The Sackler Colloquium on promises and perils in nanotechnology for medicine

Our Own Work in New Nanomaterials: Upconversion Phosphors (UCPs)

One of the authors of this paper, S.-f.L., presented her work on upconversion nanoparticles at the Sackler Conference, and we thought it would be informative to present part of that talk as an example of what is happening in new nanomaterials for biology and medicine. UCP are ceramic materials in which rare earth atoms (from the 15 elements with atomic numbers 57–71 comprising the lanthanide series) are embedded in a crystalline matrix. The rare earth ions have optical transitions arising from buried atomic 4f electronic orbitals, which typically absorb ≈970–1,000 nm in a series of lines of ≈2- to 5-nm width. Upconversion occurs via sequential excitation of a series of real, as opposed to virtual, levels as in conventional 2-photon dyes (8). The upconversion mechanism into the visible spectral region can occur as sequential excitation of the same atom or more typically excitation of 2 centers and subsequent energy transfer (9–11). The emission of UCPs consists of sharp lines characteristic of atomic transitions in a well-ordered matrix. Using different rare earth dopants, a large number of distinctive emission spectra can be obtained with colors from red to blue. Because the upconversion process occurs sequentially through real states, relatively low-power densities can be used for efficient upconversion, and UCPs have recently attracted substantial interest in both academia and industry for the purpose of labeling biological specimens (12, 13). Because the emitted light is upconverted to shorter wavelengths than the excitation wavelength, background fluorescence is virtually nonexistent.

The main advantages of UCP nanoparticle imaging for use in cells and tissue are the lack of autofluorescence by the specimen, absence of bleaching, multicolor imaging, and high-resolution cathodoluminescent imaging in a SEM (14). With the interest in imaging organisms at the subcellular level (15), nanoparticles in the sub-10-nm range are required, such as UCPs. Currently a great impediment to using UCPs in biological labeling applications is their size, a point that quantum dots and organic dyes have excelled in. Solution-stable, unaggregated, bright upconverting, oxide-based phosphors have previously been demonstrated for submicron sizes. As one might expect from the titanium dioxide story we related, the crystal structure of these nanomaterials is quite important for how they function, but unfortunately in ways we do not understand. In the case of the UCPs, the crystal matrix is of importance, because it is believed that low phonon energy can reduce nonradiative loss due to the multiphonon relaxation and thus enhance the upconversion quantum efficiency (17, 18). The phonon spectrum of a matrix strongly depends on the crystallographic structure, and as we have seen for titanium dioxide, the thermodynamically stable phase is a function of particle size and is influenced by the solvent and the surface/volume ratio increases.

Thus, achieving a combination of the sub-10-nm particle size, low aggregation of the nanoparticles, size uniformity, high crystallinity required for high upconversion efficiency, and efficient surface functionalization has proven elusive to the present. However, we have been able to synthesize sub-10-nm, low-aggregating, upconverting nanophosphors that can be suspended in a solvent at very low aggregation levels. We synthesized the sub-10-nm rare earth-doped yttrium oxide-based nanophosphors by a flame spray pyrolysis process, using a liquid precursor. To facilitate complete and even combustion at sufficient flame temperatures, we chose a precursor consisting of ethanol as a fuel and yttrium and rare-earth complexes with 2-ethylhexanoic acid (2-EHA) as a source of the host and dopants, respectively. The reaction itself proceeded in a custom-built 10-cm-long alumina Laval tube (19). The Laval tube was chosen because it achieves an ultrashort exposure of the synthesized particles to high temperatures and thus suppress Ostwald ripening and particle aggregation. In our process, the organic precursor aerosol is ignited in the apex of a Laval tube, and combustion takes place in the exit cone. The temperature of the pyrolysis as measured by a type K thermocouple was ≈1,200°C at a distance of 1 inch downstream from the apex of the Laval tube. In the process, the precursor is first atomized in O₂ using a nebulizer (Micromist, Glass Expansion). The jet is then directed at the entrance of a Laval tube made of alumina. A methane torch provides the ignition for the flame, and the reactor temperature is controlled with a resistive wire heater. Phosphors were collected on a stainless steel cylinder kept at ≈100° C.

Fig. 3 shows a high-resolution transmission electron microscopy (HRTEM) of the nanoparticles and the Fourier transform diffraction pattern. In the HRTEM image, facetted spheroidal nanoparticles with well-defined lattice structures are seen. The measured lattice spacing of the selected nanocrystal is 3.16 Å, consistent with the known lattice spacing of the (222) plane of cubic Y₂O₃ (Powder Diffr File JCPDS 65-3178, International Center for Diffraction Data). That orientation also appears to be the most frequently observed lattice orientation among the other nanocrystals shown. The corresponding particle size distribution, calculated from >220 nanoparticles, is shown in Fig. 4A. The distribution appears to demonstrate highly uniformly sized nanocrystals, with a mean size of 6.3 ± 1.3 nm.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

HRTEM image of upconverting nanophosphor from the 2-EHA based precursor, showing selected nanoparticle and its corresponding Fourier transform diffraction pattern. The nanophosphor size and distribution were analyzed by a Philips CM200 FEG-TEM. Nanophosphors for electron microscopy were collected during spray pyrolysis directly onto TEM Cu-grids coated with holey carbon film.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Analysis of nanoparticle sizes. (A) Particle size distribution of nanophosphors obtained from the 2-EHA based precursor. (B) Volume distribution of sub-10-nm nanophosphors suspended in THF, with and without SDS. Particle size distributions in suspension were measured by using dynamic light scattering on a Malvern Zetasizer Nano-ZS (Malvern Instruments).

The X-ray diffraction (XRD) spectrum in Fig. 5 shows several diffraction peaks that can be indexed to the standard Y₂O₃ body-centered cubic (bcc) phase. The bcc phase combined with the spheroidal shape of the nanophosphors obtained is a result of the high-synthesis temperature of our flame spray pyrolysis system. The nanoparticle size as calculated from Scherrer's equation (20) was ≈9 nm. The Y₂O₃ lattice constant, obtained by calculations using XRD data of the nanocrystals, is 10.5 Å, which is consistent with that of bulk Y₂O₃. The disparity of the sizes as measured by TEM and XRD can be attributed to the difference in quality of the nanophosphors, which is due to the different sample collection times during flame spray pyrolysis. For TEM analysis, only a small quantity of nanophosphors was collected directly onto TEM grids for ≈15 sec. In contrast, for XRD analysis, a much larger quantity of nanophosphors was required, which had to be collected over several hours during spray pyrolysis. Over the long time period, a slight change in spray conditions may have produced variations in size uniformity. This results in a larger size estimated by XRD analysis. However, as shown by TEM data, maintaining size uniformity is possible if spray conditions are strictly controlled. This can be achieved, for example, by installing mass flow controllers to ensure constant flame temperature and aerosol droplet size. Elemental analysis was performed on the flame spray-pyrolyzed upconverting nanophosphors by energy-dispersive X-ray spectrometry (EDX) on a Princeton Gamma Tech Integrated Dispersive X-Ray Microanalyzer (Princeton Gamma Tech, Princeton). The sub-10-nm phosphor showed preferential doping of the rare earth Yb over the rare earth Er (8.4 atomic percentage) doping over Er (6.8 atomic percentage) doping.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

The XRD spectrum showing sub-10- (broken line) and 40-nm particles (solid line).

As it is our aim to use these nanophosphors for biological imaging; the ability to suspend these particles in aqueous media stably is crucial. Hence, an understanding of the behavior of the material in water is required. The isoelectric point (IEP) of yttrium oxide ranges from 8.6 to 9.1 (21). Below this pH range, there is a strong possibility of dissolution of yttrium oxide. The rate of dissolution will be further accelerated by the larger surface area of nano-sized yttrium oxide. At a pH above the IEP, we expect dissolution to be still significant for ultrasmall particles. In most biological applications, the pH used is usually below the pH range tolerated by yttrium oxide. Therefore, these nanophosphors have to be surface-modified to confer robustness over a large pH range. To that end, we have a silanized PEG coating on our nanophosphors when used with buffer solutions. However, this step is not necessary if the nanophosphors are suspended in nonaqeous solvents. For silanization of the 10-nm nanophosphors, 1 mg of the nanophosphors was dispersed in 1 mL of deionized water by probe-tip sonication at 10–12 W (root-mean-square) output power for ≈5 sec (Model 100, Fisher Scientific Sonic Dismembrator). To the suspension, an additional 2.3 mL of deionized water was added, followed by 30 mL of methanol and 200 μL of methoxy(polyethyleneoxy)propyltrimethoxysilane (PEG-silane) (SIM6492.7, Gelest). The suspension was stirred for 2 h, heated to 60°C for 5 min, and then cooled. The nanophosphors were washed with centrifugation, twice with methanol, twice with methanol:deionized water (1:1), and twice with deionized water, and finally concentrated using centrifugal filter devices (Amicon Ultra-15, molecular weight cutoff 100,000). The resulting PEG-coated nanophosphors in water were transparent to the eye and displayed luminescence when excited with 980-nm illumination.

We have characterized in solution the PEG-silanized nanophosphors by using a combination of light scattering and zetapotential measurements. We have measured 50- to 60-nm diameter-sized PEG-silanized nanophosphors in water (Fig. 6), with a zetapotential of +17 mV at pH 5 and −37 mV at pH 9. With light scattering measurements, the hydrodynamic radius measured in solution is larger than that measured by the TEM in powder form, because of the balance of van der Waals forces, viscous flow, and Brownian motion in solution.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Volume distribution of PEG-silanized nanophosphors suspended in ultrapure water showing pH dependence. Particle size distributions in suspension were measured by using dynamic light scattering on a Malvern Zetasizer Nano-ZS (Malvern Instruments).

Of course, synthesis of nanoparticulates is one thing, but what they do in living tissue is quite another issue and, as we discussed in the sad story of asbestos, it can be quite surprising and discouraging to find that a nanomaterial can be quite toxic to living systems despite the wonderful properties. At this point, we have done only simple tests as to the influence of these nano-UCPs in tissue, in our case, nematides. The viability of the PEG-coated nanophosphors was tested in the digestive system of the nematode worm C. elegans. A similar protocol used in our previous studies (11) was followed for inoculation of the worms and transferring onto slides and for imaging using 2-photon microscopy. The dependence of the luminescence intensity was determined by integrating the emission from 1 particle in the field of view and varying the illumination intensity. Upconversion luminescence spectra were collected by using a fiber-coupled CCD spectrometer (Ocean Optics). Fig. 7A shows the visible emission from the PEG-coated sub-10-nm nanophosphors in the digestive system of the C. elegans worm under infrared illumination. The worms were monitored continuously for up to 24 h without any apparent change in the phosphors, as in blinking or bleaching, and in the condition of the worms. Because the worms were deprived of food, excretion also ceased, and the nanophosphors were still retained in them. Of course, if we are to use these UCPs as agents in applications ranging from tissue visualization to chemotherapy technologies will have to be developed for the specific labeling of the nanoparticle surface with tissue-specific probes.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 7.

Images of nanoparticles with C. elegans. (A) False-color 2-photon images of C. elegans at 980-nm excitation with green representing the bright field and red for the PEG-silanized sub-10-nm nanophosphor emission at 60× magnification. (B) Combined image of 40-nm nanophosphor-fed worm with gray representing the SEM image, and red representing the cathode luminescence (CL) image. The CL image was obtained by monitoring the total light intensity from the sample by using a photomultiplier tube, without any color filtering. The sample was tilted at a 45° angle to the incoming beam, leading to some topography-induced contrast in the CL image.