Electronically Driven Amorphization in Phase-Change In2Se3

Oct 2, 2012 - We show that the amorphization process in phase-change In2Se3 nanowires grown by chemical vapor deposition can be driven by electronic ...
1 downloads 0 Views 2MB Size
Article pubs.acs.org/JPCC

Electronically Driven Amorphization in Phase-Change In2Se3 Nanowires Elham Mafi, Afsoon Soudi, and Yi Gu* Department of Physics and Astronomy, Washington State University, Pullman, Washington 99164, United States S Supporting Information *

ABSTRACT: We show that the amorphization process in phase-change In2Se3 nanowires grown by chemical vapor deposition can be driven by electronic effects and does not require the conventional thermal melt-quench process. In particular, using transmission electron microscopy, in situ single-nanowire Raman spectroscopy, scanning Kelvin probe microscopy, and finite-element simulations, we demonstrate that the electronic amorphization can be achieved under optical excitations at temperatures far below the thermal melting point. The mechanism of this electronic amorphization is likely related to the presence of atomic bonds with different strengths in the crystalline phase In2Se3 and the weakening of the weaker bonds by nonequilibrium electrons. Our findings suggest that In2Se3 is a promising candidate for phase-change memory applications, with potential advantages including energy-efficient memory switching due to the electronic amorphization process and highly stable data storage as a result of a high melting point compared to Ge/Sb−Te alloys. On a more general level, these results indicate the need to take into account the electronic effects in modeling and understanding the phase transition processes in phase-change memories.



INTRODUCTION Structural phase transition in solids is of fundamental interest, as it reflects lattice thermodynamics and kinetics of atomic motion and provides insight into the structure−property relation. From a practical viewpoint, phase-specific material properties can enable phase-change memory applications. In particular, the crystalline−amorphous phase transformation in Ge/Sb−Te alloys and related nanostructures, accompanied by large changes in the optical reflectivity and electrical conductivity, has been extensively studied1−10 for optical and electronic data storage. It has been the common understanding that amorphization is achieved via thermal melting followed by fast quenching, with the heating provided by intense optical or electric current pulses. However, the amorphization process can also take place in the absence of melting. A recent experimental study11 based on time-resolved X-ray absorption has shown that nonequilibrium electrons generated by optical excitations can weaken certain types of atomic bonds in Ge/Sb−Te alloys. This electronic effect can lead to lattice instability, which ultimately results in the collapse of the crystalline phase and leads to amorphization at temperatures below the melting point. This is also supported by theoretical studies.12,13 As the thermal melting is not required, this amorphization process can enable lower energy consumption and faster transformations.12 Furthermore, the discovery of the electronic amorphization process provides a new perspective in searching for suitable phase-change materials.13 Particularly, a high melting temperature, which indicates a relatively high crystallization temperature14 and is thus beneficial for higher amorphous phase stability against self-crystallization, has been so far considered © 2012 American Chemical Society

disadvantageous, because a high optical or electrical power is required for the amorphization process. However, for materials in which the amorphization process can be driven by the electronic effect, a high melting point would be clearly an advantage. While this electronic effect has been demonstrated using optical means, an electrical realization of this electronic amorphization, e.g. through electron injection from electrodes, is highly desirable and more relevant to phase-change randomaccess (PC-RAM) memory applications. While most studies have focused on Ge/Sb−Te alloys, In2Se3 recently has attracted interest15,16 as a potential candidate in phase-change memory applications. In2Se3 is a polymorphic material that can take the form of several crystalline phases (α, γ, δ, and κ) at various temperatures.17−25 This polymorphism can enable multilevel memories, which can achieve higher storage density and perform higher-order logics, via phase transitions among multiple crystalline and amorphous phases within a single material system. A recent study26 has demonstrated this multilevel data storage, but the underlying structural phase transitions remain unclear. As In2Se3 thin films are often polycrystalline and have mixed phases, studies of phase transitions remain a significant challenge. Recently, synthesis of single-crystalline In2Se3 nanowires has been demonstrated.27−29 These materials with well-defined crystallinity are ideal for studying phase transitions and also hold great promise for nanowire-based memory technologies. Received: June 11, 2012 Revised: September 27, 2012 Published: October 2, 2012 22539

dx.doi.org/10.1021/jp305696w | J. Phys. Chem. C 2012, 116, 22539−22544

The Journal of Physical Chemistry C

Article

Figure 1. (a) SEM image (scale bar: 2 μm) of an In2Se3 nanowire with (b) the corresponding EDX spectra. (c) TEM image and (d) the electron diffraction of an α-phase nanowire. (e) Schematic crystal structure of the α-phase In2Se3. (f) The single-nanowire Raman spectrum of an as-grown αphase nanowire.

single crystals. The diffraction pattern can be indexed by the lattice constants of the hexagonal α-phase layered structure (schematic structure shown in Figure 1e). The growth direction of these nanowires is along [112̅0], consistent with previous reports.27 The α-phase structure is further confirmed by Raman studies. Particularly, a continuous-wave (cw) He−Ne laser emission (633 nm, corresponding to ∼1.96 eV) was used as the excitation source; the laser beam was focused to a diffractionlimited spot through a confocal optical microscope, with the power of ∼50 μW. The use of this low power level, referred to as the weak optical excitation condition in this work, minimizes the optical heating and prevents phase transitions (see also below). A typical Raman spectrum of a single as-grown nanowire (Figure 1f), obtained in the backscattering geometry, shows a dominant peak at ∼111 cm−1 and a weaker peak at ∼206 cm−1. These peaks were attributed to the A1(LO+TO) and A1(LO) phonon modes of the α-phase In2Se3 lattice,32,33 respectively, where LO (TO) represents the longitudinal (transverse) optical phonon modes. The electronic amorphization process requires the excitation of nonequilibrium electrons, which can be achieved by optical means. Specifically, the cw He−Ne laser emission was focused to a diffraction-limited spot at the power of ∼1 mW, with an exposure time of less than 1 s. This high level of power, referred to as the strong optical excitation condition in this work, was found to be necessary to induce the local electronic amorphization. By considering the nanowire diameter and the optical spot size, a single In2Se3 nanowire was locally illuminated by an optical power of ∼0.2 mW. As the α-phase In2Se3 has a band gap of 1.4 eV,34 this optical excitation generates nonequilibrium electrons. Nanowires were dispersed on thin Si3N4 windows (50 nm thick) to allow for TEM studies on locally illuminated nanowires. The optical excitation procedure was carried out in ambient air and also in an Ar environment, and no difference in the amorphization was observed in these two cases. All experiments were conducted at room temperature, unless noted otherwise. Figure 2 shows the TEM images and the corresponding electron diffractions of a nanowire that was subjected to the strong local optical excitation. The electron diffraction of the as-grown regions (Figure 2d) shows clear single-crystal patterns. In contrast, the

The phase inhomogeneity in single In2Se3 nanowires subjected to a pulsed Joule heating was investigated by microwave impedance microscopy,30 and a coexistence of crystalline and amorphous phases was observed. Another recent study,31 based on in situ synchrotron X-ray diffraction measurements of an ensemble of In2Se3 nanowires, reported a transition of the κ phase into the α phase as the temperature increases. In these cases, the phase transitions were all considered to be thermally driven, and the role of the electronic effect has not been explored. Given the high melting temperature of In2Se3 (∼1165 K vs 890 K of Ge2Sb2Te5), the presence of the electronic effect would make this material a desirable candidate for highly stable PC-RAM applications. Furthermore, if multiple phases can be stabilized at room temperature and the corresponding phase transitions can be controlled, this material would be very promising for multilevel memory applications. Here we present the evidence of an electronically driven amorphization process in In2Se3 nanowires grown by chemical vapor deposition. Specifically, using transmission electron microscopy, in situ single-nanowire Raman spectroscopy, scanning Kelvin probe microscopy, and finite-element simulations, we show that the amorphization process can occur much below the melting point, in the presence of nonequilibrium electrons that are excited by optical illuminations. These findings demonstrate the potential of In2Se3 nanowires as a promising phase-change material.



RESULTS AND DISCUSSION Figure 1a shows a scanning electron microscope (SEM) image of an In2Se3 nanowire grown by chemical vapor deposition on a Si substrate. The details of growth conditions and characterizations can be found in the Experimental Methods. The typical nanowire diameter ranges from ∼80 to 150 nm. The selectedarea energy-dispersive X-ray (EDX) spectra (Figure 1b) show that the bulk of the nanowire is stoichiometric In2Se3 while the tip of the nanowire is a Au−In−Se alloy, which is consistent with the Au-catalyzed growth mode. A typical transmission electron microscopy (TEM) image and the corresponding electron diffraction pattern, shown in Figure 1, parts c and d, respectively, demonstrate that the as-grown nanowires are 22540

dx.doi.org/10.1021/jp305696w | J. Phys. Chem. C 2012, 116, 22539−22544

The Journal of Physical Chemistry C

Article

allows us to estimate the temperature rise due to the optical heating (see also below). In any case, results shown in Figure 2 unambiguously point to the formation of the amorphous phase as a result of the local optical excitation, which is also verified by Raman studies presented below. The strong optical excitation can lead to a significantly elevated temperature, which might rise above the melting point and drive the amorphization through a conventional melt− quench process. Therefore, to exclude the possibility of a thermal amorphization process, it is necessary to evaluate the temperature rise due to the optical heating. In the following, we present simulation results and in situ Raman studies to establish that the temperature of the nanowire region under the strong optical excitation is below the melting point of In2Se3. First, a three-dimensional (3D) finite-element simulation was used to calculate the temperature of the nanowire under the local optical excitation. Details of this simulation can be found in the Supporting Information. The simulation, shown in Figure 3a, points to a maximum temperature of ∼550 K in the region under strong optical excitation. This result can be validated by considering the in situ Raman spectra of the region under strong optical excitation. It is well-known that the lattice phonon frequencies become lower as the temperature increases. Therefore, the shift of the phonon frequency can be used to obtain the temperature. Figure 3b shows an in situ Raman spectrum that was obtained from a single nanowire while it was under strong optical excitation. Interestingly, the Raman spectrum still shows a rather strong A1(LO+TO) phonon mode. This is most likely due to the presence of α-phase crystallites, as discussed above and shown in Figure 2. In addition, a rather broad peak around 240−250 cm−1 was also observed (inset in Figure 3b). This peak originates from the vibrational modes of Se chains and molecules in amorphousphase In2Se3;37,38 this further confirms the presence of the amorphous phase. The A1(LO+TO) phonon mode, under the strong optical excitation condition (∼1 mW) used to induce the amorphization, is at a frequency of ∼107.5 cm−1, as shown in Figure 3b. This frequency is lower than the ∼111 cm−1 peak obtained under the weak excitation condition (∼50 μW) at room temperature (the top curve in the inset in Figure 3c). We note that the temperature rise under the weak excitation condition was calculated by the finite-element simulation to be very small (∼9 K). To relate this shift in the phonon frequency to the temperature, we measured the A1(LO+TO) phonon frequency under the weak optical excitation at various ambient

Figure 2. (a) TEM image and the electron diffraction pattern (inset, left corner) of the nanowire area subjected to the strong local optical excitation, with an FFT image of the amorphous region (inset, right corner). (b) High-resolution TEM image showing an In2Se3 crystallite (indicated by dashed lines). (c) TEM image and (d) the electron diffraction pattern of the unilluminated nanowire area.

electron diffraction of the optically illuminated region (inset in Figure 2a, left corner) is characterized by a rather diffuse scattering with spots, indicating the presence of the amorphous phase as well as crystallites. The presence of the amorphous phase is further indicated by fast Fourier transformation (FFT) images of selected areas in the illuminated region. One example is shown as an inset (right corner) in Figure 2a, which exhibits very diffusive patterns, characteristic of the amorphous phase. The presence of the crystalline phase is confirmed by a highresolution TEM image of the optically excited area, shown in Figure 2b. The changes in the nanowire morphology, particularly the increase in the nanowire diameter (Figure 2a,c), are consistent with the formation of the amorphous phase, as the density of the amorphous In2Se3 is lower than that of the crystalline phase.35 We note that, although the presence of crystallites indicates incomplete amorphization, using a higher optical power (and thus a stronger electronic effect) can lead to a higher degree of amorphization (not shown here).36 On the other hand, the presence of crystallites is beneficial, as it

Figure 3. (a) 3D simulated temperature of a nanowire under local optical excitation on a 50 × 50 μm Si3N4 window, with the inset showing a zoomed-out and top view of the temperature distribution of the entire Si3N4 window. (b) In situ Raman spectrum of a single nanowire while it was under strong local optical excitation. (c) Frequency of the A1(LO+TO) phonon mode (solid circles) as a function of temperature, and the inset shows the in situ Raman spectra (vertically shifted for clarity) of the single nanowire at various ambient temperatures, with the dashed line representing a spline interpolation. 22541

dx.doi.org/10.1021/jp305696w | J. Phys. Chem. C 2012, 116, 22539−22544

The Journal of Physical Chemistry C

Article

Figure 4. (a) Current−voltage characteristics of the nanowire before and after the optical excitation. (b) Surface electric potential map and (c) the corresponding line profile of a nanowire subjected to local optical excitation.

possibility of the electronic amorphization based on strong and ultrafast electronic excitations. A recent experimental study has shown that the amorphization can happen below the melting point in Ge/Sb−Te alloys.11 This was suggested by a theoretical report13 to be due to a coexistence of weak and strong atomic bonds in these materials, which originates from a lattice distortion that breaks the six identical resonance Ge−Te bonds into three weak and three strong bonds.40 Upon optical excitations, the generated nonequilibrium electrons thermalize with the lattice and occupy the low-energy states corresponding to the weak atomic bonds. This leads to a further weakening and ultimately a breakdown of these bonds and, in turn, a collapse of the entire crystal structure and the loss of the longrange order. This mechanism requires a relatively small population of generated electrons,11 due to the presence of weak atomic bonds. A different perspective in ref 12 based on first-principles calculations suggests that, upon the optical excitation, the removal of the antibonding s electrons around Ge atoms leads to the crystal lattice instability and emergence of the amorphous phase. However, a large number (9−15% of the total valence electrons) of the antibonding s electrons need to be removed for the amorphization to occur.12 We note that, in our case, the optical excitation used to induce the amorphization corresponds to the removal of ∼0.01% valence electrons in In2Se3. Therefore, we propose that, for the observed amorphization in In2Se3 nanowires, the weakening and the breakdown of the weak atomic bonds by generated nonequilibrium electrons is the underlying mechanism. Indeed, the α-phase layer-structured In2Se3, as shown schematically in Figure 1e, has both strong atomic bonds, namely the covalent In−Se bonds,41 as well as weak atomic bonds, particularly the van der Waals bonds between layers. As pointed out in ref 11, there is a similarity between Ge/Sb−Te and Se crystals with regard to the electronic amorphization process; we note that our case is also very similar to the reported amorphization of Se crystals under an optical excitation below the melting point,38 in that the crystalline Se has both covalent and van der Waals bonds for the intrachain and interchain bonding, respectively.42 In addition, the surfaces might also play an important role. Particularly, surface atoms are expected to have an enhanced mobility due to broken bonds, which can facilitate the electronic amorphization process. This effect becomes more significant in nanowires, where the surface-to-volume ratio is large. Further inves-

temperatures; the results are plotted in Figure 3c. From this, the temperature of the nanowire region under the strong optical excitation was estimated to be ∼520 K. Measurements on several nanowires yield the temperature range of 480−520 K. This agrees reasonably well with the simulated results (∼550 K), as the simulation is expected to provide the upper limit of the temperature (see Supporting Information). These Raman results also verify the validity of our simulation model. An important issue that needs to be addressed is the size effect on the melting point: in general, the melting point becomes lower as the material dimension decreases. In particular, a decreased melting point of ∼950 K, compared to 1165 K in the bulk, has been reported for a 60 nm diameter In2Se3 nanowire in ultrahigh vacuum.28 This melting temperature is still significantly higher than the estimated temperature in our case. Furthermore, as the dimensions of our nanowires are generally larger (diameter ∼80−150 nm), this size effect is expected to be weaker. To verify that the melting point of our In2Se3 nanowires is at least higher than 520 K, these nanowires were heated to an ambient temperature of ∼680 K (the highest achievable temperature with our microscope heating stage) in an Ar environment, and no melting was observed from the in situ optical microscopy (not shown here). Having established that the observed amorphization does not involve conventional thermal melting, we next discuss the possible origins of this phenomenon. We note that an electronic amorphization of semiconductors under intense femtosecond laser pulses has been well-studied;39 in particular, a dense photogenerated plasma and the removal of a large quantity of valence electrons (∼ 10% of the total valence electrons) from atomic bonds weaken the crystal lattice, and this can lead to the loss of the long-range order, i.e. a formation of the amorphous phase. This type of the electronic amorphization requires that the generated electrons not be in thermal equilibrium with the lattice during the laser pulse excitation,39 and this is why femtosecond laser pulses are used in these cases. On the other hand, if the laser pulse duration is long enough (about a few picoseconds) so that a thermal equilibrium is established between the excited electrons and the lattice during the laser pulse excitation, the lattice heats up and the amorphization occurs through the conventional thermal melt−quench process. In our case, the relative weak cw optical excitation ensures a thermal equilibrium between generated electrons and the lattice, which therefore excludes the 22542

dx.doi.org/10.1021/jp305696w | J. Phys. Chem. C 2012, 116, 22539−22544

The Journal of Physical Chemistry C

Article

tigations, including first-principle studies, will provide more insight into the atomic origin of the electronic amorphization process in In2Se3 nanowires. We next discuss the effect of this electronic amorphization on the electrical properties of In2Se3 nanowires. For electrical characterizations, In2Se3 nanowires were dispersed on heavily doped Si substrates (380 μm thick) covered with a 200 nm thick Si3N4 layer.43 Upon the electronic amorphization induced by the local optical excitation, the electrical resistance of single In2Se3 nanowires increases by ∼3 orders of magnitude, as shown in Figure 4a, consistent with the formation of the amorphous phase. Using scanning Kelvin probe microscopy, we have obtained the electric potential map of single nanowires with the amorphous mark defined by the local optical excitation, as shown in Figure 4b. Interestingly, the corresponding potential profile across this amorphous mark, plotted in Figure 4c, shows that most of the applied external bias drops across the crystalline/amorphous phase boundaries, indicating that these phase boundaries, instead of the amorphous phase, have the highest electrical resistance. The potential barriers at the phase boundaries, possibly due to band structure discontinuities and/or a presence of defects at these boundaries, might contribute to this large boundary resistance. On the other hand, the presence of crystallites in the amorphous regions might provide an electrical conduction path, e.g., via carrier hopping/tunneling, which might lead to a relatively low electrical resistance in the amorphous region. For the phase-change memory applications, the capability to reversibly switch between low- and high-resistance states is required. We have attempted to “set” the devices (i.e., recrystallization) using voltage pulses and small dc voltages. However, we did not observe any significant increase in the electrical conductivity; i.e., the amorphization induced by the optical illumination appears to be irreversible. One possible explanation is that, as the recrystallization process is a thermal process driven by Joule heating, the dominant electrical resistance of the crystalline−amorphous boundary leads to a localized heating at the boundary but not inside the amorphous region. On the other hand, our preliminary results (not shown here) suggest that, for In2Se3 nanowires amorphized by voltage pulses, where the electronic excitation can be provided by injected electrons, the recrystallization can be induced by applying a low-magnitude voltage pulse. More detailed electrical transport and TEM studies will further elucidate the origin of this difference, but they are beyond the scope of this work. In summary, through TEM, Raman, and scanning probe microscopy studies as well as finite-element simulations, we have shown that the amorphization process in In2Se3 nanowires can be driven by the electronic effect and does not require the conventional thermal melt−quench process. Particularly, the electronic amorphization can be achieved through optical excitations at temperatures far below the thermal melting point. The mechanism of the electronic amorphization in In2Se3 is likely to be similar to that proposed for Ge/Te−Sb alloys, particularly the presence of atomic bonds with different strengths in the crystalline phase and the weakening of the weaker bonds by nonequilibrium electrons. In addition, the large surface-to-volume ratio in nanowires might also play an important role. Our findings suggest that In2Se3 is a promising candidate for phase-change memory applications, with the potential advantages including energy-efficient memory switching due to the electronic amorphization process and highly

stable data storage as a result of a high melting point compared to Ge/Sb−Te alloys. More generally, these results indicate the need to take into account the electronic effects in modeling and understanding the phase transition processes in phase-change memories.



EXPERIMENTAL METHODS Synthesis. In2Se3 nanowires were synthesized using a chemical vapor transport method: In2Se3 powders (Alfa Aesar, 99.99%) were heated to 820 °C and the vapor generated from the congruent sublimation was transported by a flow of H2 gas; silicon substrates coated by Au colloids were placed downstream, where the temperature was kept at 620−640 °C and the pressure was 200 Torr. The typical growth time was 5 h. Characterizations. A Leica optical microscope, coupled to a Renishaw InVia spectrometer and a 633 nm laser, was used for confocal Raman spectroscopy. For in situ high-temperature Raman experiments, an Instec microscope hot stage (HCS 302) was used. Transmission electron microscopy was performed on a FEI Tecnai transmission electron microscope. Scanning Kelvin probe microscopy (SKPM) was used to map the surface electric potential on electrically biased nanowires. The details of the SKPM can be found in ref 44.



ASSOCIATED CONTENT

S Supporting Information *

Details of the finite-element simulations. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This work was supported by the National Science Foundation through grant DMR-1206960. REFERENCES

(1) Hamann, H. F.; O’Boyle, M.; Martin, Y. C.; Rooks, M.; Wickramasinghe, K. Nat. Mater. 2006, 5, 383−387. (2) Raoux, S.; Shelby, R. M.; Jordan-Sweet, J.; Munoz, B.; Salinga, M.; Chen, Y. C.; Shih, Y. H.; Lai, E. K.; Lee, M. H. Microelectron. Eng. 2008, 85, 2330−2333. (3) Lankhorst, M. H. R.; Ketelaars, B. W. S. M. M.; Wolters, R. A. M. Nat. Mater. 2005, 4, 347−352. (4) Kang, S.; Cho, W. Y.; Cho, B. H.; Lee, K. J.; Lee, C. S.; Oh, H. R.; Choi, B. G.; Wang, Q.; Kim, H. J.; Park, M. H.; Ro, Y. H.; Kim, S.; Ha, C. D.; Kim, K. S.; Kim, Y. R.; Kim, D. E.; Kwak, C. K.; Byun, H. G.; Jeong, G.; Jeong, H.; Kim, K.; Shin, Y. IEEE J. Solid-State Circuits 2007, 42, 210−218. (5) Bedeschi, F.; Bez, R.; Boffino, C.; Bonizzoni, E.; Buda, E. C.; Casagrande, G.; Costa, L.; Ferraro, M.; Gastaldi, R. O.; Khouri, S.; Ottogalli, F.; Pellizzer, F.; Pirovano, A.; Resta, C.; Torelli, G.; Tosi, M. IEEE J. Solid-State Circuits 2005, 40, 1557−1565. (6) Rodgers, J.; Maimon, J.; Storey, T.; Lee, D.; Graziano, M.; Rockett, L.; Hunt, K. A 4-Mb non-volatile chalcogenide random access memory designed for space applications: Project status update. In Non-Volatile Memory Technology Symposium; Piscataway, NJ, 2008; p 64. (7) Jeong, C. W.; Ahn, S. J.; Hwang, Y. N.; Song, Y. J.; Oh, J. H.; Lee, S. Y.; Lee, S. H.; Ryoo, K. C.; Park, J. H.; Park, J. H.; Shin, J. M.;

22543

dx.doi.org/10.1021/jp305696w | J. Phys. Chem. C 2012, 116, 22539−22544

The Journal of Physical Chemistry C

Article

(38) Poborchii, V. V.; Kolobov, A. V.; Tanaka, K. Appl. Phys. Lett. 1999, 74, 215−217. (39) Sundaram, S. K.; Mazur, E. Nat. Mater. 2002, 1, 217−224. (40) Shportko, K.; Kremers, S.; Woda, M.; Lencer, D.; Robertson, J.; Wuttig, M. Nat. Mater. 2008, 7, 653−658. (41) Semiletov, S. A. Sov. Phys. Crystallogr. 1961, 5, 673. (42) Joannopoulos, J. D.; Schlüter, M.; Cohen, M. L. Phys. Rev. B 1975, 11, 2186−2199. (43) Compared to the thin Si3N4 windows used for studies presented above, this type of substrate is more efficient in dissipating heat. In fact, under the same strong optical excitation condition, simulations show that the temperature of the illuminated area is ∼430 K. (44) Soudi, A.; Aivazian, G.; Shi, S. F.; Xu, X. D.; Gu, Y. Appl. Phys. Lett. 2012, 100, 033115.

Yeung, F.; Jeong, W. C.; Kim, J. I.; Koh, G. H.; Jeong, G. T.; Jeong, H. S.; Kim, K. Jpn. J. Appl. Phys. 2006, 45, 3233−3237. (8) Lee, S. H.; Ko, D. K.; Jung, Y.; Agarwal, R. Appl. Phys. Lett. 2006, 89, 223116. (9) Park, H.; Yu, D.; Brittman, S.; Lee, J. S.; Falk, A. L. Nano Lett. 2008, 8, 3429−3433. (10) Cui, Y.; Meister, S.; Peng, H. L.; McIlwrath, K.; Jarausch, K.; Zhang, X. F. Nano Lett. 2006, 6, 1514−1517. (11) Fons, P.; Osawa, H.; Kolobov, A. V.; Fukaya, T.; Suzuki, M.; Uruga, T.; Kawamura, N.; Tanida, H.; Tominaga, J. Phys. Rev. B 2010, 82, 041203. (12) Li, X. B.; Liu, X. Q.; Liu, X.; Han, D.; Zhang, Z.; Han, X. D.; Sun, H. B.; Zhang, S. B. Phys. Rev. Lett. 2011, 107, 015501. (13) K., A.; Kolobov, A. V.; Krbal, M.; Fons, P.; Tominaga, J.; Uruga, T. Nature Chem. 2011, 3, 311−316. (14) Raoux, S.; Wuttig, M. Phase Change Materials: Science and Applications; Springer: NewYork, 2008. (15) Lee, H.; Kim, Y. K.; Kim, D.; Kang, D. H. IEEE Trans. Magn. 2005, 41, 1034−1036. (16) Lee, H.; Kang, D. H.; Tran, L. Mater. Sci. Eng. B-Solid 2005, 119, 196−201. (17) Julien, C.; Eddrief, M.; Balkanski, M.; Hatzikraniotis, E.; Kambas, K. Phys. Status Solidi A 1985, 88, 687−695. (18) Julien, C.; Eddrief, M.; Kambas, K.; Balkanski, M. Thin Solid Films 1986, 137, 27−37. (19) Deblasi, C.; Drigo, A. V.; Micocci, G.; Tepore, A.; Mancini, A. M. J. Cryst. Growth 1989, 94, 455−458. (20) Pfitzner, A.; Lutz, H. D. J. Solid State Chem. 1996, 124, 305− 308. (21) Marsillac, S.; CombotMarie, A. M.; Bernede, J. C.; Conan, A. Thin Solid Films 1996, 288, 14−20. (22) Lu, C. Y.; Shamberger, P. J.; Yitamben, E. N.; Beck, K. M.; Joly, A. G.; Olmstead, M. A.; Ohuchi, F. S. Appl. Phys. AMater. 2008, 93, 93−98. (23) Jasinski, J.; Swider, W.; Washburn, J.; Liliental-Weber, Z.; Chaiken, A.; Nauka, K.; Gibson, G. A.; Yang, C. C. Appl. Phys. Lett. 2002, 81, 4356−4358. (24) de Groot, C. H.; Moodera, J. S. J. Appl. Phys. 2001, 89, 4336− 4340. (25) Kambas, K.; Julien, C. Mater. Res. Bull. 1982, 17, 1573. (26) Lee, J. Y.; Sun, K.; Li, B. Y.; Xie, Y. H.; Wei, X. Y.; Russell, T. P. Appl. Phys. Lett. 2010, 97, 092114. (27) Peng, H. L.; Schoen, D. T.; Meister, S.; Zhang, X. F.; Cui, Y. J. Am. Chem. Soc. 2007, 129, 34−35. (28) Sun, X. H.; Yu, B.; Ng, G.; Nguyen, T. D.; Meyyappan, M. Appl. Phys. Lett. 2006, 89, 233121. (29) Zhai, T. Y.; Fang, X. S.; Liao, M. Y.; Xu, X. J.; Li, L.; Liu, B. D.; Koide, Y.; Ma, Y.; Yao, J. N. A.; Bando, Y.; Golberg, D. ACS Nano 2010, 4, 1596−1602. (30) Lai, K. J.; Peng, H. L.; Kundhikanjana, W.; Schoen, D. T.; Xie, C.; Meister, S.; Cui, Y.; Kelly, M. A.; Shen, Z. X. Nano Lett. 2009, 9, 1265−1269. (31) Sun, X. H.; Li, Y.; Gao, J.; Li, Q. L.; Peng, M. F.; Li, Y. Y.; Yuan, G.; Wen, W.; Meyyappan, M. J. Mater. Chem. 2011, 21, 6944−6947. (32) Kambas, K.; Julien, C.; Jouanne, M.; Likforman, A.; Guittard, M. Phys. Status Solidi B 1984, 124, K105−K108. (33) Lewandowska, R.; Bacewicz, R.; Filipowicz, J.; Paszkowicz, W. Mater. Res. Bull. 2001, 36, 2577−2583. (34) Becla, P.; Gumienny, Z.; Misiewicz, J.; Pawlikowski, J. M. Opt. Appl. 1982, 12, 143−149. (35) Kohary, K.; Burlakov, V.; D. Nguyen-Manh, Pettifor, D. Modeling In−Se amorphous alloys. European Phase Change and Ovonics Symposium; England, 2005. (36) The incomplete amorphization might be due to the nonuniform optical power distribution along the radial as well as the axial directions. (37) Weszka, J.; Daniel, P.; Burian, A.; Burian, A. M.; Nguyen, A. T. J. Non-Cryst. Solids 2000, 265, 98−104. 22544

dx.doi.org/10.1021/jp305696w | J. Phys. Chem. C 2012, 116, 22539−22544