The sensitivity of grating-coupled Surface Plasmon Polaritons (SPPs) on metallic surface has been exploited to investigate the correlation between ripples formation under ultrashort laser exposure and SPPs generation conditions. Systematic examination of coupling of single ultrashort laser pulse on gratings with appropriate periods ranging from 440 nm to 800 nm has been performed. Our approach reveals that a surface plasmon is excited only for an appropriate grating period, the nickel sample exhibits fine ripples pattern, evidencing the plasmonic nature of ripples generation. We propose a systematic investigation supported by a comprehensive study on the obtained modulation of such a coupling efficiency by means of a phenomenological Drude-Lorentz model which captures possible optical properties modification under femtosecond irradiation.
© 2011 Optical Society of America
Laser-Induced Periodical Surface Structures (LIPSS) were first observed by Birnbaum on semiconductor surfaces . Pioneering works to determine the physical processes responsible for periodic micro-and nanostructuring in laser interaction were started thirty years ago. They reveal that LIPSS depend on the irradiation conditions as well as the material properties [2–4]. Among them, the interference phenomena between the incident pulse and a secondary surface electromagnetic wave generated by scattering of laser radiation has been mostly accepted . More recently, ultrafast laser irradiated wide band-gap dielectrics, semi-conductors or metals exhibit periodic structures often much smaller than the laser wavelength λ, in contrast with structures formed under long-pulse irradiation with periods close to λ [6, 7]. When solids are exposed to ultrashort laser pulses, the exact mechanisms responsible for the ripples formation are still debated and various effects associated with short timescales have been proposed to explain the observed structures [8–13]. As ripples have been found to grow along the direction perpendicular to the laser polarization, even if high spatial frequency LIPSS parallel to the incident electric field have been reported, the hypothesis has been made that ripples can be attributed to the excitation of surface plasmon polaritons [10–12]. Surface plasmons are collective longitudinal oscillations of electrons propagating along a metal-dielectric interface at optical frequencies and it is well known that a surface plasmon is necessarily excited by a TM polarized wave [14, 15]. Due to the effective coupling of the incident electromagnetic radiation with the plasmon oscillation, a significant enhancement of the field in the vicinity of the structure can be produced, explaining the polarization dependence of ripples. In the initial stage of ripples formation, the plasmonic wave is supposed to be the main mechanism and the ripples period ΛLIPPS is related to the plasmon coupling theory and derived from previous long-pulse laser experiments :Eq.(1) reduces to ΛLIPSS ≃ λ. However, most experimental observations of ripples pattern show a spatial period ΛLIPSS smaller than λ [6,7,16], suggesting the implication of other processes as discussed above.
In experiments performed on clean surfaces, roughness provides the required momentum to excite surface plasmons, acting as a broadband random grating. In this work, to investigate the mechanism of surface plasmon generation with ultrashort pulse, we propose to provide an experimental evidence that plasmon resonance initiates ripples formation on a grating surface of wave vector kG = 2πN/ΛG, where N is an integer which represents the diffraction order and ΛG the period of the grating. A grating acts as a coupler between an incident plane wave and the surface plasmon field. An ultrafast laser beam impinging onto the surface at normal incidence should be coupled into a surface plasmon field for a very well-defined coupling grating period. Gratings diffract light in different orders, one of which being used to couple the free-space wave to the plasmon. The spatial harmonics along x of an incident beam impinging on a grating of period ΛG at a resonant angle θ satisfies the relation:Eq.(2) into Eq.(1), a maximal coupling is obtained for ΛLIPSS = ΛG = λ/η. In this frame, the purpose of this work is to investigate surface plasmon resonance conditions for ultrashort single pulse irradiation, validating this grating equation for a given set of grating periodicities.
The paper is organized as follows. The experimental section is devoted to the details of grating preparation on nickel surface, methodological approach and the irradiation procedure. The following section describes the results obtained by illuminating the different gratings, showing ripples formation for a well-defined period. Experimental evidence is presented that plasmons are definitively involved in the formation of LIPSS. Finally, to explain quantitatively why the coupling resonance period is lower than expected, the formation of LIPSS is discussed in the framework of the conventional surface plasmon theory and in terms of the substantial modifications of the optical properties under ultrashort exposure. A conclusion section summarizes the results.
2. Experimental procedure
2.1. Sample preparation and characterization
A Mach-Zehnder interference scheme was used to create a number of corrugation gratings of different period on different nickel plates coated with a thin photoresist (Shipley SPR505A) layer, as displayed on Fig. 1(a). The resist layer was deposited on a Ni-substrate by spin coating. The standard photoresist process chemistry was adapted to permit grating exposure on such high-reflectivity metallic substrate. The formation of a standing wave within the resist layer leads to interference fringes of poor contrast, therefore higher exposure dose is needed, which makes the fabrication of deep gratings on these reflective substrates difficult. In order to prevent the formation of a field antinode within the resist layer, the resist thickness should be rather thin and 140 nm thickness was chosen. A difficulty faced with a metal surface is that the electric field at the surface is close to zero since one has to use the TE polarization to create a high contrast interferogramme. The risk is therefore high that there still are some nanometers of unexposed and undeveloped photoresist at the groove bottom after development which will prevent the wet etching of the metal substrate in the grooves to take place. To get round this difficulty one first makes a preliminary uniform exposure. This first flood-exposure with non-structured light delivers small but non-zero energy dose everywhere within the resist layer. This dose offset is adjusted to remain just below the photomodification threshold of the resist. The sample is then placed in the interferogramme of a conventional Mach-Zehnder scheme fed by a 30 mW s-polarized He-Cd laser emitting at 442 nm. In the presence of the initial dose threshold all points of the resist layer located where an interference fringe is created get photomodified however small is the dose delivered by the interference fringe. Thus all grooves are open down to the nickel substrate. The physical transfer of the grating from the resist layer to the nickel substrate was made by wet etching using an acidic solution. Nickel requires an acid plus an oxidizer to etch properly. Diluted nitric acid (HNO3) contains both. A 10:1 dilution gives an etching speed of 1 nm/sec. A particular attention was paid to the verification of the absence of any contamination of the nickel surface due to sample preparation. The grating periods range from ΛG = 440 nm to ΛG = 800 nm with 10 nm increment. This range is supposed to cover all possibilities of plasmonic coupling. An example ΛG = 560 nm is shown on Fig. 1(b). All gratings have the same depth. A depth σ = 10 nm was chosen to be large enough to dominate the plasmon coupling effect of the surface roughness and small enough to just represent a seed for plasmon excitation [17, 18].
2.2. Ultrafast irradiation
The laser system used in the experiment was an amplified Ti:sapphire laser (Thales) with a pulse duration of 150 fs full width at half-maximum (FWHM) at a repetition rate of 1 kHz and a wavelength of 800 nm. The pulse rate used for samples irradiation was adjusted down to single pulse using a Pockels control unit cell. The experiments presented here have all been performed with a single pulse irradiation to avoid all accumulative phenomena which would erase the memory of plasmon resonance. The feedback mechanisms which would also change the initial periodicity by coupling with the incident light are suppressed by performing single pulse irradiation [19, 20]. The laser beam is focused normally, through an achromatic lens of 50.8 mm focal length, onto the sample that is vertically mounted on an X-Y-Z motorized translation stage. The horizontal polarization is controlled by a plate-polarizer. The dimension of the beam (at 1/e2 intensity) has been determined by the classical linear regression of the impact surface versus the energy logarithm, as expected with the rough assumption of a Gaussian beam . All the samples were placed at the image point of an aperture (diameter 2.8 mm) fixed in the set up. The on-target delivered power was finely controlled by density located in the optical path of the laser beam. Ablation of the samples was conducted in the air. The analysis of the modification of the surfaces under the action of femtosecond light pulses was performed using a scanning electron microscope (FEI, NovaNanoSEM). The topography of surface structures was studied in the Acoustic AC mode of an atomic force microscope AFM (Agilent 5500), which is a resonant mode, equivalent to tapping mode of Veeco’s AFM.
Samples with gratings periods in the range 440 to 800 nm have been irradiated with a single shot femtosecond laser beam, either TM or TE polarized as reported on Fig. 2. The small grating depth of 10 nm avoids enhancement of ripples formation resulting from a high roughness. The irradiation of the same sample of nickel without gratings does not lead to ripples formation whatever the polarization of the laser beam, at a laser fluence of 1.42 J/cm2, with a single shot exposure. Fig. 3 shows the SEM images of the samples with initial periods of 710, 760 and 790 nm, irradiated with a TM polarized laser beam at a fluence of 1.42 J/cm2. No fine ripples perpendicular to the polarization are observed for a grating period of 710 nm nor for a grating period of 790 nm, for which we would expect a classical surface plasmon resonance with a femtosecond laser wavelength of 800 nm as discussed in the following section. Interestingly, as reported on the SEM images of the 760 nm period grating, one observes clearly ripples formation for well-defined periods. The period of these fine ripples is close to 760 nm. This is observed only with the TM polarization, which evidences clearly the role of the polarization of the laser beam. This phenomenon gives the experimental evidence of grating assisted surface plasmon laser coupling .
In order to go beyond these first qualitative results, a systematic examination of all grating periods has been done. A quantification has been made through the measure of the density of ripples over the surface of the impact, measured as the ratio between the surface covered with ripples and the overall surface of the spot. These results are reported on Fig. 4(a,b) for a fluence of 1.42 J/cm2. For a TE polarization, no ripples are evidenced whatever the period of the gratings on the nickel surface. For TM polarization, a strong increase of the density of ripples is observed for periods of 760 nm, with zero values below 720 nm and above 770 nm. This sharp resonance will be discussed in the following section from theoretical findings on the observed period. These phenomena occur not only for this value of fluence (1.42 J/cm2), but also for a lower laser fluence of 0.97 J/cm2, as reported on Fig. 4(c,d). The same behavior is observed at this lower energy, attesting that in a well-defined energy range and under single laser shot exposure, grating assisted surface plasmon coupling is likely to play a precursor role in ripples formation under ultrashort laser irradiation.
In its simplest form, a surface plasmon polariton is an electromagnetic excitation that may exist at the interface of two media with dielectric constants of opposite signs, for instance nickel and air. The charge density wave is associated with bound TM-polarized electromagnetic wave at the metal dielectric interface. The amplitude of this field decays evanescently into each medium from the considered interface. This field can interfere with the incident laser beam and then can lead to a modulated energy deposition, forming ripples at the surface. To attempt an explanation of the observed behavior regarding the resonant coupling for 750 nm and 760 nm period gratings, we discuss hereafter the effects of ultrashort irradiation on the optical properties leading to a shift in the expected surface plasmon resonance.
4.1. Calculated plasmon wavelength for nickel
Applying the appropriate boundary conditions to the fields at the dielectric/metal interface results in the familiar expression for the plasmon wave vector . Provided ɛr < −ɛd, a condition satisfied at the air/Ni interface for 800 nm in the Palik’s data set , the plasmon wavelength λSP corresponds to the real part of kSP and can be written:
As the gratings have been designed in terms of aspect ratio to avoid any shift in the resonance period, calculations start from optical properties applicable for smooth and uncontaminated surfaces. The tabulated data from Refs. [22, 23] give for λ = 800 nm, ɛr = −13 and ɛi = 21.7, and a calculated surface plasmon wavelength λSP = 792 nm. If a Drude-type absorption dependence is found at low photon energy (λ > 20 μm) for nickel, the interband transitions dominate the spectral optical properties in the infrared range. Consequently, any treatment of the complex dielectric constant close to this region must involve splitting it into two parts, one part corresponding to intraband excitations described by the Drude model, and an interband part corresponding to resonant absorptions based on a Lorentz oscillator model, i.e. ɛ̃m = ɛ̃D + ɛ̃IB.
4.2. Parametric study of the Drude-Lorentz model
From the modeling standpoint, Eq.(3) requires a corrected dielectric function due to nickel excitation and should capture the essence of any ultrashort excitation effect compatible with the observed plasmon resonance shift from 792 nm to 750–760 nm. At a given frequency, the surface plasmon wavelength λSP can be tuned by the dielectric constant which require to evaluate material excitation. At our laser wavelength, the dominant changes of the optical properties is expected to originate from the modification of the electronic states. For nickel, the interband term ɛ̃IB is dominated by transitions from the narrow d bands to the Fermi surface states. Laser energy injection into the electron gas results in a spreading of the electron distribution around the Fermi energy, and thus drastically affects intraband and interband transitions occuring in the electron system.
The frequency-dependent complex dielectric constant ɛ̃m is described by the Drude-Lorentz model as :24]. According to these tabulated data, plasma frequencies have been set to , corresponding to two free electrons per atom.
In this calculation, as the two parameters ɛr and ɛi are varied, the surface plasmon wavelength is calculated and plotted as contour diagrams in Fig. 5(a). Fig. 5(b) shows the dependence of λSP with the plasma frequencies related to change in electron number densities for both intraband and interband contributions. The parametric study is discussed as a function of ωp but similar qualitative arguments can be deduced from changes of relative oscillator strengths.
The surface plasmon wavelength, and subsequently the ripples period, has already been supposed to decrease with the decreasing electron density for silicon [19, 25]. For Ni, a transition metal, the shift of the resonance from 792 nm to 750–760 nm can result from a complex change in electronic densities inside both the d-band and the s-band, due to ultrashort photoexcitation, as shown on Fig. 5(b). λSP = 760 nm solutions represented by the upper blue dashed curve on Fig. 5(b), are reachable for a d-band contribution lower than for room temperature case. λSP = 750 nm isocontour is represented by the lower green solid curve on the same figure, surrounding the available solutions. The distinction of the onset of interband transitions is not clear in Ni where d-bands overlap and have electron energies higher than the Fermi energy, as is apparent in the ab-initio calculations of electron band structure by Lin et al. . The overlapping of the high density of states d-band with the Fermi energy ensures that the 3d band electrons can be easily excited to the 4s band. Due to the photon energy greater than the interband transition threshold, electrons in the d-bands are promoted to empty states in the s/p-band near the Fermi level through direct interband transitions . As the temperature of the electron system increases, empty states will become available below the Fermi level due to Fermi smearing. The 4s band has a much smaller density of states as compared to the density of states at the Fermi level, leading to a shift of the chemical potential to higher energies, reducing the contribution of d-band electrons . The contribution of these electrons available to undergo a transition decreases, causing a decrease of the plasma frequency Ωp. Accordingly, the resultant surface plasmon wavelength get reduced and agrees with our measurements in Fig. 4. This analysis allows us to identify the range of surface plasmon resonance that can be effectively reached after an utrashort excitation. The decrease of the spatial period for the experimentally obtained surface ripples could be due to the decrease of the surface plasmon wavelength λSP with the decreasing plasma frequency in the solid.
In conclusion, we have demonstrated that laser-induced surface plasmons lead to the generation of fine ripples on a nickel surface undergoing femtosecond laser irradiation. Single pulse TM irradiations have been performed on a large range of grating periods with properly designed groove geometry to determine for which grating spatial frequency the incident beam and the surface plasmon wave vectors are matched. On the basis of this experimental procedure, our results reveal that ripples only appear for a well-defined grating period of 750–760 nm, showing that the coupling with surface plasmons is responsible for the ripples formation. Grating spacing which exhibits ripples differs from the one expected from dielectric constant at λ = 800 nm. A Drude-Lorentz model has been applied with various relative intraband and interband contributions. The dielectric constant of nickel is supposed to undergo a transiently change due to ultrashort excitation. The model shows that nickel surface irradiated by fs laser pulses can lead to a reduction of the surface plasmon wavelength compared to its room-temperature value for particular conditions of plasma frequencies. Therefore, this result is consistent with the data and provides information about the varying electronic density of the sample undergoing ultrashort irradiation.
References and links
1. M. Birnbaum, “Semiconductor surface damage produced by ruby lasers,” J. Appl. Phys. 36(11), 3688–3689 (1965). [CrossRef]
2. D. C. Emmony, R. P. Howson, and L. J. Willis, “Laser mirror damage in germanium at 10.6 μm,” Appl. Phys.Lett. 23(11), 598–600 (1973). [CrossRef]
3. Z. Guosheng, P. M. Fauchet, and A. E. Siegman, “Growth of spontaneous periodic surface structures on solids during laser illumination,” Phys. Rev. B 26(10), 5366–5381 (1982). [CrossRef]
4. J. E. Sipe, J. F. Young, J. S. Preston, and H. M. van Driel, “Laser-induced periodic surface structure,” Phys. Rev. B 27(2), 1141–1154 (1983). [CrossRef]
5. A. E. Siegman and P.M. Fauchet, “Stimulated woods anomalies on laser illuminated surfaces,” IEEE J. Quant. Elect. 22(8), 1384–1403 (1986). [CrossRef]
6. A. Borowiec and H. K. Haugen, “Subwavelength ripple formation on the surfaces of compound semiconductors irradiated with femtosecond laser pulses,” Appl. Phys. Lett. 82(25), 4462–4464 (2003). [CrossRef]
7. E. M. Hsu, T. H. R. Crawford, H. F. Tiedje, and H. K. Haugen, “Periodic xurface xtructures on gallium phosphide after irradiation with 150 fs 7 ns laser pulses at 800 nm,” Appl. Phys. Lett. 91(11), 111102 (2007). [CrossRef]
8. J. Reif, F. Costache, M. Henyk, and S. V. Pandelov, “Ripples revisited: non classical morphology at the bottom of femtosecond laser ablation craters in transient dielectrics,” Appl. Surf. Sci. 197–198, 891–895 (2002). [CrossRef]
10. M. Huang, F. Zhao, Y. Cheng, N. Xu, and Z. Xu, “Origin of laser-induced near-subwavelength ripples: interference between surface plasmons and incident laser,” ACS Nano 3(12), 4062–4070 (2009). [CrossRef] [PubMed]
11. J. Wang and C. Guo, “Formation of extraordinarily uniform periodic structures on metals induced by femtosecond laser pulses,” J. Appl. Phys. 100(2), 023511 (2006). [CrossRef]
12. J. Bonse, A. Rosenfeld, and J. Krger, “On the role of surface plasmon polaritons in the formation of laser-induced periodic surface structures upon irradiation of silicon by femtosecond laser pulses,” J. Appl. Phys. 106(10), 104910 (2009). [CrossRef]
13. M. Huang, F. Zhao, Y. Cheng, N. Xu, and Z. Xu, “Mechanisms of ultrafast laser-induced deep-subwavelength gratings on graphite and diamond,” Phys. Rev. B 79(12), 125436 (2009). [CrossRef]
14. A. D. Boardman, ed., Electromagnetic Surface Modes (Wiley, 1982).
15. H. Raether, “Surface plasmons on smooth and rough surfaces and on gratings,” in Springer Tracts in Modern Physics, (Springer, 1988), Vol. 111.
16. A. Y. Vorobyev and C. Guo, “Femtosecond laser-induced periodic surface structure formation on tungsten,” J. Appl. Phys. 104(6), 063523 (2008). [CrossRef]
17. T. Tomita, K. Kinoshita, S. Matsuo, and S. Hashimoto, “Effect of surface roughening on femtosecond laser-induced ripple structures,” Appl. Phys. Lett. 90(15), 153115 (2007). [CrossRef]
18. Y. Yang, J. Yang, L. Xue, and Y. Guo, “Surface patterning on periodicity of femtosecond laser-induced ripples,” Appl. Phys. Lett. 97(14), 141101 (2010). [CrossRef]
19. Y. Han and S. Qu, “The ripples and nanoparticles on silicon irradiated by femtosecond laser,” Chem. Phys. Lett. 495(4–6), 241–244 (2010). [CrossRef]
22. D. E. Gray, American Institute of Physics Handbook, 3rd. ed. (McGraw Hill, 1972).
23. E. Palik, Handbook of Optical Constants of Solids (Academic Press, 1985).
24. A. D. Rakic, A. B. Djurisic, J. M. Elazar, and M. L. Majewski, “Optical properties of metallic films for vertical-cavity optoelectronic devices,” Appl. Opt. 37, 5271–5283 (1998). [CrossRef]
25. J. Bonse, A. Rosenfeld, and J. Kruger, “Implications of transient changes of optical and surface properties of solids during femtosecond laser pulse irradiation to the formation of laser-induced periodic surface structures,” Appl. Surf. Sci. (to be published). [CrossRef]
26. Z. Lin and L. V. Zhigilei, “Electron-phonon coupling and electron heat capacity of metals under conditions of strong electron-phonon nonequilibrium,” Phys. Rev. B 77(7), 075133 (2008). [CrossRef]