1. Introduction

Nonlinear optical (NLO) materials have been studied extensively for their wide applications from broadband optical limiting devices to high-speed data-processing communication systems. With optical nonlinearities, scientists and engineers can manipulate properties of light, including frequency, phase, and power to a large extent. Many researchers have intensively studied the silicon (Si) based optical nonlinear systems, which demonstrate most of the NLO mechanisms, including nonlinear absorption, refraction and scattering [1–4]. Si is thus considered as a cheap alternative for fabricating future optoelectronic and photovoltaic devices. To fully characterize its nonlinear performance, Si in different matrices was intensively studied [5]. In these studies, stability of the Si material dispersion inside a matrix is always an issue to be concerned, since it directly relates to the effective operation time of the system. Dispersions of Si nanomaterials in aqueous and organic solutions are mostly unstable if untreated [6–8], therefore stabilizers, including surfactants are required if liquid matrices are used. A solid state matrix is hence advantageous for a nanoparticles system since it enables high stability of the dispersion of nanoparticles, and therefore enables a higher upper limit of nanoparticle concentration without any aggregation inside the matrix. Furthermore, a solid matrix lowers the risk of toxicity posed by using liquid organic matrices. More importantly, comparing to its fluid counterpart, a solid based system is much more convenient to be converted into practical devices, which is easier to be maintained and operated.

Among all the solid matrixes, gel wax is a favorable choice due to its transparency, chemical stability under laser irradiation and low toxicity. Gel wax (GW) is a kind of organic materials made up of mineral oils and polystyrene [9], it is transparent in visible and near IR spectra. Both the two main components display chemical stability under laser irradiations [10, 11], thus the GW can serve as an adequate matrix material for laser devices. Moreover, under the high incident energy of the laser, the GW is easy to be melted due to its relatively low melting point. Furthermore, the wax can gradually solidify back to its original state when the laser is turned off, and thus regains its functions. Tuning optical nonlinearities of NLO materials is also crucial for many applications, which has been reported in the previous researches [5, 12–14, 19]. Wide tunability of an NLO material, for instance from optical limiting effect to saturable absorption effect under constant incident laser fluence, would provide higher possibility of manipulation to various applications. In this work, a new composite of nanoparticles dispersed into a gel wax matrix is developed. Its optical nonlinearity properties are investigated using the open aperture 𝑍-scan method at different concentrations. Experiments demonstrated the capability to manipulate the optical nonlinearity of the system, switching from optical limiting to saturable absorption was achieved by changing the concentration of silicon nanoparticles (SiNPs) inside the matrix. The system is also self-recovering regarding to its optical nonlinearity after excess exposure to the laser irradiation.

2. Experimental methods

Pulsed laser ablation was applied to synthesize the silicon nanoparticles [15]. In each experiment, a 1 cm × 1 cm clean silicon wafer was positioned at the center of a 10 mL beaker, immersed inside 1 mL water. The beaker was then put on the processing stage, where the incident laser beam would be focused. The experimental setup and the synthesis mechanisms were discussed in the previous works [16, 17]. The Nd:YAG fiber laser (IDI Laser Service) operating at the 1064 nm wavelength and 10 ns pulse width was employed for the laser ablation. The laser beam was programmed to scan over a 1 mm × 1 mm area for 300 times with a rectangular scanning pattern, and the average duration for one scan is 0.52 seconds. The laser fluence was kept at 8.9 J/cm² throughout the experiment at the focal point. The entire process took about 156 seconds. 10 rounds of experiments were conducted, and aqueous dispersion of Si nanoparticles was collected. By weighing the mass of the Si wafer before and after the experiment, the concentration of the aqueous dispersion of Si nanoparticles was calculated.

To mix Si nanoparticles with the GW matrix, a certain amount of GW was first melted in a beaker on a hotplate at 180°C, and then cooled down. A chosen amount of aqueous nanoparticles dispersion was added into the solidified wax, with heating at 95°C until all water was evaporated, and SiNPs were deposited into the solid wax. Then the wax was melted at 170°C on the hotplate while stirring with a magnetic stirrer at 2000 rpm for 5 minutes, and then ultra-sonicated for 1 minute. This step was repeated for 15 times, and the mixture was left to cool down in ambient temperature until the wax matrix was solidified. The nanoparticles were held in their positions. The sample was settled for 48 hours and the stability was proved to be sustained. Three samples at concentrations of 60 mM, 30 mM and 20 mM were prepared.

The absorption spectrum of the GW was characterized by a UV-3600 Shimadzu UV-VIS-NIR spectrophotometer. From the obtained absorption spectrum, around the wavelength of the Z-scan testing laser (800 nm), the normalized absorbance of the GW is only about −0.5%. This means that the GW matrix has no significant absorption of the light energy used for the Z-scan measurement. For optical nonlinearity characterization, Z-scan measurements [18] were performed using the Spectra-Physics femtosecond laser operating at 800 nm wavelength and 0.1 J/cm² laser fluence (It was observed that when the laser fluence is below 0.05 J/cm², there was no noticeable nonlinear effect), with a pulse duration of around 100 fs and a repetition rate of 1 kHz. The spot size was ~34 µm (diameter). The sample is contained in the optical glass cell with the optical light path of 1 mm whose absorption is very small (less than 0.5%) at the 800 nm. The same setup parameters were used to study optical nonlinearities of graphene oxide [19]. In order to reduce the experimental error, all Z-scan experimental measured data were averaged over 1 second of measurement.

3. Results and discussion

3.1 Morphology and size distribution of Si nanoparticles

The size distribution and morphology of the synthesized Si nanoparticles were characterized by SEM images. A drop of the dispersion was dipped onto a polished side of a Si substrate. Then the Jeol JSM7001F SEM was used for sample observation. The morphology of SiNPs is presented in Fig. 1(a). The synthesized nanoparticles possess spherical shapes, SiNPs with different sizes are also observed from the SEM image. To characterize the size distribution, randomly selected nanoparticles were measured and the distribution is shown in Fig. 1(b). The high energy of the laser beam during laser ablation accounts for the wide range of size distribution. During the ablation process, energy is rapidly transferred from incident photons to the silicon substrate surface, causing evaporation of Si materials into energetic species and thus generating plasma that expands outwards. The energetic plasma is then cooled down with energy absorbed by the surrounding liquid. The interactions promote nucleation process to form SiNPs of various sizes, until a suspension of the nanoparticles is formed [16].

 

Fig. 1 (a) SEM image of synthesized Si nanoparticles ((SiNPs of 6 mM concentration), Scale bar = 1 µm. (b) Size distribution of synthesized SiNPs, calculated from a randomly selected sample of nanoparticles.

Download Full Size | PDF

3.2 Z-scan results of the matrix-SiNPs system

To characterize the nonlinear performance, Z-scan technique was adopted. Figure 2 illustrates Z-scan plots for the pure wax matrix and SiNPs-GW composite samples at different concentrations. The normalized transmittance is plotted against the sample position along Z-axis. Normalized transmittance T_m(Z) at position Z is defined by the transmittance at that position divided by the transmittance at the two end points far away from the focus point (Z₀) [20]:

 

Fig. 2 Calculated normalized transmittance from Z-scan measurement results, for samples at 20mM, 30 mM and 60 mM of SiNPs respectively, and for a pure gel wax sample.

Download Full Size | PDF

All plotted results are under the same laser fluence and incident light wavelength. Each data point presented in the graph is an averaged value from 3 sets of Z-scan results for the same sample. A symmetric valley centered at 0 point on the Z-axis in the curve indicates a decrease in transmittance of the sample as it approaches to the focal point, where the laser fluence reaches the highest value. Meanwhile, the nonlinear absorbance of the sample increases. The symmetric peak in the curve then implies an increased transmittance and a decreased absorbance of the corresponding sample as it approaches to the focal point.

The set of square data points (black) in Fig. 2 indicates the Z-scan measurement results of the pure GW matrix, the deviation of T_m is within the range of ± 1.66%, which is small as compared to the three samples’ results. The wax matrix serves as the supporting role in the system with its low optical absorption property. It does not demonstrate significant optical nonlinearities in Z-scan measurements. It can be observed that the nonlinear performance of the hybrid system largely depends on the concentration of the SiNPs. When the concentration of Si nanoparticles increases from 20 mM to 30 mM, the peak transmittance T_peak decreases from 0.924 to 0.868. Such a decrease in transmittance with laser fluence is attributed to optical limiting effect [21]. This nonlinear behavior is due to the nonlinear scattering of the laser irradiation on the surfaces of nanoparticles. Within this concentration range, additional SiNPs in the solid composite system impose more nonlinear scattering of the incident light, and thus the optical limiting effect is enhanced when the concentration increases.

The mechanism of this nonlinear scattering (NS) can be explained as a well-established theory. NS occurs when the incident laser power is high, a rapid energy transfer from photons to the SiNPs triggers the increase of the temperature locally, and the surrounding wax matrix also absorbs the heat and melts. Spaces are given for the gel wax to expand from the surfaces of SiNPs, which results in the formation of localized micro-bubbles. The tiny bubbles cause the incident light to scatter much stronger, which results in lower transmitted power intensity. An illustration of this mechanism is shown in Fig. 3. With this mechanism, the system behaves like an optical limiter, which decreases the optical transmittance with incident laser fluence.

 

Fig. 3 Schematic diagram for the nonlinear scattering (NS) mechanism. (a) At a low fluence, no significant influence on output light; (b) at a high fluence, melting of matrix generates localized micro-bubbles around each nanoparticle, incident light is nonlinearly scattered and output fluence is reduced.

Download Full Size | PDF

Figure 2 shows another very interesting experimental result. When the concentration of SiNPs further increases to 60 mM, the nonlinear performance changes from optical limiting to saturable absorption. The direction of the curve peak is switched from downward to upward with a peak transmittance of 1.187. The upward peak is the saturable absorption (SA) property of silicon materials [5]. The physics behind can be explained as follow: at a concentration higher than a threshold, the smaller inter-particle spacing of nanoparticles gives higher possibility for scattered laser light to interact strongly with neighboring SiNPs, such that the saturable absorption of Si has enough time to take larger effect. Therefore, the SA effect dominates the Z-scan process.

The mechanism of SA happens in one single laser pulse and can be better understood using the following physics. Silicon’s saturable absorption property largely depends on its carrier dynamics. With the 3D nanoparticles’ structure, the relaxation process of photo-excited carriers in a semiconductor is divided into three regimes: coherent, cooling and recombination regimes [22]. In the coherent regime (≤200 fs), photo-excitation is caused by the laser pulse. In the relaxation process, recombination regime is time-dominating, and is to a large extent related to the carrier lifetime of the Si materials. The laser beam used for the Z-scan measurement has a pulse duration of 100 fs. The SiNPs, generated from the Si wafer at a doping concentration of 2.8 × 10¹⁵ cm⁻³, can be assumed to have a carrier lifetime over tens of microseconds [23]. Thus the carrier lifetime of the SiNPs is significantly longer than the laser pulse duration. When the free carriers in Si nanoparticles are mostly excited in the initial stage of the laser pulse, the rest of laser energy arrive before the excited free carriers fall back to ground states. Thus the free carriers’ ability to absorb photon energy greatly reduces. The transmittance of SiNPs is thus increased, and this is attributed to the SA mechanism.

These results imply that the solid state SiNPs composite system has tunable optical nonlinearities, and the key parameter is the concentration of Si nanoparticles inside the gel wax matrix. Hybrid mechanisms of NS and SA co-exist in this solid system. Simply changing the concentration of SiNPs not only can adjust the extent of effect brought by the dominating optical nonlinearity mechanism, but also switch the dominating mechanism from one to the other. Due to the tunable optical characteristics of the system, optical limiting and saturable absorption devices can be fabricated with the exact same set of component materials. Moreover, with stable SiNPs dispersion inside the low-melting-point solid matrix, fabrication of NLO nanostructures for optoelectronic and photovoltaic devices would be possibly compatible with 3D printing techniques. However, one extra point must be examined is the consistency of optical nonlinearity performance, this is important to investigate the repetition rate effect of the system’s operation.

3.3 Self-recovering ability of the optical nonlinearity

A further set of experiments were carried out to explore the consistency of the optical nonlinearity of the system. The 30 mM sample was placed on the experimental setup for one round of Z-scan test (denoted as Test A). By keeping the laser irradiation for 4 more minutes, following by the second round Z-scan test (Test B), then the aperture was closed to block the incident laser beam for about 30 s, and the third round of Z-scan test (Test C) was conducted. The results of these three rounds of tests are presented in Fig. 4. In Test A, the data agrees with the previous results regarding to the shape and the peak value (0.872) of the curve. The downward peak of the transmittance curve at the focal point indicates optical limiting effect. In Test B, however, after excess exposure to incident laser, the downward peak switches to an upward peak at the transmission peak of 1.103. Then when the incident laser irradiation is removed for a while, the follow-up Test C gives a downward peak curve again, it has the same trend with Test A, but the magnitude of ΔT becomes smaller, at the transmission peak of 0.915.

 

Fig. 4 Normalized transmittance of (A) the synthesized sample at the concentration of 30 mM; (B) after the sample was exposed to incident laser for 4 more minutes; (C) after closing the aperture to prevent the sample from exposure to the laser irradiation for 30 s, then re-opened the aperture and conducted Z-scan Test C.

Download Full Size | PDF

This set of results demonstrates the self-recovering ability of the nano-composite system, which can help maintain the consistency of the system’s optical nonlinearity performance. The reversed result in Test B is due to the excessive exposure to the laser irradiation, which results in the excited states of SiNPs suspended in the matrix gradually being saturated. Thus the SA mechanism dominated the NS effect. The wide full width at half maxima (FWHM) of this curve implies a full bleaching of ground state electrons, a small increase in incident laser fluence causes significant increment in normalized transmittance. Blocking the laser allows the excited electrons in SiNPs to transfer absorbed energies to the surrounding matrix and fall back to the ground states, then the system regains its original function as the optical limiter. Test C justifies the self-recovering ability of the system. From the experimental results, this system has limited active duration for its optical limiting performance in each operation, while this change in functionality is reversible once the light energy input is removed for a sufficiently long period.

4. Conclusions

Silicon nanoparticles were synthesized by the laser ablation and the dispersion inside a solid state gel wax matrix was prepared at different concentrations. The Z-scan results depict the switching of the system’s optical nonlinearity from optical limiting effect to saturable absorption effect. Experiments also show that the system is able to preserve its performance consistency with the self-repairing property. This research reveals the feasibility of optical nonlinear nanomaterial system based on a solid state matrix with low melting point, and the possibility to tune the optical nonlinearity by simply manipulating the concentration of SiNPs in the matrix, which can be easily achieved by adding SiNPs into the system.