**3. Time-resolved investigation of nanosecond crystal growth in rapid-phasechange materials**

Using the X-ray pinpoint structural measurement system described in previous section, we investigated the crystallization process of GST and AIST (Fukuyama et al., 2008a), which are thought to exhibit different crystallization behavior. To reveal the crystallization process of the amorphous phase in these rapid-phase-change materials, it is necessary to know (i) the time constants of both crystallization and optical reflectivity changes, and (ii) crystallization behavior. We therefore employed avalanche photodiode (APD)/multi-channel scaling (MCS) measurement with a time resolution of 3.2 ns coupled with optical reflectivity measurement for (i) and imaging plate (IP)/pump-probe measurement of a time resolution of 40 ps for (ii) as schematically shown in Fig. 9(a). Figure 9(b) shows the experimental method and the time chart for pump-probe measurement using 40 ps X-ray pulses and synchronous femtosecond laser pulses. The DVD samples were prepared by depositing GST and AIST on a SiO2 glass substrate (12 cm diameter and 0.6 mm thick) with a 2-nm-thick 80 mol% ZnS-20 mol% SiO2 cap layer. The thickness of the sample was 300 nm, which gave sufficient X-ray diffraction intensities for structure characterization and sufficient stability of the amorphous phase to avoid spontaneous crystallization.

Fig. 9. (a) Schematic diagram of APD/MCS and IP/pump-probe measurement and (b) scheme of DVD rotating system and the time chart of pump-probe measurement.

Figure 10 shows the time-profiles of the X-ray diffraction intensity of Bragg peaks (black and blue lines) together with those of the optical reflectivity profiles (red lines) of (a) GST and (b) AIST, respectively, measured by APD/MCS measurement. These show good

Here we briefly summarize the X-ray pinpoint structural measurement system for investigation of optical recording process (see Fig. 2). The time-resolution and time precision are ~40 ps and ±8.4 ps, respectively. The spatial resolution is determined by the area of the laser beam of 30 μm, which is larger than the X-ray beam size of 3 μm. The maximum repetition rate is 1 kHz, in which the fresh sample area on the rotating disk appears at the irradiating position for every measurement. All parameters shown here are described in full

**3. Time-resolved investigation of nanosecond crystal growth in rapid-phase-**

Using the X-ray pinpoint structural measurement system described in previous section, we investigated the crystallization process of GST and AIST (Fukuyama et al., 2008a), which are thought to exhibit different crystallization behavior. To reveal the crystallization process of the amorphous phase in these rapid-phase-change materials, it is necessary to know (i) the time constants of both crystallization and optical reflectivity changes, and (ii) crystallization behavior. We therefore employed avalanche photodiode (APD)/multi-channel scaling (MCS) measurement with a time resolution of 3.2 ns coupled with optical reflectivity measurement for (i) and imaging plate (IP)/pump-probe measurement of a time resolution of 40 ps for (ii) as schematically shown in Fig. 9(a). Figure 9(b) shows the experimental method and the time chart for pump-probe measurement using 40 ps X-ray pulses and synchronous femtosecond laser pulses. The DVD samples were prepared by depositing GST and AIST on a SiO2 glass substrate (12 cm diameter and 0.6 mm thick) with a 2-nm-thick 80 mol% ZnS-20 mol% SiO2 cap layer. The thickness of the sample was 300 nm, which gave sufficient X-ray diffraction intensities for structure characterization and sufficient stability of

Fig. 9. (a) Schematic diagram of APD/MCS and IP/pump-probe measurement and (b) scheme of DVD rotating system and the time chart of pump-probe measurement.

Figure 10 shows the time-profiles of the X-ray diffraction intensity of Bragg peaks (black and blue lines) together with those of the optical reflectivity profiles (red lines) of (a) GST and (b) AIST, respectively, measured by APD/MCS measurement. These show good

**2.5 Performance of the X-ray pinpoint structural measurement system for** 

**investigation of optical recording process** 

the amorphous phase to avoid spontaneous crystallization.

width at half maximum.

**change materials** 

accordance, which prove the close relation between the X-ray diffraction intensity and the optical reflectivity of the phase-change materials, *i.e.,* the structure and the electronic properties.

Fig. 10. Time-evolution of the optical reflectivity and X-ray diffraction intensity of (a) GST and (b) AIST.

From the intensity profile measured by the APD/MCS method, the delay times, *τ*, were determined for the IP/pump-probe method. Figure 11(a) shows the diffraction patterns obtained from the IP/pump-probe method for a 40 ps snapshot. Since the intensity of each diffraction peak has a uniform time-dependent increase, there is no crystal-crystal phase transition in GST and AIST during the crystal growth. We estimated the grain sizes from the line width of Bragg reflection, as shown in Fig. 11(b). In GST, the grain size is nearly constant (=70 nm), while the grain size significantly increases up to 1 μs in AIST. It is remarkable that the volume fraction of the crystal phase is almost saturated at 300 ns (see Fig. 10). Thus these observations suggest the coalescence of the crystal domains after 300 ns.

Fig. 11. (a) Snapshots of X-ray diffraction patterns and (b) the changes in peak width.

From the above experimental findings, we propose schematic models for the crystallization processes of GST and AIST as shown in Fig. 12. In the case of GST, nucleation takes place in the whole area in the amorphous phase after laser irradiation (A), and the number of newly formed crystallites of 70 nm diameter increases during the cooling process until 300 ns (B-

Time Resolved Investigation

crystalline form.

line, crystalline form.

accompanied by small changes in bond lengths.

of Fast Phase-Change Phenomena in Rewritable Optical Recording Media 269

AIST (2.86 Å) is only slightly shorter than that found in c-AIST (2.93 Å), whereas the first peak in a-GST (2.79 Å) is shorter than that in c-GST (2.97 Å). We note also that the shoulder on the second peak in a-AIST (3.50 Å) is near that observed in the crystalline form (3.30 Å). The pronounced difference between the diffraction patterns of the two materials is strong evidence that they crystallize differently. The atomic motion and/or diffusion accompanying the phase change are larger in GST than in AIST, where the phase change is

Fig. 13. Total structure factors, *S*(*Q*), for AIST and GST. Blue line, amorphous form; red line,

Fig. 14. Total correlation functions, *T*(*r*), for AIST and GST. Blue line, amorphous form; red

C). The crystal growth is then disturbed by the impingement of crystallites with each other (D). Our schematic model is consistent with the nucleation-driven crystallization process (Zaou et al., 2000).

Fig. 12. (a) Schematic models with TEM image for the crystallization processes in GST and AIST.

On the other hand, the nuclei of AIST can immediately transform to smaller crystallites (<< 60 nm), which form domains (F). These domains are enlarged by edge-growth crystallization (F-G) and the crystallites coalesce at the final stage (300 ns~) of crystallization in AIST (G-H). These proposed schematic models are consistent with the TEM pictures, in which GST has a grainy texture filled with 100-nm-size grains, whereas AIST has a fine texture.
