Lanthanide Luminescence Enhancement of Core-Shell Magnetite-SiO2 Nanoparticles Covered with Chain-Structured Helical Eu/Tb Complexes.

Oligomeric-brush chains of helical lanthanide (Ln) complexes retain their structural and luminescent behavior after coating onto magnetic nanoparticles (MNPs) consisting of Fe3O4 covered with silicate. It is one of the type of bifunctional NPs exhibiting luminescence of Ln and superparamagnetism of Fe3O4. In comparison to a simple monolayer of complexes adsorbed on a modified surface, a layer made of luminescent chains allowed us to obtain a more intensive red/green luminescence originating from Eu3+/Tb3+ ions, and at the same time, no visible increase in particle size (compared to Fe3O4@silica particles) was observed. The luminescent properties of the Tb3+ complex were altered by MNPs; the decrease of the luminescence was not as large as expected, the excitation spectrum changed significantly, and the average luminescence lifetime was much longer at room temperature. Surprisingly, this phenomenon was not observed at 77 K and also did not occur for the Eu3+ complexes. The possibility to stack building blocks in a chain using complexes of different lanthanide ions can be used to design novel multifunctional nanosystems.

Introduction

Luminescent–magnetic nanoparticles (MNPs) have been developed over the last few decades. The popularity of these materials stems from their very small size particles, unique properties, and a huge number of their possible bioapplications, such as bimodal imaging, biolabeling, bioconjugation, photothermal therapy, drug delivery, and so forth.[1−5] So far, many luminescent–magnetic nanomaterials of different structures and luminescent phases have been proposed. Most of them are based on magnetic Fe2O3 or Fe3O4 nanoparticles (MNPs) because of their well-known superparamagnetic properties, low price, and relatively facile synthesis methods. Magnetite (Fe3O4) nanoparticles can be prepared using various methods, including quick and simple precipitation with an ammonia solution,[6,7] as well as thermal decomposition,[8,9] resulting in particles with a well-defined shape and size. By covering MNPs with a silica or polymer shell, they can be easily functionalized with organic compounds and inorganic particles.[10−12]

One example of MNP functionalization is coating them with a luminescent layer to prepare luminescent–magnetic nanosystems. Depending on the material used for coating, luminescence emission can be obtained, with parameters like color, lifetime, stability, and so forth, and selected based on the future application of multifunctional particles. Various organic and inorganic materials were proposed as the luminescent phase like quantum dots (QDs),[10,13−17] inorganic nanophosphors based on lanthanide(III) ions (Ln3+), organic dyes, and d- or f-block metal complexes.[18−21] Each of these luminophores show different structural, physicochemical, and luminescent properties, which strongly affect the properties of the final multifunctional material, mainly its particles size, biocompatibility or overall strength, and stability of the luminescence. Because of the unique luminescence properties of luminophores based on Ln3+ ions, for example, narrow emission bands, relatively long luminescence lifetimes, multicolor emission, and large energy gaps between excitation and emission bands,[22] which result from 4f–4f electron transitions, luminescent–magnetic nanomaterials utilizing the luminescent properties of lanthanide ions are of particular interest to scientists as good materials for bioapplications.[23−25] These luminescent features can be further altered by external factors, such as high pressure,[26,27] temperature,[28,29] and plasmonic effects,[30] making them useful in bioimaging and sensing (e.g., pressure, temperature). In addition, inorganic lanthanide-based phosphors show lower cytotoxicity compared to heavy metal-based QDs and much higher photophysical stability than organic dyes, QDs, or gold nanoclusters, making them a promising replacement for these materials.

In this paper, we provide a new concept for improving bifunctionality, luminescence of Ln, and magnetism of Fe3O4 by using oligomeric-linear chains of helical Ln3+ complexes[31−38] as the luminescent component, which was assembled using a facile one pot method on the surface of Fe3O4@silica core–shell nanoparticles. These complexes and their analogous were shown for the first time in 2014.[31−33] The authors proposed and prepared a series of organic ligands consisting of two bipyridine moieties bridged by ethylendiamins, which after binding with different Ln3+ ions took a specific helical shape. What is more, researchers discovered that the luminescence of the Tb3+ complex is strongly temperature dependent. These studies were strongly supplemented with computational studies done by Hatanaka et al.[37] These complexes can be linked together with terephthalic acid (bdc) into chains, whose structure was confirmed by the authors.[35] By using different ratios of Eu3+-to-Tb3+ complexes and drastic thermal sensitivity of luminescence properties of the Tb3+ complex mentioned above, the mixed-Ln chains exhibited various luminescence colors with decreasing temperatures. Recently, these molecular oligomeric-brush chains were aligned on a quartz substrate, and their luminescence and structural properties were examined.[38]

The challenge and successful concept of our work is shown in Figure . We used oligomeric-brush chains of helical complexes of Eu3+ and Tb3+ ions as a luminophore coating for Fe3O4@silica nanoparticles to realize their bifunctionality. The chain structure of the material allowed us to deposit more luminescent phases on the surface of nanoparticles and resulted in stronger luminescence (compared to a simple monolayer of complexes). The luminescence properties of the prepared luminescent–MNPs were examined at room temperature (RT) and temperature of liquid nitrogen (77 K). Unusual behavior was observed for Tb3+ complexes deposited on Fe3O4@silica nanoparticles in RT, where the decrease of the luminescence intensity was not as large as could be predicted, the excitation spectrum was significantly different, and luminescence lifetimes become longer. These differences were not observed for the Eu3+ complex. The deposition of the luminescent phase did not visibly affect the size of particles, which is a great advantage of these structures compared to inorganic luminescent materials and other coordination polymers used for coatings. Coatings with the presented chain complexes can be used for temperature-sensing[35] applications and for luminescent layers, where small sizes and intensive luminescence are needed at the same time.

Figure 1

Concept of the one-pot preparation of bifunctional MNP with Ln-luminophores. The structure of the helical ligand allows the linkers to be connect only from two sides, which makes the chain structure the only one possible. Silica prevents the intermetal/intercomponent interactions between luminophores and Fe3O4.

Results and Discussion

Transmission electron microscopy (TEM) images comparing each sample are presented in Figure (for additional images see Figure S1). Magnetic cores made of magnetite (Fe3O4) have a roughly similar size, about 10 nm (Figure a), and tend to accumulate in larger clusters. After coating them with silica with surface amine groups (Fe3O4@SiO@amine), several dozens of them can be encapsulated in a single silica NP (Figure b). The size of these nanoparticles is approximately 40–60 nm and no noncovered MNPs were observed. MNPs coated with silica and chains of Eu3+ complexes (Fe3O4@SiO@EuL–bdc) are shown in Figure c. No further increase in size was observed in this case. Energy-dispersive X-ray (EDX) mapping (Figure S2) reveals the correct distribution of silica on magnetic cores and Eu3+ complexes on the silica shell. Zeta potentials for Fe3O4@SiO@amine and Fe3O4@SiO@LnL–bdc were estimated to be −8.2 and +9 mV, respectively. The change in zeta potential was the result of replacing negatively charged terminal amine groups with positively charged carboxyl groups. Also, the positively charged dicarboxylic bridges (bdc) were used for stacking helical Ln complexes into chains.

Figure 2

Images of the prepared samples. TEM images for (a) Fe3O4, (b) Fe3O4@SiO@amine, and (c) Fe3O4@SiO@EuL–bdc nanoparticles. (d) Photographs illustrating at the same time magnetic and luminescent properties of the prepared samples with Eu3+ (up) and Tb3+ (down) ions in acetonitrile in daylight and under UV light (λex = 254 nm). Video clips of the luminescent–magnetic samples under UV light are available in Supporting Information. TEM images were made by K.G. and T.T.; photographs under daylight and UV light were taken by S.G.

Figure d shows the prepared samples of MNPs coated with Eu3+ and Tb3+ in acetonitrile in daylight and UV light (λex = 254 nm). As can be seen, the prepared nanomaterials simultaneously exhibit attraction to a weak magnet and intensive luminescence. After synthesis, the luminescent phase is strongly bound to Fe3O4@SiO and cannot be washed out despite repeated rinsing with acetonitrile.

Synchrotron PXRD patterns for MNPs coated with Eu3+ and Tb3+ chain complexes are shown in (Figure a,b respectively). The patterns for both samples are almost equal to each other and equal to the pattern for silica-coated MNPs (Figure c). Coating with chain complexes does not affect PXRD results, due to the small amount of the luminescent material deposited on the surface of nanoparticles (PXRD patterns of chain complexes are shown in Figure S3). The patterns for these 3 samples consist of the broad reflection at around 8–25° for amorphous silica and multiple narrow reflections for cubic Fe3O4 (Figure d); Fd3̅m space group (COD #96-722-8111).[39−44] The broadening of iron oxide reflections is associated with the small sizes of magnetite particles. Using the Scherrer equation[45]where D is the grain size, k is the shape factor (0.9 for spherical particles), λ is the radiation wavelength, θ is the diffraction angle, β is the full-width at a half maximum, and β′ is an apparatus effect, the average size of magnetic cores was estimated to be 9.8 ± 0.2 nm, which is consistent with the particle size observed earlier on TEM images (Figure a).

Figure 3

Synchrotron PXRD patterns. (a) Fe3O4@SiO@EuL–bdc, (b) Fe3O4@SiO@TbL–bdc, (c) Fe3O4@SiO@amine, and (d) Fe3O4 with the reference pattern.

Attenuated total reflectance (ATR) UV–vis absorption spectra of Fe3O4@SiO@LnL–bdc for both Eu3+ and Tb3+ complexes show the band localized at around 300 nm (Figure a), which is the sum of the absorption of the ligand and bdc linkers (Figure b). These results confirm that both Ln3+ complexes and linkers are present on the surface of Fe3O4@SiO NPs. The bands visible for the Fe3O4 sample, which can be assigned to the Fe2+ (≈245 nm) and Fe3+ (≈460 nm) ion species, are not visible in the final product because of the characteristics of the measurement method, in which the signal is collected mainly from the surface the particles. In the configuration we used, the evanescent wave does not reach the magnetic cores covered with the ∼20 nm thick silica shell.

Figure 4

ATR UV–vis absorption spectra. (a) Fe3O4@SiO@LnL–bdc, where Ln = Eu3+ or Tb3+, and (b) Fe3O4 NPs, the ligand of EuL and bdc linkers.

The emission spectra of the prepared samples show bands typical for Eu3+ (5D0 → 7F, where J = 0, 1, 2, 3, 4) and Tb3+ (5D4 → 7F, where J = 6, 5, 4, 3) ions, resulting from 4f–4f transitions (Figure ). A large decrease in luminescence intensity is observed for the Fe3O4@SiO@LnL–bdc samples, which is the result of a low content of chain complexes in relation to the mass of the whole particle and high absorption properties of magnetic cores. Interestingly, the decrease in the luminescence intensity in the case of Fe3O4@SiO@TbL–bdc is ≈10 times smaller than for Fe3O4@SiO@EuL–bdc measured at RT. We did not expect to observe any luminescence visible by the naked eye for this sample because TbL–bdc exhibits high emission intensity only at a low temperature (77 K) [quantum yield ≈ 64% in 77 K and 22% in RT,[31] which should decrease even more in the Fe3O4@SiO@TbL–bdc system. The difference in luminescence intensity between Tb3+ and Eu3+ samples may be based on the different 4f–4f electronic structures of these two ions. In the case of the TbL complex, the difference in energy levels between the triplet level of the ligand (≈21,000 cm–1)[31] and the excited level of the Tb3+ ion (≈21,700–16,600 cm–1) is very small and may be affected by the changes in the local environment caused by magnetic particles. At the same time, the difference in energy levels is significantly greater for the EuL complex and therefore no unusual behaviour is observed for this case. The nature of the TbL complex makes its luminescence sensitive to changes of temperature, which is a consequence of backward energy transfer from Tb3+ ions to the ligand (Figure ).

Figure 5

Emission spectra. (a, black) Fe3O4@SiO@EuL–bdc at RT and (b, black) at 77 K (λex = 332 nm), (c, black) Fe3O4@SiO@TbL–bdc at RT, and (d, black) at 77 K (λex = 338 nm); the red and green lines correspond to EuL–bdc and TbL–bdc, respectively. The intensity of the signals from the luminescent–magnetic samples in the spectra was multiplied to allow the comparison of the band shape.

Figure 6

Scheme of possible energy transitions. In the proposed system, containing Eu3+ or Tb3+ ions a backward energy transfer was observed only for the Tb3+ samples. IC and ISC are the internal conversion and intersystem crossing, respectively.

The excitation spectra of the luminescent–magnetic samples differ from the spectra of the initial complexes not bound to the surface of the NPs (Figure S4). In the case of the Eu3+ sample, a slight change in the excitation band can be observed, which results from the π-electronic system of the bipyridine moieties. In the case of Fe3O4@SiO@TbL–bdc, a significant change in the shape of the excitation spectrum is observed compared to the unbound TbL–bdc in both RT and 77 K. The right side of the main excitation band begins to decrease earlier, which may directly affect the phenomenon of emission intensity described earlier. Comparing the excitation spectra of the TbL–bdc complex from our work and Marets et al.,[38] we can conclude that this unusual effect was not caused by the immobilization of oligomeric complexes on the silica surface, because in both cases we can observe a maximum around 300–325 nm, but rather magnetic NPs inside the silica. Some changes in TbL–bdc excitation bands measured at RT and 77 K (green lines on Figure S4c,d) are caused by the small distance between ligand donor levels and acceptor levels in Tb3+ ions, which makes Tb3+ helical complexes luminescence sensitive to temperature changes. In the spectra, we can also observe absorption bands related typically to the f–f transitions in Eu3+ (∼363, ∼376, and ∼395 nm) and Tb3+ ions (∼330, ∼345, and ∼385 nm).

Luminescence decay curves for almost all samples (Figure S5) were biexponential. The average luminescence decay times measured for the nanostructures obtained were shorter than for helical chains not bound to the surface of the nanoparticles. Interestingly, the Fe3O4@SiO@TbL–bdc sample in RT exhibits a much longer average lifetime than the initial TbL–bdc, whose lifetime may be affected by a stronger backward energy transfer in RT. Additionally, very short (≈0.19 ms) and very long (≈2.6 ms) components of luminescence decay times could no longer be observed (Table ). In the case of the Fe3O4@SiO@EuL–bdc sample at RT, the immobilization of the chained complexes on Fe3O4@SiO nanoparticles promotes the existence of a component with a shorter lifetime but not at 77 K. This observation also applies to the Fe3O4@SiO@TbL–bdc sample at 77 K. What is quite unexpected, in this case, there was also a third, very short lifetime component (≈0.18 ms) typical for TbL–bdc in RT. In a previous report, Marets et al.[38] suggested that changes in lifetimes between LnL–bdc and SiO@LnL–bdc and their biexponential character may be due to the immobilization of chains on the silica surface and the mutual interactions between them. These interactions promote shorter lifetimes for longer chains, as some unwanted nonradiative complex–complex energy transfers may occur. At the same time, the earlier mentioned phenomenon of emission intensity and changes in the excitation spectra may work the opposite and promote longer lifetimes in the Fe3O4@SiO@TbL–bdc sample, which is drastically different from TbL–bdc.

Table 1

Luminescence Lifetimesa

	RT	77 K
EuL–bdc	τ1 = 0.88 ± 0.04 ms (23.3%)	τ1 = 0.77 ± 0.04 ms (22.2%)
	τ2 = 1.53 ± 0.08 ms (76.7%)	τ2 = 1.61 ± 0.08 ms (77.8%)
	τav = 1.38 ± 0.07 ms	τav = 1.42 ± 0.07 ms
Fe3O4@SiOx@EuL–bdc	τ1 = 1.41 ± 0.07 ms (28.4%)	τ1 = 0.68 ± 0.03 ms (27.5%)
	τ2 = 0.66 ± 0.03 ms (71.6%)	τ2 = 1.54 ± 0.08 ms (72.5%)
	τav = 0.87 ± 0.04 ms	τav = 1.30 ± 0.06 ms
TbL–bdc	τ1 = 0.19 ± 0.01 ms (93.1%)	τ1 = 0.50 ± 0.03 ms (17.4%)
	τ2 = 2.55 ± 0.13 ms (6.9%)	τ2 = 1.32 ± 0.07 ms (82.6%)
	τav = 0.35 ± 0.02 ms	τav = 1.18 ± 0.06 ms
Fe3O4@SiOx@TbL–bdc	τ1 = 0.43 ± 0.02 ms (62.5%)	τ1 = 0.18 ± 0.01 ms (24.0%)
	τ2 = 1.66 ± 0.08 ms (37.5%)	τ2 = 0.66 ± 0.03 ms (35.2%)
	τav = 0.89 ± 0.05 ms	τ3 = 1.38 ± 0.07 ms (40.8%)
		τav = 0.84 ± 0.04 ms

Lifetimes were measured for LnL–bdc and Fe3O4@SiO@LnL–bdc (λex = 340 nm, λem = 617 and 543 nm for Eu3+ and Tb3+ ions, respectively).

The magnetic measurements carried out show that the chains of helical complexes have weak paramagnetic properties (Figure S6), the sample with Tb3+ ions is characterized by higher values of magnetization (M) than the Eu3+ sample in relation to the applied magnetic field (H). However, the prepared luminescent–MNPs show a magnetization curve with a shape typical of superparamagnetic nanomaterials (Figure ). This is a consequence of using nanosized single-domain magnetite cores (9.8 ± 0.2 nm), whose superparamagnetic properties dominate over the weak paramagnetism of the luminescent phase. The magnetization curves for all samples do not show any hysteresis loop, and the differences in the magnetization values between them are the result of a lower content of Fe3O4 per sample. On the basis of magnetic measurements, the mass ratio for Fe3O4/SiO/chains was estimated to be 1.0:2.7:0.5. Chains include both complexes and linkers in the 1:1 M ratio.

Figure 7

Magnetization curves of prepared NPs. Fe3O4 (brown), Fe3O4@SiO@amine (purple), Fe3O4@SiO@EuL–bdc (red), and Fe3O4@SiO@TbL–bdc (green) at 300 K; the red and green lines are overlapping.

Conclusions

Nanoparticles with luminescence easily visible to the naked eye and with good magnetic properties were prepared by anchoring the chains composed of helical complexes and terephthalic acid on the surface of Fe3O4@SiO NPs. The observed decrease in luminescence intensity for the sample with Tb3+ ions at RT compared to the initial helical chain was much smaller than that detected for the same sample at 77 K and the sample with Eu3+ ions at both RT and 77 K. Unusual behavior was also noted for luminescence lifetimes and excitation/emission spectra. The mean luminescence decay time measured for nanostructures with Tb3+ ions in RT was longer than that measured for complex chains unbound to the silica surface. We associated the observed changes in luminescence properties with the immobilization of complexes on the silica surface and the magnetic particles inside silica beads. The described phenomenon was caused by the difference in the 4f–4f electronic structure between these lanthanide ions, where even small changes in the local environment may affect the luminescence properties of Fe3O4@SiO@TbL–bdc, which has a triplet level of the ligand and the excited state of the Tb3+ ion placed very close to each other.

Magnetic measurements show that chains built of helical complexes have paramagnetic properties, while the cores show superparamagnetic properties. The prepared luminescent–magnetic samples are superparamagnetic because the mass of the superparamagnetic magnetic core is about half of the total mass of the sample. What is significant, no increase in particle sizes was observed. This is important in biomedical applications where strong luminescence and small size of particles are required.

The chained structure of the luminophore allows the construction of more refined nanosystems because the building blocks of the complexes can be coordinated with another central Ln3+ ion, even within a single chain.[35,38] What is more, the chains can be terminated with any molecules having functional groups that can strongly bind to lanthanide ions (e.g., carboxyl, N-oxides) present in the helical molecule. This feature may be important for multifunctional nanomaterials and has high potential for the development for structures in the near future. Prepared luminescent–magnetic materials show large possible applications in bioimaging, where for observing cells one can use both intensive luminescence resulting from 4f–4f electron transitions in Ln3+ ions, as well as good magnetic properties of magnetite. Additionally, nanoparticles functionalized by chains of helical complexes are capable of further development through exchange/modification of building blocks or linkers.

Experimental Section

Materials

Glutaric anhydride (98%) was purchased from Tokyo Chemical Industry. Terephthalic acid (95%) and triethylamine (99%) were obtained from Wako Pure Chemical Industries. Dimethylformamide (DMF), acetonitrile (anhydrous), and methanol were of reagent grade and purchased from Tokyo Chemical Industry and Wako Pure Chemical Industries.

Synthesis of Fe3O4@SiO Nanoparticles with the Surface Carboxyl Group

Fe3O4, Fe3O4@SiO@amine nanoparticles and helical complexes of lanthanide(III) ions were prepared based on previously reported procedures.[6,7,31,46]

To functionalize Fe3O4@SiO@amine with carboxyl groups, 3.3 g of glutaric anhydride and 100 mg of Fe3O4@SiO@amine nanoparticles were dispersed in 100 mL of DMF. The solution was sonicated for 5 min and subsequently heated for 6 h on a water bath (60 °C). With the help of a magnet, the product was washed 5 times with DMF and dried for 1 h in a vacuum desiccator.

Synthesis of Fe3O4@SiO@LnL–bdc

Previously synthesized Fe3O4@SiO@carboxyl nanoparticles (20 mg) were immersed for 1 h in 10 mL of 100 mM NaOH aqueous solution. Subsequently, with the help of a magnet, the NPs were washed 5 times with acetonitrile and immersed in 10 mL of a 1 mM LnL acetonitrile solution for 30 min. In a separate volumetric flask, 1.6 mg of terephthalic acid and 2.8 μL of triethylamine were dissolved in 10 mL of ethanol (1 mM acid solution). The terephthalic acid solution was slowly added dropwise to the earlier prepared mixture of MNPs@silica@carboxyl and LnL in acetonitrile. After adding all the acid solution, stirring was continued for 1 h. With the help of a magnet, the product was washed 5 times with acetonitrile and dried overnight in a vacuum desiccator. For comparison purposes, chains of helical complexes (LnL–bdc)[35] not bound to the NP surface were also prepared by dropwise addition of 10 mL of terephthalic acid and triethylamine ethanol solution to a stirred 10 mL of 1 mM acetonitrile solution of LnL. After the dropwise addition was completed, stirring was continued for 2 h. The white precipitate was filtered using a filtration membrane and washed with a small amount of cold methanol.

Characterization

TEM was performed on a JEM-2100F (JEOL) at a maximum accelerating voltage of 200 kV. EDX spectra were collected using a Si(Li) detector as implemented in the TEM apparatus (JEOL, JEM-2100F). Luminescence spectra were recorded on a HORIBA Jobin Yvon Fluorolog 3-22 with excitation via a xenon flash lamp with a band-path filter (λex = 340 nm), using the same amounts of powder samples for measurements. Emission decay curves were obtained using a Quantaurus-Tau C11367-12 (Hamamatsu Photonics K.K.) with a LED light source for the measurement of emission lifetimes, respectively. Zeta potential in acetonitrile was measured using a Malvern Zetasizer Nano ZS (Smoluchowski approximation was used for calculations). Synchrotron X-ray powder diffraction (XRPD) data were acquired using a large Debye–Scherrer camera installed at the SPring-8 BL02B2 beamline, employing an auto sampler with a multiple MYTHEN detector.[47] ATR UV–vis absorbance spectra were collected with a SIC SIS-5100 surface and interface spectrometer system with a quartz slide used as a sample support (an incident Xe light was entering the quartz slide at an angle of 60°). In ATR techniques, the deepness of analysis depends on how far the evanescent wave can propagate through the sample, which depends on many factors, such as incident light angle, a sample support type (e.g., glass, quartz, zirconia, sapphire), light wavelength, sample types, and other.[48] Magnetization measurements were performed at 300 K for each sample, wrapped with a nonmagnetic paper, using a SQUID magnetometer (Quantum Design MPMS). All measurements were performed on the powder samples.