Comprehensive analysis of degradation mechanisms in 18650 Li-Ion cells under prolonged cycling conditions

Resea ch Pape s
Comp ehensi e analysis o deg ada ion mechanisms in 18650 Li-Ion cells
unde p olonged cycling condi ions
Pa el Blaˇ
zek
a,c
, Ondˇ
ej Kl aˇ
c
b,e
, Ma in ˇ
Sedina
b
, Ondˇ
ej ˇ
Cech
b
, Ma k´
e a Tkadleco ´
a
a
,
Zuzana S a o ´
a
a
, Tom´
aˇ
s Kazda
b
, Tom´
aˇ
s Zikmund
a,*
, Robe H. Schmi
c,d
, Joze Kaise
a
a
Cen al Eu opean Ins i u e o Technology, B no Uni e si y o Technology, B no, Czech Republic
b
Depa men o Elec ical and Elec onic Technology, Facul y o Elec ical Enginee ing and Communica ion, B no Uni e si y o Technology, Czech Republic
c
WZL | RWTH Aachen Uni e si y, Aachen, Ge many
d
F aunho e Ins i u e o P oduc ion Technology IPT, Ge many
e
The mo Fishe Scien ic, Vlas imila Pecha 12, B no 627 00, Czech Republic
ARTICLE INFO
Keywo ds:
Li hium-ion ba e y deg ada ion
18650 cylind ical cell
Elec ode delamina ion
Aging
ABSTRACT
Li-ion ba e ies a e essen ial o applica ions like elec omobili y and s a iona y ene gy s o age, whe e long- e m
pe o mance and sa e y a e c i ical. While p e ious s udies ha e explo ed deg ada ion mechanisms in Li-ion
cells, de ailed insigh s in o s uc u al changes du ing ex ended cycling emain limi ed. This s udy aims o
in es iga e he deg ada ion mechanisms in a Samsung 18650 cylind ical Li-ion cell o e 800 cycles (10 % o 90 %
S a e o Cha ge) o unde s and pe o mance ade and s uc u al changes. U ilizing pe iodic mic o-CT scans, we
obse ed signican geome ic al e a ions in he elec ode s ack, including delamina ion and bending owa ds
he cell axis, co ela ing wi h capaci y loss. Pos -mo em analyses using b oad ion beam (BIB), scanning elec on
mic oscopy (SEM), ene gy-dispe si e spec oscopy (EDS), and synch o on CT con med oids in he ca hode
ac i e ma e ial, coppe deposi ion on he anode su ace, and ele a ed le els o phospho us and uo ine, likely
om elec oly e decomposi ion and SEI laye o ma ion. These insigh s shed ligh on he s uc u al changes and
ailu e modes in cylind ical Li-ion cells du ing p olonged cycling and highligh he need o design imp o emen s
and op imized manu ac u ing p ocesses o enhance mechanical obus ness and educe de ec s, he eby ex ending
ba e y cycle li e and sa e y.
1. In oduc ion
Li hium-ion ba e ies (LIBs) ha e become a key echnology o en-
e gy s o age due o hei high ene gy densi y, e ciency, and e sa ili y,
p o iding a su cien ly long li e ime. They play a pi o al ole in a ious
echnologies, om small po able elec onic de ices o ene gy s o age
sys ems and elec ic ehicles [1–3]. One o he mos common ypes o
LIB cells is 18650- o ma , which is a cylind ical cell o 18 mm diame e
and 65 mm leng h. I is used as a s anda d choice in nume ous appli-
ca ions like powe ools, elec ic bicycles and elec ic ca s [4].
Ba e y aging mani es s i sel h ough a mul i ude o complex
symp oms, which eec changes in mic o and mac oscopic le el. I
leads o a dec ease in capaci y, an inc ease in in e nal esis ance, and
physical changes such as swelling and delamina ion o elec ode laye s.
I can lead no only o de e io a ion o pe o mance bu also damage
ba e y s uc u e, a ec he sa e y o ba e y usage, and e en culmina e
in comple e ba e y ailu e [5–15]. The issue o ba e y aging is a e y
b oad opic; u he desc ibed a e hose p ocesses ha ha e been he
subjec o esea ch in his pape .
Du ing cycling, he elec odes unde go olume ic changes man-
i es ed a he le el o whole elec odes as well as indi idual g ains o
elec oac i e ma e ial [16–18]. This is mainly due o elec ochemical
p ocesses in ol ing li hium-ion in e cala ion and also la ice s uc u al
changes [19]. Tempe a u e changes induced by high C- a e also play an
essen ial ole [20]. These e ec s in he cylind ical cells lead o he
de o ma ion o he elec odes and hei buckling owa ds he cen e o
he ba e y. This can esul in delamina ion o he elec oac i e ma e ial
and, in ex eme cases, dis up ion o he sepa a o . This phenomenon can
be pa ially coun e ac ed by he use o a cen al pin - a me al ube ha
lls he space in he cen e o he cylind ical cell, which, howe e , in-
c eases he weigh [16].
Volume changes a e also a cause o pa icle c acking [15,21,22]. I
* Co esponding au ho .
E-mail add ess: [email p o ec ed] (T. Zikmund).
Con en s lis s a ailable a ScienceDi ec
Jou nal o Ene gy S o age
jou nal homepage: www.else ie .com/loca e/es
h ps://doi.o g/10.1016/j.es .2025.117436
Recei ed 11 Feb ua y 2025; Recei ed in e ised o m 3 May 2025; Accep ed 12 June 2025
Jou nal o Ene gy S o age 130 (2025) 117436
A ailable online 18 June 2025
2352-152X/© 2025 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY license ( h p://c ea i ecommons.o g/licenses/by/4.0/ ).
can lead o he ma e ial losing con ac wi h he su ounding ma ix,
making i inac i e and dec easing he capaci y. Mo eo e , du ing
c acking, he exis ing SEI laye is dis up ed and s a s o g ow on he
newly o med su aces. This consumes li hium ions, esul ing in u he
capaci y loss. An addi ional isk lies in he possibili y o li hium dend i e
o ma ion mani es ing i sel mainly a highe loads and lowe empe -
a u es [10,23,24].
In he applica ion o e y high cu en s and ope a ion o he cell
ou side he ol age ope a ing window, dissolu ion o he cu en col-
lec o s may also occu . The cases o anodic oxida ion o he coppe
collec o and deposi ion o Cu ions on he ca hode su ace inc easing i s
olume ha e been desc ibed [25]. Also, coppe can o m an in e nal
sho ci cui h ough he o ma ion o Cu dend i es. Ra ely has he
deposi ion o coppe on he su ace o he anode o indi idual g ains
been desc ibed. The educ ion o coppe dissol ed in he elec oly e by
eac ion wi h he SEI laye compounds is assumed [26,27].
E iden ly, he p ocesses con ibu ing o ba e y deg ada ion and
aging a e complex, and hei deepe unde s anding is c ucial o he
u he de elopmen o LIBs. X- ay-based echniques a e sui able o
hei in es iga ion due o hei non-des uc i e cha ac e and can also be
used o in-si u analysis, whe e a scan is pe o med a e a ce ain
numbe o elec ochemical cycles. This includes X- ay adiog aphy [28]
as well as ad anced CT me hods [29,30]. I is possible o achie e eso-
lu ion om mic ome e s when analyzing whole cells down o sub-
mic ome e esolu ion o small pieces o elec ode ma e ials and
de ec e en mild s uc u al changes [31,32].
Wi h scanning elec on mic oscopy (SEM), e en highe esolu ion
can be achie ed o obse e indi idual g ains, hei c acking, and he
g ow h o he SEI laye [11,33–38]. In combina ion wi h elec on
dispe si e spec oscopy (EDS), i is also possible o de e mine he
elemen al composi ion and i s changes a e elec ochemical cycling
[39,40]. These analyses a e usually pe o med only on he su ace o he
elec odes and i i is a c oss-sec ion, again i is a c oss-sec ion h ough
only one side o one elec ode, which canno ully desc ibe he p ocess
aking place in he whole ba e y [41]. A he same ime, elec o-
chemical da a gi e an idea o he change in capaci y and o he p ope ies
o he ba e y, bu do no p o ide accu a e in o ma ion abou he
deg ada ion p ocess, bo h in e ms o de o ma ion o he in e nal cell
s uc u e and changes a he mic oscopic le el in he ba e y elec odes
[42].
Since he cell mus be disassembled be o e analysis, i is ad isable o
discha ge i o sa e y easons, o a oid sho -ci cui ing du ing u he
handling, and o minimize con ac wi h ai . A gon glo eboxes a e
ypically used o his pu pose. The sepa a ed elec odes a e hen
examined om he su ace o c oss-sec ion. The ma e ials a e o en
b i le and can be damaged by mechanical s ess. Fo his eason,
ocused ion beam (FIB) also allowing FIB/SEM 3D econs uc ion
[43–45] o b oad ion beam (BIB) [8,46–49] echniques a e o en used
o sample p epa a ion, which spu e he ma e ial wi hou mechanical
o ce and allow he samples o be obse ed in hei na i e s a e [7,8].
This s udy in oduces a comp ehensi e, mul i-scale, and mul i-modal
app oach o cha ac e izing deg ada ion in 18650- ype LIBs. We combine
mic o-CT, sub-mic on CT, SEM, and EDS wi h elec ochemical pe o -
mance da a ac oss 800 cha ge-discha ge cycles o c ea e a ull s uc u al
and chemical p ole o ba e y aging. No ably, we apply i ual un olling
echniques o mic o-CT da a—enabling quan i a i e acking o elec ode
de o ma ion in cylind ical geome y, a echnique a ely used in ba e y
s udies. CT scans we e conduc ed e e y 200 cycles o obse e geome ic
e olu ion in pa allel wi h elec ochemical me ics, including capaci y,
hys e esis, and impedance ia Elec ochemical Impedance Spec oscopy
(EIS).
By co ela ing la ge-scale s uc u al de o ma ion wi h localized
chemical changes, his wo k p o ides a uniquely in eg a ed iew o LIB
aging. I no only highligh s deg ada ion mechanisms a mul iple leng h
scales bu also demons a es how ad anced CT me hods can se e as
p edic i e ools o ba e y heal h assessmen . Ou me hodology o e s a
eplicable bluep in o u u e s udies aiming o b idge he gap be ween
imaging and pe o mance in eal-wo ld ba e y sys ems.
2. Ma e ials and me hods
2.1. Samples
We in es iga ed comme cially a ailable LIB Samsung INR18650-29E
(manu ac u e s a ed capaci y 2860 mAh) based LiNi
x
Mn
y
Co
1-x-y
O
2
(NMC), he mos used ca hode ma e ials in LIBs. We used wo accu-
mula o s om he same manu ac u ing ba ch. The  s one, esh cell,
was used o des uc i e analysis be o e cycling, and he o he one was
used o long- e m cycling and deg ada ion s udy. A e he cycling was
nished, we used i o des uc i e analysis o he e ec s o aging by
submic on CT, SEM and EDS. Visualiza ion o analyzed cell componen s
and inne s uc u e is shown in.
2.2. Aging and elec ical analysis
The es cell was cha ac e ized by a baseline es using wo cycles in
he ull manu ac u e -dened ol age ange o 2.5–4.2 V a 0.1C/0.1C,
0.2C/0.2C, 0.2C/0.5C and 0.2C/1C (cha ge/discha ge). EIS was pe -
o med a 100 % SoC du ing 0.1/0.1C cycle in a equency ange om 1
MHz o 30 mHz, wi h an ampli ude o 10 mV, ollowed by CT scanning.
The CCCV me hod was used wi h he limi ing cu en se o 0.02C. The
ini ial es was ollowed by long- e m cycling a 1C using he CCCV
me hod wi h a limi cu en o 0.02C in he ange o 10 % o 90 % SoC o
200 cycles, ollowed by a baseline es and subsequen ly by CT scanning.
This p ocess was epea ed un il 800 cycles we e eached. A e he las
baseline es and CT analysis, he cell was disassembled and analyzed by
SEM. The elec ochemical measu emen s we e pe o med wi h a Bio-
Logic ba e y cycle BCS-815 wi h an EIS module.
2.3. Mic o CT analysis
The whole LIB was analyzed by mic o CT in he esh s a e, a e 200,
400, 600, and 800 cycles. Fo hese measu emen s, we used a The mo
Scien ic HeliScan mic oCT sys em equipped wi h a a panel de ec o
wi h a esolu ion o 3072 ×3072 px
2
and pixel size 139 um and 160 kV
mic o ocus ube. The whole ba e y was analyzed by one scan wi h a
space-lling helical ajec o y, which allowed he use o highe
geome ical magnica ion and inc eased he signal- o-noise a io while
a oiding cone-beam a i ac s. The X- ay ube was se o 150 kV, and he
beam was l e ed wi h 0.2 mm o s ainless s eel and 0.5 mm hick Sn
oils o educe beam-ha dening. To ob ain su cien signal, he exposu e
ime was se o 0.65 s, and  e adiog aphs we e a e aged in each o he
5400 p ojec ions pe scan.
The da a we e econs uc ed wi h he i e a i e algo i hm in so wa e
p o ided by he manu ac u e . An algo i hm o sample d i co ec ion
and he sel -calib a ion algo i hm [50] we e used o co ec o
geome ical e o s. The econs uc ed olume size was app oxima ely
22 ×22 ×73 mm
3
wi h (8
μ
m)
3
oxel size.
The i ual un olling echnique was applied o all he whole-cell
da ase s in The mo Scien ic A izo. We used he p ocedu e desc ibed
in [51]. The ca hode was segmen ed using he mul i h esholding
me hod, and mo phological ope a ions such as opening/closing we e
applied. We analyzed he dis ance om he co e on he segmen ed
ca hode and compa ed esul s be ween da ase s. An un olled dis ance
map hen allows o he assessmen o ca hode posi ion and de o mi ies
in he en i e cell olume.
2.4. Sample p epa a ion o SEM, EDS, and submic on CT
Be o e disassembly, ba e ies we e deeply discha ged o 1 V wi h a
cu en o 0.05C. Inside he A -lled glo ebox, he me al case was cu
abou 1 mm below he edge o he posi i e pin using a manual ube
P. Blaˇ
zek e al.
Jou nal o Ene gy S o age 130 (2025) 117436
2
cu e . The CT inspec ion e ealed ha he e is a ee space inside wi h
minimal isk o elec ode damage. The posi i e pin was hen emo ed
using plie s, and he me al case was un olled o app oxima ely hal he
cell’s heigh . The elec oly e was d ied by lea ing he cell unde a ac-
uum in he glo ebox an echambe o abou 1 h.
Pieces o he indi idual elec odes we e cu om he un olled sec ion
using scisso s. Fu he manipula ion ook place ou side he glo ebox.
The cell was cu ac oss on he nega i e pin side wi h a hacksaw abou 1
cm abo e he pin edge. The me al case in his a ea se es as a mechanical
xa ion and p e en s s uc u e shi ing du ing c oss-sec ion p epa a ion
h ough he en i e s uc u e. This specimen was hen mechanically
g ound by 30, 15, 8, and 5
μ
m g ain size SiC g inde pape s wi h iso-
p opanol. The B oad Ion Beam (BIB) polishe model 1061 SEM Mill
(Fischione Ins umen s) was used o u he sample p epa a ion.
The indi idual elec odes we e p epa ed in c oss-sec ion mode, and
he whole s uc u e in plana mode using C yo cooling. SEM imaging
and EDS analysis we e pe o med on a Scios 2 scanning elec on mi-
c oscope (The mo Fische Scien ic). These samples we e ob ained om
he uppe hal o he cell (nea he posi i e pin) app oxima ely in he
middle o he winding (a ea be ween he cen e o he cell and i s edge).
Fo nanoCT, pieces o app oxima ely 0.5 ×2 mm we e cu om
indi idual elec odes wi h a azo blade om simila egion. The sam-
ples we e inse ed in o a Kap on ube and xed wi h mol en e hylene
ca bona e. Fo SEM/CT co ela i e analysis, he sample was  s pol-
ished using BIB, hen SEM analysis was pe o med, and nally placed in
he Kap on ube.
2.5. Submic on CT
We used he unpolished pa o he cell and un olled he elec odes. A
piece o i was cu o dimensions app oxima ely 0.5 ×2 mm. This piece
was hen xed on a special holde . I was inse ed in a Kap on ube and
pou ed wi h e hylene ca bona e o educe oxida ion. The sample was
hen imaged using nanoCT Rigaku nano3DX wi h Mo a ge , 50 kV ube
ol age, 35 s exposu e ime, 800 p ojec ions, and oxel size o 0.54
μ
m.
3. Resul s and discussion
Con inuous measu emen s e ealed capaci y dec ease du ing cell
cycling (Fig. 2 a). We also obse ed ha a e in e up ing he cycling o
analysis wi h CT, egene a ion occu ed, and capaci y was inc eased.
This e ec was mos signican a e 200 and 400 cycles. The mos
conside able capaci y educ ion o 20.8 % occu ed a e he  s 200
cycles. A e ha , he capaci y egene a ed o 90.8 % o he o iginal
alue and, du ing ano he 200 cycles, dec eased by a simila alue as in
he  s 200 cycles. This e ec epea ed also be ween 400 and 600 cy-
cles. A no able capaci y educ ion o 11.3 % occu ed be ween 600 and
800 cycles as well. In o al, capaci y d opped by 30.5 % a e 800 cycles
(see Table 1).
The discha ge cha ac e is ics a di e en C- a es be o e and a e
cycling a e shown in Fig. 2 b. A all C- a es a e 800 cycles o cycling,
he e was a dec ease in he achie ed capaci y and, a he same ime, a
dec ease in he discha ge pla eau. The capaci y a 0.1C load has been
educed om 2865 mAh, which is equi alen o he capaci y decla ed by
he manu ac u e , o 2442 mAh (14.8 % capaci y d op). A 0.2C load,
he capaci y d opped om 2807 o 2344 mAh (16.5 % capaci y d op). A
loads o 0.5C and 1C, he capaci y d opped by 19.2 % and 22.1 %,
espec i ely. I is e iden ha cycling did no only lead o a dec ease in
capaci y bu also o a dec ease in load abili y a highe C- a es.
The change in capaci y du ing he baseline es s is shown in Table 2.
When compa ing he capaci ies o he cha ac e iza ion cycles, a signi -
ican change can be seen a e 200 and 800 cycles, whe e a e 200
cycles, he e was a signican change in capaci y bu s abili y a highe
loads was main ained. Thus, a a load o 1C, he e was a 4.2 % dec ease
be o e cycling and a 4.5 % dec ease a e cycling compa ed o a cu en
o 0.1C. Ano he signican change in his pa ame e occu ed a e 800
cycles.
The hys e esis changes a di e en C- a es du ing cycling, as shown
in Fig. 2 c. I is e iden ha he hys e esis inc eases wi h inc easing C-
a e and is highes a 1C. A e 200 cycles, he e was an inc ease in
hys e esis a all C- a es. I s alue hen emained s able du ing he
ollowing cycling un il he las 200 cycles, when he alue inc eased
again signican ly o all C- a es. This inc ease o hys e esis is in co -
ela ion wi h he signican capaci y d op du ing he las 200 cycles o
cycling and co esponds o he la ge capaci y d op a a cu en load o 1C
ha was obse ed a he end o cycling.
The EIS analysis da a including he equi alen ci cui is shown in
Fig. 1 d. The equi alen ci cui used o in e p e a ion o he measu ed
EIS da a includes se ial esis ance Rs ea u ing ohm in e nal esis ance,
Rc ep esen cha ge ans e esis ance o he cha ge ans e p ocess, Q
is cons an phase elemen ep esen s in e nal capaci ance and Zw is
Wa bu k impedance which ep esen s di usion and mass anspo . This
ci cui was chosen because he cu e o ms one semici cle as in publi-
ca ions o Liu e al. [52] and San oni e al. [53]. The EIS analysis
e ealed ha he Rc a he beginning o cycling was double (12.3 mΩ)
ha a e 200 cycles (5.6 mΩ), which is isible in Nyquis plo s in Fig. 2
d. This is due o he changes in he s uc u e a he beginning o cycling,
which lead o imp o ed con ac wi h he elec oly e. Liu e al. gi e a
simila desc ip ion o his change [54]. The Rc alue was almos
iden ical be ween 200 and 400 cycles (5.5 mΩ), co esponding o he
obse ed capaci y egene a ion and nea ly iden ical change in hys e e-
sis. Subsequen ly, Rc inc eases wi h cycling, and a e 800 cycles, i s
alue app oaches he o iginal alues be o e cycling (9.4 mΩ).
The dV/dQ analysis be o e cycling and a e a di e en numbe o
cycles (200, 400, 600, and 800) is isualized in Fig. 3. The gu e shows
Table 1
Capaci y and capaci y e en ion in di e en s ages o cycling.
Cycle numbe Capaci y/mAh Capaci y e en ion/%
1 2165 –
200 1714 20.8 %
400 1668 23.0 %
600 1717 20.7 %
800 1505 30.5 %
Table 2
Deg ada ion unde di e en loads du ing long- e m cycling.
Capaci y/mAh Capaci y d op/%
Cycle
numbe
0.1C 0.2C 0.5C 1C 0.2C s
0.1C
0.5C s
0.1C
1C s
0.1C
0 2865 2807 2742 2744 98.0 % 95.7 % 95.8 %
200 2660 2591 2591 2541 97.4 % 97.4 % 95.5 %
400 2583 2555 2503 2490 98.9 % 96.9 % 96.4 %
600 2478 2459 2400 2363 99.2 % 96.9 % 95.4 %
800 2442 2344 2215 2139 96.0 % 90.7 % 87.6 %
Fig. 1. A 3D ende o he op pa o an 18650 cell wi h a sec ion showcasing
i s in e nal s uc u e. A de ailed iew highligh s he componen s o he elec-
ode s ack in a omog aphic c oss-sec ion.
P. Blaˇ
zek e al.
Jou nal o Ene gy S o age 130 (2025) 117436
3
he ac i i y o he con ained ma e ials a he anode and a he ca hode
and hei g adual deg ada ion. An anodic peak a 3.42 V and ca hodic a
3.38 V is ela ed o li hia ion o g aphi e anode [55]. A e 200 cycles,
bo h peaks shi ed o a highe ol age, and peaks we e less e iden .
Du ing he cycling, peaks become mo e e iden and, a he same ime,
change posi ion close o 3.5 V. The magni ude o he anode- ela ed peak
a e 200, 400, and 600 cycles was simila ; howe e , a e 800 cycles,
he peak magni ude dec eased, indica ing a dec ease in he ac i i y o
he anode ma e ial. This d op migh be associa ed wi h a highe d op in
capaci y a a high C- a e a e 800 cycles. Ano he e y signican
anodic peak can be obse ed a a po en ial o abou 3.65 V. This peak is
ela ed o he ansi ion om a hexagonal o a monoclinic la ice o he
NMC532 ca hode [56]. I s ac i i y g adually dec eases, wi h he mos
signican change can be obse ed a e he  s 200 cycles and hen
a e he las 200 cycles o cycling. The anodic peak a ound 4.1 V is hen
associa ed wi h high Ni ca hode ma e ials such as NMC811 o NCA [57]
and ep esen s he H2 o H3 phase ansi ion. The small anodic peak
loca ed a ound 3.8 V ep esen s he ansi ion om he M o he H2
phase. The peak a ound 4.1 V was ela i ely s able o he  s 400
cycles, bu i d opped signican ly a e 600 and 800 cycles. This
dec ease may hen be ela ed o he capaci y d op du ing he subsequen
cycling, when he pa ial egene a ion o he ba e y capaci y no longe
occu ed when cycling s a ed again.
P olonged cycling o he cell caused damage o he inne geome y o
he cell. The elec odes a e de o med, and he elec oac i e ma e ial is
c acked and delamina ed om he cu en collec o . These de o ma ions
a e clea ly isible in omog aphic c oss-sec ions (Fig. 4) as well as an
SEM image o he c oss-sec ion h ough he whole s uc u e (Fig. 5). The
de o ma ion is  s isible a e 200 cycles on he  s h ee windings o
he elec ode s ack and is ge ing mo e p onounced wi h cycling. A e
800 cycles, hey each up o he se en h winding o elec odes. SEM
image (Fig. 5) shows, in addi ion o elec ode wa ing, c acks, and loss o
con ac wi h he cu en collec o , also loss o con ac o he anode wi h
he sepa a o and, he e o e, wi h he opposi e ca hode.
The i ual un olling echnique allowed us o isualize he geome y
Fig. 2. a) Capaci y dec eases due o cycling, b) Discha ging p oles a di e en C- a e be o e and a e cycling, c) Changes o hys e esis a di e en c- a es du ing
cycling, d) EIS analysis in di e en s ages o long- e m cycling.
Fig. 3. dQ/dV analysis o INR18650-29E cell be o e cycling and a e 200,
400, 600 and 800 cycles.
P. Blaˇ
zek e al.
Jou nal o Ene gy S o age 130 (2025) 117436
4
o he ca hode ac oss he en i e olume a di e en cycling s ages.
Elec ode de o ma ions appea as e ical s ipes in he dis ance- om-
co e map. S ipes isible in he esh da ase (Fig. 6 a) ha pe sis in
da a om he cycled cell indica e ha he elec ode de ia es om an
ideal spi al. De o mi ies a ising in he inne winding due o cycling a e
also e iden in he un olled da a as addi ional s ipes in he
Fig. 4. Tomog aphic c oss-sec ions, de ail o he de o ma ion’s de elopmen h ough cycling. Top ow: Axial iew, bo om ow: axial iew. Red a ows poin o he
highes de o ma ion p esen in he inne mos winding, which is he same posi ion as show in Fig. 6. (Fo in e p e a ion o he e e ences o colo in his gu e legend,
he eade is e e ed o he web e sion o his a icle.)
Fig. 5. De ail o c oss-sec ion h ough he whole ba e y s uc u e. Elec odes
a e bending owa ds he cen al pin, a ows poin o loss o con ac wi h
sepa a o and cu en collec o .
Fig. 6. Visualiza ion o 10 inne mos windings o he un olled ca hode as an un olled dis ance om he cell co e colo map. Red a ows poin o he highes
de o ma ion p esen in he inne mos winding, which is he same posi ion as shown in Fig. 4. (Fo in e p e a ion o he e e ences o colo in his gu e legend, he
eade is e e ed o he web e sion o his a icle.)
Fig. 7. Leng h o ca hode a e un olling.
P. Blaˇ
zek e al.
Jou nal o Ene gy S o age 130 (2025) 117436
5

co esponding sec ions. These de o mi ies ex end h ough he en i e
heigh o he elec ode spi al and become mo e p onounced wi h an
inc easing numbe o cycles. A limi a ion o his un olled map is he
p esence o a i ac s in he o m o sha p bo de s be ween ho izon al
sec ions. These a i ac s s em om he un olling algo i hm in A izo,
which equi ed di iding he da a in o sec ions o educe compu a ional
demands and in oduced in e pola ion e o s ha likely s uggled o
handle de ia ions om he ideal spi al. Due o hese a i ac s, his da a
canno be used o quan i a i e measu emen s, such as compa ing dis-
ance om he co e be ween da ase s. Howe e , he un olled da a show
a ca hode leng h inc ease o 3.5 mm (0.5 %) o e 800 cycles (see Fig. 7),
likely esul ing om inc eased bending and de o ma ions. The da a
shows ha he mos p ominen change occu ed again a e he  s 200
cycles and a e he las 200 cycles, which co ela es wi h he esul s o
elec ochemical measu emen s. The change o hys e esis occu ed a a
simila ime; a he las 200 cycles, he e was a signican dec ease in
Fig. 8. C oss-sec ion SEM image o esh ba e y a) O e iew o he c oss-sec ion b, c) de ails o he manu ac u ing de ec s.
Table 3
Elemen al composi ion o egions 1–4, as shown in Fig. 9 a).
A omic %
Elemen Region 1
(NCA)
Region 2
(NMC)
Region 3
(NMC)
Region 4
(NCA)
C 2.9 1.7 1.6 5.7
O 59.1 59.1 56.0 55.6
Mn 0.7 11.2 13.0 0.2
Co 3.0 7.6 7.5 3.2
Ni 33.2 20.1 21.4 34.3
Al 0.9 0.3 0.4 0.8
P 0.2 0.0 0.1 0.2
Fig. 9. Resul s o EDS analysis o esh ca hode a) Regions o analysis, he elemen al composi ion o each egion is shown in Table 3. Figu es show he weigh
pe cen age o elemen s om emission line Kb) Mn c) Co d) Ni.
P. Blaˇ
zek e al.
Jou nal o Ene gy S o age 130 (2025) 117436
6
capaci y a high C- a es.
C oss-sec ion images o he esh cells ob ained bo h by CT (Fig. 4)
and SEM (Figs. 5 and 8) show he manu ac u ing de ec s in he s uc u e.
These a e unlled a eas wi hin he elec oac i e laye s uc u e and
bends o he cu en collec o . These a eas co espond o he ‘ oids’ in
he CT model. The cause is p obably due o he inhomogeneous
composi ion o he ca hode pas e, whe e a bubble o ms on one side o
he cu en collec o . Subsequen p essing leads o i s den ing and
damage o he ca hode ma e ial pa icles. De ails o hese a eas a e
isible in SEM images shown in Fig. 8. Analysis o h ee pieces o he
ca hode by synch o on CT (see Supplemen al Ma e ial A.1) shows he
p esence o app oxima ely 9 de ec s / mm2 o he ca hode wi h a
diame e o (72 ±19)
μ
m and olume (26,000 ±19,000)
μ
m
3
. The e is a
high a iance in de ec s shape and size, bu no dependency on he po-
si ion in he ba e y o inuence due o cycling was disco e ed.
EDS elemen al analysis o esh cell ca hode p epa ed indi idually
e ealed ha he elec oac i e ma e ial comp ises wo ypes o ma e ials
wi h di e en elemen al abundances. As can be seen in he EDS analysis
esul s shown in Table 3 and Fig. 9 below one o he ca hode ma e ials is
NMC532 (LiNi
0.5
Mn
0.3
Co
0.2
O
2
) and he o he ca hode ma e ial is mos
likely NCA (LiNi
0.88
Co
0.1
Al
0.02
O
2
). In Fig. 9, he ma e ials can be
dis inguished based on he p esence o manganese in he gi en a ea - i
manganese is p esen , i is NMC; i manganese is absen , i is NCA. This is
consis en wi h he ndings om elec ochemical es ing o he ba e y,
whe e a peak associa ed wi h he NCA ca hode was obse ed a dQ/dV
spec a.
Piece o ca hode a e 800 cycles was analyzed wi h mic oCT and
SEM. Same egion a he su ace o he ca hode piece is isualized om
bo h echniques in Fig. 10. The analysis o he cycled ca hode c oss-
sec ion by EDS showed a 3 % dec ease in Ni, a 2 % dec ease in Mn, a
1 % dec ease in Co o NMC g ains, and a 4 % dec ease in Ni in NCA
g ains (Table 4) compa ed wi h EDS esul s o p is ine ca hode (Table 3).
On he o he hand, he oxygen ac ion inc eased by app oxima ely he
same amoun . Thus, i is p obably no a dissolu ion o ansi ion me al
oxides. I can be con med by he ac ha hey we e no p esen on he
anode. The pa icles appea o be mo e c acked and agmen ed a e
Fig. 10. Analysis o a piece o ca hode a e 800 cycles. a) c oss-sec ion om submic on CT, b) 3D ende o he CT, c) Image o ca hode sec ion on om SEM, d)
De ail o SEM image (posi ion shown by g een a ows) wi h elemen al composi ion om EDS. (Fo in e p e a ion o he e e ences o colo in his gu e legend, he
eade is e e ed o he web e sion o his a icle.)
Table 4
Elemen al composi ion o egions (shown in Fig. 10 d).
A omic %
Elemen Region 1
(NCA)
Region 2
(NCA)
Region 3
(NMC)
Region 4
(impu i y)
C 2.6 2.6 2.2 8.6
O 63.2 62.9 64.2 64.7
Al 0.6 0.5 0.3 19.7
Mn 0.6 0.6 10.1 1.3
Co 3.1 3.2 6.6 0.9
Ni 30 30.2 16.6 2.9
S 0 0 0 1.9
Fig. 11. SEM images o ca hode c oss-sec ion; esh cell (a); cycled cell (b).
P. Blaˇ
zek e al.
Jou nal o Ene gy S o age 130 (2025) 117436
7
cycling, especially he NMC pa icles, compa ed wi h esh cell (Fig. 11).
The close analysis showed he local p esence o ba ium sul a e a some
g ain edges o bo h, p is ine and cycled cell. Also, pieces o aluminum
a e p esen be ween g ains (impu i y in egion 4 Fig. 9d). These a eas
ha e also inc eased amoun s o sul u and each abou 2 %.
In he anode case, a ypical laye ed s uc u e o g aphi e g ains was
obse ed. The analysis also showed mino amoun s o uo ine and
phospho us p esen a he edges o he g ains and in he space be ween
hem. The o igin may be a binde , li hium sal , o a o med SEI laye . The
compa ison o he anode o he esh cell and he cycled cell (Fig. 12)
shows ha he gaps be ween he g aphene laye s a e mo e expanded in
he cycled cell. A he same ime, he e a e pieces o coppe be ween he
g ains. Coppe is also ound in a la ge amoun on he su ace o indi-
idual g ains as pa o he SEI laye (Fig. 13).
Highe concen a ions o coppe we e addi ionally de ec ed in a
c oss-sec ion h ough he en i e cell s uc u e on he su ace o he anode
as a ligh s ipe (Fig. 14). This was con med by EDS analysis showing a
dec easing coppe concen a ion om he su ace owa ds he cen e o
he anode (Fig. 15). The su ace o he anode o a esh cell analyzed in
he same way does no show an inc eased coppe concen a ion.
4. Conclusions
This s udy p esen s a de ailed analysis o deg ada ion in Samsung
18650 cylind ical cells du ing long- e m cycling. While o e all capaci y
ade aligned wi h da ashee expec a ions and EIS esul s, CT imaging
e ealed signican in e nal mechanical deg ada ion—specically
delamina ion and inwa d bending o elec ode laye s despi e he cen al
pin. No ably, de o ma ion ex ended o he eigh inne mos windings
a e 800 cycles, highligh ing he need o imp o ed mechanical design,
such as a hicke cen al pin.
A no el nding includes he iden ica ion o oids in he ca hode
s uc u e, wi h BIB/SEM/EDS analysis linking hese egions o inc eased
NMC g ain c acking. These insigh s sugges op imiza ion o he slu y
mix u e and coa ing p ocess could mi iga e such de ec s.
The p ima y changes due o aging include he expansion o g aphi e
g ains and ac u ing o NMC g ains a ound de ec s. In he cycled ma-
e ial, ele a ed P and F le els ypical o SEI o ma ion and unexpec edly
high Cu concen a ions (up o 20 %) on he anode and g aphi e g ains.
Since sample p epa a ion was uled ou , cycling- ela ed Cu deposi ion is
likely—a phenomenon no ully explained and wa an ing u he s udy.
Li e a u e [26,27] sugges s his could in ol e in e ac ions be ween Cu
ions and SEI laye componen s. Di e en elec ode composi ion and
op imized cycling p o ocols may p e en hese issues. Addi ionally,
elec ode de o ma ion om cu en collec o expansion led o c acking
and delamina ion, inc easing hys e esis and educing capaci y due o
pa icle de achmen . These changes may comp omise pe o mance a
high C- a es. Ou ndings suppo p edic i e modeling o ba e y
heal h and unde sco e how design and cycling p o ocol op imiza ion can
mi iga e ailu e modes and ex end ba e y li e.
Fig. 12. SEM images o anode c oss-sec ion; esh cell (a); cycled cell (b).
Fig. 13. EDS mapping o cycled anode c oss-sec ion.
Fig. 14. SEM images o whole s uc u e c oss-sec ion, s udy o inc eased amoun s o coppe on he anode su ace; esh cell (a); cycled cell (b).
P. Blaˇ
zek e al.
Jou nal o Ene gy S o age 130 (2025) 117436
8
CRediT au ho ship con ibu ion s a emen
Pa el Blaˇ
zek: W i ing – e iew & edi ing, W i ing – o iginal d a ,
Me hodology, In es iga ion, Concep ualiza ion. Ondˇ
ej Kl aˇ
c: W i ing
– e iew & edi ing, W i ing – o iginal d a , Me hodology, In es iga ion,
Concep ualiza ion. Ma in ˇ
Sedina: Me hodology, In es iga ion. Ondˇ
ej
ˇ
Cech: Supe ision, Me hodology, In es iga ion. Ma k´
e a Tkadleco ´
a:
Me hodology, In es iga ion. Zuzana S a o ´
a: In es iga ion, Fo mal
analysis. Tom´
aˇ
s Kazda: W i ing – e iew & edi ing, Valida ion, Su-
pe ision, Da a cu a ion, Concep ualiza ion. Tom´
aˇ
s Zikmund: Valida-
ion, Supe ision, Resou ces, P ojec adminis a ion, Concep ualiza ion.
Robe H. Schmi : Supe ision, Resou ces. Joze Kaise : Supe ision,
Resou ces, P ojec adminis a ion, Funding acquisi ion.
Decla a ion o compe ing in e es
The au ho s decla e he ollowing nancial in e es s/pe sonal e-
la ionships which may be conside ed as po en ial compe ing in e es s:
Joze Kaise epo s a ela ionship wi h Cac uX s. .o. ha includes:
boa d membe ship and equi y o s ocks. Tomas Zikmund epo s a
ela ionship wi h Cac uX s. .o. ha includes: boa d membe ship and
equi y o s ocks. Zuzana S ubiano a epo s a ela ionship wi h Cac uX
s. .o ha includes: employmen . Ma ke a Tkadleco a epo s a ela-
ionship wi h Cac uX s. .o ha includes: employmen . Pa el Blazek e-
po s a ela ionship wi h Bake Hughes Digi al Solu ions GmbH ha
includes: employmen . I he e a e o he au ho s, hey decla e ha hey
ha e no known compe ing nancial in e es s o pe sonal ela ionships
ha could ha e appea ed o inuence he wo k epo ed in his pape .
Acknowledgmen s
We acknowledge CzechNanoLab Resea ch In as uc u e suppo ed
by MEYS CR (LM2023051). We acknowledge Ele a Sinc o one T ies e
o p o iding access o i s synch o on adia ion acili ies and o
nancial suppo unde he IUS in e nal p ojec , and we hank Lucia
Mancini o assis ance in using beamline SYRMEP. This wo k was sup-
po ed by he p ojec The Ene gy Con e sion and S o age unded as
p ojec No 02.01.01/00/22_008/0004617 by P og amme Johannes
Amos Comenius call Excellen Resea ch. This wo k was de eloped in
coope a ion wi h The mo Fishe Scien ic B no.
Appendix A. Supplemen a y da a
Supplemen a y da a o his a icle can be ound online a h ps://doi.
o g/10.1016/j.es .2025.117436.
Da a a ailabili y
Da ase is a ailable a DOI 10.5281/zenodo.14833100
Re e ences
[1] B. Sc osa i, J. Ga che, Li hium ba e ies: s a us, p ospec s and u u e, J. Powe
Sou ces 195 (2010) 2419–2430, h ps://doi.o g/10.1016/j.jpowsou .2009.11.048.
[2] R. Schmuch, R. Wagne , G. H¨
o pel, T. Placke, M. Win e , Pe o mance and cos o
ma e ials o li hium-based echa geable au omo i e ba e ies, Na . Ene gy 3
(2018) 267–278, h ps://doi.o g/10.1038/s41560-018-0107-2.
[3] D. Deng, Li-ion ba e ies: basics, p og ess, and challenges, Ene gy Sci. Eng. 3
(2015) 385–418.
[4] V. Muenzel, A.F. Hollenkamp, A.I. Bha , J. de Hoog, M. B azil, D.A. Thomas,
I. Ma eels, A compa a i e es ing s udy o comme cial 18650- o ma li hium-ion
ba e y cells, J. Elec ochem. Soc. 162 (2015) A1592, h ps://doi.o g/10.1149/
2.0721508jes.
[5] M. Pin, J. Choi, J.H. Chang, A.S. Schenk, J. Han, S. Wacławek, Y. Kim, J.Y. Cheong,
In si u X- ay based analysis o anode ma e ials o li hium-ion ba e ies: Cu en
s a us and u u e implica ions, Ene gy S o age Ma e . 73 (2024) 103798, h ps://
doi.o g/10.1016/j.ensm.2024.103798.
[6] L. Willenbe g, P. Dechen , G. Fuchs, M. Teube , M. Ecke , M. G a , N. Kü en, D.
U. Saue , E. Figgemeie , The de elopmen o jelly oll de o ma ion in 18650
li hium-ion ba e ies a low s a e o cha ge, J. Elec ochem. Soc. 167 (2020)
120502, h ps://doi.o g/10.1149/1945-7111/aba96d.
[7] T. Waldmann, A. I u ondobei ia, M. Kaspe , N. Ghanba i, F. Aguesse, E. Bekae ,
L. Daniel, S. Genies, I.J. Go don, M.W. L¨
oble, E.D. Vi o, M. Wohl ah -Meh ens,
Re iew—pos -mo em analysis o aged li hium-ion ba e ies: disassembly
me hodology and physico-chemical analysis echniques, J. Elec ochem. Soc. 163
(2016) A2149, h ps://doi.o g/10.1149/2.1211609jes.
[8] Ko ache , Sch ¨
o ne , Gs ein, Aiello, Hanzu, Wilkening, Foi zik, Wellm, Sinz,
Elle sdo e , Analy ical dissec ion o an au omo i e Li-ion pouch cell, Ba e ies 5
(2019) 67, h ps://doi.o g/10.3390/ba e ies5040067.
[9] S.J. An, J. Li, C. Daniel, D. Mohan y, S. Nagpu e, D.L. Wood, The s a e o
unde s anding o he li hium-ion-ba e y g aphi e solid elec oly e in e phase (SEI)
and i s ela ionship o o ma ion cycling, Ca bon 105 (2016) 52–76, h ps://doi.
o g/10.1016/j.ca bon.2016.04.008.
[10] T. Waldmann, M. Wilka, M. Kaspe , M. Fleischhamme , M. Wohl ah -Meh ens,
Tempe a u e dependen ageing mechanisms in Li hium-ion ba e ies – a pos -
mo em s udy, J. Powe Sou ces 262 (2014) 129–135, h ps://doi.o g/10.1016/j.
jpowsou .2014.03.112.
[11] J. Cheng, X. Li, Z. Wang, H. Guo, Mechanism o capaci y ading o 18650
cylind ical li hium ion ba e ies, T ans. Non e ous Me . Soc. China 27 (2017)
1602–1607, h ps://doi.o g/10.1016/S1003-6326(17)60182-1.
[12] M. Ecke , N. Nie o, S. K¨
abi z, J. Schmals ieg, H. Blanke, A. Wa necke, D.U. Saue ,
Calenda and cycle li e s udy o Li(NiMnCo)O2-based 18650 li hium-ion ba e ies,
J. Powe Sou ces 248 (2014) 839–851, h ps://doi.o g/10.1016/j.
jpowsou .2013.09.143.
[13] K. Dai, Z. Wang, G. Ai, H. Zhao, W. Yuan, X. Song, V. Ba aglia, C. Sun, K. Wu,
G. Liu, The ans o ma ion o g aphi e elec ode ma e ials in li hium-ion ba e ies
a e cycling, J. Powe Sou ces 298 (2015) 349–354, h ps://doi.o g/10.1016/j.
jpowsou .2015.08.055.
[14] G. Kwak, J. Pa k, J. Lee, S. Kim, I. Jung, E ec s o anode ac i e ma e ials o he
s o age-capaci y ading on comme cial li hium-ion ba e ies, J. Powe Sou ces 174
(2007) 484–492, h ps://doi.o g/10.1016/j.jpowsou .2007.06.169.
[15] H.-J. Kim, T. K ishna, K. Zeb, V. Rajangam, C.V.V.M. Gopi, S. Sambasi am, K.V.
G. Ragha end a, I.M. Obaida , A comp ehensi e e iew o Li-ion ba e y ma e ials
and hei ecycling echniques, Elec onics 9 (2020) 1161, h ps://doi.o g/
10.3390/elec onics9071161.
[16] T. Waldmann, S. Go se, T. Sam leben, G. Schneide , V. Knoblauch, M. Wohl ah -
Meh ens, A mechanical aging mechanism in li hium-ion ba e ies, J. Elec ochem.
Soc. 161 (2014) A1742–A1747, h ps://doi.o g/10.1149/2.1001410jes.
[17] J.H. Lee, H.M. Lee, S. Ahn, Ba e y dimensional changes occu ing du ing cha ge/
discha ge cycles— hin ec angula li hium ion and polyme cells, J. Powe Sou ces
119–121 (2003) 833–837, h ps://doi.o g/10.1016/S0378-7753(03)00281-7.
[18] M. Ko, S. Chae, J. Cho, Challenges in accommoda ing olume change o Si anodes
o Li-ion ba e ies, ChemElec oChem 2 (2015) 1645–1651, h ps://doi.o g/
10.1002/celc.201500254.
[19] N. Zhang, H. Tang, Dissec ing anode swelling in comme cial li hium-ion ba e ies,
J. Powe Sou ces 218 (2012) 52–55, h ps://doi.o g/10.1016/j.
jpowsou .2012.06.071.
[20] S. Yun, C. Kim, Inuences o he mal olume changes in LiMn2O4 pa icles on he
pe o mance o a Li-ion ba e y, J. Mech. Sci. Technol. 32 (2018) 3233–3239,
h ps://doi.o g/10.1007/s12206-018-0625-7.
[21] C. Lin, A. Tang, H. Mu, W. Wang, C. Wang, Aging mechanisms o elec ode
ma e ials in li hium-ion ba e ies o elec ic ehicles, J. Chem. 2015 (2015) 1–11,
h ps://doi.o g/10.1155/2015/104673.
[22] A.M. Boyce, E. Ma ínez-Pa˜
neda, A. Wade, Y.S. Zhang, J.J. Bailey, T.M.M. Heenan,
D.J.L. B e , P.R. Shea ing, C acking p edic ions o li hium-ion ba e y elec odes
by X- ay compu ed omog aphy and modelling, J. Powe Sou ces 526 (2022)
231119, h ps://doi.o g/10.1016/j.jpowsou .2022.231119.
[23] M. B oussely, Ph. Biensan, F. Bonhomme, Ph. Blancha d, S. He ey e, K. Neche , R.
J. S aniewicz, Main aging mechanisms in Li ion ba e ies, J. Powe Sou ces 146
(2005) 90–96, h ps://doi.o g/10.1016/j.jpowsou .2005.03.172.
Fig. 15. The concen a ion o Cu on line p oles ma ked in Fig. 14 in he di-
ec ion owa ds coppe cu en collec o . De ailed esul s o eleme nal analysis
a e in Supplemen al ma e ial A.2.
P. Blaˇ
zek e al.
Jou nal o Ene gy S o age 130 (2025) 117436
9