Eu . Phys. J. C (2024) 84:1004
h ps://doi.o g/10.1140/epjc/s10052-024-13393-2
Regula A icle - Expe imen al Physics
Measu emen o he absolu e e iciency o he X-ARAPUCA
pho on de ec o o he DUNE Fa De ec o 1
R. Ál a ez-Ga o e1, C. B izzola i2,3, A. Can o1,E.Cal o
1, C. M. Ca ado i2, C. Cues a1, A. de la To e Rojo1,
I. Gil-Bo ella1,C.Go i
2, D. Gu an i2,3, A. A. Machado5, S. Man hey Co chado1, I. Ma ín1,C.Massa i
2,3,
L. Meazza2,3,a, C. Paloma es1, L. Pé ez-Molina1,b, E. Seg e o4, F. Te ano a2,3, A. Ve dugo de Osa1,
H. Viei a de Souza7, D. Wa ne 6
1CIEMAT, A da. Complu ense 40, 28040 Mad id, Spain
2Dipa imen o di Fisica “Giuseppe Occhialini”, Uni e si á degli S udi di Milano-Bicocca, Piazza della Scienza 3, 20126 Milan, I aly
3INFN Sezione di Milano-Bicocca, Piazza della Scienza 3, 20126 Milan, I aly
4Dipa imen o di Fisica, Uni e si á degli S udi Fede ico II, 80126 Naples, I aly
5Ins i u o de Física “Gleb Wa aghin”, UNICAMP, Campinas, SP 13083-859, B azil
6CSU, Limeligh A e, Cas le Rock 80109, USA
7Labo a oi e As opa icule e Cosmologie, Rue Alice Domon e Léonie Duque 10, 75013 Pa is, F ance
Recei ed: 23 May 2024 / Accep ed: 20 Sep embe 2024 / Published online: 7 Oc obe 2024
© The Au ho (s) 2024
Abs ac The DUNE a de ec o has been designed o
de ec pho ons and elec ons gene a ed by he cha ged p od-
uc s o he in e ac ion o neu inos wi h a massi e liquid
a gon (LA ) a ge . The pho on de ec ion sys em (PDS) o
he i s DUNE a de ec o (FD1) is composed o 6000
pho on de ec ion uni s, named X-ARAPUCA. The de ec-
ion o he p omp ligh pulse gene a ed by he pa icle
ene gy elease in LA will complemen and boos he DUNE
LA Time P ojec ion Chambe . I will imp o e he non-
beam e en s agging and enable a low ene gies he ig-
ge and he calo ime y o he supe no a neu inos. The X-
ARAPUCA is an assembly o se e al componen s. I s pho-
on de ec ion e iciency (PDE) depends on he design o
he assembly, on he g ade o he indi idual componen s
and on hei coupling. The X-ARAPUCA PDE is one o
he leading pa ame e s o he PDS sensi i i y, ha in u n
de e mines he sensi i i y o he DUNE o he de ec ion o
co e-collapse supe no a wi hin he galaxy and o nucleon
decay sea ches. In his wo k we p esen he inal assess-
men o he absolu e PDE o he FD1 X-ARAPUCA base-
line design, measu ed in wo labo a o ies wi h independen
me hods and se ups. P elimina y esul s we e epo ed in
Paloma es (JINST 18(02):C02064, h ps://doi.o g/10.1088/
1748-0221/18/02/C02064, 2023). One hund ed six y uni s
o hese X-ARAPUCA de ices ha e been deployed in he
NP04 acili y a he CERN Neu ino Pla o m, he 1:20 scale
ae-mail: [email protected] n.i
be-mail: [email p o ec ed] (co esponding au ho )
FD1 p o o ype, and will be ope a ed du ing he yea 2024.
The assessed alue o he PDE is a key pa ame e bo h in he
NP04 and in he DUNE analysis and econs uc ion s udies.
1 In oduc ion
The Deep Unde g ound Neu ino Expe imen (DUNE) [2]
is a dual-si e expe imen ha aims o measu e he neu ino
oscilla ion pa ame e s wi h a p ecision and sensi i i y ha
will allow o es he CP iola ion in he lep onic sec o and
de e mine he neu ino mass o de ing [3]. I will also pe -
o m nucleon decay and beyond s anda d model sea ches
and i will con ibu e o he de ec ion o as ophysical neu-
inos om he galaxy, he Sun and co e-collapse supe no a
wi hin he galaxy [4]. The DUNE Fa de ec o (FD) will con-
sis o ou 17 k LA TPC modules. This liquid a gon (LA )
echnology will make possible o econs uc neu ino in e -
ac ions wi h image-like p ecision. The design o he ou
iden ically sized modules is su icien ly lexible o s aging
cons uc ion and e ol ing he LA TPC echnology. The i s
FD Module (FD1) will use he ho izon al d i (HD) ech-
nology, in which ioniza ion cha ges d i ho izon ally in he
LA unde he in luence o an elec ic ield owa ds a e ical
anode, whe e hey a e ead ou . Fou 3.5 m d i olumes a e
c ea ed be ween i e al e na ing anode and ca hode walls,
each wall ha ing dimensions o (58 ×12) m2, and ins alled
inside a c yos a . A schema ic o he TPC is shown in Fig.1.
123
1004 Page 2 o 14 Eu . Phys. J. C (2024) 84 :1004
Fig. 1 10 k DUNE a de ec o module, showing he al e na ing
58.2 m long, 12.0 m high anode and ca hode planes, as well as he
ield cage ha su ounds he d i egions be ween he anode (APA) and
ca hode plane (CPA) [2]
LA p oduces abundan VUV scin illa ion ligh , emi -
ing 51,000 pho ons/MeV when exci ed by minimum ioniz-
ing pa icles in he absence o a d i ield [5,6]. The pa -
icle’s ene gy losses popula e single and iple s a es o
A dime s (A ∗
2), ha de-exci e wi h cha ac e is ic imes o
6 ns and abou 1.6 µs espec i ely, emi ing 127 nm pho-
ons. The high ligh yield and i s e icien de ec ion will
enhance DUNE de ec o capabili ies. To ul ill he supe -
no ae neu ino p og am and o e icien ly agging nucleon
decays, DUNE equi es an a e age ligh yield o >20 pho o-
elec ons (PE)/MeV wi h a minimum >0.5PE/MeV,which
co esponds o a collec ion e iciency o 1.3% and 2.6%,
espec i ely [7].
The pho on de ec o sys em o FD1 consis s o ligh col-
lec o modulesplacedin heinac i espacebe ween heinne -
mos wi e planes o he anode planes.
These la ge-a ea ligh collec o s con e inciden scin il-
la ion pho ons in o pho ons in he isible ange ha end up in
pho o-senso s. A e in es iga ing many ligh collec o mod-
ules, DUNE has success ully eached a design o a ligh
ap ha can be deployed as ei he single- ace o dual- ace
eadou , he so-called X-ARAPUCA. In o de o alida e he
DUNE echnology, p o o ypes a e being designed, buil and
es ed a CERN, he P o oDUNEs, LA TPC p o o ypes o
0.77k o al mass. P o oDUNE-HD will es he inal design
o FD1 componen s a scale 1:1 in 2024. In pa icula , he
X-ARAPUCAs will be es ed o he i s ime he e.
2 The X-ARAPUCA de ice
The X-ARAPUCA (XA) is an imp o emen o he o iginal
concep o pho on apping inside a highly e lec i e box
while using a wa eleng h shi ing (WLS) ba o inc ease
he p obabili y o collec ing apped pho ons on o a SiPM
a ay. The ARAPUCA echnology has been alida ed in he
un-1 o he P o oDUNE de ec o , ha showed i s supe io
pe o mances in he pho on de ec ion e iciency compa ed
o dip-coa ed shi ligh guides (single o double) [8]. In he
XA he en ance window is a glass coa ed on he ex e nal
side wi h a laye o p-Te phenyl (pTP) (∼500µg/cm2) o
con e he inciden 127 nm scin illa ion ligh in o iso op-
ically e-emi ed 350 nm pho ons wi h an e iciency la ge
han 95%. The e-emi ed ligh eaches he WLS ba whe e
i is u he downshi ed o he isible ange (430 nm), as
illus a ed in he schema ic o Fig.2a. The WLS e-emi ed
ligh can be apped by o al in e nal e lec ion (θ>56◦)
o escape he WLS ba (θ<56◦) and be e lec ed by a
dich oic il e and by a e lec i e su ace. The il e is a mul-
ilaye hin- ilm coa ing deposi ed on o he inne side o he
en ance window, whose cu o is a 400 nm (Fig.2b). These
apped pho ons p opaga e owa ds he edges o he module
whe e he silicon pho o-mul iplie (SiPM) pho o-senso s a e
placed. Simula ions o he X-ARAPUCA de ice [9]showed
he e ec i eness o pho on apping inside he ligh guide
and i s dependence on ac o s such as ligh guide bulk ans-
mi ance, su ace e lec i eness and op ical coupling wi h he
SiPMs. In DUNE, wo di e en a ian s will be deployed:
a double sided op ical window a ian , o equip he middle
anode plane which collec s scin illa ion pho ons om bo h
he cen al d i olumes and a single sided a ian o equip
he wo ex e nal anode planes. In his pape he single sided
a ian has been es ed, in which he side opposi e o he
op ical window is an opaque backplane lined inside wi h an
ex ended specula e lec o (ESR). Two di e en models o
he WLS ha e been es ed, he EJ-286 manu ac u ed by Eljen
Technology [10] and he o he cus om designed and manu-
ac u ed by glass o powe (G2P) [11] in collabo a ion wi h
INFN [12]. Figu e2b shows he emission spec a o bo h he
G2PWLSand hep ima ywa eleng hshi e .Ea lie s udies
conduc ed a INFN showed he imp o ed PDE pe o mances
o a smalle (o hal -size) X-ARAPUCA de ice, embedding
he G2P WLS han wi h he EJ286 [13]. The SiPM a ays
a e acing wo o he sides o he WLS ba pe pendicula o
he en ance window. An image aken du ing he moun ing
p ocess o an XA ins alled in Neu ino Pla o m 04 (NP04) is
showed in Fig.2c. The SiPMs a e e enly spaced and an ESR
is placed be ween hem on he le emp y space on he moun -
ing boa ds, so ha pho ons eaching he edge o he module
no hi ing a SiPM a e e lec ed back in o he op ical module.
TheSiPM ac i ea eais 6x6mm while he ligh guide is 4mm
hick so ha 1/3 o he a ea is exposed o he LA allowing
he de ec ion o pho ons apped by he dich oic il e s.
The module dimensions a e 2m long and 12cm wide and
i consis s o 4 XA as shown in Fig.3a, also known as Supe -
cells.InFig.3b we can see ha each XA is composed o
6 dich oic il e s, 1 WLS ba and 48 SiPMs g ouped in 8
PCB boa ds (6 SiPMs each); he bias and signals a e ou ed
123
Eu . Phys. J. C (2024) 84 :1004 Page 3 o 14 1004
Fig. 2 (a) Schema ic o he XA wo king p inciple [14]. (b)The
dich oic il e cu o ( ed dashed line), he pTP (pu ple line) and he
G2P (blue line) emission spec a. (c) Image aken while moun ing an
FD1 XA o dimensions (50 ×12)cm2
o he on end elec onics by signal leading boa ds cus om
designed by INFN Sezione di Milano [15].
Fou con igu a ions o hese XAs depending on he SiPMs
and WLS ba ype will be es ed in P o oDUNE. Two di e -
en models o SiPMs, om Fondazione B uno Kessle (FBK)
and Hamama su Pho onics K.K (HPK), will be employed.
The pa icula models a e: FBK T iple-T ench (TT) [16],
which pixel sizeis abou 50 μm, and HPK (S13360-6075HS-
HRQ) 75 μm High Quenching Resis ance (HQR) [17] bo h
wi h a o al e ec i e a ea o abou 36mm2and speci ically
designed o being used a c yogenics empe a u es (CT).
P o oDUNE-HD also makes use o wo di e en WLS ba s,
he EJ-286PS and he ba cus om designed by G2P. Fo all
he ou con igu a ions OPTO [18] dich oic il e s we e used.
In Table 1we summa ize he ou combina ions ins alled in
P o oDUNE-HD ha ha e been cha ac e ized in his wo k.
3 Me hodology and ins umen a ion
Quan i ying he absolu e e iciency is undamen al o
DUNE, as his pa ame e is needed o ully cha ac e ize he
PDS. The PDE mus be assessed a he ope a i e condi ions,
a CT and wi h 127 nm pho ons. Fo his eason, we sub-
me ged he XA in LA oge he wi h a low-ac i i y elec o-
deposi ed241Amalphasou ce.Twoexpe imen alse upsha e
been es ablished o ca y ou his measu emen , a CIEMAT
(Mad id, Spain) and a Milano-Bicocca Uni e si y (Milan,
Fig. 3 (a) PDS module including 4 XA cells. (b) Schema ic o he
assembly p ocess wi h he di e en componen s [14]
Table 1 The ou XA con igu a ions ins alled in P o oDUNE-HD
SiPMs WLS ba Tes ing si e
(A) FBK TT EJ-286PS-1 CIEMAT + MiB
(B) FBK TT G2P-FB165A MiB
(C) HPK 75HQR EJ-286PS-1 CIEMAT
(D) HPK 75HQR G2P-FB165A CIEMAT + MiB
I aly). To measu e he absolu e e iciency o he XA, he
numbe o pho ons a i ing a i s su ace need o be known.
Two di e en me hods ha e been conside ed: compa ing he
amoun o ligh collec ed by he XA wi h he ligh de ec ed by
a calib a ed pho osenso , and es ima ing he ligh in he XA
om he α-sou ce ene gy and he known numbe o scin illa-
ion pho ons pe MeV p oduced in LA once he solid angle
sus ained by he sou ce is de e mined. Bo h se ups lique y
high pu i y a gon gas (GA ) inside a c yogenic essel.
Theda ap ocessingandacquisi ion(DAQ) sys emcon ig-
u a ion, in ol ing he so wa e and he elec onics, is simila
in bo h se ups and a diag am is shown in Fig.4
The 48 SiPMs a e passi ely g ouped a he inpu o a an-
simpedance ampli ie loca ed a he op o he XA, he so-
called cold-ampli ie . The ampli ie is based on a SiGe bipo-
la ansis o (In ineon BFP640) ollowed by a ully di e -
en ial ope a ional ampli ie (Texas Ins umen s THS4531),
and is designed o low noise a low powe , gi ing a ol age
123
1004 Page 4 o 14 Eu . Phys. J. C (2024) 84 :1004
Fig. 4 Schema ic o he da a acquisi ion sys em
whi e noise densi y o 0.37 nV/√Hz a 2.4 mW pe channel.
To educe he noise ha may be induced in signal cables, he
ou pu o he cold-ampli ie is di e en ial. A oom empe a-
u e, he so-called wa m-elec onics con e s he di e en ial
signal o single ended and in oduces a second ampli ica ion
ac o . The eadou scheme eplica es he one planned o he
DUNE FD1, whe e he di e en ial o single ended con e -
sion is pe o med in AC using a ans o me , as desc ibed
in [19]. This gi es an unde shoo on he ail o he signals,
which will need o be conside ed in he analysis. Bo h he
cold and he wa m elec onics a e he same as in he MiB
and CIEMAT se ups, while he adop ed digi ize di e s, as
epo ed in he ollowing Sec s. 3.1 and 3.2.
3.1 CIEMAT se up
The CIEMAT neu ino g oup made use o a 300l c yogenic
essel wi h di e en concen ic olumes, whose schema ic
is shown in Fig.5. The la ge and ex e nal one (100l) is illed
wi h liquid ni ogen (LN2) and he smalle one (18l) con-
ains he GA and i is whe e he XA is loca ed. In his 18l
con aine , he GA is lique ied by he mal con ac wi h he
LN2o he su ounding olume. This is achie ed by con-
olling p essu e pa ame e s and egula ing he empe a u e
alues necessa y o ca y ou he lique ac ion. The sys em is
designed o pe o m he au oma ic illing o he 100l essel
wi h LN2 om a 400l ank which is a an o e -p essu e o
4 ba . A he end, GA g ade 6.0 is lique ied wi h LN2a
2.7 ba . To educe con amina ion om ma e ial ou gassing,
we pe o m successi e acuum cycles in he essel be o e
in oducing he op ical and elec ical componen s ha will
be used o pe o m he measu emen s. A las acuum cycle
is done once all elemen s a e in place. Each XA con igu a ion
is es ed in da a aking campaigns las ing 3–4 days.
The XA is in oduced in he inne essel oge he wi h wo
e e ence HPK VUV4 SiPMs (S13370-6075CN) [17] and
a pho o-mul iplie ube (PMT) (R6836-Y00) [20] as com-
plemen a y pho o-senso s. The VUV4 SiPMs a e designed
o ha e a high sensi i i y o VUV ligh and s able pe o -
mance a CT, making possible he de ec ion o LA scin illa-
Fig. 5 Se up scheme used o ob aining he absolu e e iciency o he
XA. (a) C yogenic essel wi h i s concen ic olumes whe e GA is
liqui ied. (b) Diag am o he black box holding he senso s oge he
wi h he 241Am sou ce
ion ligh . The PMT is no speci ically designed o measu e
a CT, al hough i is also sensi i e o VUV ligh , allowing us
o s udy he scin illa ion ligh p o ile easily and moni o he
LA pu i y.
Toge he wi h he e e ence de ec o s a 241Am sou ce is
held wi h an opaque box as shown in Fig.5.I emi sα-
pa icleswi h5.485MeV(84.45%)and5.443MeV(13.23%)
ene gies wi h an ac i i y o (54.53 ±0.82) Bq [21]. The
pa icles deposi hei ene gy inside he 4cm sized black
box, and he p oduced pho ons each he XA h ough a hole
(∅= 23mm). We ensu e ha no o he pho ons a e being
de ec ed by co e ing he es o he XA wi h a black shee .
On he o he aces o he box,we place he woVUV4 SiPMs,
he PMT and a di use connec ed o a ibe . The op ical ibe
123
Eu . Phys. J. C (2024) 84 :1004 Page 5 o 14 1004
Table 2 Rela i e dis ances om he α-sou ce o he senso s and sizes
o he box’s holes
XA Re . SiPM
E ec i e a ea (mm2) 415.47 36.00
Dis ance o he sou ce (mm) 29.0 ±0.7 26.9 ±0.3
p o ides ligh om LEDs and lase s o check he esponse o
he sys em. In Table 2we summa ize he de ec o s’ dimen-
sions and posi ions in he box.
The VUV4 SiPMs we e calib a ed a oom empe a u e
by he manu ac u e ; howe e , se e al s udies [22–24]ha e
shown a dec ease o abou 50% in he PDE a CT and a
dependence wi h he inciden angle. Conside ing he esul s
o heses udiesand hemeasu emen sca iedou inCIEMAT
labs [25]wi h hese senso s exposed o VUV ligh a di e en
angles, we will assume a PDE o (11.17 ±1.3)% a 127 nm
and, 87 K and 4 V o e ol age (OV) o he model S13370-
6075CN o VUV4 SiPMs om HPK. The c oss- alk p ob-
abili y compu ed in [25] o he e e ence SiPMs a CT is
PXT =(14.84 ±0.24)% which is in pe ec ag eemen wi h
he esul p esen ed in [22].
The XA signal p o ided by he wa m-elec onics is digi-
ized by he ADC (model CAEN DT5725S [26]). The signals
o he wo e e ence SiPMs a e ampli ied and digi ized by he
same CAEN module. The igge is done a ADC le el and
is p o ided by he signal in coincidence o he wo e e ence
SiPMs. The inal ou pu consis s o 20 μs wa e o ms wi h
4 ns sampling. A p e- igge o 2 μs allows he de e mina-
ion o he baseline on a e en -by-e en basis. Fo calib a ion
uns, a pulse gene a o p o ides an ex e nal igge and syn-
ch onously pulses he ligh sou ce (lase o LED). Th ee-μs
wa e o ms a e eco ded in he calib a ion uns wi h he same
ime sampling.
3.2 INFN Milano Bicocca se up
The INFN Milano-Bicocca (MiB) se up is an ex ension o
he one used o he PDE measu emen s o he XA de ice
adop ed by he SBND p ojec [27]. The wo k [13] allowed
us o p ecisely assess he supe io pe o mance o he SBND-
XAde ice equippedwi h hecus omp oducedPMMA based
WLS [11], ha is now adop ed as he baseline componen o
bo h heDUNE FD1andFD2 Pho onDe ec ionsys ems. The
se up, he p ocedu es and he me hod a e desc ibed in de ail
in [13].
Figu e 6shows he XA loca ed a he cen e o he s ain-
less s eel ∼25l cylind ical chambe o 250mm diame e and
550mm heigh . The closed chambe is loca ed in an open 70l
dewa . Thechambe is i s ou gassed down o O(10−4)mba
and hen connec ed o a bo le o GA g ade 6.0. The open
dewa is illed wi h LA and he GA lique ac ion p ocess
Fig. 6 (a) The scheme o he INFN-MiB se up o he longi udinal
scanning and PDE de e mina ion o he XA. (b) Pic u e be o e closing
he chambe
s a s inside he chambe . Bo h he GA low and i s lique-
ac ion a e in he chambe a e sus ained by he egula ion
o he bo le p essu e educe . The exposed 241Am α-sou ce
(3.7 kBq) is moun ed on he ip o a magne ic manipula o
( o o asla o ) and allowed o slide on a e ical ail, acing
he XA a he dis ance o (55±1) mm. This allows o scan he
PDEo heXAalongi sz-axisand omoni o wi hhighp eci-
sion he LA le el inside he chambe du ing he whole illing
p ocess, as he alpha ligh pulse ampli ude g ea ly inc eases
when he sou ce is in LA .
The a ge ope a ional condi ion is eached when he XA
and he on end eadou ci cui a e subme ged in LA . Due
o he lacko GA ,in ew uns wecouldn’ each he opo he
de ice, hence only he da a om he scanning posi ions ully
subme ged by LA a e conside ed in he PDE da a analysis.
The digi ize (CAEN DT5725 250 MS/sec 14 bi s) is sel -
igge ed by se ing a h eshold, ha p o ides a igge a e
o abou 1 kHz o he α-pa icles and 100 Hz o he muons
uns espec i ely.
123
1004 Page 6 o 14 Eu . Phys. J. C (2024) 84 :1004
Fig. 7 Example o e en s gene a ed o (a) CIEMAT and (b) Milano-
Bicocca se ups
3.3 Mon e Ca lo simula ions
To p ope ly assess he PDE o he XA p o o ype, a dedica ed
GEANT4 [28] Mon e Ca lo (MC) simula ion was de eloped
o each se up. Fo bo h o hem, he scin illa ion pho ons
gene a ed by he alpha pa icle ene gy loss in LA a e emi -
ed uni o mly and iso opically, and he MC p o ides he
numbe o pho ons eaching he XA accep ance window. Fo
he CIEMAT se up he e e ence pho osenso s solid angles
a e also de e mined by he simula ions and o bo h MiB
and CIEMAT he geome ical accep ances unce ain ies a e
compu ed by a ying bo h he posi ion and dimensions o he
se up elemen s wi hin hei p ecision e o s.
Figu e 7a shows he simula ion o he CIEMAT se up
wi h he dimensions p esen ed in Table 2. The wo e e ence
VUV4 SiPMs ( ed), he PMT (blue) and he XA (g een) a e
con igu ed as sensi i e ma e ials o e ie e he numbe o
de ec ed pho ons depending on hei posi ion wi h espec
o he sou ce. The su ounding black box is designed wi h
a black plas ic ma e ial ha ully abso bs he pho ons. Each
e en o he alpha sou ce is simula ed o ha e he numbe o
pho ons in a andom posi ion wi hin he sensi i e a ea. We
ha e de e mined a sys ema ic e o due o he geome ical
accep ance unce ain y o 10.8%.
Figu e 7b shows a side iew o he XA long edge side o
he Milano Bicocca MC geome y. Abou 500 pho ons a e
d awn o isualiza ion pu poses. The 241Am deposi ed su -
ace exposed o LA is embedded in a sou ce holde whose
shape and size is included in he MC model o de e mine
he e ec i e ligh cone. The geome ical accep ance is hen
d i en by he de ice- o-sou ce dis ance, ha is measu ed
(5.5±0.1)cm, along he whole ail leng h. The unce ain y
on he geome ical accep ance accoun s o ∼7% and ep-
Fig. 8 (a) Pe sis ence his og am o selec ed wa e o ms oge he wi h
hea e agewa e o mo onesinglePE.(b)Cha gehis og amo selec ed
peaks, he gainisde inedby hedis ancebe ween he i s and hesecond
peak
esen s he majo sys ema ic e o o he MiB me hod. The
LA op ical p ope ies a e simula ed oo bu o op ical pa h
in LA o (O(10 cm)) bo h he abso bance and he Rayleigh
sca e ing a e negligible.
4 Da a analysis
4.1 Calib a ion
To measu e he PDE, he ligh pulse gene a ed in LA by
241Am alpha pa icles and collec ed by he pho osenso s
mus be calib a ed by he in eg a ed cha ge o he single
pho on elec on, i.e. by he gain ac o o he SiPMs. The
gain depends on he ope a i e empe a u e and bias ol age,
and can be a ec ed by he a igue e ec , like o he pho o-
123
Eu . Phys. J. C (2024) 84 :1004 Page 7 o 14 1004
Table 3 Expe imen ally measu ed SNR o he XA SiPMs a a gi en
o e ol age (OV) and co esponding pho on de ec ion e iciency (PDE)
OV PDE SNR
(a) FBK TT
3.5 40 2.40 ±0.08
4.5 45 3.42 ±0.07
7.0 50 3.76 ±0.05
(b) HPK HQR75
2.0 40 2.90 ±0.03
2.5 45 3.55 ±0.02
3.0 50 4.24 ±0.02
mul iplie ubes [29]. In bo h he CIEMAT and he MiB se up
he calib a ions a e pe o med wi h low in ensi y blue ligh
emi ing sou ces.
Typical calib a ion wa e o ms o he XA a e displayed
in Fig.8a. The gain is de e mined om he in eg a ed cha ge
dis ibu ion, as shown as an example in Fig.8b and de ined as
Gain =(μ2−μ1)whe eμnis hemean alueo heGaussian
co esponding o npho oelec ons. The good signal- o-noise
a io(SNR)allows he iden i ica iono hepeaks co espond-
ing o 1, 2, 3 o N PE and hen he gain de e mina ion.
The SNR quali ies he capabili y o de ec a single
PE (SPE) o e he sys em noise: i depends bo h on he
XA elec onics and on he se up ela ed dis u bances. We
de ine SNR =μ1−μ0
σ2
0+σ2
1
om he in eg a ed cha ge dis ibu-
ion whe e σnis he Gaussian wid h o he n h peak (0 is he
baseline noise peak, 1 is he 1PE peak). Fo each pho osenso
ype, Table 3 epo s he SNR measu ed o h ee o e ol age
bias alues. In all cases a SNR >2 was measu ed, hence in
all he measu emen s he SPE de ec ion capabili y is e i ied.
Figu e 9 ea u es he measu ed gain e sus o e - ol age
o XAs equipped wi h di e en SiPM models. The gain is a
cha ac e is ic o he SiPM and independen o he es o he
XA elemen s.
Dedica ed c oss- alk s udies ha e been pe o med by he
PDS Conso ium o he wo SiPMs models deployed in he
XAs.Themeasu ed c oss- alk p obabili ies(PXT, p obabili y
o ha e a second pixel ac i a ion a e a ue pho o elec on)
a e p esen ed in Table 4 o h ee OV bias alues and a e used
o compu e he co ec ion o he e iciency.
The c oss- alk co ec ion ac o s ( XT) used o he anal-
ysis a e compu ed as ollows:
XT =1
1+PXT ±PXT
(1+PXT)2.(1)
To asses he XA PDE and compa e he pe o mances o
he di e en con igu a ions, we choose he bias OV alue o
4.5 OV o FBK TT and 3.0 OV o HPK HQR75, since o
Fig. 9 Gain e susbias ol age.Dashedlines ep esen linea i so he
expe imen al poin s. Uni con e sion o ge adimensional gain alues
was made using he ampli ie ’s gain and he elec on cha ge
Table 4 Expe imen ally measu ed c oss- alk p obabili ies a CT
OV PDE PXT (%)
(a) FBK TT
3.5 40 12.68 ±0.27
4.5 45 16.05 ±0.32
7.0 50 32.47 ±0.47
(b) HPK HQR75 om [30]
2.0 40 6.6 ±0.7
2.5 45 9.0 ±1.0
3.0 50 11.0 ±1.0
hese alues he wo models exhibi simila gain (see Fig.9)
while keeping he PXT <20%.
Thea e agewa e o ms o he SPEa e compu ed o each
XAcon igu a ion and he esul s a eshown inFig.10 o bo h
MIB and CIEMAT se ups.
The on -end elec onics p o ides bipola signals wi h
a cha ac e is ic ze o c ossing ime o 1.0 µs. Due o his
beha iou we in eg a e only he posi i e pa o he signal.
The LA ligh emission ime p o ile has a iple componen
wi hcha ac e is ic emission imeτ ∼1.6µs>1.0µs,so o
assess he in eg al o he cha ge los in he nega i e lobe wo
me hods a e conside ed: ei he he alpha/muon wa e o ms
a e decon ol ed (CIEMAT) by he SPE esponse, o he scin-
illa ion ime p o ile is con olu ed wi h he SPE esponse
(MiB). The ac ion o posi i e lobe cha ge is co ec ed a e
his compu a ion.
123
1004 Page 8 o 14 Eu . Phys. J. C (2024) 84 :1004
Fig. 10 Single pho o-elec on no malized esponse ob ained du ing
calib a ion in CIEMAT and Milano-Bicocca se ups. XAs wi h (a)FBK
and (b) HPK SiPMs
4.2 Milano Bicocca analysis
In he me hod adop ed a MiB and used as seconda y me hod
byCIEMAT, heXAe iciency(MiB(XA))iscompu ed om
he a io o he de ec ed (#PE(XA)) o he expec ed (#Ph)
ligh :
MiB(XA)=#PE(XA)
#Ph · co .(2)
whe e #Ph is:
#Ph =LYLA Eα=35700 ph/MeV ·5.48 MeV ·, (3)
The maximal LA ligh yield o alpha pa icles in LA
LYLA =(35700 ±2157)pho ons/MeV including he
αquenching ac o qα=(0.70 ±0.04)is om [31,32],
while he geome ical accep ance () is de e mined by he
Mon e Ca lo simula ions discussed in Sec . 3.3. #PE(XA) is
de i ed om he i o he ull ene gy peak o he calib a ed
alpha spec a as desc ibed la e in his sec ion.
The alpha spec a a e he his og ams o he alpha wa e-
o msselec edbypulseshapedisc imina ionc i e ia (PSD)[13]
and cha ge in eg a ed o e 1000 ns. Figu e11ashows he
capabili y o he MiB se up and me hod o pa icle iden-
i ica ion by PSD when cu ing on he ac ion o he
p omp (cha ge in eg al <600 ns) o e he o al (cha ge in e-
g al <1000 ns), named he ea e p omp . Muons and alphas
a e clea ly sepa a ed in he ( p omp ) s o al cha ge plane:
alphas ha e p omp >0.7 and muons p omp <0.7. Mo e
de ails on he alpha spec a analysis a e p o ided la e on in
his sec ion.
The co ec ion ac o ( co ) is he p oduc
co = XT · in · pu i y.(4)
and akes in o accoun he c oss- alk ( XT), he ac ion o
ligh alling in he wa e o m posi i e lobe ( in ) and he e ec-
i e LA ligh yield ( pu i y), he la e being ela ed [33] o
heligh quenchingimpu i y (e.g.N2)concen a ion ha may
a y a each illing o he expe imen al chambe .
The co ec ion o he e ec i e LA yield (LYe )is ele-
an o he absolu e PDE de e mina ion and o ai ly com-
pa e he PDE o he di e en XA con igu a ions. Up o se -
e al ppm alues he impu i ies a ec only he iple (o
slow) componen o he LA emission ha o alpha pa icles
accoun s o only 23% o he LYLA . The e-no maliza ion
ac o ( pu i y) accoun s o he ac ion o he ac ual iple
(τexp) o he maximal (τpu e) componen [33]
−1
pu i y =Aslow τexp
τpu e +A as ,(5)
whe eA as =0.77andAslow =0.23a e heli e a u esingle
and iple con ibu ions o alphas and τpu e =1600 ns is
he iple ime cons an o he maximal LA LY [13,32].
A each illing o he expe imen al chambe i.e. o each
o he es ed XA con igu a ions, he τexp is ex ac ed om
he muon wa e o ms analysis. A muon un is aken wi h he
sou ce loca ed a he op posi ion o limi he numbe o alpha
e en s igge ing heDAQandmuonswa e o msa eselec ed
by PSD c i e ia: he selec ed muon wa e o ms a e indi id-
ually decon olu ed by he SPE wa e o m empla e and he
esul ing no malized muon a e age wa e o m is i ed by a
wo exponen ial unc ion con olu ed wi h a Gaussian, p o-
iding τexp as shown in Fig.11b. The τexp, hence he quali y
o he LA , is moni o ed and ound o be s able along he
alpha da a aking ime.
The τexp measu ed anges a e epo ed in Table 5 o
each XA con igu a ion oge he wi h he he co esponding
in · pu i y.
123
Eu . Phys. J. C (2024) 84 :1004 Page 9 o 14 1004
Fig. 11 (a) F ac ion o p omp o e o al in eg a ed cha ge o an alpha
un; alpha and muons popula ions a e clea ly dis inguishable by cu ing
on he p omp ac ion. (b) The decon ol ed muon wa e o m measu ed
wi h he MiB se up
Table 5 The LA iple decay ime cons an τexp anges de i ed om
he muons analysis and he co esponding combined co ec ion ac o
o he h ee measu emen s wi h he MiB se up, compu ed o he bes
i in he ange
Measu emen τexp [ns] in · pu i y
(A) FBK +Eljen 910–1113 0.843
(B) FBK +G2P 910–1115 0.843
(D) HPK +G2P 1407–1507 0.853
Fig. 12 Compu a ion o he cha ge in eg a ion co ec ion ac o o
τexp =963 ns. F om op o bo om: simula ed ime p o ile, ime p o ile
con ol ed wi h elec onics esponse, ac ion o in eg a ed cha ge (y
axis) as a unc ion o he in eg a ion ime (x axis). in =92.77% is
ound o a 1000 ns in eg a ion window
The wa e o m bipola shape equi es o assess he ac ual
cha ge in eg a ed wi hin he posi i e lobe o he alpha wa e-
o ms ( in ), ha is in u n an i-co ela ed o pu i y and τexp;
in ac a he inc ease o he la e and since i is g ea e han
he wa e o m ze o c ossing ime, a la ge ac ion o he la e
pho ons a e los since hey all in o he wa e o m nega i e
lobe. The e o e he in · pu i y is nume ically compu ed as
ollows and desc ibed in Fig.12: he alpha pa icle scin illa-
ion ime p o ile o a gi en τexp, ( op panel), is con olu ed
wi h heSPE empla eand p o ides heexpec edpulseshown
in he middle panel. The in eg a ion co ec ion ac o ( in )is
inally epo ed in he Fig.12 (bo om) as a unc ion o he
in eg a ion ime: a each ime, i is he a io o he g ey a ea
o he middle panel ( he cumula i e o he posi i e lobe o he
con olu ed wa e o m) o he o ange a ea o he op panel ( he
in eg al o he en i e aw scin illa ion signal). As an example,
aτexp =963 ns gi es in =92.77% o a 1000 ns in eg a-
ion ime and pu i y =90.84%. The in · pu i y ac o is hen
applied o he alpha pulse heigh spec a o p ope ly asses
he absolu e PDE. The e o on in · pu i y is compu ed by
a ying τpu e and τexp and ound o be <2%; as epo ed in
Table 5 hei p oduc is s able o τexp anging om 900 o
1500 ns.
Figu e 13a shows he module geome ical accep ance as
a unc ion o he sou ce posi ion; he alues a e e ie ed by
he Mon eca lo GEANT4 simula ions desc ibed in Sec . 3.3.
Two calib a ed (by SPE cha ge) alpha spec a a e also shown
in Fig.13b; he di e ence o he de ec ed pho oelec on peak
123
