Extending the CRESST-II commissioning run limits to lower masses

0.39 MB·English

by  Andrew Brown

OUTP-11-50-P
Extending the CRESST-II commissioning run limits to lower masses
Andrew Brown,1,∗ Sam Henry,1 Hans Kraus,1 and Christopher McCabe2
1Department of Physics, University of Oxford, Keble Road, Oxford OX1 3RH, UK
2Rudolf Peierls Centre for Theoretical Physics, University of Oxford,
1 Keble Road, Oxford OX1 3NP, UK
(Dated: January 24, 2012)
Motivated by the recent interest in light WIMPs of mass ∼O(10GeV/c2), an extension of the
elastic, spin-independent WIMP-nucleon cross-section limits resulting from the CRESST-II com-
missioning run (2007) are presented. Previously, these data were used to set cross-section limits
from 1000GeV/c2 down to ∼17GeV/c2, using tungsten recoils, in 47.9 kg-days of exposure of cal-
cium tungstate. Here, the overlap of the oxygen and calcium bands with the acceptance region of
the commissioning run data set is reconstructed using previously published quenching factors. The
2
resulting elastic WIMP cross-section limits, accounting for the additional exposure of oxygen and
1 calcium, are presented down to 5GeV/c2.
0
2
3 Jan Among the oIu.tstIaNnTdiRngODpUroCblTemIOsNin astro-particle Light yield00..34 ATCATCuuccaannccllccggeeiissuuppttmmtteeaann nn11 cc11sseess rreeggiioonn ooff [[22]]
2 physics is to identify the non-baryonic dark matter that OOxxyyggeenn 11ss
makesup∼80%ofthematterintheUniverse[1]. Awell- 0.2

] motivated class of dark matter candidates are Weakly
O
Interacting Massive Particles (WIMPs), which may be 0.1
C directly detected via their elastic scattering with nuclei
h. in terrestrial detectors. 0
In the CRESST-II commissioning run of 2007 [2], lim-
p
its on elastic, spin-independent, WIMP-nucleon cross-
-
o sections for WIMP masses from 1000GeV/c2 down to -0.1
tr ∼17GeV/c2 were presented. However, light WIMPs,
as with a mass ∼O(10GeV/c2) have been suggested as -0.20 10 20 30 40 50
Energy [keV]
[ a possible interpretation to the experimental results of

DAMA, CoGeNT and CRESST-II (2011) [3–5]. At the
2 FIG. 1: Acceptance region and nuclear recoil band diagram
same time, several other experiments [6–11] partially or
v for Zora/SOS23. The black line indicates the light yield
9 completely exclude the light WIMP interpretations of (L /E)limitfortheacceptanceregionin[2],andthevertical
γ
8 [3–5]. Therefore, it is of interest to see how data from dashed lines the acceptance region’s lower and upper energy
5 the commissioning run compare to these results when thresholds. 1σregionsforobservingrecoilingoxygen(green),
2 extended to examine such light WIMP scenarios. calcium (red) and tungsten (blue) nuclei are indicated, using
9. light detector resolution parameters in Table I and quench-
0 ing factors from [13]. A similar diagram may be drawn for
1 Verena/SOS21.
II. INCLUDING OXYGEN AND CALCIUM
1
:
v
The CRESST-II commissioning run data set [2] con-
i tron recoils, produce more light than nuclear recoils. By
X sists of 47.9 kg-days exposure of calcium tungstate
r (CaWO4),takenbetweenthe27th ofMarchandthe23rd convention the light produced per unit energy (the ‘light
a ofJuly,2007. Thedatawereobtainedfromtwoindepen- yield’) of electron recoils Lγ/E, is calibrated to those of
gamma interactions at 122.06keV, which is normalised
dent modules, labelled by their phonon detector / light
to a value of one. Nuclear recoils see a reduction in light
detector names: “Zora/SOS23” and “Verena/SOS21”,
compared to electron recoils, quantified by a quenching
collecting 23.8 and 24.1 kg-days respectively [12].
factor, Q , dependent upon the species i of recoiling nu-
In CRESST-II, the phonon detector is used to mea- i
cleus.
sureenergyandthelightdetectortodistinguishbetween
event types, as different types of interaction in CaWO The limits of [2] resulted from assuming all recoils
4
producedifferentamountsoflight, relativetotheenergy in the acceptance region, the region in which WIMP-
deposited. Gamma and beta interactions, causing elec- nucleon interactions are searched for, were from tung-
sten alone. However, due to the effects of a finite light
detector resolution, recoils from both calcium and oxy-
gen may also be seen within the same acceptance re-
∗Electronicaddress: [email protected] gion. This effect is illustrated in Figure 1. The addi-
2
Parameter Zora/SOS23 Verena/SOS21 n 1
o
σ0(keV) 0.784 1.508 gi
e
σ1(keV12) 1.064 0.610 e r0.8
c
σ2 0.192 0.154 an
pt
P0(keV) 0.56 0.11 ce0.6
c
P 0.0040 0.0065 a
1 h
wit
p 0.4
a
TABLEI:Resolutionvaluesusedinthiswork. σ0,σ1 andσ2 erl
are light detector resolution parameters determined from [2], v Tungsten in [2]
O
usingthemethodoutlinedinAppendixA.P0 andP1 arethe 0.2 Tungsten
Calcium
phonon energy resolution parameters from [14].
Oxygen
0
0 10 20 30 40 50
Energy [keV]
tional exposure provided by the parts of calcium and
oxygen bands that fall within the acceptance region can
FIG. 2: Estimates of the fraction of recoils of a given target
strengthen cross-section limits for light WIMPs, with
nucleusthatfallwithintheacceptanceregionforZora/SOS23
mass ∼O(10GeV/c2). In [2], the acceptance region was
in [2] as a function of energy. Oxygen is shown by green
chosensothattungstenrecoilswouldhavebeenseenwith diamonds,calciumbyredcrossesandtungstenbybluecircles,
minimalelectronrecoilbandoverlap. Inenergy, thiswas withblackshowingtheinitialassumptionof90%oftungsten
between 10 and 40keV. In light yield, the upper limit recoils falling within the acceptance region. Vertical dashed
was set so that 90% of tungsten recoils would occur in lines at 10 and 40 keV indicate acceptance region limits in
theacceptanceregion. Here,weusethissameacceptance recoil energy.
region, so that we do not introduce non-blind elements
into the analysis.
To consider oxygen and calcium recoils in this region, for calcium and 3.91+0.478% for tungsten.1 It should be
−0.430
thefractionofeachnuclearspecies’recoilsthatfallwithin notedthatthemorerecentmeasurementsoflightoutput
the acceptance region must be estimated. This requires fortungstenrecoilsinCaWO in[13]arehigherthanthe
4
severalpiecesofinformation. Thefirstisthelightdetec- 2.5%usedin[2]. Thismeansthattheamountoflightfor
tor resolution of the observed electron recoil band, as a tungstenrecoilsisonaveragehigherthanthatwhichwas
function of energy. This resolution is expressed as: expectedin[2], causinglessthantheexpected90%ofall
tungsten recoils to fall within the acceptance region, as
σ2(E)=σ2+σ2E+σ2E2, (1) can be seen in Figure 2 for Zora/SOS23.
γ 0 1 2
One last piece of information would be required for a
where E is the energy in the phonon channel. The reso- complete description of detected light. This is the small
lutionoftheelectronrecoilbanddependsonthreeterms: deviation of observed light in the electron recoil band
σ , reflecting electronic noise; σ , related to the Poisson from the normalisation of one unit of light per unit en-
0 1
distributionoftheexpectednumberofdetectedphotons; ergy. Two effects can cause this deviation: the depen-
andσ ,incorporatingpositiondependenceandotherpos- dence of light yield on energy in inorganic scintillators
2
sible effects seen in CRESST-II light detectors. [15], and an overall calibration error. Such adjustments
as used in the analysis of [2] are unavailable, although it
An additional piece of information needed is the res-
isstatedin[2]thatthelightyieldisalwaysnearthenor-
olution of each quenched band. At energy E, events in
malisation of one. Here we use the approximation that
the quenched band produce an average amount of light
themeanelectronrecoillightyieldisoneeverywhere. In
L (E) = Q L (E). The resolution of a quenched band
Qi i γ an independent analysis of the commissioning run data
is assumed to be equal to the resolution of the electron
recoil band at energy E(cid:48), where L (E) = L (E(cid:48)). The [14], the electron recoil band behaviour and light detec-
Qi γ tor resolution were parameterised, with results repeated
exact light detector resolutions used in [2] are unavail-
inAppendixB.Asacheckonourresults,wealsoconsid-
able. However, these resolutions may be obtained by
ered the resulting WIMP cross-section limits with these
fitting to the acceptance region figures in [2], using the
parameters. They are consistent with those presented
methodoutlinedinAppendixA.Theselightdetectorres-
here to within a few percent.
olutions are given in Table I. Separately, the energy res-
olution of the phonon detector was modelled in [14] by
∆E =P +P E, with energy resolution parameters also
0 1
shown in Table I.
1 Anoxygenquenchingfactorof10.4+0.5%wasusedin[5]. Limits
The next piece of information is the quenching factor −0.5
calculatedusingthisquenchingfactorareslightlystrongerthan
of each target nucleus. For this, the measurements in
thosepresentedhere,aresultofalargerfractionofoxygenrecoils
[13] are used, with 11.09+0.909% for oxygen, 6.38+0.619% beingobservablewithintheacceptanceregion.
−0.908 −0.653
3
pb] 10-1 ATulln rgesctoeinls pb]
n [10-2 Calcium n [10-4
ctio Oxygen ctio (CloDwM Sen I.I)
e e
s-s10-3 s-s10-5
s s
o o
Cr10-4 Cr (XSE2N oOnNly1)0 Extended
10-6 CRESST-II (2007)
10-5
10-7 C(2R0E1S1)S 2Ts-II CDMS II
10-6 CoGeNT 90%
10-8 DAMA 90% XENON100
10-7
10 102 103 5 10 20 50 100
WIMP mass [GeV/c2] WIMP mass [GeV/c2]
FIG. 3: 90% confidence limits on elastic, spin-independent, FIG. 4: The combined 90% confidence limit on the elastic,
WIMP-nucleon cross-sections from data in [2], considering spin-independent WIMP-nucleon cross-section from extend-
all possible nuclear recoils within the acceptance region. ing the analysis of commissioning run data to lower WIMP
The green dashed-dotted, red dashed and light-blue dotted masses (solid blue). For comparison, the favoured regions
lines result from considering WIMP interactions with oxy- from DAMA, derived from [3], CoGeNT [4] and CRESST-
gen, calcium and tungsten individually, and the solid blue II (2011) [5] are shown. Also shown are WIMP cross-
line the total rate. The WIMP halo properties used are section limits from, CDMS II [6], CDMS II (low energy) [7],
ρ = 0.3GeV/cm3, v = 544km/s, v = 220km/s and XENON100 [8] and XENON10 (S2 only) [9].
DM esc 0
v = 232km/s. Resolutions from Table I and quenching
sun
factors from [13] were used to derive these limits.
well modelled by considering interactions from tungsten
alone. At lower masses, calcium, then oxygen recoils be-
With these pieces of information, the fraction of re- come dominant, such that below ∼7GeV/c2, nearly all
coiling nuclei from each constituent of CaWO that falls
4 WIMP-nucleon interactions in the acceptance region are
within the acceptance region of [2] can be estimated, as
with oxygen nuclei. Considering all possible nuclear re-
showninFigure2forZora/SOS23. Withthesefractions,
coils then provides a significant strengthening of cross-
the interaction rate of WIMPs with oxygen and calcium
sectionlimitsatlowmassescomparedtotungstenalone.
in the commissioning run acceptance region may now be
In Figure 4, a comparison of the combined limit is
calculated. Herewefollowamethodanalogoustothatin
made to the elastic WIMP interpretation of other ex-
[2]. The elastic, spin-independent WIMP-nucleon inter-
periments [3–9]. The CoGeNT, CRESST-II (2011) and
actionratesarecalculatedfollowing[16], usingtheHelm
DAMAresultswerealreadyintensionwiththeresultsof
form factor parameterisation as suggested in [17]. The
XENON100,XENON10(S2only),CDMSII,andCDMS
total rate expected from all target nuclei is then:
II (low energy). The extended CRESST-II commission-
ingrunlimitsintroducefurthermildtensionwithDAMA
dR dR dR dR
dETot =AW dEW +ACa dECa +AO dEO , (2) and CRESST-II (2011).
As the commissioning run and CRESST-II (2011) re-
for the fractions Ai of each species’ nuclear recoils that sults are with the same target nuclei with similar en-
may be seen in the acceptance region. This rate is con- ergy thresholds, it is difficult to reduce this mild tension
volved with the observed phonon energy resolution, ∆E, by choosing different astrophysical parameters or parti-
as described in [17]. clephysicsmodels. However, itshouldbenotedthatthe
CRESST-II commissioning run and CRESST-II (2011)
run do not use the same acceptance region definitions.
III. RESULTS In thiswork, we have used the acceptanceregion defined
in the original commissioning run analysis. While this
Three events are observed in [2], at 16.89keV, ensures that we have not introduced non-blind elements
18.03keV and 33.09keV. The Maximum Gap method into the analysis, this region has not been optimised for
[18] is used to calculate the resulting elastic, spin- lightmassWIMPdiscovery. Anadditionaldifferencebe-
independent, WIMP-nucleon cross-section limits. The tween runs is the design of clamps in direct contact with
results are shown in Figure 3. The extended 90% confi- thetargetcrystals,whichasnotedin[5]introducedaddi-
dencelimitimprovessensitivitytolowmassWIMPswith tional backgrounds into the CRESST-II (2011) data set.
commissioningrundata. InteractionswithWIMPsheav- Since the commissioning run live-time is much smaller
ier than ∼17GeV/c2 in the acceptance region are dom- than in the CRESST-II (2011) run, repeating the com-
inated by tungsten recoils, and the cross-section limit is missioning run experimental conditions for a longer pe-
4
riod would allow stronger conclusions to be drawn. oxygen quenching factor, given as ∼11.1% in the text of
[2]isallowedtovaryinthefit. Avalueof11.0%isfound
from fitting to both modules’ acceptance regions.
IV. CONCLUSIONS
The WIMP cross-section limits for the 47.9 kg-days Appendix B: Alternative light detector resolutions
exposure of CaWO in the CRESST-II commissioning
4
run [2] have been extended down to a WIMP mass of In [14], the electron recoil behaviour was modelled by:
5GeV/c2. Our analysis has accounted for possible oxy-
gen and calcium recoils within the commissioning run L (E)= l1E , (B1)
acceptance region, using light and phonon detector reso- γ 1+e−leE
lutions in Table I and quenching factors from [13]. The
giving electron recoil band behaviour parameters
improvementofcross-sectionlimitsatlightmassesoccurs
and light detector resolutions for Zora/SOS23:
because recoiling oxygen and calcium nuclei dominate
over tungsten recoiling nuclei for light WIMPs. Extend-
Parameter Zora/SOS23
ing the commissioning run limits results in mild tension
l 1.068±0.003
with the recent CRESST-II [5] and DAMA [3] results. 1
l (keV−1) 0.180±0.007
e
σ (keV) 1.1±0.3
0
Acknowledgments
1
σ1(keV2) 0.46±0.03
σ 0.178±0.004
We wish to thank Franz Pr¨obst for providing the 2
favoured region contours of recent CRESST-II results,
Jens Schmaler for numerous helpful comments on this Light detector resolution for Verena/SOS21 was mod-
work, and Felix Kahlhoefer for useful discussions. We elled by three time-separated noise regions with differ-
alsowishtoacknowledgetheScienceandTechnologyFa- inglightdetectorresolutions,labelled“High”,“Medium”
cilities Council, UK who funded this work. and “Low” noise regions. The electron recoil band be-
haviour and light detector resolution parameters were:
Appendix A: Reconstructing light detector Parameter Verena/SOS21
resolutions
High Low Medium
l 1.021±0.005 1.035±0.003 1.036±0.002
To estimate light detector resolutions, oxygen and 1
l (keV−1) 0.169±0.001 0.171±0.008 0.22±0.01
tungsten acceptance regions were modelled by: e
σ (keV) 3.5±0.6 1.18±0.37 1.26±0.13
0
Acc(Qi,E)= QiLγ(E)+ENsigσQi(E), (A1) σ1(keV12) 1.00±0.15 0.72±0.07 0.76±0.04
σ 0.03±0.08 0.09±0.01 0.106±0.006
2
where Q is the quenching factor of the considered nu-
i
cleus, and N ≈1.28 is the number of standard devia- The live-time distribution between regions is taken
sig
tions allowing 90% of quenched recoils to be seen in the from Figure 9.22 of [14] at 20%, 32% and 48% respec-
acceptance region. These equations were fitted simulta- tively. However, when using these light detector resolu-
neously to the tungsten and oxygen nuclear recoil accep- tions, it should be noted that there are some differences
tance regions in Figure 8 of [2]. The tungsten quenching between the exact cuts and data selection between [14]
factor is set at 2.5% and L /E is taken to be one. The (47.5 kg-days) and [2] (47.9 kg-days).
γ
[1] G. Bertone, D. Hooper, and J. Silk, Phys. Rept. 405 327 (2010), 1619–1621, [0912.3592].
(2005), 279–390, [hep-ph/0404175]. [7] The CDMS-II Collaboration, Z. Ahmed et al.,
[2] G.Angloheretal.,AstroparticlePhysics31(2009),270– Phys.Rev.Lett. 106 (2011), 131302, [1011.2482].
276, [0809.1829]. [8] The XENON100 Collaboration, E. Aprile et al.,
[3] R. Bernabei et al., Eur. Phys. J. C67 (2010), 39–49, Phys.Rev.Lett. 107 (2011), 131302, [1104.2549].
[1002.1028]. [9] The XENON10 Collaboration, J. Angle et al.,
[4] C.E.Aalsethetal.,Phys.Rev.Lett.107(2011),141301, Phys.Rev.Lett. 107 (2011), 051301, [1104.3088].
[1106.0650]. [10] TheCDMSandEDELWEISSCollaborations,Z.Ahmed
[5] G. Angloher et al., (2011), 1109.0702. et al., Phys. Rev. D84 (2011), 011102, [1105.3377].
[6] The CDMS-II Collaboration, Z. Ahmed et al., Science [11] M.Felizardo,T.Girard,T.Morlat,A.Fernandes,F.Giu-
5
liani, et al., (2011), 1106.3014. [15] R. Murray and A. Meyer, Phys. Rev. 122 (1961), 815.
[12] R. F. Lang and W. Seidel, New J. Phys. 11 (2009), [16] F.Donato,N.Fornengo,andS.Scopel,Astropart.Phys.
105017, [0906.3290]. 9 (1998), 247–260, [hep-ph/9803295].
[13] P. Huff, The Detector Parameters Determining the Sen- [17] J. D. Lewin and P. F. Smith, Astropart. Phys. 6 (1996),
sitivityoftheCRESST-IIExperiment,Ph.D.thesis,Max 87–112.
Planck-Institut for Physik, March 2010. [18] S. Yellin, Phys. Rev. D66 (2002), 032005,
[14] R. Lang, Search for Dark Matter with the CRESST ex- [physics/0203002].
periment, Ph.D. thesis, Max Planck-Institut for Physik,
October 2008.