|Views: 95|Likes: 0

Published by Blaise_Mouttet

Physics model for some forms of ReRAM

Physics model for some forms of ReRAM

See more

See less

June 25, 2011 (ver.3)

Dynamic Systems Model for Ionic Mem-Resistorsbased on Harmonic Oscillation

Blaise Mouttet

Abstract

—

Memristive system models have previouslybeen proposed to describe ionic memory resistors.However, these models neglect the mass of ions andrepulsive forces between ions and are not wellformulated in terms of semiconductor and ionicphysics. This article proposes an alternative dynamicsystems model in which the system state is derivedfrom a second order differential equation in the formof a driven damped harmonic oscillator. Applicationis made to Schottky and tunneling barriers.

Keywords- mem-resistor, non-linear dynamic systems

I.

I

NTRODUCTION

In 1971 t

he “memristor”

had been proposed byLeon

Chua as a “missing” fundamental circuit

element exhibiting charge-dependent resistance [1]and was

recently claimed to be “found” by

researchers of HPLabs in the form of TiO

2-x

thinfilms [2]. In actuality there were earlier physicalexamples of charge-dependent memory resistors [3]and it had been demonstrate

d since the 1960’s that

TiO

2

thin films exhibit a memory resistance effect[4]. Commercial interest in memory resistors hasgradually been

increasing since the 1990’s and

hasbeen under development in numerous forms byAMD, Axon Technologies, Micron Technologies,Samsung, Sharp, Unity Semiconductor, and othercompanies with applications directed toward non-volatile memory [5]. The recent efforts of HPLabs tomarket

Chua’s

“memristor” as a

newly discovered

“4

th

fundamental circuit element”

are arguably anattempt to take credit for the work of these othercompanies rather than representative of any newdiscovery [6].The researchers at HPLabs have made someattempts to model their TiO

2

memory resistor usingionic drift/diffusion of oxygen vacancies [7]formulated in terms of memristive systems [8].However, the memristor ionic drift model is atminimum incomplete (if not entirely incorrect)since:a) it neglects hysteretic capacitive effects that wouldoccur as ions move in a thin semiconductor film,b) it neglects the mass and inertial effect of themobile ions,c) it neglects repulsive forces between ions, andd) it neglects dynamic effects of the built-in voltage.In order to incorporate the above effects thememristor model would need to be modified (orcompletely replaced) with a different dynamicsystems framework. There have been somesuggestions of broader dynamic systems modelssuch as mem-admittance systems (Fig. 1b) whichincorporate memory capacitive effects found incertain perovskite and nanocrystal films [9]. Inaddition mem-transistor systems (Fig. 1c) have beensuggested to model 3-terminal memory devices suchas the Widrow-Hoff memistor or synaptic floatinggate memory cells [10]. Yet another conceivabletype of mem-electronic system is mem-activesystems (Fig.1d) in which an active source of voltage or current is included in a resistance memorycell and exhibits independent memory effects. Oneexample of this type of mem-electronic system maybe found in a thin film having a charge storingmedium providing non-volatile voltage storage [11].Another example of mem-active systems may berechargeable batteries which are resistive whenacting as an electrolytic cell but serve as a powersource when acting as a galvanic cell.The present article presents a new dynamic modelwhich incorporates aspects of these different mem-electronic systems to properly address the physics of nano-ionic thin films.

June 25, 2011 (ver.3)

Memristive Systems (Fig.1a):

(

) =

,

,

(

)

(

)

=

(

,

,

)

Mem-Admittance System (Fig.1b):

(

) =

,

,

(

) +

,

,

(

)

(

)

=

(

,

,

)

Mem-Transistor System (Fig. 1c):

1

(

) =

,

1

,

2

2

(

) =

,

1

,

2

(

)

=

(

,

1

,

2

)

Mem-Active System (Fig. 1d):

(

) =

,

,

(

(

)

−

,

,

)

(

)

=

(

,

,

)

II.

IONIC DYNAMICS IN THIN FILMS

The assumption by [8] that the dynamics of ionicmotion can be treated solely by drift-diffusionequations is flawed. It neglects the inertia of ions ina solid and the repulsive forces between ions. Abetter model for ion dynamics in a solid is providedby Bisquert et al. [12]. This model expresses ionicdynamics

in terms of Newton’s 2

nd

law of motion(5) relating the acceleration

d

2

x/dt

2

of an ion havingeffective mass

m

ion

to the sum of the forces

F

i

actingon it.

2

2

=

When an external electric field is applied to anionic thin film sandwiched between two electrodesthere are three principle forces which act on the ion.The first force (

F

c

) is due to collisions of the ion as itmoves through the thin film. The product of thisforce and the average time between collisions

c

canbe equated to the change in the ion momentum.

=

−

The second force (

F

r

) is due to the internalelectric field (

E

r

) produced by other ions in the thinfilm. Assuming an equilibrium position

x

0

where theforces of the other ions all balance and an ion donordensity of

N

d

this force may be expressed usingGauss

’s

Law as (7).

=

=

−

(

)

2

0

(

−

0

)

where

0

is the vacuum permittivity,

r

is the relativepermittivity,

e

is the unit charge, and

z

is the valenceof the ion.The third force

F

a

is that produced by an appliedvoltage bias. According to the ion hopping modelreviewed in [13] the ionic current (i.e.

dx/dt

)dependence on the external field is given by a sinhfunction which is approximated by an exponential athigh fields and a linear function at low fields. It maythus be inferred from (6) that the ionic driving forcewould take the form

= (

) 2

−

2

where

a

is the jump distance of the ions betweenhopping,

is a frequency factor,

W

a

is the energybarrier,

k

is Boltzmann’s constant,

T

is absolutetemperature, and

E

a

is the electric field in the thinfilm. Combining (5)-(8) produces:

2

2

+

+

(

)

2

0

−

0

=(

) 2

−

2

For cases in which the thin film is bounded betweenelectrodes separated by a distance

D

and in whichthe ions are unable to penetrate the electrode aboundary condition

0≤x(t)≤

D

may be added.In order to analyze the collective motion of anaggregate number of ions in the thin films it will benecessary to modify (9). Fig. 2a illustrates a uniformdistribution of ions within a region of a thin filmhaving thickness D. If there is sufficient ionicconductivity in the medium the ions will graduallyredistribute themselves due to electrostatic repulsion

June 25, 2011 (ver.3)

leading to Fig. 2b. However the total number of ionsshould not change due to conservation of charge

1

.

0

=

0

As a voltage is applied to the thin film as in Fig. 2cthe mean value of the ionic distribution drifts and isindependent of redistribution of the collective ions.In light of the above discussion (9) may be rewrittenin terms of the aggregate motion of the ions where

x

d

(t)

is the dynamic mean of the actual positions of the ions and

x

0

(t)

is the dynamic mean equilibrium

2

position of the ions.

2

(

)

2

+

(

)

+

2

0

−

0

(

)

=

2

−

2

This may be easily simplified to:

2

(

)

2

+

1

(

)

+

2

(

)

0

(1

−

0

(

)

(

)

)= (

2

)

−

(

)2

Further simplification of (12) may be achieved inthe case when the aggregated ions are adjacent theleft or right electrode. In Fig. 3a-c a dynamicevolution of the ion distribution is illustrated in asimilar manner as in Fig. 2 in which the ions areaccumulated at a boundary. In this case it is possibleto relate the dynamic mean

x

d

(t)

with the dynamicwidth

x

d0

(t)

. For a uniform ion distribution asillustrated

x

d

(t)

=

x

d0

(t)/2

. The dynamic equation(12) may now be further simplified using (10) as:

1

It is notable that in the special case of solid electrolyte thinfilms ions are temporarily produced from an electrochemicallyactive electrode and are critical to the formation of metalfilaments important to resistance switching. This case will bediscussed in section III in the sub-section on active junctions.

2

The concept of “dynamic mean equilibrium” may seemconfusing to some. It is “equilibrium” in the sense it is the

mean position of the ions in steady state when the forcingfunction is zero. It is dynamic in the sense that a sufficientlyhigh forcing function will change its value and transientdynamics will still be in play.

2

(

)

2

+

1

(

)

+

2

0

2

0

(1

−

0

(

)

(

)

)= (

2

)

−

(

)2

Still further simplifications can be made under theassumptions thata) the applied field E

a

(t) is sufficiently small that themean equilibrium is time-independent(

x

0

(t)

=

x

0

=x

d0

/2

) and the sinh function can beapproximated as linear,b) the variation

x

d

(t)

is small compared to

x

0

,

∆

=

−

0

≪

0

Under these conditions (13) may be approximatedas:

2

∆

(

)

2

+

1

∆

(

)

+

2

0

∆

(

)= (

2

)

−

(

)

Finally we arrive at a tractable form! This is of course the familiar driven damped harmonicoscillator differential equation.

2

2

+ 2

0

+

02

=

(

)

We can immediately obtain a potentially usefulresult in terms of the resonant angular frequency

0

.

0

=

(

)

2

0

By using (17) the effective mass of ions or oxygenvacancies in mem-resistor thin films might be ableto be estimated from the frequency at whichhysteresis effects are at a maximum.III.

COUPLING IONIC AND ELECTRONIC DYNAMICS

a) Dynamic Schottky junction

The behavior of electrons in semiconductor thinfilms is well studied and reviewed such as in [14].