PowerPoint-Pr

1
Scaling Projections for Resistive Switching
Memories
Rainer Waser
JARA-FIT _at_ FZJ Forschungszentrum Jülich RWTH
Aachen University
Outline

• Introduction - Features and Classification of
  Resistive RAM
• Electronic switching effects- including FTJ
  (brief)
• Fuse-Antifuse switching effect (brief)
• Ionic switching effects- cation migration redox
  systems(electrochem. metallization cells)
• Ionic switching effects anion migration redox
  systems

Center ofNanoelectronic Systems forInformation
Technology
Forschungszentrum Jülich
2
1 Introduction - Features and Classification
of Resistive RAM
- 2
3
Evolution of the memory density geometry aspects
R. Waser (ed.), Nanoelectronics and Information
Technology, 2nd ed. Wiley, 2005
- 3
4
RandomAccessMemories
Challenges

• Fundamentals
• Physics Chemistry

Resistive
Ferroelectric
? Superparaelectric Limit Interface Scaling
Effects
DRAM
Limit Dielectric Area
Mega-Bit Era
Giga-Bit Era
Tera-Bit Era
- 4
5
Basic Definitions of Resistive RAM
Electrical switching between ON(LRS) and OFF(HRS)
state
Operation

• write-Operation by large voltage pulses
  (typically with current compliance)
  
• read operation by small (sensing) voltage
  pulses

Polarity modes of RRAM
Unipolar (symmetrical) - URS
Bipolar (antisymmetrical) - BRS
I
RESET
CC
V
CC
SET
CC
Read
Two-terminal element between electrodes
Memory cell

• Active Matrix 6...10 F2 , 1 cell / 1 T
• Passive Matrix array size e.g. 8 x 8, approx
  4 F2

History

• many reports since the 1960s
• mainly binary oxides, mainly unipolar switching

- 5
6
Requirement of bipolar switching

• asymmetry of the system (e.g. different
  electrode materials)

Forming process

• electrical stress as a precondition for
  switching
• sometimes formation during first cycle(s)

Location of the switching event - In the
electrode area

• homogeneously distributed effect?
• effect confined to filaments?

Location of the switching event - Between the
electrodes

• along the entire path / in the middle area?
• at one of the interfaces?

- 6
7
Classification of the switching mechanisms
Resistive Switching
Thermaleffect
Electroniceffect
Ioniceffect
Cation migration - electrometallization
Charge trap Coulomb Barrier
Phase Change Effect (well known)
Anion migration - redox effect
Charge injection IMT transition effect
Fuse-Antifuseeffect
Ferroelectric Tunneling barrier
R. Waser and M. Aono, Nature Materials, 2007
- 7
8
3 Electronic Switching Effects - including
ferroelectric tunnel junctions
- 8
9
Electronic Switching Mechanisms
General Effect
Charge injection or movement of displacement
charges ? modification of the electrostatic
barrier, bipolar switching
Variants

1. Charge trap effects / Coulomb repulsionTrapping
   at an interface e. g. Taguchi Sensor effector
   at internal traps (compare Flash) e.g.
   polymer - embedded metal nanoclusters
2. Insulator-Metal Transition effectsCharge doping
   ? change of band structure
3. Ferroelectric tunneling barrier effects
   tunneling parallel / antiparallel FE
   polarization ? modification of the tunnel
   barrier(thickness or height)

Geometry
Homogeneous current density ? scaling limits (?)
- 9
10
Ferroelectric Tunnel Junction
Basic Effect
Different Tunneling currents parallel and
antiparallel to the ferroelectric polarization
(no lateral confinement required) References
Esaki (1968), Tsymbal, Kohlstedt et al, Science
(2006)
? No reliable experimental evidence yet !
Results
11
Projections Ferroelectric Tunnel Junction
Memory
State-of-knowledge
No experimental verification yet competition by
redox-based processes.
Scaling limits
worse than MTJ, because the ferroelectricity
fades below approx. 2 nm thickness (ab-initio
theory HRTEM experiment)
Speed
no inherent speed limits because polarization
reversal inferroelectrics lt 1ns
Energy dissipation
Recharging of a tiny FE capacitor (much smaller
than in FeRAM) low energy switching expected
Challenges
1. realization of a pronounced effect unlikely2.
expected scaling limits not favorable
12
2 Fuse-Antifuse Switching Effect
- 12
13
Fuse-Antifuse Switching Mechanism
Materials
MIM thin film stack with I transition metal
oxide showing a slight conductivity
e. g. Pt/NiO/Pt
SET process
Controlled dielectric breakdown e. g. by thermal
runaway? formation of a conducting
filament
RESET process
Thermal dissolution of the filament (fuse blow)
? disconnected filament
I. G. Baek et al. (Samsung Electronics), IEDM
2004
- 13
14
Temperature profile - Thermal effect assisting
other switching types?
FEM simulation (Ansys ) of metallic TiO filament
(3 nm) in TiO2 matrix
3 filaments
1 filament
390 K
1100 K
Relationship to other switching effects
Toggle between bipolar and unipolar switching has
been possibleby adjusting the current
compliancedemonstrated for TiO2 thin films
(Jeong et al. 2006) and CuTCNQ (Kever et al.
2006) High current compliance ? unipolar
fuse/antifuse switching
- 14
15
FEM Simulation of the RESET process
160 nm thick NiO film on n-Si with Au top
electrodes
U. Russo et al., IEDM 2007
- 15
16
Transition from the BRS to the URS mode
Pt/TiO2(27nm)/Pt stack, sputter deposited,
electroformated at 1mA for BRS operation

• Higher currentcompliance ( 3 mA) leads to
  transition of the switching mode (BRS ? URS)
• This transition to URS is irreversible (URS ?
  BRS)

LRS of a stable URS
Details see talk H. Schroeder D.-S. Jeong
(F6)and Jeong, Schroeder, Waser, APL89, 082909
(2006)
17
Projections Fuse-antifuse effect
State-of-knowledge
- thermo-chemical effect - filamentary nature
confirmed- nature of ON-state filament still
unknown (Ni? NiOxlt1? else?)
Scaling limits
equal or larger than filaments in redox
systems(perhaps larger because of the thermal
nature of the effect) 10 nm diameter ??
Speed
approx. 10 ns (speculative value)
Energy dissipation
Relatively high, because of thermal nature no
scaling expectedbecause of filamentary nature
material optimization will lead to reduction
Challenges
better understanding required in order to
optimize material and processing
18
4 Ionic switching effects - Cation-migration
induced redox systems
- 18
19
Cation Migration Redox Systems
Names
Electrochemical Metallization Cell (ECM) PMC
CBRAM
Working principle
ON-switching Reduction _at_ cathode ? Ag filament
formation Ag e ? Ag or
OFF-switching Oxidation _at_ anode Ag ? Ag e
or
M. Faraday (1834)
Required ingredients
Electrolyte Ag ion conductor Question
is the presence of Ag ions required? What
concentration?
Type of cations M/Mz in a moderate area
of the electrochemical potential series
high exchange current density (Butler-Volmer
equation) Examples Ag, Cu, Ni, (and few
more)
- 19
20
Electrochemical deposition/dissolution
Faradays law
Phase generation / dissolution rate
Growth speed of a (cylindrical) nanofilament
Example Ag filament of 10nm diameter at I 1 mA
- 20
21
Working Principle of the ECM Cell
Ag
Ag
Ag
Ag
Ag
Pt
Pt
Pt
Pt

• Oxidation of
• top electrode
• Ion migration

• Reduction at
• bottom electrode
• Electrodeposit
• formation

• Non-volatile
• conductive
• connection

• Dissolution of the conductive path under reverse
  bias

22
Prototypical Electrochemical Cell Pt / H2O /
Ag
I-V bipolar switching
Time evolution during switching on formation of
a Ag dendrite tree at the Pt electrode
Ag
Ag
Ag
Pt
Pt
Pt
X. Guo, C. Schindler, S. Menzel, R. Waser, APL
(2007)
- 22
23
Example Ag electrochemical GeSex cells (M.
Kozicki, 1997)
Electrolyte GeSex (or GeSx ) dissolution of
Ag by photoassisted or thermally assisted
oxidation of adjacent Ag phase high driving
force in case of excess Se in GeSex (as
polyanions). Structure Ag2Se rich
nanocrystals in amorphous GeSex matrix What is
special?? stable reservoir of Ag ions
(buffer)? high Ag mobility activation
energy (easy migration path in
amorphous matrix?)
Ag
50nm GeSe on W
GeSexAg2Se
Pt or W
- 23
24
Special case the Atomic Switch
Quantized conductance atomic switch tunneling
gap as an electrode
K. Terabe, M. Aono, et al.,Nature (2005)
- 24
25
Aono group
Cu/Cu2S X-bar
Qimonda
Mbit demonstrator
- 25
26
Low-Write Currents in Ag electrochemical GeSex
cells


Ag/Ag-Ge-Se/Pt cells fabricated in vias of a
Si3N4 layer. Via diameter 2.5 ?m by
photolithogrpahy.
I-V curves with current compliance set to 1 nA
(C. Schindler, R. Waser, et al., NVMTS 07)
Latest results (unpublished) 10 pA !
- 26
27
Prospect of Multi-Bit Data Storage
2.5 µm via
2.5 µm via

• The write current determines the ON resistance
  and erase current.
• By varying the write current over 5 orders of
  magnitude, the ON state
• is variable over the same range increasing the
  opportunity for multi-bit data storage.

C. Schindler et al., NVMTS 07
28
Cell Size Dependence
Although the cross sectional area varies by a
factor of 400, the ON resistance does not vary
by a factor of more than 5 between the smallest
and the largest cell.
The ON state is cell size-independent.
C. Schindler et al., NVMTS 07
29
Unconventional electrolyte SiO2
For comparison
Sputtered SiO2 thin film
Ag-Ge-Se
Von 400 mV
Voff -100 mV
Iwrite 25 µA
Ierase -12 µA
-100 mV lt Vread lt 400 mV
Similar switching characteristics as observed in
Ag-Ge-Se!
C. Schindler, M. Kozicki, R.Waser, submitted
30
Modeling MD approach to Browian Dynamics
40 nm
S. Menzel et al., to be published
No
Yes
Yes
No
31
Projections ECM effect
State-of-knowledge
- reasonably well understood - filamentary
nature confirmed- some open questions with
respect to nanoclusters in GeSe OFF switching
Scaling limits
Very thin filaments conceivable (lt 2 nm ??) i.
e. min. F 5 nm ??
Speed
lt 30 ns ? (perhaps even faster compare to
fuse-antifuse)
Energy dissipation
ON switching electrochem. reduction of Ag
within one filament (or fraction of) migration
energy (Faraday transfer number 1
assumed)OFF switching ??
Challenges
improvement of the voltage speed bargain
32
5 Ionic switching effects - Anion-migration
induced redox systems
- 32
33
Capacitor-like structure with ? Cr-doped SrZrO3
thin films ? (Ba,Sr)TiO3 thin films? SrTiO3
Single crystals as resistive element 1
Redox-based resistive switching type
Examples of Resistive Switching in Oxides
300 nm
SrZrO3(0.2 Cr)

• Characteristics
• after forming process reversible
  switching between stable impedance states
• non-volatile
• non-destructive read-out
• multilevel memory

- 33
34
High temperature defect equilibria
Example SrTiO31at La, T1100C
Schottky equilibrium (for SrO lattice)
D
Redox equilibrium
law of mass action
Preconditions

• local electroneutrality
• homogeneous material

- 34
35
Electrical characteristics of dislocationsPrefer
ential conductivity paths in SrTiO3 single
crystals
5mm
NC-AFM of etched (100) surface of strontium
titanate(estimated density of dislocations of
4109/cm2 )
c-AFM mapping of local conductivity of (100)
surface of thermally reduced strontium titanate
-gt hot spots density of strong hot spots
51010/cm2density of weak hot spots
51011/cm2
K. Szot, 1999, 2004
- 35
36
Thermal preformation by reduction
annealingconductive Tip AFM Mapping types of
I-V Characteristics
10
SrTiO3 s.c. thermally reduced at 850 C, pO2
10-20 bar
Current (nA)
1.0
1000
0.1
metallic
I/V1
semicond.
100
insulating
0.01
10
I/V2
(nA)
1
I/V2
0.1
I/V1
50nm
0.01
0.001
-0.5
0.0
0.5
Applied bias (V)
K. Szot, W.Speier, G. Bihlmeyer, R. Waser,
Nature Materials, 2006
- 36
37
a
c
non-metallic
off
metallic
1000
b
d
n15
800
n7
n6
600
on
n5
Current (nA)
400
n4
n3
n2
200
n1
0
1
2
3
4
5
0
Applied bias (V)
K. Szot et al., Nature Materials, 2006
- 37
38
Edge dislocations in SrTiO3 crystal (stacking
fault)
Jia et al PRL.(2006)
39
Extended defects after reduction
Formation of localized metallically conducting
sub-oxides by electroreduction
- 39
40
Redox Reactions at the Electrodes
- Interconnected network of extended defects-
Switching ON oxygen vacancy accumulation near
the surface conduction through the Ti(4-x)
sublattice
- 40
41
Learning from lateral cells
K. Szot, etc. Nature Materials, 2006
SrTiO3
Optical micrograph and CAFM (above)and Cr K-edge
XANES (right) mapping
M. Janousch et al. Adv. Mater. 19, 2232 (2007)
42
K. Szot, to be published
43
LC-AFM characterization of epi-STO (10nm) / SRO /
STO
K. Szot, R. Dittmann, R. Waser, rrl-pss, 2007
- 43
44
LC-AFM write/erase processes onepi-STO (10nm) /
SRO / STO
Write /erasepatterning
K. Szot et al., rrl-pss, 2007
- 44
45
Temperature dependence
K. Szot, R. Dittmann, et al., rrl-pss, 2007
- 45
46
Current questions
Switching model Formation and resistive
switchingof redox-active dislocations
Fast transport tracks ?
Electronic chargeinjection mode (Schottky
emission, FN ?)
3-D network ofextended defects?
Nanoscale effects ?
Phase formation orfrozen kinetics?
K. Szot, 2004
- 46
47
Switching model former results
T. Baiatu, K. H. Härdtl, R. Waser, J. Am. Cer.
Soc. (1990)
48
Switching model recent results
D.-S. Jeong et al., to be published
Pt/TiO2(27nm)/Pt stack, sputter deposited,
electroformated at 1mA for BRS operation
49
Projections redox-type memory effect
State-of-knowledge
- redox mechanism suggested during formation -
details of the switching mechanism need to be
studied (thermal assisted?)- nature of ON-state
filament conducting (sub-)oxide?
Scaling limits
estimated filament diameter 2 10 nm ?
Speed
lt 30 ns ? (perhaps even faster compare to
fuse-antifuse)literature reports of values of
approx. 10 ns
Energy dissipation
perhaps similar to ECM possibly thermal
assisted, i.e. additional dissipation
Challenges
Much better understanding required in order to
optimize material and processing (and to make
more precise projections)
50
Thank You!