1. Introduction

Artificially engineered nanotextured coating exhibiting unique optical, thermal, mechanical and chemical properties, which cannot be found in conventional natural materials, will become a milestone for future fabrication technologies of active and passive optical elements and devices, as well as medical and security sensors [1–3]. Among them metal nanotextured surfaces have been of great interest due to their ability to support a highly efficient, localized surface plasmon resonance and produce significantly enhanced and confined electromagnetic fields [4]. Such enhanced local electromagnetic fields have many applications such as biosensors, near-field scanning optical microscopy, fluorescence spectroscopy, enhanced Raman scattering, etc [5].

Ion- or electron-beam lithography are well-developed techniques used to fabricate predetermined nano-relief on the metal film surfaces [6,7]. However, these techniques become relatively time-consuming and extremely expensive, when the size of the nanotextured area reaches millimeter-scale level. Meanwhile, many important applications including routine biosensing measurements do not require utilization of highly regular nanorelief [8]. For such purposes, chemical synthesis methods [9,10] can be used to fabricate relatively large and statistically homogeneous nanotextured surfaces ensuring high uniformity of the electromagnetic field enhancement over their entire area. However, chemical contamination of such surfaces during the fabrication process can disimprove the characteristics of such functional surfaces requiring also the complex post-processing utilization [11]. Therefore, high-performance and cheap techniques for large-area surface nanotexturing are of great demand.

Nanostructuring of metal films surfaces by short (nano-) and ultrashort (femtosecond) laser pulses became widespread during the past two decades [12–27]. The impact of such pulses initiates ultrafast phase transition of the solid metal film into the liquid. Behavior of this liquid in a short time interval before recrystallization is defined by a sequence of well-established thermal and hydrodynamic processes providing formation of nanoscale structures under certain experimental conditions [12,13]. Despite considerably high productivity of such approach being compared with the ion- and electron-beam techniques, its application usually requires utilization of costly high-NA lenses to shape the laser beam onto the few-micron spot as well as expensive micropositioning platform to move this spot along the sample surface as well as to maintain the sample withing the focal depth of the lens. These requirements significantly increase the cost of the produced functional surfaces and limit performance of the laser-assisted nanotexturing techniques [14,15]. Accordingly, different interference lithography techniques utilizing interfering laser beams or special phase masks to form the raster containing periodically arranged diffraction-limited light spots are usually used for fast recording of large-area arrays [16,17]. However, fabrication of such regular phase masks is a difficult technological task [16,17].

As it was mentioned above, highly ordered arrays of nanostructures are rarely required in routine biosensing applications. Accordingly, for these purposes, the phase mask, which forms the output beam containing optical spot pattern, also can be irregular, while the main requirement needed to be satisfied for surface nanotexturing consists in a tight focusing of each optical spot as well as their statistically uniform arrangement on the sample surface. In this paper, we show for the first time that simple phase mask in the form of a diffusely transmitting object meets this requirement providing the formation of a speckle-modulated pattern at a certain distance from the mask surface with the lateral size of the the individual spots comparable to the scattered radiation wavelength. Despite some irregularities of the optical shape of each individual speckle, such speckle-modulated patterns can be utilized to pattern a thin gold film with closely packed micron-sized structures - nanojets, nanobumps and through holes - previous reported only for single-shot nanoablation with tightly focused laser beams. We have found and implemented an optimal energy-tunable regime to pattern mm-scale areas of Au film with the nanojets at an averaged recording density of 35∙10³ nanojets per square millimeter and an averaged recording speed of 4.5·10³ nanostructures per pulse. The fabricated nanotextured Au substrates were shown to demonstrate spatially uniform surface-enhanced fluorescence signal from the Rhodamine 6G organic dye with an averaged 5.3-fold enhancement factor being compared with non-treated Au surface.

2. Experimental

Large-scale nanopatterned areas were fabricated by irradiation of the thin Au film under ambient conditions by third-harmonic (355 nm) 7-ns pulses generated by the Nd:YAG laser with the maximal output pulse energy of 15 mJ. For these experiments we have used the 75-nm thick Au films deposited onto the optically smooth bulk glass substrates by e-beam evaporation procedure (Ferrotec EV M-6) at a pressure of 5·10⁻⁶ bar and an average speed ~8 Å/sec. To form a speckle-modulated pattern, laser pulses were passed through a diffusely transmitting object (diffusor), a 2-mm thick quartz plate with a smooth input side and polished output side, having the average RMS roughness of ~1 µm (Fig. 1(a)). RMS roughness slightly larger than the laser wavelength was chosen to provide the best contrast of the speckle pattern [28]. The pulse energy was varied by means of a tunable filter and controlled by a sensitive photodetector (J-10SI-HE Energy Sensor, Coherent EPM2000). For nanotexturing of 75-nm thick Au film averaged laser fluencies ranging between 60 – 250 mJ/cm² were used. The central 2.1-mm diameter part of the speckle pattern having almost constant average intensity distribution was isolated from the low-energy shoulders by a pinhole placed in the image plane (Fig. 1(a)). The sample was placed at the 2-mm distance from the diffusor. Such positioning of the sample will be shown to provide the minimal average size of the individual speckles equal to 1.57λ (λ – radiation wavelength). Fabricated nanopatterned areas were characterized using scanning electron microscope (SEM, Hitachi S3400N). To increase the recording density of the nanostructures, Au film was repeatedly exposed by a sequence of laser pulses with varying speckle-modulated lateral energy distribution. To do this, the diffusor was placed at the rotating stage to stochastically change the position of individual speckles minimizing the probability of undesirable overlaying. Nanostructure recording density on the nanopatterened film areas obtained under irradiation of the sample surface with varied number of applied pulses was estimated by counting number of nanojets, nanobumps and through holes on large-area SEM images having equivalent size for each N value.

 

Fig. 1 Experimental details. a) Schematic of the speckle-modulated laser nanotexturing procedure. Insets show the typical laser intensity distribution before the diffusor (I) and speckle-pattern formed in the sample plane after passing laser radiation through the diffusor (II). The scale bars in the insets are not equal for better displaying; b) Detection scheme for photoluminescence enhancement study.

Download Full Size | PDF

Dark-field optical images of the fabricated surfaces were recorded by means of a high-resolution optical microscope (Hirox KH7700) equipped with a 0.5 NA dry lens and a charged-couple device (CCD) camera. Illumination of the patterned areas was carried out by unpolarized white light from a calibrated stabilized halogen bulb (Ocean Optics, HL-2000) using a dark-field condenser. To measure the dark-field back-scattering spectra we used a highly sensitive spectrometer (Andor, Shamrock 303i) integrated into the home-build confocal optical microscope and equipped with a cooled CCD-camera (Newton 971). Each averaged dark-field back-scattered spectrum measured from the nanotextured areas was normalized on both the back-scattered spectrum from a smooth Au film surface of the same size and the spectrum of the white light source. To measure photoluminescence (PL) enhancement we have used the Rhodamine 6G (Rh6G) dye molecules. Fabricated nanotextured surfaces covered with the 10-nm thick overlayer of Rh6G molecules (Fig. 1(b)) were obliquely irradiated with a linearly-polarized radiation from a CW semiconductor laser source (Milles Griot, λ = 532 nm) at an incidence angle of ~80° with respect to the sample’s normal (Fig. 1(b)).

The pump radiation scattered from the nanostructures array was completely blocked by means of the suitable long wavelength pass filter (with a band edge ~545 nm) provided the pump wavelength absorption of about 10⁶. The resulting PL optical images were recorded by means of the CCD camera, while the PL spectra were measured using the abovementioned spectrometer. The enhancement coefficient EF^PL was estimated as a ratio of PL intensity measured from the nanotextured area of a certain size to the intensity collected from the Rh6G layer on the non-treated metal film area of the same size.

3. Results and discussion

The speckle-modulated intensity distribution is formed by passing the coherent laser radiation through the randomly inhomogeneous medium (see Experimental section for detail), a polished quartz plate, resulting from the interference of light scattered by the diffusor irregularities (Insets I and II in Fig. 1(a)). The average size of the individual speckle can be assessed by its full width at half maximum (FWHM) δ_cal = λR/D (D - beam diameter, R - distance between the image plane and the diffusor [28],). This simple formula indicates that at R≈D/2 the FWHM value is equal to ~λ/2, which is close to the diffraction limit. However, this formula holds true only at distances R exceeding D, which is a necessary condition for complete formation of the speckle pattern [28]. Therefore, to measure the actual minimum size of the individual speckles provided by our optical system we have performed additional experimental verification varying the distance between the image plane and the diffusor and repeatedly recording the speckle-modulated patterns formed by passing the 1.5-mm-diameter laser spot with λ = 405 nm through the diffusor (Inset II in Fig. 1(a)). Under such experimental conditions, the minimal experimentally measured FWHM value for the “stable” speckle pattern δ_exp ~1.57λ is observed at the distance R = 2 mm. Figure 2(a) shows SEM image of the 75-nm Au film surface irradiated by single speckle-modulated 15-mJ laser pulse at R = 2 mm. Average density of speckles N_sp in this case was found to be as high as 36·10³ speckle/mm², while the average energy in the individual speckle <E> estimated by taking into account diffusor absorption as well as the total number of speckles in the laser beam cut by the aperture was about 0.35 nJ. As can be seen from Fig. 2(a), irradiation of the metal film surface with the single speckle-modulated pulse results in formation of almost all common types of microstructures (Fig. 2(d) (6-9)) previously reported for single-pulse nanoablation experiments including nanojets [12–14,18–20], nanobumps [21,22], through holes [12,23] and nanocrowns [24–27]. The actual density of the recorded nanostructures N_str was found to be equal to 14·10³ structures/mm², which is almost three times lower than the speckle density N_sp. It is reasonable to assume that the N_sp/N_str ratio describes the probability that the pulse energy in the individual speckle exceeds the ablation threshold energy E_th for 75-nm thick Au film. In its turn, the threshold energy E_th can be estimated using the exponential law for the light intensity density in the speckle pattern [28]:

 

Fig. 2 Side-view SEM images (at an angle of 40°) of the 75-nm thick gold film surface irradiated by a single speckle-modulated laser pulse at pulse energy of 15 mJ (a) and 7.5 mJ (b). Insets shows correspondent color-coded energy distributions in the speckle-modulated pattern with their z-axis reflecting the thesholds required to produce through hole (I) and nanojet (II). Scale bars in the insets correspond to 1 µm; c) SEM images of the 75-nm Au film surface irradiated by twenty consequent speckle-modulated laser pulses at pulse energy of 7.5 mJ. Inset III demonstrates photograph of the large-scale nanotextured area of Au film (scale bar corresponds to 1 cm); d) SEM images of the main types of structures obtained by speckle-field nanopatterning (1-5) and by single-shot nanoablation with tightly focused beams (6-9) (Scale bars correspond to 1 µm); e) Dark-field optical image of the nanotextured area section marked with a blue square in Fig. 2a. White contours illustrate the actual size and shape of the though holes, while the white arrows indicate nanojets position.

Download Full Size | PDF

For the abovementioned values of N_sp, N_str and <E>, this estimation gives 0.28 nJ, which is in a satisfactorily good agreement with the values reported for ablation experiments performed with the tightly focused Gaussian beams [22] considering significant difference in the absorption coefficients of Au film at 355-nm and 532-nm wavelength [29]. Different types of the surface nanostructures (Fig. 2(d) (1-5)) appearing after irradiation of the Au film with the speckle-modulated laser pulse clearly indicate pronounced variation of the energy concentrated in each individual speckle as well as some distortions of its lateral energy distribution. The most intense speckles are responsible for appearance of the largest nanostructures - the through holes, crowns, and even the largest irregular-shaped structures (see Fig. 2(d) (3-5)). The latters apparently repeat the uneven lateral energy distribution in the individual speckles as well as can indicate few neiboring speckles merged together. However, for low-intense speckles which also have slightly irregular lateral energy distribution, the appeared nanostructures – nanobumps, nanojets and 1-μm-wide through holes surrounded with the resolidifying ring – demonstrate almost regular circle shape (Fig. 2(d) (1-2)), typical for single-shot nanoablation with the tightly focused Gaussian beams [22]. Apparently, this indicates that the shape and size of the heat-affected zone originating from the lateral heat diffusion from the small micron-size central area with the maximal intensity rather than the initial “optical” shape of the individual speckle governs the resulting lateral shape of the fabricated nanostructures. For low-intense speckles, smoothing of the heat-affected zone turns out to cause more pronounced effect.

 

Fig. 3 a) Dark-field optical image of the nanotextured Au film surface patterned with closely packed nanojets at N = 20 and E = 7.5 mJ; b) Normalized dark-field back-scattering spectra measured from three randomly chosen 20-μm-diameter areas marked with grey circles in Fig. 3a. Inset - the same spectra from three randomly chosen areas of a smaller size with their dark-field optical images presented in the insets I, II, III (scale bars correspond to 1 µm); c) Corresponding PL image of the Rh6G layer covering the nanotextured area; d) PL spectra measured from the 20-μm-diameter areas marked with grey circles in Fig. 3(c) (curves A,B,C). PL spectrtum from the Rh6G overlayer covering non-treated metal film surface of the same size (curve D) is presented for comparison. Inset shows enhancement factor EFPL versus pulse number N applied.

Download Full Size | PDF

Figure 2(e) demonstrates dark-field optical image of the small section of the nanotextured area marked with a blue square in Fig. 2(a). As one can see, bright scattering signal is observed only for areas containing relatively small-scale features, including edges of through holes, microbump tips and nanojets, while the inner area of the through holes occupying considerable part of the presented area (marked with white curves in Fig. 2(e)) appears black. Four brightest colored scattering spots (marked with white arrows in Fig. 2(e)) attributed to the nanojets are apparently associated with the excitation of strong localized plasmon resonances. This seems to indicate that the nanojets are the most favorable candidates for plasmon-mediated biosensing applications among other types of nanostructures appearing after speckle-modulated pulse irradiation.

Considering this finding, it is important to achieve the conditions for Au film nanopatterning, which provide both minimized number of the through holes and dense recording of the nanojets. As it follows from the data for single-pulse lens-focused ablation at λ = 532 nm [22], the energy required to produce through hole in the 75-nm thick Au film is approximately 3 times higher than the threshold energy E_th. Using Eq. (1), one can estimate that 2 times lower averaged energy of the individual speckle <E> provides probability of only 1% to overpass 3E_th energy level required to produce a through hole. Accordingly, at such <E> the recording density of through holes will not exceed ~300 units per mm². At such conditions, the probability of appearance of other nanostructure types, which require surpassing E_th energy level, becomes equal to 15%, according to the Eq. (1). In this respect, the estimated recording density of the low-energy nanostructures, including nanobumps and nanojets, becomes ~5·10³ units per mm².

Following the above-mentioned considerations, the nanotexturing of the 75-nm thick Au film surface by speckle-modulated light was further performed at twice lower (~7.5 mJ) input pulse energy. As one can see, under single-pulse irradiation of the Au film surface with the speckle-modulated pattern (Fig. 2(b)) the nanojets and nanobumps mainly appear on the film surface, while the density of the through holes turns out to be even less (~70 units per mm2) than the estimated value (see also inset I and inset II in Fig. 2(a) and Fig. 2(b)). This value can be even lower if we will further reduce the pulse energy, however, the recording density of the nanojets will be negligibly small in its turn. The experimentally measured recording density of the nanojets and nanobumps was found to be equal to ~4.5·10³ units per mm² which is close to the estimated one. Considering the fact that among all produced nanostructures two thirds are the microbumps, the single 7.5-mJ laser pulse produces as high as 1.5·10³ nanojets per mm². To increase the recording density of the nanojets we have performed multiple irradiation of the Au film surface with the speckle-modulated laser pulses rotating the diffusor to stochastically change the lateral distribution of the speckles. The surface of the 75-nm thick Au film after its irradiation with twenty speckle-modulated laser pulses is illustrated in Fig. 2(c), reflecting the gradual increase of the nanojet density with the number of pulses N applied. It also should be stressed that the appearance of the nanobumps on the film surface doesn't prevent further formation of nanojets on their place. The recording density of the nanojets demonstrates 23-fold increase at N = 20 reaching maximum value of 35∙10³ units per square millimeter. Some discrepancy between the number of applied pulses and the nanostructure density can be explained in terms of pulse-to-pulse stochastic changes of speckles density inside the counting area. Additional averaging procedure performed over 5 random surface areas of the same size demonstrates 19-fold averaged increase of the nanojet recording density. Further increase of N is limited by increasing probability of the structure overlapping, which leads to destruction of the nanojets, in its turn. Considering the maximal repetition rate of our laser system of 10 Hz as well as an isolation of the central part of the speckle pattern by the pinhole (see Experimental section for details), the fabrication of 3.53 mm² nanotextured area covered with densely packed nanojets takes only 2 sec. In our paper 100-mm² nanopatterned surface was fabricated by step-by-step movement of Au film relative to modifying speckle pattern for approximately 2 minutes (see inset III in Fig. 2(c)). The performance of the presented approach with an averaged recording speed reaching 4.5∙10³ nanostructures per pulse owing to multi-pulse exposure condition is smaller comparing to nanotexturing rates for interferometric lithography techniques [30], however this simple approach requires neither optical elements nor the complex optical alignment. Apparently utilization of the lasers operating at higher pulse repetition rates (up to MHz), higher output pulse energies as well as beam expanders can significantly improve the obtained indicators.

Figure 3(a) demonstrates typical dark-field scattering from the Au film surface patterned with closely packed (35∙10³ units per mm²) nanojets with the bright spots associated with plasmon-mediated scattering from the individual nanojets. Normalized dark-field back-scattering spectra (Fig. 3(b)) measured from three randomly chosen 20-μm-diameter areas (marked with semi-transparent grey circles and letters “A”, “B”, “C” in Fig. 3(a)) have almost identical shape repeating the spectrum measured from the overall 3.53-mm2 nanotextured area (data is not presented), while the similar spectra measured from the smaller 3-μm-diameter areas (see Insets in the Fig. 3(b)) can demonstrate significant differences depending on the accumulation area.

This feature clearly shows that although characteristic scattering signal from the few-micron areas can differ resulting from the stochastic concentration of the nanojets with similar plasmonic response, moderate-size nanopatterned areas of demonstrate statistically homogeneous plasmon-mediated scattering. As one can see, the averaged scattering spectrum (Fig. 3(b)) with a broad maximum near 570 nm overlaps considerable part of the visible spectral range. Additional overlapping with the blue part of the visible spectrum apparently can be obtained by pattering the surface of silver or aluminium (or even alloyed [31]) film in a similar way. It should be stressed that in this paper we have limited our study to visible spectral range, although considering relatively large height of the nanojets (up to 1.5 μm), strong plasmonic response in the near- and mid-IR spectral range can be expected [14].

The characteristic scattering spectrum of the nanopatterned Au film coincides with both absorption and emission bands of many organic dyes, quantum dots, etc, which apparently makes such substrates attractive for plasmon-mediated biosensing applications. To support this idea, we have covered the nanopatterned surface of Au film with 10-nm-thick overlayer of Rh6G molecules and measured its surface-enhanced PL signal. Figure 3(c) demonstrates the PL image of Rh6G layer covering the nanopatterned area produced by applying 20 consecutive laser pulses. As can be seen, the strongest photoluminescence signal is observed from the dye molecules located near the nanostructures, while smooth parts of the metal film appear black indicating substantially weaker signal. PL spectra (curves A, B, C in Fig. 3(d)) measured from three randomly chosen 20-μm-diameter areas (marked with semi-transparent grey circles and letters “A”, “B”, “C” in Fig. 3(c)) demonstrate averaged 5.3-fold enhancement of Rh6G fluorescence signal in comparison to the similar spectrum measured from the non-treated area of the same size (curve D in Fig. 3(d)). Finally, the dependence of the averaged PL enhancement factor EF^PL (for details see Experimental section) on the applied pulse number N (Inset in Fig. 3(d)) shows monotonous increase up to its maximal value of 5.3 times expectedly achieved at N = 20 (i.e. when the nanojet density is maximal) supporting the previous suggestion on the favorable utilization of the nanojets for biosensing substrates fabrication. At further increase of the applied pulse number (N>20), EF^PL value decreases apparently owing to overlapping of the high-energy speckles resulting in the nanojet destruction and appearance of the through holes.

4. Conclusion

To conclude, speckle-modulated patterns obtained by passing the laser nanosecond pulse through the simplest diffusive object were found suitable for very fast large-scale texturing of metal films with micron-sized structures - nanojets, nanobumps and submicron through holes, previous reported only for single-shot nanoablation with tightly focused beams. The optimal energy-tunable regime for pattering the mm-scale areas of Au films with nanojets with an averaged recording density of 35∙10³ nanojets per square millimeter and printing rate of 4.5·10³ nanostructures per pulse were found and implemented to fabricate mm-scale nanotextured areas. The fabricated nanotextured Au substrates were shown to demonstrate spatially uniform surface-enhanced fluorescence signal from the Rhodamine 6G organic dye with an averaged 5.3-fold enhancement factor being compared with non-treated Au surface.