This work presents time-resolved images of femtosecond-laser-induced melt dynamics in 60 nm gold films on glass substrates. Melt dynamics induced by laser radiation with focus diameters of 6 μm and 8 μm (FWHM) at constant laser fluence is investigated with a temporal resolution of 10 ns. In both cases, the formation of the microbumps and gold jets takes at least 250 ns. It is shown that the formation process can be compared to jetting behavior induced by cavitation bubbles near a free liquid surface. This is confirmed by SEM illustrating a re-entrant spike through a hole in the microbump.
© 2012 Optical Society of America
Nanostructuring of thin metal films by femtosecond laser irradiation and possible formation mechanisms have evoked wide interest in the recent years. By applying single ultrashort and tightly focused laser pulses with pulse energies below the ablation threshold, microbump and nanojet structures were demonstrated [1–4]. Furthermore, beam interference methods [5–7] or even nanosecond pulses [8, 9] were applied to generate similar structures.
Femtosecond laser absorption in metals is a non-equilibrium process, usually described by the one-dimensional, two-temperature diffusion model, in which laser pulse duration is shorter than the electron cooling time [10, 11]. At low laser intensities, which are considered in this paper, the free electrons absorb the incident laser energy and are heated up quickly. Subsequently, the energy of these heated electrons is transfered to the lattice on the time scale of the order of several ps . Therefore, femtosecond laser ablation is a machining technique with smallest heat-affected zone. Usually, melting can be neglected because most of the heat is led away by convection with the hot ablated particles. However, when working below the ablation threshold this type of convection cannot occur and melting can become important. Dynamics in the molten material become possible. Additionally, in noble metals, like Au and Ag, with weak electron-phonon coupling compared to Cr, Mo or W , energy of the electrons is transferred to the lattice much slower. As a consequence melt dynamics are much slower, too. If melting effects are reproducible depends also on the laser fluence, machining strategy and the material properties. In this work, we apply laser fluences between the melting and ablation thresholds. Besides, Au shows low brittleness and high ductility leading to high degree of deformation even at low temperatures. Those material properties of Au allow controlled and repeatable generation of microbumps with jet-like protrusions on top . Furthermore, it was demonstrated that the transfer of spherical gold droplets from the front of the nanojet is possible [13–15].
Since the formation mechanism has not been visualized yet, various explanations were proposed: Marangoni convection flow of the melted material , evaporation of gold below the surface  and relaxation of compressive stresses generated by fast laser heating and melting of the film [16–19]. Microbump formation was attributed to thermoplastic deformation of thin films promoting nanojet development upon melting . The simulations of Ivanov et al. [16, 18, 19] were based on a hybrid atomistic-continuum model which is well-suited for investigations of laser-induced metal melting processes, laser spallation, and ablation. However, these simulations are time consuming and only delays up to 250 ps were studied.
Nanosecond lasers also show the ability to produce jet-like structures with droplets on top. In , Seifert and Betz presented a numerical model of the jet and droplet formation in 200 nm Au film. The droplets were generated after several laser pulses, whereas the material cooled down between every pulse. The model was based on the hydrodynamic behavior where the droplet formation is governed by a balance between inertial forces and the surface tension restraining forces. Willis et al.  showed that the droplet formation and transfer are also possible after one single laser pulse with Al and Ni films. The laser fluence of the ns laser has to be above the melting threshold. In  the same group introduced a model explaining the surface deformation of 1 μm thick Al layers on a glass substrate with the help of volumetric expansion which induced fluid motion in the liquid metal. The whole process took 135 ns and resulted in a bump structure. Compared to the fs laser-induced structures, the bump produced by ns laser was not hollow. However, in recent publications of Moening [23, 24] and Riedel  the formation of hollow bumps or hollow microbumps with jets achieved with ns laser irradiation was presented. In  the influences of film adhesion to the substrate and the film thickness were studied. It was concluded that strong adhesion between the metallic film and substrate, as well as metallic layer thickness above 300 nm, prevented bump formation.
Femtosecond laser interaction with metals has been studied in detail for several years by means of pump-probe setups. However, most of the publications show data only up to a temporal delay of a few ps because those studies were focused on the investigation of atomistic scale effects, e. g. [26, 27]. Whereas, publications of Mingareev et al.  and Domke et al.  cover a long temporal range but at high fluences , large focal sizes, for Al, Cu  and Mo  targets. In the study of Domke et al. Newton’s rings appear, which were also observed in the work of von der Linde et al. [12, 30], where ablation of a thin Al film was studied by pump-probe microscopy up to a delay of 10 ns.
This paper presents time-resolved side-view microscopic shadowgraphic images of the formation dynamics of fs laser-induced microbumps and jets on thin gold films at a fluence close to the ablation threshold. An electronically controlled pump-probe setup is applied to investigate this process up to a delay of several 100 ns with a temporal resolution of 10 ns. The dependence of two different foci on the morphology and time-scale of the process is shown. Based on the time-resolved images, a scheme of the jet formation is developed and compared with simulations of laser ablation and jetting processes.
Jet formation dynamics was studied by pump-probe experiments. Figure 1 shows a scheme of the setup. The imaging setup is a modification of the one already published in [31, 32]. It consists of a frequency-doubled Nd:YAG laser (Quanta-Ray DRC-11, Spectra Physics) with 532 nm wavelength and 9 ns pulse duration as process illumination. A 100-X microscope objective (NA 0.75, Zeiss) and a SLR camera (EOS 450D, Canon) are applied to image the process. The jets are 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. Two different focusing optics were chosen: a 5-X microscope objective (Leica, NA 0.15) or an achromatic lens with a focal length of 80 mm. To ensure high reproducibility, an auto-focus system kept a constant distance between focusing optics and the substrate surface. Edge-polished substrates were made of quartz-glass and delivered by Hellma Optics, Jena. Subsequently, the substrates were sputter-coated (Cressington 208HR) with a 60 nm thick gold layer. The temporal delay between both lasers was set by a real-time PC, presented in , and continuously measured by a fast-rising photodiode. This setup allows the generation of time delays in the ns range and longer with a temporal resolution of approximately 10 ns. One image per formation process was taken. For every formation process, a new spot on the gold-coated surface was chosen. Ex-situ measurements of the microbumps and jets were conducted by SEM imaging (FEI Quanta 400F).
3. Results and discussion
Figure 2(a) displays time-resolved images at a constant laser pulse energy of 160 nJ focused by a 5-X microscope objective. This leads to a 6 μm (FWHM) focus diameter on the metal film surface and corresponds to a laser fluence of 0.39 J/cm2. The substrate edge is visible at the bottom of the image. The black horizontal lines in the pictures are caused by the coherent laser probe radiation which is deflected at the glass substrate and creates interference. After 5 ns, the thin gold film shows a protrusion with approximately 2.4 μm height and a base width of 4.5 μm. Subsequently, the protrusion tapers at the front while the base width stays constant and the height is about 2.9 μm. At a delay of 25 ns, the bulge appears in a triangular shape while the base width is still constant and the height increases to 3.2 μm. Only 10 ns later, the bulge structure seems to collapse and a jet starts to form from the middle of the triangular bulge. While the bump decreases in height, the jet grows to approximately 4.2 μm length. In the following images the bump is barely detectable but the jet is continuously growing till a delay of 125 ns. The jet diameter after 100 ns is approximately 1 μm. After 150 ns, the jet detaches from the base and due to Plateau-Rayleigh instability the jet falls into two parts: one small spherical droplet at the front and a bigger elongated droplet. A very thin spike remains at the base of the structure. In the following images the shape of the droplets becomes spherical and both droplets move away from the substrate. The velocity of the first droplet is around 30 m/s whereas the velocity of the second droplet is approximately 20 m/s. Note, that even though one image represents one single process, the process stability enables the measurement of droplet velocities. The stability of the process and the dependence on the laser pulse energy is illustrated in Fig. 3. With increasing pulse energy more material is molten and the jet has more time till it solidifies.
The Plateau Rayleigh instability is responsible for the liquid jet breaking into droplets. To calculate the critical jet breakup time τcrit, the following formula  can be applied when neglecting viscous forces and gravity:35] suggested the value of . Taking for liquid gold at the melting point (1336 K): ρ = 17.4 g/cm3 , σ = 1.113 N/m  and the liquid jet radius a of approximately 500 nm, the critical breakup time τcrit is calculated as 63 ns. From the image series in Fig. 2(a) one can estimate that the jet starts at a delay of 50 ns and falls into droplets at a delay of 125 ns. Therefore, the experimental jet life time before the breakup is approximately 75 ns. This means that the observed liquid gold jetting is in very good agreement with the theory of droplet breakup due to Plateau Rayleigh instability. When the acceleration of the jet is neglected, the critical breakup length of the jet can be estimated with the help of the following formula:
Furthermore, an image series with an 80 mm achromatic lens as focusing optics which led to 8 μm (FWHM) focus diameter was recorded (see Fig. 4(a)). Those images were obtained on the same substrate as the images shown in Fig. 2(a) to preclude influences of the gold layer properties. Due to the larger focus, the pulse energy required to obtain visible laser-induced jet formation had to be increased up to 255 nJ for a comparable fluence of 0.35 J/cm2. The images in Fig. 4(a) show that the bulge formation velocity is slower than in Fig. 2(a). The bulge height after 25 ns is around 3.2 μm. After 100 ns the jet thickness is around 0.85 μm. 50 ns later, the bulge is no longer visible and the jet has not increased its length. Subsequently, a gold sphere at the tip of the jet is formed and the jet length decreases. It seems as if the still not solid jet is compressed by the weight of the sphere on top. At a delay of 250 ns the formation is finished because no changes at delays of 300 ns and longer can be detected. The diameter of the gold sphere on top of the short gold jet in the SEM picture shown in Fig. 4(b) is in good agreement with the dimension measured from the time-resolved images. In the SEM picture, the diameter is 1.5 μm whereas the time-resolved images show a diameter of approximately 1.6 μm. The SEM picture shows that the gold film which has been stretched for bulge formation solidified with radial structures.
Figure 5 shows energy-resolved SEM images at a constant focus diameter of 8 μm. Compared to the small focus results shown in Fig. 3 the lateral dimension of the microbump is too large and therefore, very unstable to solidify in an elevated shape.
The formation dynamic of the gold jets and microbumps seems very similar to other hydrodynamic processes, e.g. the cavitation-bubble-induced dynamic near a free liquid surface [38, 39] or the jetting behavior of liquids during laser-induced forward transfer [31, 32]. In the latter cases, μs timescale instead of ns timescale applies because of the usually larger focus spot size and liquid layer thickness in the μm range.
The scheme in Fig. 6 summarizes the formation steps. The red areas qualitatively illustrate molten gold because it is not possible to tell exactly which parts of the gold film are molten and which are only heated up at a certain delay. After the femtosecond laser pulse, the gold film is radially heated by the absorbed laser radiation (Fig. 6(a) and (b)). The heat is mainly conducted by the gold film because thermal conductivity of the gold (318 W·(m·K)−1 at 20°C ) is much higher than that of the quartz glass (1.4 W·(m·K)−1 ) and convection can be neglected. Therefore, a 2D model for heat conduction can be assumed. Molten gold shows a lower density (17.3 g/cm3) than solid (19.3 g/cm3) gold. This results in stresses in the gold film and subsequently the gold expands. Due to the constraints imposed by the glass slide and the solid gold, the liquid gold is accelerated perpendicular to the glass substrate (Fig. 6(c)).The adhesion between gold and the glass substrate is low. Therefore, the complete film detaches from the glass and forms a protrusion. Please note, that based on the time-resolved images it is impossible to tell if the gold film is already completely molten when the protrusion forms. However, for jet formation the gold film needs to be in molten state. At the tip of the protrusion, the jet forms due to the high surface tension of the gold, inertia and the atmospheric pressure (Fig. 6(d) and (e)). While in cavitation bubble experiments [38, 39] and laser-induced transfer of liquids [32, 42], a counterjet in the opposite direction of the main jet was visible, the existence of this counterjet was unclear in the case of the fs-laser-induced gold jets. Figure 7 shows a gold microbump and jet which has been produced on the same sample but with a 20 mm achromatic focus lens. This focusing optic led to a focus spot size of around 2.5 μm (FWHM). Accidentally, the bubble was partially opened at some point of the formation process. The counterjet is clearly visible through the hole inside the bump. It seems that the counterjet was liquid when it touched the glass substrate because of the flattened shape at the end of the jet. This proves the resemblance to the jetting processes induced by cavitation bubbles or laser-induced forward transfer. Another similarity is breaking of the jet into droplets. Due to the high surface tension of molten gold (1130 mN/m ) the aspect ratio of the jet is smaller when it starts to breakup into droplets compared to water jets. Another process limiting the aspect ratio is the cooling of the thin film which starts from the outside. This leads to constantly changing fluid properties, mainly surface tension and viscosity. Compared to ablation of bulk Al with over 4.5 times higher laser fluence , the fs laser-induced microbump formation process can be considered as a process where the droplet ablation starts very late. In  material removal can be observed already after 14.6 ns whereas in our study, the material removal starts after 150 ns. Basically, it can be stated that the formation speed increases with increasing laser energy at a constant focus diameter. Additionally, the wall thickness of the microbump is always thinner than the original layer thickness.
In this work, the generation mechanism of femtosecond-laser-induced microbumps and jets has been demonstrated with the help of time-resolved imaging. It was shown that a single fs-laser pulse induces melting inside the focus diameter of the 60 nm gold layer. The melted material undergoes a reproducible jetting behavior that is comparable to cavitation-bubble-induced liquid jets near a free surface or liquid jetting initiated by laser-induced forward transfer. This theory is supported by the observation of a counterjet. The time duration required for solidification of the bump and jet can take several 100 ns and depends on the laser fluence. The findings of this study serve as a basis for new and more accurate simulations.
The authors are grateful for financial support provided by Deutsche Forschungsgemeinschaft (DFG) within the SPP 1327 ”Sub-100nm structures for optical and biomedical applications”.
References and links
1. F. Korte, J. Koch, and B. N. Chichkov, “Formation of microbumps and nanojets on gold targets by femtosecond laser pulses,” Appl. Phys. A 79, 879–881 (2004). [CrossRef]
2. J. Koch, F. Korte, C. Fallnich, and B. N. Chichkov, “Direct-write subwavelength structuring with femtosecond laser pulses,” Opt. Eng. 5, 051103 (2005). [CrossRef]
3. J. Koch, F. Korte, T. Bauer, C. Fallnich, and B. N. Chichkov, “Nanotexturing of gold films by femtosecond laser-induced melt dynamics,” Appl. Phys. A 81, 325–328 (2005). [CrossRef]
4. A. I. Kuznetsov, J. Koch, and B. N. Chichkov, “Nanostructuring of thin gold films by femtosecond lasers,” Appl. Phys. A 94, 221–230 (2008). [CrossRef]
5. Y. Nakata, T. Okada, and M. Maeda, “Nano-sized hollow bump array generated by single femtosecond laser pulse,” Jpn. J. Appl. Phys. 42, L1452–L1454 (2003). [CrossRef]
6. 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, 6555–6557 (2007). [CrossRef]
7. Y. Nakata, T. Hiromoto, and N. Miyanaga, “Mesoscopic nanomaterials generated by interfering femtosecond laser processing,” Appl. Phys. A 101, 471–474 (2010). [CrossRef]
8. D. A. Willis and V. Grosu, “Microdroplet deposition by laser-induced forward transfer,” Appl. Phys. Lett. 86, 244103 (2005). [CrossRef]
9. J. P. Moening, S. S. Thanawala, and D. G. Georgiev, “Formation of high-aspect-ratio protrusions on gold films by localized pulsed laser irradiation,” Appl. Phys. A 95, 635–638 (2009). [CrossRef]
10. S. Anisimov, B. Kapeliovich, and T. Perel’man, “Electron emission from metal surfaces exposed to ultrashort laser pulses,” Sov. Phys.-JETP 39, 375–377 (1974).
11. B. N. Chichkov, C. Momma, S. Nolte, F. von Alvensleben, and A. Tünnermann, “Femtosecond, picosecond and nanosecond laser ablation of solids,” Appl. Phys. A 63, 109–115 (1996). [CrossRef]
12. D. von der Linde, K. Sokolowski-Tinten, and J. Bialkowski, “Laser- solid interaction in the femtosecond time regime,” Appl. Surf. Sci. 109/110, 1–10 (1997). [CrossRef]
13. A. I. Kuznetsov, J. Koch, and B. N. Chichkov, “Laser-induced backward transfer of gold nanodroplets,” Opt. Express 17, 18820–18825 (2009). [CrossRef]
15. A. I. Kuznetsov, C. Unger, J. Koch, and B. N. Chichkov, “Laser-induced jet formation and droplet ejection from thin metal films,” Appl. Phys. A 106, 479–487 (2012). [CrossRef]
16. 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, 064114 (2003). [CrossRef]
18. 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, 791–796 (2008). [CrossRef]
19. D. S. Ivanov, Z. Lin, B. Rethfeld, G. M. O’Connor, T. J. Glynn, and L. V. Zhigilei, “Nanocrystalline structure of nanobump generated by localized photoexcitation of metal film,” J. Appl. Phys. 107, 013519 (2010). [CrossRef]
20. 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, 363–368 (2006). [CrossRef]
21. N. Seifert and G. Betz, “Computer simulations of laser- induced ejection of droplets,” Appl. Surf. Sci. 133, 189–194 (1998). [CrossRef]
22. D. A. Willis and V. Grosu, “The effect of melting-induced volumetric expansion on initiation of laser-induced forward transfer,” Appl. Surf. Sci. 253, 4759–4763 (2007). [CrossRef]
23. J. P. Moening, D. G. Georgiev, and J. G. Lawrence, “Focused ion beam and electron microscopy characterization of nanosharp tips and microbumps on silicon and metal thin films formed via localized single-pulse laser irradiation,” J. Appl. Phys. 109, 014304 (2011). [CrossRef]
24. J. P. Moening, “Formation of nano-sharp tips and microbumps on silicon and metal films by localized single-pulse laser irradiation,” Ph.D. thesis, The University of Toledo (2010).
25. S. Riedel, M. Schmotz, P. Leiderer, and J. Boneberg, “Nanostructuring of thin films by ns pulsed laser interference,” Appl. Phys. A 101, 309–312 (2010). [CrossRef]
27. T. Ao, Y. Ping, K. Widmann, D. F. Price, E. Lee, H. Tam, P. T. Springer, and A. Ng, “Optical properties in nonequilibrium phase transitions,” Phys. Rev. Lett. 96, 055001 (2006). [CrossRef] [PubMed]
28. I. Mingareev and A. Horn, “Time-resolved investigations of plasma and melt ejections in metals by pump-probe shadowgrpahy,” Appl. Phys. A 92, 917–920 (2008). [CrossRef]
30. D. von der Linde and K. Sokolowski-Tinten, “Physical mechanisms of short pulse laser ablation,” Appl. Surf. Sci.154–155, 1–10 (2000). [CrossRef]
31. C. Unger, M. Gruene, L. Koch, J. Koch, and B. N. Chichkov, “Time-resolved imaging of hydrogel printing via laser-induced forward transfer,” Appl. Phys. A 103, 271–277 (2011). [CrossRef]
32. M. Gruene, C. Unger, L. Koch, A. Deiwick, and B. Chichkov, “Dispensing pico to nanolitre of a natural hydrogel by laser-assisted bioprinting,” Biomed. Eng. Online 10:19 (2011). [CrossRef] [PubMed]
33. J. Koch, E. Fadeeva, M. Engelbracht, C. Ruffert, H. H. Gaatzen, A. Ostendorf, and B. N. Chichkov, “Maskless nonlinear lithography with femtosecond laser pulses,” Appl. Phys. A 82, 23–26 (2006). [CrossRef]
34. N. Chigier and R. Reitz, Recent Advances in Spray Combustion: Spray Atomization and Drop Burning Phenomena (AIAA, 1996), chap. Regimes of Jet Breakup and Breakup Mechanisms (Physical Aspects), 109–135.
35. P. O’Rourke and A. Amdsen, “The tab method for numerical calculations of spray droplet breakup,” Society of Automotive Engineers, Paper 872089 (1987).
36. N. Seifert, G. Betz, and W. Husinsky, “Droplet formation on metallic surfaces during low-fluence laser irradiation,” Appl. Surf. Sci. 103, 63–70 (1996). [CrossRef]
37. I. Egry, G. Lohoefer, and G. Jacobs, “Surface tension of liquid metals: Results from measurements on ground and in space,” Phys. Rev. Lett. 75, 4043–4046 (1955). [CrossRef]
38. A. Pearson, E. Cox, J. R. Blake, and S. R. Otto, “Bubble interactions near a free surface,” Eng. Anal. Bound. Elem. 28, 295–313 (2004). [CrossRef]
39. 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, 8225–8237 (2001). [CrossRef]
40. B. S. Mitchell, An Introduction to Materials Engineering and Science for Chemical and Materials Engineers (John Wiley & Sons, 2004).
41. H. Vogel, Gerthsen Physik (Springer, 1995).
42. M. S. Brown, N. T. Kattamis, and C. B. Arnold, “Time-resolved dynamics of laser-induced micro-jets from thin liquid films,” Microfluid. Nanofluid. 11, 199–207 (2011). [CrossRef]