Controlled formation of gold nanoparticles with tunable plasmonic properties in tellurite glass

doi:10.21203/rs.3.rs-3118468/v1

Download PDF

Article

Controlled formation of gold nanoparticles with tunable plasmonic properties in tellurite glass

https://doi.org/10.21203/rs.3.rs-3118468/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 07 Dec, 2023

Read the published version in Light: Science & Applications →

You are reading this latest preprint version

Silicate glasses with metallic nanoparticles (NPs) have been of intense interest in art, science and technology as the plasmonic properties of the metallic NPs equip glass with light modulation capability. The so-called striking technique has enabled precise control of the in-situ formation of metallic NPs in silicate glasses for applications from coloured glasses to photonic devices. Over the past two decades, there has been a large amount of work to adapt the striking technique to form gold or silver NPs in tellurite glasses that exhibit the unique combination of easy fabrication, low phonon energy, wide transmission window and high solubility of luminescent rare earth ions. Nevertheless, the control of the in-situ formation of metal NPs and hence their plasmonic features in tellurite glasses has remained insufficient for photonic applications. Here, we first uncover the challenges of the traditional striking technique to create gold NPs in tellurite glass. Then, we demonstrate precise control of the size and concentration of gold NPs in tellurite glass by developing new approaches to both steps of the striking technique: a controlled gold crucible corrosion technique to incorporate gold ions in tellurite glass and a novel powder reheating technique to subsequently transform the gold ions to gold NPs. Using the Mie theory, the size, size distribution and concentration of the gold NPs formed in tellurite glass were determined from the plasmonic properties of the NPs. This fundamental research provides guidance to designing and manipulating the plasmonic properties in tellurite glass for photonics research and applications.

Physical sciences/Optics and photonics/Optical materials and structures/Nanoparticles

Physical sciences/Physics/Optical physics/Nanophotonics and plasmonics

The fascination of silicate glasses containing metallic nanoparticles (NPs, e.g., Au or Ag) in decoration and art has lasted for centuries (if not millennia). Because of their applications as optical elements (such as polarisers) ¹ as well as the roles they are expected to play in applications such as optoelectronics, photonics, sensing, electrochemistry and biomedicine ² they imply a technological relevance and thus huge scientific interest. The intriguing functionalities of silicate glasses with gold or silver NPs are intimately related to their unique optical properties originating from the localized surface plasmon resonance (LSPR). It is now well understood that the LSPR features – the spectral shape, position and intensity – are determined by the metallic NP properties (composition, size, shape and number density) as well as the refractive index of their surrounding medium ³. Therefore, tailoring LSPR performance requires being able to control the metallic NP properties in a glass matrix through glass composition and manufacturing processes.

The LSPR features of Au and Ag NPs in silicate glasses can be well controlled by the well-established traditional striking technique (i.e., in-situ formation of metallic NPs in glass matrix): (1) The glass batch is doped with Au and Ag compounds and in most cases also with polyvalent ions (typically SnO/SnO₂). (2) Melting the glass batch dissolves the Au and Ag compounds as colourless Au⁺ and Ag⁺ ions in the glass matrix – as proved by many indirect and direct experimental results in traditional oxide glasses such as silicate, germanate and borate as well as in geological silicate melts ^4–6. Quenching the melt yields the colourless precursor glass. (3) In the final colour striking step, reheating the precursor glass above the glass transition temperature induces the dissolved Au and Ag ions to be chemically reduced by the co-doped polyvalent ions to form metallic atoms that grow further into Au and Ag NPs that endow the entire glass with the desired colour and LSPR features ⁷.

Over the past two decades, the striking technique has been applied to form Au and Ag NPs in a range of other glass systems to add LSPR based functionality to these glass types ^8–15. Specifically, there has been a large amount of work to introduce Au or Ag NPs into tellurite glasses as they have unique combination of properties: easy fabrication (i.e., low melting temperature and high crystallization stability), relatively good physical and chemical durability, low phonon energy (~ 800 cm^− 1), wide transmission window (0.4–5 µm) and high solubility of luminescent rare earth (RE) ions for hosting and guiding most of the RE fluorescence ¹⁶. The tellurite glass type has been one of the most popular host matrix for studying the plasmonic effect of Ag or Au NPs on enhancing the fluorescence of RE dopants ^17–21. However, most of the published research only provided TEM images of the Au/Ag doped tellurite glasses showing the formation of crystalline NPs ^17–19, with no further evidence confirming the elemental nature of the observed crystalline NPs. In addition, distinct LSPR features in tellurite glasses have been demonstrated in only a few reports ^20,21. The ability to tune the location and intensity of the LSPR band remains a challenge. However, this ability is critical for optimizing the fluorescence enhancement of RE dopants ²².

A decade ago, unintentional formation of Au NPs in tellurite glass was discovered when developing the so-called melt doping technique to incorporate diamond nanocrystals in tellurite glasses.²³ In this technique, a batch of glass raw materials without NPs was melted in a gold crucible at a sufficiently high temperature (called melting temperature) to obtain a homogenized glass melt. Then, the temperature was lowered to a point (called doping temperature) that yields a glass viscosity which is sufficiently low to disperse the added diamond nanocrystals (and to cast the glass melt) while being sufficiently high to suppress the dissolution of the diamond nanocrystals in the glass melt ²³. Surprisingly, Au NPs were found in such a diamond nanocrystal doped tellurite glass, as evidenced by a distinct LSPR extinction band in the spectral range of 500–800 nm, causing the glass to have blue colour in transmission. The Au NP formation was proposed to be as follows ²⁴: 1) a small amount of gold was dissolved into the hot tellurite glass melt in the form of gold ions at the glass melting temperature as a result of gold crucible corrosion; 2) the added diamond nanocrystals acted as reducing agents for the reduction of the gold ions into gold atoms, which further nucleated and grew into Au NPs. The formation of Au NPs via doping diamond nanocrystals into tellurite glass melts was found to be a stochastic process and hence unsuitable for controlled Au NP formation because of the complex chemistry between the diamond nanocrystals and the tellurite glass.

Here, we demonstrate the controlled in-situ formation of Au NPs in tellurite glass via new ways to control incorporation of Au⁺ ions into the glass melt and their chemical reduction to Au⁰ atoms. The incorporation of Au⁺ ions at targeted number densities is achieved through control of the gold crucible corrosion, while the controlled reduction of the Au⁺ ions to Au⁰ atoms is realised through a novel ‘striking’ technique utilising powder reheating. Compared with the traditional gold doping technique of adding gold salt to the glass batch, our technique of controlled gold crucible corrosion during tellurite glass melting enables higher concentration of Au⁺ ions in tellurite glass. The traditional Au NP formation technique of using reducing agent (SnO/SnO₂) for the formation of Au atoms is shown to be not applicable to tellurite glasses due to the formation of reduced tellurium species (Te⁰ and/or Te²⁻). By contrast, our reheating technique of glass powder free of reducing agents prevents reduced tellurium species formation while enabling the formation of Au⁰ atoms. By combining the controlled gold crucible corrosion technique and the novel powder reheating based ‘striking’ technique, the concentration, size and size distribution of Au NPs in tellurite glass can be tailored. This opens up, for the first time, tuning of both LSPR position and intensity in tellurite glass.

2.1. Incorporation of Au⁺ ions into tellurite glass

As aforementioned in the Introduction, the first necessary condition for the formation of Au NPs is the incorporation of Au⁺ ions into the glass melt. For silicate glasses, this is achieved by melting a glass batch with added gold salt. We tested the viability of this traditional technique for TZN glass. In addition, we developed a novel technique of Au⁺ ion doping via control of the corrosion of the gold crucible in which the tellurite glass is melted.

2.1.1. Traditional technique: Adding gold salt to the glass batch

For testing the traditional approach of introducing Au⁺ ions into the glass matrix by adding gold salt such as HAuCl₄ to the glass batch, we used 10 ppm Au (in the form of HAuCl₄) and melted the glass batch at 750 ^oC for 1h in a silica crucible. The cast glass sample S1 is colourless (Fig. 1a), but a sub-mm sized Au metal particle is found in the residual glass at the bottom of the silica crucible (Fig. 1b), which indicates not all Au from the gold salt (HAuCl₄) was incorporated as Au⁺ ions in the glass. Indeed, solution ICM-MS analysis confirms that the cast glass sample S1 contains only ~ 3 ppm of Au.

2.1.2. Controlled gold crucible corrosion technique

As described above, previous work found that melting of tellurite glass in a gold crucible led to the dissolution of Au into the tellurite glass melt in the form of Au⁺ ions. In this study, the Au dissolution for TZN glasses melted in a gold crucible is quantified, and the influence of the melting temperature on the amount of dissolved Au in TZN is determined. Three glass samples (S2, S3 and S4) were made by melting undoped TZN batch in a gold crucible at 750 ^oC, 800 ^oC, and 850 ^oC (all for 1h). These samples contain ~ 6, 19, and 75 ppm of Au, respectively, as measured by solution ICP-MS analysis (Table 1). This result indicates that the dissolution of Au via gold crucible corrosion increases with increasing glass melting temperature.

As a control experiment, we made a glass sample S5 by melting an undoped TZN batch in a silica crucible at 750 ^oC for 1h. As expected, no Au above the detection limit of 0.1 ppm was measured via ICP-MS in this glass (Table 1).

2.1.3. Comparison of the mechanisms of both gold doping techniques

All four Au containing glass samples, i.e., the sample S1 made from Au-doped batch melted in a silica crucible as well as the three samples S2, S3 and S4 made from undoped batch melted in gold crucibles, are colourless (Fig. 1c). This indicates that all Au dissolved in the glasses via traditional doping technique of using gold salt (S1) or novel doping technique of gold crucible corrosion (S2, S3, S4) is incorporated in ionic form in the glasses. This agrees with silicate glass doped with Au via the traditional technique, where both Mossbauer and XAS (XANES/EXAFS) spectroscopy proved that Au exists as Au⁺ ions in colourless precursor silicate glass ^5,25.

For the same batch melting temperature and time (750 ^oC, 60min), the traditional doping technique using gold salt leads to only ~ 3ppm Au in the glass (S1), whereas our doping technique of controlled gold crucible corrosion leads to a higher content of ~ 6ppm Au (S2). The amount of dissolved Au via gold crucible corrosion was further increased using higher melting temperature; up to 75ppm Au for 850 ^oC (S4). These different Au solubility of the gold doping techniques are attributed to two effects.

One effect is the thermal decomposition of gold salt upon heating to metallic gold. Pure HAuCl₄ thermally decomposes to pure Au metal at 320 ^oC ²⁶. For a silicate glass batch doped with 200 ppm Au as aqua regia solution, thermal decomposition during heating up to the melting temperature (1400–1450 ^oC) led to the formation of Au NPs in the heated batch ⁷. Based on these findings, we hypothesize that during heating up of the TZN glass batch from room temperature to a critical temperature (below 750 ^oC), the HAuCl₄ in the glass batch is thermally decomposed to Au metal, resulting in the formation of Au particles by diffusion of the Au⁰ atoms along the surface of the raw material particles of the glass batch and in the viscous melt. The thermal decomposition of ionic gold compounds during batch heating makes the oxidation of Au⁰ at the melting temperature also the critical step to dissolve Au in ionic form for the traditional doping technique.

The second effect is that the Au oxidation is highly dependent on melting temperature and amount of oxygen. For silicate melts, oxidation of Au⁰ in the melt to Au⁺ was found to increase with increasing melt temperature (at fixed oxygen partial pressure) and/or increasing oxygen partial pressure (at fixed temperature) of the melt ^4,27. In addition, the use of oxidizing nitrate raw materials instead of carbonate in the glass batch is known to facilitate oxidation of Au⁰ in silicate glass ⁷. According to these results, we propose that for the traditional doping technique, at temperatures higher than the critical temperature, the oxygen dissolved in the viscous glass leads to the oxidation of the Au metal particles, which were formed below the critical temperature from the added gold salt, to Au⁺ ions, which become dissolved into the glass matrix. The observation of micron sized Au particle in the TZN glass at the bottom of the crucible reveals limited oxidation during glass batch melting, which is attributed to the formation of large Au particles before reaching the critical temperature as well as the low amount of oxygen present in the TZN glass melt.

For the novel gold doping technique via gold crucible corrosion, the dissolution of Au in the melt also occurs via the oxidation of Au⁰. However, both techniques differ in the amount of oxygen available for the oxidation of Au⁰. For the traditional technique of doping gold salts into the batch, oxygen dissolved in the melt is available for oxidation, but this amount is limited due to low solubility of oxygen in a glass melt. By contrast, for the gold crucible corrosion technique, we propose that the Au oxidation proceeds at the gold crucible-air-TZN melt interface as there is unlimited amount of oxygen available in the air atmosphere above the glass melt. Due to convection and swirling and the low viscosity of the TZN glass melt, the Au⁺ ions formed at this interface are readily dispersed in the glass melt. The larger amount of oxygen available for Au oxidation for the novel doping technique explains the higher amount of dissolved Au in the glass compared to the traditional technique. The increase of gold crucible corrosion and hence higher amount of dissolved Au with increasing melt temperature can be explained by three possible reasons; (i) reduced viscosity and hence enhanced diffusion, (ii) increased oxidation reaction rate and (iii) enhanced Au⁺ ion solubility.

The different mechanisms of Au dissolution for the traditional and novel gold doping techniques are summarized in the schematic shown in Fig. 2. For the traditional gold doping technique, the TZN glass batch contains Au⁺ ions via addition of HAuCl₄. With increasing batch temperature, first the Au⁺ ions are thermally decomposed to Au metal particles, which are then oxidised above the critical temperature. The amount of Au⁺ ions in the final glass melt is limited by the size of the Au microparticles formed at lower temperatures and the limited amount of oxygen in the glass melt. For the novel gold doping technique, the TZN glass batch does not contain Au⁺ ions. At high temperatures of the liquid glass melt, the unlimited amount of oxygen in the air atmosphere above the glass melt leads to dissolution of Au from the crucible as Au⁺ ions in the glass melt, whereby the amount of dissolved Au is controlled via melting temperature (and hence melt viscosity, oxidation rate and/or Au⁺ ion solubility) and melting time.

2.2. Transformation of dissolved Au⁺ ions to Au NPs

The second necessary condition for the formation of Au NPs is the reduction of the Au⁺ ions in the glass to Au atoms, which nucleate and grow to Au NPs. For silicate glasses, this is achieved by heating the bulk starting glass doped with both Au⁺ ions and polyvalent ions at elevated temperature. We tested the viability of this traditional technique for TZN glass. In addition, we developed a novel technique of Au NP formation via heating powderized starting glass doped only with Au⁺ ions. Note that for both techniques, Au⁺ ions were incorporated into the glass melts using the novel gold doping technique of gold crucible corrosion.

2.2.1. Traditional technique: reducing agent Sn²⁺

The traditional Au NP formation technique relies on the reducing power of polyvalent elements (PV) that are co-doped into the starting glass together with Au⁺ ions for the controlled formation of Au NPs via reheating the starting glass:

PV^x+ + yAu⁺ → PV^(x+y)+ + yAu⁰

where the polyvalent elements are normally selected from Sn, Sb, Pb, etc. that have more than one valence state in glass ⁷.

Sn is the most widely used polyvalent element for creating Au NPs in glass, whereby both SnO and SnO₂ have been doped into glass batches at the level of 0.02-2 wt% to generate Sn²⁺ at appropriate concentrations in the glass melt ^7,28,29. Here, both SnO and SnO₂ have also been used to investigate the Au NP formation efficacy and effect on the optical properties of TZN glass. 2 wt% of SnO or SnO₂ was employed to ensure enough reducing species Sn²⁺. The TZN glass batch doped with 2 wt% SnO or SnO₂ was melted in a gold crucible at 750 ^oC for 1h under air atmosphere, equivalently to the starting glass sample S2.

As shown in Fig. 3a, both the SnO doped glass sample (S6) and the SnO₂ doped glass sample (S7) show a distinct colouration, dark brown and light brown colour, respectively. The extinction spectra of both samples exhibit a pronounced redshift of the UV edge (more for S6 in agreement with the darker colour) compared to the co-dopant-free sample (S2) after melting under the same conditions. Similar black/brown coloration and high loss in the visible was observed in an earlier study on TZNL (TeO₂-ZnO-Na₂O-La₂O₃) based tellurite glass made by adding diamond nanocrystals, which acted as reducing agent, in high concentrations (> 20 ppm) to the glass melt at 700 ^oC ²⁴, as well as in a multicomponent tellurite glass film prepared under oxygen deficient atmosphere ³⁰. Such high loss in the visible range was ascribed to the formation of reduced tellurium species (Te⁰ and/or Te²⁻). Thus, the brown colour and high extinction magnitude of the S6 and S7 glass samples in this work are also assigned to the formation of reduced tellurium species during the glass batch melting process by the addition of SnO/SnO₂.

2.2.2. New technique: powder reheating based ‘striking’

The glass powder doping technique has been used to incorporate NPs in phosphate glasses ¹. The new technique of Au NP formation in this work was discovered when investigating the suitability of the glass powder doping technique for homogeneous nanocrystal dispersion in tellurite glass. For this study, we selected ruby (Al₂O₃:Cr³⁺) nanocrystals as they are chemically non-reducing to avoid the formation of Au NPs as found for chemically reducing diamond nanocrystals. First, bulk tellurite glass (S8) was made by melting the batch (without ruby nanocrystals) in a gold crucible. Secondly, the bulk glass was powderized into fine powder which was subsequently well mixed with ruby nanocrystals to obtain a well homogenized mixture of glass powder and ruby nanocrystals. Finally, this mixture was reheated at low temperature and short time (570 ^oC for 10 min) to avoid ruby nanocrystal dissolution, and then cast into a pre-heated mould. Surprisingly, the ruby nanocrystal doped reheated glass sample (p8R-1) shows a similar blue colour as well as a prominent extinction band in the spectral range of 500–800 nm (Fig. 3b) as diamond nanocrystal doped TZN glass ²⁴, indicating formation of Au NPs in the ruby nanocrystal doped TZN glass despite the non-reducing nature of ruby nanocrystals. This triggered the hypothesis that the formation of Au NPs in the ruby nanocrystal doped glass was not related to the added ruby nanocrystals but to the fabrication process and the glass itself.

To validate this hypothesis, the powder reheating process was repeated at the same heating conditions (570 ^oC for 10 min) but without mixing ruby nanocrystals into the glass powder. An almost identical homogeneously blue coloured glass sample (p2R-1) was produced as shown in Fig. 3b, indicating Au NPs were successfully created in TZN glass without the use of any reducing agents, i.e., only by reheating the TZN glass powders containing Au⁺ ions. This new process is referred to as powder reheating technique.

The existence of Au NPs in the glass samples made by powder reheating was proven via electron microscopy (EM) for p2R-1 and also p2R-0, which were made to study the impact of doping temperature and time (Table 2). The two glass samples were made using reheating for 10 min at 510 ^oC (p2R-0) and 570 ^oC (p2R-1). For sample p2R-1, SEM imaging in backscattering mode shows a number of bright particles, an example of a particle of 61 nm diameter is shown in Fig. 4a. The EDS spectrum for the bright particle (Fig. 4b) exhibits an intense gold signal, while the spectrum from the grey background only shows the elements O, Na, Zn, and Te, which constitute the glass matrix. For p2R-0, STEM imaging shows a bright particle of 36 nm diameter, and corresponding EDS elemental maps reveals that this particle is composed of Au, whereas the surrounding matrix contains the glass elements O, Na, Zn, and Te but no Au (Fig. 4c). These EM results confirm that under both reheating conditions Au NPs are formed from the ground glass powder, without additional reducing components.

2.2.3. New technique: powder reheating - control experiments

To further validate the hypothesis that the powder reheating process itself causes the Au NP formation, we conducted a few control experiments. We commenced with making a control glass sample where a bulk glass piece instead of powderized glass was used for reheating. Specifically, control sample b2R-1 was made by repeating the reheating process used for p2R-1 (showing Au NPs) using the same starting glass (S2) and the same reheating conditions (570 ^oC for 10 min in air atmosphere). As expected, the bulk reheated glass sample b2R-1 has the same colourless appearance and absence of any extinction peak in the visible range as the starting glass sample S2 (Fig. 3b). The lack of Au NPs in bulk reheated glass reveals that Au NPs cannot be effectively created in the entire glass volume by reheating Au⁺ ion containing glass in bulk form. Considering that at 570 ^oC the low viscosity of TZN glass enables it to flow easily, the absence of Au NPs in b2R-1 cannot be due to the inefficient diffusion of Au atoms (if they exist) to form Au NPs. That is, the lack of Au NPs in b2R-1 is related to the lack of reducing power for the reduction of Au⁺ ions during reheating the bulk piece of S2 glass. This behaviour agrees with previous reports noting the absence of Au NPs formation in Au⁺ ion doped silicate glasses (that contain no reducing agents) after reheating the Au⁺ ion doped starting glasses (in bulk form) in (non-reducing) air atmosphere (e.g., by S. Stookey ⁷, J. Qiu, et al., ³¹ and more recently by M. Kracker, et al. ³² and C. Thieme, et al. ³³). All these results indicate that bulk glass itself does not exhibit the ability to chemically reduce Au⁺ ions to Au atoms when reheated in non-reducing atmosphere.

By contrast, the act of milling Au⁺ ion doped TZN starting glass seems to provide reducing power for the efficient reduction of Au⁺ ions to Au⁰ atoms (either during the milling or the reheating process), which further nucleate and grow into Au NPs during reheating of the glass powders. As the melting of TZN glass in a gold crucible leads to the incorporation of Au⁺ ions into the glass, it is conceivable that the reheating of glass powder in a gold crucible at lower temperature may lead to further gold crucible corrosion in such a way that Au NPs are formed immediately from the dissolved Au⁺ ions. To exclude such possibility, S5 glass (undoped TZN starting glass melted in a silica crucible and thus without Au⁺ ions) was powderized and then heated in a gold crucible under the same reheating conditions as used for p2R-1. The resulting glass sample p5R-1 is colourless (though strictly speaking it is white) and has no extinction band in the visible, indicating the absence of Au NPs in this glass (Fig. 3b). The elevated background extinction generating the white appearance of the sample originates from the light scattering of air bubbles in the glass due to incomplete fusion of the glass particles as also observed for other powder reheated glasses. As for the corresponding starting glass sample S5, there was no Au measured above the detection limit of 0.1 ppm in p5R-1 via solution ICP-MS analysis. This result reveals inefficient Au dissolution into TZN glass at temperatures < 600 ^oC, at least for the 10 min duration used, which confirms that the powderizing and subsequent reheating of Au⁺ ion containing TZN glass induces the apparent reduction of Au⁺ ions to Au⁰, which then grow into Au NPs. Investigations to reveal the underlying mechanisms for this peculiar phenomenon are still in progress.

2.3. Impact of gold doping and powder reheating conditions on Au NP size, size distribution and number density

The ability to control the LSPR features of band location, width and intensity is of paramount importance for applications such as enhancing rare earth fluorescence through plasmonic interaction. These features are determined by Au NP size, size distribution and number density. Here, we investigate the impact of dissolved Au content (via gold crucible corrosion during batch melting at different temperatures T₁) as well as powder reheating temperature and duration (T₂ and t₂) on the Au NP features of powder reheated glass samples.

The photographs of the samples (Fig. 5a) show that they all exhibit a more or less pronounced blue colour in transmission and orange colour in reflection, except for p2R-0 which did not show the reflection feature. The colour and dichroic features indicate Au NP formation in all samples. The difference in transmission and reflection behaviour demonstrates formation of different size Au NPs in the samples.

The Au NP size in the various reheated glass samples (listed in Table 2) was derived using two methods; (i) EM imaging analysis and (ii) Mie theory based calculations of extinction spectra. The EM image analysis yielded Au NP size in the range of 33–89 nm diameter (Fig. 5b). For the Mie theory based analysis, calculated extinction coefficient spectra of monodisperse Au NPs were fitted to the peak position and height of the measured spectra as detailed in the Materials and methods section. The Au NP diameters with best match of the peak position are in the range of 36–90 nm (Fig. 5c), which agrees well with the Au NP size determined via EM image analysis.

The calculated extinction coefficient spectra allow deconvolution into the LSPR absorption and scattering components (Fig. 5c). The scattering component lies in the orange wavelength region and, for p2R-1 to p2R-5 samples, takes up a considerable proportion of the entire LSPR extinction spectrum. This is consistent with these samples showing considerable colour specific scattering of orange light when illuminated with white light (Fig. 5a). It is also consistent with dark field optical microscopy (Figure S5), where the Au NPs appear as bright orange particles that are dispersed homogeneously, with a dark orange background originating from the scattered light of Au NPs out of the focal plane.

The EM image analysis allowed to determine the Au NP size distribution as detailed in the Materials and methods section. For samples p2R-0 to p2R-4 and p3R-1 with low Au concentration (6–19 ppm) and made using short reheating time (10-20min), a narrow size distribution (±~ 10 nm) was found. Not surprisingly for these samples, a good agreement between calculated extinction spectra of monodisperse Au NPs with measured extinction spectra is observed as noted above. Accordingly, the size distributions determined via EM image analysis and Mie theory based analysis show a good agreement as detailed in SI. This demonstrates the suitability of our Mie calculation method for determining average Au NP size for glasses with narrow size distribution.

In contrast to the samples with lower Au concentration and made using shorter reheating times, p2R-5 made using long reheating time (30 min) and p4R-1 with highest Au content (75 ppm) exhibit larger size distributions of ± 28 nm and ± 23 nm, respectively. The widening of the size distribution in these two samples can be attributed to aggregative NP growth ^34–37, whereby longer reheating time for p2R-5 and higher initial Au NP content (due to higher total Au content) for p4R-1 are the driving factors, respectively. Note that for p4R-1, the larger size distribution explains the larger deviation of the calculated extinction spectrum of monodisperse Au NPs with the measured extinction spectrum. For p2R-5, despite also having larger size distribution, the good agreement between calculated extinction spectrum of monodisperse distribution with measured extinction spectrum of polydisperse distribution is attributed to the more symmetric, flat top distribution, whereas p4R-1 has an asymmetric distribution with pronounced peak on the smaller NP side.

To gain further insight into the impact of Au NP size distribution shape on the position of the extinction spectra calculated assuming monodisperse distribution, Mie spectra were calculated using the size distribution width values of the EM analysis (details of the calculation are given in the SI). Figure S4 shows good agreement of this calculated spectrum with the measured spectrum, which validates the suitability of the Mie theory for calculation of extinction spectra of Au NPs in glass. It also shows that for p2R-5, the symmetric and flat size distribution leads to good agreement between spectra calculated assuming monodisperse distribution and EM measured distribution. By contrast, the asymmetric distribution with pronounced peak for p4R-1 leads to significant difference between spectra calculated assuming monodisperse distribution and EM measured distribution. This result demonstrates that comparison of calculated spectra (assuming monodisperse size distribution) with measured spectra is suitable to deduce broad Au NP size distribution for asymmetric NP size distributions.

The Au NP number density (listed in Table 2) of the samples was determined using both EM image analysis and Mie theory based calculations of extinction spectra of monodisperse Au NPs. In addition, dark field optical microscopy was used to derive the Au NP number density in the focal plane (Table 2). The good agreement between the Au NP number density values for these three methods as shown in Table 2 and the SI section 5 demonstrates the consistency of the measurements and analyses.

The impact of the Au content in the starting glasses and the powder reheating conditions on Au NP size, distribution and number density is summarised below. For the samples p2R-§ with fixed Au content of ~ 6ppm, the Au NP size increases and hence the Au NP number density decreases with increasing reheating temperature T₂ and time t₂. This is confirmed by the dark field optical microscopy, where an increase in T₂/t₂ leads to an increase of particle-to-background contrast, indicating increased population of larger Au NPs with decreased number density (Figure S5). It is also confirmed by the color change of the samples; the blue colour in transmission decreases, while the orange colour in reflection intensifies, indicating formation of larger Au NPs. Enhanced diffusion either due to decreased viscosity (i.e., increased temperature) or prolonged heating (i.e., increased time) enables growth of larger Au NPs.

For the samples p2R-1, p3R-1 and p4R-2 with increasing Au content and made at fixed reheating temperature and time, the Au NP size is approximately constant, indicating that reheating conditions determine Au NP size, while the Au content has negligible effect. As expected from the observation of similar Au NP size for fixed reheating temperature and time, an increase in Au content leads to an increase in Au NP number density. This is confirmed by dark field optical microscopy, where the increase of Au content leads to a higher number density of orange Au NPs accompanied by brighter orange background (Figure S5). Furthermore, the intensification of both the blue colour in transmission and the orange colour in reflection, i.e. of both absorption and scattering, indicates larger number of Au NPs of similar size for the glasses made from starting glasses with higher Au⁺ ion concentration and using the same powder reheating conditions.

We investigated the incorporation of Au⁺ ions into tellurite glass via the traditional doping technique of melting gold salt containing tellurite glass batch in a silica crucible and our novel technique of melting dopant-free tellurite glass batch in a gold crucible. For both techniques, the critical step for Au dissolution is the oxidation of Au⁰ to Au⁺. For the traditional technique, the Au⁰ occurs in the form of Au metal particles formed via gold salt decomposition upon batch heating. For the novel technique, the Au crucible is the source of Au⁰. The different amount of dissolved Au⁺ ions in tellurite glass made under the same melting conditions (temperature and duration) for the two techniques is attributed to different level of available oxygen for the oxidation of Au⁰. For the traditional technique, the limited amount of oxygen dissolved in the glass melt is available for oxidation. By contrast, for the novel doping technique, unlimited amount of oxygen in the melting atmosphere leads to enhanced oxidation and hence higher amount of dissolved Au⁺ ions, which can be tuned by controlling the melting temperature.

We found that the traditional technique of using reducing agents in the glass batch – which has been extensively used to create Au NPs in silicate glass – is not suitable for tellurite glass as the addition of the polyvalent compounds SnO or SnO₂ into the tellurite glass leads to the formation of reduced tellurium species (Te⁰ or Te²⁻) which give rise to undesired absorption in the visible that impart the glass with a black/grey colour. Inspired by a serendipitous observation, we discovered that reheating powderized Au⁺ ion doped tellurite glass (without SnO or SnO₂) could effectively create Au NPs homogeneously within the glass matrix, while reheating corresponding bulk glass could barely transform the Au⁺ ions in the glass into the metallic form. Most importantly, we demonstrated that the formation of Au NPs with tunable size, size distribution and number density can be controlled by the amount of Au⁺ ions introduced into the starting glass using the novel gold doping technique of gold crucible corrosion together with the reheating temperature and duration of the powderized starting glass. This allows control of LSPR features in terms of band location and intensity. Specifically, increasing reheating temperature or prolonging reheating duration leads to the formation of larger Au NPs with lower number density (considering most if not all Au exists as NPs in the reheated glass), resulting in a redshift of the LSPR band and decrease of the peak intensity. For the scenario of fixed reheating temperature and duration, higher amount of Au⁺ ions in the starting glass predominantly contributes to the formation of more Au NPs in glass with similar size, giving rise to more intense LSPR band.

The ability to tune the LSPR features in tellurite glass paves the way towards plasmonics based research and applications, such as comprehensive investigation of plasmonic effects on luminescence of rare earth ions in tellurite glass.

There are several pathways for further investigations to gain insight into the mechanisms of the Au NP formation in tellurite glass, which opens up enhanced control of the Au NP formation and hence LSPR-related properties. One option is to uncover the effect of other processing parameters on the Au NP size, size distribution and number density in tellurite glass, such as batch melting atmosphere (i.e., oxygen partial pressure) for Au⁺ incorporation and particle size of glass powders for Au NP formation. Understanding of the underlying mechanism for the efficient reduction of Au⁺ in tellurite glass via the novel powder reheating based ‘striking’ technique would open up further avenues for control of Au NP formation. Furthermore, the cause for the one order of magnitude lower amount of dissolved Au in tellurite glass compared to silicate is still to be explored, specifically whether the gold solubility of gold in glass is connected to the solubility of oxygen.

4.1. Sample fabrication

All tellurite glass samples were fabricated using the TZN composition of 75TeO₂-15ZnO-10Na₂O (in mol%). Commercially sourced raw materials were used: TeO₂, ZnO, Na₂CO₃, SnO, SnO₂ and HAuCl₄ with 99.9% or higher purity. HAuCl₄ aqueous solution (10 mg/ml) was prepared by dispersing HAuCl₄ salt into corresponding amount of ultra-pure water (Millipore Milli-Q lab water system) in a thoroughly cleaned glass vial (with reagent grade 69% HNO₃ and ultra-pure water), which was stored at low temperature (~ 4 ^oC) for later use.

All precursor bulk glass samples (hereafter referred to as Starting glass samples, S1 – S8) were fabricated by melting the ~ 100 g TZN glass batches (which were undoped, or doped with SnO or SnO₂, or doped with gold as HAuCl₄ aqueous solution) in a gold or silica crucible at temperature T₁ for a duration of t₁ in air atmosphere, followed by casting the melt in a brass mold (preheated to 240 ^oC), which was then annealed at ~ T_g (293 ^oC for TZN) for 2h and slowly cooled down to room temperature. Note that for the glass batch doped with HAuCl₄ aqueous solution, it was homogenized by manually grinding in an agate mortar with an agate pestle for 10 min in air atmosphere, before melting and the subsequent processes. The dopant type and melting conditions are provided in Table 1.

Powders of the bulk starting glass samples were prepared by manually crushing and grinding in an agate mortar with an agate pestle for 10 min in air atmosphere. All starting glass powders used for reheating have particle sizes mostly in the range 1–100 µm (Figure S3).

Reheating of starting glasses in bulk or powder form resulted in so-called bulk Reheated glass samples (b#R-§) and powder Reheated samples (p#R-§), respectively, where # represents the number of the starting glass sample and § represents a certain reheating temperature and time schedule (Table 2).

The Reheated glass samples were fabricated by reheating a bulk piece (~ 4 cm³) or the powder of the corresponding starting glass (~ 20 g) in a gold crucible at temperature T₂ for a duration of t₂ in air atmosphere and subsequent quenching of the melt in a brass mold, which was then annealed at ~ T_g (293 ^oC for TZN) for 2h and slowly cooled down to room temperature.

Table 1

Melting conditions and Au content for the TZN starting glasses. ppm refers to parts per million by mass. n/m refers to not measured.
Sample	Crucible	Batch Doping	T₁ (^oC)	t₁ (min)	Au content (ppm)
S1	silica	10 ppm Au^a	750	60	3 ± 0.9^b
S2	gold	none	750	60	6.3 ± 1.4^b
S3	gold	none	800	60	19.1 ± 3.1^b
S4	gold	none	850	60	75.2 ± 3.5^b
S5	silica	none	750	60	< 0.1^b
S6	gold	2 w% SnO	750	60	n/m
S7	gold	2 w% SnO₂	750	60	n/m
S8 ^c	gold	none	750	60	n/m

^a Au in the form of HAuCl₄ aqueous solution doped into the TZN glass batch.

^b Au content measured via solution inductively coupled plasma mass spectrometry (ICP-MS), where < 0.1 ppm refers to Au content below the detection limit (of 0.1 ppm)

^c Although this starting glass was fabricated under the same conditions as S2, a different sample name was used considering this starting glass was further powderized and mixed with ruby (Al₂O₃:Cr³⁺) nanocrystals for fabricating another sample (refer to footnote ^a in Table 2).

Table 2

Reheating conditions for reheated TZN glasses, along with the measured LSPR peak wavelength, the Au NP diameter determined using electron microscopy or Mie simulation (\({D}_{EM}\), \({D}_{Mie}\)), the particles’ volume number density determined using electron microscopy or Mie simulation (\({N}_{EM}\), \({N}_{Mie}\)) and planar number density determined using optical microscopy (\({N{\prime }}_{OM}\)). For the sample name, the first number refers to the starting glass used for reheating, while the second number represents the reheating temperature and time schedule used to make the sample. n/a refers to not applicable due to lack of Au NPs, and n/m refers to not measured.
Sample	T₂ (^oC)	t₂ (min)	LSPR peak (nm)	D_EM (nm)	D_Mie (nm)	N_EM (10¹⁰/cm³)	N_Mie (10¹⁰/cm³)	N’_OM (10⁷/cm²)
b2R-1	570	10	n/a	n/a	n/a	n/a	n/a	n/a
p2R-0	510	10	605	33 ± 13	36	6.88 ± 1.53	4.97	n/m^g
p2R-1	570	10	645	62 ± 9	62	1.23 ± 0.27	0.91	0.39
p2R-2	590	10	680	71 ± 12	78	0.83 ± 0.18	0.56	0.26
p2R-3	610	10	710	84 ± 12	90	0.52 ± 0.12	0.41	0.19
p2R-4	570	20	660	65 ± 10	69	1.07 ± 0.24	0.54	0.28
p2R-5	570	30	698	89 ± 28	86	0.35 ± 0.08	0.19	0.16
p3R-1	570	10	642	60 ± 11	60	4.29 ± 0.70	2.77	1.09
p4R-1	570	10	638	75 ± 23	58	7.21 ± 0.34	4.56	1.65
p5R-1	570	10	n/a	n/a	n/a	n/a	n/a	n/a
p8R-1^a	570	10	642	n/m	n/m	n/m	n/m	n/m

^a The starting glass powder was mixed with ruby nanocrystals prior to reheating. This glass was initially fabricated for another purpose but it also led to the discovery of a new method for the in-situ formation of gold NPs in TZN glass as detailed in the Results section.

4.2. Characterization and Calculations

Solution inductively coupled plasma mass spectrometry (ICP-MS)

Solution ICP-MS was used to determine the total Au concentration (\({w}_{Au})\) in the starting glass samples (S1 – S5) in ppm-weight. Specifically, for each glass sample, ~ 0.1g in pulverized form was digested in 10 ml of aqua regia (37% HCl: 69% HNO₃ = 3:1, in volume fraction) in a Perfluoroalkoxy alkane vial with cover for 48 h at room temperature. After digestion, the sample solution was further diluted 20 fold with ultra-pure water (Millipore Milli-Q lab water system). The Au standard solutions with concentrations of 500, 200, 100, 50, 20 and 10 µg/L prepared by diluting the 10 mg/L Au stock solution with diluted aqua regia (20 fold by ultra-pure water), together with a blank solution of the diluted aqua regia were used for calibration. The sample solutions, together with the calibration solutions, were analyzed by solution ICP-MS on an Agilent 7500cx ICP-MS (Adelaide Microscopy) to quantify the Au concentrations in TZN glass samples.

Electron Microscopy (EM)

EM imaging via Scanning Transmission Electron Microscopy (STEM) or Scanning Electron Microscopy (SEM) and elemental analysis via Energy-Dispersive X-ray Spectroscopy (EDS) were used to identify and measure the size Au NPs in glass.

STEM and EDS

For the sample p2R-0 with smallest NP size, STEM coupled with EDS was employed to identify Au NPs, using an ultra-high-resolution, aberration-corrected FEI Titan Themis 80–200 operated at 200 kV (Adelaide Microscopy). The glass piece used for testing was pulverized in an agate mortar (in air atmosphere) in the presence of isopropanol for 30 min. The isopropanol solution containing dispersed glass powders (0.1–10 µm) was drop-cast onto a standard copper grid and dried for further STEM and EDS analysis. STEM images were taken in a high-angle angular dark field (HAADF) configuration to identify Au NPs in glass matrix by the Z-contrast related to the atomic number Z (i.e., brighter pixel indicates higher atomic number). STEM-EDS maps were acquired to further confirm the elemental nature of Au NPs.

SEM and EDS

For all other samples, SEM coupled with EDS was employed to identify Au NPs, using a FEI Quanta 450 FEG Environmental Scanning Electron Microscope operated at 10 kV (Adelaide Microscopy). Each glass piece used for testing was fractured in air atmosphere and then coated with a thin graphite layer (~ 15 nm) prior to SEM imaging to minimize charging from electron accumulation on the sample surface. SEM images were taken in backscattering mode to identify Au NPs near the surface of samples by the elemental Z-contrast. EDS spot analysis on different regions of each sample was further conducted to confirm the elemental nature of Au NPs.

EM image analysis

For the sample p2R-0 with smallest Au NP size, the size distribution of Au NPs in the sample could only be determined based on 7 different Au NPs due to the difficulty in find Au NPs under STEM, leading to large uncertainty in the Au NP size distribution. For the other samples, the size distribution of Au NPs was determined from images showing 20–30 different Au NPs.

The Au NP number density determined via EM image analysis (N_EM) was obtained by first measuring the total Au concentration (\({w}_{Au}\)) in the sample under consideration via ICP-MS and the volume averaged Au NP diameter (\({D}_{EM-VA}=\sqrt[3]{\frac{\sum _{i=1}^{n}{D}_{i}^{3}}{n}}\), where \({D}_{i}\) is the measured diameter of the i^th Au NP) in the sample via EM image analysis. From these two measured values and assuming that all Au⁺ ions were transformed to Au NPs during the powder reheating, the Au NP number density is calculated using known values for Au density (\({\varrho }_{Au}\)=19.32 g/cm³), TZN glass density (\({\varrho }_{g}\)=5.15 g/cm^{3 38}) via \({N}_{EM}=6\bullet {w}_{Au}\bullet {\varrho }_{g}/(\pi \bullet {\varrho }_{Au}\bullet {D}_{EM-VA}^{3})\).

UV-Vis spectroscopy

For both the starting and the reheated glass samples, the extinction spectra (\(E=-{log}_{10}T\), where \(T\) is transmittance) were measured for polished samples of ~ 2 mm thickness using Agilent Cary 5000 UV-VIS-NIR spectrophotometer. Note that the step at 800 nm in the extinction spectra of most samples is due to detector changeover.

For the reheated glass samples showing distinct LSPR bands, in order to separate LSPR contribution from all other light-attenuation effects such as Fresnel reflection or scattering by gas bubbles, surface defects and possibly non-metallic crystals of the glass matrix and impurities (hereafter referred to as background contributions), the extinction value at 1000 nm (\(E\left(1000nm\right)\)) was used as a measure of these background contributions for all wavelengths. By subtracting a constant background value from the whole measured extinction spectrum across 350–1000 nm, a spectrum (\(\varDelta E\left(\lambda \right)=E\left(\lambda \right)-E\left(1000\text{nm}\right))\) is obtained which represents the LSPR extinction. By normalizing the LSPR extinction spectrum (\(\varDelta E)\) to the sample thickness (\(d\)), the measured extinction coefficient spectrum of LSPR (\(\varDelta {\epsilon }_{meas}=\varDelta E/d\)) is obtained. Detailed justification of the background subtraction strategy is provided in the SI.

Mie theory based calculations

The calculated LSPR extinction cross section (\({\sigma }_{calc}\)) spectra were obtained using the MiePlot version 4.6.04 software provided by Philip Laven ³⁹. Input parameters include particle size and size-dependent complex refractive index of Au NPs, as well as real refractive index of the glass matrix.

The use of appropriate dielectric functions (refractive index) of Au NPs is critical to accurately calculate the extinction spectra of Au NPs. To determine if quantum correction beyond the extended Drude model is required in this work, the extinction spectrum of 20 nm Au NPs (lower size limit for the Drude model) in TZN glass was calculated using the extended Drude model based refractive index of Au via the program provided in Ref ⁴⁰. The peak position of the calculated extinction cross spectrum for 20 nm Au NPs in TZN glass is at 590 nm, which is at shorter wavelength compared to the measured LSPR peak wavelengths for all TZN glass samples in this work. This indicates the size of the Au NPs in all the TZN glass samples is larger than 20 nm and, therefore, the extended Drude model is sufficient to correct the refractive index of Au NPs for calculating the LSPR extinction cross section \({\sigma }_{calc}\).

To allow comparison between calculated and measured LSPR spectra, the same method of background subtraction was used, namely the value of calculated \({\sigma }_{calc}\) at 1000 nm was subtracted from the whole calculated \({\sigma }_{calc}\) spectrum over 400–1000 nm. In this way, the calculated extinction cross section spectrum \(\varDelta {\sigma }_{calc}\) is obtained, which shows zero intensity at 1000 nm as for the measured extinction coefficient spectra \(\varDelta {\epsilon }_{meas}\).

The Mie calculations were employed to find the \(\varDelta {\sigma }_{calc}\) spectrum that fits best to the \(\varDelta {\epsilon }_{meas}\) spectrum under consideration, which enabled to predict the Au NP size corresponding to the measured LSPR spectrum, based on the following relationship between extinction cross section, \(\sigma\), (based on natural logarithm) and extinction coefficient, \(\epsilon\), (based on common logarithm) ⁴¹:

\(\epsilon =\frac{{\sigma \bullet N}_{Mie}}{2.303}\) 1

where \({N}_{Mie}\) is the number density of Au NPs. The details of the LSPR spectrum fitting and Au NP size prediction are as follows.

First, the calculated LSPR extinction coefficient spectrum \({\varDelta \epsilon }_{calc}\) that fits best to the measured spectrum \({\varDelta \epsilon }_{meas}\) was obtained as follows: The\(\varDelta {\epsilon }_{meas}\) spectrum shows a distinct LSPR peak at a specific wavelength (such as 645 nm for sample p2R-1 shown in Fig. 5b). Mie calculation was done by setting a certain particle size (along with the corrected refractive index of Au and refractive index of TZN glass), which yielded \({\sigma }_{calc}\) spectrum and subsequently \(\varDelta {\sigma }_{calc}\) spectrum with its peak wavelength best matching that of the measured peak wavelength (by trial-and-error). Then, the Mie calculation based Au NP number density (N_Mie) was determined via:

\({N}_{Mie}=\frac{2.303(\varDelta {\epsilon }_{meas peak})}{{\varDelta \sigma }_{calc peak}}\) 2

where \(\varDelta {\epsilon }_{meas peak}\) is the peak intensity of the \(\varDelta {\epsilon }_{meas}\) spectrum, and \({\varDelta \sigma }_{calc peak}\) is the peak intensity of the \({\varDelta \sigma }_{calc}\) spectrum. Finally, the whole \({\varDelta \sigma }_{calc}\) spectrum was multiplied with \({N}_{Mie}\) to obtain the calculated (i.e., fitted) LSPR extinction coefficient \({\varDelta \epsilon }_{calc}\) spectrum according to Eq. 1.

For a specific Au NPs size distribution, the calculations were done according to:

\({\varDelta \epsilon }_{calc}=\frac{\sum {{\varDelta \sigma }_{calc k}N}_{Mie k}}{2.303}\) 3

where\({\varDelta \sigma }_{calc k}\) and \({N}_{Mie k}\) are the calculated extinction cross section and number density of Au NPs with size range \(k\), respectively.

To deconvolute the absorption and scattering contributions of the LSPR extinction of monodisperse Au NPs, the calculated (fitted) LSPR absorption and scattering cross sections \({{\Delta }\sigma }_{calc abs}\) and \({\varDelta \sigma }_{calc sca}\) were extracted from the calculated LSPR extinction spectrum \(\varDelta {\sigma }_{calc}\). \({N}_{Mie}\) obtained via Eq. 2 was then used to multiply the \(\varDelta {\sigma }_{calc abs}\) and \(\varDelta {\sigma }_{calc sca}\) spectra to obtain the calculated (i.e., fitted) LSPR absorption and scattering coefficient spectrum \({\varDelta \epsilon }_{calc abs}\) and \({\varDelta \epsilon }_{calc sca}\) according to Eq. 1.

Due to mismatch between the experimentally observed non-monodispersed nature of Au NPs in all samples relative to the theoretically monodispersed Au NPs in the Mie calculations, the N_Mie values calculated according to Eq. 1 represent an undervalued Au NP number density. For the comparison of the Au NP number densities obtained via Mie calculation and EM image analysis, the N_Mie values were normalised to the N_EM value of the p2R-1 sample.

Dark-field optical microscopy

Dark-field optical microscopy was employed as a non-destructive optical technique to identify Au NPs (> 50 nm) showing strong LSPR scattering in TZN glass samples p2R-1 – p2R-5, p3R-1 and p4R-1, using an inverted Olympus BX51 microscope equipped with a dry dark-field condenser (U-DCD, N.A. up to 0.80). The scattered light from Au NPs within the focal volume was collected by a MPlanFL N 50x / 0.80 objective (MPlanFL N); and the images were taken using a DP50 digital camera employs a 1/2 inch, 1.5 million pixel CCD, with fixed exposure time of 200 ms.

Dark field optical microscopy was used as an alternative method to determine a measure of the Au NP number density. Specifically, the planar Au NP number density (\({N{\prime }}_{OM}\)) represents the number of Au NPs detected per cm² in the focal plane of a dark-field optical microscope image. To allow comparison of the N’_OM values with the N_EM values, the N’_OM values were normalised to the N_EM value of the p2R-1 sample as done for the N_Mie values.

Note that similar dark field optical microscope setting (except using 5x / 0.15 and 20x / 0.45 objectives, with auto-adjusted exposure time) was employed to demonstrate the size range of the TZN glass powders used for reheating.

Acknowledgements

This work has been supported by Australian Research Council (ARC) grants DP210102442 and LE190100124 and the ARC Centre of Excellence for Nanoscale BioPhotonics (CNBP) (CE140100003). This work was performed in part at the OptoFab node of the Australian National Fabrication Facility utilizing Commonwealth and SA State Government funding. The authors acknowledge the instruments and scientific and technical assistance of Microscopy Australia at Adelaide Microscopy, The University of Adelaide, a facility that is funded by the University, and State and Federal Governments. H. E.-H. is supported by a South Australian Government Future Industry Making Fellowship.

Author Contributions

Y. W., J. Z., and H. E. conceived the research; Y. Wei conducted the experiments, calculations and data analysis; H. E. supervised the project; and all authors contributed to the writing of the manuscript and the interpretation of the results.

Conflict of Interest

The authors declare no competing interests.

Wei, Y., Ebendorff-Heidepriem, H. & Zhao, J. Recent Advances in Hybrid Optical Materials: Integrating Nanoparticles within a Glass Matrix. Adv. Opt. Mater. 7, 1900702, doi:https://doi.org/10.1002/adom.201900702 (2019).
Fonseca, J. Nanoparticles embedded into glass matrices: glass nanocomposites. Front. Mater. Sci. 16, 220607, doi:10.1007/s11706-022-0607-7 (2022).
Li, N., Zhao, P. X. & Astruc, D. Anisotropic Gold Nanoparticles: Synthesis, Properties, Applications, and Toxicity. Angew. Chem., Int. Ed. 53, 1756-1789, doi:10.1002/anie.201300441 (2014).
Borisov, A. & Palme, H. Experimental determination of the solubility of Au in silicate melts. Mineral. Petrol. 56, 297-312, doi:10.1007/bf01162608 (1996).
Wagner, F. E. et al. Before striking gold in gold-ruby glass. Nature 407, 691-692 (2000).
Brenan, J. M., McDonough, W. F. & Ash, R. An experimental study of the solubility and partitioning of iridium, osmium and gold between olivine and silicate melt. Earth Planet. Sci. Lett. 237, 855-872, doi:10.1016/j.epsl.2005.06.051 (2005).
Stookey, S. D. Coloration of glass by gold, silver, and copper. J. Am. Ceram. Soc. 32, 246-249, doi:10.1111/j.1151-2916.1949.tb18957.x (1949).
Kassab, L. R. P., da Silva, D. M. & de Araujo, C. B. Influence of metallic nanoparticles on electric-dipole and magnetic-dipole transitions of Eu³⁺ doped germanate glasses. J. Appl. Phys. 107, 113506, doi:10.1063/1.3431347 (2010).
Wu, Y. et al. Silver Nanoparticles Enhanced Upconversion Luminescence in Er³⁺/Yb³⁺ Codoped Bismuth-Germanate Glasses. J. Phys. Chem. C 115, 25040-25045, doi:10.1021/jp207035c (2011).
Shakhgildyan, G. Y. et al. Thermally-induced precipitation of gold nanoparticles in phosphate glass: effect on the optical properties of Er³⁺ ions. J. Non-Cryst. Solids 550, 120408, doi:https://doi.org/10.1016/j.jnoncrysol.2020.120408 (2020).
Marcondes, L. M. et al. Structural and optical characterization of tungsten phosphate glasses containing silver and erbium. Opt. Mater. 132, 112717, doi:https://doi.org/10.1016/j.optmat.2022.112717 (2022).
Rajaramakrishna, R., Karuthedath, S., Anavekar, R. V. & Jain, H. Nonlinear optical studies of lead lanthanum borate glass doped with Au nanoparticles. J. Non-Cryst. Solids 358, 1667-1672, doi:https://doi.org/10.1016/j.jnoncrysol.2012.04.031 (2012).
Jagannath, G. et al. Influence of gold nanoparticles on the nonlinear optical and photoluminescence properties of Eu₂O₃ doped alkali borate glasses. Phys. Chem. Chem. Phys. 22, 2019-2032, doi:10.1039/C9CP05783H (2020).
Som, T. & Karmakar, B. Core-shell Au-Ag nanoparticles in dielectric nanocomposites with plasmon-enhanced fluorescence: A new paradigm in antimony glasses. Nano Res. 2, 607-616, doi:10.1007/s12274-009-9061-4 (2009).
Som, T. & Karmakar, B. Surface Plasmon Resonance and Enhanced Fluorescence Application of Single-step Synthesized Elliptical Nano Gold-embedded Antimony Glass Dichroic Nanocomposites. Plasmonics 5, 149-159, doi:10.1007/s11468-010-9129-8 (2010).
Tao, G. et al. Infrared fibers. Adv. Opt. Photonics 7, 379-458, doi:10.1364/AOP.7.000379 (2015).
Rai, V. K. et al. Surface-plasmon-enhanced frequency upconversion in Pr³⁺ doped tellurium-oxide glasses containing silver nanoparticles. J. Appl. Phys. 103, doi:10.1063/1.2919566 (2008).
de Campos, V. P. P., Kassab, L. R. P., de Assumpcao, T. A. A., da Silva, D. S. & de Araujo, C. B. Infrared-to-visible upconversion emission in Er³⁺ doped TeO₂-WO₃-Bi₂O₃ glasses with silver nanoparticles. J. Appl. Phys. 112, doi:10.1063/1.4754468 (2012).
Dousti, M. R. et al. Enhanced frequency upconversion in Er³⁺-doped sodium lead tellurite glass containing silver nanoparticles. Eur. Phys. J. D 66, 237, doi:10.1140/epjd/e2012-30089-1 (2012).
Mechergui, I., Fares, H., Mohamed, S. A., Nalin, M. & Elhouichet, H. Coupling between surface plasmon resonance and Sm³⁺ ions induced enhancement of luminescence properties in fluoro-tellurite glasses. J. Lumin. 190, 518-524, doi:10.1016/j.jlumin.2017.06.013 (2017).
Yu, C. Y., Yang, Z. W., Qiu, J. B., Song, Z. G. & Zhou, D. C. Enhanced photoluminescence property and mechanism of Eu³⁺-doped tellurite glasses by the silver and gold nanoparticles. J. Am. Ceram. Soc. 101, 612-623, doi:10.1111/jace.15222 (2018).
Xu, W., Chen, X. & Song, H. W. Upconversion manipulation by local electromagnetic field. Nano Today 17, 54-78, doi:10.1016/j.nantod.2017.10.011 (2017).
Henderson, M. R. et al. Diamond in Tellurite Glass: a New Medium for Quantum Information. Adv. Mater. 23, 2806-+, doi:10.1002/adma.201100151 (2011).
Ebendorff-Heidepriem, H. et al. Nanodiamond in tellurite glass Part I: origin of loss in nanodiamond-doped glass. Opt. Mater. Express 4, 2608-2620, doi:10.1364/ome.4.002608 (2014).
Saiyasombat, C. Preparation, structure and properties of gold-based ‘ruby’sodium silicate glass: A model metal-glass nano-composite Doctor of Philosophy thesis, Lehigh University (2015).
Otto, K., Acik, I. O., Krunks, M., Tonsuaadu, K. & Mere, A. Thermal decomposition study of HAuCl₄·3H₂O and AgNO₃ as precursors for plasmonic metal nanoparticles. J. Therm. Anal. Calorim. 118, 1065-1072, doi:10.1007/s10973-014-3814-3 (2014).
Borisov, A. & Palme, H. Solubilities of noble metals in Fe-containing silicate melts as derived from experiments in Fe-free systems. Am. Mineral. 85, 1665-1673 (2000).
Ruivo, A. et al. Gold nanoparticles in ancient and contemporary ruby glass. J. Cult. Herit. 9, E134-E137 (2008).
Pellerin, N. et al. Control of selective silicate glass coloration by gold metallic nanoparticles: structural investigation, growth mechanisms, and plasmon resonance modelization. Gold Bull. 46, 243-255, doi:10.1007/s13404-013-0121-x (2013).
Ogbuu, O. et al. Impact of Stoichiometry on Structural and Optical Properties of Sputter Deposited Multicomponent Tellurite Glass Films. J. Am. Ceram. Soc. 98, 1731-1738, doi:10.1111/jace.13534 (2015).
Qiu, J. R. et al. Manipulation of gold nanoparticles inside transparent materials. Angew. Chem., Int. Ed. 43, 2230-2234, doi:10.1002/anie.200352380 (2004).
Kracker, M. et al. Redox effects and formation of gold nanoparticles for the nucleation of low thermal expansion phases from BaO/SrO/ZnO/SiO₂ glasses. Rsc Adv. 8, 6267-6277, doi:10.1039/c7ra12855j (2018).
Thieme, C. et al. The acceleration of crystal growth of gold-doped glasses within the system BaO/SrO/ZnO/SiO₂. J. Eur. Ceram. Soc. 39, 554-562, doi:10.1016/j.jeurceramsoc.2018.08.041 (2019).
Meli, L. & Green, P. F. Aggregation and Coarsening of Ligand-Stabilized Gold Nanoparticles in Poly(methyl methacrylate) Thin Films. ACS Nano 2, 1305-1312, doi:10.1021/nn800045s (2008).
Dai, Z., Sun, S. & Wang, Z. Phase Transformation, Coalescence, and Twinning of Monodisperse FePt Nanocrystals. Nano Lett. 1, 443-447, doi:10.1021/nl0100421 (2001).
Grillo, F., Van Bui, H., Moulijn, J. A., Kreutzer, M. T. & van Ommen, J. R. Understanding and Controlling the Aggregative Growth of Platinum Nanoparticles in Atomic Layer Deposition: An Avenue to Size Selection. J. Phys. Chem. Lett. 8, 975-983, doi:10.1021/acs.jpclett.6b02978 (2017).
Woehl, T. J. et al. Direct Observation of Aggregative Nanoparticle Growth: Kinetic Modeling of the Size Distribution and Growth Rate. Nano Letters 14, 373-378, doi:10.1021/nl4043328 (2014).
Ruan, Y. L. et al. Nanodiamond in tellurite glass Part II: practical nanodiamond-doped fibers. Opt. Mater. Express 5, 73-87, doi:10.1364/OME.5.000073 (2015).
Laven, P. MiePlot, <http://www.philiplaven.com/mieplot.htm> (2019).
Bohren, C. F. & Huffman, D. R. in Absorption and Scattering of Light by Small Particles Ch. Appendixes: Computer Programs, 475-476 (Wiley-VCH, 2007).
Haiss, W., Thanh, N. T. K., Aveyard, J. & Fernig, D. G. Determination of size and concentration of gold nanoparticles from UV-Vis spectra. Anal. Chem. 79, 4215-4221, doi:10.1021/ac0702084 (2007).

(Not answered)

SupplementaryInformationLSA20230876.docx

Download PDF

Journal Publication

published 07 Dec, 2023

Read the published version in Light: Science & Applications →

Editorial decision: revise
02 Aug, 2023
Review #2 received at journal
23 Jul, 2023
Reviewer #2 agreed at journal
16 Jul, 2023
Review #1 received at journal
09 Jul, 2023
Reviewer #1 agreed at journal
30 Jun, 2023
Reviewers invited by journal
30 Jun, 2023
Submission checks completed at journal
30 Jun, 2023
First submitted to journal
28 Jun, 2023
Editor assigned by journal
28 Jun, 2023

You are reading this latest preprint version

Controlled formation of gold nanoparticles with tunable plasmonic properties in tellurite glass

Status:

Journal Publication

Version 1

Abstract

Figures

1. Introduction

2. Results

2.1. Incorporation of Au⁺ ions into tellurite glass

2.1.1. Traditional technique: Adding gold salt to the glass batch

2.1.2. Controlled gold crucible corrosion technique

2.1.3. Comparison of the mechanisms of both gold doping techniques

2.2. Transformation of dissolved Au⁺ ions to Au NPs

2.2.1. Traditional technique: reducing agent Sn²⁺

2.2.2. New technique: powder reheating based ‘striking’

2.2.3. New technique: powder reheating - control experiments

3. Discussions

4. Materials and methods

4.1. Sample fabrication

4.2. Characterization and Calculations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1

Controlled formation of gold nanoparticles with tunable plasmonic properties in tellurite glass

Status:

Journal Publication

Version 1

Abstract

Figures

1. Introduction

2. Results

2.1. Incorporation of Au+ ions into tellurite glass

2.1.1. Traditional technique: Adding gold salt to the glass batch

2.1.2. Controlled gold crucible corrosion technique

2.1.3. Comparison of the mechanisms of both gold doping techniques

2.2. Transformation of dissolved Au+ ions to Au NPs

2.2.1. Traditional technique: reducing agent Sn2+

2.2.2. New technique: powder reheating based ‘striking’

2.2.3. New technique: powder reheating - control experiments

3. Discussions

4. Materials and methods

4.1. Sample fabrication

4.2. Characterization and Calculations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1

2.1. Incorporation of Au⁺ ions into tellurite glass

2.2. Transformation of dissolved Au⁺ ions to Au NPs

2.2.1. Traditional technique: reducing agent Sn²⁺