We report a method for controlling the size of a Ge (germanium) nanostructure by changing the angle between the ultrafast laser polarization and the crystal axis of Ge. The nanostructure size dependence on the laser polarization with respect to the Ge crystal axis exhibits a sinusoidal function with a minimum size at (100) axis. Moreover, the measurement of transient reflection reveals the presence of large anisotropies in both its amplitude and its relaxation dynamics with a minimum at (100) crystal axis. This implies that the observed anisotropic dependence of nanostructure size of Ge is followed by a different carrier density as well as its relaxation process, depending on the orientation of the Ge crystal axis only at near and above threshold fluence.
© 2006 Optical Society of America
Much attention has been given to the preparation of luminescent nano-sized semiconductor materials [1, 2] in the interest of fundamental research and application. In some applications, including the optoelectronic, microelectronic, and micro- and nano-biosensors, it is critical to create local nanostructures directly on the devices or on an assembled integrated chip . The advent of reliable generation and amplification of the ultrafast laser enables it to be used for laser-induced microfabrication with high quality [4–7]. A significant improvement of a size and its dispersion reduction of a colloidal gold nanoparticle was reported in a case of ultrafast laser ablation over a nanosecond laser process. Recently, T-H Her et al.  reported that the irradiation of 500 laser pulses with 100 fs pulse width on silicon surfaces under 500 torr SF6 or Cl2 creates conical spikes capped by a 1.5 mm ball at the tops. We have also recently reported that the irradiation of an ultrafast laser on a Ge single crystal with a laser fluence far above the ablation threshold results in the formation of a room-temperature photoluminescent Ge nanostructure dangled on the microstructures .
In this report, we demonstrate that relative angle change between laser polarization and the Ge crystal axis can be applied to control the size of Ge nanostructures accompanied with a strongly polarized amorphization near and above the ablation threshold. Measured changes in anisotropic reflectivity during ablation of Ge prove that the relative angle between laser polarization and the crystal axis governs the ablation process with encapsulated oxidation and particle formation with critical size.
Samples of the Ge nanostructure were prepared with femtosecond laser ablation of an undoped (001) Ge single crystal with diamond cubic structure. A linearly polarized femtosecond pulse was used to ablate the Ge wafer, which was cut from Ge (001) single crystal wafers (EaglePicher Technologies, USA) with a resistivity higher than 30 Ωcm. To prevent the effect of successive subpulses, single-shot configuration was adapted in the current work with the use of a fast mechanical shutter. The ablation of the Ge wafer was performed by a Ti:sapphire laser (Quantronics, USA) in air. This laser delivers pulses with energy of up to 1 mJ at 800 nm and a repetition rate of 1 kHz. The fundamental output of the laser was delivered to the galvanometer scanner (Scanlab AG, Germany). The topological changes of the Ge surface after laser ablation was measured with an atomic force microscope (AFM) (PSIA, Korea), and then the images were put into a personal computer for further analysis of particle size distribution by using a commercial image processor (Wright, Canada). Because the laser output profile is not flat, the size distribution was spatially inhomogeneous within the single-shot processed spot. Therefore we took special care not to limit our measurements to similar areas of topography. Instead we measured more than nine spots within every configuration of experiments, thus minimizing the variation of particle size distribution in spot by spot. It should be noted that the variation of the analysis is less than 20% between each shot. The crystal axis was determined by measuring the XRD pattern (X’pert MRD, Philips) for Ge wafers used in the current study.
In the reflectivity measurement, while the polarization of the pump (800 nm) and probe (super continuum) beam were set to be orthogonal, the intensity of the probe beam is kept at less than 0.1% of the photo excitation of the pump. The pump pulse was controlled by a triggered mechanical shutter during the probe beam irradiation with normal incidence.
3. Results and discussion
The whole-area image and near-the-focus image of the laser-processed sample surface are shown in Figs. 1(a) and 1(b), respectively. The volcano-like surface formed by single pulse consists of numerous nanostructures on the center of the ablated region, and the nanostructure shape closely resembles a hemisphere. In addition, there is no doubt of the presence of GeOx (2-3 nm thick) from the characterization of the nanostructures observed with an energy dispersive spectrometry (EDS) in Fig. 1(c) and high-resolution transmission electron microscopy (HRTEM) in Fig. 1(d). In Fig. 1(d), the gray parts that encapsulate the structure represent GeOx amorphous layers. The nanostructure is crystallized and covered with amorphous layers the thickness of a few nanometers. It should be noted that abundant oxygen encapsulates the Ge crystal, and the particle cannot grow anymore after the capsulation, which determines the particle size quantitatively.
Typical AFM images and the normalized size distributions of Ge nanostructures formed by femtosecond laser irradiation are shown in Fig. 2(a). In Fig. 2(a), the numbers denote relative angles between laser polarization and the Ge (100) axis. The average particle size shows a sinusoidal function with a minimum at the (100) axis. The particle size distribution was measured as a function of an angle between the crystal axis and laser polarization at the two different laser fluences of Fth  (hollow circle) and 2Fth (filled circle), and the results are plotted in polar coordinate (Fig. 2(b)).
The size distribution exhibits a sin(2φ) dependent behavior where φ denotes the angle between the laser polarization and the (100) crystal axis. The optical polarization parallel to the (100) axis of the Ge crystal results in the smallest particle size, while the polarization parallel to (110) results in the largest particle size. For the lower laser fluence, Fth, the mean particle size varies from 7 nm to 11 nm. Further increase in the laser fluence results in a slightly larger particle between 10 nm to 17 nm. With the increasing of the laser fluence, the nanoparticle size increases whereas its number density decreases, as shown in Fig. 2(c). The particle number density decreases due to the rapid expansion of the molten layer into air, void nucleation, or the removal of matter during amorphization as power increases .
When fluence becomes extremely high, we observe significant dispersion of size distribution. Therefore improvement in size dispersion and stability of nanostructure size are considered to occur only near and slightly higher than the ablation threshold. To help determine which mechanism governs this anisotropy in size distribution and what determines nanostructure size, we measured the reflectivity changes during the ablation process.
In Fig. 3 transient reflectivity changes are shown with different fluences near zero delay time. At the laser fluence less than the ablation threshold, F = 0.6 Fth, the reflectivity changes essentially exhibit polarization independence. As the fluence increases toward the ablation threshold and beyond, however, strong incident polarization dependence begins to develop with resulting anisotropies as large as 13%. In depth, the photo-induced reflectivity has a maximum when the optical polarization is parallel to the (100) crystal axis with sinusoidal dependence. While the reflectivity change is negligible at laser fluence less than ablation threshold, the changes are remarkably enhanced in dynamic ranges when the fluence exceeds the threshold.
For a Ge single crystal, each Ge atom has four immediate neighbors in a tetrahedral shape. While the conduction orbital is anti-bonding, the valence orbital between immediate neighbors is bonding . The inter-band excitation of a dense electron-hole plasma leads to disorder and a break in the bonding in a crystal during ablation. The excitation of electron-hole pairs weakens the bonding, and then nonthermal molten layers build up . These occurrences, the amorphization process and the amorphous-atomistic structures of the nanoparticle, have already been well-investigated elsewhere . This initial amorphization, which is induced by short pulse laser, essentially governs transient polarization-dependent reflectivity change, as shown in Fig. 3.
Temporal profiles of the reflectivity changes at different incident fluences and different polarization angles are shown in Fig. 4. By close analysis of the temporal profiles, it is noted that there are at least more than two decaying components: one of them has a time constant less than our time resolution of 150 fs; the other is longer than 1 ps. It is very interesting to note that the two components are evidently dependent on the azimuthal angle as well as the incident fluence. In this time-resolved reflectivity measurement, we observed a transition to high reflectivity change, which indicates that the liquid state has anisotropically occurred even though our time resolution is not sufficient. The amorphization process is known to be very fast (t≪ps) which is consistent with our measurement result  and directly related to the thickness of molten layer at the surface of the crystal during the ablation. This ultrafast initial decrease in reflectivity, due to photo-induced absorption by an optical phonon, provides clear evidence that a portion of the Ge crystal undergoes nonthermal melting within subpicoseconds [11, 14]. The difference in the amplitude of the ultrafast components, which is strongly dependent on the azimuthal angle even at the same incident fluence, suggests that the thickness of the molten layer is dependent on the polarization. With this assumption, it is estimated that the molten layer thickness for the polarization parallel to the (110) axis is larger than that of the polarization parallel to the (100) axis. Furthermore, with the assumption that larger absorption leads to a thicker molten layer , larger Ge nanoparticle size with increasing fluence, seen in Fig. 2(c), can be understood.
The following decay component represents the anisotropic melting state, which comes from thermalization of the optical energy, proving different oriented particle formation. Just after amorphization, complex physicochemical processes including thermal melting, recrystallization, encapsulation by an oxide layer , and nanoparticle formation  etc. will contribute the reflectivity changes within several picoseconds. Among them, the oxide layer formation might play an important role in preventing inter-particle coalescence or aggregation . These time constants, τ 1 and τ 2 , responsible for the above-mentioned complex physicochemical processes should increase as the thickness of the molten layer increases. The result indicates that the molten layer thickness, which is analogized from reflectivity changes, plays an important role in nucleation and particle growth and finally determines the size of structures regardless of the oxidation velocity. Moreover, the measured reflectivity after 1 ms represents the terminal state of a Ge nanoparticle for the irreversible process.
With these different initial, amorphous responses it should be discussed why under the same laser fluence, the changes in the azimuthal angle between the laser polarization and the crystal axis could give large variations in nanostructure size with polarization. The discussions given in the previous section led us to suppose that the exposure of laser fluence with laser polarization parallel to the (110) crystal axis results in a thicker molten layer, which follows the formation of larger nanostructures in comparison to the case parallel to the (100) crystal axis. If we noted that an additional absorption just after photoexcitation should play an important role in the overall laser ablation process, the pulse width used in the current work of about 150 fs is quite long enough to result in an anisotropic amorphization process even in single-shot duration. Even if we are not able to know any conclusive underlying mechanism within the current observations, the supposed strong anisotropy, likely originating from higher-order contributions to an optical absorption near and beyond the threshold intensity, is consistent with the observed anisotropy in size distribution and fluence-dependent particle numbers in Figs. 2(b) and 2(c).
We greatly acknowledge the financial contribution from the Ministry of Commerce, Industry and Energy of Korea.
References and links
2. L. T. Canham, “Silicon quantum wire array fabrication by electrochemical and chemical dissolution of wafers,” Appl. Phys. Lett. 57, 1046 (1990). [CrossRef]
3. F. Korte, S. Nolte, B. N. Chichkov, T. Bauer, G. Kamlage, T. Wagner, C. Fallnich, and H. Welling, “Far-field and near-field material processing with femtosecond laser pulses,” Appl. Phys. A 69, S7 (1999).
4. P. P. Pronko, S. K. Dutta, J. Squier, J. V. Rudd, D. Du, and G. Mourou, “Machining of sub-micron holes using a femtosecond laser at 800 nm,” Opt. Commun. 114, 106 (1995). [CrossRef]
5. J. P. Sylvestre, A. V. Kabashin, E. Sacher, M. Meunier, and J. H. T. Luong, “Stabilization and size control of gold nanoparticles during laser ablation in aqueous cyclodextrins,” J. Am. Chem. Soc. 126, 7176 (2004). [CrossRef] [PubMed]
6. F. Mafune, J.-Y. Kohno, Y. Takeda, and T. Kondow, “Full physical preparation of size-selected gold nanoparticles in solution: laser ablation and laser-induced size control,” J. Phys. Chem. B 106, 7575 (2002). [CrossRef]
7. J.-P. Sylvestre, S. Poulin, A. V. Kabashin, E. Sacher, M. Meunier, and J. H. T. Luong, “Surface chemistry of gold nanoparticles produced by laser ablation in aqueous media,” J. Phys. Chem. B 108, 16864 (2004). [CrossRef]
8. T.-H. Her, R. J. Finlay, C. Wu, and E. Mazur, “Femtosecond laser-induced formation of spikes on silicon,” Appl. Phys. A 70, 383 (2000). [CrossRef]
9. S. C. Jeoung, H. S. Kim, M. I. Park, J. Lee, C. S. Kim, and C. O. Park, “Preparation of room-temperature photoluminescent nanoparticles by ultrafast laser processing of single-crystalline Ge,” Japn. J. Appl. Phys. 44, 5278 (2005). [CrossRef]
10. The threshold energy 0.492J/cm2 is defined as the energy showing as amorphous layer on the part of focusing area after ablation.
11. K. Sokolowski-Tinten, C. Blome, C. Dietrich, A. Tarasevitch, M. H. Hoegen, and D. Linde, “Femtosecond x-ray measurement of ultrafast melting and large acoustic transients,” Phys. Rev. Lett. 87, 225701 (2001). [CrossRef] [PubMed]
12. C. V. Shank, R. Yen, and C. Hirlimann, “Time-resolved reflectivity measurements of femtosecond-optical-pulse-induced phase transitions in silicon,” Phys. Rev. Lett. 50, 454 (1983). [CrossRef]
13. D. C. Sayle and S. C. Parker, “Encapsulated oxide nanoparticles: the influence of the microstructure on associated impurities within a material,” J. Am. Chem. Soc. 125, 8581 (2003). [CrossRef] [PubMed]
15. A. Murali, A. Barve, V. J. Leppert, S. H. Risbud, I. M. Kennedy, and H. W. H. Lee, “Synthesis and characterization of indium oxide nanoparticles,” Nano Lett. 1, 287 (2003). [CrossRef]
16. J. Bonse, S. M. Wiggins, and J. Solis, “Dynamics of femtosecond laser-induced melting and amorphization of indium phosphide,” J. Appl. Phys. 96, 2352 (2004). [CrossRef]