Crystalline silicon (Si) nanoparticles present an extremely promising object for bioimaging based on photoluminescence (PL) in the visible and near-infrared spectral regions, but their efficient PL emission in aqueous suspension is typically observed after wet chemistry procedures leading to residual toxicity issues. Here, we introduce ultrapure laser-synthesized Si-based quantum dots (QDs), which are water-dispersible and exhibit bright exciton PL in the window of relative tissue transparency near 800 nm. Based on the laser ablation of crystalline Si targets in gaseous helium, followed by ultrasound-assisted dispersion of the deposited films in physiological saline, the proposed method avoids any toxic by-products during the synthesis. We demonstrate efficient contrast of the Si QDs in living cells by following the exciton PL. We also show that the prepared QDs do not provoke any cytoxicity effects while penetrating into the cells and efficiently accumulating near the cell membrane and in the cytoplasm. Combined with the possibility of enabling parallel therapeutic channels, ultrapure laser-synthesized Si nanostructures present unique object for cancer theranostic applications.
The interaction of inorganic nanomaterials with biological systems has recently become a hot topic due to a variety of attractive applications in biomedicine. The interest to these materials is caused by a series of their unique properties (optical, photochemical, magnetic, electrical, mechanical etc.), which promise a drastic improvement of the current state-of-the-art imaging and therapeutic modalities1,2. In particular, some semiconductor-based nanostructures called quantum dots (QDs) can emit photoluminescence (PL) with much improved characteristics compared to organic fluorophores, including a wider absorption band, longer fluorescence lifetime, good photostability and spectral tuneability of the PL band etc., making them promising candidates for spectral multiplexing with a variety of biological imaging applications3,4. CdSe/CdS and some others compound QDs have been in the focus of many studies due to high quantum yield (QY) of their PL, which can exceed 50% for some structures3. QDs could be successfully conjugated to peptides5, proteins6 and DNA7 and tested in multi-color fluorescent imaging in cellular and animal models8,9. However, almost all conventional QDs cause critical toxicity problems. As an example, CdSe-CdS QDs provide extremely toxic Cd2+ ions under the interaction with biological environment. Although the short-term toxicity of QDs conjugates can somewhat be reduced by the proper polymeric or inorganic coating10, such nanostructure compounds are hardly accessible for clinical practice.
Silicon-based nanostructures present a viable alternative to compound QDs and they could offer a solution of the toxicity problem. Chemically pure silicon (Si) is a unique inorganic material, which is not only low toxic11, but also biodegradable as in biological environment it transforms into orthosilicic acid Si(OH)4, which is naturally excreted from the body with the urine12. Si nanostructures are also known as promising theranostic agents, which can simultaneously combine therapeutic and diagnostic (imaging) functionalities (theranostics = therapy + diagnostics). Indeed, from one side Si-based nanostructures can demonstrate efficient exciton PL with tunable emission band inside the human body transparency window (750–900 nm). In this case, a much longer emission lifetime (1–100 μs) of this PL band13 compared to tissue autofluorescence signals makes possible the development of imaging modalities based on time-gated suppression of noises14. On the other hand, Si nanostructures can be used as sensitizers for photodynamic therapy (PDT) to efficiently generate singlet form of molecular oxygen under photoexcitation and thus treat malignant tumors15,16,17. In addition, Si nanocrystals can sensitize hyperthermia, i.e. local heating of a tumor tissues above 42 °C leading to an efficient destruction of cancer cells, under their irradiation by infrared radiation18, ultrasound19 and radio-frequency waves20.
It is widely accepted that to obtain water-dispersible brightly luminescent Si QDs, one has to fabricate small (2–4 nm) Si nanocrystals with an appropriate passivation (coating) of their surface to remove non-radiative recombination centers21,22,23,24,25. The latter procedure usually requires a “hydrogenation” step by using solutions of hydrofluoric acid (HF) or a mixture of HF and nitric acids (HNO3), which inevitably leads to a surface contamination by acid derivatives drastically enhancing toxicity of the QDs (such toxicity arising as a result of the fabrication procedure can be called “secondary toxicity”). Here, a straightforward approach implies a mechanical milling of porous silicon, which is preliminarily produced on a crystalline Si (c-Si) wafer by anodic etching in HF/ethanol solutions and then oxidized by storage in air16,18 or ethanol13,17. Such a combination of hydrogenation and oxidation processes provides a hydroxyl-based passivation of the nanocrystals and the upper surface layer SiOyHx (where y = 1…2; x = 0…2) ensures a good dispersion of QDs in aqueous solutions. In principle, brighly luminescent Si nanocrystals can be produced by much cleaner “dry” fabrication methods, including laser or plasma pyrolysis of silane26,27,28,29,30,31, thermal annealing of amorphous SiOx films23,24,27, or laser ablation from a c-Si target32,33,34. However, the prospects of such nanocrystals as bioimaging markers are not clear, as the strong PL of solid nanostructured films does not necessarily guarantee a high PL efficiency of those nanocrystals after the dispersion in aqueous media. As an example, nanocrystals synthesized by the pyrolysis22,26,27 and thermal annealing of amorphous SiOx films23,24,27 typically need to be subjected to a wet chemistry etching step in HF or HF-HNO3 solutions to spatially separate the nanocrystals, decrease their size, and condition the required hydroxyl-based passivation in order to obtain the desired PL emission. As a result, the secondary toxicity of QDs could not be avoided.
Here, we report on a solution of the secondary toxicity problem of brightly luminescent water-dispersible Si-based QDs by employing a laser-assisted synthesis. Formed by laser ablation from a solid c-Si target in gaseous He and grown in laser-plasma environment, Si nanoclusters experience nearly perfect crystallization, followed by surface passivation by oxygen during their exposure to air (oxide passivation). After the dispersion in aqueous solutions, the oxide-passivated Si QDs exhibit the strong exciton-based PL with QY of several percent without additional wet chemistry procedures. The QDs do not show any sign of toxicity, demonstrate biodegradability and excellent cellular uptake, which makes them ideal candidates for bioimaging applications.
Results and Discussion
The laser-ablative synthesis consists of two steps. As the first step, we use a technique of pulsed laser ablation of a solid c-Si target in gaseous ambience to deposit a thin nanostructured Si-based film on a substrate. Briefly, a c-Si wafer is irradiated by UV radiation of a KrF excimer laser in residual He gas (see Methods Section for details). The laser irradiation leads to ablation of material in the form of atoms and smallest nanoclusters, which move perpendicularly to the target surface, as shown schematically in Fig. 1a. He gas under the pressure of 1–5 Torr is used to finely control the growth and crystallization of nanoclusters: collisions of the nanoclusters with He atoms lead to their cooling, condensation and crystallization in the vapor phase32,33. The nanocrystals are then deposited on optically polished substrates of c-Si wafers, which are placed 2–3 cm from the target, to form a nanostructured film with the thickness of 1–5 μm (Fig. 1a). As shown in Fig. 1b, such films are highly porous with the estimated porosity of 65–75% (see details in Supplementary Information, Fig. S1). The films also contain hemispherical μm-size droplets due to the ejection of large target fragments, but these droplets do not affect the integrity and optical quality of the films. It is worth noting that the laser-ablated nanostructured Si-based films can be deposited on an arbitrary substrate (including a gold surface) and used in various biosensing configurations, including Surface Plasmon Resonance35,36. As the second stage, we perform an ultrasound-assisted breakage (sonification) of the laser-deposited films in deionized water or physiological saline (see Methods Section). The nanostructured porous films are detached from the substrate as a result of this sonification and dispersed in the solution forming a colloidal Si nanocrystal-based suspension.
Figure 1c presents a typical transmission electron microscopy (TEM) image of Si nanoparticles (NPs) obtained after the sonification step and related size distribution of the NPs before any size filtering (inset). One can see that despite the presence of big agglomerates the majority of laser-ablated Si NPs (LA-Si NPs) have a diameter less than 100 nm, which makes them suitable for biomedical applications. Moreover, large LA-Si NPs present aggregations of small nanoparticles (Fig. 1d) and can be disintegrated during their storage in aqueous media (Fig. 2c). Fine structure of typical LA-Si NPs can be seen in the HRTEM image of Fig. 1e. It is visible that these LA-Si NPs are composed of randomly distributed crystalline grains (denoted by orange circles) incorporated into a porous matrix. Such a matrix is mostly composed of amorphous SiOx, as it was earlier evidenced33,37,38. As follows from the FTIR data (see Fig. S2 in Supplementary Information), the upper layer composition is close to silicon dioxide, i.e. x = 1.95 ± 0.05, but deeper layers evidence the presence of a fraction of the suboxide phase close to SiO1.538. This could be interpreted as a non-uniform coating, implying nearly SiO2 compound at the surface and SiOx (x < 2) for deeper layers, where x decreases with the increase of distance related to the crystalline Si core. According to NPs size distribution shown in the inset of Fig. 1e, the mean crystal size is 2.5 ± 0.5 nm, while QDs with sizes of 5–8 nm are also present. It should be noted that such small nanocrystal sizes are consistent with the manifestation of the quantum confinement effect in semiconductor nanostructures13.
The laser-ablated films exhibit strong PL signals just after their exposition to ambient air. As shown in Fig. 2a, the recorded PL band is spectrally centered at 1.5 eV (810 nm). While the band position is commonly explained by the quantum confinement of charge carriers in Si QDs13, the PL origin can be attributed to the radiative recombination of excitons either in the whole volume of QDs13,39 or on Si-O bonds at the SiO2/Si interface40. Note that the position of 1.5 eV exciton band corresponds to the average nanocrystal size 2.5 ± 0.5 nm, which is in good agreement with previous data on exciton PL of Si QDs embedded in a SiO2 matrix41. It is important that the exciton PL band does not disappear after the sonification of the LA-Si films and their subsequent dispersion in pure water or physiological saline (Fig. 2a). The storage of NPs in aqueous media leads to a slight blue-shift of the exciton PL band to 1.6 eV, which is probably explained by a partial oxidation of LA-Si NPs, and consequently, to a decrease of the size of pure Si core. The storage of LA-Si NPs in aqueous environment also leads to the appearance of the second PL band located at 2.7 eV, which is usually attributed to the electronic states of point defects in the SiOx phase41,42. The most relevant candidates for the origin of blue band are neutral oxygen vacancies (≡Si-Si≡), which are molecular-like centers in silicon-rich or oxygen-deficient silicon oxides43,44. It is also important that the observed PL bands have quite different time decay dependences. Whereas the time decay of the defect-related band is very fast (tens of nanosecond for the 10-fold decrease of PL intensity), the relevant parameter for the exciton band is much slower (tens of μs), which is in agreement with previous studies of the exciton emission in porous silicon13 and nanostructured Si-based films23,24. The external QY of the spectrally integrated PL was found to be about 5% and 3% for the films and aqueous suspensions of LA-Si NPs, respectively. It should be noted that such QY values are orders of magnitude higher than in the case of Si NPs prepared by laser ablation in deionized water45,46.
The efficient PL from laser-ablated films and suspensions of Si QDs is a pleasant surprise, taking into account that Si nanostructures produced by most alternative dry methods such as laser pyrolysis of silane1,22,26 do not provide luminescence before an additional wet chemical treatment step in solutions of HF-HNO3. We believe that that the formation of strongly luminescent Si QDs in our case is due to particular conditions of crystal growth in laser-plasma plume and their subsequent passivation in ambient air. It is known that the plume of ablated Si nanoclusters is capable of efficiently ionizing atoms of ambient He gas and thus form plasma of a relatively long (>1 ms) lifetime47. Visible by a naked eye due to characteristic luminescence emission, such a region of the ionized gas plasma can reach the distances of 1.5–2 cm from the target. Although the nanoclusters move with a relatively high velocity (~10 m/s)47, they can get a significant amount of energy while propagating through the long living plasma before reaching the substrate. Such energy gain can be due to collisions of Si nanoclusters with He ions or hot electrons, which have a significant excess of energy of the order of tens eV in laser plasma47. It is known that even a single electron-ion recombination event in non-equilibrium plasma can increase the temperature of a small nanocluster up to 1000 K48. Therefore, we can expect a significant heating of laser-ablated Si nanoclusters during their propagation through the plasma of the ionized gas. We suppose that a subsequent cooling of the ultra-hot clusters under their collisions with cold He atoms (outside the plasma area) creates conditions for the formation of nearly-perfect low-defect Si QDs. Thus, the formation of high-quality nanocrystals requires the involvement of heating and cooling stages before the nanocrystals reach the substrate. As it was found in ref. 33, such conditions are possible under a relatively narrow range of He pressures. The optimal pressure depends on the laser energy and target-substrate distance, but for our experimental parameters it was typically in the range of 0.5–5 Torr. It is interesting that within this pressure range the nanocrystal size is almost independent of the pressure, while the film morphology is critically sensitive to this parameter. In particular, the increase of He pressure from 0.5 to 4–5 Torr leads to gradual increase of the porosity of the laser deposited films from 20% to 90%33. Here, for relatively high pressures of 3–4 Torr the substrate is covered by highly porous powder rather than films, suggesting an efficient crystallization of Si nanoclusters in the vapor phase before reaching the substrate. In our experiments pressures around 2 Torr were optimal from the point of view of maximization of PL signals. We believe that such a pressure provides optimal conditions for heating and cooling processes in the formation of Si QDs.
The oxygen–based passivation of the laser-ablated films is another important process to achieve a bright emission of Si QDs. In our experiments, the passivation of Si QDs by Si-O bonds takes place just under the exposition of the laser-ablated films to ambient air. As we showed in ref. 33, the PL intensity is further 5–10-fold enhanced during a subsequent storage of the crystals in dry air, which suggests an improved passivation due to the growth of SiOx shell. Since the laser-ablated films and NPs consist of networks Si QDs (see Fig. 1d,e), their PL intensities should be dependent on the exciton migration between neighboring Si QDs. Here, the exciton migration can result in the decrease of total PL QY and such a process is probably controlled by non-radiative defects in large Si QDs and SiOx shell24. On the other hand, the single exponential decay of the exciton PL intensity in the time scale above 1–5 μs after pulsed photoexcitation (see Fig. 2b) evidences a nearly perfect passivation of certain fraction of brightly luminescent Si QDs23. The surface passivation of laser-synthesized Si QDs by the Si-O bonds does not look very stable in aqueous ambience that is illustrated by the decrease of the PL QY and appearance of additional PL band around 2.7 eV (see Fig. 2a). This effect is probably related to a well-known dissolution property of Si NPs in aqueous media. However the dissolution time of LA-Si NPs having a native surface oxide in neutral pH solutions can last for several days20, which still makes possible efficient bioimaging studies.
To assess the potential of the prepared LA-Si NPs for biological imaging tasks, we carried out a series of tests on the incubation of nanoparticles in a cellular model. Figure 3 shows confocal fluorescent microscopy images of cancer cells with added LA-Si NPs under different magnifications. Cell nuclei are colored blue and the cytoplasm is colored green. Figure 3a shows the cells in different proliferation states, including mitotic cells in the metaphase. The last ones can be distinguished by metaphase plates, which look as blue rods substituting normal nuclei. These data evidence normal cell proliferation in the medium containing LA-Si NPs (red or red-blue spots) that confirms excellent biocompatibility of the prepared Si QDs. Figure 3b,c represent detailed views of cells before and after washing out of the nutrient solution with dispersed LA-Si NPs, correspondingly, and Fig. 3d represents a control group without NPs. It is visible that most LA-Si NPs emit both red and blue PL bands, which are in good agreement with their PL spectra in water (see Fig. 2b). In the images of Fig. 3 one can distinguish several groups of luminescent NPs: (i) free-floating NPs outside the cells; (ii) NPs stuck on the cell membrane; (iii) NPs penetrated into the cell cytoplasm; (iv) NPs concentrated near the cell nuclei. The observed variety of NPs locations evidences a non-selective mechanism of their penetration into the cells, i.e. endocytosis. Note that LA-Si NPs located near the nucleus membrane are hardly able to penetrate inside the nuclei, because the size of NPs is larger than the pore size of the nucleus membrane. The penetration of LA-Si NPs inside the cells is more easily seen in Z-scan imaging (see video file in Supplementary Information). The poor penetration of LA-Si NPs into the cell nuclei can be important to ensure their low genotoxicity similar to what was observed for NPs formed from porous Si49.
Thus, the incubation of Si QDs into living cells does not provoke any toxicity effects, while the NPs easily penetrate into the cells and concentrate in different cell regions except the nuclei. The presence of the NPs can be efficiently tracked by the red and blue emission of the PL bands of Si QDs.
In conclusion, we prepared and investigated laser-ablated Si nanoparticles composed of small Si QDs as novel contrast agents for photoluminescent bioimaging. The nanoparticles were produced by methods of the pulsed laser ablation from a c-Si target in gaseous (He) ambience, followed by ultrasonic grinding of the laser-deposited films in aqueous solutions. By following the photoluminescence signals from Si QDs we evidenced the excellent uptake of Si nanoparticles by cancer cells and their efficient accumulation in different cellular regions. The employment of such QDs does not reveal any sign of residual cytotoxicity, making them extremely promising candidates for biological imaging tasks.
We believe that the laser plasma-assisted growth of small, low defect nanocrystals and their subsequent oxide passivation in ambient atmosphere is a very promising strategy for the creation of brightly luminescent, water-dispersible and non-toxic Si QDs. In the absence of wet chemistry step the Si QDs are exposed only to a clean environment (residual He gas, air, physiological solutions) and should have ultraclean surface. Although the photoluminescence QY of Si QDs (typically 3–5%) is slightly lower compared to Si-based nanostructures prepared by wet chemistry methods, it is sufficient to obtain efficient contrast in cellular or tissue imaging. In addition, the exciton emission appears to be exactly in the window of relative tissue transparency, which simplifies the implementation of imaging configurations. Another advantage is related to the μs-scale time decay of the PL signal from LA-Si NPs. Indeed, one can profit from such a long time decay to suppress much faster autofluorescence signals from biological molecules in the time-gated regime, as it was demonstrated in vivo in ref. 14. In fact, such lifetime multiplexing can increase signal-to-noise ratio by the factor of k, which is equal to contrast agent lifetime divided to the lifetime of autofluorescence (typically 1…50 ns). In our case the factor k can reach 1000. Finally, the prepared LA-Si NPs are water-soluble as it was revealed by a series of the DLS, TEM and Raman spectroscopy tests. As an example, the DLS analysis showed a six-fold decrease of the NPs size just after several days of nanoparticle storage in physiological solutions (Fig. 2c). Raman spectroscopy additionally revealed an increase of the amorphous Si phase up to 80% after 11 days of storage in aqueous suspension (see details in Supplementary Information, Fig. S3), which is obviously related to an efficient dissolution of Si NPs. This process excludes possible long-term toxicity (including genotoxicity) of the laser-ablated Si NPs. It should be noted that conditions of nanocrystal growth (plasma temperature and time scale of processes) in our case are similar in many respects to experiments on RF-decomposition of silane, which can also lead to the formation of brightly photoluminesent nanostructured Si films without any additional wet chemistry step50. Although we may not conclude on a complete similarity of these two cases, we can guess that that such RF radiation-formed Si nanocrystals can also be sonificated and water dispersed to obtain bright mobile QDs for bioimaging.
At the first step, we used a conventional geometry of pulsed laser deposition in gaseous ambience. The radiation of a pulsed KrF laser (wavelength: 248 nm, pulse length: 17 ns, repetition rate: 10Hz) was used for the ablation of material from a rotating c-Si target ((100)-oriented c-Si wafer of n-type, specific resistance of 10 Ohm·cm)33. The radiation was focused on a focal spot of 2 mm2 on the target at the incident angle of 45° giving the radiation intensity of about 5*108 W/cm2. c-Si wafer-based substrates, identical to the target, were placed on a rotating substrate holder at 2 cm from the target. The experimental chamber was pumped down to residual pressure of 10−7 Torr before filling with helium (purity 99.9995%) for a deposition at a constant pressure about 1–2 Torr. The film thickness after ten thousands laser shots was about 1 μm.
Aqueous suspensions of LA-Si NPs were obtained by ultrasonic treatment of the laser-ablated films in deionized water or saline (0.9% NaCl in H2O) for 1 h. The power density and frequency of ultrasound were 5 W/cm2 and 44 kHz, respectively.
Scanning Electron Microscopy (SEM) images were obtained by using a Tescan Lyra 3 XM microscope with resolution of 1.2 nm and accelerating voltage of 30 kV. Transmission electron Microscopy (TEM) images were obtained by means of a Zeiss Libra 120 microscope with resolution of 0.5 nm and accelerating voltage of 120 kV. High resolution TEM images were provided by a JEOL JEM-2100F microscope with resolution of 0.8 Å and accelerating voltage of 200 kV. Size distributions of NPs and QDs were calculated by using ImageJ software.
PL spectra were measured by using a SOLAR spectrometer equipped with a CCD-unit from Hamamatsu. All spectra were corrected for the spectral response of the measurement systems. The PL transients were detected by a R928 photomultiplier tube (Hamamatsu Photonics, Hamamatsu, Shizuoka, Japan) under pulsed laser irradiation by a nitrogen laser (excitation wavelength of 337 nm and pulse duration of 10 ns). Time response of the detection system was shorter than 1 μs. PL quantum yield was measured by comparing the PL intensity and absorption of the samples with solutions of Rhodamine 6 G (PL QY about 100%).
To minimize noises in Raman measurements, we deposited a droplet of aqueous LA-Si NPs solution on a stainless steel samples. In addition, in a separate experiment we deposited nanostructured Si layers on CaF2 substrates. Raman spectra were measured by using a micro-Raman spectrometer from Horiba Jobin Yvon, excitation wavelength was 488 nm, maximum excitation power density was 10 W/cm2, spectral resolution was 0.1 cm−1. In order to avoid unfavorable heating a set of attenuating filters (0.3, 0.6, 1 and 2 dB) was used. Size distributions of NPs were measured by using a dynamic light scattering (DLS) Zetasizer ZS from Malvern. Porosity of LA-Si films and composition of LA-Si NPs were studied with a Fourier-transform infrared (FTIR) spectrometer Bruker IFS 66v/S. Before measuring the FTIR spectra, the suspensions of LA-Si NPs were deposited on an ATR crystal and then dried in air and evacuated at 10−3 Torr.
In vitro bioimaging experiments were carried out with CF2Th (dog thymus) cells infected with a green fluorescent protein (GFP) gene. The latter was induced by RSL-1 inducer added to the CF2Th culture 30 h prior the bioimaging analysis. The GFP was characterized by a fluorescence line at 515 nm (green light). LA-Si NPs were introduced into the cell culture 5 h after the injection of the inducer and 25 h prior the experiment. 30 min prior the experiment, the cell nuclei were imbued with 5 mg Hoechst, which was luminescent near 460 nm (blue light). The cells containing LA-Si NPs were studied using a Leica TCS SP5 confocal fluorescent microscope with multicolor illumination at 488 nm, 496 nm, 514 nm, 543 nm and 633 nm to ensure simultaneous excitation of both the Si QDs and cells colored with GFP and stained with the Hoechst dye.
How to cite this article: Gongalsky, M. B. et al. Laser-synthesized oxide-passivated bright Si quantum dots for bioimaging. Sci. Rep. 6, 24732; doi: 10.1038/srep24732 (2016).
MS and AVK acknowledge the support of the AMIDEX project (no. ANR-11-IDEX-0001-02) funded by the ‘Investissements d’Avenir’ French Government program, managed by the French National Research Agency (ANR) of the LASERNANOBIO project (ANR-10-BLAN-919) of the ANR, of the LASERNANOCANCER project of the ITMO “Plan Cancer 2014–2019” of INSERM (No. PC201420), and by CNRS PICS project (N6577). MBG, LAO and VYuT acknowledge the support of the Russian Foundation for Basic Research (Grant No. 15-52-15041). ANV acknowledges the support of the Russian Science Foundation (Grant No. 14-50-00029). MBG is thankful to A.V. Pavlikov for assistance with FTIR measurements.