CALOCUBE SVILUPPO DI CALORIMETRIA OMOGENEA AD ALTA ACCETTANZA PER ESPERIMENTI DI RAGGI COSMICI NELLO SPAZIO

Call nell’ambito della CSN5 dell’INFN Oscar Adriani, Responsabile Nazionale

PHYSICS MOTIVATION And the starting point for the CaloCube proposal….

Some of the Cosmic-Ray ‘mysteries’

1 particle / m2×second

1 particle / m2×year

1 particle / km2×year

High energy nuclei ●  “Knee” structure around ~ PeV

●  Upper energy of galactic accelerators (?) ●  Energy-dependent composition

●  Structures in the GeV – TeV region recently discovered for p and He ●  Composition at the knee may differ substantially from that at

TeV ●  Spectral measurements in the knee region up

to now are only indirect ●  Ground-based atmospheric shower detectors ●  High uncertainties

A direct spectral measurement in the PeV region requires great acceptance (few m2sr), good charge measurement and good energy resolution for hadrons (much better than 40%)

High energy Electrons+Positrons ●  Currently available measurements show some

degree of disagreement in the 100 GeV – 1 TeV region

●  Cutoff in the TeV region? Direct measurements require excellent energy resolution (~%), a high e/p rejection power (> 105) and large acceptance above 1 TeV

Our proposal for an ‘optimal’ CR detector

●  A 3-D, deep, homogeneous and isotropic calorimeter can achieve these design requirements:

–  depth and homogeneity to achieve energy resolution (but please remember that anyway a full containment HCAL is impossible in space!!!!)

–  isotropy (3-D) to accept particles from all directions and increase GF ●  Proposal: a cubic calorimeter made of small cubic sensitive elements

–  can accept events from 5 sides (mechanical support on bottom side) → GF * 5 –  segmentation in every direction gives e/p rejection power by means of

topological shower analysis –  cubic, small (~Molière radius) scintillating crystals for homogeneity –  gaps between crystals increase GF and can be used for signal readout

●  small degradation of energy resolution –  must fulfill mass&power budget of a space experiment

●  modularity allows for easy resizing of the detector design depending on the available mass&power

The starting point •  Exercise made on the assumption that

the detector’s only weight is ~ 1600 kg •  Mechanical support is not included in the

weight estimation

•  The chosen material is CsI(Tl) Density: 4.51 g/cm3

X0: 1.85 cm Moliere radius: 3.5 cm λI: 37 cm Light yield: 54.000 ph/MeV τdecay: 1.3 µs λmax: 560 nm

•  Simulation and prototype beam tests used to characterize the detector

N×N×N 20×20×20

L of small cube (cm) 3.6*

Crystal volume (cm3) 46.7

Gap (cm) 0.3

Mass (Kg) 1683

N.Crystals 8000

Size (cm3) 78.0×78.0×78.0

Depth (R.L.) “ (I.L.)

39×39×39 1.8×1.8×1.8

Planar GF (m2sr) ** 1.91

(* one Moliere radius) (** GF for only one face)

See for example: N. Mori, et al., Homogeneous and isotropic calorimetry for space experiments NIMA (2013) http://dx.doi.org/10.1016/j.nima.2013.05.138i

Mechanical idea

Selection efficiency: ε ~ 36% GFeff ~ 3.4 m2sr

Electrons

Electrons 100 – 1000 GeV

(Measured Energy – Real Energy) / Real Energy

Crystals only

Crystals + photodiodes

Non-gaussian tails due to leakages and to energy losses in carbon fiber material

RMS~2%

Ionization effect on PD: 1.7%

Protons Energy resolution (correction for leakage by

looking at the shower starting point)

Selection efficiencies: ε0.1-1TeV ~ 35% ε1TeV ~ 41% ε10TeV ~ 47%

GFeff

0.1-1TeV ~ 3.3 m2sr Gfeff

1TeV ~ 3.9 m2sr Gfeff

10TeV ~ 4.5 m2sr 100 TeV

40%

(Measured Energy – Real Energy) / Real Energy

10 TeV

39%

(Measured Energy – Real Energy) / Real Energy

100 – 1000 GeV

32%

(Measured Energy – Real Energy) / Real Energy

1 TeV

35%

(Measured Energy – Real Energy) / Real Energy

Proton rejection factor with simple topological cuts: 2.105-5.105 up to 10 TeV

The prototype 14 Layers 9x9 crystals in each layer 126 Crystals in total 126 Photo Diodes 50.4 cm of CsI(Tl) 27 X0 1.44 λI

A glance at prototype's TB data SPS H8 Ion Beam: Z/A = 1/2, 12.8 GV/c and 30 GV/c

2H 4He

For deuterium: S/N ~ 14

Please note: we can use the data from a precise silicon Z measuring system located in front of the prototype to have an exact identification of the nucleus charge!!!!

Energy deposit for various nuclei

Charge is selected with the placed-in-front tracking system

Good Linearity even with the large area PD!

Preliminary

Total particle energy (GeV)

Charge selected with a silicon tracker located in front (non interacting ions)

AN INTENSE R&D AND NEW TECHNOLOGIES ARE NECESSARY TO MOVE FORWARD FROM THIS STARTING POINT!

From the idea to the real case • A large R&D effort and development of new technologies

are necessary to go from the simple idea to a real detector (or better, a real prototype…)

• The project has a real and strong interest in the CSN2 in view of possible future experiments (GAMMA-400, HERD….) •  Please keep in mind the importance of the expertise and of the

technologies in this field (see DAMPE and GAMMA-400 tracking systems, INFN is receiving money from the space agencies to design and construct the detectors)

The CaloCube idea Improve the existing Cubic Calorimeter concept to: 1.  Optimize the overall calorimeter performances, in

particular the hadronic energy resolution •  Build up a really compensating calorimeter

•  Cherenkov light •  Neutron induced signals

2.  Optimize the charge measurement •  Make use of the excellent results from the SPS test •  Smaller size cubes on the lateral faces •  Materials to reduce back scattering effect

3.  Build up a prototype fully space qualified •  Mechanics •  Thermics

by developing highly innovative techniques that are one of the core interest of the INFN CSN5

Optimization of the overall calorimetric performances (I)

•  Optimize the hadronic energy resolution by means of the dual - or multiple - readout techniques and/or using cubes made by different materials •  Scintillation light, Cherenkov light, neutron related signals

•  Innovative analysis techniques (software compensation) •  Possible, due to the very fine granularity

•  Development of innovative light collection and detection systems •  Optical surface treatments directly on crystals, to collect/convert the UV

Cherenkov light •  Dichroic filters •  WLS thin layers

•  UV sensitive SiPM and small/large area – twin - Photo Diodes •  New development of front end and readout electronics

•  Huge required dynamic range (>107) •  Fast, medium and slow (delayed) signals together •  New CASIS chip ASIC with integrated ADC

vertical proton 3 TeV

The Principle of compensation: fem

Cherenkov light detection is a tool to estimate fem in a CsI(Tl) CaloCube!

And it works!

ΔE/E: 35% à 15%

BTF Test beam of the prototype with 500 MeV electrons at the end of September

Scintillation light

Cherenkov Light!?!

It seems that we could disentangle Ch from scintillator in CsI(Tl)!

Angular dependence is an hint for Cherenkov detection

CsI cube wrapped in black, 2 phototubes, both with UV filters (DREAM-like)

Signal time profile averaged over many events

+30°

PT2

PT1

PT1

-30°

PT2

PT1 PT2 Beam

Great Thanks to DREAM groups!!!!

CsI(Tl) + Plastic scintillators 24x24x24 (1:1)

The number of neutrons Nn is correlated to the energy release And delayed signal in plastic scintillator is a good tool to estimate Nn!

And it (partially) works!

ΔE/E: 32% à 26%

The technological challenge: life is not as easy as simulation…. • How technically optimize the Cherenkov or n detection in

a difficult environment? •  Tiny gaps in between the crystals •  3-D homogeneity should be maintained •  Cherenkov signal is tiny wrt scintillation signal •  Reduced power consumption

• We are planning detailed R&D activities to obtain the best possible performances •  WLS or Dichroic UV filters optically deposited directly on the

crystals •  UV sensitive light sensors •  Dedicated front end chips to handle:

•  Huge required dynamic range (>107) •  Fast, medium and slow (delayed) signals together •  Small power consumption

The work inside the call - II Optimization of the charge identifier system - CIS

•  The integration of the charge identifier system inside the calorimeter is a real break through for a space experiment •  Huge reduction of mass and power •  Huge reduction of cost •  Great simplification of the overall structure

•  Thinner size scintillators crystals/Cherenkov radiators •  Pixel structure •  Multiple readouts in the ion track •  Back scattering problem to be carefully studied

•  R&D on the front end electronics to reach: •  large dynamic range •  excellent linearity •  low power consumption

Positive hints are coming from the SPS beam test

The work inside the call - III Space qualification • Basic idea:

•  Demonstrate that such a complex device can be built with space qualified technologies •  Necessary step for a real proposal for space •  Production of a space qualified medium size prototype (~700 crystals) •  Composite materials mechanics •  Thermal aspects

•  Microcooling technologies to cool down sensors and/or electronics

•  Radiation damage issues

THE ORGANIZATION

The groups involved 1.  INFN

1.  Firenze 2.  Catania/Messina 3.  Pavia 4.  Pisa 5.  Trieste/Udine

2.  External institutions 1.  IMCB-CNR Napoli à Surface treatments and WLS depositions 2.  CNR-IMM-MATIS Catania à Dichroic filters depositions

3.  External companies 1.  FBK 2.  ST Microelectronics

Work Packages •  WP1: System design, software and simulation, prototype construction

•  Responsible: Oscar Adriani - Firenze •  WP2: Charge identifier system

•  Responsible: Paolo Maestro – Pisa

•  WP3: Crystals/radiators and optical treatments •  Responsible: Sergio Ricciarini – Firenze

•  WP4: Photodetectors and electronics •  Responsible: Valter Bonvicini – Trieste

•  WP5: Beam tests •  Responsible: Sebastiano Albergo – Catania

•  WP6: Space qualification •  Responsible: Guido Castellini - Firenze

Conclusions • The CaloCube concept is really necessary for the next

generation of cosmic ray experiments devoted to the HECR physics •  Large geometrical acceptance to probe the knee region •  Excellent electromagnetic energy resolution •  Excellent hadronic energy resolution with a thin detector (<2λI)

• An intense R&D activity is necessary to go from the idea to a real experiment: •  I: Optimization of the calorimetric performances •  II: Optimization of the charge identifier system •  III: Space qualification

• This R&D will allow INFN CSN5 to develop new widely appreciated and useful techniques, and INFN CSN2 to profit of these techniques for next generation experiments

BACKUP

The WP organizations !"#$ %&'()#* +#(,#-./ 0&1,#-./ 231,#-./ 4(5,#-./ 6(5,#-./ 7(5,#-./

!"# $%&'(')'*(+*&+),%+-./%0'(%+1%*2%)34 56 7893'.(' :);94+*&+),%+*<)'2.0+=34/).0/ 56 7

:);94+*&+>,+.(9+(+9%)%=)'*( 56 7?<)'2'@%9+9%/'1( 56A+"6 7BC+.(9+.(.04/'/+)**0/ 56A+"D 7 7 7CCEFG+<3*)*)4<%+9%/'1( 56A+"6A+"D 7 7CCEFG+<3*)*)4<%+.//%2-04 56 7 7 7"3*H%=)+=**39'(.)'*(+.(9+2.(.1%2%() 56 7 7 7 7 7 7

!"I BC+/);9'%/+J1%*2%)34A+-.=K/=.))LM "6 7B.%/)3* N%.9*;)+/);9'%/ "6 7 7

C,.3.=)%3'@.)'*(+.(9+0.-L+)%/)/ "6 7 7

8//%2-04+*&+#/)+<3*)*)4<% "6 7"3%0'2'(.34+-%.2+)%/) "6 7"3*)*)4<%+9%/'1(+.(9+.//%2-04+'(+CCEFG "6 7 7F%.2+)%/)+.(9+9.).+.(.04/'/ "6 7

!"O $%P%0*<2%()+*&+0'1,)+=*00%=)'*(+/4/)%2/ 56 7 7N'=='.3'(' $'=,3*'=+&'0)%3+<3*)*)4<%+.(9+)%/)/ CQR56 7 7 7

!S:+<3*)*)4<%+.(9+)%/)/ T8R56 7 7 7F.)=,+<3*9;=)'*(+*&+&'0)%3/+.(9+!S:+&*3+CCEFG CQRT8 7 7

!"U ?<)'2'@.)'*( *& <,*)*9%)%=)*3/ &*3 ),%9'&&%3%()+/'1(.0/

Q: 7 7

F*(P'='(' T%V+P%3/'*(+*&+),%+C8:6:+=,'< Q: 7 7 7 7E('&'%9+%0%=)3*('=+=,.'(/+&*3+-%.2+)%/)/L Q:R56RCQ 7 7 7 7N%.9W*;) %0%=)3*('=/ &*3 ),% /0*V (%;)3*(=*2<*(%()/

Q: 7 7 7

5./)+9'1')'@%3+&*3+>,+<;0/%+/,.<%+9'/=3'2'(.)'*( "D 7 7 7

!"X C34/).0/+-%.2+)%/)/ CQA56A"6A"DAQ: 7 7 780-%31* :4/)%2+-%.2+)%/)/ CQA56A"6A"DAQ: 7 7 7

!"Y :);94+*&+1%(%3.0+3%Z;'3%2%()/+.(9+)%/)+<3*=%9;3%/

56 7 7

C./)%00'(' Q,%32.0+.(.04/'/+.(9+=**0'(1 56 7 7B%=,.('=.0+/)3;=);3%+9%/'1( "6 7 7CCEFG+2%=,.('=.0+/)3;=);3%+<3*9;=)'*(+ "6 7 7

89#(-.,1-#':&;,#<=(>"3-,"&1,#'.?@"('<&;,A3<(<(9A-,B<&#(3?B('<&

8A"B-,C?"@'='B"('<&

D-".,(-#(#

E5"3:-,'1-&('='-3,#9#(-.

F5<(<1-(-B(<3#,"&1,-@-B(3<&'B#

E39#("@#G3"1'"(<3#,"&1,<A('B"@,(3-"(.-&(#

FTE in the various INFN RU !"#$ %&'$ (!"#$%&'()$#*+,-*.' /%'01(2%3-.*%-' 456*"7-(8-9+$::$ /%'01(!++'7-*,' 45;*<#$%,'(!.,'.-' /%'0$++'%$(*(7'.,%*,,' =5>$??*+*"<*(!.&$"* /%'0$++'%$(!++'7-*,' @5A*::*??'(8*$,*.* /%'0$++'%$(*(7'.,%*,,' @5B%-0-%'C(!.,'.-' A-7$%7*,'%$ 45

)*) +&,"'-.+/

!3%-*.-(2+7*% /%'01(2%3-.*%-' D5E'.&-(>*++-<' A-7$%7*,'%$ F5E',,*-()$%&-' A-7$%7*,'%$ @5G*+,$""-.-(89-3' !++1)$.-'% 5GH-*%-(>*++-<' A-7$%7*,'%$ 45;$.?-(/-$%&-9"-' !++$&.-+,* @5A-77-*%-.-()$%&-'(E%9.' A-7$%7*,'%$ F5),*%'39#,+$I(2"$J+*.3% !++$&.-+,* F5K$,,-()$#*+,-*.' G'""*#'%*,'%$(B$7.-7'(L1A1 D5

0*1 +&,"'-.+/

G*,,*.$'(/*'"'(M*",$% /%-<'(A-7$%7*,'%$ =5N*%3'C(A'#$%,' B$7.'"'&'(L1/1 45A*::'"3-(!.3%$* /%-<'(B$7.'"'&' 45

2*3 +&,"'-.+/

>*$+,%'(/*'"' A-7$%7*,'%$ O5>*%%'77H$+-(/-$%()-<'.$ /%'01(2%3-.*%-' @5)9"*P(!%,* !++$&.-+,* 455

)*4 +&,"'-.+/

E'.I-7-.-(Q*",$% /%-<'(A-7$%7*,'%$ 45G*9?(K-$&' A-7$%7*,'%$ =58%$&'%-'(!..* A-7$%7*,'%$ =5/*9"$,,*(8-'I*..- /%'01(2%3-.*%-' =5

2*4 +&,"'-.+/

4*4 56"78-,&,"'-.+/

/-+*R)-$.*

G*,*.-*R>$++-.*

6-%$.?$

/*I-*

B%-$+,$RS3-.$

Money Matrix

!"#$%& '()*+,-.,/0 12$,03/43(5 6789 678: 678; !"#$%%&8 <,0=4$-+/./.>-($5(.(*./+ ?(@$5(.(*./+)$A/+$.B($A"44$-+/./.>-($C$;77$*+>).=4) <D 67 E7 67 &'(

'(),F0$)."5,() G+>).=4$=05$5(.(*./+)H+=5,=./+)$$-+/*"+(I(0.)$A/+$(3(+>$%& <DH&DHJ!C2'HGJ K7 87 7 )(L(0(+=4$-+/./.>-,0F$*/).) <DH&DHJ!C2'HGJ 87 87 87 *( !"#+,-&

&.(

%&6 <,0=4$-+/./.>-($5(.(*./+ '(.(*./+)$A/+$GD! &D 88 9 E '*M(*B=0,*)$A/+$GD! &D ; ; N &.&B/./)(0)/+) &D ; 9 : &/<+/0.$(05$=05$+(=5/".$(4(*.+/0,*)$A/+$GD! &D ; 9 : &/ !"#+,-'

0'

%&K &+/./.>-($/A$*=4/+,I(.(+$I/5"4($ O,FB.$-,-()P$/-.,*=4$F4"() <D K 7 7 *!"+A=*($.+(=.I(0.)$Q2R$@=3(4(0F.B$)B,A.(+S ?T 87 7 7 &(!"+A=*($.+(=.I(0.)$Q2R$5,*B+/,*$A,4.(+S GJ E 7 7 1

T0=4>)>)$/A$I=.(+,=4)$=05$.+(=.I(0.) GB=+=*.(+,U=.,/0$/A$I=.(+,=4)$@,.B$5,AA(+(0.)$.(*B0,V"() GJ N 7 7 0GB=+=*.(+,U=.,/0$/A$)"+A=*($.+(=.I(0.)$@,.B$DWT$=.$OTWXG <D : 7 7 /O,FB.$-,-()P$/-.,*=4$F4"()P$)"+A=*($.+(=.I(0.) <DH?THGJ 7 87 7 &(M=))$-+/5"*.,/0$/A$5,*B+/,*$A,4.(+$QN77$A,4.(+)S GJ 7 6E 7 '1M=))$-+/5"*.,/0$A/+$2R$%=3(4(0FB.$)B,A.(+ ?T 7 8E 7 &1 !"#+,-*

1.

%&9 &B/./)(0)/+)$A/+$)*,0.,44=.,/0$4,FB. !,$-C,C0$-B/./5,/5() J!C2'H<D 87 67 87 )(?1X$A/+$.B,0$-B/./5,/5()$-=*Y=F,0F J!C2'H<D 8: 8: 7 *(

&B/./)(0)/+)$A/+$GB(+(0Y/3$4,FB. 2RC(0B=0*(5$-C/0C0$-B/./5,/5()$=05$T&') J!C2' : 7 7 /2RC)(0),.,3($!,&M J!C2' 87 7 7 &(

<+/0.C(05$T!DG) GT!D!8Z6[G=4/G"#( J!C2' 67 K7 7 /('1!$A=).$5,F,.,U(+$A/+$GB(+(0Y/3$),F0=4) &R N 8: E *(

X4(*.+/0,*)$#/=+5) <+/0.(05$=05$+(=5/".$(4(*.+/0,*) J!C2'H<D 87 67 7 *( !"#+,-)&./

%&: D0.(F+=.,/0$@,.B$.().$#(=I$=+(=)$ 1(I/.($B=054,0F$$Q+/.=.,/0H.+=0)4=.,/0S$/A$*+>).=4) GJ 87 7 7 &(&G$A/+$)4/@$*/0.+/4$=05$'T\ GJ 6 7 7 'M(*B=0,*=4$)"--/+.)$A/+$.().$#(=I) GJ 8Z: 7 7 &2/M(*B=0,*=4$)"--/+.)$A/+$.().$#(=I) <D 7 6 7 '6$L]U$^)*,44/)*/-($A/+$*+>).=4$),F0=4$*B=+=*.(+,U=.,/0 GJ 67 7 7 '( !"#+,-/

*/2/

%&; !-=*($V"=4,A,*=.,/0$.().) R,#+=.,/0$.().$QA,0=4$-+/./.>-(S <D 7 7 87 &(^".F=)),0F$.().)$Q)=I-4()S <D : : 7 &(JB(+I=4C3=*""I$.().$QA,0=4$-+/./.>-(S <D 7 7 87 &(D++=5,=.,/0$.().)$QI=.(+,=4)P$)"+A=*($.+(=.I(0.)P$(4(*.+/0,*)S <D 87 87 7 '(XMGHXMD$.().)$/0$A,0=4$-+/./.>-($C$,0).+"I(0.$*=4,#+=.,/0 <D 7 7 87 &(

T"_,4,=+>$.().$)>).(I) M(*B=0,*)P$(4(*.+/0,*)$#/=+5)P$*=#4() <D 6 ; 6 &(M,*+/*//4,0F$)."5,() M,*+/-"I-)P$*/0.+/44(+)P$*//4,0F$4,V",5)P$-,-() <DH&D 87 7 7 &(<,0=4$-+/./.>-($5(.(*./+ M(*B=0,*)$/A$.B($*=4/+,I(.(+ &D 8: K: ;7 &&( !"#+,-3

&.(

G/44=#/+=.,/0$I((.,0F)$ GJ K K K .<D : : : &/&D : 9 K &'&R 6 6 6 3

J!C2' 9 K K &(J().$#(=I) </+()((0$#(=I$.().)$3)$>(=+) MX$O,0=*P$WJ<O?!$P$!&! !&!

K$-(/-4($A/+$$:$5=>)$$A/+$3=+,/")$#(=I$.(). GJ K N 9Z: &)2/9$-(/-4($A/+$$:$5=>)$$A/+$3=+,/")$#(=I$.(). <D 87 86 ; '16$-(/-4($A/+$$:$5=>)$$A/+$3=+,/")$#(=I$.(). &D 87 K K &38$-(/-4($A/+$$:$5=>)$$A/+$3=+,/")$#(=I$.(). &R : 8Z: 8Z: 16$-(/-4($A/+$$:$5=>)$$A/+$3=+,/")$#(=I$.(). J!C2' : K K && !"#+!4$56%

&'.2/

!"#$%+'(&)!"#$%+'(&/!"#$%+'(&374$89+!"#$%*'32/ *0/2/ &.. .(&

J+=3(4)

!>).(I$5(),F0P$)/A.@=+($=05$),IZP$-+/./.>-($

GB=+F($,5(0.,A,(+$)>).(I

G+>).=4)H+=5,=./+)$=05$/-.,*=4$.+(=.I(0.)

&B/./5(.(*./+)$=05$(4(*.+/0,*)

W(=I$J().)

!-=*($V"=4,A,*=.,/0

O,FB.$*/44(*.,/0$)>).(I$A/+$A,0=4$*=4/+,I(.(+$-+/./.>-(Z

Mechanical idea

The readout sensors and the front-end chip •  Minimum 2 Photo Diodes are necessary on each crystal to cover the

whole huge dynamic range 1 MIPà107 MIPS

•  Large Area Excelitas VTH2090 9.2 x 9.2 mm2 for small signals •  Small area 0.5 x 0.5 mm2 for large signals

•  Front-End electronics: a big challenge! •  The CASIS chip, developed in Italy by INFN-Trieste, is very well

suited for this purpose •  IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 57, NO. 5, OCTOBER 2010

•  16 channels CSA+CDS •  Automatic switching btw low and high gain mode •  2.8 mW/channel •  3.103 e- noise for 100 pF input capacitance •  53 pC maximum input charge

MC simulations ●  Fluka-based MC simulation

–  Scintillating crystals –  Photodiodes

●  Energy deposits in the photodiodes due to ionization are taken into account

–  Carbon fiber support structure (filling the 3mm gap) ●  Isotropic generation on the top surface

–  Results are valid also for other sides ●  Simulated particles:

–  Electrons: 100 GeV → 1 TeV –  Protons: 100 GeV → 100 TeV –  about 102 – 105 events per energy value

●  Geometry factor, light collection and quantum efficiency of PD are taken into account

●  Requirements on shower containment (fiducial volume, length of reconstructed track, minimum energy deposit)

–  Nominal GF: (0.78*0.78*π)*5*ε m2sr= 9.55*ε m2sr

A glance at prototype's TB data

H: Z=1 <ADC>=330 He: Z=2 <ADC>=1300 Li: Z=3 <ADC>=3000 Be: Z=4 <ADC>=5300 B: Z=5 <ADC>=8250 C: Z=6 <ADC>=12000 N Z=7 <ADC>=16000

He

Li

Be

B

C

N

Please remind that this is a calorimeter!!!! Not a Z measuring device!!!!

<ΔX> = 1.15 cm

Shower starting point resolution

Protons

Shower Length (cm)

Sig

nal /

Ene

rgy

Shower length can be used to reconstruct the correct energy

100 – 1000 GeV

Red points: profile histogram Fitted with logarithmic function

Energy estimation

ΔE = 17%

Proton rejection factor Montecarlo study of proton contamination using CALORIMETER INFORMATIONS ONLY

q  PARTICLES propagation & detector response simulated with FLUKA

q  Geometrical cuts for shower containment q  Cuts based on longitudinal and lateral development

LatRMS4

protons

electrons

LON

GIT

UD

INA

L

LATERAL

q  155.000 protons simulated at 1 tev : only 1 survive the cuts

q  The corresponding electron efficiency is 37% and almost constant with energy above 500gev

q  Mc study of energy dependence of selection efficiency and calo energy distribution of misreconstructed events

10TeV 1TeV

λ1

"#(%,&)/"% 

E(GeV)

E3 d

N/d

E(G

eV2 ,

s-1 )

Protons in acceptance(9,55m2sr)/dE

Electrons in acceptance(9,55m2sr)/dE

vela

Electrons detected/dEcal

Protons detected as electrons /dEcal

Contamination : 0,5% at 1TeV 2% at 4 TeV

An upper limit 90% CL is obtained using a factor X 3,89

= = 0,5 x 106

X Electron Eff. ~ 2 x 105

Proton rejection factor

Response uniformity of the crystals

~14% Uniformity

ARTICLE IN PRESS

A chi-squared that imposes optimum linearity (but notoptimum resolution) is

w2 !X

n

X

i

En "P

m amBm;i

sn

! "2

:

By setting partial derivatives to zero,

0!@w2

@ak! " 2

X

n

X

i

En "P

m amBm;i

s2n

# Bk;i $k! 1; . . . ;4%

and, arranging terms into a set of linear equations in am,

X

n

X

i

EnBk;i

s2n

!X

m

amX

n

X

i

Bm;iBk;i

s2n

we obtain a set of linear equations ck !P

m amMm;k,

with ck !X

n

X

i

EnBk;i

s2n

and Mm;k !X

n

X

i

Bm;iBk;i

s2n

:

The solution for the constants ak is ~a !M"1~c .We used a large variety of electron data for this calibration.

Beams with energies of 10, 20, 30, 50, 100, 150 and 200GeV weresteered into each of the points a, b and c indicated in Fig. 4. Thecalibration constants were derived from all the data collected inthese 21 different runs.

The signals Bm;i used in this procedure represented theintegrated charge collected in the first 115ns after the start ofthe BGO signals. They were thus expressed in units of [mVns]. Thefour constants ak derived from this procedure all share the samerelationship between an energy unit and 1mVns; they areintercalibrated.

However, in order to determine the final calibration constants(expressed in GeVs per millivolt-nanosecond), the BGO signalsfirst had to be unraveled into their scintillation and Cherenkovcomponents. This procedure is described in Section 3.2, where wealso finalize the description of the calibration procedure for theseseparate signals.

3. Experimental data and methods

3.1. Experimental data

Most of the measurements described in this paper wereperformed with pion beams. Negatively charged pions of 20 and50GeV, and p& beams of 100, 150, 200 and 300GeV were steeredinto each of the positions a, b and c indicated in Fig. 4. In each run,50 000 events were recorded. In addition, multiparticle ‘‘jets’’ of100, 200 and 300GeV were created with p& beams steered intothe center of the calorimeter system. The polyethylene target (seeSection 2.2) was placed 35 cm upstream of the ECAL for thesemeasurements, and we selected interactions with a minimummultiplicity of 10 by means of a threshold on the signals from theITC counter downstream of this target. At each energy, 200 000events were collected this way. In order to investigate possiblebiases, we also collected 50 000 events without such a threshold,for each energy. The pion beams contained some muons, at thefew-% level. These muons were easily recognized (using the muoncounter) and removed from the event samples.

In order to study the performance of the BGO crystal matrix inthis unusual geometry, we used electron beams with energies of10, 20, 30, 50,100, 150 and 200GeV.

3.2. Exploiting the BGO signals

BGO is a bright scintillator, Cherenkov radiation representsonly a tiny fraction of the light generated by high-energy particle

showers. Yet, the very different optical spectra and time structuresoffer good possibilities for distinguishing between these twocomponents. In a previous paper, we have demonstrated that anultraviolet filter, combined with a detailed measurement of thetime structure makes it possible to measure the contributions ofscintillation and Cherenkov light to the crystal signals event byevent with excellent precision [7].

We have applied the same techniques in the present series ofmeasurement. The four PMTs that detected the light produced inthe BGO crystal matrix were equipped with UV filters.11 Thesefilters were transparent for light in the wavelength region from250 to 400nm, which harbors a large fraction of the Cherenkovlight, plus a small fraction of the scintillation light, which peaksaround 480nm. The time structure of the signals from the PMTsclearly exhibited these two components, as illustrated in Fig. 5.The (prompt) Cherenkov component is represented by the sharppeak, whereas the long tail has the same characteristic timestructure as pure scintillation signals generated in this crystal, i.e.,an exponential decay with a time constant of 300ns.

The signals from the PMTs that detected the light transmittedthrough the UV filters thus contained event-by-event informationabout the relative contributions of both Cherenkov and scintilla-tion photons. We have used the oscilloscope data to extract thisinformation, as follows. For every event, the integrated chargecollected in the time interval from 50 to 115ns after the start ofthe pulse was used as a measure for the scintillation signalproduced in that event, while the charge collected from 0 to 16nswas used as the basis for the measurement of the Cherenkovsignal. However, there was always some scintillation light thatcontaminated the latter signal. From detailed studies of the timestructure of the unfiltered (i.e., almost pure scintillation) signals,we concluded that the integrated charge due to scintillation lightcollected in the time interval from 0 to 16ns after the start of thepulse amounted to 20% of the charge collected from 50 to 115ns.

Fig. 5. The time structure of a typical shower signal measured in the BGO emcalorimeter equipped with a UV filter. These signals were measured with asampling oscilloscope, which took a sample every 0.8ns. The UV BGO signals wereused to measure the relative contributions of scintillation light (gate 2) andCherenkov light (gate 1).

11 UG11glass transmission filter (Schott). See Ref. [7] for details on theproperties of this filter.

N. Akchurin et al. / Nuclear Instruments and Methods in Physics Research A 610 (2009) 488–501492Dual readout –> BGO: scintillation + Cherenkov

Hardware compensation

power, although not very impressive, is fairly independent of theabsorber thickness.

In conclusion, we see that methods intended to extractinformation on the Cherenkov component of the PbWO4 signalsthat are based on the time structure are less sensitive to theabsorber than methods based on the directionality. Whereasthe directionality of the Cherenkov component tends to fade asthe shower develops, the prompt character of the Cherenkov lightis not affected and allows measuring the contribution of thiscomponent, albeit not with a very impressive precision.

4. Experimental results for BGO

4.1. The Cherenkov component in the BGO signals

The time structures of the signals from the BGO crystalobserved with the yellow filter and the UV filter were verydifferent.

This is illustrated in Fig. 13, which shows the time structuresmeasured from both sides of the crystal (i.e. with the twodifferent filters) for a typical shower developing in it. Thesefeatures can be understood from the properties of the filters andof the light that is converted into an electric signal.

Fig. 14 shows the transmission characteristics of the two filtersas a function of wavelength, as well as the scintillation spectrumof BGO, the spectrum of the Cherenkov light generated in thecrystal and the wavelength dependence of the quantum efficiencyof the photocathode used in the PMTs. The scintillation spectrumof BGO is centered around a wavelength of 480nm, i.e. in theyellow/green domain. The decay time of the scintillation processis !300ns. The yellow filter is highly transparent for this type oflight, as reflected by the signal shape in Fig. 13a.

The UV filter is transparent for light in the wavelength regionaround 300nm, and also has a window around 700nm, where thetransmission coefficient is a few percent of that in the ultravioletregion, and the quantum efficiency of the photocathode is also atthe level of 1% of that around 350nm. As a result, this filter ishighly transparent for Cherenkov light in the 300–400nm range,and for wavelengths 320–400nm the probability that photonsreaching the photocathode produce a photoelectron exceeds 10%.On the other hand, only a very small fraction (o0:1%) of thescintillation light penetrates this filter.

Even though Cherenkov light represented a very small fractionof the total light production in this BGO crystal, it was thereforeprominently present in the signals from the PMT that read outthe side where the UV filter was mounted. This is illustrated by

the time structure of the signals in Fig. 13b, where the sharp peakrepresents the prompt Cherenkov signal component.

In order to see if this prompt peak was indeed caused byCherenkov light, we studied its angular dependence. The crystalwas rotated around the y-axis, from y " #45$ to %60$, in steps of5$ (see Fig. 1). In order to limit the contribution of scintillationlight to the UV signals, the ADC gate was adjusted so that thesignals were integrated up to t " 120ns (i.e. over the first !10nsafter the start of the pulse, see Fig. 13), Fig. 15 shows the ratio ofthe ADC signals from the UV and Y sides of the crystal as afunction of y. It clearly illustrates the directional nature of thelight contained in the ‘‘prompt’’ UV signal component. It peaksnear y & 28$ " 90$ # yC, as one would expect for Cherenkov light.

More detailed, quantitative information was derived from thetime structure of the signals. The oscilloscope measurementsmade it possible to determine the contribution of scintillationlight to the UV signal in a narrow gate around the prompt peakevent-by-event. This could be done by normalizing the shape ofthe time structure of pure scintillation light, which was wellknown from the signals measured with the yellow filter, to the tail

ARTICLE IN PRESS

Fig. 13. The time structure of a typical 50GeV e# signal measured in the BGO crystal equipped with a yellow filter (a), and with a UV filter (b). These signals were measuredwith the sampling oscilloscope, with a time resolution of 2.0 ns. The crystal was oriented perpendicular to the beam line (y " 0).

Fig. 14. Light transmission as a function of wavelength for the two filters used toread out the BGO crystal. The light emission spectrum of the crystal, the spectrumof the Cherenkov light generated in it and the quantum efficiency of the PMTs usedto detect this light are shown as well. The vertical scale is absolute for thetransmission coefficients and the quantum efficiency, and constitutes arbitraryunits for the light spectra.

N. Akchurin et al. / Nuclear Instruments and Methods in Physics Research A 595 (2008) 359–374 367

Filter: 250 ÷ 400 nm for Cherenkow light >450 nm for Scintillator light

Even better for CsI(Tl) since the scintillation light emission is very slow

Dual-readout calorimetry with a full-size BGO electromagnetic section

N. Akchurin a, F. Bedeschi b, A. Cardini c, R. Carosi b, G. Ciapetti d, R. Ferrari e, S. Franchino f, M. Fraternali f,G. Gaudio e, J. Hauptman g, M. Incagli b, F. Lacava d, L. La Rotonda h, T. Libeiro a, M. Livan f, E. Meoni h,D. Pinci d, A. Policicchio h,1, S. Popescu a, F. Scuri b, A. Sill a, W. Vandelli i, T. Venturelli h, C. Voena d,I. Volobouev a, R. Wigmans a,!

a Texas Tech University, Lubbock, TX, USAb Dipartimento di Fisica, Universit !a di Pisa and INFN Sezione di Pisa, Italyc Dipartimento di Fisica, Universit !a di Cagliari and INFN Sezione di Cagliari, Italyd Dipartimento di Fisica, Universit !a di Roma ‘‘La Sapienza’’ and INFN Sezione di Roma, Italye INFN Sezione di Pavia, Italyf INFN Sezione di Pavia and Dipartimento di Fisica Nucleare e Teorica, Universit !a di Pavia, Italyg Iowa State University, Ames, IA, USAh Dipartimento di Fisica, Universit !a della Calabria and INFN Cosenza, Italyi CERN, Gen!eve, Switzerland

a r t i c l e i n f o

Article history:Received 9 April 2009Received in revised form29 June 2009Accepted 5 August 2009Available online 6 September 2009

Keywords:CalorimetryCherenkov lightCrystalsOptical fibers

a b s t r a c t

Beam tests of a hybrid dual-readout calorimeter are described. The electromagnetic section ofthis instrument consists of 100 BGO crystals and the hadronic section is made of copper in whichtwo types of optical fibers are embedded. The electromagnetic fraction of hadronic showersdeveloping in this calorimeter system is estimated event by event from the relative amounts ofCherenkov light and scintillation light produced in the shower development. The benefits andlimitations of this detector system for the detection of showers induced by single hadrons andby multiparticle jets are investigated. Effects of side leakage on the detector performance are alsostudied.

& 2009 Elsevier B.V. All rights reserved.

1. Introduction

In recent years, dual-readout calorimetry has emerged as apromising new solution for the need to detect both leptons andhadrons with excellent precision in high-energy particle physicsexperiments [1]. The Dual Readout Method (DREAM) is based on asimultaneous measurement of different types of signals whichprovide complementary information about details of the showerdevelopment. It has been argued [2,3] and experimentallydemonstrated [4] that a comparison of the signals produced byCherenkov light and scintillation light makes it possible tomeasure the energy fraction carried by the electromagneticshower component, fem, event by event. Since fluctuations in femare responsible for all traditional problems in calorimetric hadrondetection, this may lead to an important improvement in theperformance of hadron calorimeters [5].

The first calorimeter of this type was based on a copperabsorber structure, equipped with two types of active media.In this detector, scintillating fibers measured the total energydeposited by the shower particles, while Cherenkov light,generated by the charged, relativistic shower particles, was pro-duced in undoped optical fibers [4,6]. It was recently demon-strated that the signals from certain high-density crystals (PbWO4,BGO) can also be unraveled into Cherenkov and scintillationcomponents [7], and that such crystals, when used in conjunctionwith the fiber calorimeter mentioned above, offer in principlethe same advantages for hadronic shower detection as thelatter [8].

In this paper, we describe high-energy beam tests of a hybridcalorimeter system that consisted of a full-size electromagneticsection made of BGO crystals, backed up by a dual-readout fiberhadronic section, and surrounded by a system of lateral leakagecounters. In Section 2, the detectors and the experimental setup inwhich they were tested are described. In Section 3, we discuss theexperimental data that were taken and the methods used toanalyze these data. The experimental results are presented inSection 4, and conclusions are given in Section 5.

ARTICLE IN PRESS

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/nima

Nuclear Instruments and Methods inPhysics Research A

0168-9002/$ - see front matter & 2009 Elsevier B.V. All rights reserved.doi:10.1016/j.nima.2009.08.074

! Corresponding author. Fax: +18067421182.E-mail address: wigmans@ttu.edu (R. Wigmans).1 Now at Department of Physics, University of Washington, Seattle, WA, USA.

Nuclear Instruments and Methods in Physics Research A 610 (2009) 488–501

f=Fem

If S/E is significantly different from C/Eà Good energy resolution!

Expected number of events

Electrons Effective

GF (m2 sr) σ(E)/E E>0.5 TeV E>1 TeV E>2 TeV E>4 TeV

3.4 ~1% 181.103 35.103 5.103 6.102

Assumptions: •  10 years exposure •  p/e rejection factor ~ 105 •  Depth: 39 X0 – 1.8 λ

Protons and Helium – Polygonato Model

Effective GF (m2 sr)

σ(E)/E

E>0.1 PeV E>0.5 PeV E>1 PeV E>2 PeV E>4 PeV

P He P He P He P He P He

~4.0 35% 7.8.103 7.4.103 4.6.102 5.1.102 1.2.102 1.5.102 28 43 5 10

Heavier Nuclei (2<Z<25)– Polygonato Model

Effective GF (m2 sr)

σ(E)/E

E>0.1 PeV E>0.5 PeV E>1 PeV E>2 PeV E>4 PeV

P He P He P He P He P He

~4.8 32% 4.3.103 4.5.103 3.0.102 3.2.102 1.0.102 1.0.102 29 34 8 10

Knee

Response as function of impact point

Non interacting ions

H: Z=1 <ADC>=330 He: Z=2 <ADC>=1300 Li: Z=3 <ADC>=3000 Be: Z=4 <ADC>=5300 B: Z=5 <ADC>=8250 C: Z=6 <ADC>=12000 N Z=7 <ADC>=16000

He

Li

Be

B

C

N

Please remind that this is a calorimeter!!!! Not a Z measuring device!!!!

Preliminary

Charge linearity on the single crystal (non interacting ions)

Schott UG11 UV Filter