1. Introduction

Glass composite materials containing silver nanoparticles (NPs) are attractive for various photonic applications owing to their linear and nonlinear optical properties, mainly governed by the surface plasmon resonance (SPR) of the NPs. The spectral position and shape of the absorption band due to SPR strongly depend on the size, shape, and spatial distribution of the NPs. Submitting a silver-exchanged soda-lime glass to pulsed laser exposure leads to the formation of small silver NPs (few nanometers in diameter) only, and cluster fragmentation has been reported in the case of prolonged exposure [1]. Processing a composite glass, which contains thermally formed silver NPs, by pulsed laser exposure appears more flexible to tailor the optical properties of the NPs. A wide range of experimental results have been reported, such as the change of the size, size distribution, and shapes of the silver NPs, depending on the intensity, pulse duration, repetition rate, and wavelength of the pulsed laser [2–7]. Large-aspect-ratio prolate or oblate NPs have been formed from initially spherical NPs by pulsed laser exposure, and their optical dichroism has been used for the production of micropolarizers [6–8]. Fewer studies report on the processing of soda-lime composite glasses containing silver NPs or silver-exchanged soda-lime glasses by use of a continuous wave (cw) laser. The formation of gratings [9] or microlenses [10,11] has been reported in composite glasses by cw laser exposure at 488 nm. This wavelength is almost resonant with the absorption band because of the SPR of the silver NPs and is therefore strongly absorbed.

We have previously reported about the growth of silver NPs of controlled diameter by cw laser exposure at 244 nm, coupled to simultaneous heating at 450°C [12], of a silver-exchanged glass. In this case, no annealing of the glass is required to form silver NPs before exposure, since the silver ions are reduced by the laser exposure before merging into NPs. The laser-induced reduction of silver ions in silica-based glasses has been previously explained by the formation of color centers in the glass during the laser exposure, which simultaneously leads to the release of electrons and holes in the glass [13]. These free electrons are then trapped by the silver ions, which are thus reduced into neutral silver Ag⁰. In the work presented here, a silver-exchanged glass has been irradiated at room temperature with the same cw laser at 244 nm but with a much higher UV power density than that previously used [12]. UV-written lines have been inscribed by focusing the laser beam at the glass surface and moving the glass sample with a constant speed. Highly concentrated large NPs, of a diameter around 100 nm, are formed in the border regions of the UV-written lines, spread over several micrometers. Small NPs are formed within the first hundreds of microseconds of the laser exposure and then experience a very large temperature rise, leading to their diffusion a few micrometers away from the laser spot, where they coalesce into larger NPs. Simultaneously, a surface relief is observed, explained by the formation of liquid drops at the glass surface, characteristics of a marked temperature rise. For the highest power density available, we have observed the formation or ripples of a period of about 150 nm at the glass surface, together with the disappearance of the large NPs that are replaced by nanoholes. It is worth noting that ripples have not been reported in the literature in silica-based glasses.

2. Experimental procedure

Silver-exchanged soda-lime glasses have been obtained by immersing 1-mm-thick commercial soda-lime glass slides for 20 min at 340°C into a molten salt bath of molar concentration 0.1% AgNO₃ in NaNO₃. The refractive index of the exchanged zone, measured using the M-lines technique, varies from 1.61 to 1.52 across an 8 µm depth. A 244 nm laser was focused on the glass surface with a power density of about 100 kW/cm² to inscribe lines at a width of about 12 µm. Absorption and reflection spectra were measured with a spatial resolution of 8 µm by use of a confocal spectrometer, used in the transmission and reflection modes, respectively. Optical microscopy (OM), scanning electron microscopy (SEM), atomic force microscopy (AFM), and profilometry have been used to characterize the NPs and nanoholes.

3. Results and discussion

The effects of the laser exposure performed with a sample displacement speed of 4000 µm/s can be observed in the inset of Fig. 1(a) , which shows an optical microscope picture, taken in the transmission mode of a UV-written line. The central part of the UV-written line exhibits a yellow color, surrounded by two darker bands at a width of about 7 µm. According to the coloration of these different areas, the absorption is higher in the outer regions than in the central part of the UV-written line. Pulsed-laser exposure at 248 nm of the same silver-exchanged glass, performed for comparison, leads to a homogeneous coloration of the irradiated area, which gives evidence of the different character of the effects induced by the two kinds of lasers.

 

Fig. 1 (a) Optical density measured within (1) the central part of the irradiated line, (2) the border region of the irradiated line. The inset shows an optical microscope picture of a UV-written line taken in the transmission mode. (b) Reflectivity spectrum measured within the border region (3). The inset shows an optical microscope picture of a UV-written line taken in the reflection mode.

Download Full Size | PDF

The dark color in the border regions becomes increasingly clear away from the edge of the UV-written line, owing to a decreasing concentration of the NPs. The absorption spectra recorded separately from these two regions (curves (1) and (2) in Fig. 1(a)), despite a slight recovery due to the 8µm spatial resolution of the measurements, are ascribed to the SPR of silver NPs [13]. Curve (1) can be correctly fitted by use of Mie’s theory [14] with an average diameter of about 5 nm. The asymmetric shape of the absorption band measured in the border region and its high intensity are due to a large distribution of sizes of highly concentrated NPs [12]. Figure 1(b) shows the reflectance spectrum measured from the border region. Since enhanced reflectivity in a composite glass is noticeable in case of high filling factor only [15], the almost flat and high reflectivity between 500 and 850 nm confirms the high filling factor of the NPs in the border regions. The 36% reflectivity is much higher than that reported in silver-composite glass containing densely packed percolated silver NPs formed by thermal poling [15], which gives evidence of the efficiency of the cw UV laser exposure to grow highly concentrated silver NPs. On the contrary, the reflectance spectrum measured from the center of the UV-written lines (not shown in Fig. 1(b)) is identical to that of a non-exposed glass, owing to the low concentration of the NPs in this region.

SEM analysis is well suited to further investigate the sizes and concentrations of the NPs in the UV-written line and in the border regions. Figure 2(a) shows an SEM picture, centered on the edge of a UV-written line, shown by a transverse black line: only few NPs, which appear as white spots in the SEM picture, are observed in the irradiated part, while highly concentrated NPs of almost spherical shape are observed in the very near border region.

 

Fig. 2 SEM images: (a) showing nanoparticles near the border of the line (shown by a transverse black line) and the lack of nanoparticles in the irradiated part and (b) zoom within the area containing nanoparticles (1 µm away from the edge of the line). Circled areas show discontinuities due to coalescence.

Download Full Size | PDF

The lack of NPs within the UV-written line indicates either that the remaining absorption in this region (Fig. 1(a)) is due to silver NPs buried beneath the glass surface or that the NPs are too small to be clearly observed by SEM. The border regions contain densely packed spherical NPs, whose distribution of sizes contain three main classes of diameters (Fig. 2(b)): large NPs (mainly around 100 nm in diameter) are surrounded by NPs of diameter 40 to 50 nm and smaller ones of roughly 20 nm in diameter. This self-regular arrangement between NPs is retained over a width of 4 µm from the edge of the line, but few discontinuities can be observed in the circled areas in Fig. 2(b), due to the coalescence of two close large NPs into a larger asymmetric one (up to 200 nm in diameter). This arrangement is progressively replaced by less concentrated larger NPs of more chaotic shapes when looking further away from the edge of the UV-written line.

At this point of the paper, OM and SEM have been used to fully describe the spatial repartition of the silver NPs, but the corresponding photoinduced mechanisms are yet not fully elucidated and their explanation requires further experiments. Short laser pulses have been generated with a 30 slots blade optical chopper wheel rotating at frequency up to 4 kHz, whose all slots except one were covered with nontransparent black tape, while moving the sample at high speed (15mm/s) to avoid spatial recovery between the laser spots. This setup allows the generating of short pulses of temporal from 250 µs up to 4 ms, depending on the rotating frequency of the chopper wheel. The corresponding laser spots have then been examined by OM and profilometry. A yellowish coloration, homogeneously distributed over the whole laser spot surface and due to the SPR of the silver NPs, is observed in case of exposure times shorter than 1 ms. The corresponding absorption band is similar to that reported in curve (1) in Fig. 1(a). No surface relief is observed in the laser spots. Therefore, rather small NPs formed during the early stages of the laser irradiation near the glass surface. For exposure times longer than 1 ms, a surface relief, whose height and width increase with the irradiation time, is observed (curve (a) in Fig. 3 ).

 

Fig. 3 Surface profile, recorded by profilometry, after laser exposure of a (a) silver-exchanged glass and of a (b) pure soda-lime glass.

Download Full Size | PDF

Simultaneously, a ring that contains large and highly concentrated NPs, similar to those shown in Fig. 2, is observed about 10 µm away from the center of the spot. The width and height of the bump raised above the glass surface grow with the exposure time up to saturation for exposures times longer than 10 ms. The formation of bumps above the glass surface is explained by the formation of glass liquids drops, in which the initially formed small silver NPs move toward the periphery, following the temperature distribution and coalescing into larger NPs. The formation of glass drops at the surface requires obviously a very large temperature rise during the laser exposure. It is worth notincg that an almost similar explanation has been proposed for soda-lime glasses containing thermally formed silver NPs submitted to intense cw laser exposure at 488 nm [10,11]. The marked temperature rise under cw laser exposure is mainly governed by the absorption of the glass at the laser wavelength. In [10,11], the 488 nm laser wavelength, which is resonant with the absorption due to SPR of the NPs, is tightly focused on the glass surface, giving rise to a very large power density of roughly 2000 kW/cm². In our case, the very large absorption of the glass at 244 nm comes from the glassy matrix and overcomes the lack of power density (100 kW/cm²). It is worth noting that we have observed a surface relief of the same height in a virgin glass submitted to the same UV irradiation (curve (b) in Fig. 3), which gives evidence of the same order of magnitude of the maximum temperature in both virgin and silver-exchanged glasses. The smaller bumps located on each side of the glass liquid drops (curve (a) in Fig. 3) are due to the highly concentrated large NPs located in the ring mentioned above. Several approaches have been used to estimate the maximum temperature reached by the glass, based on the model developed in [10,11], which requires the measurement of the absorption coefficient of the glass at the laser wavelength. However, commercial soda-lime glasses are known to be completely nontransparent in the UV range so that the absorption coefficient at 244 nm is not given by glass manufacturers. Polishing of the glass slide down to a thickness of 135 µm, measured by optical microscopy on the side of the sample, has been achieved. Absorption at 244 nm has then been measured with a spectrophotometer, whose highest measurable optical density is 4, but saturation occurred. Direct measure of the laser transmitted power during intense UV exposure (180 mW incident power) through the polished sample with a detector placed behind the glass slide was also unsuccessful. Therefore, modeling of the temperature rise is not yet possible.

A different behavior of the glass has been observed in case of UV irradiation performed with an increased laser power density (120 kW/cm²), which is expected to enhance the maximum temperature reached by the glass, according to the above-mentioned results. The SEM pictures displayed in Figs. 4(a) and 4(b) reveal the formation of ripples (or laser-induced periodical surface structures) at the glass surface, located on a disk of few micrometers width surrounding the laser spot.

 

Fig. 4 SEM images showing the formation of ripples on the edge of the laser spot: (a) ripples can be observed together with few large NPS formed further away from the laser spot and (b) zoom inside the region containing the ripples.

Download Full Size | PDF

Ripples have been observed at the surface of various materials, such as semiconductors [16], polymers [17], or more recently metals [18] or metallic glasses [19], after intense cw or pulsed laser exposure. To our knowledge, ripples have not been reported in silica-based glasses. The mechanisms explaining the formation of ripples are still in debate, and various effects have been proposed, but it is commonly admitted that the formation of ripples requires a high temperature increase, which can lead to the melting, sublimation, and evaporation of the surfaces, depending on the materials [19]. Consequently, the formation of ripples gives evidence of the really high maximum temperature, since evaporation is most probably involved. AFM characterization of the ripples has revealed a period of roughly 150 nm, in good agreement with SEM observations. More details and explanations will be given in another paper. We have observed the disappearance of the most of the highly concentrated large NPs located in the ring few micrometers away from the center of the laser spot (Fig. 5(a) ).

 

Fig. 5 (a) SEM image showing the bleaching of the NPs (white spots) to the benefit of nanoholes (dark spots) and (b) AFM height scan of the nanoholes.

Download Full Size | PDF

Only few NPs remain, and instead, dark spots are observed. The AFM height scan displayed in Fig. 5(b) confirms the formation of nanoholes. The dissolution of silver NPs embedded in a glass matrix has been demonstrated either by thermal annealing at around 550 °C [13] or by thermal poling through electric field associated dissolution [20–22]. In our case, their dissolution is clearly linked with the ripples formation, according to the simultaneity of these two phenomena. We propose thus that the maximum temperature has not yet been reached, while the larger nanoparticles have been formed during the first microsecond of the laser exposure. Therefore, these NPs experience an increased temperature, which leads to their rapid evaporation, and explains that the NPs leave their imprints at the glass surface. In the case of annealing performed in a furnace, the dissolution is rather slow and does not result in the formation of nanoholes [13]. It is worth noting that we have observed the presence of some silver dust (identified by X-ray analysis) at the glass surface after the UV exposure, which indicates that evaporation of the metal occurred.

4. Conclusion

Glasses doped by silver ions have been submitted to an intense cw laser exposure at 244 nm to promote the formation of silver NPs. Large and highly concentrated silver NPs are formed in the border regions surrounding the laser-treated area, and their growth is explained by a two-steps mechanism, implying a large laser-induced temperature rise, owing to the high absorption coefficient of the glass at 244 nm. We demonstrate that performing the exposure with a higher laser power density leads to the disappearance of these large NPs, which leave their imprints at the glass surface through the formation of nanoholes. The bleaching of the NPs is explained by their vaporization, as evidenced by the simultaneous observation of ripples at the glass surface. The 150 nm spatial period of the ripples is shorter than those usually reported after pulsed laser exposure. Our technique might provide a potential method to patterned glass surfaces with ripples and nanoholes. We are also currently working to grow large gold nanoparticles in gold-exchanged soda-lime glasses [23] by the same technique, in order to broaden the potential applications of such composite glasses.