Fabrication of uniformly dispersed nanoparticle-doped chalcogenide glass

Photodarkening, photobleaching, and photodoping are well-known light-induced phenomena in chalcogenide glasses.^1–3 The photodoping effect has drawn much attention recently due to its potential application in various areas, such as mid-infrared communications, holographic data storage, diffractive elements, and sensing devices.^4–6 However, in conventional photodoping procedures, which rely on thermal evaporation and sputtering to create a thin layer of metal on top of the chalcogenide film, the thickness of the doped layer is limited by the diffusion depth of the silver.⁷ Also the uniformity of the doped layer is difficult to control, thereby setting limitations in fabricating application-favorable bulk structures.

One possible method to resolve these difficulties in fabricating chalcogenide films is to employ a solution-based process.⁸ In particular, by incorporating nanoparticles into the deposited films, we can hope to achieve a uniform distribution of silver throughout the entire depth, rather than having silver concentrated only at the surface. Previous researchers have shown the ability to distribution quantum dots into chalcogenide systems using solution-based methods.^9,10 In addition, these approaches can readily realize large area or large thickness films, and the same solutions can also be adopted for other precision dispensing techniques such as mold casting, ink jet or laser direct write, allowing spatial control over the added material. Nevertheless, uniform doping of silver nanoparticles (Ag NPs) into a chalcogenide glass matrix without agglomeration remains challenging, due to the tendency of Ag NPs to aggregate.

In this paper, we present experimental results of fabricating uniformly dispersed nanoparticle-doped chalcogenide glass using laser ablation^11,12 and solution processing methods.^13,14 In the fabrication process, we first focus a pulsed laser beam onto the surface of a bulk metallic sample within an arsenic sulfide/propylamine solution and ablate the material. The ejecta expandby carrying out the nanoparticle generation steps directly in the solution of interest instead of ideal solvents such as water or ethanol, the laser ablation methods into the liquid solution, and condenses into in a suspension of silver nanoparticles. In contrast to prior studies which use ideal and/or pure solvents such as water or ethanol, or other organics,^15–17 we perform this process directly in the glass solution of interest. Ag-doped chalcogenide films are then fabricated by spin-coating the resulting solution. The prepared solution and films are analyzed using UV-Vis spectroscopy, dynamic light scattering (DLS), energy-dispersive X-ray spectroscopy (EDX), and transmission electron microscopy (TEM) to investigate the particle structure and distribution. The results demonstrate that this process can avoid the aggregation and additional processing steps associated with other nanoparticles generation techniques, such as wet chemical approaches,¹⁸ and a uniformly dispersed nanoparticle-doped chalcogenide glass has been fabricated.

Arsenic sulfide (As₂S₃) solution was prepared by grinding bulk As₂S₃ pieces into a fine powder and dissolved into n-propylamine solvent at a concentration of 0.8 mol/l. Dissolution was carried out inside a sealed glass chamber to prevent solvent evaporation. The dissolving process usually took more than three days and a magnetic stirrer was used to expedite this process. Exposure of solution to atmospheric moisture was kept to a minimum throughout preparation procedure since water can lead to precipitate formation.¹⁹ 

To obtain the nanoparticles in As₂S₃ solution, we used the experimental setup shown in Fig. 1. A silver target (>99.99% purity, thickness 5.0 mm) was placed at the bottom of a Teflon cell, which was filled with As₂S₃ solution afterwards. The silver target was covered by a layer of solution 4 mm deep. The cell was then transferred into an air-tight vacuum chamber. All these operations were carried out inside a nitrogen-filled glovebox, to avoid the influence of oxygen.

The chamber was then placed under the focus of a 30 ps, 1000 Hz, 1064 nm Q-Switched Nd:YAG laser beam, with maximum energy of 1 mJ/pulse. The Gaussian laser beam was focused, through the liquid layer, onto the target surface using a lens with a focal length of 50 mm. Since absorption by the arsenic sulfide solution was very small at the operating wavelength, the ablation beam passed through without significant loss. The position of the target relative to the laser beam was controlled by an X-Y-Z stage to maintain proper focus and to avoid deep hole drilling in the target. Ablation was performed for 10, 25, and 40 min, and we denote the obtained solutions as 10-, 25-, and 40-min solutions in this paper, respectively. Ablation was accompanied by the emission of plasma light from the surface, and bubbling was observed immediately upon irradiation. The solution gradually turned brownish-yellow as the concentration of Ag NPs increased. Finally, the chamber was moved back to the glovebox, where the solution was collected.

Immediately after synthesis of the solutions, UV-Vis absorption measurements of silver were conducted using an Ocean Optics HR4000 high-resolution spectrometer. In order to characterize morphology in thin films, TEM experiments were carried out on a Philips CM200 microscope operating at 200 kV. High resolution images were acquired using a Gatan Orius 200 camera. Samples for TEM characterization were prepared by pouring a droplet of the colloidal solution onto 400-mesh copper grids covered with a holey carbon film (from Ted Pella). The excess solvent was then allowed to evaporate.

These suspensions were then deposited onto silicon wafers for SEM and EDX analysis. Specifically, the solution was first pipetted onto a silicon substrate, and the substrate was immediately spun at 1000 rpm for 30 s. Resulting films were soft-baked under vacuum at 60 °C for 1 h to remove most of the solvent and then hard-baked at 180 °C to remove excess solvent and to further densify the glass.²⁰ Films prepared under these conditions typically were approximately 1 μm thick and ready for subsequent analysis.

As photos of the generated Ag NPs solution in Fig. 2 (inset) show the color of these solutions gets noticeably darker due to the increasing Ag NPs concentration resulting from higher number of laser shots. Absorption of these solutions is measured by UV-Vis spectroscopy using arsenic sulfide solution as a control. In the results shown in Fig. 2, different curves denote spectra from solutions experiencing different ablation duration. The existence of Ag NPs introduces surface plasmon resonance bands, but due to the absorption of the As₂S₃ solution, only tails of these resonance bands are observable. Moreover, these measurements disclose a redshift behavior of the absorption edge associated with the increasing concentration of Ag NPs, which agrees well with the literature.^21–23 

UV-Vis measurements indicate the existence of Ag NPs without revealing information about size of those Ag NPs. In order to determine the size distribution of the Ag NPs, DLS measurements are performed on the 40-min solution. The signal from pure As₂S₃ solution is also acquired for comparison. As shown in Fig. 3, background from As₂S₃ peaks around 8 nm, which originate from the dissolved As₂S₃ units.²⁰ The peak from Ag NPs solutions is centered around 14 nm, which reveals the average size of the majority of nanoparticles. The DLS measurements are also performed on 10- and 25-min solutions, and the same distribution of Ag NPs are obtained, revealing the size is independent with ablation time. These values are in good agreement with the literature,^24,25 where nanoparticles average size of 5–25 nm are generated in ideal solutions like water.

Measurements on the solution phase demonstrate that the size of the nanoparticles is independent of ablation time, while the concentration of particles is clearly rising with increasing ablation time. However, the ultimate test of the particle distribution is the fabrication of a film with a uniform dispersion of nanoparticles. Fig. 4 shows EDX results from spin-coated films using the 40-min solution. The relative weight of silver is measured to be 1.3%. Armed with this value, the concentration of As₂S₃ solution (2 g/10 ml), and the average particle size from DLS measurements, the generation rate of Ag NPs is calculated to be 1.52 × 10⁸ nanoparticles formed per pulse, which is equivalent to 7.8 mg/h in terms of the particles weight.

To determine the structure and distribution of the Ag NPs, we use high-resolution TEM, shown in Fig. 5. The TEM images (Figs. 5(a) and 5(b)) demonstrate the existence of Ag NPs and indicate their uniform distribution. Fig. 5(a) is the dark field TEM image, in which only the diffraction signal from crystalline structures is collected. Since arsenic sulfide is in an amorphous state, the bright spots in Fig. 5(a) are the silver nanoparticles. The diffraction pattern from the Ag NPs (Fig. 5(c), inset) exhibits hexagonal symmetry that can be attributed to the basal plane of the hexagonal phase. This agrees well with the literature where it has been shown the hexagonal phase of silver stabilizes only in the nanocrystalline form for particles less than 30 nm.²⁶ Fig. 5(d) shows a high resolution TEM image of silver nanoparticles which demonstrates a high degree of crystalline order. The spacing between the lattice fringes is measured to be 2.5 Å.

The TEM image from Fig. 5(a) is transformed into a high contrast image, as shown in Fig. 5(b), which more clearly shows the spacing distribution of nanoparticles. We quantify uniformity and dispersion of the distribution using the nearest neighbor index (NNI), a standard method in spatial analysis that is used to determine the degree of spatial dispersion in a population.²⁷ In general, if the distribution of the points is clustered together, the average distance between nearest neighbors will be shorter than if the particles are scattered throughout the sample. The NNI is defined as the ratio of the average inter-point distance between nearest neighbors $d¯$ to the expected value of the average inter-point distance if the sample were randomly dispersed $E(d¯)$⁠.^28,29 The value of the NNI can range between the theoretical extremes of 0 (where all points are at the same location) and 2.1419 (where points have a perfectly uniform distribution).²⁸ The equations for these parameters are given by

For the data in Fig. 5(b), n = 265 is the total number of particles, obtained using the “particles analysis” utilities from Image J, and A = 2156 × 2156 nm² is the area of the studied region. d_i is the distance from ith particle to its nearest neighbor, determined by measuring all neighboring distances and taking the minimum. $d¯$ is calculated to be 66.68 nm using ImageJ. $E(d¯)$ is determined to be 66.21 nm. The NNI is then readily evaluated to be 1.01, which reveals a random dispersion.

Finally, to test if the calculated NNI is statistically different from that of a random process it is necessary to calculate the standard normal deviate of the distribution (Z) using $Z=(d¯−E(d¯))/σd¯$⁠, where $σd¯=0.0683A/n2$ is the standard deviation of $d¯$⁠.²⁹ If the value of Z is within [−1.96, 1.96] the distribution of points is considered to be random at the 95% confidence level.²⁹ Z is calculated to be 0.218. The value of NNI and Z evidences a uniform spatial dispersion of the particles.

In summary, we report the fabrication of uniformly dispersed Ag NP/chalcogenide glass heterogeneous material prepared through a combined laser- and solution-based process. We are able to obtain uniform distribution of nanoparticles in both the solution and thin film phases, which is evidenced by the high resolution TEM measurement and the NNI analysis. The clustering or agglomeration that is typically associated with solution based methods in nanoparticles fabrication are avoided through this approach. We believe the process is applicable to other metals and other chalcogenide glass solutions.^30,31 These materials have a great potential for applications in diffractive elements and sensing devices, particularly in cases where thick films uniformly doped by Ag NPs are crucial. Furthermore, they have ability to be photo-responsive, and could be used for direct writing, as well as recording of optical information, such as holographic data storage.

We thank Joseph Buttacci for assistance with experimental setup construction. We thank Gerald Poirier and Yao-wen Yeh for assistance characterizing samples. We gratefully acknowledge support by NSF through the MIRTHE Center (Grant No. EEC-0540832), as well as support from the Princeton-University of São Paulo partnership program. J.M.P.A. acknowledges the São Paulo research foundation for the financial support. N.Y. acknowledges the partial support of the NSF-MRSEC program through the Princeton Center for Complex Materials (DMR-0819860).