scieee Science in your language
[en] (orig)

Distinguishing the Rhombohedral Phase from Orthorhombic Phases in Epitaxial Doped HfO2 Ferroelectric Films

Abstract

Epitaxial strain plays an important role in the stabilization of ferroelectricity in doped hafnia thin films, which are emerging candidates for Si-compatible nanoscale devices. Here, we report on epitaxial ferroelectric thin films of doped HfO2 deposited on La0.7Sr0.3MnO3-buffered SrTiO3 substrates, La0.7Sr0.3MnO3 SrTiO3-buffered Si (100) wafers, and trigonal Al2O3 substrates. The investigated films appear to consist of four domains in a rhombohedral phase for films deposited on La0.7Sr0.3MnO3-buffered SrTiO3 substrates and two domains for those deposited on sapphire. These findings are supported by extensive transmission electron microscopy characterization of the investigated films. The doped hafnia films show ferroelectric behavior with a remanent polarization up to 25 μC/cm2 and they do not require wake-up cycling to reach the polarization, unlike the reported polycrystalline orthorhombic ferroelectric hafnia films.

Read accessible full text

Distinguishing the Rhombohedral Phase from Orthorhombic Phases in Epitaxial Doped HfO2 Ferroelectric Films

Author: Petraru, Adrian,Gronenberg, Ole Frederick,Schürmann, Ulrich,Kienle, Lorenz,Droopad, Ravi,Kohlstedt, Hermann
Year: 2024
DOI: 10.1021/acsami.4c10423
Source: https://macau.uni-kiel.de/servlets/MCRFileNodeServlet/macau_derivate_00006590/petraru-et-al-2024-distinguishing-the-rhombohedral-phase-from-orthorhombic-phases-in-epitaxial-doped-hfo2-ferroelectric.pdf
Dis inguishing he Rhombohed al Phase om O ho hombic Phases
in Epi axial Doped H O2Fe oelec ic Films
Ad ian Pe a u,*Ole G onenbe g, Ul ich Schu mann, Lo enz Kienle, Ra i D oopad,
and He mann Kohls ed
Ci e This: ACS Appl. Ma e . In e aces 2024, 16, 42534−42545
Read Online
ACCESS Me ics & Mo e A icle Recommenda ions *
sı Suppo ing In o ma ion
ABSTRACT: Epi axial s ain plays an impo an ole in he
s abiliza ion o e oelec ici y in doped ha nia hin ilms, which
a e eme ging candida es o Si-compa ible nanoscale de ices. He e,
we epo on epi axial e oelec ic hin ilms o doped H O2
deposi ed on La0.7S 0.3MnO3-bu e ed S TiO3subs a es,
La0.7S 0.3MnO3S TiO3-bu e ed Si (100) wa e s, and igonal
Al2O3subs a es. The in es iga ed ilms appea o consis o ou
domains in a hombohed al phase o ilms deposi ed on
La0.7S 0.3MnO3-bu e ed S TiO3subs a es and wo domains o
hose deposi ed on sapphi e. These indings a e suppo ed by
ex ensi e ansmission elec on mic oscopy cha ac e iza ion o he
in es iga ed ilms. The doped ha nia ilms show e oelec ic
beha io wi h a emanen pola iza ion up o 25 μC/cm2and hey
do no equi e wake-up cycling o each he pola iza ion, unlike he epo ed polyc ys alline o ho hombic e oelec ic ha nia ilms.
KEYWORDS: e oelec ic H O2, ul a hin ilm epi axy, ha nia, epi axial H 0.5Z 0.5O2 hin ilms, Y-doped e oelec ic H O2.
1. INTRODUCTION
Fe oelec ic ma e ials a e o inc easing echnological
impo ance ha ing con inuously ex ending applica ions includ-
ing e oelec ic andom-access memo ies
1,2
o s o age,
senso s and ac ua o s,
3,4
op ics,
5,6
and elec onic de ices o
neu o-inspi ed elec onics such as mem is o s, necessa y o
building up elec onic synapses o neu omo phic compu -
ing.
7−9
Recen ly disco e ed e oelec ici y in doped ha nia
ul a hin ilms opens new pe spec i es o he ealiza ion o
special de ices like e oelec ic ield-e ec ansis o s and
e oelec ic unnel junc ions
10,11
due o hei compa ibili y
wi h he Si echnology.
Single-c ys al Y-doped H O2 e oelec ic ilms deposi ed on
y ium-s abilized zi conia ha e al eady been epo ed,
exhibi ing he o ho hombic pola phase wi h he space
g oup Pca21(o-phase) and a pola iza ion o 16 μC/cm2,
gene ally conside ed esponsible o e oelec ici y in hin
doped ha nia ilms.
12
On he o he hand, epi axial H 0.5Z 0.5O2
ilms deposi ed on La0.7S 0.3MnO3/S TiO3(subs a e) by
pulsed lase deposi ion (PLD) ha e ecen ly been e-
po ed.
13,14
These ilms do no c ys allize in he commonly
epo ed o-phase, bu in a pola hombohed al phase ( -phase),
s abilized ia epi axial s ain, and was iden i ied wi h a la ge P
o 34 μC/cm2. Mo eo e , hese hombohed al ilms do no
equi e “wake-up” cycling o es ablishing e oelec ic swi ch-
ing unlike mos o he epo ed polyc ys alline ilms wi h an
o ho hombic pola phase. Recen ly, polyc ys alline H -
(Z )1+xO2c ys allizing in a hombohed al e oelec ic phase
wi h a low coe ci e ield ha e been epo ed.
15
Meanwhile,
epi axial doped ha nia ilms we e epo ed by some au ho s,
being deposi ed on se e al single c ys al subs a es like y ia-
s abilized zi conia (111)YSZ, (100)YSZ,
16
LSMO-bu e ed
LaAlO3,
14,17,18
S TiO3,
13,14,18,19
NdScO3, NdGaO3, MgO,
18
YAlO3, (LaAlO3)0.3−(S 2AlTaO6)0.7 (LSAT), DyScO3,
14,18
S TiO3-bu e ed Si,
20
and GaN-bu e ed Si.
21
An in e es ing
s udy on epi axial doped ha nia ilms deposi ed on a ious
La0.7S 0.3MnO3(LSMO)-bu e ed subs a es e eals ha
TbScO3and GdScO3a e e y good candida es o epi axial
s ess s abiliza ion o he e oelec ic phase.
18
In mos o hese
s udies, he s ess-s abilized me as able pola o ho hombic
phase (Pca21) was epo ed o be esponsible o he
e oelec ici y in doped ha nia ilms,
12,17,19,20,22−26
whe eas
he hombohed al (space g oups R3mo R3) phases a e less
epo ed. The eason migh be he ac ha he o ho hombic
phase is ha d o dis inguish om he hombohed al phase. In
his a icle, we ocus on he challenge o dis inguishing
be ween hese phases.
Recei ed: June 24, 2024
Re ised: July 23, 2024
Accep ed: July 28, 2024
Published: Augus 5, 2024
Resea ch A iclewww.acsami.o g
© 2024 The Au ho s. Published by
Ame ican Chemical Socie y 42534
h ps://doi.o g/10.1021/acsami.4c10423
ACS Appl. Ma e . In e aces 2024, 16, 42534−42545
This a icle is licensed unde CC-BY 4.0
Downloaded ia CHRISTIAN-ALBRECHTS-UNIV KIEL on Sep embe 9, 2024 a 12:30:29 (UTC).
See h ps://pubs.acs.o g/sha ingguidelines o op ions on how o legi ima ely sha e published a icles.
Apa om he epi axial ha nia ilms deposi ed on sapphi e,
ITO-bu e ed YSZ, and GaN, only epi axial ilms deposi ed on
LSMO-bu e ed subs a es ha e been epo ed o be e o-
elec ic.
A s udy on o he bu e laye s like LaNiO3, La0.5Ca0.5MnO3,
S RuO3, and Ba0.95La0.05SnO3showed a low o no s abilized
e oelec ic phase, excep o hose wi h mangani e elec o-
des.
27
In addi ion o he epi axial s ain imposed by he subs a e,
ac o s such as deposi ion empe a u e, pa ial gas p essu e,
and ilm composi ion also play an impo an ole in
de e mining he o ma ion o c ys al phases. Fo example,
Kaise e al. epo ed ha ei he a monoclinic o a
hombohed al phase is s able in H O2 hin ilms g own by
molecula beam epi axy on c-cu sapphi e depending on he
oxygen pa ial p essu e. DFT simula ions ha e shown ha he
hombohed al phase is s abilized by oxygen acancies, no due
o he epi axial s ain.
28
Fu he mo e, i was discussed ha a
zi conium subs i u ion may ha e he same e ec .
Supe la ice and laye s uc u es p o ide a u he imp o e-
men o he e oelec ic pe o mance o ha nia-based ilms.
Thus, “wake-up” ee e oelec ic capaci o s based on H O2/
Z O2supe la ices
29
o capaci o s wi h imp o ed e oelec ic
and dielec ic pe o mances ha e been epo ed.
30,31
In his wo k, he ocus will be on e oelec ic epi axial
doped ha nia ilms deposi ed on h ee di e en subs a es o
compa e hei s uc u al p ope ies and o de e mine hei
c ys al s uc u e. Thus, Z /Y:H O2/La0.7S 0.3MnO3/S TiO3
(100), Z /Y:H O2/La0.7S 0.3MnO3/S TiO3/Si (100), and Z /
Y:H O2/Al2O3(0001) sys ems a e in es iga ed he e. The
s uc u al cha ac e iza ion o each o hese h ee sys ems
(comp ising XRD and TEM esul s) is p esen ed in a sepa a e
subsec ion, and he elec ical cha ac e iza ion is shown in he
las sec ion o he esul s.
2. EXPERIMENTAL METHODS
2.1. Ma e ials. H 0.5Z 0.5O2and H 0.93Y0.07O2 hin ilms o
hicknesses a ying om 2.5 o 12 nm we e deposi ed by pulsed
lase deposi ion (PLD) on La0.7S 0.3MnO3-bu e ed S TiO3(001)
(STO) subs a es, La0.7S 0.3MnO3-bu e ed Nb (0.5%)-doped S TiO3
(001), S TiO3-bu e ed Si subs a es, and sapphi e (0001) subs a es.
The doped H O2and he LSMO ilms we e deposi ed in a single
p ocess wi hou b eaking he acuum. The H 0.5Z 0.5O2-and
H 0.93Y0.07O2-sin e ed ce amic a ge s we e pu chased om EVO-
CHEM whe eas he LSMO a ge was pu chased om P axAi . A K F
excime lase o 248 nm in wa eleng h was used o abla ion in a
comme cial PLD sys em om Su ace GmbH. The La0.7S 0.3MnO3
ilms we e deposi ed a a subs a e empe a u e o 780 °C, unde 0.15
mba O2, and a a lase luence and equency o 1.4 J/cm2, and 2 Hz,
espec i ely. The doped H O2 ilms we e deposi ed a a subs a e
empe a u e o 800 °C, unde 1.5 ×10−2mba O2, and a a lase
luence o 1.4 J/cm2and a equency o 5 Hz. A e deposi ion, he
ilms we e cooled a 5 C/min o oom empe a u e unde 3 mba o
oxygen p essu e. Epi axial 20 nm S TiO3bu e laye s we e deposi ed
on Si(100) subs a es by molecula beam epi axy (MBE) a Texas
S a e Uni e si y. The epi axial oxide g ow h on silicon was achie ed
using a codeposi ion p ocess in which bo h he alkaline ea h me al
and he Ti shu e s we e opened in a con olled oxygen en i onmen .
Since, unde hese g ow h condi ions, he s icking coe icien o he
indi idual elemen s is uni y, ca e ul calib a ion o he luxes was
pe o med o s oichiome ic oxide ilms. The g ow h a e used was
app oxima ely 2A/min and was calib a ed using RHEED in ensi y
oscilla ions.
2.2. XRD Cha ac e iza ion. The XRD measu emen s including
wide- ange ecip ocal space maps (RSMs) and pole igu es we e
acqui ed wi h a Sma Lab di ac ome e (Rigaku) equipped wi h a 9
kW Cu anode X- ay ube and a 2D HyPix-3000 X- ay de ec o . A wo-
bounce monoch oma o was used o high- esolu ion XRD scans.
2.3. TEM Sample P epa a ion and Analysis. C oss-sec ional
TEM samples we e p epa ed om he S TiO3and Al2O3subs a es
using ocused ion beam (FIB) milling in an FEI Helios Nanolab
sys em. Be o e Ga-ion e ching, a p o ec i e P coa ing was deposi ed
wi h a gas injec ion sou ce inside he FIB sys em. A plan- iew sample
on a Si subs a e wi h epi axial S TiO3and LSMO laye s was p epa ed
wi h a P ecision Ion Polishing Sys em (PIPS, Model 691 om Ga an
Inc.).
The FIB samples we e analyzed using a JEOL JEM-2100 wi h
elec ons accele a ed o 200 kV ex ac ed om a LaB6 ca hode, while
he PIPS sample was analyzed wi h an FEI Tecnai F30 G2a 300 kV
and a ield emission gun.
2.4. De ice Fab ica ion. Fe oelec ic capaci o s we e ab ica ed
by deposi ing Cu op elec odes by he mal e apo a ion h ough a
s encil mask. The de ice size a ies om 25 o 225 μm2. O he
e oelec ic capaci o s on Si subs a es we e pa e ned by using UV
op ical li hog aphy and A +ion beam e ching.
2.5. Elec ical Cha ac e iza ion. The e oelec ic hys e esis
loops (P−Vloops) o he e oelec ic capaci o s we e measu ed by
using a Radian Technologies P emie II e oelec ic es e . The es
signal used he e consis s o a linea amp wa e o m wi h a pe iod in
he ange o 0.1 o 150 ms. Addi ional e oelec ic hys e esis loops
we e measu ed wi h an aixACCT TF Analyze 3000.
3. RESULTS AND DISCUSSION
3.1. S uc u al Cha ac e iza ion. 3.1.1. H 0.5Z 0.5O2
(HZO) Films Deposi ed on LSMO-Bu e ed (001)-O ien ed
Nb:STO Subs a es. 3.1.1.1. X- ay Di ac ion Analysis. An X-
ay di ac ion (XRD) scan o a H 0.5Z 0.5O2(HZO) ilm on
LSMO-bu e ed (001)-o ien ed Nb:STO subs a es is depic ed
in Figu e 1a. The specula e lec ions 001, 002, and 003 o he
Nb:STO subs a e a e he mos in ense ones, ollowed by he
specula e lec ions o he epi axial LSMO ilm. The 003 and
006 e lec ions o he HZO ilm a e also p esen . Figu e 1b
shows he magni ied 003 HZO egion, whe e he hickness
oscilla ions a e clea ly isible, demons a ing good c ys alline
Figu e 1. (a) The XRD pa e n o he H 0.5Z 0.5O2/La0.7S 0.3MnO3/
S TiO3(subs a e) he e os uc u e; (b) he de ailed H 0.5Z 0.5O2003
egion wi h hickness oscilla ions; he blue line ep esen s he
expe imen al da a, and he ed line is he simula ion o he
di ac og am showing a ilm hickness o 8 nm. (c) Simula ion o
he La0.7S 0.3MnO3001 egion ( ed line), he ilm hickness is ound
o be abou 71 uni cells, i.e., 27 nm. (d) Rocking cu e o he
H 0.5Z 0.5O2003 e lec ion; FWHM = 0.032°.
ACS Applied Ma e ials & In e aces www.acsami.o g Resea ch A icle
h ps://doi.o g/10.1021/acsami.4c10423
ACS Appl. Ma e . In e aces 2024, 16, 42534−42545
42535
quali y and well-de ined in e aces. An a e age HZO ilm
hickness o 8 nm is es ima ed om he i o he 003 e lec ion
egion, using he o mula o he heo e ical di ac og am,
32
adap ed o in e plana spacing:
I n( ) sin( d sin / )/sin( d sin )/ 2
[ ]
, whe e dis he
in e plana spacing o he XRD peak in ol ed ( he planes a e
Figu e 2. (a) Wide ange RSM o he HZO/LSMO/STO (subs a e) he e os uc u e. The spo s belonging o LSMO/STO a e deno ed in da k
blue, whe eas he spo s om he HZO ilm a e deno ed in ed. The angles ma ked nex o he HZO spo s ep esen he in-plane o a ion o he
co esponding HZO domains; he hkl Mille indices a e assigned he e o he R3mphase. (b) Pole igu e o he HZO {201} and HZO {022} spo s;
he black, ed, blue, and g een a ows indica e he ou domains wi h he amilies o spo s associa ed wi h hem. Pole igu e simula ion o he HZO
ilms: (c) phase R3m, conside ing (001) ou -o -plane o ien a ion and p esence o ou domains, wi h an in-plane o a ion o 90°wi h espec o one
ano he ; (d) phase Pca21conside ing (111) ou -o -plane o ien a ion and p esence o ou domains, wi h an in-plane o a ion o 90°wi h espec o
one ano he , and he p esence o {200}, {020}, and {002} spo s.
Figu e 3. HZO/LSMO/STO (subs a e) he e os uc u e. (a) 2θscans o he 12 poles a chi = 71.1°and o he specula spo (chi = 0°in black). 2θ
scans o he 12 poles a (b) chi = 55.7°and (c) a chi = 35°.
ACS Applied Ma e ials & In e aces www.acsami.o g Resea ch A icle
h ps://doi.o g/10.1021/acsami.4c10423
ACS Appl. Ma e . In e aces 2024, 16, 42534−42545
42536
pa allel o he sample su ace), θis i s co esponding
di ac ion angle, λis he X- ay wa eleng h, and nis he ilm
hickness in he numbe o in e plana spacing. The heo e ical
di ac og am is plo ed wi h a ed line as shown in Figu e 1b.
A simila analysis o he LSMO ilm e ealed an LSMO ilm
hickness o 27.4 nm, as illus a ed in Figu e 1c. The ocking
cu e o he H 0.5Z 0.5O2003 e lec ion illus a ed in Figu e 1d
shows a FWHM o 0.032°, indica ing a good c ys alline quali y
o hese ilms.
Nex , wide- ange ecip ocal space maps (RSMs) and pole
igu es we e measu ed o he same s ack o de e mine he
c ys allog aphic phase and he o ien a ion o he HZO ilms.
Thus, in he ecip ocal space map o he H 0.5Z 0.5O2/
La0.7S 0.3MnO3/S TiO3(subs a e) he e os uc u es shown in
Figu e 2a, he subs a e e lec ions 001, 002, and 003 a e
iden i ied and labeled in da k blue colo on he RSM. The
ecip ocal space spo s co esponding o he epi axial LSMO
laye could no be dis inguished om hose o he STO
subs a e, due o lack o esolu ion in he wide ange RSM
measu emen s. He e, we like o speci y ha he sample was on
pu pose in-plane o a ed by an angle φ= 15°(wi h espec o
he in-plane a-di ec ion o he STO subs a e) o a oid he
con ibu ion o nonspecula spo s o he STO/LSMO). Fo
mo e cla i y, a wide- ange RSM o he same sample measu ed
a φ= 45°is shown in Figu e S1, whe e he 111, 112, 113, and
221 spo s o he STO/LSMO a e also p esen . The HZO ilms
g ow epi axially ou -o -plane on he LSMO/STO empla e,
and he ilms appea o consis o ou kinds o domains,
belonging o he same c ys allog aphic R3mphase, i s
epo ed by Wei e al.,
13
ha ing ou kinds o in-plane
o ien a ions, o a ed by 90°wi h espec o one ano he . A
simila domain s uc u e was epo ed by Nukala e al. o such
he e os uc u e.
14
The eason o his domain con igu a ion
could be explained by he 4- old symme y imposed by he
subs a e, as also desc ibed by Es andi
a e al.
18
and obse ed in
he e oepi axy o oxides
33
o semiconduc o s.
34
The esul s a e
illus a ed in Figu e 2a. The angles ma ked nex o he HZO
spo s in he wide- ange RSM ep esen he in-plane o a ion o
he co esponding HZO domains. F om he 022 HZO spo ,
we ex ac ed he la ice pa ame e alues o a= 7.17 and c=
Figu e 4. (a) HRTEM mic og aph showing an LSMO back elec ode
in ZA [
54
2] g owing along (012) planes. An FFT analysis shown
below in (b) and (c) was pe o med on he wo HZO egions ma ked
wi h ci cles. The le g ain is in [100] ZA, and he igh FFT shows a
supe posi ion o a [100] and a [
1
00] ZA which could be explained by
a win wi h a 2- old o a ion a he [003] di ec ion. These domains a e
indica ed by yellow and ed ec angles. The o ange ellipse in he le
FFT in (b) ma ks he addi ional [
2
22] mino i y g owing di ec ions o
he main [003] g owing di ec ion. The SAED pa e n in (d) is simila
o ha o FFT in (c).
Figu e 5. (a) Wide- ange RSM o a 5.6 nm hick HYO ilm deposi ed on La0.7S 0.3MnO3/S TiO3/Si (subs a e); he e, he obse ed spo s om he
HYO ilm a e assigned o (001)-o ien ed ilms o he e oelec ic R3mphase, consis ing o ou ypes o domains wi h 0°, 90°, 180°, and 270°in-
plane o ien a ion. The measu emen was acqui ed a φ= 0°. (b) Pole igu e o he HYO {022}, HYO {201}, STO/LSMO {011}, and Si {111}
e lec ions; he yellow, ed, whi e, and g een a ows indica e he ou domains wi h he amilies o spo s associa ed wi h hem. The measu ed 2θ
ange is 24.9−37.4°. The adial di ec ion ep esen s he χaxis anging om 0° o 90°, whe eas he azimu hal di ec ion ep esen s he φaxis wi h a
ange om 0° o 360°.
ACS Applied Ma e ials & In e aces www.acsami.o g Resea ch A icle
h ps://doi.o g/10.1021/acsami.4c10423
ACS Appl. Ma e . In e aces 2024, 16, 42534−42545
42537
8.82 Å. Mo e p ecise alues will be ex ac ed om he high-
esolu ion in-plane and ou -o -plane scans.
A pole igu e o he HZO/LSMO/STO (subs a e)
he e os uc u e is shown in Figu e 2b. The adial di ec ion
ep esen s χ, which anges be ween 51°and 85°, while he
azimu hal di ec ion ep esen s φ, which anges be ween 0°and
360°. Please no e ha he pole igu e p esen ed he e is no
measu ed o a single e lec ion bu o a wo- he a ange
be ween 23°and 37°. Thus, he HZO {201} and HZO {022}
amilies o e lec ions a e seen he e. The 12 poles o he HZO
{201}, measu ed a a 2θangle o 30.5°and he 12 poles o he
HZO {022} measu ed a a 2θangle o 35.6°a e clea ly isible
in he pole igu e. F om he c ys al symme y o he R3m
phase, one expec s h ee poles o each, spaced by 120°, bu
he p esence o ou kinds o c ys allog aphic domains o a ed
90°wi h espec o one ano he would gi e us he obse ed 12
poles. To ep oduce he measu ed da a, we conside ed ou
domains o he R3mphase ha ing he (001) o ien a ion and an
in-plane o a ion o 90°wi h espec o one ano he and
simula ed he pole igu e using he MTEX
35
simula ion ool.
The simula ed pole igu e is p esen ed in Figu e 2c. As can be
seen, bo h {201} (a chi = 71.15°) and {022} (a chi = 55.7°)
spo amilies could be ep oduced in he simula ion o he R3m
phase. I one conside s he o ho hombic Pca21phase wi h a
single (111)-o ien ed domain, h ee poles o a ed by 120°a e
expec ed o he HZO{
1 1
1}, om he c ys al symme y, as
simula ed wi h MTEX and also expe imen ally epo ed,
18
and
one pole o each o he {020}, {200}, and {002} e lec ions.
The HZO{
1 1
1} o he Pca21phase would be he
co esponding HZO{201} spo s o he R3mphase, and he
HZO{020} o he Pca21phase would be he co esponding
HZO{022} spo s o he R3mphase. Fu he mo e, i one
conside s (111) g own ilms consis ing o ou domains ha ing
ou in-plane o ien a ions o a ed 90°wi h espec o one
ano he and simula es he pole igu e, one can ep oduce he
12 spo s co esponding o HZO {
1 1
1}, as shown in he
simula ed pole igu e om Figu e 2d. To enhance cla i y, he
poles om each domain in he simula ion om Figu es 2c,d
a e indica ed wi h an a ow o a speci ic colo . Thus, we ha e
black, ed, blue, and g een a ows o indica e he ou domains.
Addi ionally, 12 poles a e expec ed a a wo- he a angle o
abou 35.6°, each o he {020}, {200}, and {002} e lec ions
con ibu ing wi h ou spo s in he simula ion, sepa a ed by a
90°angle in phi. The co esponding wo- he a angles o he
nonsymme y equi alen 020, 002, and 200 e lec ions o he
Pca21phase a e close o each o he . Thus, i is di icul o
dis inguish be ween he R3mand Pca21phases based on he
(low esolu ion) wide- ange RSM and pole igu es. One way o
dis inguish he R3mphase om he Pca21phase is o compa e
he in e plana spacing o he 12 spo s measu ed a a chi angle
o 71.15°wi h he in e plana dis ance o he specula 111 spo
o he Pca21phase. In he case o he Pca21phase, all spo s
should ha e he same d-spacing (due o he symme y),
whe eas o he R3mphase, he specula spo ( he 003
e lec ion) should ha e a la ge d-spacing alue.
13,14
Symme ic
wo- he a scans o he 12 poles measu ed a chi = 71.15°a e
plo ed oge he wi h he ou -o -plane e lec ion (chi = 0°) as
shown in Figu e 3a. The 12 spo s sha e nea ly he same 2θ
alues, whe eas he ou -o -plane e lec ion has a smalle 2θ
(la ge d-spacing), which is an indica ion o he R3mphase.
Ano he possibili y o dis inguish be ween he wo
polymo phs is o measu e symme ic 2θscans a a ound 35°
o he 12 spo s a a chi angle o 55.7°. In he case o he R3m
phase, he {022} spo s should ha e he same 2θ alues,
whe eas o he Pca21phase, he {020}, {200}, and {002}, each
con ibu ing wi h ou spo s ( om he ou domains)
sepa a ed by 90°in φwi h espec o one ano he (see Figu e
2d), should ha e sligh ly di e en 2θ alues. Simila ly, one can
analyze he 12 spo s measu ed a chi = 35°a a 2θo abou
50°. Thus, o he R3mphase, he {204} spo s should show he
same 2θ alues whe eas o he Pca21phase, he {022}, {202},
and {220} each con ibu ing wi h ou spo s sepa a ed by 90°
in φwi h espec o one ano he should ha e sligh ly di e en
2θ alues. The symme ic wo- he a scans o he 12 poles
measu ed a chi = 55.7°and a 2θ alue a ound 35°a e
p esen ed in Figu e 3b, and he 12 poles measu ed a chi = 35°
and a 2θ alue o a ound 50°a e shown in Figu e 3c. To
imp o e he ease o unde s anding o he eade , he 2θscans
o each {020}, {200}, and {002} as in Figu e 3b and {022},
{202}, and {220} e lec ions as in Figu e 3c co esponding o
he Pca21phase a e depic ed using dis inc colo s: ed, g een,
and blue. Thus, in he case o he Pca21phase, he e lec ions
o he same colo (sepa a ed in phi by 90°) should sha e he
same 2- he a alue and should be sligh ly di e en o each
colo . When hese 2θ alues a e all equal, assuming he
e oelec ic Pca21phase, would imply a=b=c, which
con adic s he epo ed me ics o his c ys allog aphic phase,
making i unlikely ha he in es iga ed ilms a e in he Pca21
phase. These esul s a e ins ead in good ag eemen wi h he
me ics o he R3mphase. Howe e , he exis ence o a sepa a e
Figu e 6. (a) Plan- iew SAED pa e n o Si/STO/LSMO/HYO. The
subs a e Si is o ien ed in he [100] zone axis as well as he STO
which is o a ed by 45°wi h espec o Si. HYO shows mainly wo
ypes o e lec ions. The ed ci cles depic he HYO [001] zone axis
wi h {220} e lec ions aligned wi h he STO (101) planes ( ed a ow)
while he e lec ions ma ked by he g een ci cles a e aligned wi h he
STO (110) planes (g een a ow). (b) shows a high- esolu ion TEM
mic og aph in he same iewing di ec ion which is con i med by he
FFT in (c) showing he same di ac ion pa e n as in a). In (c), he
g een, ed, and yellow ci cles we e used o il e ing he FFT and
building a colo ed in e se FFT shown in (d). The majo i y o he
domains ha e an [003] ou -o -plane o ien a ion wi h {220} e lec ions
ma ked by ed and g een ci cles. The mino i y (yellow) is o ien ed
wi h STO (200).
ACS Applied Ma e ials & In e aces www.acsami.o g Resea ch A icle
h ps://doi.o g/10.1021/acsami.4c10423
ACS Appl. Ma e . In e aces 2024, 16, 42534−42545
42538

e oelec ic hombohed al phase which is no jus a s uc u al
dis o ion caused by epi axial s ess dis o ion o he Pca21
o ho hombic phase is s ill unde deba e, as sugges ed by Fina
and Sanchez.
36
To gain a clea e unde s anding o how he scans om
Figu e 3c ela e o pa icula e lec ions and domains,
simula ed pole igu es o bo h R3mand Pca21phases o he
12 spo s a chi = 35°and a 2θo abou 50°a e shown in Figu e
S5.
3.1.1.2. T ansmission Elec on Mic oscopy Cha ac e -
iza ion. The same HZO ilm on he LSMO-bu e ed (001)-
o ien ed Nb:STO subs a e analyzed by XRD was used o
TEM cha ac e iza ion. The HRTEM mic og aph in Figu e 4a
shows he LSMO back elec ode g owing along (012) planes
while he HZO hin ilm g ows mainly in he [003] di ec ion o
he R3mphase. Howe e , a mino i y g owing di ec ion along [
2
22] was also obse ed (ma ked wi h an o ange ellipse in he
FFT in Figu e 4b, as also epo ed by Wei e al).
13
This
mino i y g ow h di ec ion is no obse ed in XRD measu e-
men s. Di e en domains o he HZO we e p esen wi h (02
2
)
planes o ien ed o he igh o he le , hese planes a e
indica ed by a black and a ed a ow, espec i ely. F equen ly
bo h o ien a ions supe impose, as can be seen in he FFT in
Figu e 4c. This pa e n can be explained by a 2- old win
o a ion a ound he [003] g owing di ec ion. Acco dingly, we
can see he e a leas wo o he ou domains obse ed in XRD.
The missing wo domains canno be dis inguished in his
o ien a ion because hey ha e he same di ac ion pa e n. Fo
ins ance, he [110] zone axis (ZA) is equi alen o he [100]
ZA in a hombohed al symme y. A clea iew o he bounda y
o he domains, howe e , was no ound in many mic og aphs
examined.
The selec ed a ea elec on di ac ion (SAED) pa e n in
Figu e 4d was aligned pa allel o he [100] ZA o he STO. In
Figu e 7. (a) High- esolu ion 2θ/ωscan o a 4.7 nm hick HZO ilm deposi ed on sapphi e. (b) Wide ange RSM o he same ilm on sapphi e.
The spo s belonging o sapphi e a e deno ed in black, whe eas he spo s om he HZO ilm a e deno ed in ed. The angles deno ed nex o he
HZO spo s ep esen he in-plane o a ion o he co esponding HZO domains assigned o he R3mphase. The sapphi e subs a e was o a ed by
phi = 30°du ing he RSM measu emen . (c) Pole igu e o he HZO {201}, HZO {022}, sapphi e {10
2
} and sapphi e {104} spo s. The adial
di ec ion ep esen s he χaxis anging om 0° o 90°, whe eas he azimu hal di ec ion ep esen s he φaxis wi h a ange om 0° o 360°.
ACS Applied Ma e ials & In e aces www.acsami.o g Resea ch A icle
h ps://doi.o g/10.1021/acsami.4c10423
ACS Appl. Ma e . In e aces 2024, 16, 42534−42545
42539
his o ien a ion, he LSMO is in he [
2
2
1
] ZA and he
epi axial g ow h o all ilms is clea ly isible. The di ac ion
pa e ns o STO and LSMO a e indica ed by black ho izon al
lines. Fo HZO, he (003) e lec ions in he g owing di ec ion
a e mos p ominen , while in-plane no e lec ions can be
e idenced, and also he diagonal planes a e e y ain (ma ked
wi h o ange ci cles).
Fu he mo e, om his SAED pa e n, a di e ence in he d-
alues o (006) and (042) can be obse ed, as illus a ed by he
dashed ing a ound he (006) e lec ions. These e lec ions
would co espond o he {222} planes o he e oelec ic
o ho hombic phase, which canno ha e a d- alue di e ence
(in a elaxed s uc u e) due o he o ho hombic symme y. An
e ec o an ellip ical dis o ion o he di ac ion pa e n by
unp ecise adjus men s o he p ojec o lens sys em can be
Table 1. aand cLa ice Pa ame e s o he S ained Doped
H O2Films Deposi ed on Di e en sys ems in his Wo k
We e Ex ac ed om High-Resolu ion 2θ/ωScans and In-
Plane Scans
a
he e os uc u e a(Å) c(Å)
8 nm HZO/LSMO/STO 7.194 8.899
7.7 nm HYO/LSMO/STO 7.170 8.953
7.6 nm HZO/LSMO/STO/Si 7.236 8.822
5.6 nm HYO/LSMO/STO/Si 7.212 8.883
6.2 nm HYO/Al2O37.207 8.879
2.6 nm HYO/Al2O37.239 8.773
4.7 nm HZO/Al2O37.219 8.805
H O2,R3m(no. 160); a=b;α= 90°;β= 90°;γ= 120°
a om x y z occupancy
H 0.83335 0.16665 0.25089 1
H 0.00000 0.00000 0.58415 1
O 0.14966 0.85034 0.15904 1
O 0.48699 0.51301 0.32788 1
O 0.00000 0.00000 0.85998 1
O 0.00000 0.00000 0.35364 1
a
The c ys allog aphic s uc u e is he one epo ed by Wei e al. in e
13 adap ed o he la ice pa ame e s measu ed in his wo k.
Figu e 8. Pole igu e simula ion o he HZO ilms using MTEX. (a) Phase R3m, conside ing (001) ou -o -plane o ien a ion and p esence o wo
domains (indica ed by he ed and he blue a ows), wi h an in-plane angula o a ion o 180°wi h espec o one ano he . (b) Phase Pca21
conside ing (111) ou -o -plane o ien a ion, and he p esence o wo domains, wi h an in-plane angula o a ion o 180°wi h espec o one ano he .
Figu e 9. In-plane XRD o he HYO(6.2 nm)/Al2O3: (a) in-plane
scan showing he HYO(220) and Al2O3(300) e lec ions; he la ice
pa ame e s and he FWHM a e gi en in he inse . (b) In-plane phi-
scan o he HYO(220) egion showing a 6- old in-plane symme y;
he FWHM is 4.1°.
ACS Applied Ma e ials & In e aces www.acsami.o g Resea ch A icle
h ps://doi.o g/10.1021/acsami.4c10423
ACS Appl. Ma e . In e aces 2024, 16, 42534−42545
42540
uled ou , because he e lec ions o he polyc ys alline P om
he FIB p epa a ion a e no ellip ically dis o ed.
Supe imposed a e he [100] ZA o STO and he [
2
2
1
] ZA
o LSMO which a e illus a ed by black e ical lines. The
e lec ions o HZO in he diagonal a e e y ain . The e o e,
he ec angles om he FFT a e used he e double sized o
illus a e he di ac ion pa e n o HZO. The ings in he
SAED pa e n in (d) o igina e om he p o ec i e P coa ing
used in he FIB.
3.1.2. H 0.93Y0.07O2(HYO) Films Deposi ed on S TiO3-
Bu e ed Si (100)-O ien ed Subs a es. 3.1.2.1. X- ay
Di ac ion Analysis. Y ium-doped Ha nia ilms
H 0.93Y0.07O2(HYO) we e deposi ed on Si (100)-o ien ed
subs a es on which a 10 nm S TiO3(STO) was deposi ed by
molecula beam epi axy (MBE) a Texas S a e Uni e si y.
37,38
The H 0.93Y0.07O2/La0.7S 0.3MnO3epi axial s ack was deposi ed
by PLD, using he same deposi ion pa ame e s as hose
deposi ed on Nb:S TiO3subs a es. The wide- ange RSM o
he doped ha nia ilms deposi ed on La0.7S 0.3MnO3/S TiO3/
Si (subs a e) shows ea u es simila o hose deposi ed on
La0.7S 0.3MnO3/S TiO3(subs a e), an ou -o -plane o ien a-
ion, and he p esence o ou kinds o domains wi h di e en
in-plane o ien a ions, as desc ibed be o e. The wide- ange
RSM was acqui ed a φ= 0°, and besides he specula Si(004),
he Si(022) spo was isible in his di ac ion geome y. The
nonspecula STO/LSMO di ac ion spo s
1
11,
1
12, and
2
21
a e also p esen . The esul s a e shown in Figu e 5a. A pole
igu e o HYO{022}, HYO{201}, STO/LSMO{011}, and
Si{111} measu ed o a 2θ anging om 34.8° o 47.8°is
shown in Figu e 5b. The 12 spo s o he HYO{022} and
HYO{201} a ising om he ou domains a e p esen on he
pole igu e, simila o he esul s om doped ha nia ilms
deposi ed on LSMO-bu e ed STO subs a es.
3.1.2.2. T ansmission Elec on Mic oscopy Cha ac e -
iza ion. The TEM p epa a ion o plan- iew samples g own
on he STO subs a es was no success ul because in 3 ials
he samples b oke apa du ing mechanical polishing be o e
inishing he p epa a ion. As a esul , a Z /Y:H O2sample on
S TiO3-bu e ed Si was p epa ed. This sample should be
Figu e 10. (a) HRTEM mic og aph o Al2O3in [210] ZA and HZO
in [100] ZA. Two di e en domains we e ound ia FFT analysis in
(b) and (c); again, he [100] ZA is mi o ed a he (003) plane. The
SAED pa e n in (d) shows bo h domains, and he Al2O3[210] ZA is
indica ed by e ical black lines. In his sys em, he (003) e lec ions
a e a he weak compa ed o he STO/LSMO sys em, and he (060)
e lec ion is isible in-plane. In FFTs e en he (030) planes. The
di ac ion ings o igina e om he P coa ing which was used o
p o ec du ing FIB p epa a ion.
Figu e 11. Fe oelec ic cha ac e iza ion. Dynamic e oelec ic pola iza ion hys e esis loops o he (a) Cu/H 0.5Z 0.5O2/La0.7S 0.3MnO3/S TiO3,
(b) Cu/H 0.93Y0.07O2/La0.7S 0.3MnO3/S TiO3/Si, and (c) Cu/H 0.93Y0.07O2/La0.7S 0.3MnO3/Nb:S TiO3capaci o s. (d) “Wake-up” e ec
in es iga ion on a p is ine e oelec ic capaci o .
ACS Applied Ma e ials & In e aces www.acsami.o g Resea ch A icle
h ps://doi.o g/10.1021/acsami.4c10423
ACS Appl. Ma e . In e aces 2024, 16, 42534−42545
42541
ep esen a i e, as indica ed by he XRD analysis. A SAED
pa e n o he plan- iew sample wi h Si o ien ed in he [100]
ZA is shown in Figu e 6a. He e, he HYO o ms a simila
di ac ion pa e n as epo ed by Wei e al.
13
wi h 12 {120}
e lec ions which can be explained by HYO ha ing a leas wo
domains o a ed in plane by 90°. The g een and ed ci cles in
Figu e 6a indica e wo domains. The domains ma ked wi h
g een ci cles a e o ien ed wi h he STO (110) planes while he
domains ma ked wi h ed ci cles a e o ien ed wi h he STO
(101) planes. Adding he 2- old win o a ion obse ed in he
c oss-sec ional TEM analysis esul s in he ou domains
obse ed in XRD pole igu es. These ou ypes o domains
may o m du ing he c ys al nuclea ion du ing PLD, whe e
each nucleus can be o ien ed wi h STO (110), (−110), (1−
10), o (−1−10). The same colo code is used in he FFT
shown in Figu e 6c o he HRTEM mic og aph in shown
Figu e 6b. These colo ed ci cles we e used o he il e ed
in e se FFT in Figu e 6d which illus a es he g anula
mic os uc u e o he HYO hin ilm. Howe e , all indi idual
g ains a e epi axial.
A TEM image in a di e en o ien a ion, il ed away om he
[100] zone axis o Si, is shown in Figu e S4. He e, s ong
Moi e inges appea which we e again used o a il e ed
in e se FFT. F om hese alse colo images, he a e age wid h
o 20 domains can be oughly es ima ed o be 10 nm ±3 nm.
The high c ys allini y o he HZO, as indica ed by he ocking
cu e, a gues o cohe en in e aces be ween hese 10 nm-
sized domains as would be he case o wins.
In addi ion o he obse ed majo i y e lec ions in he SAED
pa e n shown in Figu e 6a om HYO wi h he [003] ou -o -
plane o ien a ion, a mino i y is obse ed in he FFTs (see
yellow ci cles in Figu e 6c and blue ci cles in Figu e S4b) ha
is o ien ed wi h he STO (200) planes.
3.1.3. H 0.5Z 0.5O2Films Deposi ed on Al2O3(0001)
Subs a es. 3.1.3.1. X- ay Di ac ion Analysis. Nex , Y- and
Z -doped H O2 ilms we e deposi ed on igonal Al2O3(0001)
single c ys al subs a es. A 5 nm hick H 0.5Z 0.5O2 ilm
deposi ed on Al2O3(0001) is cha ac e ized by XRD
expe imen s, as p esen ed in Figu e 7. The high- esolu ion
ω-2θscan shown in Figu e 7a allowed o he p ecise
measu emen o he cla ice pa ame e , lis ed in Table 1,
and de e mina ion o he ilm hickness based on he hickness-
oscilla ion pe iod, which in his case was 4.7 nm. A wide- ange
RSM shown in Figu e 7b was measu ed a an in-plane o a ion
angle o φ= 30°. Thus, in addi ion o he specula 006 spo ,
he 113, 116, and 119 spo s o he Al2O3subs a e a e obse ed
a his in-plane phi angle. The HZO ilm appea s o consis o
wo kinds o domains, belonging o he same -phase R3mno.
160, o a ed in-plane by 180°wi h espec o one ano he and
by 30°wi h espec o he Al2O3c ys allog aphic cell. The
co esponding angles o hese wo domains, which we e
ob ained om he simula ion o he RSM a e deno ed nex o
indexed HZO spo s in he igu e. The pole igu e o he same
sample is p esen ed in Figu e 7c. The HZO {201}, HZO
{022}, sapphi e {10
2
} and sapphi e {104} spo s om he
la ice planes symme ically equi alen a e ma ked on he pole
igu e. The 2θ anges om 25° o 37.8°.
The sapphi e {10
2
} and {104} spo s e lec he igonal
symme y o he phase, esul ing in h ee spo s on he pole
igu e o each amily. The HZO {201} and HZO {022} spo s
should also show h ee spo s each, acco ding o he
c ys allog aphic symme y o he R3mphase, bu six spo s
a e obse ed expe imen ally. This can be explained by he
p esence o wo kinds o domains, belonging o he same
igonal phase R3m, o a ed in-plane by 180°wi h espec o
one ano he , as deduced o he wide- ange RSM simula ion
discussed abo e. This domain con igu a ion has been epo ed
in p io s udies on HZO ilms deposi ed on GaN and
sapphi e,
14
also being imposed by he symme y o he
subs a e. To e i y ha , a pole igu e o he HZO ilms was
simula ed using MTEX, acco ding o he p oposed assump ion,
and he esul s a e shown in Figu e 8a. The simula ion
co esponds o he indings ob ained om he expe imen s.
The simula ion o he Pca21phase o a (111) o ien ed
monodomain ilm gi es h ee poles om he {1
1
1}, and one
poles om each {020}, {002} and {200} e lec ion amilies. I
one conside s wo (111) domains o a ed in-plane by 180°
wi h espec o one ano he , he simula ions ep oduce he
obse ed six {111} poles, as well as he six poles o he {020},
{002}, and {200} ones, as shown in Figu e 8b. Also, in his
case, i is di icul o dis inguish be ween he R3mand Pca21
c ys allog aphic phases om he pole igu e XRD measu e-
men s.
Conside ing he ela i ely uni o m in ensi y dis ibu ion o
he e lec ions co esponding o a speci ic amily o spo s, one
can s a e ha he olume ac ion o he wo ypes o HZO
domains appea s o be equally dis ibu ed.
The measu ed wide- ange RSM o he HZO deposi ed on
sapphi e could also be assigned o a e oelec ic o ho hombic
phase like, o example, he one epo ed by Xu e al.
39
Wi h
his assump ion, he ilms a e epi axial and ha e an (111) ou -
o -plane o ien a ion. Such an example is illus a ed in Figu e
S2, whe e he obse ed spo s in he RSM om he H O2 ilms
deposi ed on sapphi e a e assigned o he abo e-men ioned
c ys allog aphic phases.
To p ecisely de e mine he in-plane la ice pa ame e s o he
deposi ed ilms, we pe o med in-plane X- ay di ac ion
measu emen s. Thus, in he case o doped ha nia ilms
deposi ed on sapphi e, he HYO (220) and Al2O3(300)
e lec ions we e obse ed, as can be seen in Figu e 9a. This
means ha he (220) c ys allog aphic planes o he epi axially
doped ha nia ilms a e pa allel wi h he (300) planes o he
unde lying sapphi e subs a es and pe pendicula o he HYO
(003) planes which a e pa allel wi h he sample su ace. The
epi axial ela ionship in his case conside ing he R3m
c ys allog aphic phase is [220] HZO(001) //[300]
Al2O3(001). I one would conside he e oelec ic o ho-
hombic phase Pca21ins ead o he igonal phase (R3m), he
(111) c ys allog aphic plane is pa allel wi h he sample su ace,
whe eas he (20
2
) plane o ms an angle o 89.84°wi h he
(111) plane, e y close o 90°. Fo his eason, based on he in-
plane XRD measu emen s, one canno dis inguish be ween he
e oelec ic o ho hombic phase and he hombohed al one.
F om hese high- esolu ion in-plane scans, one can p ecisely
de e mine he in-plane la ice pa ame e o he hombohed al
HYO ilms deposi ed on sapphi e, which is 7.207 Å. Nex , he
2-The aChi/Phi axis was ixed a he HYO(220) peak posi ion,
and a phi-scan was pe o med, wi h a comple e 360°phi-axis
scan ange. The 3- old in-plane symme y expec ed o he
hombohed al ilms combined wi h he p esence o he wo
180°in-plane o a ed domains is consis en wi h he obse ed
six e lec ions spaced by 60° om his scan and wi h he esul s
om he pole igu es o he HZO ilms deposi ed on sapphi e,
discussed abo e. The in-plane phi scan shown in Figu e 9b is
equi alen o an in-plane ocking cu e, and he wid h o he
e lec ions gi es us a measu e o he in-plane mosaici y. In he
ACS Applied Ma e ials & In e aces www.acsami.o g Resea ch A icle
h ps://doi.o g/10.1021/acsami.4c10423
ACS Appl. Ma e . In e aces 2024, 16, 42534−42545
42542