Experimental and theoretical investigation on fs-laser-induced nanostructure formation on thin gold films

Transcription

1 Experimental and theoretical investigation on fs-laser-induced nanostructure formation on thin gold films Dirk Wortmann, Jürgen Koch, Martin Reininghaus, Claudia Unger, Conny Hulverscheidt, Dmitry Ivanov, and Boris N. Chichkov Citation: Journal of Laser Applications 24, (2012); doi: / View online: View Table of Contents: Published by the Laser Institute of America Articles you may be interested in Effects of ion and nanosecond-pulsed laser co-irradiation on the surface nanostructure of Au thin films on SiO2 glass substrates J. Appl. Phys. 115, (2014); / Nanoparticle formation after nanosecond-laser irradiation of thin gold films J. Appl. Phys. 112, (2012); / Investigation of nonlinear optical properties of gold nanograins embedded in indium oxide films by reflection Z- scan using continuous laser J. Appl. Phys. 108, (2010); / Ultrafast dynamics of periodic nanostructure formation on diamondlike carbon films irradiated with femtosecond laser pulses Appl. Phys. Lett. 89, (2006); / Correlations between embedded single gold nanoparticles in SiO 2 thin film and nanoscale crater formation induced by pulsed-laser radiation J. Appl. Phys. 92, 5720 (2002); /

2 JOURNAL OF LASER APPLICATIONS LASER GENERATED SUB-100 NM STRUCTURES JULY 2012 Experimental and theoretical investigation on fs-laser-induced nanostructure formation on thin gold films Dirk Wortmann Lehrstuhl fuer Lasertechnik, RWTH Aachen University, Steinbachstrasse 15, Aachen, Germany Jürgen Koch Laser Zentrum Hannover e.v., Hollerithallee 8, Hannover, Germany Martin Reininghaus Lehrstuhl fuer Lasertechnik, RWTH Aachen University, Steinbachstrasse 15, Aachen, Germany Claudia Unger Laser Zentrum Hannover e.v., Hollerithallee 8, Hannover, Germany Conny Hulverscheidt Lehrstuhl fuer Lasertechnik, RWTH Aachen University, Steinbachstrasse 15, Aachen, Germany Dmitry Ivanov Lehr-und Forschungsgebiet Nichtlineare Dynamik, RWTH Aachen University, Steinbachstrasse 15, Aachen, Germany Boris N. Chichkov Laser Zentrum Hannover e.v., Hollerithallee 8, Hannover, Germany (Received 20 October 2011; accepted for publication 22 June 2012; published 16 July 2012) In this paper, the authors report on the formation of nanobumps and nanojets on thin gold films, induced by single fs-laser pulse irradiation. Experimental results on the structure size and shape depending on the pulse energy and the pulse duration are presented. For the first time, the process of short laser pulse nanostructuring on thin metal films was modeled by molecular dynamic simulations on the scale directly accessible in the experiments. Additionally, pump-probe experiments were performed for in-situ visualization of the structure formation. VC 2012 Laser Institute of America. Key words: fs-laser nanostructuring, nanojets, molecular dynamic simulation, nanostructure formation dynamics I. INTRODUCTION Micro and nanoprocessing of materials is one of the most intensively studied applications for ultrashort pulsed laser radiation. Pulse durations in the femtosecond regime are smaller than the heat diffusion time and enable an energy deposition only in the irradiated region and a depth of a few nanometers with negligible heat affected zone. Therefore, fs-laser radiation is a distinguished tool to produce micro and nanostructures with sharp contours in metals as well as in dielectrics. Taking advantage of well-defined ablation and modification threshold and nonlinear absorption effects, one can beat the diffraction limit and produce smaller structure sizes by precise control of the laser pulse energy. Fs-laser processing enables structure sizes even in the sub-100 nm regime: one example is so called nanobumps and nanojets on noble metal films. 1,2 While the experimental work on nanostructuring on material surfaces has achieved a significant success, especially in application for Bio-technologies, 3 the mechanisms of surface restructuring with short laser pulses are still controversially discussed. In the particular case of thin metal films, possible mechanisms include Marangoni convection, 2 peeling and expansion of the film due to the pressure of evaporating material, 4,5 as well as thermo-elastic and plastic deformation of the film, suggested based on recent continuum simulations. 5 Significant efforts were made to conclude on the mechanism responsible for nanostructuring as full or partial combination of the mechanisms pointed out above. 6 Experimentally, nevertheless, the analysis of nanobump and nanojet formation mechanism is commonly limited to observations of nanostructures formed after the laser pulse, which does not reveal the complete process kinetics. Therefore, a combination of experiment and numerical simulation is needed for a complete process understanding. As it was demonstrated in Refs. 7 and 8, one attractive computational method is provided by atomistic-continuum MD-TTM simulation (molecular dynamics coupled to two temperature model). Fast nonequilibrium phase transformations at atomic scale are taken into account with MD and in continuum include the description of fast electron heat conduction mechanism as well as the laserinduced electron-phonon nonequilibrium and laser light absorption with TTM. In this paper, we demonstrate for the first time the simulation on an experimentally accessible scale. Time-resolved pump-probe observation of the structure forming process also helps to explain them by using the relevant time-constants as input for the numerical simulation. 9 II. NANOSTRUCTURE FORMATION In metals, nanostructures can be fabricated based on the laser-induced dynamics in a small volume of molten material. Slow energy transfer to the lattice due to the weak X/2012/24(4)/042017/6/ VC 2012 Laser Institute of America

3 J. Laser Appl., Vol. 24, No. 4, July 2012 Wortmann et al. FIG. 1. SEM-images of nanostructures, produced with single pulses of fs-laser radiation (t ¼ 100 fs, l ¼ 800 nm) focused with a 50 mm lens on a thin gold film. electron-phonon coupling in noble metals leads to a relatively long living molten phase, enabling surface structuring without ablation. In the experiment, single pulses of fs-laser radiation (k ¼ 800 nm) with Gaussian intensity distribution have been focused on a thin gold film (d ¼ 40 nm) on a silicon substrate. At small pulse energies, a bumplike structure appears, with a characteristic size much smaller than the laser spot size smallest feature sizes are below 100 nm (Fig. 1). With increasing pulse energy, the diameter of the bump increases. At a certain pulse energy, the bump collapses and a jet of molten material grows from the middle. A further increase of the pulse energy leads to the ablation of a molten droplet and a destroyed film surface. The characteristic pulse energy for the different features depends on the focus diameter, but the principal behavior is the same and independent if using a 50 mm lens or microscope objectives with numerical apertures of 0.2 to 1.2. Calculation of the absorbed fluence using the pulse energy, the focus diameter, and the reflectivity of gold leads to a threshold for the bump formation of about 5 mj/cm 2, a nanojet formation between 10 and 20 mj/cm 2, and an ablation threshold of about 20 mj/cm 2. Changing the pulse duration by changing the compressor length of the chirped pulse laser amplifier does not lead to significant changes in the resulting structures (Fig. 2). While most material structuring or ablation processes using continuous wave or long pulsed laser radiation are intensity dependent, the formation of nanobumps and nanojets only depends on the fluence, up to a pulse duration of at least 8 ps. This behavior can be explained by an ultrafast melting of the material. The laser energy is deposited in the film on a time scale below the electron-phonon coupling time, and no significant heat diffusion occurs before the melting of the irradiated material occurs. III. MOLECULAR-DYNAMIC SIMULATION To compare the modeling with experimental results, modeling on an experimental scale is necessary. For feasible calculation times, symmetrical fabrication of nanostructures was utilized. A flat top pulse, elongated in one direction, was used for generation of a nanobarrier kind of structure on the surface of 60 nm Au film situated on a silica substrate (Fig. 3, top). Since we do not expect a strong temperature gradient in the lateral (X, Fig. 3) direction, for an identical simulation/experiment comparison, we constructed the FIG. 2. SEM images of nanobumps and nanojets produced with different pulse durations as indicated above. Pulse energy is 150 nj in the top row and 180 nj in the bottom.

4 J. Laser Appl., Vol. 24, No. 4, July 2012 Wortmann et al FIG. 3. SEM-image of nanobumps and a nanobarrier on a 60 nm gold film, generated with a 30 fs flat top pulse (top) and schematic representation of the computational cell used in large scale modeling (bottom). FIG. 4. The atomic configurations from the modeling of 30 fs-laser pulse nanostructuring on 60 nm Au film on a silica substrate. A 0.5 nm thick slice of MD- TTM domain covering 2 lm width is shown in each snapshot. The atoms are colored according to the CSP for distinguishing between crystalline surrounding (CSP < 0.11) and liquid ambient (CSP > 0.11).

5 J. Laser Appl., Vol. 24, No. 4, July 2012 Wortmann et al. FIG. 5. Measured reflectivity of a gold film depending on the pulse energy (blue: d ¼ 200 nm, red: d ¼ 40 nm) and calculated reflectivity depending on the electron temperature. computational cell schematically shown in Fig. 3, bottom. The modeled gold sample consisting of atoms with dimensions of nm 3 was taken in X, Y, and Z axes correspondingly representing, therefore, a 60 nm Au film on a silica substrate. The MD-TTM domain is covering 5 lm in Y direction ending with nonreflective boundaries (NRBs) as they were described in Refs. 7 and 10. Consisting of three perfect fcc plains, NRBs were devoted to behave dynamically to mimic bulk material behavior beyond MD- TTM domain and to absorb pressure waves resulting from relaxation of the laser-induced pressure waves. The NRBs, however, are transparent for the heat flux, and only the TTM model was solved beyond them, subject to the electron and phonon temperatures on the scale of 30 lm in Y direction. In the x direction, we apply periodic boundary (PB) conditions for the simulation of the material slice of 10 nm thickness. The cohesive energy between the Au sample and the silica substrate was approximated as 10% from that of Au bulk material. This number was chosen in accordance with statistics and experimental measurements. The free surface atop was exposed to a 30 fs flat top pulse at the incident fluence of 275 mj/cm 2. The MD-TTM domain was divided to 3D mesh with 1 nm 3 unit cell, and the thermophysical properties of the atomic subsystems (such as temperature and pressure) were averaged over the closest 26 cell s neighbors (2500 atoms) to suppress the fluctuations due to MD part. The results of the simulation can be observed in the following pictures. In Fig. 4, several snapshots from the modeling are shown up to 1500 ps for visualization of the atomic configuration of the sample. The atoms are colored according to the central symmetry parameter (CSP) which was used here to distinguish between atoms with crystal structure from those submerged in the liquid. The irradiated laser energy induces melting of the material among with acceleration in the upward direction. After about 500 ps, a detachment of the film is observed, and even after 1500 ps, a significant amount of molten material possessing a strong momentum is visible, assuming a much longer time of cooling and resolidification. In a similar process, a Gaussian intensity distribution leads to the formation of spherical nanodroplets, which fly away from the gold film surface. Following a similar approach as demonstrated in Ref. 8, we can extract the physical mechanisms involved into nanostructure formation on the experimental scale of several microns. In general, the dynamics of plotted quantities confirms two essential mechanisms responsible for the nanostructure formation. The relaxation of laser-induced stresses, due to strong laser heating under conditions of inertial stress confinement leads to acceleration of the molten matter in the upward direction. As a result, the established hydrodynamic motion is responsible for the nanostructure growth. The resolidification of this structure, on the other hand, is associated with the flux of thermal energy away from the central part due to fast electron heat conduction process. Discrepancies in the experimentally observed threshold fluences for the structure formation and the absorbed fluences used as input parameters for the simulation can be explained by a change in the material absorptivity during the pulse due to high electron temperatures. At high electron temperatures induced by strong laser fields, the reflectivity of the material decreases and the absorptivity increases. Figure 5 shows the measured reflectivity of a gold film versus the pulse energy, normalized to the reflectivity based on the Fresnel equations only. For comparison, the electron temperature dependent reflectivity based on the Hopkins model is FIG. 6. Scheme of the pump-probe setup. TM: turning mirror, DM: dichroic mirror, GS: gold-coated quartz glass substrate, MO: microscope objective, and BD: beam dump.

6 J. Laser Appl., Vol. 24, No. 4, July 2012 Wortmann et al FIG. 7. (a) Pump-probe microscope images of the nanobump and nanojet formation with different time delays; (b) SEM image of the final structure. visualized. In future simulations, a reflectivity and, therefore, an absorptivity change depending on the incident pulse energy will be included. IV. FORMATION DYNAMICS Studying of the formation dynamics was achieved by pump-probe experiments. The imaging setup (Fig. 6) was a modification of the one already published in Refs. 11 and 12. It consists of a frequency-doubled Nd:YAG laser (Quanta- Ray DRC-11, Spectra Physics) with 532 nm wavelength and 9 ns pulse duration, lenses for beam collimation and focusing, microscope objectives (20 Zeiss, 100 Zeiss), and a SLR camera (EOS 450D, Canon). The jets were generated by an amplified Ti:Sa fs-laser (Femtopower Compact Pro, Femtolasers Produktions GmbH) which delivers sub-30-fs pulses at 800 nm with 1 khz repetition rate. An achromatic lens with a focal length of 80 mm was used to focus the fsradiation on the gold-coated quartz glass slide. Coating with a 60 nm thick gold layer was performed by a sputter-coater (Cressington 208HR), and the substrate was provided by Hellma Optics, Jena. The temporal delay between the both lasers was set by a real-time PC, presented in Ref. 13, and continuously measured by a fast-rising photodiode. This setup allows time delays in the ns range up to ms range with a temporal resolution of approximately 10 ns. Ex-situ measurements of the nanojets have been conducted by scanning electron microscope (SEM) imaging (FEI Quanta 400F). Figure 7(a) shows the time-resolved image series of the nanojet generation at a fluence of 0.41 Jcm 2. The first bump structure can be visualized after a delay of 10 ns with the dimensions of approximately 6.5 lm width at the base and 2.5 lm maximum height. The image at 25 ns delay shows an elongated bump with a height increase of almost 15% and a width decrease of about 15%. After 25 ns, the bump has grown to 3.6 lm height and tapered at the front tip. The width at base stays constant compared to the image at 25 ns delay. After 75 ns, the tapering of the bump continued and formed a jetlike structure. A growth in height to 4.4 lm and a constant base width are measured. The bump at the base of the jet decreases at a delay of 100 ns, whereas the jet itself thickens at the top without further increase of length. This thickening transforms into a spherical droplet structure at a delay of 200 ns. After 300 ns, the length of the jet decreases to approximately 3.2 lm. The jet-forming dynamic is completed after 500 ns since the appearance of the jet at a delay of 4 s is the same. Figure 7(b) shows a SEM image of the final structure under 45. The diameter of the bump is 6 lm which is the same value as in the image after 10 ns. The height of the jet is approximately 3.1 lm in the SEM image. This value is in good agreement to the value measured in the time-resolved image at a delay of 500 ns. V. CONCLUSION AND OUTLOOK Nanobump and nanojets on a thin gold film are produced by single fs-laser pulses. Feature sizes smaller than the laser spot size are observed. Variations of the pulse duration show an independence of the feature size and shape on the pulse duration between 100 and 8000 fs, indicating an ultrafast melting process in the absence of heat diffusion, although feature sizes in the 100 nm regime are surprisingly small for a structure formation by melt dynamics. For the first time, the process of short laser pulse nanostructuring on thin metal films was modeled by molecular dynamic simulations on the scale directly accessible in the experiments. The relaxation of the laser-induced stresses inside of the molten target serves as a main driving force for the nanojetlike structure growth. The electron heat conduction, on the other hand, is providing the conditions for its fast resolidification. Optical pump-probe measurements indicate a formation time of the bump of several ten nanoseconds. The formation mechanism appears similar to jet formation which occurs during cavitation bubble interaction with a free surface. 14,15 In future work, we will include an absorptivity change due to the increased electron temperature in the simulation to FIG. 8. Scheme of an EUV-pump-probe microscope for the observation of nanobump and -jet formation on a sub-100 nm scale.

7 J. Laser Appl., Vol. 24, No. 4, July 2012 Wortmann et al. further improve the comparability of experimental and modeling results. Additional, an extreme ultraviolet (EUV)-pumpprobe microscope (Fig. 8) will be set up with a spatial resolution of about 30 nm to observe the structure formation on a smaller scale. ACKNOWLEDGMENTS The authors acknowledge the Juelich Super Computer Facility (Juelich, Germany) teams for the technical support provided for super large scale parallel simulations. The presented work was done under financial support due to DFG grant within the SPP 1327 Sub-100 nm structures for optical and biomedical applications. 1 J. Koch, F. Korte, C. Fallnich, A. Ostendorf, and B. Chichkov, Directwrite subwavelength structuring with femtosecond laser pulses, Opt. Eng. 44(5), (2005). 2 F. Korte, J. Koch, and B. N. Chichkov, Formation of microbumps and nanojets on gold targets by femtosecond laser pulses, Appl. Phys. A 79, (2004). 3 E. Fadeeva, V. K. Truong, M. Stiesch, B. N. Chichkov, R. J. Crawford, J. Wang, and E. P. Ivanova, Bacterial retention on superhydrophobic titanium surfaces fabricated by femtosecond laser ablation, Langmuir 27, (2011). 4 Y. Nakata, N. Miyanaga, and T. Okada, Effect of pulse width and fluence of femtosecond laser on the size of nanobump array, Appl. Surf. Sci. 253, (2007). 5 Y. P. Meshcheryakov and N. M. Bulgakova, Thermoelastic modeling of microbump and nanojet formation on nanosize gold films under femtosecond laser irradiation, Appl. Phys. A 82, (2006). 6 A. I. Kuznetsov, J. Koch, and B. N. Chichkov, Nanostructuring of thin gold films by femtosecond lasers, Appl. Phys. A 94, (2009). 7 D. S. Ivanov and L. V. Zhigilei, Combined atomistic-continuum modeling of short pulse laser melting and disintegration of metal films, Phys. Rev. B 68, (2003). 8 D. S. Ivanov, B. Rethfeld, G. M. O Connor, T. J. Glynn, A. N. Volkov, and L. V. Zhigilei, The mechanism of nanobump formation in femtosecond pulse laser nanostructuring of thin metal films, Appl. Phys. A 92, (2008). 9 I. Mingareev and A. Horn, Time-resolved investigations of plasma and melt efections in metals by pump-probe shadowgraphy, Appl. Phys. A 92, (2008). 10 C. Schäfer, H. M. Urbassek, L. V. Zhigilei, and B. J. Garrison, Pressuretransmitting boundary conditions for molecular dynamics simulations, Comput. Mater. Sci. 24, (2002). 11 C. Unger, M. Gruene, L. Koch, J. Koch, and B. N. Chichkov, Timeresolved imaging of hydrogel printing via laser-induced forward transfer, Appl. Phys. A 103, (2011). 12 M. Gruene, C. Unger, L. Koch, A. Deiwick, and B. N. Chichkov, Dispensing pico to nanolitre of a natural hydrogel by laser-assisted bioprinting, Biomed. Eng. Online 10, 19 (2011). 13 J. Koch, E. Fadeeva, M. Engelbrecht, C. Ruffert, H. H. Gaatzen, A. Ostendorf, and B. N. Chichkov, Maskless nonlinear lithography with femtosecond laser pulses, Appl. Phys. A 82, (2006). 14 P. B. Robinson, J. R. Blake, T. Kodama, A. Shima, and Y. Tomita, Interaction of cavitation bubbles with a free surface, J. Appl. Phys. 89, (2001). 15 A. Pearson, E. Cox, J. R. Blake, and S. R. Otto, Bubble interactions near a free surface, Eng. Anal. Boundary Elem. 28(4), (2004).