EDELWEISS-II, direct Dark Matter search experiment, first data analysis and results

HAL Id: tel-00490990

https://tel.archives-ouvertes.fr/tel-00490990

Submitted on 10 Jun 2010

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

first data analysis and results

Siliva Scorza

To cite this version:

N

d’ordre 199-2009

LYCEN – T 2009-19

Thèse

présentée devant

l’Université Claude Bernard Lyon-I

Ecole Doctorale de Physique et d’Astrophysique

pour l’obtention du

DIPLOME de DOCTORAT

Spécialité : Physique des Particules

(arrêté du 7 août 2006)

par

Silvia SCORZA

EDELWEISS-II, direct Dark Matter search experiment :

first data analysis and results

Soutenue le 6 novembre 2009

devant la Commission d’Examen

Jury : M.

B.

Ille

Président du jury

M.

J.

Gascon

Directeur de thèse

M.

G.

Gerbier

M.

A.

Giuliani

Rapporteur

N

d’ordre 199-2009

LYCEN – T 2009-19

Thèse

présentée devant

l’Université Claude Bernard Lyon-I

Ecole Doctorale de Physique et d’Astrophysique

pour l’obtention du

DIPLOME de DOCTORAT

Spécialité : Physique des Particules

(arrêté du 7 août 2006)

par

Silvia SCORZA

EDELWEISS-II, direct Dark Matter search experiment :

first data analysis and results

Soutenue le 6 novembre 2009

devant la Commission d’Examen

Jury : M.

B.

Ille

Président du jury

M.

J.

Gascon

Directeur de thèse

M.

G.

Gerbier

M.

A.

Giuliani

Rapporteur

Many people have followed dire tly and indire tlymy work and my ven-tures during these years in Lyon. I owe mu h to them, but rst of all I wouldliketothankmysupervisor JulesGas onforbeingalwayspresentand patient enough toanswermy questions, willingto help me with suggestions and en ouragementwhenever I needed it. I amalsovery gratefulto himfor ommuni atingto me the greatest enthusiasm forresear h.

IamgladtothanktheEDELWEIS S ollaborationespe iallythe astropar-ti le group in Lyon for the warm wel ome, for always giving me valuable suggestions about my workand areer and for their help and patien e.

I would liketo thank the members of the ommittee, Bernard Ille, Gilles Gerbier and Jules Gas on and in parti ular a spe ial thank to the referees, Andrea Giulianiand Josef Jo hum, for having read and evaluated my work and provided me with fruitfuland valuable omments.

I willremaingrateful tomyo emate Mar -Antoine forputtingup with me in omi and tragi ir umstan es, and for amusing me with interesting dis ussions in both ases. I am also grateful to the whole third oor for thoughtfully tolerating the aforementioned dis ussions and my singing per-forman esfromthe adja ent o es. Ungrand mer iauxservi es te hniques du laboratoire,en parti ulierl'équipe informatique, pour leur soutien.

I would like to greatly thank all my friends and olleagues at IPNL and elsewhere for being always ni e and fun and helping me to enjoy the past three years.

Tout d'abord je remer ie les do torants et non-do torants qui ont su transformer en moments très onviviaux les repas de midi, pauses, goûters, et , et , toutes es hoses qui rendent le quotidien agréable e un parti o-lare ringraziamentoall'irridu ibile gruppo di italianidel laboratorio (Silvia, Federi a, Giulia,Mar o, Mauro, Gia omoe Silvano) semprepresente.

Un mer i spe ial à Myriam, une amie très hère toujours prête à rigoler et qui a su me remonter le moral dans mes phases de down. Les soirées, le Martini, lesbières, lamusique, lesen eintes, la hevalerie et laritirata.

ria e Marta divise tra Padova e Genova ma sempre presenti e la oppia Fran es hini-Ponzano he hanno saputo rallegrare le ormai rare serate gen-ovesi.

Introdu tion 1

1 Dark Matter Challenge 5

1.1 Theoreti alframework . . . 5

1.1.1 Standard Cosmology . . . 5

1.1.2 Historyof the universe ina nutshell . . . 8

1.2 Motivationsand eviden es for Dark Matter . . . 10

1.2.1 The gala ti s ale . . . 10

1.2.2 Galaxy luster s ale. . . 12

1.2.3 Cosmologi als ale . . . 13

1.3 The WIMP Hypothesis . . . 19

1.3.1 Parti le Candidate . . . 21 1.4 DM Dete tion . . . 24 1.4.1 Indire t Sear h . . . 24 1.4.1.1 GammaRays . . . 25 1.4.1.2 Antimatter . . . 25 1.4.1.3 Neutrinos . . . 26 1.4.2 Dire tSear h . . . 27 1.4.2.1 Dete tion Te hniques. . . 28

1.4.2.2 Theoreti alre oil spe trum . . . 34

1.4.2.3 Ex lusion plot . . . 38

2 The EDELWEISS Experiment 41 2.1 Expe te d Ba kground . . . 42

2.1.1 Intera tions inside dete tors . . . 42

2.1.2 Startingpoint: EDELWEISS-I . . . 45

2.2 The EDELWESS -II setup . . . 46

2.3 Dete tors . . . 47

2.3.1 Ge-NTD . . . 51

2.3.1.1 Ionizationmeasurement . . . 52

2.4 Ele troni s and data a quisition . . . 55

3 Data Analysis 57 3.1 Signal Pro essing . . . 58

3.2 Energy Calibration . . . 58

3.3 Resolution of heat and ionization hannels . . . 64

3.4 Ele tron and nu lear re oil zones standard deviations . . . 64

3.5 Thresholds . . . 68

3.6 Fidu ialvolume . . . 69

3.7 Analysis strategy and quality uts . . . 70

3.8 WIMP andidate sele tion . . . 72

4 Physi s Run: 8

th

ool down 75 4.1 Dete tor performan e and sele tion . . . 75

4.2 WIMP sear h . . . 81

4.2.1 Datasele tion . . . 81

4.2.2 Limitsonthe rossse tionforspin-independent WIMP-nu leonintera tions. . . 86

4.3 Ba kground interpretation . . . 91

4.3.1 Gammarays . . . 91

4.3.1.1 Highenergy gammarays . . . 92

4.3.1.2 Lowenergy gamma rays . . . 99

4.3.2 Alpha and beta ba kgrounds . . . 102

4.3.2.1 Response ofEDELWEIS Sdete tor toa

210

Po sour e . . . 103

4.3.2.2 Measured alpha ba kground . . . 108

4.3.2.3 Predi tion of beta leakage in the nu lear re- oilband . . . 110

Con lusions 113

A 117

One of the greatest mysteries of the universe that, for the present, puzzles the mind of most astronomers, osmologists and physi ists is the question: What makes up our universe?. This is due to how a ertain substan e named Dark Matter ame under spe ulation. It is believed this enigmati substan e,oftypeunknown,a ountsforalmostthree-quartersofthe osmos within the universe, ould be the answer to several questions raised by the modelsof theexpanding universe astronomershave reated, and even de ide the fate of the expansion of the universe.

Agreatdealofeort hasbeenmadesin e1687,whenNewtonintrodu ed the notion of gravity dis ussing it in terms of for es between bodies (i.e. visible baryoni obje ts) stating in the introdu tion of his  Philosophiae Naturalis Mathemati a that I have no regard in this pla e to a medium, if any su h there is, that freely pervades the intersti es between the parts of bodies. Sin e then, the deviations of observed motions from expe te d traje tories have proved very ee tive in deepening our understanding of universe. Whenever anomalieswere observed inthemotionofplanetsinthe Solar system, the question arose: should su h anomalies be regarded as a refutation of gravitation laws or as an indi ation of the existen e of unseen obje ts?

The modern problem of dark matter is on eptually very similar at the old one about unseen planets: we observe in large astrophysi al system at alllengths ales,fromgala ti to osmologi alone,somein onsisten ies that an only be explained either by assuming the existen e of a large amount of invisible, dark , matter, or by assuming a deviation from the well-known gravitation laws and the generalrelativity theory.

the abundan e of matter to lie in the range

Ω

M

h

2

_{= 0.1358}

+0.0037

−

_0.0036

and the abundan e of baryoni matter tobe in the range

Ω

b

h

2

_{= 0.02267}

+0.00058

−

_0.00059

[1℄, ingoodagreement with predi tions fromBig Bang Nu leosynthesis

0.018 <

Ω

b

h

2

_{< 0.023}

[2℄.

Itis ommonlybelievedthatsu hanon-baryoni omponent ould onsist ofnew,asyetundis overed,parti les, usuallyreferredtoasWIMPs(Weakly Intera ting MassiveParti les). Someextensions ofthe standardmodel(SM) of parti le physi s predi t the existen e of parti les that would be ex ellent DM andidates. In parti ular great attention has been dedi ated to andi-dates arisingin supersy mmetri theories: the Lightest Supersymmetri Par-ti le (LSP). In the most supersymmetri s enarios, the so- alled neutralino seems to be a natural andidate, being stable in theories with onservation of R-parity and having masses and ross se tionsof typi alWIMPs.

Oneway ofprobingthenatureofdarkmatterparti lesistolookfortheir annihilations signal [3℄. A wide literature exists dis ussing the prospe ts of observingannihilationradiationfromthegala ti entre, highenergy neutri-nos fromthe Sun, gamma-rays and syn hrotronfromdarkmatter lumpsin thegala ti halo, gamma-raysfromexternalgalaxies, positronsand antipro-tons and more. Large un ertainties are asso iated with predi tions of anni-hilationuxes, duetoourpoorknowledgeofthedistributionofdarkmatter, espe ially in the innermost regions of Galaxy. Other promising strategies in lude ollider sear hes for dark matter and asso iated parti lesand dire t dete tionexperimentsdesigned toobservetheelasti s attering ofdark mat-ter parti leswith nu lei.

I would like to emphasize that the dete tion of dark matter parti les in any one of the experimental hannels dis ussed will not alone be su ient to on lusively identify the nature of dark matter. The dire t or indire t dete tion of the darkmatter parti lesmaking up our galaxy's halo ishighly unlikelytoprovideenoughinformationstorevealtheunderlyingphysi s (su-persymmetry, et .) behind these parti les. In ontrast, ollider experiments mayidentify along-lived, weakly intera tingparti le,but willnot beable to test its osmologi alstability orabundan e. Weshould solve the mysteryof darkmatter parti le nature onlyby ombining the information provided by manydierent experimental approa hes.

TheEDELWEIS S ollaborationisadire tdarkmattersear hexperiment, aiming to dete t dire tly a WIMP intera tion in a target material, high purity germanium rystalworking at ryogeni temperatures. Itreliesinthe measurementofnu learre oilsthatprodu emeasurableee ts inthe rystal su hionizationand heat.

followingwork. ThenatureofDMhasbeenoneofthemost hallengingtopi s in ontemporary physi s sin e the rst eviden es of its existen e had been found in the 1930s. Cosmologists and astrophysi ists on one side, together with parti le theorists on the other have put a lot of eort into this eld: I willbrieya ountfortheira hievementsandfortheexperimentalstrategies whi h an be set in this s enario. Sin e this thesis work was arried out within the EDELWEIS S-II dire t dark matter experiment, I will fo us the next hapteron this topi , des ribing the main features.

The se ond hapter is relatedtothe set-upof the EDELWEIS S-II, the urrent stage of the EDELWEIS S experiment ne essary after a rst phase that a hieved the best upper limit on the WIMP elasti s attering on nu- leon as afun tion of WIMP mass in2004 [4℄. The set-up was on eived to redu e radioa tiveba kgroundobserved intherstexperimentphase. Thus, des ribing the starting point for this se ond stage, I will present dete tors involved in,withape uliarregardtotheGe-NTDtype,the sameimpliedin EDELWEIS S-I, onwhi h I havefo used my thesis work.

Inthethird haptertheperformedGe-NTDanalysis hainispresented. Startingwiththesignalpro essingofthere ordeddata,Iwillenterinthe es-sentialanalysissteps from alibrationsignalspassingthrough measurements ofthresholds and resolutionsinordertopredi tnu lear andele troni re oil band and denition of du ial zone to on lude determining a sele tion for likelyWIMP andidate.

These suggestions are applied in the fourth hapter , whi h presents the analysis and the results of the 8

th

Dark Matter Challenge

1.1 Theoreti al framework

In this se tion I will briey present theoreti al ingredients to make up a osmologi almodel inorder to understand why we need adark matter on-tributionin universe. In ludedshort history of universe isdrawn.

1.1.1 Standard Cosmology

Following the progress in te hnologies of experiments measuring osmolo-gi al parameters, a lot of osmologi al models have been proposed even if most osmologistsagreeonafundamentalpi ture,the Big-Bang s enario. It des ribes the universe as a system evolving from a highly ompressed state existingabout10

10

yearsago, experimentallywellmotivatedbyHubble'slaw dis overy[5℄. This modelexplainsinasatisfa torywaymost ofproperties of universe, su h as its thermal history, ba kground radiation and abundan e of elements. A osmologi almodel is omposed by three fundamental ingre-dients:

•

Einstein'sequation , relatingthe geometryofthe universewith its mat-ter and energy ontent;

•

Metri s , des ribing the symmetries of the problem;

•

Equation of state , spe ifyingthe physi alproperties of the matterand energy ontent

equationisofse onddierentialorderandlinearinthe se ondderivative[6℄. The resultingequation isgenerally written:

R

µν

+

1

2

g

µν

R = −

8πG

N

c

4

T

µν

+ Λg

µν

(1.1) where R and R

µν

are respe tively the Ri is alar and tensor (obtained by ontra tion of the Riemann urvature tensor), g

µν

is the metri tensor, G

N

the Newton's onstant, T

µν

is the energy-momentum tensor and

Λ

is the so- alled osmologi al onstant. If we ignore for a little while the term on erning the osmologi al onstant, the equation Eq. (1.1) is pretty well understood. It relates the geometry of the universe, des ribed by the left-handed side terms, to its energy ontent. It hara terized by the energy-momentum tensor on the right-handed side resulting in the well known re-lationship between matter ontent and geometry of the universe: the key on ept of general relativity.

The osmologi al onstant term, rstly introdu ed by Einstein to have a stationary solution for the universe and afterward abandoned due to the universe's expansion dis overy, represents a va uum energy more related totime-spa e itself rather than its matter ontent. It is a sour e of gravita-tionaleldevenintheabsen eofmatter. Usuallyweassumeanuniversewith propertiesofhomogeneityandisotropyassymmetryoftheproblem,madefor mathemati al onvenien e. Su hpropertieshavebeen onrmedbymany ob-servations; in parti ular,observations of the Cosmi Mi rowave Ba kground (CMB)haveshown remarkable isotropy, on ethe ontributionfromgala ti planeandthedipole omponentweresubtra ted. Isotropy aloneif ombined with Coperni anprin iple

1

would imply homogeneity. However, dire t evi-den eofhomogeneity omesfromgalaxysurveys, suggestingahomogeneous distribution inex es s at s ales of

∼

100 Mp ; it means that any > 100 Mp diametersphere enteredinanypla eoftheuniverseshould ontainthesame amount of matter.

The properties of isotropy and homogeneity and the assumption that spatial omponent of metri an be time dependent imply a spe i metri . Thelineelementinredu ed- ir umferen epolar oordinates anbeexpressed as:

ds

2

_{= −c}

2

dt

2

+ a(t)

2



dr

2

1 − k r

2

+ r

2

_dΩ

2



,

(1.2) 1

where

dΩ

2

_{= dΘ}

2

_{+ sin}

2

_Θdφ

2

and

r

,

Θ

and

φ

are the (xed) omoving oordinates arriedby the fundamental observers. In su ession,

a(t)

is the s ale fa tor and the onstant k, des ribing the spatial urvature an vary between k= -1,0, +1. Forthe simplest ase, k=0, the Eq. (1.2) omes ba k to the metri of ordinary (at) Eu lidean spa e. The Einsteinequation an be solved with this metri and its time- omponent lead to the Friedmann equation, havingthe followingform:

 ˙a

a



2

+

k

a

2

=

8πG

N

3

ρ

tot

,

(1.3)

where

ρ

tot

isthe total average energy density of the universe: it onsists in the sum of matter, radiation and va uum energy density ontribution

ρ

tot

= ρ

m

+ ρ

r

+ ρ

Λ

. In parti le physi s units,

~

= c = 1

, the Newton's onstant

G

N

has the same order of magnitude of an inverse squared mass, the Plank mass whose value is 1.22

×

10

19

GeV.

Sin e only relative hanges have measurable ee ts and thus the whole s alefa tor value isarbitrary,itis ommontointrodu eaparameterH that depends ontime a ording to the followingformula

H(t) =

˙a(t)

a(t)

(1.4)

and that manages the lo al expansion as stated by the Hubble's law,

v = Hd

,wherev isthere essionvelo ityandd thephysi aldistan e. Re ent astrophysi alobservations[7℄setsthe presentvalue ofthe Hubbleparameter to

H

0

= 73 ± 3

kms

−

₁

Mp

−

₁

(1 Mp

∼

3.1 10

22

m). We an noti efromthe Eq. (1.3), we have a at universe (k=0) when the energy density is equal to a riti aldensity,

ρ

c

:

ρ

c

≡

3H

2

8πG

N

(1.5)

Usually the abundan e of a substan e in the universe (matter, radiation or va uum energy) is normalized to

ρ

c

, so we thus dene a quantity

Ω

i

of a substan e of spe ies i and density

ρ

i

as

Ω

i

≡

ρ

i

ρ

c

(1.6)

So forthe total amount inthe universe itis normalto dene

ρ < ρ

c

Ω < 1

k = -1 Open

ρ = ρ

c

Ω = 1

k =0 Flat

ρ > ρ

c

Ω > 1

k =1 Closed

Table1.1: Classi ationof osmologi almodelsa ordingtoaverageddensity value (

ρ

) inunits of the riti al density (

ρ

c

).

that allows a re-stylingof the Friedmann equation, Eq. (1.3), asfollow

Ω − 1 =

_H

k

₂

_a

₂

.

(1.8) Thesignofk isthereforedeterminedbythevalue that

Ω

anassume(see Tab. 1.1).

1.1.2 History of the universe in a nutshell

A ording to the Big Bang model, universe originates in an explosion after whi heverysingleparti lestartedtoqui kly moveawayfromotherparti les behavinglikeahotgas offundamentalparti lesinfastexpansion. Theearly universe des ription is based on the extrapolation of the known physi s up to the Plank era at time

t ≃ 10

−

₄₃

s, that means in terms of energy the gravitational intera tion began strong at the order of Plank mass M

P l

≃

10

19

GeV. Startingat this epo h the universe evolution an bedes ribed by the followingphase transitions:

* T

∼

10

16

GeV First phase transition: The Grand Uni ation (GUT: GrandUniedTheory)epo hendsup. Thesingleuniedeldinwhi h the strong nu lear, the weak nu lear and the ele tromagneti for es were fused breaks down intoStandard Model groups. The ele troweak epo hbegan: strong andele troweakfor es (weak nu learand ele tro-magneti for e) are dierent;some models predi tthat su hepo hget underway withanhyperexpansionofthe universe (ination). Quarks andleptonsaredistin tentitiesandbosons ande ayintothemleading anasymmetry between matterand anti-matter.

* T

∼

10

2

* T

∼

10

1

-10

3

GeV Quark-antiquark (

q

-

q

¯

) annihilation o urs: weakly intera ting dark matter andidates with GeV-TeV s ale masses freeze-out.

* T

∼

0.3 GeVQCDphasetransition: onnement ofquarksand gluons intohadrons.

* T

∼

1 MeV Neutrinosde oupling,neutrons freeze-outand e

+

-e

−

anni-hilationso ur.

* T

∼

100 keV Nu leosynthesis: protons and neutrons fuse into light elements (D,

3

He,

4

He, Li). The Standard Big Bang nu leosynthes is provides by far one stringent onstraints to the Big Bang theory and predi tionsagree with observations, see Fig. 1.7.

* T

∼

1 eV Matter and radiation densitiesare equal: the stru ture for-mationbegins.

* T

∼

0.4 eV Photon de oupling produ es osmi ba kground radiation (CMB).

* T=2.7K

∼

10

−

₄

eV Today.

1.2 Motivations and eviden es for Dark Matter

The rst interest in a matter whi h does not emit radiation having no ou-plingwithphoton(dark)andthus an beobserved onlyby itsgravitational ee ts dates ba k toÖpik's 1915 studies about dynami almatterdensity in the Solar vi inity [10℄. The urrent onnotation of dark matter appears in 1933,intherstZwi ky'sworkonthedynami sofgalaxiesintheComa lus-ter [11℄. Sin e then astrophysi al eviden es of the presen e of some mass ex es s with respe t tothe visiblefra tion a umulate throughout the ages, while osmologistsandparti lesphysi ists seektotobservationsintoa the-oreti al stru ture. The old dark matter s enario was then established as a likely osmologi alexplanation [12℄, while on the parti le side the debate is still wide open. Here I outline the main dark matter eviden es from an observational pointof view [13℄.

1.2.1 The gala ti s ale

ir-Figure1.2: Rotation urveof NGC6503. Figurefrom [14℄.

(Fig. 1.3 gives a visual explanation of this eviden e). A simple Newtonian approa h gives:

v(r) =

r

GM (r)

r

(1.9) where

M (r) ≡ 4π

R ρ(r) ˙r

2

_dr

and

ρ(r)

is the mass density prole that should befalling

∝ 1/

√

r

beyond the opti al dis . If the matterdensity was given only by the known visible mass, one would expe t the velo ity to fall like

1/

√

r

outside the gala ti disk. Instead, experimental data show that the velo itykeeps a onstant trendeven farbeyond the visibledisk,probing the existen e of a dark matter halo with

ρ(r) ∝ 1/

√

r

2

. The density prole ofthe innermost partof the spheri al haloisstillunknown. Howeveradding a darkmatter halo allows a goodt todata.

The standard method used to measure masses is to balan e the kineti and potentialenergies using the virialtheorem

2 .

Limitationsofrotations urvesarise fromthefa t thatone an onlylook out as faras thereis lightor neutralhydrogen (21 m), namely todistan es oftensofkp . Therefore, we anseethe beginningsofDM halos,but annot tra ewheremostofDMis. Lensingexperimentsgobeyondtheselimitations,

2

Figure 1.3: Explanation of the rotation urves of galaxies (http://phili a. om/uploads/images/145/Image/image006.jpg).

see next se tion.

1.2.2 Galaxy luster s ale

hT i = −

1

₂

hV i

(1.10) where

hT i

is the average kineti energy derived from the dispersion in the velo ities and

hV i

is the average potential energy. The latter is used to determine the mass of the luster.

The X-rayemissionofhot intra- lustergas, assumedtobeathydrostati equilibrium, an also be used to estimate the mass of luster [17℄. X-ray proles of the gas are measured and then t to temperature and density distribution models todetermine the mass of the luster.

The mass of a luster an also be determined by gravitational lensing method [18℄: general relativityin fa t states that ompa t gravitational bo-dies bend nearby photon paths (namely, make a spa e-time geodesi urve) andthusa taslensesforlightsour esbehindtheminthelineofsight. Thus, when lightrays pass through large gravitationalmasses su h asgalaxy lus-ters they are dee ted by the gravitationaleld produ ed by the luster, in a manner similar to the way an opti al lens bends light to form an image. By analyzing the amount of bending of light, we an be able to determine the luster mass. The estimates show that there is far more mass exerting gravitationalee tthansuggestedbyluminousmatter. Figure1.4illustrates the gravitationallensee t produ edbyCluster0024+1654 asseenthrough Hubble Spa e Teles ope.

1.2.3 Cosmologi al s ale

Theexperimentaleviden epresentedearlierindi atesthatasubstantialpart ofthe universeismadeup ofanonluminous omponent. By onsideringthe onstraints set on the amount of luminous matter in the universe based on astronomi alobservations of galaxiesand galaxy lustersalone, we on lude that

Ω

lum

∼ 0.005

. Measurements of lusters are onsistent with a matter density of the universe

Ω

m

= 0.341

+0.031

−

_0.029

[19℄. In summary, the eviden es are overwhelming for the existen e of an unknown omponent of DM that omprises 95% of the mass in galaxies and lusters.

Figure 1.4: Image of galaxy luster 0024+1654 taken by the Hubble Spa e Teles ope demonstrates gravitational lensing by large galaxy luster. Light fromdistantgalaxiespassesbythegravitational luster'smassandgetsbent, reatinga lensingee t. Figure from[18℄.

The Cosmi Mi rowave Ba kground (CMB) Further eviden e for darkmatter omesfrommeasurementson osmologi als ales ofanisotropies intheCMB. TheCMB istheremnant isotropi radiationfromthehot early days of the universe.

By studying this radiation, we are able to examine the onditions in the universe 400000 years after the Big Bang, known as the surfa e of last s attering. Though the virtual isotropy of the CMB isbetter thanone part in 10

5

, there exist tiny ripples in the temperature of mi rowave sky whi h provide us information about the seed u tuations that existed at the time of the de oupling of matter and radiation, mu h prior to the formation of galaxy stru tures. These seed u tuations grew by gravitationalattra tion into the stars, galaxies and galaxy lusters we see today. Measurements of the anisotropies in the CMB may thus be employed to determine various osmologi alparameters.

Figure 1.5: Full sky maps of the osmi mi rowave ba kground anisotropy. Figurefrom http://map.gsf .nasa.gov/.

of the de oupling of matter and radiation, these pressure waves left tra es oftheir existen einthe temperatureu tuations seeninthe CMB radiation today, see Fig. 1.5.

Cosmologi al models predi t the existen e of a ousti peaks in the an-gular power spe trum, see Fig. 1.6. The relative position and height of the peaks provide estimates on

Ω

and

Ω

m

. Studies of the peaks in the CMB angular anisotropy power spe trum by experiments su h as DASI [21℄ and MAXIMA [22℄ point to

Ω

m

+ Ω

Λ

≃ 1

, suggesting a at universe. This has been orroborated by the Wilkinson Mi rowave Anisotropy Probe laun hed in 2001 to study the mi rowave ba kground sky with unpre edented a u-ra y. From WMAPdata ombinedwith the distan e measurements ofType Ia Supernovae (SN) and Baryon A ousti Os illations (BAO) in the galaxy distribution,wend(at68%C.L.. un ertainties)thefollowingvalues forthe abundan e of baryons and matter [1℄:

Ω

m

h

2

_{= 0.1358}

+0.0037

−

_0.0036

; Ω

b

h

2

_{= 0.02267}

+0.00058

−

_0.00059

(1.11)

thatistosay,roughlyspeaking,thatordinarymatter(baryons)a ounts only for1/6 of the total matter density inthe universe, the other 5/6 being attributed to dark matter. WMAP ollaboration derives also the value of

Ω

m

= 0.27 ± 0.04

and of the va uum energy density

Ω

Λ

= 0.73 ± 0.04

ompatiblewithaatuniverse. Thevalueof

Ω

b

h

2

thusobtainedis onsistent with predi tions fromBig Bang Nu leosynthesis [2℄.

0.018 < Ω

b

h

2

_{< 0.023}

Figure 1.7: Evolution of light elements abundan es over time (or tempe-rature). When the universe has ooled su iently, the light elements are not disso iated by theenergeti photons. Nu leosynthesis pro eedsuntilthe supply of free neutrons is exhausted. Figurefrom [25℄.

Primordial Nu leosynthesis To determine the nature of most of this dark matter, we turn to Big Bang Nu leosynthesis. This term refers to the al ulations of the abundan es of the light elements su h as

2

_H

,

3

_H

,

4

_He

and

7

_Li

relative to photons within the framework of the Big Bang model des ribing the universe [23℄, [24℄. Less than one se ond after the Big Bang, theneutron-to-protonratioismaintainedinthermalequilibriumthroughthe followingrea tions:

p + e

−

←→ n + ν

(1.13)

Figure 1.8: The predi ted abundan e of the light elements as a fun tion of baryon density. The verti al band indi ates the narrow range of baryon densities onsistentwiththedeuteriummeasurements;theboxes(thearrows for

3

_He

)indi atethe rangeinbaryondensity(horizontalextentofbox)that is onsistent with the measured light-element abundan e (verti al extent of box). Theoverlap ofthe boxes withthe deuteriumand indi atesthe general onsisten yoftheobserved abundan esoftheotherlightelementswiththeir predi ted abundan esfor this baryondensity. Figurefrom [25℄.

menally hot

10

32

Kelvin to

10

9

Kelvin, below the nu lear binding energies, suppressingthenumberofphotons withenergies highenoughtodisasso iate these nu lei. Light elements begin to form. Figure 1.7 gives the evolution of the abundan esof lightelements overtime (and temperature) inthe rst fewminutes after the Big Bang.

BythetimethetemperatureoftheuniversefelltoT

≈

0.003MeV,thelight elements

4

_He

,

3

_He

,

3

_H

,

7

_Li

and

7

_Be

an, therefore, pla e limits on the baryoni density. In reased a ura y in astrophysi al measurements havepla ed tighter onstraintson the baryoni matter-density. These in lude studies of quasars [25℄ and the abundan e of

2

_H

in high redshift louds [26℄. Current measurements orroborate the WMAP results onstraining the baryoni matter density in the range de-du ed by Eq. (1.12).

Given our knowledge of the matter density

Ω

m

≃ 0.3

, this suggests that most ofthe matter inthe universe isnon-baryoni innature. Thedensity of ordinarybaryonswithinanarrowrangeisthepredi tedprodu tion onsistent withwhat isa tuallymeasured,see Fig.1.8. BBN theoryand baryoni dark matter density measurements, ombined with CMB measurements, suggest that non-baryoni dark matter is an important omponent of matter inthe universe.

DarkEnergy Eviden eforthe70%darkenergyintheuniverse omesfrom observations of distant supernovae ([27℄ and [28℄). The further supernovae are dimmer than expe te d, as is most easily explained by an a elerating universe. There are twodierent approa hes tothe darkenergy:

•

a va uum energy su h as a osmologi al onstant or time dependent va uum may be responsible [29℄;

•

itis possible that GeneralRelativity isin omplete and that Einstein's equationsneed to be modied([30℄and [31℄).

Note, however, that this dark energy does not resolve or ontribute to the question of darkmatter ingalaxies.

1.3 The WIMP Hypothesis

In this thesis, I limit the dis ussion to dark matter andidates whi h are heavy, ele tri ally neutral and weakly intera ting. This lass of parti les, known as WIMPs, is parti ularly well motivated by their thermal history. At su iently early times after the Big Bang, when the temperatures are greater than the mass of the parti le,

T ≫ m

χ

, the equilibrium number density of su h parti les is

n

χ

∝ T

3

, but for lower temperatures,

T ≪ m

χ

, the equilibrium abundan e is exponentially suppressed,

n

χ

∝ exp

−

_m

_χ

_/T

Figure 1.9: A thermal reli starts in LTE (Lo al Thermodynami Equilib-rium) atT  m

X

. When the rates keeping the reli in hemi alequilibrium be ome smaller than the expansion rate, the density of the reli relative to the entropy density be omes onstant. This is known as freeze out . [33℄

At high temperatures (

T ≫ m

χ

),

χ

parti le is abundant and rapidly onverting to lighter parti les and vi e versa (

χχ → ll

, where

ll

are quark-antiquarkand lepton-antileptonpairs, and if

m

χ

isgreaterthan the mass of the gauge and/or Higgsbosons,

ll

ould be gauge and/orHiggs boson pairs aswell). Shortlyaftertemperature

T

dropsbelow

m

χ

,thenumberdensityof

χ

drops exponentially and the rate

Γ = hσvi n

χ

for annihilation of WIMPs, where

hσvi

is the thermalaverage ross se tion

σ

forannihilationof

χχ

into lighter parti les times relative velo ity

v

, drops below the expansion rate,

Γ ∼ H

.

Ω

χ

h

2

=

m

χ

n

χ

ρ

c

≃ 0.1(

3 × 10

−

₂₆

cm

3

_sec

−

₁

hσvi

)

(1.15)

The result is to a rst approximation independent of the WIMP mass and is xed primarily by the annihilation ross se tion.

TheWIMPvelo itiesatfreeze-outaretypi allysomeappre iablefra tion of the speed of light. Therefore, from the Eq. (1.15), the WIMP will have a osmologi alabundan e

Ω

χ

h

2

_{∼ 0.1}

today if the annihilation ross se tion times

c

is roughly 3

×10

−

₂₆

m

3

se

−

₁

, orin parti le-physi sunits (obtained using

~

c = 2 × 10

−

₁₄

GeV-fm),

10

−

₈

GeV

−

₂

. Curiously, this is the order of magnitude one would expe t froma typi alele troweak ross se tion,

σ

weak

≃ (

α

2

m

2

weak

)

(1.16)

where

α ≃

O(0.01)isthene-stru ture onstantand

m

weak

≃

O(100GeV). The numeri al onstant in Eq. (1.15) needed to provide

Ω

χ

h

2

_∼

0.1 omes essentially from the expansion rate (whi h determines the riti al density). This relationbetween the expansion rate and the ele troweak s ale suggests that if a new, as yet undis overed, stable massive parti le with ele troweak intera tions exists, it should have a reli density suitable to provide a good andidate for the darkmatter.

This has been the reason en ouragingthe massiveexperimentaleort to dete t WIMPs.

1.3.1 Parti le Candidate

The undeniableexisten e of some kindof substan e whoseee ts are, as far as we know, onlygravitational in nature,brought immediately the question ofitsparti le omposition. Namely,the darkmatterparti le andidatemust showthe followingobserved proprieties [34℄:

1. It must have extremelyweak ornoele tromagneti nor strong intera -tions. As a onsequen e, darkmatter annot oolbyradiatingphotons and thus, unlike baryons, does not ollapse to the enter of galaxies. In other words, one ould state that dark matter is very nearly dissi-pationless.

small ones ( top-down formation), whi h ontradi t the astrophysi al data.

3. It must a ount for the measured density

Ω

m

.

A wealth of WIMP andidates has been proposed, fromStandard Model neutrinos tothe most exoti ones. In fa t,while astrophysi al and osmolo-gi al onstraintsare moreorless denite, thegreatun ertaintiesaboutwhat dire tiontotakeinordertogobeyond thelimitsofStandardmodelof Parti- lemakeroomforfantasy. Iwillfo usmyattentiononaverylimitednumber of darkmatter andidates, ons ious anyhow that every hoi e issomewhat arbitrary, and that the great favor en ountered today by the neutralino and othersupersymmetri parti les, ouldvanishsuddenlyiftomorrowanewbig dis overy points elsewhere.

Neutrino. The rst WIMPs onsidered were massive Dira neutrinos (parti le dierent from its antiparti le) or Majorana neutrinos (parti le o-in identwith itsown antiparti le)withmasses inthe rangeof afew GeV to afewTeV; duetothe Yukawa ouplingwhi hgivesaneutrino itsmass, neu-trinointera tionsbe omestrongaboveafewTeV,andtheneutrinonolonger remainsasuitableWIMP andidate[35℄. TheLEP(LargeEle tron-Positron) olliderex ludeneutrinomassesbelowhalfZ

0

mass. Moreover, heavierDira neutrinoshavebeen ruledout astheprimary omponentoftheGala ti halo by dire t dete tion experiments [36℄, and heavier Majorana neutrinos have been ruled out by indire t dete tion experiments, e.g. [37℄ and [38℄, over mu h of their mass range. Therefore, Dira neutrinos annot omprise the halo dark matter [39℄; Majorana neutrinos an, but only overa small range of fairly largemasses.

Supersymmetri Candidates Amu hmorepromisingWIMP andidate omes from ele troweak-s ale supersy mmetry [40℄ (SUSY). SUSY was hy-pothesized in parti le physi s to solve Standard Model in onsisten ies su h asthenaturalnessproblemwithfundamentalHiggsbosonsattheele troweak s ale. In theGUTtheory,the parameterthat ontrols theHiggs-bosonmass mustbeextremely small,butitmaybe losertounity(thatitmeansin par-ti le theory idiom more natural) in supersymmetri theories. Uni ation ofthe strong and theele troweak oupling onstant atthe GUTs aleseems to need SUSY and SUSY seems to be indispensable in theories that unify gravity with the other three fundamental for es.

bosonsandfermionsare oupledin ommonmultiplets. Everyknownparti le isthenprovidedasuperpartner withthesamequantumnumbers,ex eptspin whi hdiers for1/2. Sin e nobosons with the same harge and mass ofthe ele tronnor anyothersuperpartnershaveeverbeen observed, itis lear that SUSY is broken in the low energy world in whi h we are supposed to live, and that supersy mmetri parti les must have masses above urrent lower bounds (

∼

100 GeV). Anyhow, this s enariointrodu es asevere imperfe tion allowinga supersymmetri parti le to mediate a qq

→ ˜l˜q

pro ess, providing thus an e ient hannel for proton de ay. Current experimental limits on proton lifetime are however on the order of 10

33

years. Therefore, it has been proposed to add to the SUSYa new dis rete symmetry, the R-parity, to distinguish ordinary parti les (R=+1) from their superpartners (R=-1). If R-parity holds with a broken SUSY, supersymmetri parti les an de ay only inan odd number of superpartners, plus ordinary parti les, preventing thustheprotonfromde ayingandguaranteeingthelightestsupersy mmetri parti le(LSP) tobe stable.

Wewill onsider the MinimalSupersymmetri extensionof the Standard Model (MSSM) that ontains the smallest possible number of elds to give rise orre tlytotheStandard Modelwhenthe symmetryisbroken. Without entering intodetails, here we listthe populations inthe MSSM

•

allordinary quarks with their spin 0superpartners ( squarks

q

˜

);

•

allknown leptonsand theirbosoni ounterparts ( sleptons

˜l

):

•

all gauge bosons (gluons, W

i

and B) and their fermioni partners (re-spe tively gluinos, Winos and Binos) ommonly alledgauginos;

•

the standard Higgsbosons, anadditionalHiggs doublet with opposite

hyper harge and a ouple of spin 1/2 Higgsinos.

Considering mass eigenstates, ele troweak gauginos mix into eight die-rentstates: the harged partsofWinos and Higgsinosappearastwo ouples of harginos (

χ

±

1

,

χ

±

2

), while Bino,the neutral harge Wino and the neutral states of the Higgsinos formfour neutralinos (

χ

0

1,2,3,4

).

Axinoand Gravitino. Theyaresuperpartnersofparti lesintrodu ed inextensionsofStandardModelandhen etheyarenotpresentintheMSSM frame. The spin 1/2 partnerof axion, the axino , and the spin 3/2 gravitino , superpartners of the unseen graviton, the gauge boson that mediates gravi-tationalintera tion, showsimilarphenomenologyas WIMP andidates. De-pending on the SUSY model adopted and on the early universe onditions, axino or gravitino an be the LSP, although their lightness ould originate rathera warm dark matter.

Lightest Neutralino. The lightest ofthe fourneutralinos,

χ

0

1

,usually referred to simply as the neutralino (

χ

), is at the present day regarded as the most suitable WIMP andidate. The features of su h a parti le, om-pletely developed in a parti le physi s framework, t well the astrophysi al onstraintsfordarkmatter,withoutneedforanyadho hypothesis. Itsmass an range from 150 GeV (ele troweak s ale) to several TeV. Being the LSP (Lightest Supersymmetri Parti le)of the theory,it isstable, sin e R-parity onservationprohibitseveryde aypro essotherthanself-annihilation. Neu-tralinohas a quitelowannihilationrate and itis heavy enoughto represent agooddark matter andidate.

1.4 DM Dete tion

Weakly Intera ting Massive Parti les (WIMPs) are regarded as the most naturaldarkmatter andidate. Theynotonly anformaba kgrounddensity intheuniverse,buttheywillalso lustergravitationallywithordinarystarsin thegala ti halos. Inparti ulartheywouldbepresentinourowngalaxy,the MilkyWay,raisingthehopeofdete tingreli WIMPsdire tly,byperforming experiments on the Earth through elasti s attering of WIMPs o target nu lei or indire tly, looking for by-produ ts of WIMP-WIMP annihilation that o urs either inthe Sun orgala ti halos.

1.4.1 Indire t Sear h

propor-it means that natural interesting zones for sear hing signi ant uxes are regions hara terizedby high darkmatter density.

High density region of gala ti halo su h as the gala ti enter ould be good ampliers to dete t WIMP annihilation produ ts su h as antimatter parti les and photons. Astrophysi al obje ts likeSun and Earth ould be a good dark matter amplier to observe neutrinos produ ed by DM parti les s attering o nu lei onSun and Earth. In this se tionI des ribe the role of theseexperimentalprogramsinthestrategytorevealidentityofdarkmatter.

1.4.1.1 Gamma Rays

Traveling essentially unimpeded from their produ tion site, the photons ge-nerated in dark matter annihilation have an advantage over the other in-dire t dete tion hannel. In parti ular, gamma rays are not dee ted by magneti elds, potentially providing useful angular information, retaining theirspe tral information: thespe trumobserved onEarth isthe same that wasgeneratedinthe darkmatterannihilation. Theprospe tsforidentifying dark matter annihilation radiation from the Gala ti Center (GC) strongly depends on the WIMP nature, on the unknown dark matter density in the region around the GC and on our understanding of the astrophysi al ba k-grounds.

Theteles opes, potentially apableofdete tinggammarayradiationfrom WIMP annihilation in ludethe satellite-based experimentGLAST [41℄,[42℄ and lots of ground based Atmospheri Cerenkov Teles opes su h as HESS, MAGIC and VERITAS. GLAST will ontinuously observe a large fra tion of the sky, but with an ee tive area far smaller than the one possessed by ground-based teles opes. In ontrast, groundbased teles opes study the emissionfromasmallangulareldbutwithfargreaterexposure. Thegamma rays energy range is also dierent, while GLAST is able to dire tly study gammarayswithenergiesovertherangeof100MeVto300 GeV,theground based teles opes are onlysensitive to gamma rays with energy greater than

∼

100 GeV. As a result of this dierent energy ranges a essible by these experiments, sear hes for WIMPs lighter than a few hundred GeV are most promising with GLAST while ground based teles opes are better suited for heavierWIMPs.

1.4.1.2 Antimatter

and losingenergy resulting inadiuse spe trumonEarth. Bystudying the osmi anti-matterspe tra,satellite-basedexperimentssu hasPAMELA[43℄ and AMS-02 [44℄ may be able to identify signatures of dark matter. As ompared to antiprotons and antideuterons, osmi positrons are attra tive probes of dark matter: the spe trum samples only the lo al dark matter distribution and is thus subje t to onsiderable un ertainty than the other anti-matterspe ies.

UnlikegammaraymeasurementsoftheGala ti enter ordwarfgalaxies, observationsof the osmi positronspe trum(as wellasthe antiprotonsand antideuterons spe tra) ould potentially provide ameasurementof the dark matterannihilationrateoverlarge volumesof spa e. Therefore, su ha mea-surement ouldbeusedtodeterminetheprodu tofthe WIMPsannihilation ross se tionand itsdensity squared,averaged over the sampledvolume.

1.4.1.3 Neutrinos

Althoughdarkmatterannihilationsinthe gala ti haloprodu etoofew neu-trinostobe dete ted [45℄,annihilationswhi ho ur inthe enter ofthe Sun ould potentially generate anobservable uxof high energy neutrinos [46℄.

Darkmatterparti less atterelasti allywithand be ome aptured inthe Sun. WIMPs an generate neutrinos through a wide range of annihilation hannels. Annihilationstoheavyquarks,tauleptons,gaugeandHiggsbosons an generate neutrinos in the subsequent de ay. On e produ ed, neutrinos antraveltotheEarthwherethey anbedete ted: muonneutrinos produ e muons in harged urrent intera tion with i eor water nu lei inside ornear the dete tor volume of high energy neutrino teles ope.

ExperimentslikeMACRO[37℄,AMANDA[47℄andSuper-Kamiokande[38℄ set upperlimitsonneutrino uxes omingfrom the enter of Earth and the Sun. Super-K upper limit on neutrino-indu ed muons above 1 GeV from WIMP annihilations in the Sun is approximately 1000 to 2000 km

2

_·

y for WIMPs heavier than 100 GeV, and approximately 2000 to 5000 km

2

_·

1.4.2 Dire t Sear h

Whatkindofassumptionsweneedtosu eedinadire tdarkmattersear h? All that we have to know for dire t dete tion is that the Galaxy ontains a halo of WIMPs normally assumed tobe of spheri al isothermal form with a lo al density 0.3 GeV

·

−

₂

·

m

−

₃

, anes ape velo ity of 650 km

·

s

−

₁

, with rms velo ity 279 km

·

s

−

₁

and relativehalo-Earth velo ity of 235 km

·

s

−

₁

[50℄. The dire t sear h prin iple repose on the elasti s attering of these neu-tral, non-relativisti parti les(WIMPs), otarget nu lei ofa suitable dete -tor, su h that the energy transferred as the resulting nu lear re oil passes through the material an be observed, usually as either ionization, s intilla-tionorheat(phonons). Kinemati sandthe likelymassrangeand velo ityof the parti leimpliesanu lear re oilspe trumwith energies below

∼

100 keV, with an exponential form rising to low energies and with no spe tral fea-tures. This hara teristi s, together with the expe te d lowintera tions rate ofabout1-10

−

₆

eventkg

−

₁

·

d

−

₁

,imposethreeessentialrequirementsofWIMP dete tor te hnology: low energy threshold (

≤

10 keV

recoil

); large dete tor mass (> 10 kg) and low parti le ba kground of any kind. The latter is arried out by passive and a tive gamma and neutron shielding, by using materials with greatly redu es radioa tive U, Th and K ontent during the dete tor onstru tion and preferring deep underground sites to redu e os-mi raymuon-indu edneutrons that ould otherwiseprodu enu learre oils indistinguishable fromWIMP.

The oupling of these non-relativisti WIMPs has two terms: a s alar Spin-Independent(SI)partandanaxialSpin-Dependent(SD)part[51℄. For most SUSY models, SI provides the dominant oupling and thus highest rate. This is be ause although neutralino-nu leon ross se tion are substan-tially mu h smaller for SI ase [52℄, oheren e a ross the nu leus results in onstru tive interferen e whi h greatly enhan es the WIMPs-nu leus elasti rossse tionfor highAtargets. TheoppositeisrightforSD wheretheaxial ouplingtonu leus whi hdiers for spininterferesdestru tively: sensitivity toSDintera tionsrequires atargetisotopewith anunpaired nu leon,either proton orneutron.

Knowingthat, forinstan e, typi alambientenvironmentalgammauxes an produ e event rate >10

5

perature ionization/phonon or s intillation/phonon dete tors in whi h the ratio event-produ ed ionization or s intillationto phonons is measured in a suitable ryogeni materials su h asGeor Si(ionization)and CaWO

4

(s in-tillation); noble liquid gases, notably Xenon and Argon, in whi h ionization and s intillationare measured simultaneously. A moderatelevelof dis rimi-nation an be also a hieved in spe i s intillators su h as NaI(Tl), Cs(Tl), liquidAr and liquidXe.

Sin ethe re oildis riminationandthe ba kgroundredu tion seemstobe feasibleinsu hte hnologies, the mainissue thatremainstodeal with, given the la k of spe tral featuresin the re oil spe trum, is howto determinein a learway whether any remaining ountsare due to WIMPs fromthe galaxy and not either nu lear re oils form an una ounted ba kground (neutrons and surfa e events) or a dete tor artifa t. As WIMP intera ts weakly with ordinarymatter,nomultipleintera tionsare expe te d,informationthat an be used in order to identify neutron, sin e these latter reate as well as a WIMP nu lear re oil. It is alsoknown that WIMP event rate is modulated with a maximum value in June and a minimum in De ember. Some exper-iments use this signature. In fa t, at least for a standard halo model, the Earth's motionthrough Galaxyimpliesan expe ted seasonal modulationin the re oil spe trum (shape and ux) [53℄, [54℄. This is be ause the om-ponent of the Earth's solar orbital velo ity in the dire tion of our gala ti motion (orbital plane in lined at 60

o

) either adds to or subtra ts from the gala ti orbital velo ity dependingon the season. Unfortunately the annual modulationee t is very small,typi ally afew %, requiring already atleast ton-s aledete tors toobtainsu ientevent statisti foraviablesear h[50℄. Besides the annual modulation there is also a diurnal modulation proof of thegala ti originofthe signal: thanks toourgala ti orbitalmotion(

∼

235 km

·

s

−

₁

), we would expe t the dire tion of nu lear re oil indu ed by WIMPs intera tion within a target to be dominantly opposite to our dire tion of motion(in gala ti oordinates), [55℄, [56℄.

1.4.2.1 Dete tion Te hniques

Semi ondu tors Ionization dete tors, in the form of low ba kground germanium (HPGe) and sili on diodes used for double beta de ay sear hes, provided the rst limits on WIMP intera tions [57℄. Su h experiments were vital to ruling out early andidates for WIMPs but as a te hnology they suer from the in apability to dis riminate gamma ba kground events from nu lear re oilevents of interest.

S intillators Whenaparti leintera tswithas intillating rystal,light is being emitted with photon numbers that are proportional to the energy of the in oming parti le. After, light signal is dete ted by photomultipliers thatamplifythelightand onvert itinanele tri signal. TheDAMA exper-iment, lo ated in the Gran Sasso Underground Laboratory, using nine low ba kground 9.7kg NaI(Tl) rystals found eviden eforamodulation, report-ingthedis overyoftheWIMPin1997[58℄. Thissignalwas onrmedbynal DAMAresultsfromatotalof107.73kg

·

d[59℄remainingtheonly laimed di-re tobservationofWIMPs, orrespondingtoaWIMPmassofabout50GeV anda rossse tiononprotonof7.2

×10

−

₆

pb. However, thisresultappearsin ontradi tionwithseveralotherexperiments,infa tthiskindofWIMPmass and ross se tion have been ex luded rstly by EDELWEIS S [60℄ and then CDMS[61℄. Nonstandard-WIMPmodels[62℄,su haslightneutralinos,have beeninvestigatedinordertore on ileDAMA laimswiththelimitsfromthe sear hes, but WIMPs with masses below urrent limits are ex luded by the high-resolution germanium experiment CoGeNT [63℄. In addition, assum-ingWIMPswith dominantspin-dependent rossse tiononprotons ontrast with the limits fromCOUPP [64℄ and KIMS[65℄ experiments.

More re entlyother inorgani s intillatorslikeCsI(Tl) and CaWO

4

have been employed inthe darkmatter sear h. The KIMS experiment [66℄, inan underground laboratory in the South Korea uses a CsI s intillating target insteadofaCaWO

4

rystalbe omeanintegralpart oftheCRESST bolome-tri experiment in whi h s intillation light is measured simultaneously with heat [67℄. The event by event dis rimination is ensured, in this ase, by the fa tthat nu learre oils have mu hsmallerlightyield than ele troni one.

For these dire t dark matter te hnology the main disadvantage is repre-sented by the la k of apowerfulre oil dis rimination.

DRIFT[56℄,MIMAC[68℄andotherlowpressure gasTimeProje tChamber (TPC) R&D programs [69℄. As urrently the only known route to signi- antre oildire tion sensitivity, TPCte hnology holdsex eptional powerfor WIMP physi s and possibly the only route to a denitive gala ti signal. Howeverthereareseveral hallengestoaddress,su hastheneedoflow pres-sure gas implying large volume dete tors and the desirability of a hieving tra k head totaildis rimination.

Bolometers Atlowtemperaturestheheat apa ityofadiele tri rys-talgoesasT

3

. ThusatmK temperatures thesmallenergy depositionfroma nu learre oil an yieldameasurableproportionalin rease in rystal tempe-rature[50℄. Thisisstartingidea forthe earliestdire tdarkmattersear hing te hniques, where energy released by parti le intera tions an be observed as phonons or quanta of latti e vibrations. However it was demonstrated, rst in Si [70℄ and the in Ge [71℄ that phonon dete tion ould be ombined with simultaneous dete tion of ionization to provide also anevent by event dis riminationagainstele tronre oils. Infa t,be auseofthe dependen e of the proportionof energyobserved inthetwo hannelsonthe eventdE/dx, a high dE/dx event, su h as a re oilingnu leus produ es proportionally more heat than ionization (the ionization is quen hed). For instan e, the ratio of ionizationtore oil energy, alled the ionization yield,forGe re oilsin Geis

∼

0.3 of the value for ele tron re oils above20 keV [61℄.

Whilethe simplephonons dete tion, thatmeans bolometers without ol-le tion of ionization have proven quite useful for dark matter sear hes, this hybrid te hnique of simultaneousionizationand phononmeasurements with its apability for ba kground reje tion has been pushed harder. Most no-table is CDMS, [72℄,[73℄, (at Soudan mine) and EDELWEIS S ollaboration (at Frejus), [4℄, [60℄.

The EDELWEIS Sexperimentwillbe ompletelydes ribed inChapter. 2 being the obje t of this thesis work.

je ting surfa e ele tron events that otherwise ontaminatethe signal region. It ounts 15Gedete tors (3.75kg) with anee tive exposure of 121.3 kg

·

d, averagedoverre oil energies of 10-100keV. A blindanalysis resulted inzero observed events, yieldinga90% C.L. spin-independent upperlimitsinGeof 6.6

×

10

−

₈

pb(6.6

×

10

−

₄₄

m

2

)foraWIMPmassof60GeV/

2

[74℄. The ba k-grounddue to surfa e events with bad harge olle tion forthat experiment wasestimated tobe 0.6

±

0.3 events.

As analternative,ROSEBUD [75℄ and CRESST [67℄have developed de-te tors in whi h, instead of ionization signal, s intillation light is measured in oin iden e with heat, in parti ular using CaWO

4

rystal [67℄. Here a sili on wafer with tungsten thermometer is used to dete t the photons and a super ondu ting evaporated lm used as the heat sensor. Although only 1%orless of theenergy depositedisdete ted asphotons thisismu hhigher thanfeasibleatroomtemperatureandissu ienttoprodu eenergy resolu-tion omparable to NaI(Tl) rystal. Results so far have been obtained with two 300 g rystals at the Gran Sasso Underground Laboratory with a total exposure ofabout 20 kg

·

drevealing 16events inthe energy range 12-40keV onsistentwith the expe te d neutron ba kground given that the experiment isnot s reened by a neutron shield.

All these ryogeni experiments are now progressing toward signi ant upgrades for instan e CDMS is proposing 25 kg and a possible move to the deeperSNOLABsiteandCRESSTisupgradingtoallow33CaWO

4

dete tors totaling10kg. However, asoutlinedinfollowingparagraphs,itis likelythat even greatertarget mass willbe needed, possibly atthe ton-s ale orlarger.

Liquid noble gases Liquid noble gas te hnology for WIMP sear hes hashadare entrapidgrowth. Mostnotablehasbeenliquidxenon(LXe),by DAMA/Xe[76℄and[77℄,but alsore entlyliquidneon[78℄and,inparti ular, liquidargon.

s intillationrelative toele tron re oils of the same energy.

Ex iting progress has been made re ently with the two phase LXe te h-nologywithbothZEPLINIII[81℄(Boulby,UK)andXENON10[82℄(LNGS, Italy) announ ing new leading limits. ZEPLIN-III onsists in 12 kg of two phase xenon; an analysis of 847 kg

·

d of data has ex luded a WIMP nu- leon elasti s attering spin-independent ross se tionabove 7.7

×

10

−

₈

pb at 55GeV/

2

WIMPmass witha90%C.L.. The greatXENON10advantageis averylowenergy thresholdallowingtodis riminatesignalfromba kground down to4.5 keV nu lear re oil energy. A blind analysis of 58.6 kg

·

dof data ex ludes previously unexplored parameter spa e, setting a90% C.L.. upper limitforthe WIMP-nu leon spind-independent ross se tionof 8.8

×

10

−

₈

pb for a WIMP mass of 100 GeV/

2

and 4.5

×

10

−

₈

pb for a WIMP mass of 30GeV/

2

.

Tonne-s ale idea WIMP experiments with target masses of kg-s ale arerea hingsensitivitiesprobestartingtowellintoSUSYfavoredparameter spa e.

It is pretty ertain, that favored spin-independent oupled dark matter doesnotexistwith rossse tions>

∼

2

×

10

−

₇

pb. Meanwhile,theoreti al pre-di tions for a neutralino-like WIMP rea h ross se tion value smaller than 10

−

₁₁

pb[83℄,[84℄. Hen e,nextgenerationexperimentsmustnotonlya hieve furtherba kgroundsuppressionbutalsobe apableofton/multi-tonmasses, simplytoensureastatisti allyobservablesignalrate. Se ondly,forsu hlarge dete torsit anbearguedthat thougha tivegammadis riminationremains important, greater emphasis is needed on materialpuri ation, passive and a tive shielding of radioa tive ba kground and on sear hes for additional features in the data showing that remaining events, in parti ular, are not neutrons. In fa t,if weassume that the experimentis situateddeep enough tos reenmuon-indu edneutrons,gammaandneutronsfromU/Th hainsin theenvironmentanddete tor willdominatetheba kground. Forthethe rel-evantenergyrange,lessthan200keV,su h ontaminationprodu estypi ally 10

5

-10

6

targets/te hnologieswithdierentA(atomi mass)anddierentsystemati s sin ethe dierentbehaviorofWIMP andneutron s attering ross se tionas a fun tion the atomi number mass, and/or studying orrelation of events with Gala ti motion by observation of annual modulation or a dire tional signal allowinga dire t identi ation of events as of extra-terrestrial origin, shouldallowadis rimination of WIMP fromneutron signals.

S ale-up to ton-s ale is planned for instan e, for ryogeni te hnologies, making best use of the high dis riminationpower demonstrated notably by CDMS,EDELWEIS S andCRESST. Twoparti ular eorts are foreseen: Su-perCDMS [85℄ and EURECA (European Underground Rare Event sear h withCalorimeterArray) [86℄. Theformer willuse Geand Siionization/heat te hnologylikeCDMSinastagedexpansionfrom27kg to145kg and even-tuallyto 1100 kg by 2015 either atthe US DUSEL, if built, or SNOLAB in Canada. EURECArepresentsa fusionofEDELWEIS S, CRESST ollabora-tions with further new group to develop a 100-1000 kg array using various targets, possibly both ionization/heat and s intillation/heat dis rimination ideas. Bothproje tswillneedtodevelopimproved dete tors,inparti ularto allow better reje tion of surfa e events, for instan e through event position re onstru tion orimproved analysis, and toredu e unit osts.

1.4.2.2 Theoreti al re oil spe trum

Sin e no WIMP signal is dete ted in urrent stage of dire t dark matter sear hes(fa ingthefa tofanunknown ba kground),thiskindofexperiment an establish an upper limit on the s attering ross se tion of WIMP on nu leon as a fun tion of WIMP mass. To su eed in this we need to know how our dete tors respond to a hypotheti al WIMP signal. Thus, the rst stepistosimulateatheoreti alre oilspe trumoftargetnu leiinthe rystal used by the experiment (Germanium for EDELWEISS) indu ed by elasti s atteringofWIMPswithgivenM

W

WIMPmassands attering rossse tion on nu leon

σ

W −nucl

. Later, we will degrade this theoreti al spe trum with experimental threshold and resolution toensure a quiterealisti spe trum.

Here, I want only to stress that the elasti s attering ross se tion de-pendsonthetypeofintera tion onsidered: eitherspin-independentor spin-dependent oupling between WIMPs and nu leons. For spin-independent intera tions, we an express the ross se tion onthe target atom (

σ

W −N

) as afun tion of the ross se tion onprotons (

σ

W −p

) as:

σ

_{W −N}

SI

= (

M

W

+ M

P

M

W

+ M

target

)

−

₂

(

A

2

) the event rate will behigher in heavier targets.

For spin-dependent intera tions, the ross se tion an be written:

σ

_{W −N}

SD

=

32

π

G

2

F

m

2

r

J + 1

J

(a

p

hS

p

i + a

n

hS

n

i)

2

· σ

SD

W −p

(1.18) where

J

is the total angular momentum of the nu leus,

hS

p

i

(

hS

n

i

) is the expe tation value of the spin ontribution of the proton (neutron) group in the nu leusand

a

p

(

a

n

)the oupling onstants between WIMPs andprotons (neutrons).

The starting theoreti al spe trum depends on WIMP s attering ross se tion o nu lei (

σ

W −nucl

), on WIMP mass (M

W

) on target rystal mass (M

A

) and on gala ti halo parameters: dark matter lo al density

ρ

0

and WIMP speed distribution in the halo (

f (v)

) and allows us to determine a theoreti al event rate in the dete tor. All the al ulations for the re oil spe trum refer to the Lewin-Smith work [50℄. Ea h event rate of whatever in omingparti les attering on atarget material an bewrite as follow:

dR =

N

0

A

σvdn,

(1.19)

where

N

0

istheAvogadronumber(=6.02

×

10

23

mol

−

₁

),Atheatomi mass oftargetnu lei,

σ

theparti le-targetmaterials attering ross se tionand

dn

the dierential density per m

3

of in oming parti le moving at

v

velo ity. If we onsider this parti le being a WIMP, its dierential density assumesthe followingform:

dn =

n

0

k

f (−

→

v , −

v

→

E

)d

3

−

→

_{v ,}

(1.20)

where

n

0

is the WIMP density in the gala ti halo,

k

is a normalization onstant(seebelow),

−

→

v

itsvelo ityinthegalaxyrestframeand

−

v

→

E

theEarth velo ity relative to Galaxy.

f (−

→

v , −

v

→

E

)

is the parti le's speed distribution in thehalo supposedtobeaMaxwellianone, Eq.(1.21)brokenoatavelo ity

v

esc

that represents the Galaxy's es ape velo ity.

f (−

→

v , −

v

→

E

) = exp

−

₍

−

→

_{v +}

−

_v

→

_E

₎

2

/v

2

0

.

(1.21)

This

v

esc

, identifying the needed speed to equal the kineti energy of anobje ttothe magnitudeof itsgravitationalpotentialenergy, orresponds thustothemaximalWIMP'svelo ityvalue: aparti lewithavelo ityhigher than

v

esc

willes apefromthehalo. The

k

variableisanormalization onstant hosen in orderthat

n

0

=

R

v

esc

Figure1.11: WIMP elasti s attering o nu lei s heme.

Sin e experiments measure a WIMP event energy spe trum, we need to express the dierential event , Eq. (1.19), as a fun tion of re oiling nu lei energy produ edby WIMP s attering o asshown inFig. 1.11

The WIMP kineti energy is des ribed by

E =

1

2

M

W

v

2

and the target nu lei re oil energy willbe:

E

R

=

1

2

rE(1 − cos θ) = rE cos

2

_θ

R

;

(1.22)

where

θ

is WIMP s attering angle referring to the enter of mass frame,

θ

R

isthere oilingnu leusangle,seeFig.1.11 and

r

isfourtimestheredu ed mass divided by the sum of the masses dened as

M

W

M

A

(M

W

+M

A

)

2

. The re oil energy for the WIMP willbe maximal for target nu lei mass (

M

A

)equal to WIMP mass (

M

W

). Assuming anisotropi distribution of s attering angles inthe enter ofmassframe,thatitmeansthatre oilsareequallydistributed as a fun tion of re oil energy in a range 0<E

R

<

rE

, we an al ulate the dierentialevent rate inre oil energy as follow:

dR

dE

R

=

Z

E

max

E

min

1

rE

dR(E)

=

1

E

0

r

Z

v

max

v

min

v

0

2

v

2

dR(v)

(1.23)