Abstract
Nanocrystals of ZnO are currently attracting great interest as potential labels for biological applications, such as theranostic devices, due to their luminescent properties and low toxicity in vivo. It has been reported that doping with Mg2+ ions could enhance the luminescence of ZnO quantum dots (QDs). In the present study Mg-doped ZnO QDs were synthesized by a hydrolysis and condensation reaction. Surface modification of the QDs was performed using oleic acid (OA) to hinder their aggregation and to provide them colloidal stability in non-polar environments. Mg2+ ions could be incorporated into the ZnO wurtzite lattice owing to the very close values of the Mg2+ and Zn2+ ion radii. However, the dopant ions strongly influenced the growth and final size of ZnO nanocrystals, as evidenced by time-resolved synchrotron SAXS measurements. The presence of Mg prevented the aggregation of the primary nanoparticles. Doping with Mg2+ ions widened the band gap of ZnO QDs and enhanced their visible luminescence. With increasing proportion of Mg2+ ions, both the absorption and emission spectra experienced a blue shift. The luminescence went through a maximum for a 20 mol% nominal concentration of Mg2+ ions in the reaction medium. The quantum yield (QY) of 20 mol% Mg-doped ZnO colloidal suspension (64%) was about 6 times higher than that of the ZnO suspension (10%). Mg-doped ZnO QDs capped by OA formed stable colloidal dispersions in chloroforme, with strong visible fluorescence (QY=38%), promising for biological imaging.
Introduction
At the nanoscale, the optoelectronic properties of semi-conductor crystals depend on their size. For crystals smaller than twice the Bohr exciton radius (the so-called quantum dots, QDs), the semi-conductor band gap widens as the nanocrystal size decreases, owing to quantum confinement effects. Consequently, the radiative recombination of an electron from the conduction band with a hole from the valence band leads to the emission of a photon whose wavelength can be tuned by changing the nanocrystal size. The creation of an electron-hole pair (or exciton) is induced by the absorption of a photon with energy above that of the band gap, resulting in a broad absorption spectrum. Both the absorption and emission spectra of QDs experience a blue shift with QD decreasing size (1).
Usual QDs are made of CdSe, CdTe, InAs and InP. Depending on their composition and size, their emission ranges from near UV to near IR. They need appropriate surface modification, such as coating with amphiphilic molecules, to be dispersible in an aqueous solution. When functionalized with ligands such as antibodies or peptides, QDs can be used to label different types of cellular targets or detect biomarkers (2, 3). Recently, nanoparticles containing QDs have been investigated as theranostic devices for simultaneous imaging and drug delivery (4). For instance, doxorubicine has been loaded, along with QDs, into micelles, aptamers, and liposomes (5–7).
However, the potential cytotoxicity of first generation QDs is a major concern for biological applications (8–11). In this context, luminescent nanoparticles of ZnO are currently attracting great interest as potential labels for bio-imaging because of their biodegradability and very low toxicity in vivo, although ZnO nanoparticles are able to release Zn2+ ions and produce destructive reactive oxygen species (ROS), as shown by cytotoxicity tests in vitro (12–14). ZnO is an n-type semi-conductor with a wide band gap of 3.37 eV at room temperature and a Bohr exciton radius of ∼2.34 nm. The typical excitonic emission, arising from recombination of photo-generated electrons with holes in the valence band or in traps near the valence band, is thus observed in the UV range. Remarkably, the photoluminescence spectrum of ZnO nanocrystals also displays a broad visible emission, more suitable for biological imaging, which has been ascribed to point defects such as O and Zn vacancies or interstitials and related to surface oxygen-containing moieties, such as OH groups (15). However, the understanding of the exact mechanism for visible emission is still lacking.
Because of the role of defects and surface chemistry, the synthesis technique is expected to strongly influence the luminescence of ZnO. In general, the visible emission of thin films or particles synthesized at high temperature or annealed is very weak whereas nanoparticles prepared by the sol-gel route exhibit stronger defect-induced visible luminescence. A typical sol-gel route is the hydrolysis and condensation of a precursor in ethanolic solution (Zn salt solution) catalyzed by a base (16). However, the obtained ZnO QDs display a low quantum yield and are unstable in water and in non-polar solvents. Surface modification is further required to prevent growth and aggregation of QDs and to ensure their colloidal stability in different environments (17). Obtaining ZnO QDs with high quantum yield while preserving their visible luminescence upon surface modification remain a challenge because of the sensitivity of photoluminescence to subtle changes of synthesis parameters and environment.
It has been reported that doping with Mg2+ ions can enhance the luminescence of ZnO. Most of previous researches have focused on Mg-doped ZnO thin films and particles annealed at high temperature, with high UV emission but weak visible luminescence (18–22).
In the present study, we report the design of ZnO-based QDs with strong visible luminescence, promising for labeling of lipidic nanoparticles which could be used as theranostic devices. Specifically, we have shown that (i) Mg-doped ZnO QDs prepared using the sol-gel method exhibited enhanced visible luminescence, (ii) the maximum quantum yield was obtained for Mg precursor concentration of 20 mol%, (iii) Mg-doped ZnO QDs capped by oleic acid formed stable colloidal suspensions in toluene and chloroform, while preserving their photoluminescence.
Mg-doped ZnO QDs were characterized by elemental analysis, transmission electronic microscopy, small- and wide-angle X-ray scattering, Raman spectroscopy, UV-Vis absorption and photoluminescence emission. The aim was to establish a relationship between the composition and structure of the QDs and their luminescent properties.
Materials and methods
Materials
Zinc acetate dihydrate, ZnAc2·2H2O (Alfa Aesar), lithium hydroxide monohydrated, LiOH·H2O (Alfa Aesar), magnesium acetate tetrahydrate, Mg(Ac)2·4 H2O (Alfa Aesar), oleic acid (Alfa Aesar), ethanol (VWR, BDH, Prolabo), chloroform (VWR, BDH, Prolabo) and heptane (Carlo Erba Reagents), were used as received, without further purification. Zinc oxide powder (Alfa Aesar) was used as standard.
Synthesis
ZnO QDs
ZnO colloidal suspensions were synthesized by hydrolysis and condensation at 60°C of a precursor solution (Zn4OAc6 ethanolic solution) upon mixing with a LiOH ethanolic solution, following the sol-gel route proposed by Spanhel and Anderson (23).
The Zn4OAc6 tetrameric precursor was first prepared by refluxing an absolute ethanol solution with 0.05 M ZnAc2·2H2O concentration for ∼2 h at 80°C. The thus-obtained transparent precursor solution was stored at ∼4°C. Meanwhile, LiOH·H2O absolute ethanol solution (0.5 M) was prepared by ultrasonic mixing. Hydrolysis and condensation reactions were carried out in a flask containing the precursor solution thermostated at 60°C. LiOH solution was dropped into the precursor solution and the reaction medium was then continuously stirred for 30 min. Afterward, the ZnO colloidal suspension was cooled and stored at ∼4°C to prevent any further growth process. The nominal molar ratio [OH]/[Zn] was 1.4.
Mg-ZnO QDs
Mg doping was achieved by adding a Mg(Ac)2 ethanol solution to the Zn precursor solution, keeping the total molar concentration [Zn+Mg] constant. The [LiOH]/[Zn+Mg] molar ratio was also kept constant. The reaction conditions were similar to those described above for preparing undoped nanoparticles. The Mg precursor concentration was increased from 2.5 to 20 mol%. Samples are referred to using nominal Mg2+ mole fraction in the initial reaction mixture (e.g. 20% Mg-ZnO).
OA-ZnO QDs
ZnO QDs capped by oleic acid (OA) were obtained by adding oleic acid to the ZnO QDs suspension (13 mM, 18.8 mM, 25 mM or 36.7 mM OA concentration) under vigorous stirring during 30 min at 60°C. The resulting turbid solutions were stored at 8°C overnight and then centrifuged (10 min at 10,000 rpm, 13000 g). The supernatant was removed and the OA-ZnO QDs were washed with ethanol, in which they could not be dispersed, in order to remove the unreacted OA. Afterward, the washed OA-ZnO QDs were dried at room temperature under vacuum and put in suspension into chloroform or toluene.
OA-Mg-ZnO QDs
The surface modification of the 20% Mg-ZnO QDs, which showed optimal photoluminescence properties, was achieved in the same way.
Powder obtention
Doped or undoped ZnO QDs could be recovered from the ethanolic suspensions thanks to an extraction procedure using heptane. The as-synthesized colloidal suspensions were mixed with a “non-solvent” heptane (24) (1:4) to induce the precipitation of the QDs, and then centrifuged at 20°C for 10 min (10,000 rpm, 13000 g). The supernatant was discarded and the powder was dried under vacuum at room temperature. The dried powder was used for X-ray powder diffraction characterization, Mg quantification and Raman spectroscopy analyses.
Characterization
Inductively coupled plasma mass spectrometry (ICP-MS)
Mg2+ concentrations in doped QDs were determined using inductively coupled plasma mass spectrometry at the Institut des Sciences Analytiques, Villeurbanne.
High resolution transmission electron microscopy (HRTEM)
HRTEM investigations were performed with a JEOL JEM-2011 UHR scanning electron microscope operating at 200 keV, at the “Service of electronic microscopy” at the University Pierre and Marie Curie, Paris, France. A drop of the dilute colloidal suspension of QDs was deposited on a copper grid. Image analysis was carried out with the ImageJ software.
X-ray powder diffraction (XRD)
XRD analysis of nanoparticles was performed on a Bruker D2 PHASER diffractometer, using the Cu Kα radiation, λ=1.5418 Å, selected by a curved graphite monochromator and a fixed divergence slit of 1/8° in a Bragg-Brentano configuration. The diffraction intensity data were measured in the 2θ range 5–70° by the step counting method (0.1° step and 3 s counting time).
Small-angle X-ray scattering (SAXS)
SAXS experiments were performed on the SWING beamline, operated at 10.5 keV, at the SOLEIL synchrotron source (Saint Aubin, France).
For two compositions, pure ZnO and 20 % Mg-ZnO, the nucleation and growth of nanoparticles have been followed in situ, using a home-made stopped-flow device. The freshly prepared precursor solution and the ethanolic LiOH solution, contained in two syringes, were simultaneously pushed by syringe pumps into two tubes, rapidly mixed at the Y-junction of these tubes and introduced into the thermostated X-ray capillary. The temperature of the sample holder was set to 60°C. Time-resolved SAXS patterns could be recorded from the first seconds of the reaction. The acquisition time of each curve was 200 ms. The dead time between two acquisitions was adjusted to take into account the evolution of the nanoparticle growth rate during synthesis.
SAXS patterns were collected by a two-dimensional CCD detector with a sample-to-detector distance of 1727 mm. The scattered intensity was reported as a function of the scattering vector q=4π sinθ/λ, where 2θ is the scattering angle and λ the wavelength of the incident beam. The calibration of the q range (0.0062–0.64 Å–1) was carried out with silver behenate.
Intensity values were normalized to account for beam intensity, acquisition time and sample transmission. Each scattering pattern was then integrated circularly to yield the intensity as a function of q. The scattered intensity from a capillary filled with ethanol was subtracted from the sample scattering curves. The analysis of the SAXS data was carried out using the software package SASFIT (25).
Raman spectroscopy
The Raman spectra were collected in the backscattering configuration at room temperature with a commercial Kaiser Optical Systems RXN1 Raman spectrometer equipped with a near-IR laser diode working at 785 nm, operating with a power output of 50 mW. All the samples, apart from liquid oleic acid, were dried powders.
Spectroscopy UV
The absorption spectra were measured using a Cary Win 4000 UV-Vis spectrophotometer with a cuvette of 1 mm optical path. The spectra of the QDs suspensions were recorded between 250 and 450 nm, with a wavelength step of 1 nm, and an average counting time of 0.2 s per point. The UV-vis spectra were corrected from the absorption spectrum of ethanol.
Photoluminescence spectroscopy (PL)
The emission (PL) and excitation (PLE) photoluminescence spectra were recorded on a Luminescence spectrometer LS50B (Perkin-Elmer) at room temperature with a Xe lamp (20 kW) as the excitation source. For PL measurements, each sample was excited at the optimal wavelength, defined by the PLE spectrum (Table 1).
Amount of Mg incorporated into ZnO QDs obtained by ICP, crystallite size deduced from X-ray diffraction peaks using the Debye-Scherrer relation, bandgap measured using UV/Vis absorption, wavelength of the PLE peak maximum, wavelength of the PL emission peak maximum, quantum yield of each sample.
| Initial Mg mole fraction | Mg incorporated into ZnO, % | Crystallite size, nm | Bandgap, eV | λex, nm | λem, nm | QY, % |
|---|---|---|---|---|---|---|
| 0 | 3.3 | 3.37 | 365 | 520 | 10.2 | |
| 2.5 | 0.41 | 2.7 | 3.42 | 363 | 520 | 5.0 |
| 5 | 0.72 | 2.3 | 3.52 | 358 | 512 | 16.7 |
| 10 | 1.98 | 2.2 | 3.60 | 350 | 476 | 27.8 |
| 20 | 2.94 | 2.0 | 3.71 | 342 | 470 | 64.8 |
The quantum yield (QY) determines the efficiency of the conversion of absorbed photons into emitted ones. Usually, the relative QY of the studied sample is compared with that of a reference fluorophore. In this work, a solution of Rhodamine 6G in ethanol (QY=95%), often used as reference for green emission, was selected as the standard (26). According to the method described by Crosby and Demas (27), the QY is given by the following formula:
where A is the absorbance at the excitation wavelength (λ), n the refractive index of the solvent used and D the integrated fluorescence intensity (the area under the corrected emission curve). The subscripts r and x refer to the reference and sample solutions, respectively. To minimize the errors, the same excitation wavelength was chosen to measure the PL spectra of the Rhodamine 6G and QD solutions and the two solutions had the same absorbance at this wavelength (A=0.05).
Results
Mg-doped ZnO QDs were synthesized by the hydrolysis and condensation reaction described above. Surface modification of the QDs was performed using oleic acid (OA) to hinder their aggregation and to provide them colloidal stability in different environments. Different quantities of OA were added to the QD suspensions. In all cases the QDs capped by OA formed stable colloidal dispersions in chloroform or toluene.
Preliminary measurements having shown that the visible luminescence of Mg-ZnO QDs went through a maximum for a 20 mol% nominal concentration of Mg2+ ions in the reaction medium, our investigations focused on the structure and properties of QDs obtained with Mg2+ concentration increasing up to 20%.
Actual atomic concentrations of Mg in doped QDs, determined using ICP-MS, are reported in Table 1. These data evidence a low incorporation of Mg2+ ions into ZnO QDs. The Mg2+ concentration varied from 0.41 to 2.94%.
Figure 1A shows powder wide-angle X-ray diffraction patterns, recorded between 2θ=5° and 70°, of ZnO reference, ZnO QDs and Mg-ZnO QDs. They display peaks characteristic of the ZnO hexagonal wurtzite structure. No peak characteristic of the MgO rock salt phase (e.g., at 42.5°) is detected whatever the concentration of Mg2+ ions.
(A) XRD profiles of ZnO standard and 0, 2.5, 5, 10 and 20% of Mg-ZnO QDs, showing the presence of the hexagonal wurtzite phase (B) zoom of the (100) peak in the 30°–34° 2θ range.
The broadening of the peaks of Mg-ZnO QDs could result from a decrease in crystallinity and/or crystal size, promoted by the presence of Mg. Assuming that this broadening is mainly due to size effects, the average size of nanocrystals could be estimated using the Debye-Scherrer relation (28) applied to the (100) reflection:
where D is the crystal size; k is a constant (shape factor, 0.89 for spherical nanoparticles), λ is the X-ray wavelength, β is the full width at half maximum (FWHM) of the diffraction peak and 2θ is the diffraction angle. The size of the nanocrystals decreased from 3.3 nm for ZnO to 2.0 nm for 20% Mg-ZnO. Figure 1B shows a zoom of the (100) diffraction peak between 2θ=30° and 2θ=34°. Mg2+ ions shift the (100) peak to slightly higher angles, suggesting a weak decrease of the lattice parameters. This small change, whose extent depends on the Mg concentration, is expected to reflect the substitution of some Zn2+ ions by Mg2+ ions in the crystalline lattice, as the Mg2+ ions have a slightly smaller ionic radius (0.57 Å for Mg2+ ions compared to 0.60 Å for Zn2+ ions). At high Mg concentrations the peak shifts are obscured by the large width of the peaks.
The size and morphology of ZnO and 20% Mg-ZnO nanoparticles were also studied using HRTEM (Figure 2). HRTEM images show nanocrystals with approximately spherical shapes and average diameters of about 4 nm for ZnO QDs. The average size of Mg-ZnO QDs, whose structure appears less ordered, is slightly smaller. Overall, these data are in agreement with XRD data.
HRTEM images of undoped ZnO QDs (A) and 20% Mg-ZnO QDs (B).
The influence of Mg on the growth of nanoparticles was further evidenced by time-resolved SAXS patterns, recorded from the beginning of the reaction. Figures 3A and 4A present three-dimensional stacked log-log plots of the SAXS curves as a function of time, corresponding to the formation of ZnO and 20% Mg-ZnO QDs, respectively. Selected curves of ZnO and 20% Mg-ZnO QDs formation are shown in Figures 3B and 4B, respectively. The two systems exhibit clearly different time evolution.
Three-dimensional stacked log-log plots of the SAXS curves as a function of time recorded in situ during the formation of ZnO QDs (A) and selected in situ SAXS profiles measured at the indicated reaction time (min) (B).
Three-dimensional stacked log-log plots of the SAXS curves as a function of time recorded in situ during the formation of 20% Mg-doped ZnO QDs (A) and selected in situ SAXS profiles measured at the indicated reaction time (min) (B).
At the beginning of the reaction, ZnO curves display a plateau at low q-range (Guinier region), indicative of the scattering of non-interacting particles. In the Guinier region, the scattered intensity can be approximated by I(q)=I(0) exp(–Rg2q2/3) where Rg is the radius of gyration (Guinier radius) of the particles (29). The Guinier plateau progressively shifts to lower q-values, reflecting the increase in the nanoparticle size with time. The curves display an oscillation at high q-values, which can arise from the form factor of spherical particles. The SAXS profiles could be fitted by the form factor of homogeneous spheres with Gaussian radius distribution. The radii of gyration Rg are consistent with the mean radii of the nanoparticles R, as deduced from the fits [for a sphere Rg is defined as Rg=(3/5)½R] (Figure 5A). The size of the nanoparticles increased steeply during the first minutes of the reaction, up to 3 nm after 3 min. After about 15 min the size of the primary particles, which has reached 4 nm, no longer evolved.
Comparison of the time evolution of the radius of gyration (Rg), determined in the Guinier region, and the radius (R) , deduced from the fit of curves by the form factor of spheres, of nanoparticles: (A) ZnO QDs and (B) 20% Mg-ZnO QDs.
At the end of the reaction, after about 50 min, the intensity scattered at low q-values strongly increases and the Guinier plateau disappears. In the low q region, a linear decay of logI is observed in the logI vs. logq plot: I(q) ∝ q–α (30) with α=2.09. This behaviour evidences the formation of fractal aggregates with fractal dimension α, defined as the exponent that relates the mass M of an object to a characteristic dimension R: M ∝ Rα. The α=2.09 fractal dimension is characteristic of structures formed by reaction-limited cluster-cluster aggregation (RLCCA). The aggregation rate is limited by the time needed to overcome the repulsive barrier between two clusters. Compared to the diffusion-limited cluster-cluster aggregation (DLCCA), the RLCCA mechanism corresponds to slower colloid aggregation, yielding more compact clusters due to the lower sticking probability (31).
This linear decrease in logI as a function of logq is observed between q-values corresponding to the primary particle size and the fractal aggregate size, respectively. As shown by the curve oscillation in the high q range, the mean radius of the primary particles remains constant, at about 2 nm, until the end of the reaction. Consequently, the formation of ZnO QDs involves two main steps: the nucleation and growth, up to a size of ∼4 nm, of elementary nanoparticles during the ten first min of the reaction, followed by the formation of fractal aggregates with fractal dimension α=2.09.
The structures giving rise to the first Mg-ZnO SAXS profiles collected during the synthesis of Mg-ZnO QDs could not be identified. After about 5 min, the SAXS curves could also be fitted by the form factor of homogeneous spheres with Gaussian radius distribution. The growth of nanoparticles is rapidly stopped after the nucleation step. Nanoparticles remain smaller than ZnO ones. A further small increase in size is observed at the end of the reaction. The intensity I(0), proportional to the number of nanoparticles in the medium, increases continuously during the first 60 min of the reaction. Of note, unlike ZnO nanoparticles, Mg-ZnO QDs do not form fractal aggregates with time, suggesting that the nanoparticles have different surface properties.
Raman scattering carried out on QDs allows obtaining information on the surface species and on the modifications of the optical phonon spectrum, as compared to ZnO bulk crystal values (Figure 6A and B).
Raman spectra of ZnO standard, ZnO QDs, Mg-ZnO QDs, OA, OA-ZnO QDs and OA-Mg ZnO QDs with indications of the principal vibrations (A) from 200 to 800 cm–1 and (B) from 800 to 1800 cm–1.
The spectrum of ZnO standard shows a prominent peak at 439 cm–1 and weak peaks at 332 cm–1, 380 cm–1, 410 cm–1, 581 cm–1 and 666 cm–1, in agreement with previous studies (32). The peak at 439 cm–1 is characteristic of the wurtzite lattice; the sharp line shape is indicative of high crystalline order. This peak becomes less intense in ZnO QDs while the intensity of the 410 cm–1 band increases, appearing as a shoulder of the peak at 439 cm–1. More importantly, in Mg-ZnO QDs the weak 439 cm–1 peak is involved in a broad band, reflecting the diminution of crystallinity induced by Mg incorporation.
The presence of oleic acid (OA) attached on the surface of QDs is unambiguously evidenced by the intense and well resolved peaks at 1265 cm–1, 1301 cm–1 and 1655 cm–1 which cannot overlap with peaks arising from other species. These peaks correspond to δ(=CH) deformations, δ(CH2)n deformations and υ(C=C) stretching vibrations of unsaturated fatty acids, respectively (33). Other peaks assigned to OA (34) are listed in Table 2.
Raman shift (cm–1) and assignment of the main vibrations observed in the following samples: ZnO standard, ZnO QDs, Mg-ZnO QDs, OA, OA-ZnO QDs and OA-Mg-ZnO QDs. Raman shifts are compared with values from References (32–35).
| Raman shift, cm–1 | Origin | Reference, cm–1 | Assignment | Reference |
|---|---|---|---|---|
| 332 | ZnO | 332 | E2 (high)–E2 (low) | Cusco |
| 380 | ZnO | 380 | A1(TO) | Cusco |
| 410 | ZnO | 408 | E1(TO) | Cusco |
| 439 | ZnO | 437 | E2(high) | Cusco |
| 581 | ZnO | 584 | E1(LO) | Cusco |
| 620 | Acetate | 621 | Y15(B2); out of plane | Koleva |
| 673 | ZnO/acetate | 666/671 | TA+LO/υ5(A1); OCO sym. bend | Cusco/Koleva |
| 941 | Acetate | 942 | υ4 (A1), C–C strech | Koleva |
| 1065 | OA | 1063 | υas(CC), chain ordered | Tandon |
| 1085 | OA | 1082 | υ(CC), chain ordered | Tandon |
| 1116 | OA | 1118 | υ(CC), chain ordered | Tandon |
| 1267 | OA | 1265 | δ(=CH) | De Gelder |
| 1301 | OA | 1301 | δ(CH2)n | De Gelder |
| 1435 | Acetate | 1435 | υ2(A1); CO sym. strech | Koleva |
| 1439 | OA | 1438 | CH2 scissoring | Tandon |
| 1655 | OA | 1655 | υ(C=C) | De Gelder |
Beside peaks originating from the optical phonon spectrum of QDs and OA molecules capping their surface, Raman spectra of QDs exhibit a prominent peak at 940 cm–1 and weak broad peaks at 1352 cm–1 and 1435 cm–1, which could be attributed to acetate ions. These peaks correspond to ν(C–C) stretching, δ(CH3) symmetric bending and ν(CO) symmetric stretching, respectively (35). Raman measurements support the incorporation of Mg2+ ions in ZnO lattice and demonstrate that the QD surface is efficiently capped with OA. They also evidence the presence of acetate surface groups.
The optical properties of the ZnO and Mg-ZnO QD colloidal suspensions were studied by UV-Vis absorption and PL spectroscopy. UV-vis absorption spectra of ZnO and Mg-doped ZnO QD suspensions are shown in Figure 7. The absorption edge shifts from 340 nm to 305 nm with Mg concentration increasing from 0 to 20 mol%. This progressive blue-shift is indicative of the band gap widening upon Mg doping. The band gap energy was estimated to be the energy at which the absorbance of the absorption edge reached half of its maximum value (36). An increase from 3.37 eV to 3.71 eV is observed, consistent with previous studies (Table 1).
Absorption spectra of ZnO and 2.5, 5, 10 and 20% Mg-ZnO colloidal suspensions.
Usually, the PL spectra of ZnO exhibit two components. One is the exciton, or band-edge, emission in the UV range, the other is visible emission, also called deep-level emission, due to the existence of defects which are mainly located on the ZnO surface. Nanoparticles prepared by the sol-gel route exhibit mainly visible luminescence whereas their UV emission is weak. On the contrary, highly crystalline ZnO displays strong band-edge luminescence. In our study, only a weak UV emission peak at ∼380 nm was observed when the ZnO QDs where excited at 342 nm. The intensity of the UV emission peak further decreased and became negligible when the concentration of Mg in QDs increased, in agreement with a previous report (37). Mg-doped QDs were excited at low wavelength, between 324 nm and 346 nm. We have therefore focused on the visible emission, by selectively exciting the defect states. Photoluminescence excitation (PLE) spectra were measured with detection at a wavelength where the visible emission was maximum (Table 1), revealing an increase in the peak energy with increasing Mg doping level (Figure 8). This dependence is in line with that of the absorption spectra.
PLE and PL spectra of ZnO and 2.5, 5, 10 and 20% Mg-ZnO colloidal suspensions. The inset shows the photography of ZnO (right) and 20% Mg-ZnO (left) colloidal suspensions under UV lamp (λexc=365 nm).
Photoluminescence emission spectra were then collected by exciting each QD suspension at the wavelength corresponding to the maximum of the PLE spectrum (Figure 8). The emission spectra exhibit two broad bands at 400–495 nm and 495–590 nm, whose relative intensities depend on Mg2+ concentration. The intensity of the first band is enhanced at high Mg2+ concentration. The luminescence increases when Mg2+ concentration rises from 5% to 20%. The quantum yields of the different samples are reported in Table 1. Remarkably, the quantum yield (QY) of 20 mol% Mg-doped ZnO colloidal suspension (64%) is about 6 times higher than that of the ZnO suspension (10%).
Figure 9 shows the emission spectra of ZnO and 20% Mg-ZnO QDs capped by OA dispersed in chloroform excited at 345 nm and 341 nm, respectively. The emission wavelength range is not affected by OA. The corresponding QY values are reported in Table 3. It can be seen that OA reduced the QY of doped ZnO QDs from 64% to about 40%.
PL spectra of the OA-ZnO QDs and OA-Mg-ZnO QDs dispersed in chloroform. The insets show the photographs of each sample under UV lamp (λexc=365 nm).
Quantum Yield (QY) of the ZnO and 20% Mg-ZnO QDs capped with oleic acid (OA).
| Sample | QY, % |
|---|---|
| ZnO 50 mM OA 13 mM | 5.76 |
| ZnO 50 mM OA 36.7 mM | 11.86 |
| Mg ZnO 50 mM OA 13 mM | 41.56 |
| Mg ZnO 50 mM OA 36.7 mM | 38.22 |
Discussion
Doping ZnO is an effective approach to modify its properties. The introduction of dopant ions can either enlarge or narrow the band gap of the semiconductor, thereby tuning the emission colors. Furthermore, doping ZnO nanoparticles with ions such as Mn2+, Ni2+ or Co2+ is known to impart magnetic properties to the material (38, 39).
Effective dopant incorporation is a critical issue. If dopant ions are excluded during ZnO crystal growth or just adsorbed on the surface, the desired properties may be compromised. The close radii of Mg2+ (0.57 Å) and Zn2+ (0.60 Å) contribute to the solid solubility of Mg2+ in ZnO since Mg2+ ions may replace Zn2+ ions in the wurtzite lattice. However, both the ratio of actual Mg amount in nanoparticles to nominal concentration in the reaction medium and the maximum Mg doping were found extremely sensitive to the conditions of sample preparation in previous studies. Generally, higher Mg2+ contents were achieved when the nanocrystal synthesis was performed at high temperature.
Yang et al reported a dopant content of 22.6 mol%, for a nominal 50% content in the reaction mixture. Doped nanocrystals with different shapes were obtained by alcoholysis of Zn stearate and Mg stearate in 1-octadecene at 270°C (40). Cohn et al achieved a dopant content of 18%, close to the nominal content, by heating nanocrystals in dodecylamine at 160°C for about 20 min after hydrolysis and condensation reaction in DMSO at 50°C (41). In contrast, Mg content as low as 0.03%, for a Mg concentration in the initial solution of 20%, was obtained by Ghosh et al using sol-gel process at 70°C (42). In all cases the wurtzite ZnO phase was maintained. In the present study, the actual Mg content was always much lower than the nominal precursor concentration in the reaction medium. When synthesis was carried out with a [LiOH]/[Zn+Mg] molar ratio r=1.4, the maximum Mg2+ substitution achieved was 2.94% for an initial Mg concentration of 20%. Synthesis performed with a different molar ratio, r=0.5, yielded a concentration of 0.56%. These findings suggested that kinetic factors, likely related to relative precursor reactivities, could limit Mg incorporation in the host ZnO crystal. Generally, it has been emphasized that the growth rate of the host crystal and the rate of deposition of dopant ions should be balanced to achieve effective doping during crystal growth (43).
The presence of Mg2+ ions affected the growth and size of the NPs, along with their structure. Mg2+ addition decreased the final size of QDs and inhibited their fractal aggregation, suggesting a modification of their surface properties. As evidenced by X-ray diffraction, HRTEM images and Raman spectra, doped QDs displayed a less crystalline, more disordered wurtzite structure than ZnO QDs.
The incorporation of Mg widened the band gap of doped ZnO QDs primarily by decreasing their size since NPs smaller than ∼6 nm are sensitive to quantum confinement. Moreover, it is known that doping with Mg widens the band gap of ZnO (37, 40). According to Cohn et al., the introduction of Mg2+ ions raises the conduction band potential and lowers the valence band potential of ZnO (41). The two effects (quantum confinement and intrinsic effect of Mg) could not be separated.
The visible emission spanned the 400–590 nm range, featuring two broad bands at 400–495 nm and 495–590 nm whose relative intensities depended on Mg2+ concentration. Different relaxation processes, radiative or non-radiative, can take place upon photoexcitation of a ZnO nanoparticle. Two mechanisms have been proposed for the ZnO visible emission: i) recombination of an electron in the valence band with a hole in a deep trap, ii) recombination of an electron in singly ionized oxygen vacancy (i.e., deeply trapped) with a photo-generated hole in the valence band (Xiong, 2010). The green-yellow luminescence of ZnO is usually attributed to singly ionized oxygen vacancies (15, 44–48). It has been shown by PL microscopy probing luminescence profile at single QD level that the visible emission of ZnO QDs is intrinsically broad. The broad bandwidth does not mainly arise from the NP size (or composition herein) distribution but from multiple transitions involving closely spaced energy levels, inherent to every QD, lying between the valence band and the conduction band (44). The increase in QY upon Mg2+ doping can be explained by the increased concentration of inner and surface defects, in particular oxygen vacancies, and by the decrease in QD size. Because of the larger surface-to-volume ratio, smaller QDs entail the formation of more numerous surface defects. The surface state is expected to strongly influence the luminescence of ZnO QDs. For instance, correlations have been observed between the presence of surface hydroxide moieties and visible luminescence intensity. As observed by Felbier et al., when surface oxygen containing species, such as OH groups, were desorbed under vacuum, the visible emission vanished while the UV emission was significantly enhanced (49). Of note, the decrease in QY when Mg-doped QDs are capped by oleic acid is consistent with the hypothesis of the role of OH groups in visible emission. The number of OH surface groups is expected to decrease if they react with oleic acid. The difference in relative intensities of the two broad bands at 400–495 nm and 495–590 nm suggests the existence of different types of defects and/or different paths for electron-hole recombination as a function of Mg concentration.
Conclusions
Mg-ZnO QDs were synthesized by a hydrolysis and condensation reaction. Mg2+ ions could be incorporated into the ZnO wurtzite lattice owing to the very close values of the Mg2+ and Zn2+ ion radii. However, the dopant ions strongly influenced the growth and final size of ZnO nanocrystals. Doping with Mg2+ ions widened the band gap of ZnO QDs and enhanced their visible luminescence. The luminescence went through a maximum for a 20 mol% nominal concentration of Mg2+ ions in the reaction medium. With increasing proportion of Mg2+ ions, both the absorption and emission spectra experienced a blue shift. Mg-ZnO QDs capped by oleic acid (OA) formed stable colloidal dispersions in chloroform and toluene, with strong visible luminescence, promising for biological imaging.
List of non-standard abbreviations
Mg-ZnO QDs, Mg-doped ZnO quantum dots; OA-ZnO QDs, ZnO quantum dots with oleic acid as surface modifier; OA-Mg-ZnO QDs, Mg-doped ZnO quantum dots with oleic acid as surface modifier.
About the authors
Eloísa Berbel Manaia is currently preparing her PhD thesis in co-direction between the Faculty of Pharmaceutical Sciences, São Paulo State University (Brazil) and the Faculty of Pharmacy, Institut Galien, University Paris Sud (France). The topic of her PhD thesis and her research interest focus on the development of QDs based on ZnO for application in bioimaging and also in theranostic systems using lipid nanocarriers. She is particularly interested in nanomedicine, diagnostic molecules and drug delivery systems.
Renata C.K. Kaminski received a Chemistry diploma from the University of Paraná (UFPR) in 2000. She completed her PhD degree in Chemistry at the São Paulo State University (2006) studying TiO2-based thin films and powder. In 2008 she had a post-doctoral position at the Synchrotron SOLEIL (Saint-Aubin), where she worked as chemical engineer and researcher for 1 year. Now, she works as researcher at University of Sergipe (UFS), her research foccuses on TiO2-based nanoparticles for photoprotection and antioxidants delivery via emulsified or liquid crystalline systems.
Bruno Leonardo Caetano received a Chemistry diploma from the University of Franca (2003). He completed his PhD degree in Chemistry at the São Paulo State University (2010) studying the formation and growth of ZnO quantum dots. He completed his first post-doctoral in 2011 at the Synchrotron SOLEIL (Saint-Aubin). Now, his research for his second post-doctorate is focused on the nucleation and growth of nanoparticles, organic-inorganic hybrid materials and magnetic nanoparticles for drug delivery systems and hyperthermia.
Valérie Briois received a Chemical Engineering diploma from the National School of Chemistry of Paris (ENSCP) in 1988. She completed her PhD degree in Chemistry at the Université Pierre et Marie Curie-Paris VI in 1991 on the study of the genesis of oxysulfate of cerium(IV) by synchrotron radiation techniques. Then she joined the CNRS as researcher first at the Laboratoire pour l’Utilisation du Rayonnement Electromagnétique (LURE, Orsay) and then at the Synchrotron SOLEIL (Saint-Aubin) where she took the responsibility of hard X-rays beamlines dedicated to the X-ray absorption spectroscopy (XAS). She focuses her research interests on nanostructured materials prepared by soft-chemistry routes aiming to establish relationships between local range order structures determined by XAS and nanomaterials properties.
Leila Aparecida Chiavacci graduated in Chemistry at São Paulo State University (UNESP) (1993), obtained her MSc in Chemistry/Physical Chemistry at UNESP (1996), and she completed her PhD in Chemistry/Physical Chemistry and in Materials Science in co-direction between UNESP and LURE, Université Paris-Sud (2001). Currently, she is a Professor at the Faculty of Pharmaceutical Sciences – UNESP – Araraquara. She has experience in the area of Chemistry and Materials, with an emphasis in Physical Chemistry and Nanomaterials, acting on the following topics: colloids, nanomaterials, controlled release of drugs, SAXS, EXAFS, and sol-gel hybrid organic-inorganic materials.
Claudie Bourgaux joined the CNRS as researcher in 1982, first at ESPCI (PARIS) and then at the “Laboratoire pour l’Utilisation du Rayonnement Electromagnétique” (LURE, ORSAY) as co-responsible of SAXS beamline. Currently, she works at “Institut Galien Paris-Sud”. Her research interests include supramolecular assemblies of lipids and amphiphiles, either for the delivery and controlled release of drugs or for the study of interactions of drugs with model membranes.
Acknowledgments
The authors thank D. Desmaële for his help during synthesis, the synchrotron SOLEIL for providing beamtime at the SWING beamline (project number: 20131299) J. Perez, Y. Liatimi and P. Roblin for their support during SAXS experiments, S. Blanchandin for WAXS measurements, the Service of electronic microscopy of the University Pierre et Marie Curie for HRTEM measurements, and FAPESP and PADC/FCF-UNESP for the financial support. This work was conducted during a scholarship of E. Berbel Manaia supported by the International Cooperation Program CAPES/COFECUB (ME 767-13) at the University Paris – Sud, financed by CAPES – Brazilian Federal Agency for Support and Evaluation of Graduate Education within the Ministry of Education of Brazil.
References
1. Reid PJ, Fujimoto B, Gamelin DR. A simple ZnO nanocrystal synthesis illustrating three-dimensional quantum confinement. J Chem Educ 2013;91:280–2.10.1021/ed300693dSearch in Google Scholar
2. Michalet X, Pinaud FF, Bentolila LA, Tsay JM, Doose S, Li JJ, et al. Quantum dots for live cells, in vivo imaging, and diagnostics. Science 2005;307:538–44.10.1126/science.1104274Search in Google Scholar PubMed PubMed Central
3. Wang C, Gao X, Su X. In vitro and in vivo imaging with quantum dots. Anal Bioanal Chem 2010;397:1397–415.10.1007/s00216-010-3481-6Search in Google Scholar PubMed
4. Xie J, Lee S, Chen X. Nanoparticle-based theranostic agents. Adv Drug Deliv Rev 2010;62:1064–79.10.1016/j.addr.2010.07.009Search in Google Scholar PubMed PubMed Central
5. Tian B, Al-Jamal WT, Al-Jamal KT, Kostarelos K. Doxorubicin-loaded lipid-quantum dot hybrids: surface topography and release properties. Int J Pharm 2011;416:443–7.10.1016/j.ijpharm.2011.01.057Search in Google Scholar PubMed
6. Park JH, von|Maltzahn G, Ruoslahti E, Bhatia SN, Sailor MJ. Micellar hybrid nanoparticles for simultaneous magnetofluorescent imaging and drug delivery. Angew Chem Int Ed 2008;47:7284–8.10.1002/anie.200801810Search in Google Scholar PubMed PubMed Central
7. Bagalkot V, Zhang L, Levy-Nissenbaum E, Jon S, Kantoff PW, Langer R, et al. Quantum dot-aptamer conjugates for synchronous cancer imaging, therapy, and sensing of drug delivery based on bi-fluorescence resonance energy transfer. Nano Lett 2007;7:3065–70.10.1021/nl071546nSearch in Google Scholar PubMed
8. Su Y, He Y, Lu H, Sai L, Li Q, Li W, et al. The cytotoxicity of cadmium based, aqueous phase – synthesized, quantum dots and its modulation by surface coating. Biomaterials 2009;30:19–25.10.1016/j.biomaterials.2008.09.029Search in Google Scholar PubMed
9. Chang E, Thekkek N, Yu WW, Colvin VL, Drezek R. Evaluation of quantum dot cytotoxicity based on intracellular uptake. Small 2006;2:1412–7.10.1002/smll.200600218Search in Google Scholar PubMed
10. Lewinski N, Colvin V, Drezek R. Cytotoxicity of nanoparticles. Small 2008;4:26–49.10.1002/smll.200700595Search in Google Scholar PubMed
11. Klaine SJ, Alvarez PJ, Batley GE, Fernandes TF, Handy RD, Lyon DY, et al. Nanomaterials in the environment: behavior, fate, bioavailability, and effects. Environ Toxicol Chem 2008;27:1825–51.10.1897/08-090.1Search in Google Scholar PubMed
12. Jia Z, Misra RD. Tunable ZnO quantum dots for bioimaging: synthesis and photoluminescence. Mater Tech 2013;28:221–7.10.1179/1753555713Y.0000000061Search in Google Scholar
13. Xiong HM. ZnO nanoparticles applied to bioimaging and drug delivery. Adv Mater 2013;25:5329–35.10.1002/adma.201301732Search in Google Scholar PubMed
14. Moussodia RO, Balan L, Merlin C, Mustin C, Schneider R. Biocompatible and stable ZnO quantum dots generated by functionalization with siloxane-core PAMAM dendrons. J Mater Chem 2010;20:1147–55.10.1039/B917629BSearch in Google Scholar
15. Zhang L, Yin L, Wang C, lun N, Qi Y, Xiang D. Origin of visible photoluminescence of ZnO quantum dots: defect-dependent and size-dependent. J Phys Chem C 2010;114:9651–8.10.1021/jp101324aSearch in Google Scholar
16. Caetano BL, Santilli CV, Meneau F, Briois VR, Pulcinelli SH. In Situ and Simultaneous UV-vis/SAXS and UV-vis/XAFS time-resolved monitoring of ZnO quantum dots formation and growth. J Phys Chem C 2011;115:4404–12.10.1021/jp109585tSearch in Google Scholar
17. Xiong HM. Photoluminescent ZnO nanoparticles modified by polymers. J Mater Chem 2010;20:4251–62.10.1039/b918413aSearch in Google Scholar
18. Vijayalakshmi K, Karthick K. Influence of Mg doping on the microstructure and PL emission of wurtzite ZnO synthesized by microwave processing. J Mater Sci Mater Electron 2013;24:2067–71.10.1007/s10854-012-1057-2Search in Google Scholar
19. Xiong HM, Shchukin DG, Möhwald H, Xu Y, Xia YY. Sonochemical synthesis of highly luminescent zinc oxide nanoparticles doped with magnesium(II). Angew Chem 2009;121:2765–9.10.1002/ange.200805590Search in Google Scholar
20. Lin YJ, Wu PH, Tsai CL, Liu CJ, Lin ZR, Chang HC, et al. Effects of Mg incorporation on the optical properties of ZnO prepared by the sol-gel method. J Appl Phys 2008;103:113709.10.1063/1.2939255Search in Google Scholar
21. Ding R, Xu C, Gu B, Shi Z, Wang H, Ba L, et al. Effects of Mg incorporation on microstructure and optical properties of ZnO thin films prepared by sol-gel method. J Mater Sci Technol 2010;26:601–4.10.1016/S1005-0302(10)60092-8Search in Google Scholar
22. Yousefi R, Jamali Sheini F, Muhamad MR, More MA. Characterization and field emission properties of ZnMgO nanowires fabricated by thermal evaporation process. Solid State Sci 2010;12:1088–93.10.1016/j.solidstatesciences.2010.04.019Search in Google Scholar
23. Spanhel L, Anderson MA. Semiconductor clusters in the sol-gel process: quantized aggregation, gelation, and crystal growth in concentrated zinc oxide colloids. J Am Chem Soc 1991;113:2826–33.10.1021/ja00008a004Search in Google Scholar
24. Meulenkamp EA. Synthesis and growth of Zno nanoparticles. J Phys Chem B 1998;102:5566–72.10.1021/jp980730hSearch in Google Scholar
25. Kohlbrecher J, Bressler I. SASfit. Paul scherrer institut, Villigen: Switzerland. 2014.Search in Google Scholar
26. Sun LW, Shi HQ, Li WN, Xiao HM, Fu SY, Caoc XZ, et al. Lanthanum-doped ZnO quantum dots with greatly enhanced fluorescent quantum yield. J Mater Chem 2012;22:8221–7.10.1039/c2jm00040gSearch in Google Scholar
27. Crosby GA, Demas JN. Measurement of photoluminescence quantum yields. Rev J Phys Chem 1971;75:991–1024.10.1021/j100678a001Search in Google Scholar
28. West AR. Solid state chemistry and its applications. New York: John Wiley and Sons, 1992.Search in Google Scholar
29. Guinier A, Fournet G. Small angle scattering of X-rays. New York: Wiley, 1955.Search in Google Scholar
30. Craievich AF. In: Handbook of sol-gel science and technology. Sakka A, Almeida R, editors. Norwell, MA: Kluwer Publishers, 2005.Search in Google Scholar
31. Brinker CJ, Scherer GW. Sol-gel sciences. The physics and chemistry of sol-gel processing. San Diego: Academic Press, Inc., 1990.Search in Google Scholar
32. Cuscó R, Alarcón-Lladó E, Ibáñez J, Artús L. Temperature dependence of Raman scattering in ZnO. Phys Rev B 2007;75:165202.10.1103/PhysRevB.75.165202Search in Google Scholar
33. De Gelder J, De Gussem K, Vandenabeele P, Moens L. Reference database of Raman spectra of biological molecules. J Raman Spectrosc 2007;38:1133–47.10.1002/jrs.1734Search in Google Scholar
34. Tandon P, Förster G, Neubert R, Wartewig S. Phase transitions in oleic acid as studied by X-ray diffraction and FT-Raman spectroscopy. J Mol Struct 2000;524:201–15.10.1016/S0022-2860(00)00378-1Search in Google Scholar
35. Koleva V, Stoilova D. Infrared and Raman studies of the solids in the Mg(CH3COO)2-Zn(CH3COO)2-H2O system. J Mol Struct. 2002;611:1–8.10.1016/S0022-2860(01)00957-7Search in Google Scholar
36. Brus L. Electronic wave functions in semiconductor clusters: experiment and theory. J Phys Chem 1986;90:2555–60.10.1021/j100403a003Search in Google Scholar
37. Yousefi R, Zak AK, Jamali-Sheini F. Growth, X-ray peak broadening studies, and optical properties of Mg-doped ZnO nanoparticles. Mater Sci Semicond Process 2013;16:771–7.10.1016/j.mssp.2012.12.025Search in Google Scholar
38. Schwartz DA, Norberg NS, Nguyen QP, Parker JM, Gamelin DR. Magnetic quantum dots: synthesis, spectroscopy, and magnetism of Co2+- and Ni2+-doped ZnO nanocrystals. J Am Chem Soc 2003;125:13205–18.10.1021/ja036811vSearch in Google Scholar PubMed
39. Norberg NS, Kittilstved KR, Amonette JE, Kukkadapu RK, Schwartz DA, Gamelin DR. Synthesis of colloidal Mn2+:ZnO quantum dots and high-TC ferromagnetic nanocrystalline thin films. J Am Chem Soc 2004;126:9387–98.10.1021/ja048427jSearch in Google Scholar PubMed
40. Yang Y, Jin Y, He H, Wang Q, Tu Y, Lu H, et al. Dopant-induced shape evolution of colloidal nanocrystals: the case of zinc oxide. J Am Chem Soc 2010;132:13381–94.10.1021/ja103956pSearch in Google Scholar PubMed
41. Cohn AW, Kittilstved KR, Gamelin DR. Tuning the potentials of “Extra” electrons in colloidal n-Type ZnO nanocrystals via Mg2+ substitution. J Am Chem Soc 2012;134:7937–43.10.1021/ja3019934Search in Google Scholar PubMed
42. Ghosh M, Raychaudhuri AK. Optical properties of Mg-substituted ZnO nanoparticles obtained by solution growth. Nanotech IEEE Trans Nanotechnol 2011;10:555–9.10.1109/TNANO.2010.2050899Search in Google Scholar
43. Buonsanti R, Milliron DJ. Chemistry of doped colloidal nanocrystals. Chem Mater 2013;25:1305–17.10.1021/cm304104mSearch in Google Scholar
44. Layek A, De S, Thorat R, Chowdhury A. Spectrally resolved photoluminescence imaging of ZnO nanocrystals at single-particle levels. J Phys Chem Lett 2011;2:1241–7.10.1021/jz200370sSearch in Google Scholar PubMed
45. van Dijken A, Meulenkamp EA, Vanmaekelbergh D, Meijerink A. The kinetics of the radiative and non-radiative processes in nanocrystalline ZnO particles upon photoexcitation. J Phys Chem B 2000;104:1715–23.10.1021/jp993327zSearch in Google Scholar
46. Xu X, Xu C, Wang X, Lin Y, Dai J, Hu J. Control mechanism behind broad fluorescence from violet to orange in ZnO quantum dots. Cryst Eng Comm 2013;15:977–81.10.1039/C2CE26550HSearch in Google Scholar
47. Liu DP, Li GD, Su Y, Chen JS. Highly luminescent ZnO nanocrystals stabilized by ionic-liquid components. Angew Chem Int Ed 2006;45:7370–3.10.1002/anie.200602429Search in Google Scholar PubMed
48. Yousefi R, Kamaluddin B. Fabrication and characterization of ZnO and ZnMgO nanostructures grown using a ZnO/ZnMgO compound as the source material. Appl Surf Sci 2009;256:329–34.10.1016/j.apsusc.2009.08.025Search in Google Scholar
49. Felbier P, Yang J, Theis J, Liptak W, Wagne A, Lorke A, et al. Highly Luminescent ZnO quantum dots made in a non-thermal plasma. Adv Funct Mater 2014;24:1988–93.10.1002/adfm.201303449Search in Google Scholar
©2015 by De Gruyter
