Journal of Nanoparticle Research

, Volume 12, Issue 1, pp 39–46

Effect of water and UV passivation on the luminescence of suspensions of silicon quantum dots

Authors

  • J. Vincent
    • DSM/IRAMIS SPAM-LFP (CEA CNRS URA 2453) CEA Saclay Bâtiment 522
    • LITEN/DTNM/L2T CEA Grenoble Bâtiment C2
  • V. Maurice
    • DSM/IRAMIS SPAM-LFP (CEA CNRS URA 2453) CEA Saclay Bâtiment 522
  • X. Paquez
    • DSM/IRAMIS SPAM-LFP (CEA CNRS URA 2453) CEA Saclay Bâtiment 522
  • O. Sublemontier
    • DSM/IRAMIS SPAM-LFP (CEA CNRS URA 2453) CEA Saclay Bâtiment 522
  • Y. Leconte
    • DSM/IRAMIS SPAM-LFP (CEA CNRS URA 2453) CEA Saclay Bâtiment 522
  • O. Guillois
    • DSM/IRAMIS SPAM-LFP (CEA CNRS URA 2453) CEA Saclay Bâtiment 522
  • C. Reynaud
    • DSM/IRAMIS SPAM-LFP (CEA CNRS URA 2453) CEA Saclay Bâtiment 522
    • DSM/IRAMIS SPAM-LFP (CEA CNRS URA 2453) CEA Saclay Bâtiment 522
  • O. Raccurt
    • LITEN/DTNM/L2T CEA Grenoble Bâtiment C2
  • F. Tardif
    • LITEN/DTNM/L2T CEA Grenoble Bâtiment C2
Special focus: Safety of Nanoparticles

DOI: 10.1007/s11051-009-9708-9

Cite this article as:
Vincent, J., Maurice, V., Paquez, X. et al. J Nanopart Res (2010) 12: 39. doi:10.1007/s11051-009-9708-9

Abstract

This article presents the evolution of the photo-luminescence (PL) of silicon quantum dots (QDs) with an average diameter of 5–6 nm dispersed in alcohol under different conditions. Two samples were considered after alcohol dispersion: freshly synthesized (kept in air for 2 days) QDs which do not exhibit luminescence and air-aged (kept in air for 2 years) QDs exhibiting red-IR luminescence. Experiments performed with addition of a small volume of water, followed by heating for different times showed that the oxidation occurs gradually until transforming totally the initial material in SiO2. The oxidation process does not enable the appearance of PL from the Si core for dispersed non-aged powders, while it results in a blue shift of the PL maximum intensity for the aged ones. The results obtained after UV illumination clearly indicate an effect of the UV irradiation on the luminescence of QDs dispersed in aqueous environment, and the treatments with acidic water lead to the conclusion of a possible enhancement of the PL by hydrogen passivation of the non-radiative defects. This result should be taken into account for post-production treatments and applications, more particularly, considering a controlled and safe use of luminescent Si QDs.

Keywords

SiliconNanoparticlesQuantum dotPassivationLuminescenceAqueous suspensionsEnvironmentEHS

Introduction

Semiconductor quantum dots (QDs) show unique size-dependent optical absorption and photo-luminescence (PL) properties arising from the quantum confinement of the electron–hole pair below the Bohr radius (Yoffe 1993; Ledoux et al. 2002). In addition to the interesting properties of nanosized objects such as high quantum yield and low photobleaching (Li and Ruckenstein 2004), the possibility of tailoring the PL response with the particle size has induced a strong research effort on semiconductor QDs. For the last twenty years, various luminescent semiconductors QDs have been synthesized using a wide variety of techniques based on chemical or physical routes (Rogach 2008; Veinot 2006). One example described in the literature is the synthesis of cadmium selenium (CdSe) QDs which can be prepared with a perfectly controlled size and luminescence. However, for most applications of luminescent QDs materials, any hazardous toxicity cannot be accepted (Hardman 2006; Shiohara et al. 2004; Fujioka et al. 2008). Recent in-vitro studies seem to indicate that the cyto-toxicity of silicon QDs is far lower than cyto-toxicity exhibited by CdSe particles (Fujioka et al. 2008). Therefore, it seems possible to use Si QDs as safe agent for luminescent applications such as labeling (anti-counterfeiting, biological markers, pollution tracing, etc.).

Silicon is an indirect band–gap semiconductor and its optical characteristics, with respect to quantum confinement, have been intensively studied (Ledoux et al. 2002; Wolkin et al. 1999; Delerue et al. 1999). However, surface of silicon QDs is very reactive, and its PL properties are still a field of active discussions. Especially quenching of luminescence by defects seems to be the most important phenomenon leading to uncontrolled or non-reproducible luminescence from Si QDs.

Therefore, the control of the surface chemistry is a key point to develop applications of these nanoparticles. A possible way of surface passivation is to grow a silica shell at the surface. Moreover, the SiO2 shell makes Si QDs water easier to disperse and enables further functionalization of QDs surface by a well-developed chemistry (Niemeyer 2001; Burns et al. 2006). Indeed, Si nanoparticles oxidized by simple exposure to ambient atmosphere develop an amorphous silica shell (Yang et al. 2005; Huisken et al. 2003) and exhibit red PL easily detected by naked eye under UV illumination. This PL appears after an exposure time ranging from few days to weeks, depending on storage conditions, humidity, and so on. This process is efficient but slow, and cannot be controlled. It is known that the quality of the oxide layer depends greatly on the growth rate, its production way, with different behavior for air, water, or specific chemical aging (Morita et al. 1990; Bychto et al. 2008; Tardif et al. 2003; Mende et al. 1983; Cerofolini and Meda 1997). This induces different PL behaviors, and this point should be clarified to achieve controlled and stable luminescence properties.

In this context, we present the results obtained on the passivation of Si QDs synthesized by laser pyrolysis of silane focusing on the particular role of water oxidation on Si QDs dispersed in pure ethanol. In order to study the passivation of Si QDs by simple chemical water oxidation and its influence on luminescence properties, several sets of experiments were performed on freshly synthesized (kept in air for 2 days) and air-aged (kept in air for 2 years) Si QDs nanoparticles. The nanoparticles were dispersed in alcohol, small amounts of water were added to the suspension, and the resulting mixtures were heated for different times. These experiments have shown the gradual oxidation of Si QDs by water and the different behaviors of freshly synthesized and air-aged QDs for the different treatment durations. Further studies of the PL stability versus addition of small amount of acid and UV irradiations illustrate the strong photochemical reactivity of aged Si QDs.

Experimental

Pyrolysis of silane SiH4 molecules by CO2 laser is an efficient process to synthesize high-quality Si QDs (Ehbrecht et al. 1995; Li et al. 2003; Lacour et al. 2007). By controlling the silane flow rate, the laser pulse intensity and duration, one can tailor the particle size with a quasi mono-disperse size distribution. The main advantage of this technique is its high production rate ≈200 mg/h for particles with a small diameter (down to 3 nm). The detailed description of the laser-driven aerosol reactor used in our experiment has been reported elsewhere (Lacour et al. 2007).

After production, the Si QDs powders are collected without preventing contact with air. From a dark brown powder showing no PL we obtained after aging in air for 2 years a slightly orange powder exhibiting red-IR luminescence after dispersion in alcohol. This powder is referenced as aged (a), while the powder aged for 2 days is referenced as freshly synthesized (fs) powder.

Si QDs are dispersed in absolute ethanol (AnalaR, Normapur, VWR) by ultra-sonication for a duration of 10 min with the following concentrations [Si QDs] = 10 mg/L, leading to perfectly stable colloidal solutions (no sedimentation or flocculation is observed within years). Experiments were performed by adding a small quantity of deionized (DI) water (Millipore), [Water] = 5 vol.%, to the dispersions. They were heated in a closed vessel at a constant temperature T of 120 °C for duration from t = 4 min to 20 h. In a first approximation, we assume that the heating results mainly in increased kinetics. For experiments using acidic water, we added some tens of mL of HCl (0.03 M) (Sigma Aldrich) or some drops of nitric acid (37%) (Merck). In general, acid dispersions undergo sedimentation within half an hour, and the doses have been adapted to prevent this phenomenon. After thermal treatment, the particles remained in the same solvent (water/alcohol).

After treatments, the PL spectra were measured by illumination of the dispersions using standard quartz with 1*1 cm2 cross section containers with a 6 W UV lamp emitting at 254 nm. The PL signal was collected through an optical fiber connected to a CCD with an acquisition system (spectrometer Ocean Optics USB 2000 covering the range 348–1138 nm). The acquisition time was set to 10 s for each sample which enables direct comparison of the spectra. FTIR measurements were recorded with a Perkin Elmer FTIR2000 spectrometer using the KBr pellet method. Absorption measurements were performed on a Spectronic-Unicam UV-500. Transmission Electron Microscopy (TEM) micrographs were obtained on a Philips CM12 microscope. X-Ray Diffraction patterns (XRD) were obtained on a Bragg–Brentano diffractometer (Siemens D 5000) in the θ–2θ geometry.

Results and discussion

The sizes of particles measured through XRD for the freshly synthesized powders and estimated from TEM micrographs for the air-aged ones are 4.9 and ≈6 nm respectively. Only air-aged QDs exhibited luminescence which can be detected by naked eye during illumination of powder by the UV lamp. The presence of SiO2 on both freshly synthesized and aged powders is confirmed by FTIR measurements as shown in Fig. 1, where three typical absorption bands of Si–O vibrations can be observed at 460, 800, and 1090 cm−1 (Kirk 1998). The Si–H vibrations appearing at 2090 and 2240 cm−1 for freshly synthesized powder have almost disappeared for air-aged powders. It can be noted that the relative intensity of the typical Si–O vibration versus Si–H bonding increases after the aging under air.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-009-9708-9/MediaObjects/11051_2009_9708_Fig1_HTML.gif
Fig. 1

FTIR spectra of aged (curve with large Si-O vibrations) and freshly synthesized (curve with Si-H vibrations) dry powders

After dispersion, it was not possible to detect any luminescence from the suspension elaborated from freshly synthesized Si QDs while the suspension elaborated from air-aged nanoparticles exhibited luminescence with a maximum intensity of 500 counts at 763 nm (see Fig. 2). These spectra are used as reference spectra in the following sections.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-009-9708-9/MediaObjects/11051_2009_9708_Fig2_HTML.gif
Fig. 2

PL spectra of aged and freshly synthesized powders after dispersion in ethanol with water (acquisition with integration duration of 10 s)

Oxidation experiments on QDs were performed with water, and the evolution of the absorption spectrum of the suspensions was studied as a function of treatment duration. These measurements are reported in Fig. 3a and b for freshly synthesized and air-aged Si QDs respectively. The absorption threshold of freshly synthesized Si QDs (Fig 3a) continuously moves to lower wavelengths with increasing treatment duration. However, for treatment duration longer than approximately 1 h, the absorption spectra are typical of silicon dioxide, and it can be assumed that the material is totally transformed into SiO2. TEM micrographs (Fig. 4a) show the presence of slightly aggregated SiO2 nanoparticles after a treatment duration of 20 h.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-009-9708-9/MediaObjects/11051_2009_9708_Fig3_HTML.gif
Fig. 3

Absorbance spectra of freshly synthesized (a) and aged (b) after dispersion and different durations of treatment ranging from 8 min to 20 h

https://static-content.springer.com/image/art%3A10.1007%2Fs11051-009-9708-9/MediaObjects/11051_2009_9708_Fig4_HTML.jpg
Fig. 4

TEM micrographs of freshly synthesized (a) and aged (b) dispersions treated for 20 h

The absorption measurements of the freshly dispersed suspensions obtained from air-aged luminescent nanoparticles (Fig. 3b) clearly show an absorption threshold at a lower wavelength compared to Fig. 3a. This confirms the presence of small Si core. Similar to the evolution observed in Fig. 3a, a shift of the threshold toward shorter wavelengths is observed for duration lower than 30 min. For longer treatment durations (about 1 h), the absorption increases, and the threshold is shifted toward higher wavelength. In addition to the peaks previously observed, these samples show a small absorbance contribution at 280 nm which may be attributed to silica defects (Salh 2007; Robertson 1988). TEM Micrograph (Fig. 4b) shows that after 20 h of treatment, the material presents a strong agglomeration, and it is no more possible to distinguish Si QDs. The increased absorption of the suspension for durations equal or longer than 1 h and the appearance of the 280 nm step may be related to this agglomerated structure.

The above-described results confirm the oxidation of Si QDs in aqueous environment for both air-aged and freshly synthesized powders. The absorption threshold of the spectra is progressively shifted toward shorter wavelengths when the treatment duration is increased. This may indicate a gradual oxidation of the Si QDs (Lockwood et al. 1994; Meier et al. 2007) for durations lower than 1 h. For longer treatment duration, the material is fully oxidized (Fig. 4b). Concerning the agglomeration problem, a different chemical behavior of freshly synthesized and air-aged nanoparticles is observed for durations longer than 1 h.

For the PL aspect, the treatment of the freshly synthesized Si QDs results only in the appearance of a PL peak at 430 nm. Figure 5a shows that its intensity increases with the oxidation duration and reaches a saturation regime at 2000 counts after 15–20 h. This peak is attributed to silica defects (Salh 2007; Robertson 1988), and it should be noted that the low intensity peak observed around 900 nm is a second-order replica of this 430 nm peak. PL spectra of air-aged Si QDs dispersions treated with water for durations ranging between 8 and 30 min and between 1 and 20 h are presented in Fig. 5b and c, respectively. After the treatment, the PL intensity is drastically reduced and the peak maximum is blue shifted (compared to the reference). Indeed, the peak observed at 763 nm before treatment, is shifted gradually to 744 nm, 640 nm, and 610 nm after 8 min, 14 min, and 30 min of oxidation, respectively. Moreover, these spectra exhibit the typical SiO2 PL peak at 430 nm with an intensity increasing with the treatment duration. After 60 min of oxidation, Fig. 5c shows that the 430 nm peak dominates and no other PL peak can be observed.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-009-9708-9/MediaObjects/11051_2009_9708_Fig5_HTML.gif
Fig. 5

PL spectra of Si QDs after dispersion (t = 0) and after treatment for different times: freshly synthesized (a), aged (b) and (c) (acquisition with integration duration of 10 s). It should be noted that the low intensity peak observed around 900 nm on (a) is a second-order replica of the 430-nm peak

From these PL results, we can infer that the water-induced oxidation of the Si QDs leads systematically to some luminescence originating from silica defects with an intensity rising with the treatment duration. We use powders that show a quite good mono-dispersity, considering that typical size dispersion for samples synthesized in similar experimental conditions was measured at ± 10%. Therefore, for duration smaller than 1 h, we exclude the possibility that the oxidation occurs first for very small particles, and it seems most likely that the treatment results in a rough core/shell structure with defect density rising as the oxidation layer increases. Despite the core/shell structure which may appear through the oxidation of silicon, no PL from Si core could be identified for freshly synthesized powders indicating the presence of radiative defects and therefore a low quality of the passivation layer.

Water is known to interact with SiO2 by different mechanisms (Bakos et al. 2004; Batyrev et al. 2008). In order to study the stability of luminescent Si QDs versus UV radiation in aqueous environment, we recorded PL spectra for different times of illumination (tillum) for aged Si QDs dispersion after the water treatments. These measurements, for tillum = 20 h, are reported in Fig. 6a and b. All the dispersions exhibit an increased PL intensity. For the suspension illuminated just after dispersion, the maximum shows a red-shift with a displacement from 763 nm to 782 nm (Figs. 5a, 6a). The PL intensity saturates at 800 counts after 16 h of illumination. The most important effects of UV illumination are seen for the oxidized dispersions. For oxidation duration of 14 min, the PL is shifted from 640 nm to 742 nm with an increase of the intensity by one order of magnitude. The PL of this dispersion saturates after 20–30 h of illumination and exhibits an increasingly higher intensity than the freshly dispersed one. In the case of 8 min of oxidation, we can observe mainly a small decrease of the intensity with a low displacement of the PL maximum compared to the reference. Finally, for 1 h of treatment, a peak at 700 nm is observed with an intensity of 1400 counts. It should be noted that the typical silica peak at 430 nm does not show variation during the UV irradiation, only a peak at 360 nm which appears for oxidation durations longer than 1 h, raised during the illumination producing a second order around 700 nm. This 360-nm peak has not been attributed. We note, however, that no change is observed in the PL of suspensions in pure ethanol (no water).
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-009-9708-9/MediaObjects/11051_2009_9708_Fig6_HTML.gif
Fig. 6

PL spectra comparison between the freshly dispersed aged Si QDs (t = 0) and water oxidized ones after 20 h of UV illumination. (a) oxidation duration ranging between 8 and 30 min and (b) oxidation durations of 1 h and 20 h (acquisition with integration duration of 10 s)

These results demonstrate that the PL peak position of suspensions of air-aged powders can be tuned from 790 to 700 nm with oxidation treatment and UV exposure. In order to explain the increase in PL intensities when suspensions of air-aged nanoparticles are illuminated with UV, one possibility is to consider a luminescence originating from quantum confinement. In this hypothesis, non-radiative defects which reduce the PL intensity are progressively passivated during UV exposure. As many defects are dangling bonds, passivation is likely to occur by bonding of the incorporated hydrogen (Morse and Pianetta 2004; Cartier et al. 1993), which may originate from the dissociation of the water under UV radiation. Therefore, more efficient passivation leads to increased PL intensity. In addition, according to the literature dealing with bulk or layered silicon, some other mechanisms can be involved. In air with oxygen and UV, an oxidation of silicon surface gives a stable SiO2 layer with a good stability (Kasi and Liehr 1992). The exposition of H+ without UV does not affect the native oxide layer (Crossley et al. 1995) in air. The article by Morse and Pianetta (2004) gives some experimental observations of increasing oxidation rate of silicon surface in water by the UV light at 254 nm compared to the water-only exposition. According to these authors, the oxidation of silicon surface comes from the Si surface radical generated by UV light (photoreaction) which reacts with the H2O or O2 to form the oxide layer. Concerning nanoparticles, there are a few reports (Savchyn et al. 2008), and more detailed studies are necessary to understand the mechanisms involved in the passivation of free standing silicon nanoparticles and evolution of PL under UV light.

In order to study the effect of the presence of H+, acidic water (HCl) was used during oxidation treatments. Contrary to the above-described results from DI water experiments, a red shift from 763 to 790 nm and an increase of the PL intensity are observed after 15 min of heating, as can be seen in Fig. 7a. If we compare to treatment with DI water for 14 min (Fig. 5b), a lower intensity of the 430 nm peak is observed. Interestingly, UV illumination experiments reported in Fig. 7b resulted in PL intensity saturation after only 3 h for acidic water treatment, and no shift can be observed, the wavelength remaining at 790 nm. Same results were obtained when HNO3 was used instead of HCl.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-009-9708-9/MediaObjects/11051_2009_9708_Fig7_HTML.gif
Fig. 7

PL spectra of aged Si QDs dispersions after acidic water treatment (a) and the same dispersion after 20H of UV illumination (b) (acquisition with integration duration of 10 s)

These acidic water experiments may confirm the fact that H+ (Cartier et al. 1993) brought by the acid plays a role in non-radiative defect passivation. Interestingly, once the dispersions have been treated with acid, the PL wavelength remains unchanged at 790 nm. Note that this wavelength is close to the position of the PL peak measured on dry powder (815 nm), indicating strong surface effects that shift the PL nanoparticles in suspension (PL of freshly dispersed nanopowders is measured at 765 nm (Fig. 5b). Using the model of Delerue et al. (1999), the shift from 790 to 700 nm can be interpreted in terms of size evolution. In these conditions, the water oxidation of aged Si QDs would be a quite low growth process with a rate in the order of ~1 nm/h.

Finally, the interaction of Si QDs with UV illumination demonstrates the high reactivity of luminescent Si QDs. Its potential interaction with the surrounding during exposure should be clarified by further experiments especially the role of H+ that may be unavoidable in biological environments.

Conclusions

Air-aged and freshly synthesized Si QDs powders dispersed in alcohol were heated in the presence of water for different durations. For both dispersions, the oxidation occurs gradually until full transformation of the initial material in SiO2, as reflected by the PL at 430 nm. However, for long oxidation durations, air-aged powder dispersions show a tendency to agglomeration that is not observed for non-aged powders. From the luminescence point of view, the oxidation leads to the appearance of a typical SiO2 defect luminescence for both dispersions. In air-aged dispersion, PL of silicon can be tuned from 790 to 700 nm. The results obtained under UV illumination indicate that the UV illumination affects the QDs radiative properties. Several mechanisms involving radical chemistry can be invoked. Treatments made with acidic water lead us to assume that H+ may also be a passivation agent. This point should be taken into account for further studies on luminescent Si QDs applications.

Copyright information

© Springer Science+Business Media B.V. 2009