United Arab Emirates UniversityScholarworks@UAEU

Theses Electronic Theses and Dissertations

2008

Preparation and Characterization of (GeSbTe))Semiconductors for Phase-Change Optical StorageHessa Ali Humeid Al-Shamisi

Follow this and additional works at: https://scholarworks.uaeu.ac.ae/all_theses

Part of the Materials Science and Engineering Commons

This Thesis is brought to you for free and open access by the Electronic Theses and Dissertations at Scholarworks@UAEU. It has been accepted forinclusion in Theses by an authorized administrator of Scholarworks@UAEU. For more information, please contact fadl.musa@uaeu.ac.ae.

Recommended CitationHumeid Al-Shamisi, Hessa Ali, "Preparation and Characterization of (GeSbTe)) Semiconductors for Phase-Change Optical Storage"(2008). Theses. 293.https://scholarworks.uaeu.ac.ae/all_theses/293

United Arab Emirates University Deanship of G raduate Studies

PREPARATION AND CHARACTERIZATION OF ( GeSbTe) SEMICONDUCTORS FOR PHASE-CHANGE OPTICAL STORAGE

By Hessa Ali Humeid AI-Shamisi

Supervised by Dr. Hassan Ghamlouche

Physics Department Faculty of Science, UAEU

Dr. Nase r Qamhieh Physics Department Faculty of Science, UAEU

Dr. Saleh Thaker Physics Department

Faculty of Science, UAEU

A Thesis Submitted to the Deanship of G raduate Studies in Partial F u lfil lment of the Require ments fo r Degree of Master of Science in

Mate rials Science and Engineering

2008

Acknowledgments

First, always and foremost, I thank almighty Allah for rus blessing and for

providing me the capabi lity to successfully complete trus work.

I would l ike to express my heartiest thanks and sincere gratitude to my supervisors, Dr.

Hassan Ghamlouche, Dr. Naser Qamrueh and Dr. Saleh Thaker for their guidance, support,

encouragement, patience, unl imited assistance, cri t ical reviewing and providing me the

required facilities during the work . 1 feel both honored and privileged to work with such a

group.

Many thanks to Mr. Mohammed Elshaer ( lab technician in the physics department)

and Mr. E sam Attia (technician in the Central Lab Unit), for their help in preparing some

experiments and analyzing the samples. Special thanks to Ms. Sadeeqa Ahmed, from the

physics department, for her assistance and support .

Special appreciation is expressed to my col leagues and friends in UAE University,

especially in the physics department .

No words can ever express my great grat itude and special thanks to my great

family, especially my parents for t heir great support and my husband for his patience and

great help throughout my study.

II

Abstract

In the last decade, the phase change optical recording based upon chalcogenide

glasses has advanced remarkably at ful l -scale commercial appl icat ions in the form of

rewritable compact disks (CD-RWs) and d igital video disks ( DVD-RWs). The principle of

phase-change optical recording is based on a thermal ly induced reversible transformation

between amorphous and crystal l ine phases. The search for recording material s possessing

fast crystall izat ion rate and long data retention has been pursued. Research resu lts have

revealed that Ge-Sb-Te compounds exhibit fast crystal l ization, which has been explained in

terms of their single-phase structure.

In the present work, electrical properties of Ge-Sb-Te thin fi lms using I mpedance

Spectroscopy, Current-Voltage, and Capacitance-Voltage techniques is presented . First, we

studied the effect of the electrical contacts, deposited on top of Ge-Sb-Te fi lms, on the

measured electrical signal ( I mpedance Spectroscopy). To do this, three different

configurations of the electrical contacts were prepared . The results show that the electrical

contacts have no effect when the silver paste is placed on the gold electrodes away from the

thin fi lm under test. This guarantees no interaction between the si lver paste and the fi l m

under test during the heat ing process. Also the resu lts show that the order of depositing

layers ( material and the gold) is not important and what matters is the locat ion of the si lver

paste on the gold electrodes.

Second, the Impedance Spectroscopy and Current-Voltage data shows similar

resistance and relaxat ion frequency temperature dependence with activation energies 0.36

and 0.39 eV, respect ively. Therefore the corresponding energy gap is est imated to be 0.72

eV and 0 .78 eY. This result agrees with the optical gap (0 . 7 eV) reported in l iterature.

This agreement confirms that Ge2Sb2 Te5 belongs to the chalcogenide fami ly. I n addition,

"

III

the I -V curves show that the amorphous-crystal l ine transit ion temperature (Tc is around

1 3 5 °C

Third, the Capacitance-Voltage measurements were performed on amorphous

samples of I no.3Ge2 Sb2 Te5 and Ge2Sb2Tes thin films in order to study the effect of doping

on the electrical properties of Ge2Sb2 Te5 . The results show that inducing I ndium to

Ge2Sb2 Tes decrease the capacitance of the sample as the bias voltage increases for the

whole range of temperatures. However, the same behavior is observed for undoped sample

(Ge2 Sb2 Te5) only at temperatures close to the amorphous-crystal l ine transition temperature

and at high appl ied bias voltages (� 1 5 volts) . The behavior is attributed to the increase in

electron concentration as the bias voltage and temperature increase. For l no.3Ge2Sb2 Te5

sample, the indium can be considered as an additional source of free electrons. This would

contribute to a further and significant decrease in the capacitance of I no.3Ge2Sb2 Te5 sample,

and eventual ly the capacitance becomes negat ive. The nonlinearity in the capacitance and

conductivity could be related to the nucleat ion mechanism as the temperature becomes

clo e to the amorphous-crystal l ine transit ion temperature. Final ly in the conclusion 1

proposed some study to be conducted on Ge2Sb2 Te5 films.

o

IV

TABLE OF CONTENTS

LIST O F FIGUR ES . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . VII

LIST O F TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IX

CHAPTER I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

INTRO DUCTION ........................................................................................................... 1

CHAPT ER II . . . . . . . . . . . . . . . . . . . . . . .. . . . . .. . . . . . . . . . . . . . . . . . . . . . .............. . . . . . . . . . . . . . . . . . . . ....................... 5

THEORETICAL BACKGROUN D . . . . . . . . . . . . . . . . . . . . . . . . .......... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 . 1 I ntroduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 .2 . Glass formation and the amorphous phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2 . 3 P hase-change materia ls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . '" . . . . . . . . . . . . . . . . . . . . . . . . . 9

2 . 3 . 1 Chalcogenides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 0

2.4 Density of states in amorphous semiconductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 3

2 5 Conduction in amorphous semiconductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 6

2 . 5 . 1 Extended state conduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 7

2 . 5 . 2 Hopping conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 9

2 . 6 AC conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

CHAPT ERll . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

SAMPLE PREPARATION AND EXPERIMENTAL TECHNIQUE S ....................... 24

3 . 1 Sample preparation . . . ... . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . 26

3 . 1 . 1 Preparation of bulk materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . · . . . . . . . . · · · · · . . . . . . . . 26

3 1 . 2 Thermal Evaporation Technique (TE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

3 . 1 . 3 Sputtering Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . · . . . . . . . . . . . . · . . . . . . . . . . . . . . . . . . · 28

3 . 1 . 4 Deposit ion of electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 28 o

v

3 2 Mea uring Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3 2 . 1 The impedance pectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3 .2 .2 Current-Voltage technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1

3 .2 . 3 Capacitance-Voltage technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1

C HAPT ER IV ......................................................................................................... 32

RESULTS AND DISCUSSIONS ................ ....................... ............................................ 32

4. 1 DC Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

4 .2 I mpedance spectroscopy measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

4 3 Capacitance-Voltage measurements of Ge2Sb2Tes and Ino.3Ge2Sb2Tes . . . . . . . . . . . . . . . . . . 42

CONCLUSION ........................................ ........................................ ......................... 5 1

RE FER ENCES ................................................. . .......... ............................................. 53

..

VI

LIST OF FIGURES

Fig. 1 Sc�ematic of the temperature-t ime �rofi les associated with recording ( left panel ) and erasure (nght panel) of amorphous marks In a crystal l ine layer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Fig. 2 Volume as a function of temperature for a l iquid, a glass and a crystal . . . . . . . . . . . . . . 7 Fig. 3 Schematic representations of a crystal J ine and an amorphous structure . . . . . . . . . . . . . 8 Fig. 4 Composit ion triangle of Sb, Te and Ge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Fig. 5 The periodic table of elements. The chalcogens appear in the sixth group of the table. . .. . . . . . . . .. . . . . . . . . . . . . .. . . . . . . . . . . .. . .. . . . . . ... . . . . . . . . . . . . . . . . . . . . . .. . . 1 1 Fig. 6 Two crystal l ine phases of Ge2Sb2 Tes . On the right hand side the pure fcc crystal, and in left hand side the pure hcp crystal .. ................................................. 1 2 Fig. 7 Density of states in amorphous semiconductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 5 (a) CFO model . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 5 (b) With bands of donor (E) and acceptor (Es) arisi ng from the same defect (Mott and Davis model ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 5 Fig. 8 I l lustration of the density of states near the band edge, together with the carrier mobi l ity �(E) the occupation probabi l ity fCE) and the conductivity cr(E) . . . . . . . . . . . . . . . . . 1 6 Fig. 9 Schematic i l lustration of the temperature dependence of conductivity expected for an amorphous semiconductor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Fig. 1 0 I mpedance vector representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1 Fig. 1 1 Impedance p lots for series and parallel combinations of Rand C. . . . . . . . . . . . . . . . . 22 Fig. 12 X-ray diffract ion spectrum of Ge2Sb2 Tes film as deposited . . . . . . . . . . . . . . . . . . . . . . . 25 Fig. 13 EDX of Ge2Sb2 Tes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Fig. 14 Schematic i l lustration of experi mental arrangement used for producing amorphous Ge2Sb2 Tes bulk samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Fig. 15 Schematic diagrams of the electrical contacts . . . · . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Fig. 1 6 Connections for an impedance measurement . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Fig. 17 1-Y curves for Ge2Sb2 Tes ( Sample 4) films at different temperatures a) in the amorphous phase and b) around and after the transit ion to the crystal l ine phase . . . . . . . . . . 33 Fig. 18 Resistance of Ge2Sb2Tes fi l m ( Sampl e 4 ) as a function of temperature. The dashed l ine is the l inear fit of the resistance in the amorphous state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Fig. 1 9 Resistance measurements as a funct ion of temperature for samples prepared at Aachen, Leuven and U AE. The amorphous-crystal l ine transit ion temperatures Tc are 145 DC, 1 40DC and 1 3 5DC, respect ively . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Fig. 20 I mpedance spectra for a Ge2Sb2Tes ( Sample 4 ) fi lm at different temperatures . .. . 37 Fig. 2 1 Frequency at the peak of the semicircle as a function of 1 03fT (K1). The slope of the dashed l ine determines t he activation energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 7 Fig.22 I maginary impedance a s a function of frequency for sample 1 at different temperatures in the amorphous state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 9 Fig.23 I maginary impedan ce a s a function o f frequ ency for sample 2 at different temperatu res in the amorphous state . . . . . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Fig. 24 Imaginary impedance as a function of frequency for sample 3 at different .......... .

temperatures in the amorphous state . . .. . .. . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . 40 Fig. 25 Imaginary impedance as a function of frequency for sample 4 at different. ......... .

temperatures in the amorphous state . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1 Fig. 26 I mpedance spectra for samples 1 and 4 at T = 80 DC .............................. 4 1 Fig. 27(a) Capacitance variation of sample S 5 as a function of appl ied bias voltage (-20V to +20Y) for temperatures: 45 , 65, 85, 1 05 and 1 1 5 °C. The frequency is 1 00 kHz . . . . . . 43

o

Vll

Fig. 27(b) Capacitance variation of sample S5 as a function of appl ied bias voltage ( -20V to +20V) for temperatures: 45, 65,85, 1 05 and 1 1 5 °C . The frequency is 1 MHz . . . . . . . . . 43 Fig. 28(a) Capacitance variation of sample S6 as a function of appJ ied bias voltage (-20V to +20V) for temperatures: 65, 85, ] 05, 1 1 5 and 1 25 °C . The frequency is 100 kHz . . . . . 44 Fig. 28( b) Capacitance variation of sample S6 as a function of appl ied bias voltage (-20V to +20V) for temperatures: 65,85, 105, 1] 5 and 1 25 °C . The frequency is ] MHz . . . . . . 44 Fig. 29(a) Capacitance variation of sample S5 as a fu nction of temperature for different appl ied bias voltages, ranging from 0-20 volts. The frequency is 1 00 kHz. Solid l ines are to guide the eye . . . . . . . . . . . . . . . . . . . . . . . . . . " ................................................. 45 Fig. 29( b) Capacitance variat ion of sample S5 as a function of temperature for different appl ied bias voltages ranging from 0-20 volts. The frequency is 1 MHz. Solid l ines are to guide the eye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Fig. 30(a) Capacitance variation of sample S6 as a function of temperature for different applied bias voltages, ranging from 0-20 volt s . The frequency is 1 00 kHz. Sol id l i nes are to guide the eye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Fig. 30( b) Capacitance variation of sample S6 as a fu nction of temperature for different applied bias voltages, ranging from 0-20 volts. The frequency is I MHz. Solid lines are to guide the eye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

Fig. 3 1 (a ) Normalized capacitance of samples S 5 and S6 as a function of temperature for appl ied bias voltage of 0 volts. The data i s shown for both 1 00 kHz and 1 MHz. Solid l ines are to guide the eye. Error bars are systematic and est imated to be about 5 % . . . . . . . . . . . . . 48 Fig. 3 1 ( b) Normal ized capacitance of sampJes S 5 and S6 as a function of temperature for appl ied bias voltage of 20 volts. The data i s shown for both] 00 kHz and 1 MHz. Solid l i nes are to guide the eye. Error bars are systematic and est imated to be about 5 % . . . . . . . 48

VIII

LIST OF TABLES

Table 1: Requirements for phase change media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Table 2: General information about the samples under investigation . . . . . . . . . . . . . . . . . . . . . 24

L I ST OF PUBL ICATI ONS

1. Sale h T. Mahmoud, H. Gbamlouche, N. Qamhieh, Hessa AJ-Sba misi "Novel preparation methods of electrical contacts in Ge-Sb-Te thin fi lms", Applied

urface Science Vol. 253, (2007), p. 7242-7245.

2 . H. G h amlouche, S. T. Mahmoud, N. Qamhieh, Sadiqa Ahmad and Hessa AI­Shamisi "Capacitance-Voltage Measurements of Ge-Sb-Te Thin Fi lms", Physica Scripta Vol 77, p. 1-5 (2007)

3. N. Qamhieh H. Ghamlouche, S. T. Mahmoud, H. AI-Shamisi and S. Ahmad "Electrical Conduct ivity M easurements in Ge2Sb2 Te5 fi lms", J Optoelectronics and AdvancedMaterials Vol . 9, No. 10, p. 3157-3160 (2007).

4. Hessa AI Shamisi, H. G hamlouche, N. Qamhieh, S. T. Mahmoud, "Electrical Measurements of Ge-Sb-Te Fi lms Using I mpedance Spectroscopy", Proc. of The 8th Annual UAE University Research Conference, Al Ain (2007)

o

IX

CHAPTER I

I NTRODUCTION

The ability to store information has been an important aspect of human development .

Many events, knowledge and information in the world are preserved for future generation.

Optical data storage, is one of the most successful new techniques for storing huge amounts of

data [ Erwin, 2006], [Libera, 1 990] and [Ovshinsky, 1 968] . Whereas the first optical storage

products only allowed reading of the encoded information, nowadays both recordable and

rewriteable storage media have been developed.

Phase-change media is a rewriteable optical storage media; recently it is used in

many rewriteable products. This success is based on the discovery [Yamada, 1 987] of alloys

containing elements such as Ge, Sb and Te, and the forming of phase-change media to meet

DVD specifications [ Satoh, 200 1 ] . Studying the characteristics of this material led to the

development of three generat ions of optical storage products : compact disks (CD), digital

versatile disks ( DV D), and Blue-ray disks (BD) or high-definition digital versatile disks (HD

DVD).

The formation of small amorphous marks in a crystalline matrix corresponds to

writing of information, whereas the re-crystallization of the amorphous spots leads to the

erasure of information. I n recent years phase-change materials have also become of interest

for developing non-volatile memories with a number of attractive features [Wuttig, 2005] and

[Luo, 2004] . I n applications as a non-volatile memory, the pronounced difference in electrical

resistivity i s used. The amorphous state has a high resistance. Applying a long voltage pulse

locally heats the amorphous region and leads to re-crystallization. Applying a higher voltage

pulse to the crystalline state with low resistance leads to local melting and the formation of an

amorphous region on rapid quenching. The phase transitions from the amorphous to

1 o

crystalline states, and ice versa, of Ge-Sb-Te and Se-Sb-Te films by applying electrical

pulses have been observed [Tanaka, 2002] and [Nakayama, 2003] .

The first material s used were good glass formers such as Te-based eutectic alloys,

represented by Te85Gel5 doped with an element such as Sb, S and P [Ovshinsky, 1 968] .

Although these materials already showed electrical switching that could be used for electronic

storage, the time for crystal lization was of the order of microsecond, partly because the first

alloys did not crystallize in a single-phase material . The first materials to show fast re­

crystal lization and good optical contrast were Ge-Te [Chen, 1 986] and Gel J TC6oSfl4Au25

[Yamada, 1 986] and [Ohno, 1 989]. This give rise to the discovery of a new al loys, such as

Ge1Sb4Te7, GelSb2Te4 and Ge2Sb2Tes [Yamada, 1 987 and 1 99 1 ] . Nowadays Ge2Sb2Tes and

related materials such as GeSbTeN, GeSnSbTe, GeB1SbTe, GeBiTe and GeI nSbTe [Kojima,

1 998 and 200 1 ], [Yusu, 2005] and [ Kusada, 2006] have been tried, some of which are

frequently used in commercial products. The Ge-Sb-Te composition has been used in

commercial optical and electrical devices due to it s fast crystaJJization. The most important

properties of Ge2Sb2 Te5 is the fast transformation between amorphous and crystal l ine phases.

Ge2Sb2 Te5 alloy exhibit best performance in data storage devices due to it's high thermal

stability at room temperature, high crystal l ine rate at high temperatures, and extremely good

reversibility between amorphous and crystalline phases [Zhiwei, 2006] . Recently, these

materials attracts the attention of the scient ist to study and understand the crystallization

phenomena [Yamada, 1 99 1 ] , [Tomigava, 1 999], [Weidenholf, 1 999], [Friedrich, 2000],

[Gonzales, 2000], [ Kolobove, 2004], [Wuttig 2005], and [\Vang, 2005] . The importance for

this research is to achieve a ( i ) high recording density, ( i i ) high data transfer rate and (i i i) high

direct overwrite cyclabil ity. The minimum recorder bit dimension can be significantly

decreased by using short wavelength lasers, which results in higher recording density [Liqiu,

1 998 ] . Optimizations of the physical propert ies of the phase change material; particularly its

o

2

cry tall ization behavior i s essential to increase the data transfer rate and overwrite cyclabil ity

[Martijn 2003 ] . These studies have pointed out that the crystal lization of the amorphous

materials requires some incubation time to form the critical nuclei, which limits the speed of

devices [Weidenholf, 200 1 ] .

Most of the experimental methods used to investigate the amorphous to crystalline

transit ion have shown that resistance measurements are among the most sensitive [Friedrich,

2000] . However, results of DC measurements frequently depend on the nature of the electrical

contacts. Morales-Sanchez et al. 2003 proposed AC impedance measurements to separate the -contribution of contacts from the measured electrical response. The impedance measurements

in amorphous Ge2Sb2 Tes t hin fi lms as a funct ion of temperature, have shown that the contacts

have a strong influence on their electrical properties. New configurations of electrical contacts

have been used in order to minimize their effect on the total measured response of chaJcogenide

a110 Ge2Sb2 Te5 using AC impedance spectroscopy technique [Saleh, 2007] .

The rapid and reversible amorphous-to-crystal l ine phase transformat ion IS

accompanied by increase in the optical reflect ivity and the electrical conductivity . However,

uncerta int ies about the opt ical band gap and electrical properties of this material have

persisted . Hereafter the optical propert ies of Ge2Sb2 Tes phase change material had been

invest igated and it is shown that the optical absorption in a l l phases (amorphous and

crystal l ine) fol lows the Touc relation (a = B(l1aJ - Eg)2 ,where B is a constant and Eg is often

taken as a definition of the optical gap, the so-called Tauc-gap), [Tauc, 1 974], which

corresponds to the optical t ransit ions in most amorphous semiconductors, [Bong Lee, 2005] .

The common feature of these glasses i s the presence o f local ized states in the mobi l ity gap.

This is due to the absence of long range order as wel l as various inherent defects . In

chalcogenide glasses, it is assumed that the localized states in the mobi l i ty gap are the charged

defect s D+ and D- with negative effect ive correlation energy [see sect ion 2.4]. This type of

3

defects u ual ly pins the Fermi energy level in the middle of the forbidden energy gap and

result in the absence of electron spin resonance signal [ Street, 1 978] and [Okamoto, 1 984] .

The investigation of electrical conductivity in chalcogenide glasses i s a valuable tool for

determining the position of the Fermi energy level in the energy gap of semiconductor

materials

The motivat ion of the present work is to investigate the electrical properties and the

indium doping effects of Ge2Sb2 Tes phase change material s using d ifferent techniques. The

electrical measurements such as Current -Voltage ( I-V) characterist ics, DC conductivity, AC --conductivity ( I mpedance Spectroscopy), and Capacitance-Voltage (C-V) analysis wi l l be

conducted. The data gathered from these measurements can give us a clear idea about the

phase transit ion temperature, the nucleat ion process and the energy gap for these materials.

This thesi s is organized as follows; after the introduction in chapter I, a basic

theoretical 0 erview of phase-change material s, specia l ly, chalcogenide ( Ge-Sb-Te) is

presented in chapter I I . In chapter TIl, the preparation of the bu lk al loy and thin fi lms of

Ge2Sb2 Tes using sputtering and thermal evaporat ion techniques in addition to impedance

spectroscopy, I -V and C-V techniques are presented . The experi mental results are di scussed in

chapter IV . The conclusion of the work is summarized at the end of the thesis .

4

CHAPTER I I

THEORETICAL BACKGROUND

2.1 Introduction

In a rewritable disc, infonnation is stored in the so-cal led phase-change layer. This

is often a chalcogenide alloy, which can be reversibly converted from the crystalline to the

amorphous tate by a laser pulse. Applying short consecutive write pulses that cause melt

quenching of the initially crystalline recording film controls writing of amorphous marks . .-Erasure of amorphous marks is achieved by heating the phase-change film to intennediate

temperature levels above glass transition temperature (T g) and below melting point (T m) to

induce re-crystallization. Characteristic temperature-time profiles that are associated with

the write and erase processes are given in figure 1 .

Write: melt-quench ing Erase: recrystallization

----------------------�

Time Time

crystalline molten amorphous amorphous crystalUne

Fig. 1 Schematic of the temperature-t ime profiles associated with recording ( left panel) and

erasure ( right panel) of amorphous marks in a crystall ine layer [Erwin, 2006] .

5

The phase-change materials used today are a result of a 30-years and still

continuing period of empirical optimization of materials. A large number of phase-change

materials have been proposed, but only a few materials meet the fol lowing requirements:

1- Writabil ity· they have to enable writ ing of data.

2- Archiving: the stored information has to be stable.

3 - Readability' easy to read.

4- Erasability' the information should be erasable.

5- CycJability: the storage medium should allow numerous write/erase cycJes. -These data storage requirements can be translated to media requirements (see Table 1 ) .

Table 1: Requirements for phase change media [Erwin, 20061-Storage requirement Material requ irement Material property Writabi l ity Glass fom1er Melt ing pointllayer design,

appropriate optical absorption

Archival storage Stable amorphous phase High act ivation energy, high crystall izat ion temperature

Readabi lity Large signal to noise ratio High optical contrast Erasability Fast re-crystallization Simple crystalline phase,

low viscosity CycJabi l ity Stable layer stack Low stresses, low melting

temperature

2.2. Glass for mation and the amorphous phase

I n phase-change optical recording, information is stored by writing amorphous

marks in a crystalline phase-change layer. This section will deal with the basics of the

process of mark formation and the structure of the resulting amorphous phase. In figure 2

the volume of a liquid i s considered that is cooled to below the melting temperature. At the

melting point (T m), crystal l ization may occur. As i l lustrated in figure 2, crystallization is

accompanied by an abrupt change in volume at (T m) . I t is, however, also possible that the

liquid will become 'undercooled', gett ing more viscous with decreasing temperature, and

6

ultimatel a solid pha e will [orm . This solid is called a glass and the temperature region in

which the undercooled liquid acquires the properties of a solid is called the glass transition

temperature (T g). Glass formation is characterized by a gradual break in the slope of the

volume vs. temperature diagram. For a given composition, the value of the glass transition

temperature depends on the cool ing rate. When cooled slowly (dashed line in figure 2), the

glass transition temperature wil l shift to lower values, as the undercooled liquid has more

time to adjust its propertie to its metastable equil ibrium values. However, cooling down

slowly also increases the chance for crystallization .

v Liqui

I Undercooled liquid I

,

� _----r�

Glas�------j Crystal

T

F ig. 2 Volume as a function of temperature for a liquid, a glass and a crystal [E lliott, 1 990].

The abi lity to form glasses is almost a universal property of condensed matter. I n

order t o produce a n amorphous material crystall ization should b e bypassed. Crystallization

7

takes time and, therefore, the amorphous phdse can be reached by cooling rapidly to below

the gJass transition temperature. I n a phase-change optical disc, rapid cooling is made

possible by the small volume that is amorphized and the special stacking of layers

comprising the di sc.

Amorphous solid, is a sol id like any other if macroscopic properties like shape,

shear stiffiless etc. are considered. But, on an atomic scale there is a difference between

crystal l ine and amorphous sol ids. The main difference between them is the length scale

over which the atoms are related to one another by periodicity. A crystal material has an -atomic structure that repeats periodically across its whole volume on the other hand,

amorphous materials, like glass, have no long-range order at all This is il lustrated in figure

3 . For example, all atoms in figure 3 (b) satisfy their coordination number, each have three

nearest neighbors at nearly the same distance. Nevertheless at short range there is variation

in bond strength and fluctuation in bond angles. On the contrary, it has been discovered

recently, that t he wel l-known phase-change material Ge2Sb2 Tes shows a substantial

difference in short-range order between the amorphous and the crystaJJine state. It still

needs to be clarified if this is a generic feature of phase change materials .

(a) (b)

Fig. 3 Schematic representations of a crystalline and an amorphous structure [ Elliott,

1 990].

8

2.3 Pbase-change materials

uitable phase-change materials for optical recording have the ability to form

glasses at time scales and temperatures that are appropriate for given conditions such as

data transfer rate and available laser power. This requires that the glass formation takes

typically less than 1 00 ns within an adequately cooled recording stack, and that the melting

point is below 1 000 DC, typical ly around 600 °C. Furthermore, the material should possess

sufficient optical contrast between the amorphous and crystalline state. Some materials that

fulfil l these demands and that are currently applied in rewritable CDs and DVDs, ar&­

indicated in figure 4 . The main constituents of these materials are antimony, tellurium and

germanium [Erwin, 2006]

Ge

Te Sb

Fig. 4 Composition triangle of Sb, Te and Ge [Erwin, 2006].

•

9

Phase-change memory uses the unique behavior of chaJcogenide glass, which can

be "switched" between two states, crystalline and amorphous with the application of heat.

early all memory devices make use of a chalcogenide alloy of germanium, antimony and

tellurium ( Ge-Sb-Te) called GST. It is heated to a high temperature at which point the

chalcogenide becomes a liquid . Once cooled, it is frozen into an amorphic glass-like state

which has a high electrical resistance. By heating the chalcogenide to a temperature above

its crystal lization point, but below its melting point, it will transform into a crystal l ine state

with a much lower resistance. This phase transition process can be completed in as quickly

as five nanoseconds, which makes chaJcogenides suitable for memory devices.

2.3.1 Chalcogenides

Chalcogens are the group 6 elements in the periodic table, as shown in figure 5.

they are somet imes known as the oxygen family. They consist of the elements oxygen (0),

sulfur (S), selenium (Se), tel lurium (Te), and the radioactive polonium ( Po ). The

compounds of the heavier chaJcogens (part icularly the sulfides, selenides and tellurides)

are col lectively known as cha1cogenides. Unless grouped with a heavier cha1cogen, oxides

are not considered chalcogenides.

Chalcogenides have special electrical properties. They have no electron spm

resonance and they cannot be doped since they have negative correlation energy defects.

Chalcogenide alloys are glasses containing a chalcogenide element (sulphur, selenium or

tellurium) as a substantial constituent [Mott, 1 979] .

•

10

IA 1 H

3 IIA

cbaJcogeos

/0 IliA IVA VA VIA VilA He

5 6 7 8 9 10 2 Li

4 Be

Periodic Table of Elements B C N 0 F He

11 12 15 16 7 18 3 Ha Mg P S CI Ar

4 19

K 20 Ca

37 38 5 Rb Sr

55 56 6 Cs Ba

87 88 7 Fr Ra

It Lanthanide Series

+ Actinide Series

58 Ce

90 Th

59 Pr

91 Pa

60 61 62 63 6<1 Nd Sm Eu Gd

92 93 94 95 96 U

65 66 67 68 1b 0)' Ho Er

97 98 99 100

69 70 Tm Yb 101 102

71 lu

103

8{) Rn

Fig. 5 The periodic table of elements. The chalcogens appear in the sixth group of the table.

Germanium-Antimony-Tellurium (Ge-Sb-Te) GST, is a phase change material from

the group of chalcogenide glasses, used in rewritable optical discs and phase-change

memory applications. The new phase-change materials are likely p-type Ge-Sb-Te

semiconductors. The melting point of this alloy is about 600 °C (900K). At 1 50 °C Ge-Sb-

Te becomes pure fcc crystal, as shown in figure (6a), then it becomes pure hcp crystal at

400 °C as shown in figure (6b).

�Sb2Te5, whose trigonal crystal structure is depicted on the right-hand side in

figure 6, consists of a sequence of hexagonal layers of Ge, Sb and Te. However, if short

l aser pulses are used to crystallize this material, a metastable structure is formed, which is

characterized by s ix fold coordination of the atoms with the cubic arrangement

characteristic for the rocksalt lattice.

11

Thi structural arrangement is shown on the left-hand side in figure 6. The Te atoms

'occupy each ite of their fcc sub-lattice in the NaCI structure. Alternating Ge, Sb and

vacancies occupy the sites of the other sub-lattice. Recent studies find compelling evidence

for a pronounced local distortion away from the six fold coordinated sites, which

considerably lowers the energy of the solid, [ Erwin, 2006] .

(a) (b)

Fig. 6 Two crystalline phases of Ge2Sb2 Tes . On the right hand side the pure fcc crystal,

and in left hand side the pure hcp crystal [ Erwin, 2006] .

As mentioned before, selenium (Se) and tellurium (Te) are considered

semiconductors. I n the amorphous phase lack of periodicity or long range order makes the

-> wave vector k no more a good quantum number to describe the electrical propert ies. The

only valid concept in amorphous phase is the density of state.

Q

12

2.4 Density of tate in a morphous semiconductors

everal models were proposed for the density of states of amorphous

emiconductors. AJI used the same concept of localized states in the band tails which is

different than the sharp bands in a crystaL The difference between these models of the

amorphous semiconductor l ies in the estimate of the extent of this tailing. An early model,

namely Cohen Fritzsche-Ovshinsky (CFO), supposed that the amorphous structure would

lead to overlapping band tai ls of localized states as shown in figure (7a). Those states

derived from the conduction band (CB) would be neutral when empty and those from the

valence band (VB) would be neutral when ful l. Such overlapping states would pin the

Fermi-level energy (EF) near the middle of the gap, a feature required by the electrical

properties of these materials. The other principal feature of this model is the existence of

the "mobi lity edges" at energies Ec and Ev, in the CB and VB tails respectively. The

mobil ity edges were introduced earlier by Mott, which is why the model is sometimes

called the M ott CFO model . The difference between the two energies is called the mobility

gap. evertheless, one of the major objections against the CFO model is the high

transparency of amorphous chalcogenides below a well defined absorption edge [Mott,

1 979]

An alternative suggest ion to the CFO model for states in the gap is shown in figure

(7b) [Mott, 1 975]. This model assumes that the tai ls of localized states should be rather

narrow, and they extend a few tenths of an electron-volt into the forbidden gap. The model

also proposed the existence of a band of states near the middle of the gap, arising from

defect centers, e .g. dangling bonds. These states may act both as deep donors ( D-) and

acceptors ( D+) (charge defects) . Single and double occupancy conditions lean to two bands

separated by an appropriate correlation energy or Hubbard U .

•

13

Adopting the abo e model for gap states in tetrahedral amorphous serruconductors

such as Ge or i clearly explains many of their electri cal properties. For example, the

observed electron spin resonance ( ESR) signal reveals the presence of unpaired spins

ascribed to electrons in dangling bond states. Moreover, at low temperatures, conduction

occurres by variable range hopping, indicating that transport properties are dorrunated by a

high density of states at EF [Qamhieh, 1 996].

For the chalcogenides, the situation is different . There is plenty of evidence that

chalcogenide glasses contain high density of gap states with EF pinned near rrud-gap.

However, none of the mentioned phenomena in the previous paragraph is normally

observed (unless one first i l lurrunates the material at low temperatures) [Mott, 1 979] .

Therefore, a rather different model i s required for gap states in chalcogenides which gives a

good description of many electrical and optical propert ies of these glasses. The model

proposed that local ized gap states are charged defects with negative effective correlation

energy between electrons. The origins of this energy were soon understood in terms of the

Valence-Alternation-Pair (V AP) model [Kastner, 1 976].

•

14

Valence band

Mobility edge

Valence band

(a)

(b)

Band tails

F ig. 7 Density of states in amorphous semiconductors

(a) CFO model

Conduction band

Mobili edge

Conduction band

(b) With bands of donor (E) and acceptor (Es) arising from the same defect (Mott

and Davis model) [ Mott, 1 979].

1 5

2.5 Conduction in a morphous semiconductors

AJthough chalcogenides are p-type ( conduction by holes), standard usage is to discuss

equivalent conduction properties in terms of electrons. Conductivity is a macroscopic quantity

which represents an average property of carriers as they move from site to site. Calculation of

the conductivity therefore involves the transfer rate, scattering and trapping processes, as well

a the appropriate averaging over the distribution of states. Conductivity is the product of the

carner density and the carner mobility

a = neJ.1 ( 1 )

The contributions to (J are summed over the density of states,

a = f N(E)eJL(E)f(E)dE (2)

Where f{E) is the Fermi-function. Figure 8 shows schematical ly the variation of N(E), �(E)

and f{E) with energy E and their contribution to the conductivity for states near the band

edge.

N(E) �(E) feE) a(E)

Fig. 8 I l lustration of the density of states near the band edge, together with the carrier

mobility 11(E), the occupatir)O probability f{E) and the conductivity aCE)·

When there is a very high defect density in the middle of the band gap, as in tetrahedrally

coordinated semiconductors like Si and Ge, conduction takes place by hopping at the Fermi

o

1 6

energy Hopping conduction can also take place in the band tail s where the density of states

again becomes large, but the carrier concentration is lower. The much lower defect density

in amorphous chalcogenides prevents tills mechanism from contributing significantly at

moderate temperatures and instead conduction takes place by electrons or holes at the band

edges where both the density of states and the mobility increase with energy [Qamrueh,

1 996] . Since our study is at relat ively illgh temperatures (above room temperatures), the

mechanism of extended states conduct ion is reviewed in the following section.

2.5. 1 Extended state conduction

The conductivity of any semiconductor can be expressed in the form of Eq .2 . The

integral contains contributions from electron transport above EF and hole transport below

EF When conductivity takes place far from the Fermi level, Boltzmann stat istics can be

applied to describe the occupancy. The conductivity due to a single type of carrier excited

beyond the mobility edge into the extended states is given by

( 3 )

For the particular case i n which the mobility increases abruptly at the mobi l ity edge energy,

Ee, according to Matt's view, then Eq. 3 gives

[ E -E ] a = a rrun exp - C KT F

Omin is referred to as minimum metall ic conductivity and is given by

(4)

(5)

where !lc, is the free carrier mobility at Ec. Mott calculated a value for Omin, given by the

expressIOn

e2 a rrun = const. -

na (6)

o

17

where a i the interatomic distance and the constant lies in the range 0 026 and 0. 1 ; <Jmin i s

usually in the order of 200-300 n-Icm-I.

The conductivity in amorphous semiconductors IS usual ly activated, and IS

described by

(7)

where a. i s the conductivity prefactor and Eo is the activation energy. Comparing Eq . 4

and 7, one might conclude that a measurement of <J(T) immediately gives the location of

t he mobi lity edge (Ec-EF), and the prefactor gives the conductivity at the mobility edge.

However, the measurements are unhelpfuJ in this regard as there is a huge variation in the

values of a ..

The most obvious mechanism by which the prefactor varies is when the activation

energy is temperature dependent,

(8)

where the constant value 8 represents the first order approximation to the temperature

dependence, and (Ec-EF)o is the value extrapolated to zero temperature. Substituting Eq . 8

into Eq. 7, the expression for the conductivity becomes

[ (£ -E ).] a = a. exp(o / k) exp - C kT F

(9)

Therefore, within the l inear approximation, the conductivity activation energy Eo measures

the value of (Ec-EF) and the prefactor <Jo contains the additional term involving the

temperature dependence, exp(8/k) . Davis and M ott have made an estimate of <J'oexp(8/k).

They found values clustering round 1 03 n-Icm-I for most amorphous semiconductors. An

est imate for 8 can be obtained from the temperature dependence of t he optical gap.

o

1 8

2.5.2 Hopping conductivity

We have seen that structural defects glve nse to localized states in the gap of

amorphous semiconductors. However one of the definitions of localized states is that the

average conduction associated with the states at energy E is zero. Thus, conduct ion

involving localized states can take place by hopping of electrons from full states to

neighboring empty states, with phonon assistance. Hopping conduction dominates when

the Fermi occupation factor in Eq. 2 gives the states nearest to the Fermi energy a

sufficiently strong weighted contribution to the total conductivity. There are two ranges of

localized tates where conduction by hopping can take place:

1 ) Conduction in band tails : When carriers are excited into localized states at the

band edges hopping conduction can occur at energies close to the conduction band

edges EA and EB, (figure 7).

2) Conduction in states around E F: I f the density of states at the Fermi level is

fi nite, then there is contribution from carriers with energies near EF which can hop

between localized states by a process analogous to impurity conduction that takes

place in heavily doped crystalline semiconductors. Hopping is to sites with the

smallest distances and is therefore sometimes called nearest-neighbor hopping.

o

19

Ln (J

Ec- EF 1 . Extended state conduction

/ 2. Conduction in band tails

. . - - - / . - -, , " 3. Conduction in states a round EF

-/ I ff

Fig. 9 Schematic i l lustration of the temperature dependence of conduct ivity expected for an

amorphous semiconductor.

This behavior has been observed in several amorphous semiconductors at low

temperatures, where the activated nature of band conduction ensures that it is smal ler than

the contribution from variable-range hopping. Figure 9 shows schematically the three

conduction mechanisms and the change in it on lowering the temperature. If the density of

defect states at EF is high then conduction in the band tails will not be dominant at any

temperature range and a direct transition from extended state to hopping conduction at EF

will result [Qamhieh, 1 996] . In the present work, onJy the first part of figure 9 at high

temperature is to be considered, where, extended state conduct ion takes place.

2.6 AC conductivity

AC conductivity i s one of the studies done on solids in order to characterize the

bulk resistance of the crystalline sample. Measurement of AC conductivity can be done by

different techniques. The currently used technique is the complex impedance spectroscopy.

This study also gives information on electrical properties of materials and their interface

with electronical ly conduct ing electrodes . The complex impedance spectroscopy

20

Q

mea urement of AC conductivity i s based on studies made on the measurement of sample

impedance/admittance over a range of temperatures and frequencies and analyzing them in

complex impedance plane. This is particularly characterized by the measurement and

analysis of impedance (2), admittance ( 11 and plott ing of these functions in the complex

plane which is known as Nyquist diagrams. Impedance is a more general concept than

resistance because it takes phase differences into account . In AC measurements, the

resistance, R, i s replaced by the impedance, Z, which i s the sum of resistance and reactance.

Impedance can be written as -Z = Z ' + Z ", ( 1 0)

where Z ' i s the real part and Z II the imaginary part of Z. Z is a vector quantity and may be

plotted in the plane with either rectangular or polar coordinates as shown in figure 1 0.

The two rectangular coordinate values are

Re(Z) = Z ' = IZI cos <p,

Im(Z) = Z II = IZI sin <p,

with phase angle <p = tan-I(Z '/Z ") and the magnitude of IZI = [(Z '/ + (Z ")2 ] 1 /2

IX

Xe I--------�

Fig. 1 0 I mpedance vector representation.

2 1

( 1 I )

( 1 2)

In polar form, Z may be written as

Z(O) = IZl exp( j 0) , ( 1 3 )

Where, 0) is the angular frequency

exp(j 0)) = cos( 0)) + j sine 0)) ( 1 4)

In general, Z is frequency dependant . Impedance spectroscopy consists of the measurement

of Z(O)) over a wide frequency range. It is from the resulting structure of Z(O)) vs 0) that one

derives information about the electrical properties of the electrode material system [ padma,

2002] . Impedance plane plots for the sample which can be considered as a capacitance (C)

in series with the resistance (R), or (C) in parallel with ( R) are shown in figure 1 1 .

R ( '

•• ----/'VV� I I _ •

Z" (.0)

f

1 R Z'(Q)

R

.-[ - 1 1 ---' (

ZIt (Q)

High frequency

Fig. 1 1 I mpedance plots for series and parallel combinations of R and C.

22

R Z'(Q)

Figure 1 1 also shows, the impedance spectra ( Imaginary part of the impedance Z"

versus real part Z' of the impedance for a certain temperature), where the resistance is equal

to the diameter of the semicircle in Nyquist plot . The semicircle starts at high frequency

then ends to a low frequency for a certain temperature but the shift on this semicircle

depends on the temperature as we will see in the next chapters. At the semicircle peak one

can get the relaxation frequency at which (1)RC = 1 .

Impedance spectroscopy technique was used to study the AC conductivity and the

capacitance of the samples with a series combination of R and C which is explained in the -next chapters

23

CHAPTER I I I

SAMPLE PREP ARATION AND EXPERI M ENTAL TECHNIQUES

Thin films of Ge2 b2 Tes were prepared by two different techniques (Sputtering &

Thermal evaporation techniques) at different places as shown in table 2 . (The bulk alloys of

glassy Ge2Sb2 Te5 were prepared by rapid melt quenching technique). These alloys have

been used as a source in the thermal evaporat ion technique to deposit films at two places,

UAE University and the University of Leuven (Belgium).

-Table 2: General informat ion about the samples under investigat ion

Sam ple Sample Contact Co m position Deposition Tc Prepa red at type Technique

1 Aachen Fig. 1 5-a Ge2Sb2Te5 Sputtering 1 45 °C University

2 Aachen Fig. 1 5 -c Ge2Sb2 Te5 Sputtering 1 45 °C University

3 Leuven Fig. 1 5-c Ge2Sb2TeS Thermal 1 40 °C University Evaporation

4 UAE Fig. 1 5-b Ge2Sb2Te5 Thermal 1 3 5 °C University Evaporation

5 UAE Fig. 1 5-b Ge2Sb2Te5 Thermal 1 3 5 °C University Evaporation

6 UAE Fig. 1 5-b l no.3Ge2 Sb2 Tes Thermal 1 50 °C University Evaporation

The amorphous structure and stoichiometery of the films were identified by X-ray

diffraction and Energy Dispersive X-ray analysis E DX as shown in figures ( 1 2 & 1 3 ) . X-

ray spectroscopy is a technique which allows checking whether the material is amorphous

or crystalline. Using this technique, the amorphicity of evaporated Ge2Sb2 Tes film was

24

confirmed. Sample was characterized by the appearance of diffraction lines due to the

formation of a crystalline state. To check the amorphicity of our sample X-ray diffraction

spectroscopy were performed using Philips diffractometer with Cu.:Kt X-radiation source.

Figure 1 2 shows the diffraction pattern in the range of 28 == 0-80°. The absence of any

sharp peaks indicates the amorphous nature of the quenched material . . EDX is an analytical

technique used for the elemental analysis, as shown in figure 1 3 three peaks appear for the

three elements in the sample used Ge2Sb2 Te5 which are characteristic of an element's

atomic structure

Gold electrodes were evaporated for electrical contacts usmg three different

configurat ions. Three different techniques were used to analyze the thin films. 1 ) The DC

four probes technique has been used to measure the I-V curves and hence the resistance of

the samples. From this technique one can determine the amorphous-crystalline transition

temperature. 2) The impedance spectroscopy technique was used to study the AC

conductivity and the capacitance of the samples. 3 ) The capacitance-voltage technique was

used to study the capacitance of the film as a function of temperature and frequency.

1 0 0 0

8 0 0

6 0 0 <J) C :J 0 4 0 0 0

2 0 0

0 0 2 0 4 0 6 0 8 0 2 8

Fig. 1 2 X-ray diffraction spectrum of Ge2Sb2Te5 film as deposited

25

Counts

2000-T e

1500-

1000-s

500-

....

3. 1 Sample preparation

3. 1 . 1 Prepa ration of bulk mate ria ls

Rapid quenching technique has been used to prepare the bulk alloys of the phase-

change material s . This technique is historically the most established one and it is sti l l the

most widely used in the preparation of amorphous chalcogenide materials [Qamhieh,

1 996] . . The method consists of weighing and mixing the high purity constituents fol lowed

by melting in sealed evacuated quartz ampoules. The melt is continuously rotated to ensure

homogenizat ion. Subsequently, the melt will vitrify when cooled by quenching in either air

or ice water. More rapid quenching has also been used in preparing amorphous

chalcogenides, e .g . rol ler quenching for Sb3S3 and melt-spinning for Te-Se, Ge-Se and Te-

Ge alloys.

The bulk al loys of glassy Ge2Sh2 Te5 were prepared at UAE University from

99.995% pure Ge, Sb, and Te by rapid melt quenching technique, as shown in figure 1 4 .

The materials with proper const ituents were sealed in an evacuated quartz tube (at 1 0-3

Pa)

and heated at 1 050 °C for 24 hours in a rotating cylindrical furnace to ensure

homogenizat ion.

26

r I

Te G --:-Jr s

1 ) Materials in a quartz tube

, • Tube holder " " , \ , ,

2) Sealed under vacuum =1 0-3 Pa

3) Melts in cylindrical furnace at 1 050°C for 24 hour

4) Quenched in iced-water

Ice-water

.---. -� Fig. 1 4 Schematic i l lustration of experimental arrangement used for producing amorphous

3. 1 .2 Thermal Evapo ration Technique (TE)

This technique is perhaps the simplest and possibly the most widely used method for

producing amorphous sol ids in t hin film fonns. TE is normally performed in vacuum and

involves resistive or electron-beam heating of a boat containing the material to be

evaporated. The material is melted then evaporated and finally condensed onto a cold

substrate. This method is suitable for relatively low melting-point compounds, such as

chalcogenides.

27

3. 1 .3 puttering Technique

puttering is a physical vapor deposition, process whereby atoms in a solid target

material are ejected into the gas phase due to bombardment of the material by energetic

Ions

The process can be thought of a large cluster of atoms, colide by energetic ion. The

first collision pushes atoms deeper into the cluster, subsequent collisions between the atoms

can result in orne of the atoms near the surface being ejected away from the cluster. The

number of atoms ejected from the surface per incident particle is called the sputter yield.

3. 1 .4 Deposition of electrodes

Gold electrodes were evaporated for electrical contacts in three different ways, as

shown in figure 1 5 . In the first method figure ( 1 Sa), gold .is evaporated on top of Ge-Sb-Te

film. I n the second method figure ( I 5b), gold is evaporated first on the glass substrates and

then Ge-Sb-Te film is deposited . In the third method the material is deposited on a part of

the glass and then the gold is deposited on top of it such that part of the gold electrode

surface touches only the substrate, as shown in figure ( ] 5c).

o

28

(a)

Fig. 1 5 Schematic diagrams of the electrical contacts.

3.2 Measuring Technique

3.2. 1 The impedance spectroscopy

electrodes

Ge-Sb-Te fihll Glass substrate

(c)

I mpedance is an important parameter used to characterize electronic circuits,

devices, and the materials used to make devices. I mpedance (Z) is generally defined as the

total opposition a device or circuit offers to the flow of an alternating current (Ae) at a

given frequency, and is represented as a complex quantity which is graphically shown on a

vector plane. An impedance vector consist s of a real part ( resistance, R) and an imaginary

part ( reactance, X) as shown in figure 1 0, [Kazunari, 2003 ] .

29

Impedance spectroscopy ( I S) is a powerful method of characterizing many of the

electrical propert ies of materials and their interfaces with electronically conducting

electrodes It may be used to investigate the dynamics of bound or mobile charge in the

bulk or interfacial regions of any kind of solid or liquid material : ionic, semiconducting,

mixed electronic-ionic and even insulators (dielectrics) [ Ross 1 987] .

Impedance spectroscopy measurements were performed in th is work using 1 260

Solartron impedance analyzer in the range of 1 Hz- 1 0MHz, with amplitude of 0.5 volt and

zero bias. Si lver paste has been used to make contact between the coaxial cable and the

gold electrodes. The measurements were performed at different temperatures from 30- 1 30

dc . The temperature was measured by a K-type thermocouple, and the samples were heated

in Argon gas environment . Figure 1 6 shows the circuit connections for an impedance

measurement. In this figure the current (I) is applied across the sample and the voltage (V)

i s then measured across it . The impedance is then given by: Z = V I

Z60 and Zview have been used for data acquisition and analysis . These pieces of software

were supplied by Solartron especially for the 1 260 solartron impedance analyzers.

Test Equipment : - -- -- - - · - - - - - - - - --FaradayShf�d - - - -- -- -- :

GPIB r Pa tc electrode

ub trate

bTc film

- - - - - - - - - - - - - - - _ . - - - - - - - - - - - - - - - - - - - - -

Fig. 1 6 Connections for an impedance measurement

3 0

3.2.2 Current-Voltage techniq ue

Four probes technique was used to measure the DC resistance of three samples

which were prepared from the same batches of samples 2, 3 and 4. The samples were

heated from room temperature up to 1 70 °C at a heating rate of 5 °C/min.

3.2.3 Capacitance-Voltage technique

Capacitance-Voltage measurements were performed on amorphous and crystalline

samples using KeithJey 590 CY meter at two operating frequencies, 1 00 kHz and 1 M Hz.

The circuit for the CV Analyzer is similar to the impedance measurement connection

shown in figure 1 6.

The samples were heated inside a Faraday-cage and Argon gas environment . The

measurements were performed for a sweep of voltages from -20 to +20 volts at different

temperatures. K-type thermocouple has been used to measure the samples temperatures.

This measurement is used to find the dependence of both capacitance and resistance on

applied bias voltage, temperature and frequency.

For the I -V and C-V measurements, we have written our own programs for data

acquisition using Labview. This later is a graphical programming tool used mostly in

instrument automation.

..

3 1

4. 1 DC Measurements

CHAPTER I V

RESULTS AND D I SCUSSIONS

Four probes technique has been used to measure I -V characteristics of sample 4

u mg the sample configuration shown in figure ( 1 5b), in addition to two other gold

electrodes.

Figure 1 7 shows the current versus voltage ( I -V) curves for Ge2Sb2 Te5 film at

different temperatures ( 70- 1 50°C) . The I-V curves are straight lines at the measured

temperature and voltage ranges. This indjcates that the material has an ohmic resistance,

which can be calculated from the slope of the I -V curve at each temperature [Qamhieh,

2007] .

The resistance of sample 4 is plotted as fu nction of temperature as shown in figure

1 8 . A sharp drop in the resistance is evident at T � 1 3 5°C Uri T= 2 .45 Kl ) . This

temperature is the amorphous-crystalline phase t ransition temperature (Tc) of the material,

[Wang, 2005] . Below this temperature the In R versus liT curve is a straight line in the

measured temperature range. The resistance R can, therefore, be expressed by the well

known Arrhenius relation [Mott, 1 979],

( 1 5 )

where a is t he conductivity, £.01 i s the activation energy for dark conductivity and kB i s the

Boltzmann constant . From figure 1 8, the value of Ev is est imated to be 0 .36 eV.

o

32

--. <t :::i. --c OJ L-L-::J

0

3 , •

2 f- 0

A

'V 1 � *

o f-� e A A

- 1 � v- v-

* *

-2 - x '\(

- 3 I - 1 0

+ 4 f-..

") f-�

I I

T = 70 °C (a ) T = 80 °C T = 90 °C T = 1 00 °C T = 1 1 0 °C

x T = 1 20 oC *'

� � � � � � # i •

. � � � i .. \) TV'

* X \l <\1 * • * )( * .(

x

I , -5 0

Vo lta ge (V)

. ,

T = 1 30 DC T = 1 40 DC T = 1 50 DC

I

Ib 'l • I

- .. • ..

x x

I

x X

x * * "*

* \l * V v * v ... 4

O � � � 'iI 'if

I 5 1 0

I

"" • • a • ill :It

-

-

-

D f- ... • • • • • .. · · t iJr l � • • • • • • • ir -•

-

-

I

i O

Fig. 1 7 J -V curves for Ge2Sb2 Tes (Sample 4 ) fi lms at different temperatures a) in the

amorphous phase and b) around and after the transition to the crystal l ine phase.

33

- - ' , -.

1 0 T

• ,-... C 1 0 6 --- I �_______________ __----------___ � --y,-Q) 0 Amorphous

c: ro

1 05 � (/J f/) OJ

Cl:: 1 04 Crystal I

� . '

• 1 0 3

2 . 3 2 . 4 2 . 5 3 . 0 3 . 1

F ig. 1 8 Resistnce o f Ge2Sb2 Te5 film ( Sample 4) as a function of temperature. The dashed

l ine is the linear fit of the resistance in the amorphous state.

The amorphous-crystalline transition temperatures ( Tc) of Ge2Sb2Tes fi lms have

been roughJy determined using resistance measurements as l isted in table 2. Sample 2 3

and 4 were heated from room temperature up to 1 70 °C at a heating rate of 5 °C/min . .

Figure ] 9 shows the sample resistance as fu nction of temperature for three samples

prepared from the same batch of samples 2, 3 and 4. The sharp drop in the resistance

appears at temperatures of 1 45 °C (sample 2 ), 1 40 °C (sample 3 ) and 1 3 5 °C ( sample 4) .

The temperature at which t he sample resistance drops by several orders of magnitude is the

amorphous-crystalline transition temperature [Wang, 2005].

34

1 05

1 0 '

• S a m p ie 2 • S a m p ie 3

- ... S a m p le 4

6 0 8 0 1 0 0 1 2 0 1 4 0 1 6 0 1 80 o Tern pe rature ( C)

Fig. 1 9 Resistance measurements as a function of temperature for samples prepared at

Aachen, Leuven and UAE. The amorphous-crystalline transition temperatures Te are

1 45 °C, 1 40°C and 1 3 5°C, respect ively.

4.2 Impedance spectroscopy measu rements

The impedance spectrum at different temperatures (T= 80- 1 20°C) for sample 4 are

shown in figure 20 each point in the impedance spectrum represent the imaginary part of

the impedance Z" versus its real part Z' at a certain frequency and temperature. It can be

observed that at the same temperature, the distribution of the data points fit with one

semicircle from low to high frequency, the so caJled Nyquist plot . A parallel resistor and

capacitor ( RC) equivalent circuit accurately fit the frequency data [Barsoukov, 2005] .

Hence the electrical properties of Ge2Sb2 Te5 film can be associated with a simple RC

equivalent circuit . R is determined from the radius of the semicircle and C is determined by

3 5

using the frequency at the peak of each semicircle (relaxation frequency, fo), where 2 1t fo

RC = 1 hould be sat isfied ( see figure 1 1 ) . As expected, the diameter of the semicircle keeps

decreasing with increasing the temperature. Moreover, the semicircle peak shifts toward

higher frequency as the temperature increases. These two behaviors are due to the thermal

activation process that occurs in semiconductor materials. The relaxation frequency

obtained is plotted against temperature, as shown in fi.gure 2 1 . It is obvious from the figure

that the Log of the frequency is increase as the temperature increase according to the

Arrhenius relation. This shows simi lar phenomena of the conductivity dependence on the -temperature.

( 1 6)

The activation energy, Ev, calculated from the slope of the straight line fit is found

to be 0.39 e V. This value is relatively close to the one obtained by I -V measurements.

The temperature behavior of the DC and AC measurements indicate that the

conduction in these glasses, Ge2Sb2 Te5, is through an activated process having single

activation energy in the investigated temperature range, where both resistance and

relaxation frequency have the same temperature dependence. The calculated activation

energy values from figures 1 8 and 2 1 are close to each other, 0 .36 and 0 .39 eV,

respectively. To first approximation, the value of the activat ion energy determines the

position of the Fermi energy level with respect to the conducting mobi l ity edge. If

Ge2Sb2 Te5 fi lm has chalcogenide characters, this means that the Fermi energy level is

pinned in the middle of the energy gap, Eg [Mott, 1 979] . Therefore the energy gap can be

estimated to be Eg= 2Ey. The corresponding energy gap obtained from I-V measurements

and AC impedance spectroscopy are 0 .72 and 0 .78 e V, respectively. This value for the

energy gap is in good agreement with the optical gap Eo= 0 .7 eV measured by [Bong Lee,

2005] . This can be reserved as an evidence for the chalcogenide nature defects in

36

amorphous Ge2Sb2 Te5 fi lms, which is characterized by the negative effecti e correlation

energy charge centers D · and D- .

7x 1 0

6x 1 0 .----. - - ..

/- • • ___ T=BO ·C

rJ/ -.- T=90'C

/ T= I OOoC � T= I IOaC • T= l 2:tC c: � •

N qu n e\' I - •

2x 1 0 • •

• • I

•

1 x 1 0

1 .2X 1 0

Fig. 20 I mpedance spectra for a Ge2Sb2Te5 ( Sample 4) film at different temperatures .

� 10' -o

.... .. ... ... ... '- ...

.2 60 - 6'5 � ;0

... .... . .... ... ... .... .... . .... ... ...

0001T ( 1)

... ... .... ... ... ... .... ... ...

: 90 : 95 3 00 3 05

Fig. 2 1 Frequency at the peak of the semicircle as a function of 1 03fT (JC l ). The slope of

the dashed line determines the activation energy.

3 7

Figures 22, 23, 24 and 25 show the imaginary part of the impedance as function of

frequency at different temperatures below Tc for samples 1 , 2, 3 and 4 respectively. Figure

22 shows the presence of two peaks. The peak at low frequency is attributed to the

electricai contacts, whereas the peak at high frequency is attributed to the contribution of

the material under test [Morales, 2003 ] . The appearance of the contacts effect in figure 22 i s

due to the fact that these measurements have been made using the contact type shown in

figure ( l 5a) . Figures 23, 24 and 25 show only one peak at high frequencies that is related to

the material contribution, whereas the contribution of the contacts d isappears at low ...,.. frequencies. The di appearance of the contacts effect in figures 23 , 24 and 25 is due to the

fact that these measurements have been done using the contact type shown in figures ( 1 5c),

( 1 5c) and ( l 5b), respectively. Figure 26 shows a typical example of the impedance spectra

of samples 1 and 4. It can be seen that the spectrum of sample 1 consists of two well

defined semicircles, whereas the spectrum of sample 4 consists of onJy one semicircle. As

has been shown above, the low frequency semicircle describes the electrical properties of

contacts and the one at high frequency describes the electrical properties of the material . I t

is evident from the above results that the place where the silver paste is put on the sample

plays a crucial role in electrical measurements . I n the configuration of figure ( l 5a), the

silver paste is placed on the gold that is deposited over the tested material . During the

heat ing process the silver paste might penetrate the gold and interact with the material .

Such interaction could be the reason for having similar behavior for the peaks due to the

contacts and the one due to the material, as shown in figure 22. The decrease in both peaks

amplitudes i s due to the decrease in the resistance of the material as the sample heats up.

Moreover, both peaks shift toward higher frequency as the temperature increases. This is .

due to the activation phenomena of the carriers in the semiconductor materials [ Barsoukov,

2005] .

38

It is worth tv note that there are no preferences in the order of deposition between

the gold and the materials since the effect of the contacts disappears in the configurations

shown in figures ( I Sb) and ( 1 Sc) . Beside that, the silver paste should be placed on the gold

film only to avoid any direct contact between the silver paste and the material . This will

guarantee that there is no interaction between the silver paste and the material during the

heating process.

1 2x 1 0& • T = 8 0 DC

1 Ox 1 0&

• T = 9 0 1leo � 0

� o. ... T = 1 0 0 DC <> T = 1 1 0 DC -

� 8. 0x 1 05 N E I 6 . 0x 1 05

..,. T = 1 2 0 DC *" T = 1 3 0 DC • ..

� . . o 0 o · 0 • 0 • 4 Ox 1 05

2. 0x 1 05

0 0 1 0D 1 0 '

F re q u e ncy ( Hz )

Fig.22 I maginary impedance as a function of frequency for sample I at different

temperatures in the amorphous state.

39

..----, C --N E I

1 . 2x 1 lt

1 . 0x 1 0s

8 . 0x 1 0 '

6. 0x 1 0 '

4 . 0x 1 0'

2. 0x 1 0 '

0. 0 1 0 0

• T - e o DC o T = 9 0 DC .& T = 1 00 DC o T = 1 1 0 DC

• T = 1 20 DC n T = 1 30 DC

1 0;2

F re qu e n cy ( Hz )

... • •

• •

• •

• cP:> o •

• 0 0 • 0 •

• ° .t.

•

Fig.23 Imaginary impedance as a function of frequency for sample 2 at different

temperatures in the amorphous state .

• T = 8 0 Cc 4x 1 0 ' 0 T = 9 0 Cc

A I = 1 0 0 DC 0 T = 1 1 0 DC

3x 1 0 ' T T = 1 2 0 Cc -a t:l T = 1 3 0 Cc -N E 2x 1 0 ' I

1 x l 0'

0

l et 1 0 1 1 0z 1 0J 1 0 '

Frequency (Hz) 1 O�

••

• •

0 0 •

0 •

0 • M 0 •

• • � <>a • . 0

0 . � o° .to o .... ..:r o 0 <:P .... � it

l Oti 1 0 '

Fig. 24 I maginary impedance as a function of frequency for sample 3 at different

temperatures in the amorphous state

40

•

2 O x 1 0& • 0 .6..

1 5 xl 0& <> - ... C * -N 1 O x1 0& E •

5 O x1 05

0 .0 1 00 1 0 '

T = 80 D C T = 90 DC T = 1 00 D C T = 1 1 0 DC T = 1 20 D C T = 1 30 DC

1 O� 1 0 3

.+. + •

• •

+ 00 0 0 0

• <> •

• 0 �� £ •

o 4 �<> • £ 0<> � 0 4 <> • <> £ <> • 0 · 0<> .. • 0 £ <> · 0 {> · 0 1 0 ' 1 05

F re q u e ncy ( H z )

1 0& 1 0 '

Fig. 25 I maginary impedance as a function of frequency for sample 4 at different

temperatures in the amorphous state

0 0 1 - 2 . 0x 1 05

-4 Ox l 05 - �o - 6 . 0x 1 05 . 0

• 0 0 0 -8 Ox 1 05 I-

0 0 0 •

- 1 . 0x 1 Oli l- • -c:: - - 1 . 2x 1 0fi f-N

- 1 .4x 1 01i f- •

- 1 . 6x 1 Oli f- • - 1 8x l 01i f- •

- 2 . 0x l 01i

- 2 2x l Oli 0 1 x l Oli

0

01% o 0

r

0 a a

0

• •

0

2x l 0·

0

Z· (0)

0 •

0 0

•

Fig. 26 Impedance spectra for samples 1 and 4 at T = 80 °C

4 1

I ,T Samp le 1 • cP Sam p le 4 • 0 0

. 0 0

o � o • 0

0 0 0 •

•

• •

I 3x l 0 1 4 x l 0 1i

.-

4.3 Capacitance-Voltage measu re ments of Ge2Sb2 Tes and Ino.3GezSbz Tes

Influence of indium doping on the crystall ization kinetics of Ge2Sb2 Tes has been

investigated by [Wang, 200S ] . The results indicate that indium might play an important

role in modifYing the crystall ization kinetics of Te-based phase change materials.

Therefore, we investigate below the influence of indium doping using Capacitance-Voltage

measurements.

Amorphous thin films of Ge2Sb2 Tes ( S S ) and l no.3Ge2Sb2 Tes ( S6) are prepared by

thermal evaporation technique. Both films have a thickness and a separation between the

gold electrodes of about 0 . 8 � and 1 mm, respectively. Pre-deposited gold electrodes on

glass substrates are used for the electrical contacts as shown in figure ( I Sb) .

The capacitance variation of Ge2Sb2 Tes ( S S ) and I no.3Ge2Sb2 Te5 ( S6) with applied

bias voltage ranged from -20V to +20V, as shown in figures 27 and 28 at different

temperatures for 1 00 kHz and 1 MHz. The two figures (27 and 28) show a bell shape curve

where the capacitance attains a maximum value, and it is more pronounced at 1 00 kHz­

frequency and at temperatures close to amorphous-crystalline transition temperature (T c) .

However a negati e capacitance is observed for ample S6 at 1 00 kHz-frequency and

1 2SoC temperature, as shown in figure (28a) .

•

42

1 2

1 0 ........ lL Q. -- 8 CD u c: .,

6 == u ., Q. .,

U 4

2

•

0

.... '*

*

T = 45 DC T = 65 D C � * *' T = 85 DC *

* *

T = 1 05DC *' ** * ** *' T = 1 1 5DC ;* ,

/ �,

a

/* #*� � *

-20 - 1 0 o Ap p l i e d b i a s ( V )

1 0 20

Fig. 27(a) Capacitance variation of sample S5 as a function of applied bias voltage (-20Y

to +20Y) for temperatures: 45, 65, 85, 1 05 and 1 1 5 °C . The frequency is 1 00 kHz.

5 .0

........ lL Q. --CD 4 .0 u c: .,

== u ., Q. ., 3 .0 0

• T = 45 ° C 0 T = 65 ° C .... T = 85 ° C '* T = 1 05 ° C * T = 1 1 5 ° C �

�

-20 - 1 0 0 1 0

Ap p l i e d b i a s (V )

b

20

Fig. 27(b) Capacitance variation of sample S5 as a function of appliecl bias voltage (-20Y

to +20Y) for temperatures: 45, 65, 85 , 1 05 and 1 1 5 °C. The frequency is 1 M Hz.

o

43

LL 0.0 a...

Q) o c � -0 .4 ·u (U a... (U

U -0 .8

-20 - 1 0 o Appl ied b ias (V)

1 0 20

Fig. 28(a) Capacitance variation of sample S6 as a function of applied bias voltage ( -20V

to +20V) for temperatures: 65, 85, 1 05, 1 1 5 and 1 25 °C . The frequency is 1 00 k Hz.

Q) o c (U ......

'u (U a... (U

U

0 .42

b

0 .40

-20 - 1 0 o Appl ied bias (V)

•

0

A

*

1 0

T = 65 °C T = 85 °C T = 1 05 ° T = 1 1 5 °c

Fig. 28( b) Capacitance variation of sample S6 as a function of applied bias voltage ( -20V

to +20V) for temperatures: 65, 85, 1 05, 1 1 5 and 1 25 °C . The frequency is 1 M Hz.

"

44

Figures 29 and 3 0 show, respectively, the capacitance variation of samples S5 and

S6 a a function of temperature for different applied bias voltages (0-20 volts) at 1 00 kHz

( figurers (29a) and (30a» and ] MHz (figures (29b) and (30b) . In figure (29a) the

capacitance increases with temperature for bias voltages less than 1 5 volts and it shows a

maximum at - 1 05 °C for voltages greater than 1 5 volts . In figure (30a) the capacitance

decreases for all bias voltages and becomes negative at temperatures greater than 1 1 5°C . At

high frequency ( 1 MHz) figures (29b) and (30b) show that the capacitance increases with

temperature for all bias voltages, and no negative capacitance was observed at high

frequency ( 1 MB).

----.. LL a.. <l> u C co ...... T5 co a.. co

o

1 2

1 0

8

6

4

• Bias = 0 V -0- Bias = 3 V

Bias = 6 V -*- Bias = 9 V

-*- Bias = 1 2 V -'\l Bias = 1 5 V

-T- Bias = 1 8 V

a

o /

Bias = 20 V / : * � ... ...... *----� ��e��� v � ! � r: =- � '-. e � O ;;;:g;;;; � � =-- �

2 ����-L�� __ �L-���-L��--��

4 0 60 8 0 1 00 1 20

Temperature (C)

Fig. 29(a ) Capacitance variat ion of sample S5 as a function of temperature for different

applied bias voltages, ranging from 0-20 volts. The frequency is 1 00 kHz. Solid lines

are to guide the eye.

Q

4 5

� c co --'0

5.0

4 .5

4 . 0

3.5

� 3.0 co

o 2 5

Bias = 0 V b -0- Bias = 3 V

.. Bias = 6 V -(r- Bias = 9 V

Bias = 1 2 V v-- Bias = 1 5 V T- Bias = 1 8 V Q Bias = 20 V

Temperature (oC)

Fig. 29(b) Capacitance variation of sample S5 as a function of temperature for different

applied bias voltages, ranging from 0-20 volts. The frequency is 1 MHz. Solid lines

are to guide the eye.

0 8

0 4

,---. LL Q. 0 0 Q) u c co ......

-0 4 "[> co Q. co

0 -0.8

- 1 .2 20

O � �

g��$""� - Bias = O V ��

-0- Bias = 3 V

Bias = 6 V

Bias = 9 V

Bias = 1 2 V

---v-- Bias = 1 5 V

---T- Bias = 1 8 V

-Q- Bias = 20 V

40 60 80 1 00

Temperature CC)

a

1 20 1 40

Fig. 30(a ) Capacitance variation of sample S6 as a function of temperature for different

applied bias voltages, ranging from 0-20 volts. The frequency is 1 00 kHz. Solid lines

are to guide the eye.

46

0 40

Bias = 0 V -0- Bias = 3 V b

Bias = 6 V Bias = 9 V

...-... 0 36 Bias = 1 2 V I • LL •

a.. -<;J- Bias = 1 5 V -0 Q '---'

p� <l> Bias = 1 8 V () c Bias = 20 V cu -'0 cu 0 . 32 a.. cu

0

Temperature (oC)

Fig. 30(b) Capacitance variat ion of sample S6 as a function of temperature for different

applied bias voltages, ranging from 0-20 volts. The frequency is 1 M Hz. Solid lines

are to guide the eye.

To compare the capacitance variation with temperature variation of the normalized

capacitance ( C / C �o ) of S5 and S6 for 0 and 20 volts bias are shown in figures (3 1 a) and 4_ c

-, 1 b), r -.;spectively. These figures summarize the three effects : temperature, applied bias

voltage and frequency. A decrease toward negative values of the capacitance for S6 at 1 00

kHz and 0 bias voltages i s observed in figure (3 1 a) . A more pronounced decrease in the

capacitance of S6 at 1 00 kHz is observed at higher bias (20 volts), as shown in figure (3 1 b) .

The capacitance of S5 starts decreasing at high temperature at 20-volts bias, as shown in

figure (3 1 b) .

47

3 -- S5 ( 1 00 k H z) --*- S5 ( 1 M H z) -<>- S6 ( 1 0 0 k H z)

2 � S6 ( 1 M H z)

o 3

-1 a

-2 L-� ____ � ____ � ______ � ______ � ______ � 4 0 6 0 80 1 00

Te mpe ratu re (aC) 1 2 0 1 4 0

Fig. 3 1 (a ) Nonnalized capacitance o f samples S5 and S 6 a s a function of temperature for

applied bias voltage of 0 volts . The data i s shown for both 1 00 kHz and 1 MHz. Solid

lines are to guide the eye. Error bars are systematic and est imated to be about 5 %.

0 -- S5 ( 1 00kH z) � u ----A- S5 ( 1 M H z) � U -0- S6 ( 1 00 kH z) 0 _1 � S 6 ( 1 M H z )

-2 b

-3 4 0 60 80 1 0 0 1 20 1 4 0

T em peratu re (aC)

Fig. 3 1 (b) N onnalized capacitance of samples S5 and S6 as a function of temperature for

applied bias voltage of 20 volts . The data is shown for both 1 00 kHz and 1 MHz.

Solid l ines are to guide the eye. Error bars are systematic and estimated to be about 5

%.

o

48

The capacitance variation and negative capacitance effect in homogenous

semiconductor (barrier free) has been studied by [ penin, 1 996] . In the case of application of

an alternating voltage across a certain semiconducting material, the equivalent capacitance

measured by a CV meter is given by [ M ajeed Khan, 2005 ] :

_ A aIX r e - ( d ) [ rm - 1 2 2 ] , + UI r ( 1 6)

where, A is the area of the plate capacitor and d is the dielectric thickness. ( r m = e

) 47ra IX

i s the Maxwellian dielectric relaxation t ime, e is the dielectric constant and r is the

dielectric relaxat ion t ime. After substitut ing r m ' Eq. ( 1 6) can be written as:

( 1 7)

At low temperatures ( T« Tc ), the conductivity of Ge-Sb-Te film is low, hence the

second term of Eq . ( 1 7) can be ignored and the capacitance is given by:

C = � e 47r d ( 1 8)

As the temperature increases and gets close to T c, crystallization occurs by the

mechanism of nucleation and growth . In such a mechanism, small crystalline nuclei fonn

initially, which subsequently grow [Erwin, 2006] . I n t his mechanism, the area of the nuclei

i ncreases and the separation between the islands decreases and hence the equivalent

capacitance of Eq . 1 8 increases.

The conductivity of a semiconductor can be increased either by doping or

increasing the temperature. Doping with indium can be considered as a source of free

electrons t hat increase the conductivity of the sample. The nonlinear behavior of the

sample' s capacitance with the applied bias voltage might be related to the nucleation

mechanism. Due to nucleation, two phases will be formed in the sample, amorphous film Q

49

and crystalline islands The interface Uunct ion) between the two phases might behave as a

potential barrier and the charge carriers are accumulated at the junction. At low bias

voltage, the electric filed i s not sufficient for the charge carriers to overcome the barrier.

Therefore, the conduct ivity of the film is not affected. As the applied bias voltage

increa es, the charge carriers gain enough energy from the appl ied electric field and

overcome the barriers and hence the conductivity increases. Therefore, at high temperature,

close to T c, and at high bias voltage the second term of Eq. ( 1 7) becomes more dominant

over the first term and hence the capacitance decreases, as shown in figure (30a). Since the

conductivity of indium doped film ( S6) is much greater than the undoped one ( S 5 ), the

second term of Eq . ( 1 7) increases. This leads to a direct decrease of the capacitance for all

appl ied bias voltages, as shown in figure (30a) . Eventually, t he capacitance becomes

4 Jr ro-DC · h Th h f · .

negati e when 2 2 IS greater t an one. e p enomena 0 negatIve capacItance £( 1 + (u r )

has been observed in other semiconducting materials such as Cd3AS6 [Gould, 1 999], CdTe

[ I smail, 1 996] and CdSe [Oduor, 1 997] .

At high frequency ( 1 M Hz), the second term of Eq. ( 1 7) decreases and the first term

is dominant for t he whole range of temperature and applied voltage. Therefore, at high

frequency the capacitance increases as shown in figures (29b) and (30b).

o

50

CONCLUS I ON

I nvestigat ion of the electrical properties of Ge-Sb-Te thin films using Impedance

Spectroscopy, Current-Voltage and Capacitance-Voltage techniques has been presented .

First, the electrical contacts effects on the electrically measured signal have been

studied using Impedance Spectroscopy technique. Three different configurations of the

electrical contacts are studied . The results show that the effect of the electrical contacts

disappears when the si lver paste is placed only on the gold electrodes away from the thin

film under test . This is related to the fact that the silver paste will not interact with the

material during the heating process. The results show that the order of depositing layers

(material and the gold electrodes) is not important, whereas the position of the silver paste

on the gold electrode is the more significant .

Second the electrical properties of Ge2Sb2Tes thin films usmg Impedance

Spectroscopy and Current-Voltage measurements have been presented. The data shows

similar resistance and relaxat ion frequency temperature dependence with activation

energies of 0 . 36 and 0. 39 eV, respectively. Therefore the corresponding energy gap is

est imated to be 0 .72 eV and 0 .78 eV (This value is in good agreement with the values

reported in literature [Bong Lee, 2005]) . This agreement between electrical and opt ical

measurements confirms that Ge2Sb2 Tes belongs to the chalcogenide family. The measured

amorphous-crystalline transition temperature (Tc) is found to be 1 3 5 °C .

T hird, Capacitance-Voltage measurements have been performed on amorphous

samples of I no.3Ge2Sb2 Tes and Ge2Sb2 Tes t hin films in order to study the indium doping

effect on the electrical properties of Ge2Sb2 Tes . The bias voltage, temperature and

frequency effects have been investigated. The results show that the capacitance of

I no.3Ge2Sb2 Tes sample decreases as the bias voltage increases for the whole range of

temperatures . However the same behavior i s observed for Ge2Sb2 Tes sample at ,

5 1

o

temperatures close to the amorphous-crystalline transition temperature and at high applied

bias voltages ( � 1 5 volts) only. The dependence of the sample's capacitance on bias voltage

and temperature i s attributed to the increase in the free electron concentration as the bias

voltage and temperature increase. For l no.3Ge2Sb2 Tes sample, the indium can be considered

as an additional source of free electrons. This would contribute to further decrease in the

capacitance of l no.3Ge2 Sb2 Tes sample. The negat ive capacitance effect might be attributed

to a significant increase of the film's conductivity due to temperature and applied bias

voltage. The nonlinearity in the capacitance and conductivity could be related to the -nucleat ion mechanism as the temperature gets close to the amorphous-crystall ine transition

temperature.

Chalcogenide glasses are versati le semiconductor and dielectric functional materials.

They possess a number of wel l known pecul i ar properties l i ke photo-induced change of

opt ical parameters and electrical switching. Research options st i l l abound to ident ify al loys

with very low switching voltages, and to demonstrate a sufficient number of read-write and

erase cycl e Seeking new materia ls with very low voltage switching and invest igat ing the

optical and electrical properties of these materia ls are attracting much attention.

Moreover, investigations of photo-physical processes in amorphous chalcogenide l ayers

were extended towards the nanostructure. The nano-Iayered fi lms were in the focus of

development of a new photosensit ive, optical recording media. These topics provide future

research opportunit ies.

52

References

Barsoukov E and M acdonald 1 . R . , " Impedance Spectroscopy: Theory, Experiment, and Applications", Wi ley-Interscience 2nd edition, (2005 ).

Bong-Sub Lee, Abelson 1 . R , Bishop S. G. , Kang D. H , Cheong B . , and Kim K B . , J . pp. Phys 97, 093 509. (2005)

Chen M . , Rubin K A , and Barton R W ., Appl . Phys. Lett 49, 502, ( 1 986)

El l iott R , "Physics of Amorphous M aterials" , 2nd edition ( Longman Group, UK, 1 990).

Erwin R. Meinders, Andrei V. Mij iritsk i i , L iesbeth van Pieterson, and Matthias Wuttig, "Optical Data Storage , Phase change Media and Recording" Vo14. (Springer, 2006) . -Friedrich I . , Weidenhof V. , Njoroge W . , Franz P . , and Wuttig M . , J . Appl . Phys. 87, 4 1 30, (2000) .

Ghamlouche H , Mahmoud S .T . , Qamhieh N . , Ahmed S . , and Al-Shamisi H . , Physica Scripta 77 1 , (2007).

Gonzales-Hernandez J . , Prokhorov E . , and Vorobiev Y, Vacuum Sci. Techno. A. 1 8, 1 694, (2000) .

Gould R D., and Din M . , Superficies y Vacio, 9, 230, ( 1 999).

I smai l B. B. and Gould R . D. , Proc. S PIE, 2780, 46, ( 1 996).

Kastner M. , Adler D . , and Fritzsche H . , Phys. Rev. Lett . 3 7, 1 504 ( 1 976).

Kazunari Okada, Toshimasa Sekino, " Agi lent Technologies I mpedance Measurement Handbook", (2003 ) .

Koj ima R . , Okabayashi S . , Kashihara T . , Horai K , Matsunaga T. , Ohno E . , Yamada N . , and Ohta T . , J pn . J . App 1 . Phys. 3 7, 2098, ( 1 998) .

Koj i ma R, and Yamada N. Jpn. J . Appl . Phys. Part 1 40, 5930, (200 1 ).

Kolobov A. V. , Fons P . , Frenkel A . , Ankudinov A. L . , Tominaga J , and Uruga T . , Nat . M ater. 3 , 703 , (2004).

Koo Y. W., Kim 1 . H . , Cho W. 1 . and Chung H B., M icroelectron. Eng. Online: doi :

1 0. 1 0 1 6/j . mee. 0 1 . 245 , (2007).

Kusada H, Hosaka T . , Kojima R . , and Yamada N. , Proc. 18th Symp. PCOS2005 32-35, (2006) .

Lee H . Y . , Kim J . K . and Chung H . B . , 1 . Non-Cryst . Sol ids 279 209, (200 1 )

Lee H Y. , Park S . H . Chun J . Y . , and Chung H . B . , J . AppI . P hys. 83 538 1 , ( 1 998).

53

Liuera M , and Chen M . , Mater Res. Soc. Bul l . 1 5 40, ( 1 990).

L iq iu Men and Fuxi Gan, Optics Communications 1 45, 2 1 , ( 1 998).

Luo M. B , and Wuttig M . , Adv Mater. 1 6, 439, (2004) .

Macdonald R . J . , " Impedance Spectroscopy Emphasizing Sol id M aterials and Systems", (John Wi ley & Sons, Canada, ( 1 987) .

M aj eed Khan M . A , Zulfequar M., and Husain M., Physica B 366, 1 , (2005 ).

Mart ijn H. R Lankhorst, Liesbeth van Pieterson Mark van Schijndel, Ben A J . Jacobs, and Jan C . . Rijpers, Jpn . 1 . App J . Phys. Part 1 , 42,863, (2003 ).

M orales- anchez E . , Prokborov E.F. , Mendoza-Galvan A, and Gonzalez-Hernandez 1 . , Vacuum 69,36 1 , (2003 ). -Mott F and Davis E . A , "Electronic Processes in Non-Crysta l l ine Materials", 2nd edition (Clarendon Press, Oxford, 1 979).

M ort N. F., Davis E . A and Street R .A, Phi l . M ag. 32, 96 1 ( 1 975) .

Nakayama K , Kojima K . , Imai Y , Kasai T, Fukushima S . , Kitagaw A , Kumeda M . , Kakimoto Y . and uzuki M , Jpn . J . App l . Phys. 42 404, (2003).

Oduor A. 0., PhD Thesi s, Keele University, UK ( 1 997).

Ohno E . , Yamada N , Kurumizawa T , Kimura K . , and Takao M ., Jpn. 1 . AppJ . Phys. Part 1 28, 1 235 , ( 1 989).

Okamoto H, Kida H . , and Hamakawa Y, Philos. M ag. B 49, 23 1 , ( 1 984).

Ovshinsky S . R . , Phys. Rev. Lett . 2 1 , 1 450, ( 1 968).

Ohno E., Yamada N., Kurumizawa T. , Kimura K, and Takao M . , Jpn. 1 . Appl . Phys. Part 1 28, 1 23 5, ( 1 989).

Padma R . , Suvama K, Raghavendra RAO, and Subbarangalah K, Bul l . Mater. Sci . , Vol . 25, No. 7, pp. 647-65 1 , December (2002) .

Penin A . , Semiconductors 3 0, 340, ( 1 996).

Qamhieh N . , PhD Thesis, Katholieke University Leuven "Electrical Properties of

Amorphous Germanium Selenides", ( 1 996) .

Qamhieh N . , Gham louche H . , Mahmoud S . T , Al-Shamisi H . and Ahmad S . , 1 .

Optoelectronics and Advanced M aterials, 9 , 3 1 5 7, (2007).

Saleh T Mahmoud, Ghamlouche H, Qamhieh N . , and Ahmed S., 1 . Non-Cryst . Sol ids 3 54,

1 976 (2008).

Saleh T M . , Ghamlouche H, Qamhieh N. and Al-Shamisi H, Appl ied Surface Sci�nce

253, 7242, (2007).

54

atoh 1 , and Yamdda . , Proc S PI E 4085, 283, (200 1 )

Street R A, Phys. Rev. B 1 7, 3984, ( 1 978).

Tanaka H., ishihara T., Ohtsuka T . , Morimoto K , Yamada N., and Morita K, Jpn. 1 . App l . Phys. 4 1 L 1 443 (2002).

Tauc J , " Amorphous and Liquid Semiconductors" ( plenum Press, New York, 1 974).

Tomigava J . , and Atoda N , Jpn. 1 . Appl . Phys. 38 , L322, ( 1 999).

Wang K , Steamer c . , Wamwangi D , Ziegler S . , and Wuttig M . , App! . Phys. A 80, 1 6 1 1 (2005 )

Weidenholf Y . , Friedrich 1 . Ziegler S . , and Wuttig M . , 1 . AppJ . Phys. 86, 5879, ( 1 999).

Weidenholf Y . , Friedrich 1 . , Ziegler S . , and Wuttig!\1., 1 . AppJ . Phys. 89, 3 1 68, (200 1 ) .

Wuttig M . , and Noboru Y , Review Art icle, Yo1 .6, (Nature Publ ishing Group, 2007).

Wuttig M . Nature Materia ls 4, 265, (2005) .

Yamada N. , E ij i Ohno, Kenichi N ishiuchi, Nobuo Akahira, and Masatoshi Takao, 1 . App. Phys. 69, 2849, ( 1 99 1 ).

Yamada N , Ohno E . , Akahiran N . , i shiuchi K , Nagata K . , Takao M . , Jpn. 1 . Appl . Phys. 26, 6 1 , ( 1 987) .

Yamada N. , Ohno E . , Nishiuchi K, Akahira N. , and Takao M., 1 . App J . Phys. 69, 2849, ( 1 99 1 )

Yamada Takenaga M. , and Takao M . , Proc. SP IE 695, 79, ( 1 986).

Yusu K, N akai T . , Ashida S . , Ohmachi N . , Mori shita N . , and Nakamura N., Proc. E'PCOSO ; avai lable at http ://www . epcos.org, (2005) .

Zhiwei Sun, J ian Z hou, and Rajeev Ahuja, Phys. Rev. Lett . 96, 055507, (2006).

5 5

lAl �)�I �YJ ...r-�I �I c!.J .l,jc �y:J1 �J.jJjiSlYl j.!>y ��j .)) .2lli � � y.J..J -:..Jy! Y-Y-J � �� ��, liAJ ;�I ..::l.i.J� I.J� y-�,!I �1.......;,,1 J.U! 1 3Ge� b� Te. � � �I � �YJI ��I 0.1 .� r-fi .)! J,..oJ 01 .)) lno.3Ge2 b1 Te ¥ ��I �I � l.;�

UWI 0-"" J..-W'JI ; J� � Y u--- Li.).fol LJS �WI .J»' � �) �.)d � jlll �.J ��I

�:U.J.JI �) �.JJ..!YJI

-

r l�1 � f.ull lf y.-a-:JI �I � � I� u� L...li:i ;; � YI �I ul�1 ..,-! ..b.)

�j �¥ lY"1�IJ �l'1 Oolle 'I � ( CD ) u,� ..;a1.;J1 � � u�fil\Sl1 C4-j .llr

� � lYL..YI yt. i.j)yJl �I �Iy' �j.J4l IJ �:U� .;:LII U:J1 0F. �4W1 �I 0) DVD

J.:.,.b <.S�J tjfJ' lfJ� � J.u..., � � ol ly> � � � � � ·lf�1 �I

� ;;.l..Aia..o J.,4il1 � � fJ' lf� Ge-Sb-Te �y ul �I �\..:ij up\ �J ' wl ... � ..bUh)l\

.-Ge-Sb-Te OolL. L>--- 0�1 �)I "L::u.1l ���I lY"I�1 Lljol JL=JI �I lfo...P-:l

�pl �I L";-!J �I �.;bJ ' ( I-V) �pl �I (.i�J }.flll �y, ' �4hJI '(!JUI rl��

(C-V)

� .

;;}_..';,'jl ..,.]c Ge-Sb-Te �)I ,,�I F � �yJl ���I u:::L::.)1 �t �Ij.l WAi iJ l' 4...jL 'lilli w · · 1:\ �lJ . . < 11 w� n � ul1.J · - G� - J)l;. . �liJl �lJ . . <11 .Y--J • � � . - . � Y' -j.jJ � lY' - . �

. �)I ,,�I � U\ 0J� ��I ���I �I � �I U�I �J Ij) ���I u:::L::.}l -illj �! -UL:. � . �I � J)l;. �)I "t...';,iJ1 idA � �I U� I Jc-tii r.lc � �I I� .l.7-! Y. �IJ ( � �I yWYIJ �)I "t...';,iJI • .llo ) uli,JJ1 �ji � �yll �I rJc. �t:i.:J1 W:......:.:.JI

�liJl � ��I yWYI � �1 0�1

: �\.J • U)k. . -.J .ut...:Li � . G ( I -V) . . <1 1 , - I I - � lilli �1..c.l'1 --ub 'u � I � . LJ . t..r!� � (j...r- C"" j _ J , - � J

lf Jt..-u �IJ �I �UJI & J)l;.. l.>..o .jlyJI 4j.l � ( Relaxation frequency ).l.lylIJ .4...., JliJl

oiA. w.I.,s lJ.Jfi! 0. 72J 0 .78 �UJI o� � j� dlJ� �?l'jiJ 'Jljill � w.I.,s UJfi! 0 39 J 0 . 36 Ge2Sb2TeS • .llA u\ .loS): liA , wJ� UJ�! 0. 7 lfJt..-u �IJ � �yJl o �1 � 0!lji �I

�I o)yJI 4j.l u� �1 0';-! c-" j4J.l1 Glj.l �JI � dlJj �! -UL.o� , �fil\Sl1 �w � . 1 3 5 DC lfJ\...;J �j.J4l1 �! �j�)l\1 �I l.>" o.lWI Jtllil lA-lL �

. t:iJ\.J • l.>..o o.).J-4l... j-:?ill �)I ,,�I l.>..o u� (C-V) ��I �I c-" �I J..r&l �I.).l �.?I

..J ���I lY"I y--::J1 � ry-,l...j'jl 4jL.:.:,! �'U uy.-J dljjJ I Oo.3Ge2Sb2Te5J Ge2Sb2Te50.lL....

<.r.�1 �1 .l.,!1'y-! �l....i:i:i l oo3Ge2Sb2Tes � ���I �1 0� �LUlI �) Ge2Sb2Tes

.l..lc. � r�.).j'll � 4l -Ut...:.:..... �I u4-\1 � � �..,LJl liA J .�liJl oJrJl U�j.l � 1 5 l.>..o �1 .l.P. 0..r! -lLJ �j.J4lI �! �.)�)lJ1 �1 0-0 Jp:ill .).? 4.).l 0-0 �y.ll o)yJl U�.).l

o�1 ��\ wl) .... ';1 �4-4LJI w�I�1 oJ�

� WI \..j � � y.

: �1jJ\ 0.0 �� ��.J �WJI � � �

i�1 �.JLl1 �1.JLa'i1 �4- �} J1.,..J1 tw.� .J ,.� � �WI 4JJ uS- �I �� ��I

2008