Con en s lis s a ailable a ScienceDi ec
Nuclea Ins . and Me hods in Physics Resea ch, A
jou nal homepage: www.else ie .com/loca e/nima
New eadou codi ica ion o la ge-a ea mul i-gap iming RPCs o Muon
Sca e ing Tomog aphy
João Sa ai a ∗, Albe o Blanco
LIP, Labo a o y o Ins umen a ion and Expe imen al Pa icle Physics, Po ugal
A R T I C L E I N F O
Keywo ds:
Gaseous de ec o s
Resis i e Pla e Chambe s
Readou codi ica ion
Muon sca e ing omog aphy
Mon e Ca lo simula ions
A B S T R A C T
A new eadou echnique decoupling he numbe o FEE channels om he de ec o a ea was es ed wi h
la ge-a ea Resis i e Pla e Chambe s (RPCs) ea u ing a sensi i e a ea o 130 × 90 cm2. Despi e using only 48
elec onic channels o ead ou 888 pick-up s ips, a 2D submillime ic spa ial p ecision was achie ed du ing a
long un wi h cosmic ays, along wi h a ime p ecision o 89 ps (𝜎). FLUKA Mon e Ca lo simula ions we e also
pe o med o e alua e he use o RPCs wi h high spa ial and empo al p ecision o scanning la ge olumes
employing he Muon Sca e ing Tomog aphy echnique.
1. In oduc ion
A no el eadou sys em o Resis i e Pla e Chambe s (RPCs) was
designed o la ge su ace applica ions equi ing high spa ial and em-
po al p ecision, such as he Muon Sca e ing Tomog aphy (MST). Con-
side ing ha he cos o RPC sys ems is o en d i en by he on -end
elec onics, a new codi ica ion was de eloped o signi ican ly educe
he dependence o he numbe o elec onic channels on he de ec o
a ea, wi hou subs an ially dec easing i s pe o mance.
The new me hod was i s es ed wi h a double-s ack mul i-gap
iming RPC ( RPC) wi h a sensi i e a ea o 30 × 30 cm2, using 24+24
p eampli ie s o ead ou 120+120 s ips. A spa ial p ecision be e
han 1 mm and ime p ecision below 100 ps we e achie ed wi h his
p o o ype, as epo ed in [1]. Subsequen measu emen s we e ca ied
ou wi h a de ec o 13 imes la ge in a ea (130 × 90 cm2), using he
same numbe o on -end elec onics (FEE) channels (24+24) o ead
ou a conside ably highe numbe o s ips: 360 longi udinally and 528
ans e sely.
In his pape , we b ie ly desc ibe he new eadou codi ica ion and
epo he esul s ob ained wi h he la ge-a ea RPCs. Mon e Ca lo sim-
ula ions a e also p esen ed, highligh ing he impo ance o a millime e
spa ial p ecision and e y good ime p ecision o he MST echnique.
2. Readou sys em
The new eadou me hod in ol es wo ypes o eadou boa ds,
one o which ha ing much wide s ips han he o he : he hin-s ip
eadou elec odes wi h a pi ch o 2.54 mm and in e s ip o 1 mm,
and he wide-s ip eadou boa d wi h a 61-mm pi ch and in e s ip
∗Co esponding au ho .
E-mail add ess: [email p o ec ed] (J. Sa ai a).
o 2 mm (see Fig. 1). The hin-s ip eadou elec ode p o ides, ia
cha ge in e pola ion, he ine posi ion o e en s in one dimension. Two
o hem we e used, wi h he s ips o ien ed in o hogonal di ec ions, o
achie e a ine 2D posi ion. The wide-s ip eadou elec ode, in u n,
simul aneously p o ides he ime and an addi ional 2D coa se posi ion
o in e ac ions, as desc ibed u he . These pick-up elec odes a e used
in combina ion wi h a double-s ack RPC, wi h he hin-s ip boa ds
posi ioned a he op and bo om o he s ack, and he wide-s ip boa d
placed a he cen e , be ween he RPCs ( e e o he laye diag am in
he nex sec ion o mo e de ails).
To educe he numbe o elec onic channels, each channel eads
ou N hin s ips p e iously connec ed in pa allel by he Signal Me ging
PCB (SMPCB), as he one depic ed in Fig. 1. The SMPCB ou es signals
om N s ips in o a single ack, ansmi ing hem o he espec i e
elec onic channel wi hou any p ocessing. To p o ec he FEE, all s ips
on he SMPCB we e g ounded h ough 20 MΩ esis o s, p e en ing
elec ical loa ing. Wi h his no el app oach, he sensi i e a ea o a
de ec o can be signi ican ly inc eased while main aining he same
numbe o p eampli ie s. This is accomplished by simply modi ying he
numbe o hin s ips connec ed in pa allel (N).
To de e mine in which hin s ip he signal was in ac induced,
he wide-s ip eadou elec ode mus be added o he se up. Wi h bo h
ex emi ies o he wide s ips connec ed o cu en -sensi i e ampli ie s,
he addi ional eadou boa d p o ides a 2D coa se posi ion o he
in e ac ions in he de ec o , needed o esol e he ambigui y aised
by adding hin s ips in pa allel. I also enables measu ing he ime o
e en s, use ul o Time O Fligh (TOF) applica ions.
h ps://doi.o g/10.1016/j.nima.2025.170466
Recei ed 29 Janua y 2025; Recei ed in e ised o m 17 Ma ch 2025; Accep ed 24 Ma ch 2025
Nuclea Ins umen s and Me hods in Physics Resea ch A 1076 (2025) 170466
A ailable online 2 Ap il 2025
0168-9002/© 2025 Else ie B.V. All igh s a e ese ed, including hose o ex and da a mining, AI aining, and simila echnologies.
J. Sa ai a and A. Blanco
Fig. 1. (a) Example o pick-up elec odes equi ed o he new eadou codi ica ion
and es ed wi h a 30 × 30 cm2 p o o ype [1]. Le : hin-s ip PCB wi h 120 s ips
measu ing 36 cm in leng h; igh : wide-s ip PCB wi h 5 s ips 38 cm long. (b) Signal
Me ging PCB (SMPCB) wi h he bo om iew showing bo h sides o he boa d and how
s ips a e wi ed in pa allel.
The wide-s ip pick-up elec ode mus be designed in such a way
ha he s ip pi ch co esponds app oxima ely o he wid h o he
g oup o hin s ips connec ed o sepa a ed elec onic channels: 24
hin s ips o he es s he e epo ed.1 Mo eo e , while he numbe
o FEE channels connec ed o he hin s ips emains unchanged when
inc easing he de ec o a ea, new channels mus be added o ead ou
he addi ional wide s ips p esen in he la ge de ec o . Howe e ,
because o he signi ican wid h o hese s ips, he inc ease in he
numbe o elec onic channels can be conside ed esidual.
3. Expe imen al se up
3.1. La ge-a ea de ec o
As ini ially desc ibed in [1], he new eadou codi ica ion based on
he SMPCB was i s es ed wi h a 30 × 30 cm2 p o o ype. The same 24-
channel SMPCB was a e wa ds used wi h he 130 × 90 cm2 de ec o
shown in Fig. 2.
Bo h de ec o s ha e an iden ical laye ed s uc u e, namely:
•a s ack o wo mul i-gap RPCs each wi h 6 gas gaps, 300 μm wide;
• esis i e elec odes made o loa glass, 1.1 mm hick;
• wo hin-s ip eadou PCBs wi h s ips o ien ed o hogonally o
achie e a 2D ine posi ion o ionizing e en s, one loca ed a he
bo om and he o he a he op o he s ack;
•a wide-s ip eadou PCB a he cen e o he s ack, o an ad-
di ional 2D coa se posi ion o e en s and p o iding also he
espec i e ime o e en s.
Wi h he 30 × 30 cm2 p o o ype, 24 elec onic channels we e
needed o ead ou 24 g oups o N= 5 hin s ips wi ed in pa allel.2
124 hin s ips × 2.54 mm/s ip = ∼61 mm, co esponding o he pi ch o
he wide-s ip eadou PCB.
2A o al o 24 × 5 = 120 s ips pe hin-s ip eadou boa d.
Fig. 2. Pa ial iew o he la ge-a ea de ec o (ac i e a ea o 130 × 90 cm2) used wi h
he no el eadou echnique. A he bo om o he s ack: h ee SMPCBs connec ed in a
daisy-chain con igu a ion, c ea ing 24 g oups o 15 hin s ips wi ed in pa allel (360
longi udinal hin s ips in o al); he wide-s ip eadou boa d is also isible jus below
he op RPC. The hin-s ip eadou boa d loca ed a he op o he s ack was emo ed
o be e isibili y o he unde lying laye s.
Fig. 3. Le : laye diag am o he la ge-a ea de ec o (1.17 m2). Righ : de ailed iew o
he s acked laye s ( op hin-s ip eadou boa d emo ed). The daisy-chain link be ween
wo SMPCBs is also isible (black jumpe s a he op o he igu e).
Fo he 130 × 90 cm2 de ec o , in u n, N inc eased om 5 o 15 o
he hin-s ip eadou boa d wi h longi udinal s ips (X) and om 5 o
22 o he boa d wi h ans e se s ips (Y). In his way, despi e he
subs an ial inc ease in a ea, he same 24 channels a e employed o
each dimension (X, Y) o ead ou he same 24 g oups o hin s ips,
hough each has addi ional s ips connec ed in pa allel.3
As p e iously men ioned, a esidual inc ease o elec onic channels
occu s due o he addi ional wide s ips equi ed wi h he la ge de ec-
o : only 20 addi ional channels we e needed ( om 58 o 78 channels,4
i.e. a 1.3- old inc ease), while he de ec o a ea was scaled up by a
ac o o 13 ( om 0.09 o 1.17 m2) and he numbe o s ips, equi ed
o ine spa ial measu emen s, inc eased om 240 o 888.
A close-up iew o he SMPCBs used wi h he wide-a ea de ec o
and i s laye diag am can be seen in Fig. 3. The de ec o was ope a ed,
wi h a gas mix u e o 99% o R-134a and 1% o SF6, du ing weeks
using a coincidence igge gene a ed ex e nally by plas ic scin illa o s5
(Bic on BC420) loca ed ew cen ime e s below and abo e he de ec o .
The educed elec ic ield was se o app oxima ely 360 Td (a ound 2.7
kV/gap, 90 kV/cm) leading o an e iciency close o 95%.
3A o al o 24 × 15 = 360 longi udinal hin s ips and 24 × 22 = 528
ans e se hin s ips.
4Numbe o elec onic channels o he 30 × 30 cm2 p o o ype: 24+24+5+5
= 58, namely: 24 channels o he longi udinal hin s ips (X), 24 o he ans-
e se ones (Y) and 5+5 o bo h ends o 5 wide s ips. Fo he 130 × 90 cm2
de ec o : 24+24+15+15 = 78, i.e. only 10+10 addi ional channels we e
needed o ead ou he ex a 10 wide s ips.
5Two pa allelepipeds o 8 × 3 × 2 cm3, coupled o pho omul iplie ubes
Hamama su H6533.
Nuclea Ins . and Me hods in Physics Resea ch, A 1076 (2025) 170466
2
J. Sa ai a and A. Blanco
Table 1
De ailed in o ma ion on he ype o p eampli ie s connec ed o he eadou elec odes,
as well as he measu ed and ex ac ed quan i ies wi h he espec i e FEE.
Cha ge-sensi i e amp Cu en -sensi i e amp
Pick-up elec ode Thin-s ip PCB Wide-s ip PCB
S ip eadou Single-ended Double-endeda
Measu ed quan i y Cha ge Cha ge, Time (T)
Ex ac ed quan i y 1D ine posi ion; wi h
wo o hogonal boa ds:
(X𝑓𝑖𝑛𝑒 , Y𝑓𝑖𝑛𝑒 )b
2D coa se posi ion:
X𝑐𝑜𝑎𝑟𝑠𝑒b, Y𝑐𝑜𝑎𝑟𝑠𝑒 = (T𝑓−T𝑏)/2
e en ime: T = (T𝑓+T𝑏)/2
a F on ( ) & back (b) sides o he s ips.
b Via cha ge in e pola ion.
3.2. Elec onics & da a acquisi ion sys em
Two ypes o FEE we e used wi h he new eadou me hod he e
desc ibed: (i) cha ge-sensi i e ampli ie s connec ed o he hin-s ip
eadou elec odes and used o in eg a e he induced signals o e ew
mic oseconds, co esponding o he ime needed o bo h componen s
o he gene a ed ion-elec on pai s o each he espec i e esis i e
elec odes; (ii) cu en -sensi i e ampli ie s o he eadou o he wide-
s ip pick-up elec ode, p o iding ime and cha ge measu emen s using
only he elec on-domina ed pa o he induced signals. The Measu ed
and de i ed quan i ies achie ed wi h his se up a e summa ized in
Table 1.
Fo he hin s ips, a cus om-designed FEE was chosen elying on he
ul alow-noise ope a ional ampli ie AD8599 con igu ed as in eg a o .
Fea u ing a 10 MHz bandwid h, he p eamp is he e o e insensi i e o
signal e lec ions ha migh occu by he ac o ha ing ins umen ed
only one end o he hin s ips. Wi h an in eg a ion ime cons an se o
100 μs, he cha ge induced by bo h he elec onic and ionic componen s
o he signal is collec ed, maximizing in his way he Signal Noise
Ra io (SNR) and he e o e enhancing he p ecision compu ing he e en
posi ion.
The FEE o he HADES-RPC TOF expe imen (GSI, Da ms ad ) was
used o he eadou o he wide s ips. I includes a as 2.1 GHz
BGM1013 p eamp eeding a high-speed MAX9601 compa a o ha
ou pu s a digi al signal p o iding he e en ime wi h a p ecision
below 35 ps (𝜎) [2]. The p eampli ied signal is also ou ed h ough
a second b anch and hen in eg a ed by a OPA690 ampli ie wi h an
RC ime cons an o 40 ns, allowing o a Time o e Th eshold (ToT)
measu emen o he in eg a ed signal. In his way, he digi al signal
gene a ed by he FEE includes bo h he ime and cha ge in o ma ion,
encoded in he leading edge and he wid h o he signal, espec i ely.
The da a acquisi ion sys em is based on he mul i-pu pose T igge
Readou Boa d - 3 (TRB3) [3], an FPGA-based pla o m wi h ou
pe iphe al FPGAs wi h a 32-channel Time- o-Digi al Con e e (TDC)
implemen ed in each o hem, ea u ing a ime p ecision below 20
ps [4]. An Analog- o-Digi al Con e e (ADC) expansion ca d6 was
connec ed o one o he pe iphe al FPGAs o digi ize he in eg a ed
signals om he cha ge-sensi i e ampli ie s.
4. Expe imen al esul s
As p e iously men ioned, he main goal o he eadou echnique
desc ibed in his wo k, is o decouple he numbe o FEE channels
om he eadou a ea, allowing o an inc ease o he su ace o
he de ec o wi hou subs an ial change in he numbe o elec onic
channels. To assess he impac o his me hod on de ec o pe o mance,
wo measu emen s we e compa ed be ween he 30 × 30 cm2 p o o ype
and he 130 × 90 cm2 RPCs he e p esen ed: (i) he ime p ecision o
6Equipped wi h wel e 4-channel AD9219, ea u ing a esolu ion o 10 bi s
and sampling a e o 40 MHz.
Fig. 4. Time p ecision o he 130 × 90 cm2 RPCs, dec easing om 109 o 89 ps (𝜎)
a e emo al o he scin illa o con ibu ion.
bo h de ec o s, in e ed by sub ac ing he measu ed ime o he RPCs
om ha o he scin illa o s used o he coincidence igge ; (ii) he
de ails ob ained in he econs uc ed image o he scin illa o s p ojec ed
on he RPCs.
The ime p ecision o he la ge-a ea RPCs, achie ed o e a 50-day
acquisi ion pe iod, is p esen ed in Fig. 4. A e applying he ime walk
co ec ion and emo ing he scin illa o con ibu ion,7 a inal p ecision
o 89 ps (𝜎) was achie ed, a ew ens o picoseconds abo e he alue
ob ained wi h he i s p o o ype [1]. The sligh deg ada ion in ime
p ecision obse ed wi h he la ge de ec o migh simply be a ibu ed
o he ini ial condi ioning o he RPCs, as he p ecision sys ema ically
imp o ed o e ime, eaching 73 ps (𝜎) when conside ing only he las
10 days o acquisi ion.
The p ojec ed shadow o he scin illa o s on he RPCs is shown
in Fig. 5. The p esence o 300 μm-diame e ishing lines be ween he
glass elec odes o he RPCs esul s in he comple e inhibi ion o
gas ioniza ion a his loca ion. The absence o e en s a he space
loca ion in he econs uc ed image o he scin illa o s was he e o e
expec ed, indica ing ha he spa ial p ecision o he RPCs should be
submillime ic. This obse a ion leads us o conclude ha he inc ease
in elec onic noise, esul ing di ec ly om adding s ips in pa allel wi h
SMPCBs,8 was no signi ican enough o comp omise he de ec ion o
he 300 μm space s.
5. Mon e Ca lo simula ions applied o muon omog aphy
5.1. Muon Sca e ing Tomog aphy (MST)
Ma e ials o high a omic numbe (Z) concealed in medium-sized
olumes can be iden i ied by he Muon Sca e ing Tomog aphy (MST)
echnique, as i s epo ed in 2003 [5]. Based on Mul iple Coulomb
Sca e ing (MCS) o na u ally occu ing cosmic- ay muons, his ech-
nique enables mapping he in e nal s uc u e o he scanned objec . To
econs uc he sca e ing e en s wi hin he iducial egion, a minimum
o wo poin s along he muon pa h mus be measu ed, bo h ups eam
and downs eam o he analyzed objec . A s ack o ou de ec o s is
he e o e equi ed, each wi h sensi i e a eas anging om a ew o
ens o squa e me e s, esul ing ine i ably in a la ge numbe o FEE
channels.
7Ob ained by compu ing he ime di e ences be ween de ec o s ( RPC and
scin illa o s 1 and 2 (SC1, SC2)) and sol ing he ollowing sys em o equa ions:
(1) 𝜎2
𝑡𝑅𝑃 𝐶 +𝜎2
𝑆𝐶1=𝜎2
𝑡𝑅𝑃 𝐶−𝑆𝐶1,(2) 𝜎2
𝑡𝑅𝑃 𝐶 +𝜎2
𝑆𝐶2=𝜎2
𝑡𝑅𝑃 𝐶−𝑆𝐶2,(3) 𝜎2
𝑆𝐶1+𝜎2
𝑆𝐶2=
𝜎2
𝑆𝐶1−𝑆𝐶2.
8Up o 22 s ips o 90 cm linked in pa allel, co esponding o a o al s ip
leng h o almos 20 m.
Nuclea Ins . and Me hods in Physics Resea ch, A 1076 (2025) 170466
3
J. Sa ai a and A. Blanco
Fig. 5. (a) 2D posi ion map o he 8 × 3 × 2 cm3 scin illa o s. Posi ioned wi h he
3 cm side aligned e ically, he scin illa o s p ojec a shadow o app oxima ely 8 cm
× 2 cm on he RPCs. The ed a ow iden i ies he loca ion o he nylon mono ilamen
space s ins alled longi udinally be ween he esis i e elec odes. (b) p ojec ion o he
2D e en map on o he X axis, showing a clea educ ion in he numbe o e en s a
he loca ion o he 300-μm gas gap space s.
RPCs wi h almos 2 m2 we e used in 2021 o in e he p esence o
small blocks o aluminum, i on and ungs en wi hin a scanned olume
sligh ly below 1 m3 [6]. A ha ime, despi e he de ec o no being
speci ically designed o he MST echnique, a 5 cm- hick ungs en
block was iden i ied in 10 min. The de ec o was equipped wi h 512
elec onic channels, ea u ing a spa ial p ecision a he cen ime e le el.
Wi h he eadou me hod p esen ed in his communica ion, a spa ial
p ecision on he submillime e scale could ha e been achie ed wi h a
conside ably lowe numbe o FEE channels.
5.2. Mon e Ca lo simula ions
To assess he impac o high spa ial and empo al p ecision on he
MST me hod, se e al Mon e Ca lo simula ions we e conduc ed using
he ollowing me hodology:
•gene a e he muon lux a sea le el using wo al e na i e me hods:
– he FLUKA a mosphe ic model composed o 100 sphe ical
laye s om 0 o 70 km abo e sea le el, each wi h di e en
ai densi ies [7,8];
–a muon gene a o de eloped o ea u e a 𝑐𝑜𝑠2(𝜃) zeni h
angle dis ibu ion and a di e en ial ene gy spec um p o-
po ional o 𝐸−2.7 ( la below 2 GeV);
•a e alida ion,9 any o he muon luxes could be loaded as sou ce
e m and anspo ed in second s ep simula ions using he FLUKA
code. Two di e en geome ies we e c ea ed10 o his pu pose
(see Fig. 6):
–a mid-sized geome y ( iducial egion o ∼0.9 m3), equi a-
len o he muon elescope used in 2021 [6];
–an ex ensi e geome y ( iducial egion o ∼154 m3), co e-
sponding o a ull-size uck ca ying a la ge sea con aine ;
• o each simula ed scena io, he ene gy and coo dina es o e e y
muon c ossing a de ec o plane we e dumped using a cus omized
mgd aw. FLUKA sub ou ine. Se e al simula ions we e ca ied ou
wi h di e en geome y a ian s, namely:
– he ull geome y: including a 10 × 10 × 10 cm3 block o
high-Z ma e ial,11 4 RPCs, he su ounding ai , and in case
o he la ge geome y: he loo , he uck pla o m and he
con aine si ua ed abo e i ;
9Being, o ins ance, in e y good ag eemen wi h he muon ene gy
spec um om [9].
10 Using he FLAIR g aphical in e ace [10].
11 Se e al ma e ial we e es ed: u anium, ungs en, i on, e c. The esul s
ob ained wi h ungs en a e he e epo ed.
Fig. 6. Geome ies used o e alua e he e ec o he ma e ial budge and he spa ial
and empo al p ecision on he MST echnique. Le : medium-sized geome y including 4
RPCs, each measu ing 1.6 m × 1.2 m, and a block o high-Z ma e ial inside he olume
be ween he wo inne planes ( iducial egion). Righ : signi ican ly la ge iducial egion
(also wi h 2 de ec o planes on he op and bo om sides), including a 15-cm- hick
conc e e loo , a uck pla o m wi h a hickness o 10 cm and s eel walls 2 mm hick,
and a shipping con aine s also made o s eel, wi h 2 mm hick walls and o dimensions:
13.7 m × 2.5 m × 2.9 m. A 10 × 10 × 10 cm3 high-Z block is also p esen inside he
con aine .
–equi alen o he p e ious geome y, bu wi hou he block
o high a omic numbe . The muon sca e ings gene a ed
om his simula ion can be conside ed as ‘noise’ since e-
la ed o all he ma e ials excep he one o in e es ;
–as opposed o he las con igu a ion, a geome y including
only he high-Z objec gene a es sca e ings exclusi ely e-
la ed o his ma e ial, hus co esponding o he expec ed
‘signal’;
–ano he a ian also conside ed, included only he su ound-
ing ai , o es ima e i s con ibu ion o he o al angula
dis ibu ion o he muons a e c ossing he iducial egion.
• he collec ed da a om he Mon e Ca lo calcula ions could hen
be used o:
–pe o m sca e ing angle dis ibu ions o muons wi hin he
iducial egion, as shown in Fig. 7;
– econs uc he 3D image o he scanned objec plo ing
di ec ly he midpoin o he common pe pendicula (also
known as poin o he closes app oach (POCA)) be ween
inciden and exi ing ajec o ies, as illus a ed in Fig. 8;
–es ima e he exposu e ime equi ed o iden i y he high-Z
ma e ial based on he SNR achie ed a e popula ing he
oxelized iducial egion wi h he POCAs, as depic ed in
Fig. 9.
5.3. Simula ion ou comes
5.3.1. Ma e ial budge e ec on angula dis ibu ion
Compa ing sca e ing angle dis ibu ions o simula ions whe e com-
ponen s a e p og essi ely emo ed om he geome y, allowed us o
assess he con ibu ion o each ma e ial o he o al sca e ing o he
muons a e c ossing he iducial egion. Fo he simula ions co e-
sponding o he muon elescope used in 2021 (mid-sized geome y o
Fig. 6), i was ound ha he impac o he RPCs12 on he cumula i e
sca e ing is indeed conside able. In ac , as shown in Fig. 7 (le ),
he MCS on he ungs en block only s a o be dominan o la ge
de lec ions (abo e ∼12◦), mainly because o he p esence o he RPCs.
This esul co obo a ed he expe imen al obse a ions o 2021, whe e
he de ec ion o he ungs en block could be imp o ed disca ding
angula de lec ions up o 11◦ [6], i.e. emo ing a subs an ial numbe o
e en s, and consequen ly inc easing he exposu e ime needed o he
ma e ial iden i ica ion.
12 Each de ec o plane, consis ed o a s ack o 2 mul i-gap RPCs wi h 2 gas
gaps each (1 mm wide) and glass elec odes 2 mm hick, hence a o al o
2 × 3 × 2 mm o glass plus 2 × 2 mm o he aluminum box.
Nuclea Ins . and Me hods in Physics Resea ch, A 1076 (2025) 170466
4
J. Sa ai a and A. Blanco
Fig. 7. (a) Sca e ing angle dis ibu ions as a unc ion o ma e ial budge o he mid-
sized geome y. (b) Same dis ibu ions bu emo ing muons wi h ene gy up o 500 MeV.
Fig. 8. 3D econs uc ion o he scanned uck, con aine and ungs en block
(10 × 10 × 10 cm3), o an exposu e ime co esponding o a du a ion o 55 min.
The POCAs we e plo ed di ec ly in he 3D olume, applying he ollowing es ic ions:
only sca e ing angles abo e 5.5◦ (le ), and emo ing also muons wi h ene gy below
1 GeV ( igh ). An equi alen econs uc ion image o he mid-sized geome y can be
ound in [6].
The simula ions also showed ha , by ejec ing low-ene gy muons (a
ew hund ed o MeV), he sca e ings om he ungs en block become
dominan o e almos he en i e sca e ing ange (Fig. 7 ( igh )). Since
muons o low ene gy a e highly sca e ed, e en in he su ounding ai ,
emo ing hem imp o es no only he econs uc ed image bu also he
exposu e ime equi ed o de ec high-Z ma e ials, as u he discussed.
Fo he muon elescope used in 2021, i would ha e conside ably
educed he ime necessa y o he ungs en iden i ica ion.
5.3.2. TOF o low-ene gy muon ejec ion
The TOF echnique can be employed o he a o emen ioned pu -
pose o emo ing low-ene gy muons. Fo ins ance, wi h a ime p ecision
o 100 ps (simila o he p ecision achie ed wi h he new eadou
me hod he e p esen ed), muons wi h ene gy up o 500 MeV can be
ejec ed o he medium-sized geome y o Fig. 6 and up o 1 GeV in
case o he uck geome y, whe e he dis ance be ween ou e planes
exceeds 5 m. The la ge dis ance be ween de ec o s also imp o es
ela i e momen um esolu ion.
Fig. 8 shows he econs uc ed image o he uck and i s con aine
wi h he ungs en block inside, a e mo e han 50 min o acquisi ion.
As p e iously s a ed, emo ing muons wi h ene gy below 1 GeV (Fig. 8
( igh )) signi ican ly imp o es he image o he scanned objec , making
i possible o iden i y he di e en elemen s o he geome y: he
conc e e loo , abo e i he uck deck suppo ing he con aine , he
walls o he con aine (be e obse ed om a on al pe spec i e) and
he ungs en block a he cen e .
5.3.3. Time equi ed o iden i y he ungs en block
To es ima e he exposu e ime necessa y o iden i y he ungs en
block in bo h es ed geome ies, he ollowing app oach was es ab-
lished:
(i) disc e ize he iducial egion in o oxels: 5 × 5 × 5 cm3 o he
mid-sized geome y and 10 × 10 × 10 cm3 o he ex ensi e one;
(ii) add a Gaussian noise o he muon posi ions o emula e he
de ec o ’s spa ial p ecision;
(iii) compu e he POCAs and alloca e hem o he co esponding
oxels;
Fig. 9. SNR ( a io be ween he highes alues o he no malized numbe o POCAs
pe oxel, inside and ou side he co ec loca ion o he high-Z ma e ial) s. exposu e
ime in minu es: (a) calcula ed o he mid-sized geome y, wi h spa ial p ecisions
up o 1.0 cm; (b) ob ained o he uck geome y using he same de ec o ’s spa ial
p ecisions. An addi ional cu e was added showing he signi ican SNR dec ease when
no es ic ions (in sca e and ene gy) a e applied.
(i ) calcula e he a e age numbe o POCAs pe oxel o iden i y he
oxels wi h signi ican ly highe concen a ions o POCAs;
( ) educe he exposu e ime and epea all he abo e s eps.
Ou lie oxels wi h a numbe o POCAs exceeding a no malized
h eshold can hen be plo ed o iden i y hei loca ion wi hin he
iducial egion. Fo he pe o med simula ions, whe e he posi ion o
he high-Z ma e ial is known, i is possible o assess whe he he plo ed
oxels a e accu a ely loca ed, and compu e he SNR by compa ing he
highes ou lie alues inside and ou side he co ec posi ion. By doing
so, i was obse ed ha :
• o all he simula ions p esen ed in his s udy, and based on he
ob ained SNR (see Fig. 9) bu also on he densi y o ou lie s in
a pa icula egion, one minu e o exposu e ime is enough o
eliably iden i y he 10 × 10 × 10 cm3 ungs en block;
• o bo h conside ed geome ies, a de ec o ’s spa ial p ecision a
he cen ime e le el is no ideal o he MST echnique, as he SNR
emains close o one o all exposu e imes, equi ing o ins ance
a highe sca e ing angle es ic ion;
• o he uck simula ions:
– es ic ions on sca e ing de lec ion and muon ene gy mus
be applied o imp o e he SNR, ein o cing he impo ance
o e y good empo al p ecision o use he TOF echnique;
–a submillime e spa ial esolu ion is no o pa amoun im-
po ance o he achie ed SNR.
6. Conclusion
A no el eadou echnique was de eloped o decouple he numbe o
elec onic channels om he de ec o a ea, wi hou subs an ial deg a-
da ion o i s pe o mance. The app oach elies on he Signal Me ging
PCB ha connec s se e al pick-up s ips in pa allel be o e being ed o a
single FEE channel. The new me hod was es ed wi h mul i-gap RPCs,
130 × 90 cm2 in a ea, whe e up o 22 s ips (each ∼90 cm long) we e
wi ed oge he and ou ed o he same elec onic channel. The RPCs
we e ope a ed in coincidence igge wi h scin illa o s, measu ing he
cosmic ay lux du ing weeks. A ime p ecision o 89 ps (𝜎) was
achie ed, co esponding o a sligh deg ada ion in compa ison o a p e-
iously es ed small p o o ype (30 × 30 cm2) ha used he same numbe
o FEE channels. This ime p ecision imp o ed sys ema ically o e ime,
indica ing ha he obse ed deg ada ion migh be ela ed o he ini ial
condi ioning o he RPCs. Mo eo e , despi e he signi ican numbe o
s ips connec ed in pa allel wi h his echnique, a 2D submillime ic
spa ial p ecision was achie ed.
FLUKA Mon e Ca lo calcula ions applied o he Muon Sca e ing
Tomog aphy we e also pe o med o assess he easibili y and ele ance
Nuclea Ins . and Me hods in Physics Resea ch, A 1076 (2025) 170466
5
J. Sa ai a and A. Blanco
o using RPCs, wi h e y good spa ial and empo al p ecision, o scan
la ge olumes (abo e 150 m3). The simula ions e ealed ha : (i) he
ma e ial budge o he RPCs has an app eciable impac on he mul iple
sca e ings su e ed by he muons, especially he low-ene gy ones; (ii)
o e y la ge geome ies, whe e he single-sca e ing app oxima ion
shows i s limi a ions, a millime e spa ial p ecision is enough o quickly
iden i y high-Z ma e ials, while he submillime e p ecision le el does
no imp o e signi ican ly his capaci y; (iii) a ime p ecision on he
o de o 100 ps is essen ial o o e come he limi a ions men ioned in
he p e ious poin s, allowing o he ejec ion o highly sca e ed low-
ene gy muons ia he TOF echnique. Finally, based on he simula ion
esul s and assuming millime e spa ial p ecision and ime p ecision
a ound 100 ps, an exposu e ime o one minu e was su icien o de ec
a 10 × 10 × 10 cm3 ungs en block in he es ed geome ies.
Decla a ion o compe ing in e es
The au ho s decla e ha hey ha e no known compe ing inan-
cial in e es s o pe sonal ela ionships ha could ha e appea ed o
in luence he wo k epo ed in his pape .
Acknowledgmen s
The au ho s acknowledge he GSI and LIP, wi h special hanks o
he membe s o he de ec o labo a o y and he mechanical wo kshop
o LIP o hei suppo . This wo k was suppo ed by he Founda ion
o Science and Technology (Po ugal) (CERN/FIS-INS/0006/2021) and
he Eu opean Union’s Ho izon 2020 Resea ch and Inno a ion p o-
g amme unde G an Ag eemen AIDAinno a n.◦ 101004761.
Re e ences
[1] J. Sa ai a, A. Blanco, NIMA 1068 (2024) 169803, h p://dx.doi.o g/10.1016/j.
nima.2024.169803.
[2] D. Bel e , e al., IEEE T ans. Nucl. Sci. 57 (2010) 2848–2856, h p://dx.doi.o g/
10.1109/TNS.2010.2056928.
[3] TRB- amily, b.gsi.de.
[4] A. Neise , e al., JINST 8 (C) (2013) 12043, h p://dx.doi.o g/10.1088/1748-
0221/8/12/C12043.
[5] K. Bo ozdin, e al., Na u e 422 (2003) 277, h p://dx.doi.o g/10.1038/422277a.
[6] J. Sa ai a, e al., NIMA 1050 (2023) 168183, h p://dx.doi.o g/10.1016/j.nima.
2023.168183.
[7] G. Ba is oni, e al., Ann. Nucl. Ene gy 82 (2015) 10–18.
[8] C. Ahdida, e al., F on . Phys. 9 (2022) 788253, h p://dx.doi.o g/10.3389/ phy.
2021.788253.
[9] PDG, Re . Pa . Phys. P og. Theo . Exp. Phys. 083C01 (2020) h p://dx.doi.o g/
10.1093/p ep/p aa104.
[10] V. Vlachoudis, M & C In e na ional Con e ence, ISBN: 9781615673490, 2009,
pp. 790–800.
Nuclea Ins . and Me hods in Physics Resea ch, A 1076 (2025) 170466
6
