Bias dependent specic contact resistance of phase change material to metal contacts

SAFE WORKSHOP, 2010 1

Bias dependent specific contact resistance of phase

change material to metal contacts

Deepu Roy, Micha in ’t Zandt, and Rob Wolters

Abstract—Knowledge of contact resistance of phase change

materials (PCM) to metal electrodes is important for scaling, device modeling and optimization of phase change random access memory (PCRAM) cells. In this article, we report the systematic determination of the specific contact resistance (ρc) with voltage

bias for doped Sb2Te to TiW metal electrodes. These data are

reported for both the amorphous and the crystalline state of the PCM.

Index Terms—PCM, specific contact resistance, Cross Bridge

Kelvin Resistor(CBKR), bias dependence.

I. INTRODUCTION

E

LECTRICAL contacts are an essential part of a device or a circuit for its operation. These contacts are character-ized by the contact resistance, which is expressed in terms of specific contact resistance (ρc). ρc is defined as the derivative

of voltage with respect to current density at the contact at zero bias[1]. For a line cell the resistance of the current path includes the resistance of the switching part of the line (RL)

and the resistance of the crystalline PCM contacts (Rc) [2].

In the ovonic type cells the PCM to electrode contact can be in the amorphous or crystalline state. In reality, during the operation of the device there will be a bias/voltage drop at the contacts. Depending on the conduction mechanism at the contact this voltage drop at the contact results in a deviation of ρc compared to the value at zero bias.

In this article this bias dependence of ρc is characterized

for phase change material (PCM) to TiW contacts. This is per-formed by current-voltage (I-V) resistance measurements (Rk)

on Cross Bridge Kelvin Resistance (CBKR) test structures as shown in Fig.1.

II. EXPERIMENT

To fabricate PCM to TiW CBKR structures, 50 nm TiW is deposited on a silicon wafer and patterned to form the bottom electrode layer at the contact. The top layer is 20 nm PCM and the contact area between the two layers is defined by contact opened in a 50 nm PECVD SiO2layer. To measure the PCM

to metal contact resistance, a current is forced from the metal

Manuscript received October 11, 2010. This research was carried out under the project number MC3.07298 in the framework of the Research Program of the Materials innovation institute M2i (www.m2i.nl).

Deepu Roy and Micha in ’t Zandt are with the NXP-TSMC Research Cen-ter, NXP Semiconductors, High Tech Campus-4, 5656AE, WAGp 02, Eind-hoven, The Netherlands. (phone:0031-4027-26872; fax:0031-4027-43352; e-mail:deepu.roy@nxp.com).

Rob Wolters is with the NXP-TSMC Research Center, NXP Semiconduc-tors, High Tech Campus-4, 5656AE, Eindhoven, The Netherlands, and also with MESA+ Institute for Nanotechnology, University of Twente, Enschede, The Netherlands.




































 
 
 
 



 



 



 



 



 



 



 



 





W

x

W

y

L

d

Metal

PCM



 



Contact

1

3

2

4

Fig. 1. Coss Bridge Kelvin Resistor structure showing upper (PCM) layer, lower (metal) layer and the contact structure (δ, L).

to PCM (1 to 3) and the voltage is measured orthogonal to the direction of current flow (2 and 4) using a semiconductor parameter analyzer. This allows measurement of the average voltage at the contact[3] and the voltage drop due to current spreading in the δ. The derivative of this I-V curve gives the measured resistance at the contact (Rk). From Rk the specific

contact resistance (ρc) of the contact interface only is extracted

by using a 2D analytical model excluding the resistance in the δregion [4]: Rk = ρc A + 4Rshδ 2 3WxWy  1 + δ 2(Wx− δ  (1) To be able to measure the contact resistance to amorphous PCM these structures should be processed below the crystal-lization temperature of PCM. The amorphous to crystalline transition temperature for this doped Sb2Te is 160◦C [5][2].

So the CBKR structures have been processed with a thermal budget of maximum 120◦C. Contact resistance measurements

are performed on these fabricated structures (PCM amorphous) and after annealing (PCM crystalline) at temperatures of 250

◦C for 5 min in N₂ ambient. Samples annealed above 250◦C

were protected against oxidation and evaporation by capping with a 500 nm PECVD SiO2 layer deposited at 250◦C.

III. RESULTS ANDDISCUSSIONS

Fig. 2 shows the I-V characteristics of CBKR structures with PCM in the crystalline and in the amorphous state. Crystalline PCM contacts show a linear current voltage (I-V) characteristic unlike the amorphous PCM. Rkwith voltage bias is calculated

-SAFE WORKSHOP, 2010 2

-500.0µ

0.0

500.0µ

-30.0m

0.0

30.0m

V

o

lta

g

e

in

V

Current in A

-2.0µ -1.0µ

0.0

1.0µ

2.0µ

-1.4

-0.7

0.0

0.7

1.4

V

o

lta

g

e

in

V

Current in A

(a) (b)

Fig. 2. I-V characteristics for the CBKR structures with (a) PCM in the crystalline state and (b) PCM in the amorphous state.

Ω Ω

(a) (b)

Fig. 3. Change in ρcwith applied bias voltage. (a) Crystalline PCM, (b) amorphous PCM.

form the derivative of the I-V curve, from which the ρc is

calculated. The extracted ρc for doped Sb2Te to TiW contacts

in the amorphous and crystalline state for positive and negative bias voltage across the contact is shown in Fig.3.

The crystalline state of the PCM results in a linear I-V of the contacts and the Rk and ρc remains unchanged with

the applied bias voltage. From Hall measurements we have derived a carrier concentration of approximately 21021

cm−3

for crystalline doped Sb2Te phase change material. Given

this carrier concentration this PCM behaves like a metal to silicon contact showing ohmic behavior with a linear I-V. Here the tunneling process dominates and the value of ρc is

hardly dependent on the work function of the metal [1]. In the crystalline state the ρc value for PCM to metal contacts is at

least two orders of magnitude higher than common metal to metal contacts [6].

In the case of amorphous PCM to metal contacts, the I-V characteristics is non-linear and here the extracted ρc

decreases with the voltage bias at the contact (Fig. 3(b)).The dependence of ρcon the sign and magnitude of the applied bias

voltage is determined to investigate the effect on the switching characteristics of PCRAM cells. This variation in ρc with

bias should be taken into account for contact characterization and device modeling. The bias dependence on reset current in PCRAM cells has been reported in the literature [7]. Our

measurements show a stronger dependence of ρcwith negative

bias compared to positive bias. A similar behavior in ρc

values is also reported for metal semiconductors contacts in the case of thermionic field emission. The ρc values are not

symmetrical with the bias voltage and the resistance peak is not occurring at zero bias [8].

IV. CONCLUSION

In this article we report the bias dependence of specific contact resistance for TiW to doped Sb2Te in the amorphous

and crystalline state of using dedicated CBKR measurement structures. The electrical properties of crystalline PCM to metal contacts show similarities with the highly doped silicon to metal contacts. In the amorphous state ρc values show a

strong dependence with sign and the magnitude of applied voltage bias. The knowledge of ρc is critical for the scaling,

device modeling and optimization of the switching character-istics for PCRAM cells.

REFERENCES

[1] S. M Sze, Physics of semiconductor devices, 2nd ed. Harlow, Wiley-Interscience, New York, 1981,ch. 5.

[2] M. H. R. Lankhorst, B. W. S. M. M. Ketelaars, and R. A. M. Wolters, ”Low-cost and nanoscale non-volatile memory concept for future silicon chips,” Nature Materials, vol. 4, pp. 347-352, April 2005.

-SAFE WORKSHOP, 2010 3

[3] S. J Proctor and L. W Linholm, ”A direct measurement of interfacial contact resistance,” IEEE Electron Device Lett., vol. 3, no. 10, pp. 294-296, Oct. 1982.

[4] T. A. Schreyer, K. C. Saraswat, ”A two dimensional analytical model of the cross bridge Kelvin resistor”, IEEE Electron Device Lett., vol-7, 1986, pp.661.

[5] I. Friedrich, V. Weidenhof, N. Njoroge, P. Franz and M. Wuttig, ”Structural transformations of Ge2Sb2Te5 films studied by electrical resistance measurements,” J. Appl. Phys., vol. 87, no.9, pp. 4130-4134, May 2000.

[6] N. Stavitski, J. H. Klootwijk, H. W. Van Zeijl, A. Y. Kovalgin and R. A. M. Wolters, ”Cross bridge kelvin resistor structure for reliable measurement of low contact resistances and contact interface character-ization,” IEEE Trans. Semiconductor manufacturing, vol. 22, no.1, pp. 146-152, Feb 2009.

[7] S. Lee, J. Jeong, T. S. Lee, W. M. Kim, and B. Cheong, ”Bias polarity dependence of a Phase change memory with Ge-doped SbTe: A method for multilevel programming,” Appl. Phys. Lett.,” 92, 243507 (2008). [8] C. R. Crowell and V. L. Reideout, ”Thermionic field resistance maxima in metal semiconductor (schottky) barriers,” Appl. Phys. Lett., vol. 14, no. 3, Feb. 1969.