [D3] from terrestrial hydrogen’, hewrotein a letter of 7 April 1869. More than three years later he wrote in another letter toLockyer about the assumption that D3 belonged to an element unknown on Earth: ‘Iremember always protesting in our conversations about the yellow line, against makingthis assumption, until we had exhausted every effort to get the line out of hydrogen’.21
Figure 1. Emission spectrum of D lines from sunspot, showing the D3 or helium line. Thetwo stronger lines to the left are the sodium lines. From Kayser (note 30), p. 188.
19 J.N. Lockyer, ‘Spectroscopic Observations of the Sun, II’, Philosophical Transactions of the RoyalSociety, 159 (1869), 425!44, and ‘Spectroscopic Observations of the Sun, III’, Proceedings of the RoyalSociety of London, 17 (1869), 350!56. Quotations from Lockyer 1874 (note 15), p. 459 and p. 478. Thesymbols C, F and D refer to the place of the lines in the Fraunhofer absorption spectrum.
20 J.N. Lockyer, ‘Spectroscopic Observations of the Sun, III’ (note 19), and Lockyer 1874 (note 15), p.486. The first part of the series on ‘Spectroscopic Observations of the Sun’ appeared in 1866, an account ofthe new metod of observing solar prominences. As a further indication of Lockyer’s uncertainty withrespect to the nature of the D3 line, in a table of bright lines in the chromosphere he added in a footnote tothe D3 line the single word ‘Hydrogen’, italicized and followed by a question mark (ibid. p. 495).
21 Letters of 7 April 1869 and 9 September 1872, quoted in Meadows 1972 (note 14), 59!60. See alsoColin A. Russell, Edward Frankland. Chemistry, Controversy and Conspiracy in Victorian England(Cambridge: Cambridge University Press, 1996), p. 436. Frankland proposed to use exceedingly longhydrogen tubes to see if the mysterious line would be produced in this way, but apparently these tubes werenever constructed. See Lockyer and Lockyer 1928 (note 14), p. 42.
163Reconsideration of Helium’s Early History
Downloaded By: [University of Pennsylvania] At: 20:38 22 September 2010
We’ve Been Here Before...
E. Frankland
Helium was first discovered by astronomers in the solar chromosphere in 1868, but not
by chemists in the lab until 1895!W. Ramsay
N. Lockyer
What is Dark Matter?Suppose you decide to search for “terrestrial” dark matter.
What do you know?If you explain the astronomy data with dark matter, then you know:
• Cross-sections for interaction between dark matter and itself/other particles are very small.(or you would have seen it already)
• Local density near Earth is around 0.3 GeV/cm3
(within a factor of 2 or 3)
• Solar system moves through dark matter halo at220 km/sec
Direct Dark Matter Searches(“looking for your lost keys under the street light”)
1. Anomalous nuclear recoils(WIMP scattering)
2. Primakoff interactions(axion-photon coupling)
3. Periodicity/Directionality(the 21st century search for the “aether wind”)
4. [Insert your clever idea here]
XENON, CDMS, CoGeNT, DEAP/CLEAN, LUX, PICASSO, COUPP, CRESST, XMASS, EDELWEISS, ...
ADMX, CAST, ...
DAMA/LIBRA, DRIFT, DMTPC, ...
Why WIMPs?
AA48CH13-Feng ARI 16 July 2010 22:3
Y
T (GeV)
ΩX
t (ns)
10–4mX = 100 GeV
10–6
10–8
10–10
10–12
10–14
10–16
100 101 102 103108
106
104
102
100
10–2
10–4
101 100
Figure 2The comoving number density Y (left) and resulting thermal relic density (right) of a 100-GeV, P-waveannihilating dark matter particle as a function of temperature T (bottom) and time t (top). The solid graycontour is for an annihilation cross section that yields the correct relic density, and the shaded regions are forcross sections that differ by 10, 102, and 103 from this value. The dashed gray contour is the number densityof a particle that remains in thermal equilibrium.
X: General darkmatter candidate
the number of dark matter particles become negligible, but interactions that mediate energyexchange between dark matter and other particles may remain efficient.
This process is described quantitatively by the Boltzmann equation
dndt
= !3H n ! "!Av#(n2 ! n2eq), (5)
where n is the number density of the dark matter particle X, H is the Hubble parameter, "!Av#is the thermally averaged annihilation cross section, and neq is the dark matter number density inthermal equilibrium. On the right-hand side of Equation 5, the first term accounts for dilutionfrom expansion. The n2 term arises from processes XX $ SM SM that destroy X particles, whereSM denotes SM particles, and the n2
eq term arises from the reverse process SM SM $ XX, whichcreates X particles.
The thermal relic density is determined by solving the Boltzmann equation numerically. Arough analysis is highly instructive, however. Defining freeze out to be the time when n"!Av# = H ,we have
n f % (mX T f )3/2e!mX /T f %T 2
f
M Pl"!Av#, (6)
where the subscripts f denote quantities at freeze out. The ratio x f & mX /T f appears in the ex-ponential. It is, therefore, highly insensitive to the dark matter’s properties and may be considereda constant; a typical value is xf % 20. The thermal relic density is, then,
"X = mX n0
#c= mX T 3
0
#c
n0
T 30
% mX T 30
#c
n f
T 3f
% x f T 30
#c M Pl"!Av#!1, (7)
www.annualreviews.org • Dark Matter Candidates 501
Ann
u. R
ev. A
stro
. Ast
roph
ys. 2
010.
48:4
95-5
45. D
ownl
oade
d fr
om w
ww
.ann
ualre
view
s.org
by U
nive
rsity
of P
enns
ylva
nia
on 0
4/25
/11.
For
per
sona
l use
onl
y.
Annu. Rev. Astron. Astrophys. 2010. 48:495–545
A new particle with
mass near the EW
symmetry-breaking
scale and weak force
gauge couplings
produces the right
thermal relic density.
Hunting for WIMPsElena Aprile Dark Matter Detection and the XENON Experiment
Figure 1: Event rates for a 100 GeV/c2 WIMP with spin-independent WIMP-nucleoncross-section of 10�44cm2 for di�erent target materials.
For a recent review of the field we refer to the report by the DUSEL S1 DarkMatter Working Group ([4], and references therein). Covering the bulk of the SUSYparameter space for WIMPs will require a sizable increase in sensitivity from thecurrent best experimental limits [7, 6]. An increase in detector mass and exposure, inaddition to a reduction in and/or improved rejection of radioactive and cosmogenicbackgrounds is necessary.
The predicted event rates for a WIMP mass of 100 GeV/c2 and a spin-independentWIMP-nucleon cross-section of 10�44cm2 are shown in Fig. 1 for Ge, Xe and Artargets. The fast fall of the event rate with increasing recoil energy demands a verylow energy threshold, around 10 keV. At this energy, the event rate for a Xe target isabout 30% higher than for a Ge target, due to the Xe larger atomic number. Cryogenicsolid state detectors, based on Ge and Si crystals, have for a long time dominated thefield of dark matter direct detection, showing the best background discrimination andreporting stringent spin-independent WIMP-nucleon cross-section (4.6�10�44cm2 ata WIMP mass of 60GeV/c2 [7]).
In recent years, however, the application of cryogenic noble liquids in dark mat-ter searches, has gained new momentum due to their promise for large target massdetectors with possibly as powerful background discrimination as cryogenic crystals.LXe and LAr are especially attractive as they are known to be good scintillators andionizers, as established in many works. The scintillation mechanism in these liquidsis well known [8]. Both excitation and electron-ion pairs recombination produce ex-
185
E. Aprile, http://www.slac.stanford.edu/econf/C080625/pdf/0018.pdf
Want:
Low energy threshold
Large, “cheap,” and clean target material
Excellent separation of nuclear recoils (induced by WIMPs) from other backgrounds
Background Discrimination
Scintillation Ionization
Heat/Phonons
DEAP/CLEAN, XMASS,
DAMA/LIBRA
XENON, LUX, WARP, ArDM
COUPP, PICASSO
CDMS, EDELWEISSCRESST, ROSEBUD
DRIFT, DMTPC,CoGeNT
Experimental Results:How are we doing so far?
In the Thick of It
arXiv:1207.5988
5
]2WIMP Mass [GeV/c6 7 8 910 20 30 40 50 100 200 300 400 1000
]2W
IMP-
Nuc
leon
Cro
ss S
ectio
n [c
m
-4510
-4410
-4310
-4210
-4110
-4010
-3910
]2WIMP Mass [GeV/c6 7 8 910 20 30 40 50 100 200 300 400 1000
]2W
IMP-
Nuc
leon
Cro
ss S
ectio
n [c
m
-4510
-4410
-4310
-4210
-4110
-4010
-3910
]2WIMP Mass [GeV/c6 7 8 910 20 30 40 50 100 200 300 400 1000
]2W
IMP-
Nuc
leon
Cro
ss S
ectio
n [c
m
-4510
-4410
-4310
-4210
-4110
-4010
-3910
DAMA/I
DAMA/Na
CoGeNT
CDMS (2010/11)EDELWEISS (2011/12)
XENON10 (2011)
XENON100 (2011)
COUPP (2012)SIMPLE (2012)
ZEPLIN-III (2012)CRESST-II (2012)
XENON100 (2012)observed limit (90% CL)
Expected limit of this run:
expectedσ 2 ± expectedσ 1 ±
FIG. 3: New result on spin-independent WIMP-nucleon scat-tering from XENON100: The expected sensitivity of this runis shown by the green/yellow band (1�/2�) and the result-ing exclusion limit (90% CL) in blue. For comparison, otherexperimental results are also shown [19–22], together withthe regions (1�/2�) preferred by supersymmetric (CMSSM)models [18].
the benchmark region fluctuates to 2 events is 26.4% andconfirms this conclusion.
A 90% confidence level exclusion limit for spin-independent WIMP-nucleon cross sections �� is calcu-lated, assuming an isothermal WIMP halo with a lo-cal density of ⇢� = 0.3GeV/c3, a local circular veloc-ity of v0 = 220 km/s, and a Galactic escape velocity ofvesc = 544 km/s [17]. Systematic uncertainties in the en-ergy scale as described by the Le↵ parametrization of [6]and in the background expectation are profiled out andrepresented in the limit. Poisson fluctuations in the num-ber of PEs dominate the S1 energy resolution and arealso taken into account along with the single PE resolu-tion. The expected sensitivity of this dataset in absenceof any signal is shown by the green/yellow (1�/2�) bandin Fig. 3. The new limit is represented by the thick blueline. It excludes a large fraction of previously unexploredparameter space, including regions preferred by scans ofthe constrained supersymmetric parameter space [18].
The new XENON100 data provide the most strin-gent limit for m� > 8GeV/c2 with a minimum of� = 2.0 ⇥ 10�45 cm2 at m� = 55GeV/c2. The max-imum gap analysis uses an acceptance-corrected expo-sure of 2323.7 kg⇥days (weighted with the spectrum of a100GeV/c2 WIMP) and yields a result which agrees withthe result of Fig. 3 within the known systematic di↵er-ences. The new XENON100 result continues to challengethe interpretation of the DAMA [19], CoGeNT [20], andCRESST-II [21] results as being due to scalar WIMP-nucleon interactions.
We acknowledge support from NSF, DOE, SNF, UZH,Volkswagen Foundation, FCT, Region des Pays de laLoire, STCSM, NSFC, DFG, Stichting FOM, WeizmannInstitute of Science, and the friends of Weizmann Insti-tute in memory of Richard Kronstein. We are grateful toLNGS for hosting and supporting XENON.
⇤ Electronic address: [email protected]† Electronic address: [email protected]
[1] N. Jarosik et al., Astrophys. J. Suppl. 192, 14 (2011);K. Nakamura et al. (PDG), J. Phys. G37, 075021 (2010).
[2] G. Steigman and M. S. Turner, Nucl. Phys. B253, 375(1985); G. Jungman, M. Kamionkowski, and K. Griest,Phys. Rept. 267, 195 (1996).
[3] G. Bertone, D. Hooper, and J. Silk, Phys. Rept. 405, 279(2005).
[4] M. W. Goodman and E. Witten, Phys. Rev. D31, 3059(1985).
[5] E. Aprile et al. (XENON100), Astropart. Phys. 35, 573(2012).
[6] E. Aprile et al. (XENON100), Phys. Rev. Lett. 107,131302 (2011).
[7] M. Aglietta et al. (LVD), Phys. Rev.D58, 092005 (1998).[8] E. Aprile et al. (XENON100) Astropart. Phys. 35, 43
(2011).[9] E. Aprile et al., Phys. Rev. C79, 045807 (2009).
[10] X. Du et al., Rev. Sci. Instr. 75, 3224 (2004).[11] E. Aprile et al. (XENON100), (2012), arXiv:1207.3458.[12] E. Aprile et al. (XENON100), Phys. Rev. D84, 052003
(2011).[13] S. Yellin, Phys. Rev. D66, 032005 (2002).[14] G. Plante et al., Phys. Rev. C84, 045805 (2011).[15] E. Aprile et al., Phys. Rev. Lett. 97, 081302 (2006).[16] M. M. Szydagis et al., JINST 6, P10002 (2011).[17] M. C. Smith et al., Mon. Not. R. Astron. Soc. 379, 755
(2007).[18] Combined region using C. Strege et al., JCAP 1203,
030 (2012); A. Fowlie et al. (2012), arXiv:1206.0264;O. Buchmueller et al. (2011), arXiv:1112.3564.
[19] C. Savage et al., JCAP 0904, 010 (2009).[20] C. E. Aalseth et al. (CoGeNT), Phys. Rev. Lett. 106,
131301 (2011).[21] G. Angloher et al. (CRESST-II), Eur. Phys. J. C72, 1971
(2012).[22] Z. Ahmed et al. (CDMS), Science 327, 1619 (2010);
Z. Ahmed et al. (CDMS), Phys. Rev. Lett. 106,131302 (2011); E. Armengaud et al. (EDEL-WEISS), Phys. Lett. B 702, 329 (2011); J. Angleet al. (XENON10), Phys. Rev. Lett. 107, 051301 (2011);M. Felizardo et al. (SIMPLE), Phys. Rev. Lett. 108,201302 (2012); E. Behnke et al. (COUPP) (2012),arXiv:1204.3094; D. Y. Akimov et al. (ZEPLIN-III), Phys. Lett. B 709, 14 (2012); E. Armengaudet al. (EDELWEISS) (2012), arXiv:1207.1815.
Summary of the data so far:
! It took astronomers many decades to sort out evidence for dark matter.
! Particle physicists are still very early in the process, and there is plenty of room for other approaches....
Can we look to past successes for inspiration?
Lessons From Neutrino Experiments
Low energy neutrino detection demands large, clean, and deep underground experiments, much like dark matter.
Scaling requires a detector that can be composed from “simple” repeatable structures.
Percent-level calibration is possible, especially when you start with a well-modeled detector.
Careful modeling of backgrounds allows for low threshold signal extraction.
Sudbury Neutrino Observatory
Ex: SNO Radon Spike
Best fit convolution of Monte Carlo by ~0.1 MeV
Rn spike data
In situ, distributed calibration sources are a powerful systematic constraint
MiniCLEAN:Searching for Dark Matter withArgon and Neon Scintillation
Scintillation in Noble Liquids
Energy deposition in noble liquids produces short lived excited diatomic molecules in singlet and triplet states.
Pulse Shape Analysis
Electronic recoil
Nuclear Recoil
Triplet state highly suppressed!
Singlet Triplet
He ~10ns 13 s
Ne <18.2 ns 14.9 μs
Ar 7 ns 1.60 μs
Xe 4.3 ns 22 ns
Rejecting Electron-like Events in Argon
Discriminate with ratio of prompt to
total light
Reject beta and gamma
backgrounds with less than 10-8
leakage
Number of photoelectrons0 50 100 150 200 250 300 350 400 450 500
prom
ptF
00.10.20.30.40.50.60.70.80.9
1
)ee (keVeffT0 20 40 60 80 100 120 140 160
FIG. 7: Fprompt versus energy distribution for neutrons and! rays from an Am-Be calibration source. The upper bandis from neutron-induced nuclear recoils in argon, the lower-band is from background !-ray interactions.
promptF0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Even
ts/0
.01
wid
e bi
n
1
10
210
310
410
510
610-ray eventsγ 7 10×1.7
100 nuclear recoil events
FIG. 8: Fprompt distribution for 16.7 million tagged !-rayevents from the 22Na calibration, and nuclear recoil eventsfrom the Am-Be calibration, between 120 and 240 photoelec-trons (approximately 43–86 keVee). No !-ray events are seenin the nuclear recoil region.
measured the triplet lifetime in DEAP-1 over the courseof the run to check that impurities did not build up inthe detector over time.
We use 22Na calibration data to measure the tripletlifetime. For each calibration run, we find all events thatpass the data cleaning cuts and contain over 200 photo-electrons. The raw traces for these events are aligned ac-cording to the measured trigger positions and summed.We then fit the following model to the average trace be-tween 500 and 3000 ns from the trigger:
f(t) = A exp(!t/!3) + B, (3)
where A is a normalization factor, !3 is the triplet life-time and B is a constant baseline term.
As a consistency check, we measured !3 for photo-electron bins of size 200 between 200 and 1600 photo-
(cm)fitZ-50 -40 -30 -20 -10 0 10 20 30 40 50
bin
(Hz)
fitR
ate
per Z
-510
-410
-310
-210
-110
1 Surface backgroundSNOLAB backgroundPSD signal
FIG. 9: Comparison of Zfit distribution for !-rays from thePSD data, and for high-Fprompt backgrounds during the run(labeled Surface backgrounds). Also shown, for reference, isthe distribution of high-Fprompt background events with thedetector operating underground at SNOLAB.
Days since Aug. 19, 20070 10 20 30 40 50 60
bg.
rate
(mH
z)pr
ompt
Hig
h F
0
2
4
6
8
FIG. 10: High-Fprompt background event rate versus time.The average background rate is 4.6 ± 0.2 mHz.
electrons and did not observe any systematic e!ect fromthe signal size. There are systematic errors associatedwith both the fit window and the linear baseline correc-tion discussed in Section III C. We estimated the size ofthe error associated with the fit window to be 40 ns bychanging the start and end times of the fit by 500 ns.We performed the fit for both corrected and uncorrectedtraces and estimated the size of the error associated withthe baseline to be 50 ns. We added the two estimatedsystematic errors to determine a combined systematicerror of 60 ns.
The measured lifetimes over the course of the runfor traces without the baseline correction are shown inFig. 12, in which the error bars shown are statistical only.We observe no significant increase in the impurity levelthroughout the run, and we measure the long time con-stant to be 1.46±0.06 (sys) µs, consistent with previousmeasurements [5, 13, 14]. Further analysis of systematic
6
M.G. Boulay et al. arXiv:0904.2930
Important to reject intrinsic 39Ar background
Observing Extreme UV
Almost everything absorbs 128 nm light!TPB can wavelength shift EUV up to 440 nm with high efficiency.
Wavelength [nm]80 100 120 140 160 180
]-1
Scin
tilla
tion
Prob
abili
ty D
ensit
y [n
m
0.02
0.04
0.06
0.08
0.1
0.12
Tran
smitt
ance
[%]
10
20
30
40
50
60
70
80
90Helium Scint. Trans.
2MgF
Neon Scint. Sapphire Trans.
Argon Scint. Synth. Sil. Trans.
Krypton Scint. UVT Glass Trans.
Xenon Scint.
NIM A 654 (2011), 116-121V. Gehman, SRS, K. Rielage, A. Hime, Y. Sun, D.-M. Mei
Wavelength [nm]120 140 160 180 200 220 240
Fluo
resc
ence
Effi
cien
cy
0.6
0.8
1
1.2
1.4
1.6
1.8
TPB Re-emission Efficiency
10% measurement resolves factor of 3 ambiguity in the literature
→ Disentangles TPB and argon scintillation efficiency, useful for other noble liquids
NIM A 654 (2011), 116-121V. Gehman, SRS, K. Rielage, A. Hime, Y. Sun, D.-M. Mei
TPB Re-emission Spectrum
Wavelength [nm]350 400 450 500 550 600
]-1
Prob
abili
ty D
ensit
y [n
m
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014 Illumination Wavelength:128 nm 160 nm175 nm 250 nm
Discovered that re-emission spectrum independent of illumination wavelength and has cutoff near 400 nm. → Direct impact on choice of optical materials
NIM A 654 (2011), 116-121V. Gehman, SRS, K. Rielage, A. Hime, Y. Sun, D.-M. Mei
Single Phase Ar/Ne DetectorsAdvantages:
• Target material is very inexpensive.• No need for electric fields to drift charge.
• Simpler detector design• Able to use a spherical geometry (but not required)• Does not require 39Ar-depleted argon for large
detectors• Neon is clean enough to use for pp solar neutrinos
Disadvantages:• Lower A2 reduces coherent scattering enhancement• Self-shielding from external backgrounds worse than other
materials• Atmospheric argon contains a high rate beta decay
isotope, 39Ar @1 Bq/kg (also a fantastic calibration!)
The DEAP and CLEAN Family of Detectors
DEAP-0:Initial R&D detector
DEAP-1:7 kg LAr2 warm PMTsAt SNOLab since 2008
picoCLEAN:Initial R&D detector
microCLEAN:4 kg LAr or LNe2 cold PMTssurface tests at Yale
MiniCLEAN:500 kg LAr or LNe (150 kg fiducial mass)92 cold PMTsSNOLAB 2013DEAP-3600:
3600 kg LAr (1000 kg fiducial mass)266 warm PMTsSNOLAB 2014
40-140 tonne LNe/LAr Detector:pp-solar ν, supernova ν, dark matter <10-46 cm2
~2018?
10-44 cm2
10-45 cm2
10-46 cm2
WIMP σ Sensitivity
MiniCLEAN CollaborationBoston University
D. Gastler, E. Kearns, S. Linden
UC BerkeleyG.D. Orebi Gann
Los Alamos National LaboratoryM. Akashi-Ronquest, K. Bingham, R. Bourque, J. Griego,
A. Hime, F. Lopez, J. Oertel, K. Rielage, L. Rodriguez
Massachusetts Institute of TechnologyJ.A. Formaggio, S. Jaditz, J. Kelsey, K. Palladino
National Institute Standards and TechnologyK. Coakley
University of New MexicoM. Bodmer, F. Giuliani, M. Gold, D. Loomba, J. Matthews, P. Palni,
J.Wang
University of North Carolina/TUNLR. Henning
University of PennsylvaniaT. Caldwell, J.R. Klein, A. Mastbaum,
S. Seibert
Royal Holloway University of LondonA. Butcher, J. Monroe, J.A. Nikkel, J. Walding
University of South DakotaV. Guiseppe, D.-M. Mei, G. Perumpilly, C. Zhang
Syracuse UniversityR.W. Schnee, B. Wang
Yale UniversityD.N. McKinsey
The MiniCLEAN Detector
Courtesy J. Griego
Inner Vessel
PMT
OuterVessel
LAr/LNe
92 8” PMTs
TPB @ R=43 cm
PMTs @ R=81 cm
Modular “Cassettes”
Test fit of 1 light guide in inner vesselat fabricator
Simplified View
Liquid Ar/NeTarget
Liquid Ar/NeShielding
Acrylic
UV fluor(TPB)
PMTs
Water Shielding
Courtesy J. Griego
InnerVessel
OuterVessel
WaterShieldTank
Deck
Veto PMTs
SNOLAB
SurfaceFacility
2 km of rock(6000 mwe)
UndergroundLaboratory
Sudbury, Ontario, Canada
Lab Area
Courtesy F. Lopez
Insert MiniCLEAN
here
Inner Vessel
Manufacturing completed, pressure and leak tested in Sept. 2012
Inner Vessel Underground
January 2013
MiniCLEAN WIMP AnalysisWill perform a maximum likelihood analysis with a blind signal box in three reconstructed observables:
Ener
gy
Radius
Part
icle
ID
WIMP Sensitivity
MiniCLEAN w/ 150 kg fiducial volume
80 = 80 keVr threshold50 = 50 keVr threshold30 = 30 keVr threshold
Small change in energy threshold
equivalent to large change in fiducial
volume!
Lowering the Energy Threshold
• Rejection of 39Ar beta decay events sets the energy threshold in single-phase argon.
• 39Ar is the only background that scales like volume, rather than surface area, so it also limits detector size.
• Important to maximize the performance of our pulse-shape rejection algorithms.
• Prior to construction, we have done this with a detailed simulation...
MiniCLEAN Simulation: RAT
• Developed for Braidwood (SRS), now used by SNO+ and DEAP/CLEAN.
• Wraps together GEANT4, ROOT, and GLG4sim (KamLAND) into full simulation and analysis package.
• Fully propagates optical photons, including PMT response and digitizer / DAQ simulation
• Includes pulse shape analysis tools, maximum likelihood position reconstruction
Comparison with MicroCLEAN
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2Energy (MeVee)
0
0.2
0.4
0.6
Co
un
ts /
sec
on
d /
keV
microCLEAN dataRAT Monte CarloAfter setting global
detection efficiency in Monte Carlo
57Co source
Modeling of Cryogenic PMT Response
T. Caldwell
Time (0.1 ns)1800 1900 2000 2100 2200 2300 2400 2500
AD
C C
ount
s
1920
1940
1960
1980
2000
2020
2040
2060
Time (0.1 ns)1800 1900 2000 2100 2200 2300 2400 2500
AD
C C
ount
s2000
2010
2020
2030
2040
2050
MiniCLEAN uses Hamamatsu R5912-02-MOD 8” PMTs, tested for use at temperatures down to 25K.
Photoelectron (PE) Counting
• Counting the number of PE on a channel is fundamental to energy and position estimation.
• PMTs produce variable size, finite width (~20ns) pulses for each PE.
• Counting peaks in a waveform is biased low when there is pulse pileup.
• Integrating charge is unbiased, but higher variance.
• How to get best of both worlds?
• (Side note: This is one of many instances where single phase detectors get better as they get bigger and add channels...)
Sample (4 ns)0 500 1000 1500 2000 2500 3000 3500 4000
AD
C C
ount
s
1500
2000
2500
3000
3500
Probably many photoelectrons
Probably single photoelectrons
(regardless of amplitude)
Bayesian PE Counting
Bayesian PE Counting
PN (n|q, t1, t2) =PQ(q |n)PN (n | t1, t2)P1i=0 PQ(q | i)PN (i | t1, t2)
Chop up waveform for each PMT into regions with pulses, and for each region, compute:
n-PE charge distribution
Probability of n-PE given pulse integral q, and pulse
ranging from t1 to t2
Poisson probability of observing n (or i) photons
between t1 and t2 in this PMT.
This is where we insert our knowledge of argon scintillation time structure and event position & energy
Position & Energy Reconstruction
No photon can travel directly from the event
vertex to a PMT!
• Rayleigh scattering length in argon is ~90 cm (or 66 cm?).
• TPB further randomizes photon paths.
• Estimating PMT hit probability based simply on solid angle tends to bias the fit inward.
• The actual light pattern is more isotropic than a direct propagation model would predict.
ShellFit: Self-Tuning Maximum Likelihood Fitting
• Automatically use Monte Carlo simulation to collapse complete space of photon histories to before and after wavelength shifting.
• Integrate over TPB surface using a GPU to compute likelihood.
• 4 sec/event w/ CPU and 0.2 sec/event w/ GPU!
UV transmission to TPB Visible
mc) / PEmc - PErec(PE-0.25 -0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25
Even
ts /
0.00
5
1
10
210
310
410
Charge Division
Bayesian PE Counting
Energy resolution after iterating PE counting
MiniCLEAN simulation, 39Ar, 75-100 PE, inside fiducial volume
Particle ID Techniques
PE0 50 100 150 200 250 300
pf
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1
10
210
310
PE0 50 100 150 200 250 300
pr
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1
10
210
310
PE0 50 100 150 200 250 300
rl
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
1
10
210
310
Prompt ratio (standard)
PE0 50 100 150 200 250 300
pf
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1
10
210
310
PE0 50 100 150 200 250 300
pr
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1
10
210
310
PE0 50 100 150 200 250 300
rl
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
1
10
210
310
Likelihood ratio (new)
MiniCLEAN simulation
Nuclear recoils
Electron recoils
Nuclear recoils
Electron recoils
The Future• The “Generation 2” concept for single phase liquid argon
and neon is the CLEAN experiment.
MiniCLEAN(~500 kg total)
CLEAN-40(44 tons total)
CLEAN-140(140 tons total)
The CLEAN Family
MiniCLEAN(~500 kg total)
CLEAN-40(44 tons total)
CLEAN-140(140 tons total)
The Future
(Front TPB plate removed)
Simpler PMT modules
The Future
• Between 40 and 140 tons total argon mass (≥15 tons fiducial).
• Optimal size with natural argon limited by pulse-shape discrimination performance and pileup.
• At ~140 tons (total), a dual-use argon and neon detector can do a percent-level precision measurement of pp solar neutrinos and observe supernova neutrinos.
• With depleted argon, could go very large...
Conclusion
• Something is out there, and it might be WIMPs!
• We’ve seen claims of direct detection, but you should continue to be skeptical.
• Single phase noble liquid detectors offer a highly scalable option for dark matter and neutrino detection.
• MiniCLEAN extends the DEAP/CLEAN series of detectors to 150 kg fiducial volume with liquid argon and neon.
• Many analysis improvements ready to test on data as it arrives.
• Construction is underway, with detector commissioning scheduled for mid-2013.
“I often look at the bright yellow ray emitted from the chromosphere of the sun, by that unknown element, Helium, as the astronomers have ventured to call it. It seems trembling with excitement to tell its story, and how many unseen companions it has. And if this be the case with the sun, what shall we say of the magnificent hosts of the stars? May not every one of them have special elements of their own? Is not each a chemical laboratory in itself?”
John William DraperInaugural Address to the American Chemical Society1876