A method to enhance the data transfer rate of eutectic Sb-Te phase-change recording
media
Tung-Ti Yeh, T.-E. Hsieh, and Han-Ping D. Shieh
Citation: Journal of Applied Physics 98, 023102 (2005); doi: 10.1063/1.1957132
View online: http://dx.doi.org/10.1063/1.1957132
View Table of Contents: http://scitation.aip.org/content/aip/journal/jap/98/2?ver=pdfcov Published by the AIP Publishing
Articles you may be interested in
Transient phase change effect during the crystallization process in phase change memory devices Appl. Phys. Lett. 94, 243504 (2009); 10.1063/1.3155200
Effect of indium on phase-change characteristics and local chemical states of In–Ge–Sb–Te alloys Appl. Phys. Lett. 93, 021905 (2008); 10.1063/1.2959730
Ultrafast crystallization and thermal stability of In–Ge doped eutectic Sb 70 Te 30 phase change material J. Appl. Phys. 103, 043501 (2008); 10.1063/1.2841724
Kinetics of crystal nucleation in undercooled droplets of Sb- and Te-based alloys used for phase change recording
J. Appl. Phys. 98, 054910 (2005); 10.1063/1.2037870
Atomic force microscopy measurements of crystal nucleation and growth rates in thin films of amorphous Te alloys
Appl. Phys. Lett. 84, 5240 (2004); 10.1063/1.1764591
A method to enhance the data transfer rate of eutectic Sb-Te
phase-change recording media
Tung-Ti Yeh and T.-E. Hsieha兲
Department of Materials Science and Engineering, National Chiao-Tung University, 1001 Ta-Hsueh Road, Hsinchu 30050, Taiwan, Republic of China
Han-Ping D. Shieh
Department of Photonics and Institute of Electro-Optical Engineering, National Chiao-Tung University, 1001 Ta-Hsueh Road, Hsinchu 30010, Taiwan, Republic of China
共Received 23 August 2004; accepted 24 May 2005; published online 19 July 2005兲
This work describes the effect of nitrogen doping to eutectic Sb-Te phase-change materials in order to enhance the speed of the amorphous-to-crystalline phase transformation. When nitrogen at a sputtering gas flow ratio of N2/Ar= 3% was doped in the eutectic Ge-In-Sb-Te recording layer, the data transfer rate was increased up to 1.6 times. When thin GeNxnucleation promotion layers were
further added in below and above the recording layer, an overall enhancement up to 3.3 times in data transfer rate was achieved. The nitrogen contents corresponding to the N2/Ar flow ratios共N2/ Ar = 0%–10%兲 were calibrated by electron spectroscopy for chemical analysis. Transmission electron microscopy revealed that nitrogen doping was able to promote the phase transformation by generating numerous nucleation sites uniformly distributed in the recording layer and hence increased the recrystallization speed. © 2005 American Institute of Physics.
关DOI: 10.1063/1.1957132兴
I. INTRODUCTION
A large storage capacity and a high data transfer rate are always demanded for optical recording media. The phase-change recording materials with higher amorphous-crystalline transition speed are hence required. Currently the eutectic Sb-Te system, or called the fast-growth phase-change alloy, has received much attention because of its good signal properties and high recrystallization speed when short-wavelength laser and high numerical aperture 共NA兲 lenses are used. The eutectic Sb-Te system is termed as the fast-growth material since its recrystallization is initiated from the crystalline-amorphous interface and the amorphous mark shrinks as the grain growth propagates toward the cen-ter of the signal marks,1–4as illustrated in Fig. 1共a兲. Adding foreign element in the eutectic Sb-Te system such as Ag, In, and Ge is a common way to improve the recrystallization characteristics and optical properties of optical disks.5 Not only the composition of the recording material but also the nucleation-promoting layer/dielectric layers enclosing the re-cording layer play important roles in the recrystallization process.6,7 The surface reactivity and chemical affinity are the major factors that enhance the recrystallization speed. In this work, we doped nitrogen during the sputter deposition of the recording material and inserted GeN nucleation-promotion layers above and below the recording layer in order to further enhance the data transfer rate of eutectic Sb-Te phase-change optical disk. The signal properties and data transfer rate of optical disks were investigated by using a dynamic test. The nitrogen contents doped in the recording layer at various sputtering N2/Ar flow ratios were determined
by electron spectroscopy for chemical analysis共ESCA兲 and transmission electron microscopy 共TEM兲 was utilized to ex-amine the change in the microstructure induced by nitrogen doping.
II. THE MODEL OF RECRYSTALLIZATION PROCESS
Figure 1共b兲 illustrates the model applied to enhance the velocity of phase transformation. It was realized by doping nitrogen in the eutectic Ge-In-Sb-Te recording layer as well as in the nucleation promotion layers. Nitrogen doping gen-erates numerous nanometer-scale precipitates uniformly dis-tributed in the recording layer. They provide the preferential sites for amorphous-crystalline transition in the recording layer so that the edge of marks is not the only site to initiate the recrystallization during signal recording. The tiny
pre-a兲Author to whom correspondence should be addressed; electronic mail: tehsieh@cc.nctu.edu.tw
FIG. 1. 共a兲 In phase-change recording materials with a growth-dominated crystallization mechanism, crystallization starts from the crystalline-amorphous mark edges and proceeds to the center of the mark. 共b兲 In nitrogen-doped fast-growth materials, it is assumed that the phase transition initiates both at the crystalline-amorphous mark edges and precipitates within the mark.
0021-8979/2005/98共2兲/023102/5/$22.50 98, 023102-1 © 2005 American Institute of Physics
cipitates not only induce the heterogeneous nucleation, they also shorten the distance of grain growth required to com-plete the transition. Hence the velocity of amorphous-crystalline transition is promoted. In addition, the precipi-tates must remain stable under laser irradiation to ensure good disk signal properties after periodic signal recording.
III. EXPERIMENTAL PROCEDURES
Figure 2 illustrates the multilayer structures of the opti-cal disk used in this work. At first, disk samples with con-ventional four-layer stack were prepared using a SFI 共Sur-face Inter共Sur-face Corp.兲 sputtering system at a background pressure better than 1⫻10−6torr. The multilayer structure was deposited on 0.6-mm-thick polycarbonate共PC兲 substrate in the sequence of ZnS-SiO2共55 nm兲/Ge-In-Sb-Te 共GIST兲-共N兲x共16 nm兲/ZnS-SiO2共11 nm兲/Al-Cr共133 nm兲, as shown in Fig. 2共a兲. During the deposition of the recording layer, the N2/Ar flow ratio was adjusted to values of 0%, 0.5%, 1%, 3%, 5%, and 10% to obtain undoped and nitrogen-doped specimens. Then the optimal composition of GIST-共N兲xwas
chosen to be the recording layer of the optical disk with
six-layer stack shown in Fig. 2共b兲. The sputtering conditions and designated sample numbers are listed in Table I.
After the optimum initialization, the disk samples were sent to a dynamic tester共DDU1000, PULSTEC Co.兲 having a pickup head with a 650-nm laser diode and a NA= 0.6 objective lens to evaluate their signal properties. Figure 3 shows the writing strategy used in this study. At the begin-ning of the dynamic test, the disks were written at various laser powers 共7–15 mW兲 with fixed erasing power 共Pe
= 6 mW兲 and reading power 共Pr= 0.7 mW兲 at various linear
velocities to identify the optimum writing power共Pw兲. Then
the tracks recorded with 8T signals were erased by irradiat-ing a dc laser beam of various laser powers共3–7 mW兲 at the same linear velocity used for recording. The carrier-to-noise ratio 共CNR兲 was measured before and after the erasing pro-cess to determine the attenuation of 8T signal carrier, and the value of the attenuation of 8T signal carrier is defined as the dc erasability.
The nitrogen contents of GIST-共N兲xrecording layer
cor-responding to various N2/Ar flow ratios were calibrated by ESCA 共Thermo VG-Scientific, Microlab 350兲. The method report by Nokukuni et al.8and Chen et al.9was adopted to prepare the plan-view TEM 共PTEM兲 specimens. After re-moving the PC substrate, the disk sample was cut into small pieces using a pair of scissors. A 3M tape was applied to the disk to peel off the Al-Cr reflection layer. After dissolution of the PC substrate by CH2Cl2 solution, the specimen was mounted on the copper mesh and transferred to a TEM 共Phil-ips, Tecnai 20 TEM兲 for microstructure observation.
FIG. 2. The cross-sectional structures of an optical disk with共a兲 four-layer and共b兲 six-layer stack.
TABLE I. Sputtering conditions and designated sample numbers.
Sample
no. N2/Ar ratio Target type
Sputtering pressure 共mTorr兲 Sputtering power 共W兲 0 / 10 ZnS-SiO2 3 250共rf兲 0 / 10 Al-Cr 3 400共dc兲 20/ 10 Ge 3 200共rf兲 N000 0 / 10 GeInSbTe 3 50共rf兲 N005 0.05/ 10共0.5%兲 GeInSbTe 3 50共rf兲 N010 0.1/ 10共1.0%兲 GeInSbTe 3 50共rf兲 N030 0.3/ 10共3.0%兲 GeInSbTe 3 50共rf兲 N050 0.5/ 10共5.0%兲 GeInSbTe 3 50共rf兲 N100 1.0/ 10共10.0%兲 GeInSbTe 3 50共rf兲
FIG. 3. Writing strategy for 8T signal.共Tsfp: start-up delay value for the first pulse. Ttop: ending delay value for the first pulse, Tsmp and Tmp: delay time, Tslp: start-up delay value for the last pulse, Tlp: ending delay value for the last pulse, and Tle: start-up delay for the off pulse兲
023102-2 Yeh, Hsieh, and Shieh J. Appl. Phys. 98, 023102共2005兲
IV. RESULTS AND DISCUSSION A. dc erasability
Figure 4 shows the dc erasability of 8T signals as a function of linear velocity and/or data transfer rate of various samples. The outputs of dc laser power were adjusted so that the dc erasability at each linear velocity reached the maxi-mum. As we know, the dc erasability higher than 25 dB is required for the direct overwriting共DOW兲 of optical disks.10 Furthermore, dc erasability of each disk sample significantly decreased with an increase in linear velocity. This result was attributed to the fact that the laser irradiation time at a high linear velocity is shorter than that at a low linear velocity. Laser irradiation time at a specific linear velocity is ex-pressed as11
T =D
V, 共1兲
where T is the laser irradiation time, D is the diameter of the laser spot, and V is the linear velocity of the disk. Equation 共1兲 depicts that a recording material must possess a suffi-ciently high recrystallization speed for phase change so that amorphous marks can be effectively erased at a specific lin-ear velocity. Therefore, it is necessary for a phase-change optical disk to have a high recrystallization speed to realize a high data transfer rate. From the results of dc erasability test shown in Fig. 4, the doping-free disk sample 共N000兲 just passes the test requirement at a linear velocity below 5.3 m/sec. However, the nitrogen-doped samples, for instance, the N030 and GeN/N030/GeN disk samples successfully pass the dc erasability test requirement at a linear velocity of 8.8 and 12.3 m/sec, respectively. This means that the recrys-tallization speeds of N030 and GeN/N030/GeN disk samples were approximately 1.6 and 3.3 times higher than that of the sample free of nitrogen. Apparently, an appropriate amount of nitrogen doping and an effective nucleation-promoting layer benefited the data transfer rate and/or recrystallization speed of GIST phase-change optical disks. Miao et al.12 pro-posed that the GeN layers might cause heterogeneous
nucle-ation at the interface of GeN and phase-change recording layer to enhance the recrystallization speed.
B. Jitter values
The method of digital versatile disk共DVD兲 jitter evalu-ation was adopted to measure the jitter for all intervals be-tween the data and the clock.13The random data was written in the track of N000, N030, and GeN/N030/GeN disk samples at 3.5-m/sec linear velocity for measuring the jitter values. A low jitter implies a very low error when 8T signal marks were written in the disk samples. Figure 5 shows the jitter values of N000, N030, and GeN/N030/GeN disk samples for various DOW cycles. It was found that N000 and N030 samples have quite similar jitter values at the same number of DOW cycles. This implies that a relatively small amount of nitrogen doping would not deteriorate the signal properties of optical disks. GeN/N030/GeN sample exhibited higher jitter value than N000 or N030 samples at the same number of DOW cycles. This might result from the differ-ence in the thermal properties between GeN/N030/GeN and N000 共or N030兲 disks because of the different disk struc-tures.
C. Nitrogen concentration analyzed by ESCA
The nitrogen contents of GIST-共N兲xrecording layer
cor-responding to various N2/Ar flow ratios共0%–10%兲 shown in Fig. 6 were calibrated by ESCA. It was found that the nitro-gen content increases with the increase of N2/Ar flow ratio and similar result was reported by Jeong et al.14 when Ge2Sb2Te5recording material was sputtered at the sputtering gas flow ratio of N2/Ar艋10%.
D. TEM observation
The micrographs of 8T signal marks in disk samples of N000 and N030 written at a linear velocity of 7 and 10.5 m/sec are shown in Figs. 7共a兲 and 7共b兲. The lamellar-like structure commonly seen in a eutectic recording alloy was observed in the crystalline area of both samples. In addition to the higher degree of structure irregularity, we observed
FIG. 4. dc erasability of 8T signals as a function of linear velocity and/or data transfer rate of various samples共the 8T carriers were written on the initialized track at optimum writing power before erasing兲.
FIG. 5. Jitter values of random data vs the number of DOW cycles at a linear velocity of 3.5 m/sec for N000, N030, and GeN/N030/GeN disk samples.
tiny precipitates uniformly distributed in N030 sample at high magnification. These precipitates with size less than 10 nm were considered to be nitride compounds, e.g., Ge-N, Sb-N, Te-N, In-N, etc. The 8T amorphous marks were formed at an optimum Pw/ Peand the amorphous marks in
N000 sample possess a more distinctive shape than those in N030 sample. Figures 7共c兲 and 7共d兲 show the micrographs of the residual marks in N000 and N030 samples, respectively, erased at the linear velocities of 7 and 10.5 m/sec. As shown in Figs. 7共c兲 and 7共d兲, the residue of marks can be easily found in the undoped sample while in the nitrogen-doped specimens, the residual marks were hardly observed and en-tirely different from that of the undoped sample, as shown in Fig. 7共c兲. This explains why the dc erasability of an undoped sample is inferior to that of nitrogen-doped samples, as shown in Fig. 5. The microstructure observation also re-vealed that the recrystallization behavior of the nitrogen-doped specimens should be different from that of the un-doped specimen due to the existence of tiny nitride precipitates.
E. Calculation of activation energy Eaof phase
transition
The eutectic GIST alloy is termed as the fast-growth material since its recrystallization is initiated from the crystalline-amorphous interface and the amorphous mark shrinks as the grain growth propagates toward the center of the mark, as illustrated in Fig. 1共a兲. The velocity of the grain growth derived from the net jump frequency of atoms across the amorphous-crystallization interface can be expressed as3
V共T兲 = V0e−Ea/R⌬T共1 − e−⌬gac/RT兲, 共2兲
where V0 is a preexponential factor, Eais the activation
en-ergy of transition from the amorphous to the crystalline state, ⌬gac is the free energy difference between an atom in the
amorphous state and that in the crystalline state, R is the gas constant, and ⌬T is the temperature difference between the interface temperature and the glass transition temperature. Equation 共2兲 implies that low Ea value benefits the phase
transition from the amorphous state to the crystalline state.
We measured five different crystallization temperatures at different heating rates 共5, 10, 20, 40 and 80 °C/min兲 using differential scanning calorimetry共DSC, Dupont 2000 Series兲 to fit into the Kissinger plot15,16shown in Eq.共2兲 to calculate the activation energy for crystallization. The Kissinger equa-tion is
FIG. 6. Nitrogen content of GIST-共N兲xrecording layer vs N2/Ar gas flow ratio of sputter deposition.
FIG. 7. The micrographs of 8T signal marks of共a兲 N000 disk sample writ-ten at a linear velocity of 7 m/sec and共b兲 N030 disk sample written at a linear velocity of 10.5 m/sec. The micrographs of residual amorphous marks of共c兲 N000 disk sample erased at a linear velocity of 7 m/sec and 共d兲 N030 disk sample erased at a linear velocity of 10.5 m/sec. A local magnified picture of共b兲 is attached at the lower right-hand corner of the micrograph.
023102-4 Yeh, Hsieh, and Shieh J. Appl. Phys. 98, 023102共2005兲
ln
冉
␣ TP2冊
= −冉
Ea R冊
TP −1 + c, 共3兲where␣ is a heating rate, TP is the crystallization
tempera-ture, Eais the activation energy for phase transformation, R
is the Boltzamann constant, and c is a constant. The activa-tion energy of transiactiva-tion from the amorphous to the crystal-line state, Ea, can be calculated from the slope of the plot of
ln共␣/ Tp
2兲 vs 1/T
p shown in Fig. 8. For N000, N005, N010,
and N030 samples, Ea were 2.462, 2.521, 2.642, and 1.686
eV, respectively, as shown in Table II. According to the re-crystallization model shown in Fig. 1共b兲, nitrogen doping might generate numerous nanometer-scale precipitates uni-formly distributed in the GIST recording layer. In addition to the amorphous-crystalline edge of marks, they were also the preferential sites for the amorphous-crystalline transition of the recording media. When sufficient numbers of tiny pre-cipitates provided the sites for heterogeneous nucleation, a substantial decrease of Eawas hence observed, e.g., in N030
sample. Furthermore, uniform distribution of precipitates also shortened the distance of grain growth required to com-plete the transition so that an enhancement of recrystalliza-tion speeds/data transfer rate was obtained.
V. CONCLUSIONS
We added nitrogen during sputter deposition of Sb-Te phase-change recording material in order to enhance the rate of amorphous-to-crystalline phase transformation. As re-vealed by TEM observation, nitrogen doping was able to
provide numerous nanometer-scale precipitates in the record-ing layer, which could serve as the preferential sites of het-erogeneous nucleation for amorphous-crystalline transition. In addition to the edge of the signal marks, these tiny pre-cipitates were also the sites to initiate the recrystallization of recording media so that the data transfer rate of optical disks was effectively increased. Nitrogen doping at a sputtering gas flow ratio of N2/Ar= 3%共corresponds to about 3 at. % of nitrogen doping as determined by ESCA兲 might enhance the data transfer rate of an optical disk up to 1.6 times without severely damaging the signal jitter values. However, the disks failed the dynamic tests when too much nitrogen 共N2/ Ar艌5%兲 was introduced. Further enhancement of the data transfer rate up to 3.3 times was achieved in the phase-change optical disk in which the GIST-共N兲xrecording layer
was enclosed between two GeNxnucleation promotion
lay-ers. The implantation of new nucleation sites not only accel-erated the recrystallization speed of amorphous-crystalline transition via heterogeneous nucleation, it also shortened the distance of grain growth so that the phase-change optical disks with high data transfer rate could be realized without a drastic change of disk structure.
ACKNOWLEDGMENTS
This work were supported by the Ministry of Education of the Republic of China with the Academic Center of Ex-cellence in “Photonics Science and Technology of Tera Era” under Contract No. 89-E-FA06-1-4 as well as the National Science Council of the Republic of China under contract No. NSC92-2216-E-009-013.
1R. E. Reed-Hill, Physical Metallurgy Principles, 2nd ed.共Van Nostrand, New York, 1991兲, p. 408.
2D. A. Porter and K. E. Easterling, Phase Transformations in Metals and
Alloys, 2nd ed.共Chapman and Hall, New York, 1992兲, p. 263.
3E. R. Meinders, H. J. Borg, M. H. R. Lankhorst, J. Hellmig, and A. V. Mijiritskii, J. Appl. Phys. 91, 9794共2002兲.
4L. H. Chou and Y. Y. Chang, Jpn. J. Appl. Phys., Part 1 40, 1272共2001兲. 5T. Kato, H. Hirata, T. Komaki, H. Inoue, H. Shingai, N. Hayashida, and H.
Utsunomiya, Jpn. J. Appl. Phys., Part 1 41, 1664共2002兲.
6G. F. Zhou and B. A. J. Jacobs, Jpn. J. Appl. Phys., Part 1 38, 1625 共1999兲.
7B. Tieke, M. Dekker, N. Pfeffer, R. V. Woudenberg, G. F. Zhou, and I. P. D. Ubbens, Jpn. J. Appl. Phys., Part 1 39, 762共2000兲.
8N. Nobukuni, M. Takashima, T. Ohno, and M. Horie, J. Appl. Phys. 78, 6980共1995兲.
9H. W. Chen, T. E. Hsieh, J. R. Liu, and H. P. D. Shieh, Jpn. J. Appl. Phys., Part 1 38, 1691共1999兲.
10H. Inoue, H. Hirata, T. Kato, H. Shingai, and H. Utsunomiya, Jpn. J. Appl. Phys., Part 1 40, 1641共2001兲.
11N. Yamada, Proc. SPIE 3109, 28共1997兲.
12X. S. Miao, T. C. Chong, L. P. Shi, P. K. Tan, J. M. Li, and K. G. Lim, Jpn. J. Appl. Phys., Part 1 40, 1581共2001兲.
13DVD Forum, The DVD Specifications for Re-recordable Disc 共Version 1.0, 1999兲, p. PHX-G2.
14T. H. Jeong, M. R. Kim, H. Seo, J. W. Park, and C. Yeon, Jpn. J. Appl. Phys., Part 1 39, 2775共2000兲.
15H. Seo, T. H. Jeong, J. W. Park, C. Yeon, S. J. Kim, and S. Y. Zkim, Jpn. J. Appl. Phys., Part 1 39, 754共2000兲.
16H. E. Kissinger, Anal. Chem. 29, 1702共1957兲. FIG. 8. The Kissinger plot of N000, N005, N010, and N030 samples.
TABLE II. Calculation of the activation energy for N000, N005, N010, and N030 samples.
Sample no. N000 N005 N010 N030
Activation energy共eV兲 2.462 2.521 2.642 1.686
