nanoscale thermal mapping of vo 2 - harvard...
TRANSCRIPT
Nanoscale Thermal Mapping of VO2
Changhyun KoProf. Shriram Ramanathan
Harvard School of Engineering & Applied Sciences
Jenny Hoffman
Contributors:
Thanks to seed funding from:
Adam PivonkaAlex Frenzel
Kevin O’ConnorHarry Mickalide
Prof. Eric HudsonHarvard Physics
V
VO2
n-type Si
100 nm
VO2 Insulator to Metal Transition
Tetragonal Metal
Monoclinic Insulator
T (K)
340 K
Goodenough, J.Solid State Chem 3, 490 (1971)
EF
3dπ (V-O-V)
3d|| (V-V)
EF
3dπ (V-O-V)
3d|| (V-V)
0.7 eV
Insulator-Metal Transition• temperature: ~340 K • uniaxial stress: ~38 kbar• optical excitation• electric field???? ~107 V/m
Transition features• tunability near room temperature• 80 fs switching time• large change in optical reflectance• high resistivity ratio: up to 105
Applications• sensors• data storage• memristors• tunable metamaterials
2
Can the transition be triggered by E-field alone??
Thermal vs. Electronic Transition?
Morin, PRL 3, 34 (1959)
500 250 167 125 100 83T (Kelvin) Considered a thermal transition…
Caveats:1. Joule heating?2. electrical breakdown of gate?
3
Stefanovich, JPCM 12, 8837 (2000)
…until 40 years later: electronic triggering?
Qazilbash, APL 92, 241906 (2008)
Motivations
4
300 320 340 360 380
102
103
104
105
Heating Cooling
T (Kelvin)
R (Ω
)
VO2
n-type Si
Pd
Bulk thermal transition
0 5 100
50
100
150
Cur
rent
(uA
)
Bias (V)
Voltage-triggered transition
VO2
n-type Si
Au
Kim, Ko, Frenzel, Ramanathan, Hoffman, APL 96, 213106 (2010)
1. VO2: E-field vs. Joule heating
2. General scanning technique for nanoscale T measurement
Question boils down to:what is the local temperature here?
Inhomogeneity → need local probe
5Qazilbash, Science 318, 1750 (2007)
insulator metal
Kim, Ko, Frenzel, Ramanathan, Hoffman, APL 96, 213106 (2010)
Thermal inhomogeneity: Electrical inhomogeneity:4.44 V
4.93 V
5.43 V
3.45 V
500 nm
3.95 V
0 50 100 150Current (μA)
40 nm
0
3.45 V
3.95 V
4.44 V
4.93 V
5.43 V
Local probe: conducting force microscope
RR ~102
V
VO2
n-type Si~ 2 nm SiOx
Cr/Aucoated tip
n-type Sicantilever: kc ~ 40 N/m
6
12 cm
built by Dr. Jeehoon Kim
0 5 100
50
100
150
Cur
rent
(uA
)
Bias (V)
187 nm
high Rinsulator
low Rmetal
IV map
0 5 10 150
50
100
150
200
Cur
rent
(uA
)
Bias (V)
Cur
rent
(uA
)
VO2
Si
Au
Bias (V)
Original data
100 nm
Topography
7
High-resolution IV map
0 5 10 150
50
100
150
200
Cur
rent
(uA
)
Bias (V)
straycapacitance
100 nm
Cur
rent
(uA
)
Bias (V)
VO2
Si
Au
New data (same sample) Topography
8
High-resolution IV map
0 5 10 150
50
100
150
200
Cur
rent
(uA
)
Bias (V) 100 nm
Cur
rent
(uA
)
Bias (V)
VO2
Si
Au
New data (same sample) Topography
9
High-resolution IV map
100 nm
TopographyV at transition
4.5 7.5 V = 187 13nmSi
VO2
TEM:
167 207 nm
=2 4 10 V/m
E-field
10But how can we get T at transition?Voltage E ( 10 V/m)
Poole-Frenkel effect
11
EF
3dπ
3d||
0.7 eV
1. VO2 at T=0, V=0:
EF
3dπ
3d||
0.7 eV
2. VO2 at room T, V=0:
Frenkel, Physical Review 54, 647 (1938)Simmons, Physical Review 155, 657 (1967)
EF
3dπ
3d||
0.7 eV
3. VO2 at room T, V>0:
= exp // /
fundamentalconstants high freq. dielectric
film thickness
temperature
12
Poole-Frenkel slope
0 5 10 150
50
100
150
200
Cur
rent
(uA
)
Bias (V)
ln = /4 1 / C
13
Poole-Frenkel slope
0 5 10 150
50
100
150
200
Cur
rent
(uA
)
Bias (V)
slope ≡ = 1.57ln = /4 1 / C
14
Poole-Frenkel slope
slope
1 3 /100 nm
slope ≡ 1.57ln /4 1 / C
Solve for local temperature
15
fundamentalconstants high freq. dielectric
film thickness
temperature
slope ≡ = 1.57ln = /4 1 / C
100 nm
high freq. dielectric
100 nm305 K
345 K
Temperature
slope
1 3 / 167 207 nm
film thickness
100 nm
16
100 nm305 K
345 K
TemperatureCompare E-field vs. Temperature
100 nm
-field
4 10 V/m
2 10 V/m
→ No special -field → = 332 K (c.f. = 339K)E ( 10 V/m) T (K)
Conclusions
17
1. Joule heating is primarily responsible for insulator-to-metal transition in VO2(2-terminal geometry on Si)
2. Poole-Frenkel mapping viaconducting AFM provides anew general scanning techniquefor nanoscale T measurementon insulating films/ /
T (K)
100 nm305 K
345 K
Temperature