Project Topics
Get Complete Project Material File(s) Now! »
Safer-by-design concept
To address the potential harms of nanomaterials while maintaining their development, safer-by-design concept has been employed recently. First appeared in the literature around 2010, it consists in integrating knowledge of potential adverse effects into the process of designing nanomaterials, and engineering these undesirable effects out of them. For this, both hazard which is the potential of a nanomaterial to cause harmful effect, and exposure which describes a certain dose that human or environment encounter over time, may be considered.34–36
Although safer-by-design is relatively new in this field, similar concepts such as green chemistry, quality-by-design, inherent safer design, etc. have been well-established in fields such as chemistry, construction, nuclear technology, aeronautics, water treatment, pharmaceutics, health facilities, etc. since the end of the last century.34,37–39 In those fields, the concept describes safety measures for the prevention of accidents, illnesses or environmental damage that are applied during the design stage of a facility, process, practice, material or product.
In general, safer-by-design is not new and has been used for years in the industry. Various fields have adopted and developed this concept differently, but the common idea is to design products or processes that possess an intrinsically low risk potential, instead of confining this potential by application of protective systems.34
Recent examples of safer-by-design approaches
In the literature, several safer-by-design approaches were reported, mainly to reduce potential hazard of nanomaterials. Most of these studies were based on aforementioned paradigms and focused on modification of either the inorganic core or the surface of the nanomaterials.
Modification of inorganic core
One of the first attempts to reduce nanomaterials hazard by a safer-by-design approach was through modification of inorganic core of nanomaterial. In 2011, Xia et al. demonstrated that the dissolution of zinc oxide ZnO nanoparticles and the Zn2+ release, that led to a series of sub-lethal and lethal toxicological responses at the cellular level, could be lessened by iron-doping.40 ZnO nanoparticles doped with up to 10 at.% of Fe were synthesized by flame spray pyrolysis. They were spherical with size ranging from 8 nm to 20 nm. It was suggested that the presence of iron as dopant reinforced the atom bonding in the particles and decreased the Zn2+ dissolution rate, thus the ZnO nanoparticles toxicity. Concrete evidence regarding the bonding reinforcement was however lacking.
Similar to ZnO nanoparticles, copper oxide CuO nanoparticles are known to exhibit toxicity due to the Cu2+ dissolution in biological media. Employing the same iron-doping strategy, Naatz et al. showed that the toxicity of CuO nanoparticles could also be reduced.41 Fe-doped (up to 10 at.%) CuO nanoparticles of size from 10 to 20 nm were synthesized by flame spray pyrolysis. The incorporation of Fe led to distortion of the monoclinic CuO lattice (Figure I.6). In fact, CuO has a square planar geometry with each Cu coordinated to four oxygen atoms. By doping with 10% of iron, the Cu-O bond lengths increased from 2.02 to 2.19 Å. The authors hypothesized that this lattice distortion enhanced the material stability. However, further explanation was not provided. Doping above 6 at.% of iron also resulted in the formation of a CuFe2O4 phase on the surface of the nanoparticles (Figure I.6). It was demonstrated that the presence of CuFe2O4 phase greatly inhibited the dissolution of CuO nanoparticles in various cell culture media. Finally, cytotoxicity of the nanoparticles towards human macrophage-like THP-1 and epithelial BEAS-2B cell lines, as well as their interference with zebrafish hatching were reduced.
Figure I.6: Structure-activity relationship describing the structural and chemical properties of pristine and iron-doped CuO nanoparticles, and their influence on the biological response assessment (taken from ref 41).
It should be noted that doping is not a universal safer-by-design strategy for all types of nanomaterials. In some cases, doping was shown to enhance toxicity. Indeed, George et al. showed that doping of titanium oxide TiO2 nanoparticles with Fe3+, using the same flame spray pyrolysis method, reduced their bandgap from 3.3 eV to 2.8 eV (Figure I.7a).42 As a result, their absorption spectrum was shifted towards the visible region, making them more suitable for photocatalysis with visible light. However, the oxidation ability of Fe-doped TiO2 upon near-visible irradiation (below 3.2 eV) was also found to be superior to that of undoped nanoparticles due to their lower bandgap. Doped TiO2 nanoparticles induced higher oxidative stress and cell death to macrophages-like RAW 264.7 cells (Figure I.7b).
In this example, even though the tailored design strategy led to enhanced functionality of the nanomaterial, it also increased its toxicity. Nevertheless, the acquired knowledge on the potential toxicity of iron-doped TiO2 nanoparticles remains useful for future development of novel nanoscale photocatalysts containing TiO2. Alongside previously discussed examples of safer-by-design ZnO and CuO nanoparticles, this study also highlighted the importance of integrated effort of nanomaterials synthesis and safety assessment in safer-by-design strategies.
Figure I.7: a) UV-visible absorption of undoped and Fe-doped TiO2 nanoparticles represented by Kubelka-Munk reflection plot over energy. b) Cytotoxic responses of macrophages-like RAW 264.7 cells exposed to the nanoparticles and near-visible illumination: mitochondrial superoxide generation (MitoSox), decreases in mitochondrial membrane potential (DMMP) and loss of cell viability (PI uptake) (adapted from ref 42).
Modification of surface
Beside modification of the inorganic core, surface passivation is another strategy to tackle the dissolution-induced toxicity of nanoparticles such as zinc oxide ZnO. Lewinski et al. recently prepared 5-nm ZnO nanoparticles– for photoluminescence properties.43 They were coated with a densely packed shell of 2 (2 methoxyethoxy)acetate (MEAA) ligands, which acted as polyethylene glycol (PEG) prototype, using an one-pot, self-supporting organometallic procedure (Figure I.8 Right). This ligand shell passivated the nanoparticles and protected them from subsequent Click Chemistry on their surface, retaining their photoluminescence properties. It was not the case for ZnO nanoparticles prepared by traditional sol-gel methods (Figure I.8 Left).44 The authors also showed that the MEAA ligand shell inhibited leaching of Zn2+ ions from the inorganic core and generation of free radicals on the surface. It was suggested that the well-passivated ZnO nanoparticles induced negligible oxidative stress and adverse effect on human lung fibroblast MRC-5 and human lung epithelial A549 cells.43
Although direct comparison with ZnO nanoparticles prepared by traditional sol-gel route was lacking, this study highlighted the surface passivation strategy to reduce nanomaterial toxicity.
Figure I.8: Comparison of characteristic properties and toxicity of zinc oxide ZnO nanoparticles prepared by traditional sol-gel method and those prepared by novel organometallic self-supporting approach (taken from ref 43).
Another example of a safer-by-design approach relying on surface modification of the nanomaterials is the case of multi-walled carbon nanotubes (MWCNTs). They were shown to cause fibrogenic effects in cells and animal lungs due to their surface hydrophobicity and high aspect ratio.45 Wang et al. demonstrated that effective surface coating of the nanotubes by biocompatible copolymers such as pluronic F108 (PF108) protected them from damaging the lysosomal membrane and initiating a sequence of cooperative cellular events that play a role in the pathogenesis of pulmonary fibrosis (Figure I.9).46 The author suggested that PF108 coating was a promising approach for designing safer MWCNTs.
Figure I.9: Comparison between bovine serum albumin (BSA) and pluronic F108 (PF108) as dispersants for multi-walled carbon nanotubes (MWCNTs) in terms of stability, interaction with lysosome membrane as well as pro-fibrogenic effects (taken from ref 46).
A reviewed vision of safer-by-design approach for nanomaterials
In the past ten years, several safer-by-design strategies were successfully implemented to prepare safer metal oxide, carbon-based nanomaterials among others. They mainly consist in modifying either the inorganic core of the nanomaterials or their surface. The examples demonstrated the essence of nanomaterials safer-by-design, i.e. determining the key physicochemical characteristics that contributed to toxicity and how to remediate them by rational design. Nevertheless, it is clear that a gap between materials research and safety assessment still exists. While most of safety studies and safer-by-design approaches have been focused on raw materials, the materials research field has evolved beyond individual nanomaterials to create sophisticated nanostructures and devices. Moreover, these studies mainly targeted well-developed nanomaterials, some of them are already present in commercial products. Current safer-by-design approaches also remain heavy on reducing the potential hazard of nanomaterials. Evaluation of intended functionality after safer-by-design modifications was often neglected. The aforementioned works on safer-by-design copper oxide CuO and zinc oxide ZnO nanoparticles are prominent examples.40,41 The authors mentioned various functionalities of both nanoparticles such as pigment and UV-absorber. However, these functionalities were not studied after the iron-doping of the nanoparticles to reduce their toxicity.
Taking into account these points, in order to take full advantage of the safer-by-design approach and maximize the potential of nanomaterials, we believe that safer-by-design could be applied in the earliest stage of nanomaterials development, i.e. to novel, emerging nanomaterials that are still in basic research phase. Furthermore, both their toxicity and functionality for a targeted application need to be considered in parallel. For emerging nanomaterials, little is known about their potential risks. Thus, determining the key physicochemical properties that contribute to both the toxicity and the functionality is of particular relevance. Altogether, we propose that the properties, toxicity and functionality of the emerging nanomaterials are to be studied in parallel. These three points can be schematically linked by a feedback loop (Figure I.10). By applying this safer-by-design feedback loop over and over, we hope to improve both the functionality and the safety of the nanomaterials by well-designed incremental modifications. As a consequence, it is expected that effects of each modification can also be rationalized.
Lastly, as the nanomaterials evolve through their life cycle, ideally, the safer-by-design concept needs to be applied at every step of nanomaterials development, from basic research to production, uses and end of life to ensure their safety.
Figure I.10: Schematic representation of safer-by-design approach for emerging nanomaterials.
Thesis outline
The purpose of this thesis is to apply the reviewed safer-by-design concept discussed above to bimetallic gadolinium-cerium oxysulfide nanoparticles of compositions Gd2(1-x)Ce2xO2S.
Despite their promising applications in catalysis and biomedical imaging,47,48 studies on toxicity of nanoscale oxysulfides remain scarce, making them a perfect model of emerging nanomaterials.
The following chapter (Chapter II) will be dedicated to the synthesis and characterization of bimetallic gadolinium-cerium oxysulfide nanoparticles. The nanoparticles are prepared by colloidal synthesis in organic solvents and their complete description (size, shape, composition, crystalline nature, surface ligand and surface state) is presented. In Chapter III, the safer-by-design approach is discussed in relation with oxysulfide nanoparticles design for photocatalysis. Following attempts to remove surface ligand, the tunable absorption properties of Gd2(1-x)Ce2xO2S nanoparticles are presented. Their photocatalytic activity under visible light is demonstrated via tests of photodegradation of organic dyes and radical production. Alongside the functionality, the toxicity of oxysulfide nanoparticles is assessed by a combination of in vitro (murine macrophage RAW 264.7 cell line) and in vivo (C57 black 6 mice) models. Cell viability, oxidative stress and inflammation response are the main discussed endpoints. In order to elucidate the cell-nanoparticle interaction and possible origins of toxicity, cells exposed to nanoparticles are finally mapped by X-ray hyperspectral imaging. Chapter IV shows preliminary results of oxysulfide nanoparticles developed for fluorescent-magnetic biomedical imaging. The photoluminescence of Eu-doped Gd2(1-x)Ce2xO2S nanoparticles are first discussed. Then, the doped nanoparticles are coated with polyvinylpyrrolidone (PVP) for water-dispersibility. Cytotoxicity of the PVP-coated and Eu-doped is finally characterized. The last part of the manuscript summarizes the results of the three last chapters and proposes perspectives for the safer-by-design approach applied to emerging nanomaterials.
Definition of metal oxysulfide
Metal oxysulfides are compounds composing of at least one metal, in addition to oxygen and sulfur both at their reduced state. The generic formula for metal oxysulfides is MxOySz. It should be emphasized that oxygen and sulfur must adopt negative oxidation state (e.g. -I or -II) for the compound to be called “oxysulfide”. This is to make a clear distinction with other compounds containing oxygen and sulfur such as metal sulfates Mx(SO4)y, where the sulfur is oxidized and at positive oxidation state +VI. In the literature, one can find reported oxysulfides under other names such as “thiooxide” and “oxidesulfide”.1–3 However, the term “oxysulfide”, proposed by Flahaut et al. in 1958 mostly for Ln2O2S (Ln = lanthanide),4 is now employed in a large majority of the works on compounds containing both reduced oxygen and sulfur.5
Properties and applications of metal oxysulfides
Metal oxysulfides are scarcer than their oxide and sulfide counterparts and are mostly artificial. They were often considered undesired byproduct in sulfidation of metal oxides. First discovered in 1827 by Carl Gustaf Mosander in cerium oxysulfide CexOxSz, other oxysulfides with transition metals, lanthanides and actinides have also been reported in the bulk as well as in the nanoscale.5 Progression in the field and keystones are shown in Figure II.1.
One of the reasons for this scarceness is likely because oxysulfides are difficult to obtain. According to their definition, both reduced oxygen and sulfur must be present. However, possessing a large number of oxidation states from -II to +VI, sulfur is easily oxidized in compounds containing both elements. This is why reported structures of metal oxysulfides often show alternate layers of metal-sulfur M-S and metal-oxygen M-O (Figure II.1).
The presence of two different types of bonding (M-O and M-S) in metal oxysulfides is an interesting feature. Oxygen and sulfur show a rather large difference in electronegativity: 3.4 and 2.6 respectively by Pauling scale. By consequence, the properties of metal oxide and metal sulfide are quite different. While the M-O bonding in metal oxides has a more ionic character, resulting in them being mostly insulators (e.g. TiO2, ZnO), the M-S bonding in metal sulfides has a more covalent character, giving rise to their semiconductor properties (e.g. CdS). This also leads to the relatively low solubility in water of metal sulfides compared to metal oxides. If one represents the three main types of chemical bonding by an equilateral triangle scheme, most metal oxysulfides should be close to the center (Figure II.2).6
Among metal oxysulfides, lanthanide oxysulfides with generic formula Ln2O2S have been one of the first reported and have been intensely studied up until now. Their applications in the bulk form have been identified since the 1980s as screens and monitors,7,8 scintillators,9 lasers,10,11 etc. For those applications, Y2O2S, La2O2S and Gd2O2S are usually used as matrices and doped with other lanthanides for desired photoluminescence properties.
At the nanoscale, Ln2O2S nanoparticles are potential contrast agents for biomedical imaging due to their good chemical and thermal stability. Moreover, similar to their bulk form, they can also host other lanthanides for a variety of photoluminescence properties (Figure II.3a).12,13 In particular, Yb/Er co-doped lanthanide oxysulfide nanoparticles were used for upconversion luminescence imaging14 while Eu-doped Y2O2S nanoparticles were reported for their persistent photoluminescence properties.15
Figure II.3: a) TEM image of lanthanide-doped Gd2O2S nanoparticles and b) their photoluminescence under 254 nm excitation (adapted from ref 13). c) SEM image of doped Gd2O2S nanoparticles and d) their gray-scaled as well as color-mapped magnetic resonance images at different concentrations (adapted from ref 14).
Lanthanides can also exhibit high magnetic susceptibility, which is of major importance for magnetic imaging techniques such as Magnetic Resonance Imaging (MRI). In fact, Gd2O2S nanoparticles have already been reported as potential contrast agents for MRI due to the 4f7 electron configuration of GdIII in the structure (Figure II.3b).14,16
Other than applications in biomedical imaging, catalysis based on redox reactions using Ln2O2S has also been explored. Notably, Ce2O2S nanoparticles on carbon have been tested for oxygen reduction reaction (ORR)17 while Eu2O2S nanoparticles catalyzed reaction of CO and water to yield CO2 and H2.18
Transition metal oxysulfide
Compared to lanthanide oxysulfides, transition metal oxysulfides are scarcer. Especially ternary metal oxysulfides containing a lone transition metal. Whereas, quaternary oxysulfides containing one transition metal and one lanthanide/rare-earth element are more common.5
In the bulk form or as thin films, ternary transition metal oxysulfides have been mostly studied for their electrochemical properties due to the often-large number of oxidation states of the transition metal compared to lanthanides. It is indeed the case for titanium,19 tungsten20 or molybdenum oxysulfides.21,22 These studies focused heavily on the development of new cathode and anode for lithium-ion batteries. Due to the larger number of phases available, quaternary oxysulfides have also been reported for other applications such as photoconversion,23 superconductivity24 and magnetism.25
At the nanoscale, very few transition metal-containing oxysulfides were obtained. They were mostly with zinc,26–28 cobalt29 and titanium.30 These nanoparticles were mainly exploited for their high surface-to-volume ratio, as the number of studies on catalysis dominated. Ishikawa et al. showed that Sm2Ti2S2O5 nanoparticles (Figure II.4a) exhibited catalytic activity for water oxidation and reduction under visible light (440 nm ≤ λ ≤ 650 nm).30 This is particularly interesting because their oxide counterpart Sm2Ti2O7 only absorbs UV light (λ ≤ 380 nm). By partially replacing oxygen by sulfur, the absorption of the materials shifts to the visible region (Figure II.4b). A similar absorption change was also found in ZnO1-xSx nanoparticles (Figure II.4c).26,27 It was shown that the absorption of the material could even be tuned by modifying the S/O ratio (Figure II.4d) and these nanoparticles showed catalytic activity in photodegradation of methyl orange.26 Other than application in photocatalysis, zinc oxysulfide nanoparticles and cobalt oxysulfide nanoparticles are active in hydrogen evolution reaction (HER) as well.28,29
Lanthanide oxysulfide nanomaterials for safer-by-design approach
As discussed in the first chapter, the essence of the safer-by-design approach is to incrementally modify the material, staying within one family of compound. Two strategies are mainly adopted: modification of the inorganic core and modification of the surface. By changing only one or a few parameters at a time, one can hope to rationalize each modification step. This is of particular importance for fundamental research. Therefore, in order to take full advantage of the safer-by-design approach, it is ideal to work with emerging nanomaterials that are multifunctional and can be easily modified.
In this view, lanthanide oxysulfide Ln2O2S nanoparticles seem quite suitable for the safer-by-design approach. In fact, bulk Ln2O2S phases have well-known crystalline structure which is advantaging for the rationalization process. The Ln2O2S nanoparticles can readily be modified for different properties and applications such as catalysis and biomedical imaging. Moreover, compared to oxide counterparts, the overall bonding in oxysulfide is more covalent. This is expected to result in a lower water solubility of the material, thus potentially a lower toxicity.31
As far as multifunctionality goes, gadolinium oxysulfide Gd2O2S nanoparticles are promising candidates as they have been used as MRI contrast agent or X-ray absorbing materials. Doped Gd2O2S:Ln nanoparticles have been shown to be potential contrast agents for multimodal biomedical imaging.14,16,32–35
In our laboratory, we recently succeeded in synthesizing gadolinium-cerium oxysulfide Gd2(1-x)CexO2S nanoparticles analogous to Gd2O2S.36 The Ce-containing bimetallic nanoparticles are expected to possess even more functions than the monometallic ones due to the distinctive properties of cerium compared to other lanthanides. In contrast to most lanthanides which are trivalent, cerium can be both trivalent CeIII and tetravalent CeIV. The Ce3+/Ce4+ redox couple is responsible for the catalytic and antioxidant properties of cerium(IV) oxide CeO2 nanoparticles.37–39 In particular, Gd-doped CeO2 nanoparticles have been used for therapeutics (antioxidant/MRI) thanks to the antioxidant properties of Ce and the magnetic properties of Gd.39 Therefore, Gd2(1-x)CexO2S nanoparticles may be useful for therapeutics as well.
Furthermore, while bulk gadolinium oxysulfide is white and only absorbs UV light, bulk cerium oxysulfide is deep brown and absorbs visible light.4 We presume that mixed oxysulfide Gd2(1-x)CexO2S nanoparticles also absorbs in the visible spectrum and the absorption is tunable by the Gd/Ce ratio. It should be noted that tunable absorption in the visible region has already been achieved with ZnO1-xSx nanoparticles using the S/O ratio (Figure II.4d).26,27 These features would make Gd2(1-x)CexO2S nanoparticles promising materials for photocatalysis using visible light.
In summary, Gd2(1-x)CexO2S nanoparticles are versatile nanomaterials due to the potential absorption properties, magnetic properties and catalytic properties (Figure II.5). We also assume that they can be doped with other lanthanides for photoluminescent properties similar to analogous Ln2O2S. Moreover, the more covalent bonding in oxysulfides may result in lower solubility in aqueous media, and thus lower toxicity compared to metal oxides. Altogether, Gd2(1-x)CexO2S nanoparticles appear particularly relevant for safer-by-design approach.
Table of contents :
Chapter I – Safer-by-design approach for emerging nanomaterials
Chapter II – Gadolinium and gadolinium-cerium oxysulfide nanoparticles: synthesis and characterization
Chapter III – Towards safer-by-design bimetallic Gd-Ce oxysulfide nanoparticles for photocatalysis
Chapter IV – Potential application of bimetallic Gd-Ce oxysulfide nanoparticles in biomedical * imaging
Conclusion and perspectives
