Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
MR41CH14-Ramanathan ARI 3 June 2011 8:32
Oxide Electronics UtilizingUltrafast Metal-InsulatorTransitionsZheng Yang, Changhyun Ko, and Shriram RamanathanSchool of Engineering and Applied Sciences, Harvard University, Cambridge,Massachusetts 02138; email: [email protected]
Annu. Rev. Mater. Res. 2011. 41:337–67
First published online as a Review in Advance onMarch 30, 2011
The Annual Review of Materials Research is online atmatsci.annualreviews.org
This article’s doi:10.1146/annurev-matsci-062910-100347
Copyright c© 2011 by Annual Reviews.All rights reserved
1531-7331/11/0804-0337$20.00
Keywords
oxide electronics, ultrafast switch, Mott transition, correlated oxides,vanadium dioxide (VO2)
Abstract
Although phase transitions have long been a centerpiece of condensed mat-ter materials science studies, a number of recent efforts focus on potentiallyexploiting the resulting functional property changes in novel electronicsand photonics as well as understanding emergent phenomena. This is quitetimely, given a grand challenge in twenty-first-century physical sciences is re-lated to enabling continued advances in information processing and storagebeyond conventional CMOS scaling. In this brief review, we discuss synthe-sis of strongly correlated oxides, mechanisms of metal-insulator transitions,and exploratory electron devices that are being studied. Particular empha-sis is placed on vanadium dioxide, which undergoes a sharp metal-insulatortransition near room temperature at ultrafast timescales. The article beginswith an introduction to metal-insulator transition in oxides, followed by abrief discussion on the mechanisms leading to the phase transition. Therole of materials synthesis in influencing functional properties is discussedbriefly. Recent efforts on realizing novel devices such as field effect switches,optical detectors, nonlinear circuit components, and solid-state sensors arereviewed. The article concludes with a brief discussion on future researchdirections that may be worth consideration.
337
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
MIT: metal-insulatortransition
VO2: vanadiumdioxide
1. INTRODUCTION
Metal-insulator transition (MIT) in oxides (1) is a topic of long-standing interest in condensedmatter materials sciences. Experimental and theoretical studies to unravel the mechanism of MIThave been ongoing for nearly half a century. A number of reviews on MIT mechanisms and mate-rials in the past 40 years indicate the consistent interest in this subject (1–10). Significant interestin MIT in oxide materials (11–19) initiated from Morin’s (11) seminal paper on phase transitionsin binary transition-metal oxides in the 1950s and was further promoted by MIT-relevant dis-coveries in correlated oxides such as, but not limited to, high-temperature superconductivity incuprates (20) and colossal magnetoresistance in manganites (21). As Morin noted, the resistancein some transition-metal oxides, such as vanadium oxides [dioxide (VO2), sesquioxide (V2O3), andmonoxide (VO)] and titanium sequioxide (Ti2O3), increases by several orders of magnitude whenthe temperature decreases from high to low across the transition temperature (TMIT) (11). Recentyears have seen major advances in growth of thin-film oxides by a variety of techniques such asmolecular beam epitaxy, sputtering, chemical vapor deposition (CVD), and pulsed laser deposition(PLD), and resurgence in the oxides MIT field is taking place, with growing emphasis on exploringdevice applications of this spectacular phenomenon. A principal focus of this review is on oxidematerials and devices utilizing MITs, along with a qualitative discussion on the mechanisms.Particular emphasis is placed on emerging research devices operable near room temperaturewherein the phase transition can be triggered by small thermal, electrical, or optical perturba-tions. We point the reader to the article by Imada et al. (8) for a detailed account on the theory ofMITs.
Figure 1a shows a number of correlated oxides with corresponding MIT temperatures(3, 6, 8, 12, 22). The elementary mechanism leading to the transition as well as the resistiv-ity magnitude change are different among these materials and are considered in later sections.Figure 1a shows only the oxides with MIT triggered by temperature, and not other correlatedoxides with non-temperature-driven mechanisms (such as band-filling control, displayed by somemanganites and cuprates). Among these correlated oxides, the TMIT of VO2 is close to roomtemperature, which is ∼340 K in bulk crystals. This unique property coupled with an impressivefive-order-of-magnitude resistance change (in single crystals) across the transition make VO2 acompelling candidate for exploratory device research. This article uses VO2 as a representativeexample for discussing the phase transition and potential devices that could be realized.
As mentioned above, a frequently employed approach to induce MIT is thermal triggering,i.e., changing the temperature by heating or cooling. As shown in Figure 1b (part i ), with atemperature increase from 341 to 344 K, the resistance of the VO2 thin film decreases by nearlyfour orders (23). Other approaches to trigger the phase transition include electrical (19, 24), optical(19, 25, 26), magnetic (19), and strain (27) excitations. An example of each of these approachesis shown in Figure 1b, parts ii–v, respectively. Figure 1c shows how one could utilize MIT asa switch. An external perturbation in the form of thermal, electrical, optical, or magnetic fieldcan trigger the phase transition, leading from a high-resistance OFF state to a low-resistance ONstate, i.e., inducing the switching behavior. In some cases, combination of two or more excitationscan lead to the transition at different thresholds.
Figure 2a points out potential applications for MITs in oxides in materials physics andsolid-state electronics, whereas Figure 2b shows representative device concepts utilizing MIT incorrelated oxides, including two-terminal electronic switches (28), three-terminal (gated) elec-tronic switch devices (29), optical devices (30), electronic oscillators (31), metamaterials (32),memristive devices (33), thermal sensors (34), and chemical sensors (35). The subsections inwhich each is discussed in the review are indicated in Figure 2b. If one considers the physical
338 Yang · Ko · Ramanathan
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
MIT
ab
cSelected oxides
T MIT
(K
)6
02
00
30
04
00
50
01
,20
01
00
20
00
40
06
00
80
01
,00
0–
20
02
04
06
0
10
5
10
4
10
3
10
2
10
1
10
5
10
3
10
1
10
7
10
9
10
5
10
7
10
9
10
6
10
8
10
10
Mo
no
clin
icin
sula
tor
Str
on
gly
corr
ela
ted
me
tal
Ru
tile
me
tal
Resistance (Ω)
T (K
)
VO
2
Resistance (Ω)
Resistance (Ω)
Ap
pli
ed
vo
lta
ge
(V
)T
ime
(μ
s)
R L
20
K
Pr 0
.7C
a 0.3
Mn
O3
Pr 0
.7C
a 0.3
Mn
O3
30
K
Lase
r p
uls
e
0.2
mm
10
V
Th
erm
al,
ele
ctri
cal,
op
tica
l, m
ag
ne
tic,
or
stra
in-d
riv
en
ex
cita
tio
ns
(ii)
Ele
ctri
cal t
rig
ge
rin
g(i
ii) O
pti
cal t
rig
ge
rin
g
(i)
Th
erm
al t
rig
ge
rin
g
34
13
42
34
33
44
ON
sta
te
Low
re
sist
ance
Hig
h r
esi
stan
ce
OF
F s
tate
Resistance
Ox
ide
MIT
trig
ge
rin
g
V4O
7 (
25
0 K
)
Nd
NiO
3 (
20
1 K
)
V6O
11 (
17
7 K
)
V2O
3 (
16
5 K
)
V6O
13 (
15
0 K
)
PrN
iO3 (
13
5 K
)
V5O
9 (
13
5 K
)
VO
(1
26
K)
VO
2 (
34
0 K
)
Sm
NiO
3 (
40
3 K
)
Ti3O
5 (
44
8 K
)
LaC
oO
3 (
50
0 K
)
Ti2O
3
Nb
O2 (
1,0
81
K)
Liq
uid
nit
rog
en
tem
pe
ratu
re
Ro
om
te
mp
era
ture
Fe3O
4 (
12
1 K
)
V8O
15 (
70
K)
(40
0–
60
0 K
)
02
46
81
01
23
30
34
03
50
36
03
70
38
0
10
–2
10
2
10
6
10
0
10
4
1.2
0.8
0.4
0.0
–0
.4
–0
.8
–1
.2
0.8
0.4
0.0
–0
.4
–0
.8
0.0
0.2
5
0.5
0
0.7
5
1.0
Resistivity (Ω cm)
Stress (GPa)
Tem
pe
ratu
re (
K)
Ma
gn
eti
c fi
eld
(T
)
Strain (%)
Pr 0
.7C
a 0.3
Mn
O3
VO
2M
I
35
K 75
K
10
0 K
(iv)
Mag
ne
tic
trig
ge
rin
g(v
) S
trai
n t
rig
ge
rin
g
Fraction of metallic phase
Figu
re1
The
met
al-i
nsul
ator
tran
sitio
n(M
IT)s
witc
h.(a
)MIT
tem
pera
ture
(TM
IT)o
fsom
ese
lect
edox
ides
(bul
kcr
ysta
ls).
Ext
erna
lstr
ess
orsu
bstr
ate-
driv
enco
nstr
aint
sca
nsi
gnifi
cant
lyin
fluen
ceth
etr
ansi
tion
tem
pera
ture
and
the
resi
stiv
itych
ange
.(b)
MIT
-tri
gger
ing
appr
oach
esin
corr
elat
edox
ides
.(i)
Tem
pera
ture
-tri
gger
edM
ITin
VO
2.(ii
)Ele
ctri
cally
trig
gere
dM
ITin
Pr 0
.7C
a 0.3
MnO
3.(ii
i)O
ptic
ally
trig
gere
dM
ITin
Pr 0
.7C
a 0.3
MnO
3.(iv
)Mag
netic
ally
trig
gere
dM
ITin
Pr 0
.7C
a 0.3
MnO
3.(v
)Str
ain/
stre
ssef
fect
son
MIT
inV
O2.
Pan
elia
dapt
edw
ithpe
rmis
sion
from
Ref
eren
ce23
;pan
els
ii–iv
adap
ted
with
perm
issi
onfr
omR
efer
ence
19;p
anel
vad
apte
dw
ithpe
rmis
sion
from
Ref
eren
ce27
.(c)
Bas
icco
ncep
tofu
tiliz
ing
MIT
inco
rrel
ated
oxid
esas
asw
itch,
with
the
high
-res
ista
nce,
insu
latin
gan
dlo
w-r
esis
tanc
e,m
etal
licst
ates
onbo
thsi
des
ofM
IT,d
efine
das
OFF
and
ON
stat
es,r
espe
ctiv
ely.
The
switc
hing
ofth
ede
vice
can
betr
igge
red
ther
mal
ly,e
lect
rica
lly,o
ptic
ally
,mag
netic
ally
,and
byst
rain
driv
e,co
rres
pond
ing
toth
eM
IT-t
rigg
erin
gap
proa
ches
show
nin
pane
lb.
www.annualreviews.org • Oxide Electronics Utilizing Ultrafast Metal-Insulator Transitions 339
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
VI
Al2O3VO2
Temperaturestage
°K
Tunablematerials
Chargecontrol circuit
elements
Thermalcomputing
Energydissipation
3-D circuits Spintronics
Quantumactuation
Hystereticmemory
Nonlinearneural netelements
Solid-oxidefuel cell
electrodes
MEMSdevices
Circuitprotection
Meta-materials
Ultrafastswitches for
logic
Thermalswitch
Thermalsensor
Chemicalsensors
Mottphysics
Memristiveelements
Opticalmodulators
Ultrafastoxide metal-
insulatortransition
Memristive device(Section 4.2.4)
Thermal sensor(Section 4.3.3)
Chemical sensor(Section 4.3.4)
Oscillator(Section 4.2.3)
Two-terminalelectronic switch
(Section 4.2.1)
Metamaterial device(Section 4.3.2)
Gated electronicswitch (Mott FET)
(Section 4.2.2)Optical device(Section 4.3.1)
VO2
10 V70 °C
2 µm
IVO2
C Inb
Vnb
VO2
NW
Pt
Ga
-In
-Sn
μhp
VO2 tab
180 nm Si
40 nm etch
Si3 µm SiO2
1 μm240 µm240 µm
GateV+ V–
I+ I–
Au
SiO2VO2
NIS
T
Oxide MITdevices in
this review
a
b
Figure 2Devices utilizing MITin correlated oxides.(a) MIT in oxides: acompelling case formaterials physics andsolid-state electronicsresearch. (b) A selectedcollection of deviceapplications that useMIT in correlatedoxides and that arediscussed in thisreview, includingtwo-terminalelectronic switchdevices, three-terminal(gated) electronicswitch devices, opticaldevices, electronicoscillators,metamaterial devices,memristive devices,thermal sensors, andchemical sensors. Thesubsection numbers ofeach device in thereview are indicated inblue text. Panel badapted withpermission fromReferences 28–35.
340 Yang · Ko · Ramanathan
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
CMOS:complementarymetal-oxidesemiconductor
FET: field effecttransistor
MOS: metal-oxidesemiconductor
scaling limits that Si complementary metal-oxide semiconductor (CMOS) field effect transistor(FET) technology is facing (36–40), novel, correlated oxide-based devices are worth exploring asalternate candidates for information processing, and ultrafast phase transitions may be interestingcomputational vectors.
2. ELEMENTARY MECHANISMS OF THEMETAL-INSULATOR TRANSITION
Mott (1, 3, 6, 41) forwarded a model for MIT considering electron-electron interactions. PerMott’s theory, a critical carrier density nc is proposed as n1/3
c aH ≈ 0.2, where aH is the Bohr radiusof the material. Once the carrier density in the material is larger than nc, a phase transition occursdue to the electron-electron interaction effect, i.e., correlation. The MIT discussed above is namedMott MIT (or Mott-Hubbard MIT), and the insulator therein is accordingly referred to as theMott-Hubbard insulator. However, MIT phenomena can also happen due to reasons other than anelectron correlation effect. For example, MIT can also occur from an electron-phonon (electron-lattice) interaction, which is referred to as Peierls MIT (42, 43). Generally, the Peierls MIT arisesfrom a lattice structural change in a material. The structure change induces a lattice deformation,which modifies the periodic ionic potential in the material and consequently results in a bandstructure change. The blue bronze K0.3MoO3 is a typical example of a Peierls insulator, whichundergoes a Peierls MIT at 181 K, with an evident conductivity change accompanied by a structurechange (43). The electron localization effect due to disorder (44–46) can also lead to AndersonMIT. In the 1950s, Anderson (44) observed that randomly distributed lattice defects could lead toan insulting state. The disorder-induced electron localizations result in the existence of a mobilityedge, which separates the localized and delocalized states in the bands. Insulating states formonce the Fermi level is between a band edge and its mobility edge. Anderson MIT commonlyappears in strongly disordered materials and materials with strong impurity scattering, such asheavily doped semiconductors (e.g., Si:P). An insulator under the frame of conventional bandtheory (without consideration of the electron-electron interactions) is termed a band insulator (orBloch-Wilson insulator). Diamond and common undoped semiconductors are examples of bandinsulators. Table 1 briefly summarizes the above four categories of insulators. Mott-Hubbardinsulator and charge-transfer insulator are other proposed classifications. The difference betweenthe two is due to the magnitudes of two parameters: Hubbard Coulomb repulsion energy U andcharge-transfer energy �. The smaller of the two parameters determines the type of classification.For example, V2O3 is a typical Mott-Hubbard insulator, whereas NdNiO3 is a charge-transferinsulator.
Generally, the control parameters for MIT can be classified into three categories: temperaturecontrol, bandwidth control, and band-filling control. Temperature control, the most straightfor-ward case, simply involves changes in temperature by heating or cooling. The oxide systems shownin Figure 1a display temperature-controlled MIT with a TMIT defined in the x axis. Bandwidthcontrol can be induced by external or internal pressure. Internal pressure can be achieved withsubstitutional doping with atoms of different sizes. A typical example is RNiO3-type (R = Pr, Nd,and Sm) material. Band-filling control is related to tuning the doping level with donors or accep-tors. MIT in most cuprates and manganites is the band-filling-control type; in these materials,high-temperature superconductivity and colossal magnetoresistance, respectively, are observed.Some materials can show more than one control parameter; for example, RNiO3 can be eithertemperature or bandwidth controlled. Table 1 summarizes the three control types. In additionto the above three control parameters, dimensionality can also be a control parameter inducing
www.annualreviews.org • Oxide Electronics Utilizing Ultrafast Metal-Insulator Transitions 341
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
Table 1 Classification of insulators and control mechanisms of metal-insulator transition (MIT)
Classifications of insulatorsType OriginBloch-Wilson insulator (also known as bandinsulator)
Under the framework of conventional band theory
Mott-Hubbard insulator MIT occurs due to electron-electron interactions (correlations)Peierls insulator MIT occurs due to electron-phonon (lattice) interactionsAnderson insulator MIT occurs due to disorder-induced localizationControl mechanisms of MITControl type Representative approach Example materialsTemperature control Heating/cooling Materials in Figure 1a
Bandwidth control Pressure tuning RNiO3 (R = Pr, Nd, and Sm)Band-filling control Doping with acceptors/donors YBa2Cu3O7−y and La1−xSrxMnO3
MIT. A typical example is stoichiometry change in layered perovskites that can induce changesbetween two and three dimensions (8), consequently altering the material’s electronic properties.
However, in some materials, the primary mechanism of MIT is still under debate. As shown inFigure 1b, the resistance of the VO2 thin film changes by approximately four orders of magnitudenear 341–344 K, simultaneously accompanied with a structure change from a monoclinic (MoO2
structure) insulating phase (M1) to a tetragonal (rutile structure) metallic phase (R), as shown inFigure 3a,b (47–51). The symmetry of the monoclinic MoO2 structure belongs to the P21/c spacegroup, whereas that of the tetragonal rutile structure belongs to the P42/mnm space group. Duringthe structure change from tetragonal to monoclinic, the vanadium atoms are displaced out of theoctahedral planes and paired with each other, and the formed V-V bond is tilted with respect tothe octahedral planes in the tetragonal structure. In the structure change, the following unit vectorrelations exist: amono = 2ctetra, bmono = atetra, and cmono = atetra − ctetra. During MIT in VO2, the bandstructure also changes, as shown in Figure 3c. From the metallic phase to the insulating phaseacross the MIT, the 3d‖ band is split into two bands—a lower-energy, filled bonding 3d‖ bandand a higher-energy, empty antibonding 3d‖∗ band—and the antibonding 3dπ
∗ band is pushed toa higher energy. As a result, a bandgap Eg ∼ 0.6–0.7 eV exists in the insulating phase.
Whether the above-mentioned structure change induces MIT in VO2 or the structure changeis an accompanying phenomenon of the carrier-induced MIT determines whether VO2 is thePeierls type or the Mott-Hubbard type. Zylbersztejn & Mott (52) initially pointed out in the1970s that MIT in VO2 may not be the simple Mott-Hubbard type on the basis of band-splittingalignment analyses. In the 1990s, Wentzcovitch et al. (53, 54) reopened the discussion on thenature of MIT in VO2. They claimed that VO2 may be more likely to be a band insulator than aMott-Hubbard insulator on the basis of local density approximation calculations on the monoclinicM1 structure, which results in a semimetal with a very small number of carriers (53, 54). Rice et al.(55) commented that Wentzcovitch et al. did not consider the other insulating phase (M2) ofVO2, that the M2 phase is the Mott-Hubbard type, and that the M1 phase is a superposition oftwo lattice distortions of the M2 type. Recently, Cavalleri et al. (56) used ultrafast spectroscopyto investigate the structural and electronic effects in VO2 in the time domain and observed thatMIT is delayed with respect to hole injection (which was used to initiate the formation of themetallic phase). Cavalleri et al. suggest that MIT in VO2 is not the Mott-Hubbard type. Kim et al.(57) used femtosecond pump-probe measurements and observed that the rutile metal phase ofVO2 does not form simultaneously with MIT in VO2; instead, a monoclinic and correlated metal
342 Yang · Ko · Ramanathan
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
b
c
Ef
3d||*3dπ*
3dπ*
3d||
3d||
bm
am
cm
br
cr
ar
a
b c+
VanadiumOxygen
a
c+
b
Eg ~ 0.6–0.7 eV
a
Figure 3Structure and band diagram of VO2. (a,b) Structure change of VO2 from the monoclinic insulating phase(M1) to the tetragonal rutile metallic phase (R) during MIT in (a) a three-dimensional view and (b) a cross-sectional view. (c) Band structure change of VO2 across the MIT. The left and right panels show the bandstructures for the insulating and metallic phases, respectively. Panel a adapted with permission fromReference 51; panels b and c adapted with permission from Reference 50.
phase exists between MIT and the structural phase transition. Kim et al. interpreted this findingas possible evidence for a Mott-Hubbard transition.
The phase transition in VO2 happens at ultrafast timescales. Table 2 summarizes experi-mentally measured phase transition time constants in VO2 using optical pump-probe, terahertzspectroscopy, time-resolved X-ray diffraction, four-dimensional ultrafast electron microscopy,and pulsed voltage measurements. The timescale of the phase transition is typically picosecondor faster, with the exception of pulsed voltage measurements that may well be limited by theresolution of the instrumentation. That the phase transition can be triggered at subpicosecondtimescales naturally opens up the intriguing possibility of creating an ultrafast switch! If one cantrigger the transition electrically, then there is indeed a compelling case to consider for high-performance logic. The hysteresis usually accompanying the structural transition qualitativelyleads to interesting memory device candidates. Looking deeper, one could exploit the dynamics ofthe transition to create artificial structured materials with nanoscale metallic and dielectric states
www.annualreviews.org • Oxide Electronics Utilizing Ultrafast Metal-Insulator Transitions 343
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
Tab
le2
Exp
erim
enta
llym
easu
red
swit
chin
gti
me
for
the
ultr
afas
tm
etal
-ins
ulat
ortr
ansi
tion
inV
O2
a
Met
hod
Ref
eren
ceV
O2
form
Pha
setr
ansi
tion
trig
geri
ng(p
ump
sour
ce)
Pha
setr
ansi
tion
tim
eC
rite
ria
for
dete
rmin
ing
phas
etr
ansi
tion
Opt
ical
pum
p-pr
obe
met
hods
127
Pol
ycry
stal
line
thin
film
sgr
own
bysp
utte
ring
780-
nmT
i:sap
phir
ela
ser
∼5ps
Ref
ract
ive
inde
xch
ange
sto
ast
eady
stat
ede
rive
dfr
omre
flect
ion/
tran
smis
sion
spec
tra
151
Cry
stal
line
thin
film
son
glas
ssu
bstr
ate
800-
nm50
-fs
optic
alpu
lse
100
fsto
50ps
Refl
ectiv
ityof
the
prob
ech
ange
(pha
setr
ansi
tion
time
isde
pend
ento
npr
obe
exci
tatio
npo
wer
)
56T
hin
film
son
Siw
ithSi
3N4
buffe
rla
yer
15-f
sop
tical
puls
e75
to45
4fs
Refl
ectiv
ityof
the
prob
ech
ange
(pha
setr
ansi
tion
time
isde
pend
ento
npr
obe
puls
edu
ratio
n)
133
Thi
nfil
ms
grow
non
sapp
hire
,MgO
,and
glas
ssu
bstr
ates
byP
LD
400-
nm13
0-fs
50-H
zT
i:sap
phir
ela
ser
puls
e10
psto
1ns
Refl
ectiv
ityof
the
prob
ech
ange
s(p
hase
tran
sitio
ntim
eis
depe
nden
ton
VO
2th
infil
mth
ickn
ess
and
prob
eex
cita
tion
pow
er)
Ter
aher
tzpu
mp-
prob
em
etho
ds15
2P
olyc
ryst
allin
eth
infil
mso
ndi
amon
dw
indo
w1.
55-e
V12
-fs
800-
kHz
Ti:s
apph
ire
lase
rpu
lse
130
fsT
empo
rald
elay
time
for
the
first
cycl
eof
elec
tron
icco
nduc
tivity
chan
ges
∼10
psSa
tura
tion
ofco
nduc
tivity
chan
ges
128
Sing
lecr
ysta
lsan
dth
infil
ms
grow
non
sapp
hire
byP
LD
800-
nm50
-fs/
150-
fs5-
kHz/
1-kH
zT
i:sap
phir
ela
ser
puls
e
0.7
psT
rans
mitt
ance
abru
ptly
chan
ges
100
psSa
tura
tion
oftr
ansm
ittan
cech
ange
s
Tim
e-re
solv
edX
-ray
diffr
actio
n(la
ser
pum
p+
X-r
aypr
obe)
151
Cry
stal
line
thin
film
son
glas
ssu
bstr
ate
800-
nm50
-fs
optic
alpu
lse
∼1ps
Rut
ile-p
hase
(110
)pea
kat
13.7
8◦
chan
ges
tom
onoc
linic
-pha
se(0
11)
peak
at13
.9◦
inX
-ray
diffr
actio
n
57T
hin
film
sgr
own
onsa
pphi
reby
sol-
gelm
etho
d78
0-nm
20-f
s92
-MH
zT
i:sap
phir
ela
ser
puls
e∼3
00fs
Met
al-p
hase
X-r
aydi
ffrac
tion
(110
)pe
akat
13.8
◦ap
pear
saf
ter
X-r
ayir
radi
atio
n
153
Hig
hly
orie
nted
thin
film
sgr
own
onsa
pphi
reby
PL
D
800-
nmfe
mto
seco
ndop
tical
puls
e25
psM
onoc
linic
-pha
se(0
20)p
eak
at39
.86◦
chan
ges
toru
tile-
phas
e(0
20)
peak
at39
.77◦
inth
eX
-ray
diffr
actio
npa
tter
n
∼100
psSt
ruct
ural
chan
ge–i
nduc
edat
omvi
brat
ions
abat
e
344 Yang · Ko · Ramanathan
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
Four
-dim
ensi
onal
ultr
afas
tele
ctro
nm
icro
scop
y(la
ser
pum
p+
elec
tron
prob
e)
50P
olyc
ryst
allin
eth
infil
ms
prep
ared
onm
ica
subs
trat
eby
sol-
gelm
etho
d
776-
nmfe
mto
seco
ndla
ser
puls
e3.
1±
0.1
psA
mpl
itude
of(3
01̄)-
Mel
ectr
ondi
ffrac
tion
peak
chan
ges
154
Pol
ycry
stal
line
thin
film
spr
epar
edon
mic
asu
bstr
ate
byso
l-ge
lmet
hod
776-
nmfe
mto
seco
ndla
ser
puls
e3
±1
psIn
tegr
ated
inte
nsity
of(3
02̄)-
Mel
ectr
ondi
ffrac
tion
peak
chan
ges
51Si
ngle
crys
tals
800-
nmfe
mto
seco
ndla
ser
puls
eO
nor
der
offs
Van
adiu
m-v
anad
ium
bond
dila
tion
isin
dica
ted
byel
ectr
ondi
ffrac
tion
On
orde
rof
psD
ispl
acem
ento
fvan
adiu
mat
oms
isin
dica
ted
byel
ectr
ondi
ffrac
tion
On
orde
rof
100
psSo
und
wav
esh
ear
mot
ion
isin
dica
ted
byel
ectr
ondi
ffrac
tion
Pul
sed
volta
gem
etho
d10
8O
rien
ted
thin
film
sgr
own
onsa
pphi
reby
PL
DV
olta
gepu
lse
∼9ns
Abr
uptv
olta
gech
ange
acro
ssV
O2
isin
dica
ted
bydi
gita
losc
illos
cope
a PL
Dde
note
spu
lsed
lase
rde
posi
tion.
www.annualreviews.org • Oxide Electronics Utilizing Ultrafast Metal-Insulator Transitions 345
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
that could be electrically tuned. The rest of this review discusses these possibilities and snapshotsof representative research being carried out regarding this exciting opportunity. High-qualityfilms with reproducible transition characteristics for probing the fundamental physics as well asexploratory devices are among what are needed to experimentally investigate these ideas. Thischallenge is discussed next.
3. CORRELATED OXIDE THIN-FILM SYNTHESIS: THE CATIONVALENCE STATE PROBLEM AND OXYGEN VACANCIES
The general problems concerning oxide syntheses in thin-film form center around epitaxial orhighly crystalline stoichiometric films on suitable substrates with controlled defect concentrations.The structure is optimized by choice of lattice-matched substrates, deposition temperature, anddeposition method, whereas stoichiometry is often adjusted through precursor deposition rates,background pressure, and postdeposition anneals. Given that oxides form in ionic lattices anddefects are often charged, nonstoichiometry can have severe consequences for functional prop-erties and, in some cases, leakage currents that significantly degrade the device performance orcomplicate interpretation of electrical measurements. The transition-metal cations can take onmultiple-valence states in such oxides and lead to challenging synthesis problems. For example,the multiple-valence states of vanadium ions lead to several vanadium oxide phases with diverseV/O ratios (58–60), such as Magneli phases (VnO2n−1) (58–60) and Wadsley phases (V2nO5n−2)(59, 61). Figure 4a shows the phase diagram of the vanadium oxide system (60). VO2 has beensynthesized in various forms as bulk single crystal (62); thin films in both epitaxial (63–65) andpolycrystalline (66–73) forms; and nanostructures, such as nanoparticles (74–76), nanobeams (27,31, 77–79), nanowires (35, 80–84), and nanobelts (85). Different synthesis methods have been em-ployed for thin-film VO2 growth, including PLD (65, 86, 87), CVD (88, 89), sputtering (66–73),sol-gel coating (90, 91), electron beam (e-beam) evaporation (92, 93), and ion beam deposition(94, 95). Panels b (69) and c (66) of Figure 4 show how sensitive VO2 thin-film growth is tooxygen flow rates and growth temperatures, respectively, through use of the dc-magnetron sput-tering technique. The growth windows for VO2 films in terms of oxygen flow rate and growthtemperature are approximately 7.8∼8.8% and ∼550◦C. Only the VO2 films grown within theoptimized growth window show a more-than-three-order-of-magnitude resistance change acrossthe MIT, indicating good stoichiometry of the film because studies of polycrystalline VO2 thinfilms show that the resistance change across MIT and MIT temperature are sensitive to the degreeof nonstoichiometry (96–98). Hence, growing high-quality homogeneous VO2 films without for-mation of nonstoichiometric phases is a challenge (96). Different substrates have been employedfor thin-film VO2 growth; such substrates include glass (88), Si (28, 73), Ge (99), Al2O3 (66, 68,70–72), TiO2 (63, 64, 87), amorphous SiO2-coated Si (29, 66, 67, 69), HfO2-coated Si (100), andyttria-stabilized zirconia (YSZ) buffer layers on Si (65). The choice of substrate can significantlyinfluence MIT characteristics, including the TMIT, in part by affecting the strain state.
4. SOLID-STATE-DEVICE CONCEPTS
4.1. The Ultrafast Oxide Metal-Insulator Transition Switch
As discussed in the introduction, a central theme in devices exploring MIT is an ultrafast switch.The high-resistance, insulating state and the low-resistance, conducting state on both sides of MITdefine the OFF and ON states of the switch, as shown in Figure 1c. The switching behavior, i.e.,changes between the OFF and ON states due to the phase transition, can be triggered by an external
346 Yang · Ko · Ramanathan
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
30 50 70 11090
Re
sist
an
ce (
kΩ
)
Temperature (°C)
10–1
100
101
102
103
b Oxygen flow effect
10p88p8
7p8 9p8
6p8
30 50 70 90
10–2
10–1
100
10–3
VO2 temperature (°C)
R/R m
ax
c Growth temperature effect
Sput. T = 350°C
Sput. T = 450°C
Sput. T = 550°C
1,200
1,000
800
600
400
200
01.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5
x in VOx
Tem
pe
ratu
re (
°C)
VO VOx
V2O3+
V3O5
V3O5
V 4O7
V 3O5
+ V 4O
7
V 5O9
+ V 6O
11
V 5O9
V6O11
V 4O7
+ V 5O
9V 5O
9 +
V 6O11
V7O13+
VO2
V6O13 V2O5
V3O7
VO2+
V6O13
V6O13+
V3O7
680°C 676°C
708°CV6O13 + liquid
VO2
VO2 + liquid
Liquid
a Vanadium oxides phase diagram
V3O7+
V2O5
V2O3 V7O13
Figure 4Growth optimization of VO2. (a) Phase diagram of the VOx system. (b) Oxygen flow effect on VO2 growth.(c) Growth temperature effect on VO2 growth. Panel a adapted with permission from Reference 60; panel badapted with permission from Reference 69; panel c adapted with permission from Reference 66.
perturbation (that is thermal, electrical, optical, or magnetic) or by removal of the excitation. Insome cases, combination of two or more excitations can also efficiently lead to the transition.Depending on the mechanism leading to the transition, the OFF and ON transitions can havedifferent time constants. Hence, in a device, the slower time constant will dominate the speed of theswitch. The energy needed for transition onset corresponds to the switching energy. Additionally,latent heat changes may need to be accounted for in thermodynamic analyses, depending on the
www.annualreviews.org • Oxide Electronics Utilizing Ultrafast Metal-Insulator Transitions 347
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
E-MIT: electricallytriggered MIT
I-V: current-voltage
MottFET: Mott fieldeffect transistor
material system and transition mechanism. The hysteresis during the transition is yet anothervariable that could be exploited for devices such as memory. In some cases, multiple steps duringthe transition (such as that in domain switching) lead to additional possibilities such as the creationof nonlinear resistor circuits. The following subsections discuss emerging phase transition–basedelectronic devices, optical devices, and thermal/chemical sensors.
4.2. Phase Transition Electronic Devices
The ability to trigger the phase transition electrically at or near room temperature in two- orthree-terminal device configurations creates the potential for novel, low-power oxide electronics.The phenomena of resistive switching behavior in VO2 can be traced back to the early 1970s, withnonlinear I-V characteristics, discontinuity of resistance change with applied voltage, and electricoscillations reported (101–103). Furthermore, several elementary switch devices were demon-strated on the basis of nonlinear I-V characteristics in VO2. However, this nonlinear resistiveswitching behavior in VO2 was perhaps not associated with electrically triggered MIT (E-MIT)until more recently. In 2000, Stefanovich et al. (104) reported that electric field or electron in-jection into VO2 can trigger MIT. Although investigators have argued that the current-inducedheating effect may also induce MIT, theoretical simulations indicate that the Joule heating effectfrom leakage current is likely insufficient to trigger MIT in the case of homogeneous currentflow (105). Recently, a variety of groups have studied E-MIT (28, 104, 106–111). A field-drivenphase transition may lead to a Mott field effect transistor (MottFET) (29, 110) and further pro-vide valuable insights into the physics of the transition. In a MottFET, a Mott insulator is em-ployed as the channel, and the external gate voltage switches the channel between the insulatingOFF state and the metallic ON state (112–115). Qualitatively, this approach offers an alternatethree-terminal switch concept to a conventional semiconductor FET. Due to the much higherelectron carrier concentration possible in a metallic channel, this approach may have additionaladvantages.
4.2.1. Two-terminal switches utilizing electrically triggered metal-insulator transition.Figure 5a shows the structure of a planar VO2 electronic switch device employed by Stefanovichet al. (104) for E-MIT studies. The VO2 channel length, width, and thickness are 0.2 mm, 5 ∼10 μm, and 0.1 ∼ 1 μm, respectively. The thickness of the SiO2 layer is ∼70 nm on the p-Sisubstrate (0.1 � cm). Al contacts were deposited via thermal evaporation. The I-V characteristicsof the VO2 channel are S shaped, as shown in Figure 5b, indicating that MIT occurred at acertain threshold voltage Vth ∼ 2 V at room temperature. Experiments showed that Vth decreasesas temperature increases. When the temperature is above TMIT, the E-MIT effect no longer occurs.Figure 5c shows the I-V characteristics of a planar VO2 switch device at various temperatures:(1) 293 K, (2) 241 K, (3) 211 K, (4) 144 K, (5) 91 K, and (6) 15 K (109). Vth increases as temperaturedecreases. The increased Vth was attributed to the decreased VO2 channel conductivity at lowertemperatures. Figure 5d shows the cross-sectional schematic of a two-terminal VO2 switch devicereported recently with improved quality of the VO2 thin films (108). VO2 thin films were depositedon r-plane sapphire substrates using PLD, and Au/Cr electrodes were deposited and patternedusing a lift-off process. Figure 5e shows the I-V characteristics of the VO2 switch device shownin Figure 5d with a load resistance at room temperature (108). An evident MIT behavior in VO2
thin films was triggered by a dc voltage, with a critical voltage VMIT of ∼7 V. The current jumpat VMIT ∼ 7 V indicates MIT from the insulator state and metal in the VO2 thin film. The sameresearch group also proposed that this device is a varistor (variable resistor) toward electrostaticdischarge protection (111).
348 Yang · Ko · Ramanathan
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
7
6
5
4
0
–1
–2
–3
–4
–5
0 1 2 3 4 5 6 7 8 9 10
0 5 10 15 20 25 30
0 1 2 3 4 50
2 4 6 8 10 12
2.0
1.5
1.0
0.5
0
160
140
120
100
80
60
40
20
80
40
1 2
Electric field (106 V m–1)
2.5
2.0
1.5
1.0
0.5
0.0
25°C40°C55°C70°C75°C85°C100°C
E-MITT-MIT
Substrate: Al2O3 (1012)
VO2
3 μm
Au/Cr Au/Cr
Cu
rre
nt
(μA
)
Voltage (V)
VoltageDS (V)
Cu
rre
nt D
S (m
A)
Cu
rre
nt
(μA
) Cu
rren
t (μA
)
Voltage (V)
JD
S (10
5 A cm
–2)
Lo
g I
(A)
Voltage (V)
VO2 1 kΩ
12
3
4 5 6
6'
5'
b
c
a e
f
d
Metal
InsulatorVMIT = 7.1 V
Vth
3
2
1
Al contacts VO2 channel
V2O5δ = 0.1–1 μm
Si
(0.2 mm)
(5–10μm)
SiO2δ = 700 Å
Figure 5Two-terminal electronic switch devices utilizing MIT in VO2. (a) Schematics of the planar structure of VO2 switch devices. (b) I-Vcharacteristics of a VO2 switch device at room temperature, showing an S shape with a threshold voltage Vth of ∼2 V for electricallytriggered MIT (E-MIT). (c) I-V characteristics of a planar VO2 switch device at various temperatures. Vth decreases with temperature.The dashed lines indicate unstable (transient) regions of the I-V curves. The right-hand axis shows the low-conductivity regions ofcurves 5 and 6 in an extended scale denoted as 5′ and 6′. (d ) A cross-sectional view of a two-terminal planar VO2 switch. (e) I-Vcharacteristics of the VO2 switch device shown in panel d with a load resistance at room temperature, showing that a dc voltage (with acritical MIT voltage of ∼7 V) triggers MIT in VO2 thin films. The inset shows the circuit diagram. The left-hand and right-hand yaxes show current and current density, respectively. ( f ) I-V characteristics of a vertical VO2 switch device at different temperatures.E-MIT is observed at 1.5 ∼ 2 V at temperatures below the thermally triggered MIT (T-MIT). Panels a and b adapted with permissionfrom Reference 104; panel c adapted with permission from Reference 109; panels d and e adapted with permission from Reference 108;panel f adapted with permission from Reference 107.
The two-terminal VO2 devices discussed in Figure 5a–e are planar junctions. E-MIT wasalso recently observed in out-of-plane VO2 devices. Figure 5f shows I-V characteristics of a VO2
capacitor at different temperatures (107). The VO2 thin films were grown on n+-Si substrates byrf-magnetron sputtering, and Pd metal was deposited as the top contacts using e-beam evaporationwith a shadow mask (107). The electrical measurements were carried out using top Pd and bottomn+-Si contacts. An abrupt change in leakage current is seen with both increasing T and E. E-MITcan be seen up to 55◦C. VMIT for E-MIT is slightly lower than 2 V at 25◦C, corresponding to
www.annualreviews.org • Oxide Electronics Utilizing Ultrafast Metal-Insulator Transitions 349
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
an electric field of ∼107 V m−1 (consistent with the typical fields observed in planar devices).The I-V curves present an ohmic relationship in the insulating phase. At larger voltages, bothlow-temperature and high-temperature states show nearly identical resistance, further suggestingthe validity of E-MIT.
Whether E-MIT persists when VO2 reaches the nanoscale is a key issue for scaled devices.Ruzmetov et al. (28) performed experiments in similar VO2 capacitor devices, but with muchsmaller top contacts. In these devices, the top contacts were ∼200-nm Au dots fabricated by e-beamlithography. The electron transport measurements were performed with a conducting atomicforce microscopy (AFM) tip. E-MIT phenomena were observed in these nanoscale VO2 capacitordevices in similar voltage ranges. Sakai & Kurisu (116) studied the effect of hydrostatic pressureon E-MIT in a two-terminal planar VO2 device. These investigators observed that the criticalcurrent to trigger MIT increases with pressure, whereas voltage at the transition is independentof pressure, which further suggests that MIT is triggered electrically, i.e., not by Joule heatingalone.
Figure 6 demonstrates controlled local phase switching in VO2 thin films through use of abiased conducting AFM tip (106). Figure 6a shows the geometry and Figure 6b shows the heightprofile in an AFM line scan. Figure 6c shows several consecutive I-V sweeps at 107 K. Panels d ande of Figure 6 show 200 consecutive I-V sweeps at two different representative locations at 293 K.Figure 6f shows an AFM image depicting the surface morphology of the VO2 thin film. Panelsg–k of Figure 6 show current mapping at different bias voltages. The metallic puddle is initiallyseeded at two grains (Figure 6f–h) at small bias. The number of grains grows with increasing biasvoltage. Single-grain switching with reproducible hysteresis demonstrates the existence of E-MITin VO2 down to the few-nanometer range. These preliminary results on voltage triggering of MITdown to the nanoscale are encouraging for further explorations of integrated devices and circuitcomponents.
4.2.2. Three-terminal gated field effect switches. In a three-terminal VO2 FET-like device,an electric field introduced by applied external gate voltage (separated from the channel by aninsulator) can modulate the resistance of a VO2 channel. The conductivity of the VO2 channelmay be switched between the high-resistivity insulating state and the low-resistivity metal state.The device behaves as a MottFET (112–115) with a VO2 channel. In other words, three-terminalelectronic switches utilizing E-MIT may potentially demonstrate functional MottFET for logicoperations in a CMOS-like architecture. Additionally, a three-terminal VO2 FET device structurefacilitates fundamental studies to clarify E-MIT or Joule heating–induced MIT in VO2 becausethe possibility of a current-induced heating effect can be further minimized in a three-terminalVO2 device compared with a two-terminal VO2 device. This subsection discusses results on three-terminal FET-like VO2 devices, although systematic studies on this important problem are stillin the early stages.
Figure 7a shows an early proposed three-terminal VO2 switch device by Chudnovskiy et al.(117). The device is composed of a VO2 channel, a gate dielectric layer on top of the VO2 channel,a gate electrode on top of the gate dielectric layer, and source and drain electrodes at both ends ofthe channel. The device operating temperature is set above TMIT, and the channel is in the metallicphase in the absence of a gate voltage, although the channel goes to the dielectric state once a gatevoltage is applied. This is similar to a MOSFET working in the depletion mode instead of theenhancement mode, which is generally employed. Panels b and c of Figure 7 show an experimentalexample of a three-terminal VO2 device by Kim et al. (110). The inset in Figure 7b shows thedevice’s structure. The VO2 thin films were deposited on SiO2/Si (and sapphire) substrates using
350 Yang · Ko · Ramanathan
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
40 nm
0
4.44 V 4.93 V 5.43 V
600
500
400
300
200
100
00 2 4 6 8 100 2 4 6 8 10
200
300
250
200
150
100
50
14
Bias (V)Bias (V)
Cu
rre
nt
(μA
)
Cu
rre
nt
(μA
)
c T = 106.75 K
e T = 293 K d T = 293 K
40
20
0He
igh
t(n
m)
0 0.5 1 1.5 2Position (μm)
Cr/Au-coated tip
VO2 film(~200 nm)
n-type Si
V
SiOx nativeoxide (3–4 nm)
b
a
f g h
i j k
3.45 V 3.95 V
Figure 6AFM studies ofelectrically triggeredMIT in VO2.(a) Schematic of theAFM tip and thesample geometry.(b) AFM line scan.(c) Fourteenconsecutive I-V sweepsat T = 106.75 K.(d,e) Two hundredconsecutive I-V sweepsat two differentrepresentativelocations at T =293 K. The color scalebar shows the times ofsweep. ( f ) AFM imageshowing the surfacemorphology of theVO2 thin film.( g–k) Currentmapping at differentbias voltages. Grainboundaries are drawnin white. Adapted withpermission fromReference 106.
www.annualreviews.org • Oxide Electronics Utilizing Ultrafast Metal-Insulator Transitions 351
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
240 µm
0 1 2 3 4 5 2.8 3.2 3.6 4.00
0.5
1
Motttransition
J DS
(A c
m–
2)
J (1
05 A
cm
–2)
Electric field (MV m–1)Electric fieldDS (MV m–1)
240 µm
b
a
10–1
102
106
101
104
100
105
103
d
c
50
0
–50
–100
–150
–0.5 0.0 0.5
Gate 2 voltage (V)
Re
sist
an
ce c
ha
ng
e, Δ
R (Ω
)
Gate 1
S D
Gate 2
SiO2
VO2VO2
Al2O3
e
Gate open
Electrode: Au/Cr
1st
2nd
3rd
4th
5th
Vgate = 0 V
Vgate = –10 V
Vgate = –20 V
Vgate = –40 V
Vgate = –60 V
Vgate = –70 V
GateV+ V–
I+ I–
Au
SiO2VO2
Gate
Dielectric Layer Dielectric layer
VO2
Gate
DS
VO2
Source Drain
Insulator SiO2
Gate: WSi
Substrate: Si
Figure 7Three-terminal gated electronic switch devices utilizing MIT in VO2. (a) An early proposed three-terminalgated VO2 switch device. (b) Current density versus electric field applied on the VO2 channel with the gateopen at room temperature. The inset shows the three-terminal VO2 device structure. (c) Current densityversus electric field applied on the VO2 channel with applied negative gate voltages at room temperature.(d ) Optical microscopy image of the top view of a three-terminal VO2 device. (e) The effect of the back-gate(gate 2) voltage on the source-drain resistance of the VO2 channel of a three-terminal VO2 device withdouble-gate structure from the top view as shown in panel d. The inset shows a cross-sectional schematic ofthe device. Panels b and c adapted with permission from Reference 110; panels d and e adapted withpermission from Reference 29.
PLD at 400◦C with gas flows of Ar + 10% O2 under 55 ∼ 60-mTorr growth pressure. TheVO2 thin films are ∼90 nm thick. The VO2 thin film (on sapphire) shows an MIT transitionat ∼340 K with a change in resistance of approximately four orders across the MIT. The SiO2
underneath the VO2 thin film was used for the gate oxide, and Au/Cr and WSi were used for thesource-drain and gate contacts of the three-terminal device. The VO2 channel length and widthare 5 and 25 μm, respectively. Figure 7b shows the source-drain current density versus electricfield applied on the VO2 channel with the gate open at room temperature. The device is turnedon and off at effective electric fields of ∼4.8 and ∼2.8 MV m−1, respectively. Figure 7c showsthe source-drain current density as a function of electric field under different applied negativegate voltages. When the magnitude of the negative gate biases increases, the turn-on voltage (for
352 Yang · Ko · Ramanathan
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
triggering MIT) increases, which Kim et al. (110) attributed to the decrease in conductivity dueto increased hole carrier density.
Panels d and e of Figure 7 present a recent study on three-terminal VO2 devices (29). Theinset in Figure 7e shows the cross-sectional schematic of the device. A uniform Al2O3 dielectriclayer was grown on a high -n+-Si substrate using atomic layer deposition. The n+-Si substrate wasused as the back gate of the device (shown as gate 2 in Figure 7e). A VO2 thin film was grown ontop of the Al2O3 dielectric layer using magnetron sputtering. The device channel was patternedusing photolithography. The top gate (shown in Figure 7e), source, and drain electrodes weredeposited using e-beam evaporation and the lift-off process. A SiO2 layer was used as the top-gatedielectric layer. Figure 7d shows a planar view optical microscopy image of the three-terminalVO2 device. Figure 7e shows how the back-gate (gate 2) voltage affects the source-drain resistanceof the VO2 channel. A systematic resistance decrease occurs when a negative gate voltage of−0.5 V is applied, whereas no evident resistance change is observed when a positive gate voltageup to 0.5 V is applied. The authors (29) noted that the magnitude of the applied voltages wasrestricted to 0.5 V due to the gate oxide leakage issue. At −0.5 V gate voltage, the decrease inVO2 channel resistance is ∼120 �. The results discussed in Figure 7e are significant in that thereversible modulation of the VO2 channel resistance by an external gate voltage was demonstratedin an all-electrical three-terminal VO2 device configuration. Future research directions in this topicinclude low-leakage gate dielectric fabrication to increase the gate voltage that could be applied toinduce a larger field in the channel or ionic electrolyte gating (118) to achieve larger critical carrierdensities in the channel as well as thinner conformal VO2 layers to increase the field penetrationinto most of the channel region.
4.2.3. Metal-insulator transition oscillators. Oscillators are used in electronic circuits in whicha periodic repetitive signal is needed, such as signal-broadcasting transmitters and clocks. Oscil-lators can be made of different components, such as piezoelectric materials (e.g., quartz), resistor-capacitor/inductor (RC/RL) circuits, transistor circuits, and Gunn diodes. In this subsection, wediscuss oscillators utilizing E-MIT in VO2 materials. Oscillation phenomena in VO2 have beenexplored since three decades ago (102, 103). Figure 8 shows an example of an oscillator deviceutilizing E-MIT in VO2 at room temperature (119). A VO2 thin film of ∼100 nm was grown on asapphire substrate using the sol-gel technique (90). Ni and Au were deposited on the lithographi-cally patterned VO2 film as contact electrodes using sputtering. The cross-sectional view andplanar view of the device structure are shown in the insets of Figure 8a. The surface morphologyof the VO2 film is shown in the right inset of Figure 8a, with a root-mean-square roughness of∼5 nm using atomic force microscopy. Figure 8a shows the I-V characteristics of a VO2 devicewith both a voltage sweeping mode (V mode) and a current sweeping mode (I mode). In V mode,the I-V curves show two threshold voltages of Vt1 ∼ 9.7 V and Vt2 ∼ 4 V, at which the currentincreases and decreases abruptly. The current readings at Vt1 and Vt2 are It1 ∼ 0.43 mA and It2
∼ 1.34 mA, respectively. In I mode, as the current flowing through the device (IO) increases, thevoltage across the device (VM ) increases exponentially until it reaches Vt1. Once VM exceeds Vt1,however, it starts to decrease until it reaches Vt2 (as shown with a negative differential resistance).Then VM stays at Vt2 regardless of the current increase. Figure 8b shows the oscillator circuit,which is a single-loop circuit composed of a two-terminal VO2 device (as shown in Figure 8a), astandard resistor REXT = 7 k�, and a voltage source VS. When a rectangular voltage pulse (witha dc offset of 10 V) was supplied from the voltage source, as shown in Figure 8d, the IO and VM
were recorded with an oscilloscope (Figure 8d). Periodic oscillations are clearly observed. Theamplitude of IO and VM and oscillation frequency fO are ∼0.8 mA, ∼5.7 V, and ∼0.317 MHz,respectively.
www.annualreviews.org • Oxide Electronics Utilizing Ultrafast Metal-Insulator Transitions 353
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
3
2
1
0
0 2 4 6 8 10
3
2
1
0
1.2
0.8
0.4
0.02520151050
0
4
8
12
Time (µs)
Vo
lta
ge
(V
)
Cu
rren
t (mA
)
Voltage (V)
Cu
rre
nt
(mA
)
Cu
rren
t de
nsity
(10
5 A cm
–2)
Au Au
Au
Au
Ni Ni
Ni
Ni
W
L
(nm)
90
80
70100 nm
P1(Vt1, It1)
P2(Vt2, It2)
S
Voltage source
Ground
VR
VM
IO
VS
VO2 device(RM, CM)
REXT
7
VS
VM
IO
VO2
Al2O3
Al2O3
VO2
a
b
d
(i)
(ii)
(iii)
V-mode up
V-mode down
I-mode up
I-mode down
Time (µs)
1.2
0.8
0.43 4 5 6
I O (
mA
)
c
MIT
Chargingprocess
P1
P3
354 Yang · Ko · Ramanathan
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
We next briefly discuss the mechanism of oscillations utilizing E-MIT of VO2 in the circuit ofFigure 8b. Initially, with insufficient external voltage, VM (VM = VS − REXTIO) is smaller thanVt1, and the VO2 film is in an insulating state with a high resistance. Once VM reaches Vt1 inducedby a sufficient external voltage supply VS, the VO2 film enters the low-resistance metallic state dueto E-MIT, accompanied with an abrupt IO jump. Simultaneously, VM decreases while VR (VR =VS − VM ) increases abruptly because RM becomes orders of magnitude smaller due to MIT. OnceIO approaches (VS/REXT), VM becomes small enough such that the VO2 film changes from themetallic state to the insulating state. Hence, VM increases whereas VR decreases accordingly dueto the increased RM . When VM increases to Vt1, the circuit enters the starting point of the nextoscillation. This is a complete period of oscillation, which is comparable in analog to a simple seriesRC circuit. Figure 8c shows a single oscillation period of the circuit. With appropriate tuning ofREXT, different oscillation frequencies can be designed (119, 120). Furthermore, MIT-inducedoscillations have also been demonstrated in nanoscale VO2-based materials, such as W-dopedVO2 nanobeams (31). These results pave the way to explore an oscillator circuit utilizing E-MIT.
4.2.4. Memristive devices utilizing metal-insulator transition. The idea of fabricating logicor memory devices from two-terminal device configurations has seen a renaissance with memris-tors (121, 122) and in a broader context may enable neuromorphic circuits (123). Correlated oxidecrossbar arrays may be interesting candidates for circuit elements, and the ability of these devicesto trigger phase transition at room temperature by a voltage makes such technology particularlyattractive. This subsection discusses an example of a VO2 device for memristors and memorycapacitors. A memristor is a two-terminal device whose resistance is not determined by the in-stantaneous applied voltage (current) but rather is reliant on the dynamic history of the system’scharge flow. Figure 9a shows a schematic of a VO2 memristor device (33). The VO2 material,grown on a sapphire substrate using the sol-gel technique (91), shows an up-to-four-order-of-magnitude resistivity change across MIT. The measurement temperature is set at 340 K, close tothe onset of MIT in VO2. At this temperature, the resistance of the VO2 film is a sensitive andhighly hysteretic function of temperature. The I-V characteristics of the device were investigatedduring a ramped voltage pulse (50 V in 5 s). Figure 9b shows three measured I-V curves (33). Twotypical memristive characteristics are evident in the I-V curves. First, the I-V curves are nonlinear,
←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−Figure 8Oscillator device utilizing electrically triggered MIT in VO2. (a) I-V characteristics of a two-terminal VO2device. The left insets show the cross-sectional and planar views of the device. The right inset shows an AFMimage of VO2 showing the surface morphology and yielding a root-mean-square roughness of ∼5 nm. Panela shows the I-V characteristics of a VO2 device with both a voltage sweeping mode (V mode) and a currentsweeping mode (I mode). In V mode, the I-V curves show two threshold voltages of Vt1 ∼ 9.7 V and Vt2 ∼4 V, at which the current increases and decreases abruptly. The current readings at Vt1 and Vt2 are It1 ∼0.43 mA and It2 ∼ 1.34 mA, respectively. In I mode, as the current flowing through the device (IO) increases,the voltage across the device (VM ) increases exponentially until it reaches Vt1 (as shown in region i ). OnceVM exceeds Vt1, however, it starts to decrease until it reaches Vt2 (as shown in region ii with a negativedifferential resistance). Then VM stays at Vt2 regardless of the current increase (as shown in region iii ).(b) Circuit diagram for generating the electrically triggered MIT oscillations. The oscillator circuit is asingle-loop circuit composed of a two-terminal VO2 device (as shown in panel a), a standard resistorREXT = 7 k�, and a voltage source VS. (c) One period of oscillation. (d ) Oscillatory electric responses forthe VO2 device at room temperature. When a rectangular voltage pulse (with a dc offset of 10 V) wassupplied from the voltage source (VS) (black curve), IO (red circles) and VM (blue squares) were recorded with anoscilloscope. Periodic oscillations are clearly observed. The amplitude of IO and VM and oscillation frequencyfO are ∼0.8 mA, ∼5.7 V, and ∼0.317 MHz, respectively. Adapted with permission from Reference 119.
www.annualreviews.org • Oxide Electronics Utilizing Ultrafast Metal-Insulator Transitions 355
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
0.6
0.5
0.4
0.3
0.2
0.1
0.0
1.020
1.015
1.010
1.005
1.000
338.5 338.6 338.7 338.810 20 30 40 50
Cu
rre
nt
(mA
)
Voltage (V) Temperature (K)
Ca
pa
cita
nce
(C 0
+ C
m)/
C 0
b c
Pulse 1Pulse 2Pulse 3
(ii)
(i)
α
(v)
(iv)
(iii)
δ
γ
β
a VI
Al2O3VO2
Temperaturestage
T0 =
°K
Figure 9A memristive device utilizing MIT in VO2. (a) Schematic of a VO2 memristor device. (b) Three I-V curvesshowing the hysteretic nonlinear memristive behavior of the VO2 memristor device. The arrows pointing upindicate increasing voltage, and the arrows pointing down indicate decreasing voltage. (c) Memorycapacitance of a hybrid VO2 device during heating and cooling. Panels a and b adapted with permission fromReference 33; panel c adapted with permission from Reference 124.
and second, the I-V curves are hysteretic, which is associated with MIT in VO2. The hysteresiscontains the memory aspect of the memristor. Figure 9c demonstrates a memory capacitancesystem utilizing MIT in VO2 (124). The temperature perturbation–induced capacitance changesdemonstrate memory behavior (as shown in Figure 9c), which is associated with hysteretic MITbehavior of VO2.
4.3. Phase Transition Optical Devices, Thermal Sensors, and Chemical Sensors
This section briefly discusses optical devices, thermal sensors, and chemical sensors utilizing phasetransitions. The sharp change in carrier concentration in the proximity of a phase transition createsunique opportunities for novel electronics that can be monolithically integrated.
4.3.1. Optical devices utilizing metal-insulator transition. In the early 1970s, VO2 was pro-posed for optical storage due to its first-order MIT transition (125, 126). From that point, opticallytriggered MIT in VO2 (56, 127–130) has been studied periodically, and models (131) have beenproposed to explain the light-induced MIT phenomena. On the basis of these studies, opticaldetector/sensor/switch devices (132–136), microwave switch devices (137, 138), and modulatordevices (30) based on VO2 material have been proposed and demonstrated. Figure 10d shows aschematic of the experimental setup of a light pulse detector (135) with a 20-nm-thick VO2 thinfilm and 330-nm-thick Si3N4 on Si. When excited with a near-infrared laser pulse, the VO2 thinfilm undergoes photo-induced MIT. Owing to this behavior, this device works as a detector forlight pulses. The optical pump-probe setup is based on a cavity-dumped Ti:sapphire oscillator.Figure 10a shows the frequency-resolved VO2 reflectivity variation, measured as a function of thedelay between the probe pulse and the VO2 switching triggered by the pump pulse. Figure 10b
shows the time traces at different wavelengths. Figure 10c shows the frequency-resolved reflec-tivity variation. The significant changes in absorption VO2 across MIT, combined with ultrafast
356 Yang · Ko · Ramanathan
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
h
i
c
d
–6 –3 0 3
1.0
0.9
0.8
0.7
50
40
30
20
10
0
–10
1.0
0.8
0.6
0.4
0.2
0.030 45 60 75
0.8 1.2 1.6 2.0 2.4
SC pulse
Pump
Time
Prism
PDA
VO220 nm
Si3N4330 nm
Si substrate
VO2optical
response
Wa
ve
len
gth
(μ
m)
ΔR/R (%)
Frequency (cm–1)
Frequency (THz)
aa
–1.5 –1.0 –0.5 0.0 0.5
1.0
0.9
0.8
0.7
Wa
ve
len
gth
(μ
m)
Delay (ps)
300.0 K340.4 K340.8 K341.2 K341.5 K
b
3
0
–3
–6–1.5 –1.0 –0.5 0.0 0.5
ΔR/
R (%
)
Delay (ps)
1,010 nm
970 nm
890 nm 695 nm
825 nm
e e e f g
Terahertzbroadband beam Gold
VO2Al2O3
Re
al
pe
rmit
tiv
ity
Re
al
pe
rme
ab
ilit
y
Figure 10Optical and metamaterial devices utilizing MIT in VO2. (a) The time- and frequency-resolved reflectivityvariation measured on a VO2 film. (b) Time traces at different wavelengths. (c) The frequency-resolvedreflectivity variation. (d ) Schematic of time- and frequency-resolved setup. Abbreviations:SC, supercontinuum; PDA, photodiode array. (e–g) Schematics of a VO2 split-ring resonator (SRR) hybridmetamaterial. In panel e, green and blue represent the metallic phase and the insulating phase, respectively.(h,i ) Experimentally retrieved permittivity and permeability bandwidths for the hybrid metamaterials. Theshaded areas show the range of values accessible with the hybrid SRR-VO2 device. Panels a–d adapted withpermission from Reference 135; panels e–i adapted with permission from Reference 140.
www.annualreviews.org • Oxide Electronics Utilizing Ultrafast Metal-Insulator Transitions 357
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
switching, make VO2 an attractive candidate for the next generation of active integrated opticalcomponents suitable for broadband applications. Briggs et al. (30) fabricated a compact Si photonicwaveguide modulator utilizing MIT of VO2 thin films. The VO2 thin films were grown using PLD,were patterned by lithography, and were integrated with Si-on-insulator photonic waveguides todemonstrate a compact absorption modulator toward photonic circuit applications. Through useof single-mode waveguides at λ = 1.55 μm, optical modulation of the guided transverse-electricmode of more than 6.5 dB with a 2-dB insertion loss over a 2-μm active device length was demon-strated. A sharp decrease in resonator quality factor is observed above 70◦C, corresponding toVO2 entering the metallic phase across MIT with an ∼78% decrease in transmission.
4.3.2. Metal-insulator transition metamaterials. Metamaterials are artificial materials com-posed of designed inclusions, which show properties that are not readily available in nature, suchas negative refractive index (139). Researchers have recently proposed several intriguing appli-cations of metamaterials, such as cloaking, ultrafast optoelectronic switching, and imaging. Theoptical properties of VO2 can be strongly and quickly tuned via external stimuli due to its MIT.Utilizing MIT in VO2 toward metamaterial applications (32, 124, 140–142) opens a new researchdirection. In this subsection, we use a simple example to discuss VO2 metamaterial applicationsfor dynamic tuning of an infrared resonance (140). Panels e–g of Figure 10 show the schematicsof a VO2 split-ring resonator (SRR) hybrid metamaterial. Figure 10e shows a close-up view ofthe SRR on top of a near-field image of a VO2 thin film during MIT. Figure 10f shows the devicelayout and experimental setup. Gold SRRs (100 nm thick) of a 20-μm period were fabricated usinglithography on top of a 90-nm-thick VO2 thin film on a sapphire substrate. Figure 10g shows asingle SRR pattern and illustration of the overlap of the SRR fields with the VO2 thin film. Theresonance frequency of the SRR metamaterial is highly sensitive to the dielectric property of thematerial placed nearby, especially in the vicinity of the SRR gaps. Panels h and i of Figure 10 showexperimentally retrieved permittivity and permeability bandwidths for the hybrid metamaterials,demonstrating the controllability of resonance in VO2-based hybrid metamaterials.
4.3.3. Thermal sensors utilizing metal-insulator transition. Because the resistivity of VO2
changes sharply with temperature, VO2 can be used for thermal switches and thermal detectors(this application is being explored in night vision systems). Here, we use a programmable criticaltemperature sensor (34) as an example. Figure 11a shows the temperature dependency of theconductivity of a VO2 thermal sensor device (34). The ∼100-nm VO2 thin films were grown onm-plane sapphire substrates using the sol-gel method. The two-terminal device dimensions areshown in the inset of Figure 11a. The critical voltage for E-MIT is ∼21 V. As applied voltageincreases from 1 to 22 V, an abrupt conductivity jump occurs above 5 V. TMIT gradually shifts from68◦C at a 1-V bias to room temperature at a 21-V bias. Through use of this property (different MITtemperatures at different voltages), a prototype thermal sensor based on VO2 was demonstrated.Recently, Yang et al. (100) demonstrated a solid-state thermal capacitor device with a VO2 thinfilm as an active layer, which shows a more-than-one-order-of-magnitude capacitance changefrom room temperature to 100◦C due to the dielectric constant change of VO2 across the phasetransition. This could possibly be explored further as thermal sensors in solid-state circuits.
4.3.4. Chemical sensors utilizing metal-insulator transition. The carrier concentrationchanges dramatically in the vicinity of TMIT. VO2 nanowire gas sensors (35, 84) have been demon-strated accordingly. Figure 11b shows a schematic of a VO2 nanowire gas sensor thermistor device(35). VO2 nanowires were grown on a SiO2/Si substrate using the vapor solid method. Droplets ofliquid metal alloys (Ga-In-Sn) were used for soft contacts on both ends of the VO2 nanowire. The
358 Yang · Ko · Ramanathan
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
1 V3 V5 V7 V9 V11 V13 V15 V28 V19 V21 V22 V
20 µm
50 µ
m
T insulator
Intermediatephase
R metal
Vapp
SiO2
Heater
I V
M domain
ArPressure
93 Torr
V+MIT
5 6 7 827 47 67 87Temperature (°C) Bias (V)
Co
nd
uct
ivit
y, σ
(Ω
cm
)–1
Cu
rre
nt N
W (
A ×
10
–4)
101
100
6
4
2
0e e
c c
d d
e
b a
c
d
f PG
PL
SPT line
MIT line
A
1.5 Torr
560 Torr
320 Torr
10 µm
0 V
10 V
20 V
VO2
I domain
Figure 11Thermal sensor and gas sensor devices utilizing MIT in VO2. (a) A VO2 thermal sensor device. Shown are the temperature and voltagedependency of the conductivity and the coplanar VO2 device. The inset shows the device structure. SPT denotes structural phasetransition. (b) Schematic of a VO2 nanowire gas sensor device. PG and PL indicate heat flux dissipating into the gas environment andmetal contacts, respectively. (c–e) Microscopic images of a VO2 nanowire with increased self–Joule heating induced by the flowingcurrent. ( f ) I-V characteristics of the VO2 nanowire gas sensor device at different Ar pressures. Panel a adapted with permission fromReference 34; panels b–f adapted with permission from Reference 35.
soft contacts were used to prevent any strain-induced TMIT shift. A resistor of 11 k� was connectedin series with the VO2 nanowire device to limit current. Joule heating is sustained by an externallyapplied periodic voltage bias on the device. Panels c–e of Figure 11 show microscopic images of theVO2 nanowire under different voltage biases (different heating) (35). During the heating process,the metallic domain and insulating domain were observed. Figure 11f shows the I-V character-istics of the VO2 nanowire gas sensor device under different Ar gas pressures (35). Thermal lossof the gas sensor to the environment increases with Ar pressure, which results in forward shiftsof the transition voltage to larger values. Hence, significantly different I-V loops are observedfor different gas pressures, as shown in Figure 11f. This is a simple demonstration of a chemicalsensor.
4.4. Devices Utilizing Metal-Insulator Transition in CorrelatedOxides Other than VO2
Although this article focuses on VO2, a plethora of transition-metal–oxide systems show correlatedphenomena that could be interesting for novel solid-state devices. High-temperature supercon-ductors, such as in the Y-Ba-Cu-O system, have been consistently studied for superconduct-ing quantum interference devices, Josephson junctions, and related low-temperature electronics(143). Colossal magnetoresistance in manganites (144), such as in the La-Ca-Mn-O family, is
www.annualreviews.org • Oxide Electronics Utilizing Ultrafast Metal-Insulator Transitions 359
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
interesting for magnetic tunnel junctions and magnetic field sensors. The recent observation of ahigh-mobility electron gas at the interfaces between LaAlO3 and SrTiO3 (145) has naturally led tospeculation as to novel devices that could exploit this exotic phenomena (146, 147). Coupling effectsfrom layered oxide heterostructures also lead to novel phenomena involving multiferroic proper-ties. Strain is yet another parameter that can magnify the response to external stimuli or in somecases lower the energy barrier for switching. The field of quantum materials in general is seeingexplosive growth. Correlated oxides are at the forefront of this revolution, and research in this areahas benefited from advances in thin-film growth techniques, lithography, and spatially resolvedcharacterization over the past few decades. The topic is interdisciplinary and offers significant intel-lectual challenges for condensed matter physicists, materials scientists, and device physicists alike.
5. CONCLUDING REMARKS
In this brief review, we discuss the exploration of correlated oxide phase transitions in novelelectronics, photonics, and related devices. Although this area has intrigued researchers for severaldecades ever since the original observation of MIT in oxides, the twenty-first-century grand chal-lenge of continued innovation in information processing sciences beyond CMOS scaling createsan opportunity to seriously consider alternate computing vectors that may offer unique advantages.Lithographic scaling is essentially close to its limits, and alternate state variables are importantto address in a timely manner (39). Strongly correlated oxides could well serve that role due tostrong electron-electron and electron-lattice interaction effects and the ability to manipulate sucheffects through synthesis and compositional control. The ultrafast nature of the phase transitionalong with spectacular changes in the electrical/dielectric properties create several possibilitiesfor logic and memory devices, some of which could potentially be transformative. A practicaladvantage could be ease of integration onto Si platforms given that the thin-film technology iscompatible with traditional semiconductor processing and near–room temperature operation.
Once the governing physics is better understood, a critical study would then be needed tobenchmark elementary devices and the projected limits of the technology, followed by reliabilityanalyses. For example, can structural distortions during a MIT be tolerated in practical devicesthat may require millions of switching operations? In terms of correlated electron physics, Mottinsulators and their interfaces present model systems for fundamental research. In addition, phasetransitions create unique and enabling opportunities for defense electronics, such as in addressingdirected energy, protection, and sensing. Finally, given the quasi-two-dimensional nature ofMIT (the high carrier concentration even in the insulating state leads to a small field penetrationdepth), one can explore interface-driven phenomena as well with thin-film model systems. Forexample, a recent theoretical study predicts semi–Dirac point behavior at VO2-TiO2 interfaces(148, 149), which could be very interesting to experimentally investigate. Furthermore, thisprediction may be generalized to interfaces between a Mott insulator and a band insulator. Thisleads to a fascinating question: Can one realize graphene-like electron transport phenomena inoxides displaying phase transitions?
SUMMARY POINTS
1. An ultrafast switch can be demonstrated utilizing metal-insulator transition in correlatedoxides, with the ON and OFF states defined as a low-resistance, metallic phase and ahigh-resistance, insulating phase of the material, respectively. Switching may be triggeredwith electronic, optical, thermal, or magnetic perturbations or a combination thereof.
360 Yang · Ko · Ramanathan
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
2. Room temperature MottFETs with correlated oxides may be potential candidates for fu-ture computing elements. Utilizing a functional correlated oxide for information process-ing may lead to three-dimensional circuits and integration on polycrystalline or arbitrarysubstrates.
3. The highly nonlinear current-voltage characteristics of two-terminal devices with phasetransition elements coupled with hysteresis lead to intriguing possibilities in varis-tors, memristive devices, neuromorphic circuits, and a variety of solid-state circuitarchitectures.
4. High-quality quantum materials are central to experimentally realizing these devices.Rigorous materials science studies would advance synthesis of correlated oxides with su-perior functionality along with a fundamental understanding of the role of microstructureand defects.
FUTURE ISSUES
Above we review published literature on phase transitions and ongoing efforts to realizenovel electron devices utilizing sharp transitions. Here, we briefly explore what else couldbe possible. Indeed, some of these ideas rely simply on the phase transition phenomenadriven by thermal excitations or in some cases by a current and hence may be possible withexisting materials.
1. Thermal switch for directing heat flow: Phase transition leads to a conductance changethat is accompanied by a large change in mobile carrier concentration. There may quitepossibly be a thermal conductivity change during this transition (150). Hence, explor-ing oxide systems wherein one can design solid-state thermal switches by exploitingphase transitions may be an interesting avenue of research. Such research may lead toself-regulating thermal barriers whose thermal conductivity dynamically adjusts to thesurrounding temperature.
2. Information processing in a thermal gradient: If an array of phase transition elementsis placed in a thermal gradient in the temperature range in which phase transition canspontaneously happen, then one may be able to design a circuit that can naturally switchand perform computation using thermal waves. The heat flow across the gradient (suchas in a solid-state device that dissipates heat) may hence be exploited for useful computa-tional work. This could be a novel approach to utilize part of the dissipated heat, whichis often a challenge to manage effectively in high-performance devices.
3. Spin-dependent metal-insulator transition: By fabricating a phase transition electrodeonto a spin tunnel junction, one may be able to realize a spin-polarized current drivinga phase transition. Hence, creating a spintronic device that incorporates ultrafast phasetransitions may be possible.
4. Materials discovery: Oxide materials that possess low dielectric loss and can be reversiblyswitched to a conducting state would be particularly interesting to realize low-standby-power electronics.
www.annualreviews.org • Oxide Electronics Utilizing Ultrafast Metal-Insulator Transitions 361
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
DISCLOSURE STATEMENT
The authors are not aware of any affiliations, memberships, funding, or financial holdings thatmight be perceived as affecting the objectivity of this review.
ACKNOWLEDGMENTS
We acknowledge grants from the Office of Naval Research (N00014-10-1-0131) and the Air ForceOffice of Scientific Research (FA9550-08-1-0203) for financial support.
LITERATURE CITED
1. Mott NF. 1968. Metal-insulator transition. Rev. Mod. Phys. 40:677–832. Doniach S. 1969. Insulator-metal transition. Adv. Phys. 18:819–483. Mott NF. 1974. Metal-Insulator Transitions. London: Taylor & Francis. 278 pp.4. Van Zandt LL, Honig JM. 1974. Theories pertaining to the semiconductor-metal transition in crystals.
Annu. Rev. Mater. Sci. 4:191–2205. Milligan RF, Thomas GA. 1985. The metal-insulator transition. Annu. Rev. Phys. Chem. 36:139–586. Mott NF. 1990. Metal-Insulator Transitions. London: Taylor & Francis. 286 pp.7. Gebhard F. 1997. The Mott Metal-Insulator Transition: Models and Methods. Berlin/London: Springer.
317 pp.8. Imada M, Fujimori A, Tokura Y. 1998. Metal-insulator transitions. Rev. Mod. Phys. 70:1039–2639. Edwards PP, Johnston RL, Hensel F, Rao CNR, Tunstall DP. 1999. A perspective on the metal-nonmetal
transition. Solid State Phys. Adv. Res. Appl. 52:229–33810. Kravchenko SV, Sarachik MP. 2004. Metal-insulator transition in two-dimensional electron systems.
Rep. Prog. Phys. 67:1–4411. Morin FJ. 1959. Oxides which show a metal-to-insulator transition at the Neel temperature. Phys. Rev.
Lett. 3:34–3612. Adler D. 1968. Mechanisms for metal-nonmetal transitions in transition-metal oxides and sulfides. Rev.
Mod. Phys. 40:714–3613. Austin IG, Mott NF. 1970. Metallic and nonmetallic behavior in transition metal oxides. Science 168:71–
7714. Goodenough JB. 1971. Anomalous properties of the vanadium oxides. Annu. Rev. Mater. Sci. 1:101–3815. Honig JM, Vanzandt LL. 1975. Metal-insulator transition in selected oxides. Annu. Rev. Mater. Sci.
5:225–7816. Ganguly P, Vasanthacharya NY, Rao CNR, Edwards PP. 1984. Composition-controlled metal-insulator
transitions and minimum metallic conductivity in the oxide systems LaNi1−xCrxO3, LaNi1−xMnxO3,LaNi1−xFexO3, or LaNi1−xCoxO3. J. Solid State Chem. 54:400–6
17. Raychaudhuri AK. 1995. Metal-insulator transition in perovskite oxides: a low-temperature perspective.Adv. Phys. 44:21–46
18. Tokura Y, Nagaosa N. 2000. Orbital physics in transition-metal oxides. Science 288:462–6819. Tokura Y. 2003. Correlated-electron physics in transition-metal oxides. Phys. Today 56:50–5520. Bednorz JG, Muller KA. 1986. Possible high-Tc superconductivity in the Ba-La-Cu-O system. Z. Phys.
B 64:189–9321. Jin S, Tiefel TH, McCormack M, Fastnacht RA, Ramesh R, Chen LH. 1994. Thousandfold change in
resistivity in magnetoresistive La-Ca-Mn-O films. Science 264:413–1522. Torrance JB, Lacorre P, Nazzal AI, Ansaldo EJ, Niedermayer C. 1992. Systematic study of insulator-
metal transitions in perovskites RNiO3 (R = Pr, Nd, Sm, Eu) due to closing of charge-transfer gap.Phys. Rev. B 45:8209–12
23. Qazilbash MM, Brehm M, Chae BG, Ho PC, Andreev GO, et al. 2007. Mott transition in VO2 revealedby infrared spectroscopy and nano-imaging. Science 318:1750–53
24. Asamitsu A, Tomioka Y, Kuwahara H, Tokura Y. 1997. Current switching of resistive states in magne-toresistive manganites. Nature 388:50–52
362 Yang · Ko · Ramanathan
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
25. Miyano K, Tanaka T, Tomioka Y, Tokura Y. 1997. Photoinduced insulator-to-metal transition in aperovskite manganite. Phys. Rev. Lett. 78:4257–60
26. Fiebig M, Miyano K, Tomioka Y, Tokura Y. 1998. Visualization of the local insulator-metal transitionin Pr0.7Ca0.3MnO3. Science 280:1925–28
27. Cao J, Ertekin E, Srinivasan V, Fan W, Huang S, et al. 2009. Strain engineering and one-dimensionalorganization of metal-insulator domains in single-crystal vanadium dioxide beams. Nat. Nanotechnol.4:732–37
28. Ruzmetov D, Gopalakrishnan G, Deng JD, Narayanamurti V, Ramanathan S. 2009. Electrical triggeringof metal-insulator transition in nanoscale vanadium oxide junctions. J. Appl. Phys. 106:083702
29. Ruzmetov D, Gopalakrishnan G, Ko C, Narayanamurti V, Ramanathan S. 2010. Three-terminal fieldeffect devices utilizing thin film vanadium oxide as the channel layer. J. Appl. Phys. 107:114516
30. Briggs RM, Pryce IM, Atwater HA. 2010. Compact silicon photonic waveguide modulator based on thevanadium dioxide metal-insulator phase transition. Opt. Express 18:11192–201
31. Gu Q, Falk A, Wu JQ, Lian OY, Park H. 2007. Current-driven phase oscillation and domain-wallpropagation in WxV1−xO2 nanobeams. Nano Lett. 7:363–66
32. Dicken MJ, Aydin K, Pryce IM, Sweatlock LA, Boyd EM, et al. 2009. Frequency tunable near-infraredmetamaterials based on VO2 phase transition. Opt. Express 17:18330–39
33. Driscoll T, Kim HT, Chae BG, Di Ventra M, Basov DN. 2009. Phase-transition driven memristivesystem. Appl. Phys. Lett. 95:043503
34. Kim BJ, Lee YW, Chae BG, Yun SJ, Oh SY, et al. 2007. Temperature dependence of the first-order metal-insulator transition in VO2 and programmable critical temperature sensor. Appl. Phys. Lett. 90:023515
35. Strelcov E, Lilach Y, Kolmakov A. 2009. Gas sensor based on metal-insulator transition in VO2 nanowirethermistor. Nano Lett. 9:2322–26
36. Schulz M. 1999. The end of the road for silicon? Nature 399:729–3037. Ieong M, Doris B, Kedzierski J, Rim K, Yang M. 2004. Silicon device scaling to the sub-10-nm regime.
Science 306:2057–6038. Brewer JE, Zhirnov VV, Hutchby JA. 2005. Memory technology from the post CMOS era. IEEE Circuits
Devices 21:13–2039. Zhirnov VV, Hutchby JA, Bourianoff GI, Brewer JE. 2005. Emerging research logic devices. IEEE
Circuits Devices 21:37–4640. Theis TN, Solomon PM. 2010. It’s time to reinvent the transistor! Science 327:1600–141. Mott NF. 1949. The basis of the electron theory of metals, with special reference to the transition metals.
Proc. Phys. Soc. Lond. A 62:416–2242. Peierls RE. 1964. Quantum Theory of Solids. Oxford, UK: Clarendon. 229 pp.43. Gruner G. 1988. The dynamics of charge-density waves. Rev. Mod. Phys. 60:1129–8244. Anderson PW. 1958. Absence of diffusion in certain random lattices. Phys. Rev. 109:1492–50545. Belitz D, Kirkpatrick TR. 1994. The Anderson-Mott transition. Rev. Mod. Phys. 66:261–39046. Evers F, Mirlin AD. 2008. Anderson transitions. Rev. Mod. Phys. 80:1355–41747. Westman S. 1961. Note on a phase transition in VO2. Acta Chem. Scand. 15:21748. Heckingbottom R, Linnett JW. 1962. Structure of vanadium dioxide. Nature 194:67849. Goodenough JB. 1971. The two components of crystallographic transition in VO2. J. Solid State Chem.
3:490–50050. Grinolds MS, Lobastov VA, Weissenrieder J, Zewail AH. 2006. Four-dimensional ultrafast electron
microscopy of phase transitions. Proc. Natl. Acad. Sci. USA 103:18427–3151. Baum P, Yang DS, Zewail AH. 2007. 4D visualization of transitional structures in phase transformations
by electron diffraction. Science 318:788–9252. Zylbersztejn A, Mott NF. 1975. Metal-insulator transition in vanadium dioxide. Phys. Rev. B 11:4383–9553. Wentzcovitch RM, Schulz WW, Allen PB. 1994. VO2: Peierls or Mott-Hubbard? A view from band
theory. Phys. Rev. Lett. 72:3389–9254. Wentzcovitch RM, Schulz WW, Allen PB. 1994. Comment on VO2: Peierls or Mott-Hubbard? A view
from band theory. Reply. Phys. Rev. Lett. 73:304355. Rice TM, Launois H, Pouget JP. 1994. Comment on VO2: Peierls or Mott-Hubbard? A view from band
theory. Phys. Rev. Lett. 73:3042
www.annualreviews.org • Oxide Electronics Utilizing Ultrafast Metal-Insulator Transitions 363
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
56. Cavalleri A, Dekorsy T, Chong HHW, Kieffer JC, Schoenlein RW. 2004. Evidence for a structurally-driven insulator-to-metal transition in VO2: a view from the ultrafast timescale. Phys. Rev. B 70:161102
57. Kim HT, Lee YW, Kim BJ, Chae BG, Yun SJ, et al. 2006. Monoclinic and correlated metal phase inVO2 as evidence of the Mott transition: coherent phonon analysis. Phys. Rev. Lett. 97:266401
58. Kachi S, Kosuge K, Okinaka H. 1973. Metal-insulator transition in VnO2n−1. J. Solid State Chem. 6:258–70
59. Demeter M, Neumann M, Reichelt W. 2000. Mixed-valence vanadium oxides studied by XPS. Surf. Sci.454:41–44
60. Katzke H, Toledano P, Depmeier W. 2003. Theory of morphotropic transformations in vanadiumoxides. Phys. Rev. B 68:024109
61. Irizawa A, Higashiya A, Tsunekawa M, Sekiyama A, Imada S, et al. 2005. Metal-insulator transition inV6O13 probed by photoernission and optical studies. J. Electron Spectrosc. Relat. Phenom. 144:345–47
62. Ladd LA, Paul W. 1969. Optical and transport properties of high quality crystals of V2O4 near metallictransition temperature. Solid State Commun. 7:425–28
63. Sambi M, Sangiovanni G, Granozzi G, Parmigiani F. 1997. Growth and the structure of epitaxial VO2
at the TiO2(110) surface. Phys. Rev. B 55:7850–5864. Lu JW, West KG, Wolf SA. 2008. Very large anisotropy in the dc conductivity of epitaxial VO2 thin
films grown on (011) rutile TiO2 substrates. Appl. Phys. Lett. 93:26210765. Gupta A, Aggarwal R, Gupta P, Dutta T, Narayan RJ, Narayan J. 2009. Semiconductor to metal transition
characteristics of VO2 thin films grown epitaxially on Si (001). Appl. Phys. Lett. 95:11191566. Ruzmetov D, Zawilski KT, Narayanamurti V, Ramanathana S. 2007. Structure-functional property
relationships in rf-sputtered vanadium dioxide thin films. J. Appl. Phys. 102:11371567. Ruzmetov D, Senanayake SD, Ramanathan S. 2007. X-ray absorption spectroscopy of vanadium dioxide
thin films across the phase-transition boundary. Phys. Rev. B 75:19510268. Ruzmetov D, Zawilski KT, Senanayake SD, Narayanamurti V, Ramanathan S. 2008. Infrared reflectance
and photoemission spectroscopy studies across the phase transition boundary in thin film vanadiumdioxide. J. Phys. Condens. Matter 20:465204
69. Ruzmetov D, Senanayake SD, Narayanamurti V, Ramanathan S. 2008. Correlation between metal-insulator transition characteristics and electronic structure changes in vanadium oxide thin films. Phys.Rev. B 77:195442
70. Ruzmetov D, Heiman D, Claflin BB, Narayanamurti V, Ramanathan S. 2009. Hall carrier density andmagnetoresistance measurements in thin-film vanadium dioxide across the metal-insulator transition.Phys. Rev. B 79:153107
71. Ko C, Ramanathan S. 2008. Stability of electrical switching properties in vanadium dioxide thin filmsunder multiple thermal cycles across the phase transition boundary. J. Appl. Phys. 104:086105
72. Ko C, Ramanathan S. 2008. Effect of ultraviolet irradiation on electrical resistance and phase transitioncharacteristics of thin film vanadium oxide. J. Appl. Phys. 103:106104
73. Ko C, Ramanathan S. 2009. Dispersive capacitance and conductance across the phase transition boundaryin metal-vanadium oxide-silicon devices. J. Appl. Phys. 106:034101
74. Lopez R, Haynes TE, Boatner LA, Feldman LC, Haglund RF. 2002. Size effects in the structural phasetransition of VO2 nanoparticles. Phys. Rev. B 65:224113
75. Suh JY, Lopez R, Feldman LC, Haglund RF. 2004. Semiconductor to metal phase transition in thenucleation and growth of VO2 nanoparticles and thin films. J. Appl. Phys. 96:1209–13
76. Donev EU, Lopez R, Feldman LC, Haglund RF. 2009. Confocal Raman microscopy across the metal-insulator transition of single vanadium dioxide nanoparticles. Nano Lett. 9:702–6
77. Wu JQ, Gu Q, Guiton BS, de Leon NP, Lian OY, Park H. 2006. Strain-induced self organization ofmetal-insulator domains in single-crystalline VO2 nanobeams. Nano Lett. 6:2313–17
78. Wei J, Wang ZH, Chen W, Cobden DH. 2009. New aspects of the metal-insulator transition in single-domain vanadium dioxide nanobeams. Nat. Nanotechnol. 4:420–24
79. Zhang SX, Chou JY, Lauhon LJ. 2009. Direct correlation of structural domain formation with the metalinsulator transition in a VO2 nanobeam. Nano Lett. 9:4527–32
80. Guiton BS, Gu Q, Prieto AL, Gudiksen MS, Park H. 2005. Single-crystalline vanadium dioxide nanowireswith rectangular cross sections. J. Am. Chem. Soc. 127:498–99
364 Yang · Ko · Ramanathan
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
81. Sohn JI, Joo HJ, Porter AE, Choi CJ, Kim K, et al. 2007. Direct observation of the structural componentof the metal-insulator phase transition and growth habits of epitaxially grown VO2 nanowires. Nano Lett.7:1570–74
82. Baik JM, Kim MH, Larson C, Wodtke AM, Moskovits M. 2008. Nanostructure-dependent metal-insulator transitions in vanadium-oxide nanowires. J. Phys. Chem. C 112:13328–31
83. Sohn JI, Joo HJ, Ahn D, Lee HH, Porter AE, et al. 2009. Surface-stress-induced Mott transition andnature of associated spatial phase transition in single crystalline VO2 nanowires. Nano Lett. 9:3392–97
84. Baik JM, Kim MH, Larson C, Yavuz CT, Stucky GD, et al. 2009. Pd-sensitized single vanadium ox-ide nanowires: highly responsive hydrogen sensing based on the metal-insulator transition. Nano Lett.9:3980–84
85. Liu JF, Li QH, Wang TH, Yu DP, Li YD. 2004. Metastable vanadium dioxide nanobelts: hydrothermalsynthesis, electrical transport, and magnetic properties. Angew. Chem. Int. Ed. 43:5048–52
86. Kim DH, Kwok HS. 1994. Pulsed-laser deposition of VO2 thin films. Appl. Phys. Lett. 65:3188–9087. Muraoka Y, Hiroi Z. 2002. Metal-insulator transition of VO2 thin films grown on TiO2 (001) and (110)
substrates. Appl. Phys. Lett. 80:583–8588. Sahana MB, Subbanna GN, Shivashankar SA. 2002. Phase transformation and semiconductor-metal
transition in thin films of VO2 deposited by low-pressure metalorganic chemical vapor deposition.J. Appl. Phys. 92:6495–504
89. Maruyama T, Ikuta Y. 1993. Vanadium dioxide thin films prepared by chemical vapour deposition fromvanadium(III) acetylacetonate. J. Mater. Sci. 28:5073–78
90. Chae BG, Kim HT, Yun SJ, Kim BJ, Lee YW, et al. 2006. Highly oriented VO2 thin films prepared bysol-gel deposition. Electrochem. Solid State Lett. 9:C12–14
91. Chae BG, Kim HT, Yun SJ. 2008. Characteristics of W- and Ti-doped VO2 thin films prepared bysol-gel method. Electrochem. Solid State Lett. 11:D53–55
92. Hood PJ, Denatale JF. 1991. Millimeter-wave dielectric properties of epitaxial vanadium dioxide thinfilms. J. Appl. Phys. 70:376–81
93. Mani RG, Ramanathan S. 2007. Observation of a uniform temperature dependence in the electricalresistance across the structural phase transition in thin film vanadium oxide (VO2). Appl. Phys. Lett.91:062104
94. West KG, Lu JW, Yu J, Kirkwood D, Chen W, et al. 2008. Growth and characterization of vanadiumdioxide thin films prepared by reactive-biased target ion beam deposition. J. Vac. Sci. Technol. A 26:133–39
95. West KG, Lu JW, He L, Kirkwood D, Chen W, et al. 2008. Ferromagnetism in rutile structure Crdoped VO2 thin films prepared by reactive-bias target ion beam deposition. J. Supercond. Novel Magn.21:87–92
96. Griffith CH, Eastwood HK. 1974. Influence of stoichiometry on metal-semiconductor transition invanadium dioxide. J. Appl. Phys. 45:2201–6
97. Youn DH, Kim HT, Chae BG, Hwang YJ, Lee JW, et al. 2004. Phase and structural characterizationof vanadium oxide films grown on amorphous SiO2/Si substrates. J. Vac. Sci. Technol. A 22:719–24
98. Narayan J, Bhosle VM. 2006. Phase transition and critical issues in structure-property correlations ofvanadium oxide. J. Appl. Phys. 100:103524
99. Yang Z, Ko C, Ramanathan S. 2010. Metal-insulator transition characteristics of VO2 thin films grownon Ge (100) single crystals. J. Appl. Phys. 108:073708
100. Yang Z, Ko C, Balakrishnan V, Gopalakrishnan G, Ramanathan S. 2010. Dielectric and carrier transportproperties of vanadium dioxide thin films across the phase transition utlizing gated capacitor devices.Phys. Rev. B 82:205101
101. Duchene J, Terraill M, Pailly P, Adam G. 1971. Filamentary conduction in VO2 coplanar thin-filmdevices. Appl. Phys. Lett. 19:115–17
102. Taketa Y, Kato F, Nitta M, Haradome M. 1975. New oscillation phenomena in VO2 crystals. Appl. Phys.Lett. 27:212–14
103. Fisher B. 1978. Voltage oscillations in switching VO2 needles. J. Appl. Phys. 49:5339–41104. Stefanovich G, Pergament A, Stefanovich D. 2000. Electrical switching and Mott transition in VO2.
J. Phys. Condens. Matter 12:8837–45
www.annualreviews.org • Oxide Electronics Utilizing Ultrafast Metal-Insulator Transitions 365
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
105. Gopalakrishnan G, Ruzmetov D, Ramanathan S. 2009. On the triggering mechanism for the metal-insulator transition in thin film VO2 devices: electric field versus thermal effects. J. Mater. Sci. 44:5345–53
106. Kim J, Ko C, Frenzel A, Ramanathan S, Hoffman JE. 2010. Nanoscale imaging and control of resistanceswitching in VO2 at room temperature. Appl. Phys. Lett. 96:213106
107. Ko C, Ramanathan S. 2008. Observation of electric field-assisted phase transition in thin film vanadiumoxide in a metal-oxide-semiconductor device geometry. Appl. Phys. Lett. 93:252101
108. Chae BG, Kim HT, Youn DH, Kang KY. 2005. Abrupt metal-insulator transition observed in VO2 thinfilms induced by a switching voltage pulse. Physica B 369:76–80
109. Boriskov PP, Velichko AA, Pergament AL, Stefanovich GB, Stefanovich DG. 2002. The effect of electricfield on metal-insulator phase transition in vanadium dioxide. Tech. Phys. Lett. 28:406–8
110. Kim HT, Chae BG, Youn DH, Maeng SL, Kim G, et al. 2004. Mechanism and observation of Motttransition in VO2-based two- and three-terminal devices. N. J. Phys. 6:52
111. Kim BJ, Lee YW, Choi S, Yun SJ, Kim HT. 2010. VO2 thin-film varistor based on metal-insulatortransition. IEEE Electron Device Lett. 31:14–16
112. Hormoz S, Ramanathan S. 2010. Limits on vanadium oxide Mott metal-insulator transition field-effecttransistors. Solid-State Electron. 54:654–59
113. Zhou C, Newns DM, Misewich JA, Pattnaik PC. 1997. A field effect transistor based on the Motttransition in a molecular layer. Appl. Phys. Lett. 70:598–600
114. Newns DM, Misewich JA, Tsuei CC, Gupta A, Scott BA, Schrott A. 1998. Mott transition field effecttransistor. Appl. Phys. Lett. 73:780–82
115. Ahn CH, Triscone JM, Mannhart J. 2003. Electric field effect in correlated oxide systems. Nature424:1015–18
116. Sakai J, Kurisu M. 2008. Effect of pressure on the electric-field-induced resistance switching of VO2
planar-type junctions. Phys. Rev. B 78:033106117. Chudnovskiy F, Luryi S, Spivak B. 2002. Switching device based on a first-order metal-insulator transition
induced by an external electric field. In Future Trends in Microelectronics: The Nano Millennium, ed. S Luryi,JM Xu, A Zaslavsky, pp. 148–55. New York: Wiley Intersci.
118. Ueno K, Nakamura S, Shimotani H, Ohtomo A, Kimura N, et al. 2008. Electric-field-induced super-conductivity in an insulator. Nat. Mater. 7:855–58
119. Lee YW, Kim BJ, Lim JW, Yun SJ, Choi S, et al. 2008. Metal-insulator transition-induced electricaloscillation in vanadium dioxide thin film. Appl. Phys. Lett. 92:162903
120. Kim HT, Kim BJ, Choi S, Chae BG, Lee YW, et al. 2010. Electrical oscillations induced by the metal-insulator transition in VO2. J. Appl. Phys. 107:023702
121. Chua LO. 1971. Memristor: the missing circuit element. IEEE Trans. Circuit Theor. 18:507–19122. Strukov DB, Snider GS, Stewart DR, Williams RS. 2008. The missing memristor found. Nature 453:80–
83123. Jo SH, Chang T, Ebong I, Bhadviya BB, Mazumder P, Lu W. 2010. Nanoscale memristor device as
synapse in neuromorphic systems. Nano Lett. 10:1297–301124. Driscoll T, Kim HT, Chae BG, Kim BJ, Lee YW, et al. 2009. Memory metamaterials. Science 325:1518–
21125. Roach WR. 1971. Holographic storage in VO2. Appl. Phys. Lett. 19:453–55126. Smith AW. 1973. Optical storage in VO2 films. Appl. Phys. Lett. 23:437–38127. Becker MF, Buckman AB, Walser RM, Lepine T, Georges P, Brun A. 1994. Femtosecond laser excitation
of the semiconductor-metal phase-transition in VO2. Appl. Phys. Lett. 65:1507–9128. Nakajima M, Takubo N, Hiroi Z, Ueda Y, Suemoto T. 2008. Photoinduced metallic state in VO2 proved
by the terahertz pump-probe spectroscopy. Appl. Phys. Lett. 92:011907129. Lysenko S, Rua AJ, Vikhnin V, Jimenez J, Fernandez F, Liu H. 2006. Light-induced ultrafast phase
transitions in VO2 thin film. Appl. Surf. Sci. 252:5512–15130. Cavalleri A, Chong HHW, Fourmaux S, Glover TE, Heimann PA, et al. 2004. Picosecond soft
X-ray absorption measurement of the photoinduced insulator-to-metal transition in VO2. Phys. Rev.B 69:153106
366 Yang · Ko · Ramanathan
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41CH14-Ramanathan ARI 3 June 2011 8:32
131. Vikhnin VS, Lysenko S, Rua A, Fernandez F, Liu H. 2006. The model of ultrafast light-induced insulator-metal phase transition in VO2. Solid State Commun. 137:615–20
132. Gea LA, Boatner LA. 1996. Optical switching of coherent VO2 precipitates formed in sapphire by ionimplantation and annealing. Appl. Phys. Lett. 68:3081–83
133. Lysenko S, Rua A, Vikhnin V, Fernandez F, Liu H. 2007. Insulator-to-metal phase transition andrecovery processes in VO2 thin films after femtosecond laser excitation. Phys. Rev. B 76:035104
134. Rini M, Hao Z, Schoenlein RW, Giannetti C, Parmigiani F, et al. 2008. Optical switching in VO2 filmsby below-gap excitation. Appl. Phys. Lett. 92:181904
135. Cilento F, Giannetti C, Ferrini G, Dal Conte S, Sala T, et al. 2010. Ultrafast insulator-to-metal phasetransition as a switch to measure the spectrogram of a supercontinuum light pulse. Appl. Phys. Lett.96:021102
136. Jorgenson GV, Lee JC. 1986. Doped vanadium oxide for optical switching films. Solar Energy Mater.14:205–14
137. Dumas-Bouchiat F, Champeaux C, Catherinot A, Crunteanu A, Blondy P. 2007. rf-microwave switchesbased on reversible semiconductor-metal transition of VO2 thin films synthesized by pulsed-laser depo-sition. Appl. Phys. Lett. 91:223505
138. Dragoman M, Cismaru A, Hartnagel H, Plana R. 2006. Reversible metal-semiconductor transitions formicrowave switching applications. Appl. Phys. Lett. 88:073503
139. Padilla WJ, Basov DN, Smith DR. 2006. Negative refractive index metamaterials. Mater. Today 9:28–35140. Driscoll T, Palit S, Qazilbash MM, Brehm M, Keilmann F, et al. 2008. Dynamic tuning of an infrared
hybrid-metamaterial resonance using vanadium dioxide. Appl. Phys. Lett. 93:024101141. Wen QY, Zhang HW, Yang QH, Xie YS, Chen K, Liu YL. 2010. Terahertz metamaterials with VO2
cut-wires for thermal tunability. Appl. Phys. Lett. 97:021111142. Seo M, Kyoung J, Park H, Koo S, Kim HS, et al. 2010. Active terahertz nanoantennas based on VO2
phase transition. Nano Lett. 10:2064–68143. Makhlin Y, Schon G, Shnirman A. 2001. Quantum-state engineering with Josephson-junction devices.
Rev. Mod. Phys. 73:357–400144. Salamon MB, Jaime M. 2001. The physics of manganites: structure and transport. Rev. Mod. Phys.
73:583–628145. Ohtomo A, Hwang HY. 2004. A high-mobility electron gas at the LaAlO3/SrTiO3 heterointerface.
Nature 427:423–26146. Mannhart J, Schlom DG. 2010. Oxide interfaces: an opportunity for electronics. Science 327:1607–11147. Cen C, Thiel S, Hammerl G, Schneider CW, Andersen KE, et al. 2008. Nanoscale control of an interfacial
metal-insulator transition at room temperature. Nat. Mater. 7:298–302148. Pardo V, Pickett WE. 2009. Half-metallic semi-Dirac-point generated by quantum confinement in
TiO2/VO2 nanostructures. Phys. Rev. Lett. 102:166803149. Pardo V, Pickett WE. 2010. Metal-insulator transition through a semi-Dirac point in oxide nanostruc-
tures: VO2 (001) layers confined within TiO2. Phys. Rev. B 81:035111150. Oh DW, Ko C, Ramanathan S, Cahill DG. 2010. Thermal conductivity and dynamic heat capacity across
the metal-insulator transition in thin film VO2. Appl. Phys. Lett. 96:151906151. Cavalleri A, Toth C, Siders CW, Squier JA, Raksi F, et al. 2001. Femtosecond structural dynamics in
VO2 during an ultrafast solid-solid phase transition. Phys. Rev. Lett. 87:237401152. Kubler C, Ehrke H, Huber R, Lopez R, Halabica A, et al. 2007. Coherent structural dynamics and
electronic correlations during an ultrafast insulator-to-metal phase transition in VO2. Phys. Rev. Lett.99:116401
153. Hada M, Okimura K, Matsuo J. 2010. Characterization of structural dynamics of VO2 thin film onc-Al2O3 using in-air time-resolved X-ray diffraction. Phys. Rev. B 82:153401
154. Lobastov VA, Weissenrieder J, Tang J, Zewail AH. 2007. Ultrafast electron microscopy (UEM): four-dimensional imaging and diffraction of nanostructures during phase transitions. Nano Lett. 7:2552–58
www.annualreviews.org • Oxide Electronics Utilizing Ultrafast Metal-Insulator Transitions 367
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41-FrontMatter ARI 4 June 2011 12:1
Annual Review ofMaterials Research
Volume 41, 2011Contents
Materials Science of Biological Systems
Advances in Drug DeliveryBrian P. Timko, Kathryn Whitehead, Weiwei Gao, Daniel S. Kohane,
Omid Farokhzad, Daniel Anderson, and Robert Langer � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1
Crystallization Pathways in BiomineralizationSteve Weiner and Lia Addadi � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �21
Deformation and Fracture Mechanisms of Bone and NacreRizhi Wang and Himadri S. Gupta � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �41
Linear and Nonlinear Rheology of Living CellsPhilip Kollmannsberger and Ben Fabry � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �75
Mussel-Inspired Adhesives and CoatingsBruce P. Lee, P.B. Messersmith, J.N. Israelachvili, and J.H. Waite � � � � � � � � � � � � � � � � � � � � � �99
Nanomechanics of the Cartilage Extracellular MatrixLin Han, Alan J. Grodzinsky, and Christine Ortiz � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 133
Plant Stems: Functional Design and MechanicsThomas Speck and Ingo Burgert � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 169
Current Interest
Elastic and Mechanical Properties of the MAX PhasesMichel W. Barsoum and Miladin Radovic � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 195
Electrocaloric MaterialsJ.F. Scott � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 229
Electrochromic MaterialsRoger J. Mortimer � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 241
Nanowire Solar CellsErik C. Garnett, Mark L. Brongersma, Yi Cui, and Michael D. McGehee � � � � � � � � � � � � 269
Nonconventional (Non-Silicon-Based) Photovoltaic MaterialsT. Unold and H.W. Schock � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 297
v
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.
MR41-FrontMatter ARI 4 June 2011 12:1
On the Future Availability of the Energy MetalsT.E. Graedel � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 323
Oxide Electronics Utilizing Ultrafast Metal-Insulator TransitionsZheng Yang, Changhyun Ko, and Shriram Ramanathan � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 337
Point Defects in Oxides: Tailoring Materials ThroughDefect EngineeringHarry L. Tuller and Sean R. Bishop � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 369
Recent Developments in Semiconductor Thermoelectric Physicsand MaterialsAli Shakouri � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 399
Thermoelectric Phenomena, Materials, and ApplicationsTerry M. Tritt � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 433
Index
Cumulative Index of Contributing Authors, Volumes 37–41 � � � � � � � � � � � � � � � � � � � � � � � � � � � 449
Errata
An online log of corrections to Annual Review of Materials Research articles may befound at http://matsci.annualreviews.org/errata.shtml
vi Contents
Ann
u. R
ev. M
ater
. Res
. 201
1.41
:337
-367
. Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Har
vard
Uni
vers
ity o
n 07
/06/
11. F
or p
erso
nal u
se o
nly.