1. Introduction
Upconversion (UC) is the process by which a system, be it an atom, an ion, or a molecule, through absorbing low-energy quanta from the environment, either simultaneously or in successive steps reaches a state from which the de-excitation, either to the ground or to an intermediate level, is achieved by emitting a higher energy quanta compared with the absorbed ones. The energy absorption process is a succession of events like ground-state absorption (GSA), and one or multiple excited-state absorptions (ESAs). The absorptions of the energy quanta could be resonant, phononic-assisted, sensitized, or cooperative sensitized etc. [
1].
Different nanoparticles with UC properties, designed for either high conversion efficiency or tunable color, have been studied for biomedical applications, such as biosensing and bioassays, high-contrast agents, drug delivery, and various therapies [
2]. In this area, the use of IR radiation has some major advantages over its counterparts that use higher energy radiation (UV or blue light) due to a deeper penetration into biological tissue and because it is much safer.
The UC of the Er
3+:Yb
3+ pair, in various crystalline and biological compatible hosts, is used in medical engineering, where it acts as an in situ nanoscale, high-fidelity temperature sensor, by probing the
2H
11/2,
4S
3/2→
4I
15/2 fluorescence emissions intensity ratio, which depends on the energy level populations, which in turn are influenced by temperature [
3,
4,
5,
6,
7,
8,
9]. Also, Er
3+:Yb
3+ upconverting nanoparticles, due to their low energy excitation wavelength, can act as 980 nm excited triggers for drug delivery when combined with drug carrier particles [
10].
In optical engineering, e.g., in fiber optics signal amplifiers, Yb
3+ is used to enhance the Er
3+ emission of
4I
13/2→
4I
15/2 (1500 nm), using 980 nm as the pumping radiation, which is cheaper to obtain and has better efficiency [
11,
12,
13,
14,
15].
Er
3+:Yb
3+ UC nanoparticles were also applied in pollutant detection and removal [
16], efficiency increasing of the photocatalytic activity of TiO
2 [
17], lamp phosphors, display panels, catalysts, sensors, and photovoltaic process enhancers [
18].
The main goal of the studies of the Er
3+:Yb
3+ activator–sensitizer pair is to increase the UC efficiency of Er
3+ in incident near-infrared (NIR) radiation. The NIR light is easily obtained today with cheap laser diodes. Many crystals were used as hosts for this ionic pair, e.g., Y
2O
3, BaGd
2ZnO
5, and NaYF
4 [
19,
20,
21,
22,
23], each having a certain efficiency that is linked to the phononic interactions between Er
3+ and Yb
3+ and the crystalline lattice host.
Nevertheless, a subtlety of the mechanism of the UC process, i.e., why the red–green emission ratio of Er
3+ ions is higher when the relative concentration of Yb
3+ to that of Er
3+ is increased, is not well understood. Only one paper was found addressing this issue [
24], but it is not explanatory enough.
In order to better grasp the causes of this phenomenon, another crystal host, namely Y
2TiO
5 ceramic doped with Er
3+ and Yb
3+, is studied in this work. This ceramic was chosen not only because, to the best of our knowledge, the literature lacks information regarding the UC of Er
3+ embedded in this matrix, but also because Y
2TiO
5 is special as the titanates display a wealth of useful characteristics like a large bandgap (≥3.3 eV [
25]), thermal stability [
25], photocatalytic properties, and they do not pose biological hazards [
26]. Also, yttrium titanate ceramics display various phase states, which are very useful in the fission energy industry [
27]. They can accommodate large defects in the crystalline matrix induced by high-intensity particle radiation or by allowing small-molecule gases (helium, hydrogen) to penetrate the lattice [
28].
We selected Er
3+ as the activator and Yb
3+ as the sensitizer because the
2F
5/2→
2F
7/2 transition energy of Yb
3+ is quasi-resonant with the
4I
11/2←
4I
15/2,
4F
7/2←
4I
11/2 or
4S
3/2←
4I
13/2 transitions of Er
3+ [
29], ensuring an easier energy transfer between these ions [
1]. Also, the Er
3+ energy levels are appropriately distanced between themselves [
29] in order for the excited states to have a sufficiently long lifetime, such that absorption of other quanta happens before de-excitation and the UC is facilitated.
The aim of this study was to investigate how the conformation of the Er3+, Yb3 doped Y2TiO5 crystalline matrix influences the UC spectral response of Er3+ upon illumination at 976 nm. Hence, we compared the upconversion spectra of Er3+ in Y2TiO5 (YTO) with those in other host crystalline matrices, either previously studied by us or reported in the literature.
During this endeavor, we found that the UC process for this Y2TiO5 matrix is not linear and displays high sensitivity to the wavelength of the incident IR laser light, with variations as small as 2 nm leading to notable changes in the spectral composition of the upconversion emission in the visible range. This behavior is peculiar for this crystalline lattice, since Er3+ present in other crystalline hosts (e.g., Y2O3, BaGd2ZnO5) displays more linear characteristics, as we observed in our research.
Herein, we report for the first time this nonlinear behavior of the UC process of Er3+:Yb3+ doped ceramics, nonlinearity that is linked to the crystalline structure of Y2TiO5 both at the nanoscale and to the shape of the surrounding coordination polyhedra at the locus of Y3+ substitution.
This research consists also in improving the understanding of the subtleties of the energy transfer between the dopant ions and with the lattice by correlating the experimental data with simulations of the dopant distributions in the crystal and their relative interionic distances.
This nonlinearity could find different applications, such as cheap laser wavelength indicators, checking of structural modifications at the nanoscale level in materials, detecting forgeries, 980 nm laser tuning, or doppler sensors for noninertial fiber gyroscopes.
2. Materials and Methods
2.1. Materials Used and Dopant Concentrations Labelling
The synthesis of Y2TiO5 ceramics, doped with Er and Yb samples, was carried out in two steps: first, by obtaining the precursor metallic oxides using the sol–gel technique; second, these oxides were then molded into pellets using high pressure, and the pellets were sintered at a high temperature and atmospheric pressure.
The reagents used, erbium(III) nitrate pentahydrate, ytterbium(III) nitrate pentahydrate, yttrium(III) nitrate hexahydrate, titanium(IV) butoxide 97% (TB), citric acid (CA), and ethylene glycol (EG) were purchased from Sigma Aldrich (Merck Group, Darmstadt, Germany) and were used as received.
Table 1 presents the chosen concentrations of the dopant ions for replacing yttrium atoms in Y
2TiO
5 and the sample labeling that will be used throughout this text.
YTO 0-0 was chosen as the reference for the crystal structure when all the conditions of synthesis were the same. Also, YTO 0-0 was used to investigate the response of the pure Y2TiO5 matrix at a 976 nm illumination (fixing the background). YTO 0-1 was chosen to investigate the response of Yb3+ ions at the 976 nm range and observe the energy that the emitted photons have. The dopant concentration values were chosen to be small enough such that the Er3+ and Yb3+ ions should replace the Y3+ ions in the Y2TiO5 matrix with the lowest possible lattice distortions. As it will be seen, the lattice is very sensitive to these dopants, and concentration values higher than 10% (relative to Y3+) strongly hinder the equilibrium of the orthorhombic phase of Y2TiO5.
2.2. Synthesis of Oxide Powders
The metal nitrates in the appropriate quantities, CA (2 moles for every mol of metallic ions) and TB were dissolved in EG. The utmost care was taken for no traces of water to be in the reaction mixture since water induces the quick hydrolysis of TB. The reaction mixture was heated at 120 °C under continuous magnetic stirring (300 rpm) for slow evaporation until a transparent and thick gel was formed. The gel was thermally treated at 400 °C for 3 h until the most carbonaceous compounds were removed, and a grayish powder was obtained. The as-synthesized powders were calcinated at 900 °C for 3 h, after which their color changed from grey to pinkish due to the presence of Er3+ ions.
2.3. Obtaining of Er, Yb Doped Y2TiO5 Ceramics
Y2TiO5:Er,Yb ceramic pellets of 13 mm diameter and 1 mm thickness were obtained in two steps: first, the powders were pressed at 200 bar for 2 min, followed by grinding the resulting pellets in an agate mortar; second, the powders were pressed at 100 bar for 1 min with a slow release of the pressure. This pressing procedure was chosen to avoid the brittleness of the resulting ceramics because the initial powders had a very low cohesiveness. The first step ensures an aggregation of the particles into larger ones, whereas the second step eliminated the voids created between them, while inserting smaller inner tensions and preventing the pellet from cracking during manipulation and thermal treatment.
Two pellets for each composition were produced and placed, one on top of the other, in alumina crucibles. This stacking prevents the accidental modification (contamination, evaporation of constituents, etc.) of the composition of the touching faces of the pellets during the prolonged oven time.
The pellets were sintered at 1250 °C for 16 h. This is the minimum time for the sintering treatment for the Y
2TiO
5 crystal lattice stabilization. The temperature was chosen according to the reported results [
30], which showed that the diffusion coefficients for oxygen ions in the Y
2TiO
5 structure are sufficiently high even at temperatures lower than the melting point. This ensures the formation of the orthorhombic phase while impending the segregation of phases with a lower energy of formation (−3.98 eV/atom for Y
2O
3 [
25]) than Y
2TiO
5 (−3.87 eV/atom [
25]) or other phases like hexagonal, pyrochlore, fluorite, or compounds like Y
2Ti
2O
7 and YTiO
3 [
31].
2.4. Characterization of Y2TiO5 Samples
The oxide powders and the ceramics were characterized using X-ray diffraction (XRD) with a Rigaku Miniflex II, Tokyo, Japan, diffractometer with Cu Kα radiation in the 2θ range of 10°–70°, at a 2 °/min rate, and a step of 0.01°. The scanning electron microscopy (SEM) was performed using a Tescan Vega 3LM, Brno, Czech Republic, microscope equipped with an EDS spectrometer. The upconversion spectra were measured using a USB4000CG-UV-NIR spectrometer from Ocean Optics (Orlando, FL, USA) and the OceanView software version 1.6.7. The LCU98E042Ap diode (Laser Components GmbH, Olching, Germany) was driven with emitted power control and thermally stabilized because the lasing wavelength varies with temperature. The thermal control allowed for the IR laser wavelength tunability with subnanometer precision. The ray was collimated with a standard laser collimator, which illuminated an area of 0.5 mm2, allowing for individual microcrystal probing. The acquisition fiber optic head entrance was at a distance of 3 mm from the illuminated spot and had a diameter of 200 µm.
The spectra acquisition periods for the photon counting were 30 ms, and the laser illumination power was 150 mW. All the measurements were averaged over 10 full-spectrum acquisitions, such that a single averaged spectrum was measured in 0.3 s.
These power and time parameters were chosen so as not to saturate the spectrometer sensor in the spectral range of interest (visible). The measurement setup for the upconversion spectra is depicted in
Figure S1 (Supplementary Materials).
3. Results
3.1. Powder Oxides XRD Characterization
The XRD analysis of the resulting oxide powders reveals smooth and wide peaks, indicating the formation of a cubic phase, which is the precursor of the orthorhombic Y
2TiO
5 (formed after the sintering), and is presented in
Figure 1.
3.2. XRD Analysis of Y2TiO5:Er,Yb Ceramics
The XRD patterns of the YTO 0-0, 0-1, 1-0, 1-2, 1-4, and 1-8 ceramics, as well as the simulated orthorhombic YTO 0-0 diffractogram using the VESTA 3 software [
32], can be seen in
Figure 2. The XRD patterns of the other Y
2TiO
5:Er,Yb samples are presented in
Figure S2 (Supplementary Materials). It was found that the sintering process at 1250 °C for 16 h is the optimum thermal treatment to obtain the orthorhombic Y
2TiO
5 phase. Other attempts, i.e., a lower calcination temperature or shorter time, resulted in a mixture of cubic and orthorhombic phases, whereas higher temperatures gave rise to other phases segregation.
In the case of YTO 1-4, the matrix begins to be distorted (a small peak at 2θ = 30° can be observed), and for YTO 1-8, the distortions are even higher. These deformations are compatible with the undoped structure because of the large interstitial (no bond) spaces and the fact that Yb3+ ions have almost a 4% smaller ionic radii than those of Y3+ ions.
3.3. Crystallinity of the Ceramic Pellets
After sintering, the protected face of each ceramic pellet was first sanded with 2000 grit, then polished with 6000 grit until the face was gloss-shining. The procedure revealed, only under frontal illumination with a bright white light, the microcrystal structures of the pellets. An example of the polished face of YTO 1-0 is shown in
Figure 3 (see also
Figure S3), and one can see that the sizes of the microstructures are uniform.
The uniformity of the microcrystal size for each dopant concentration shows a successful sintering process. The variation in the average size of the microcrystals shows how the dopant concentration influences the crystallization process. There is a variation in the microstructure sizes across the cases of dopant concentrations, with the lowest feature sizes being for YTO 1-0 (
Figure 3) and the largest ones for YTO 1-2 (
Figure S3). This is an indication that the doping ions (even if the total dopant concentration relative to Y is below 10%) strongly influence the crystallization process and the crystalline phase formation (
Table S1) revealed by the XRD patterns (
Figure 2).
3.4. Structure of the Y2TiO5:Er,Yb Crystal Unit Cell
Y
2TiO
5 crystallizes in the orthorhombic system with P1 symmetry. The unit cell parameters are:
a = 3.72 Å,
b = 10.45 Å, and
c = 11.35 Å. The CIF data [
25] regarding the Y
2TiO
5 compound were visualized using the VESTA 3 software [
32] and reveal the positions of the ions in the unit cell, as depicted in
Figure 4.
Yttrium ions are coordinated by seven oxygen ions, and Ti4+ ions are coordinated by five oxygen ions. All Y3+ ions have the same geometry of the coordination polyhedra.
It can be observed (
Figure 5 and
Figure S4) that the orthorhombic matrix has large voids, which can easily give rise to lattice defects, nanostructural instability, the accommodation of other kinds of ions, or being permeable to gases with low atomic radii. The ionic radii for the involved metallic ions, coordinated by seven oxygen ions, are provided in
Table 2. Yb
3+ has a 4% smaller radius compared with Y
3+ [
33]. Therefore, larger concentrations of Yb
3+ hinder the shape and stability of the orthorhombic Y
2TiO
5.
The XRD phase matching analysis of the samples performed using the Match! software [
34] shows that, for each ceramic, there is a combination of phases that are compatible with the orthorhombic structure of Y
2TiO
5 (
Table S1). This is due to the interstitial spaces of the structure, as seen in
Figure 4 and
Figure 5, which allow for such a variability of combinations of phases with the same structure of the unit cell but with slightly different geometrical parameters; the percentage of the combinations is also influenced by the dopant concentrations. In the case of YTO 1-8 and YTO 3-6, which have high dopant concentrations, fluorite Y
2Ti
2O
7 and even hexagonal Y
2TiO
5 were formed.
Consulting the phase diagram for Y
2O
3–Y
2Ti
2O
7 [
31], one can conclude that the phase composition is very sensitive to the types of doping ions and their concentrations. Even low concentrations of doping ions can significantly alter the phase equilibrium, and this is due to the metastable conformation of the matrix.
In the cases of YTO 1-8 and YTO 3-6 with a high concentration of Yb
3+, the XRD analysis showed (
Figure 2 and
Figure S2), beside the expected orthorhombic phase, a hexagonal phase and the Y
2Ti
2O
7 fluorite phase. The YTO 1-8 ceramic was obtained also trying slightly different sintering conditions (at a maximum temperature of 1300 °C), but the hexagonal and fluorite phases appeared in all cases, which is an indication that Yb
3+ promoted the formation of the Y
2Ti
2O
7 phase, supposedly because of its lower ionic radius (0.925 Å) than that of the Y
3+ radius (0.960 Å).
3.5. SEM Investigation of the Sintered Powders and Ceramics
The SEM images show that the powders (calcined also at 1250 °C) and the ceramics have very similar aspects, with the pellets being more compact (
Figure 6 and
Figure S5). Comparing with other morphologies shows that the closest matching one is that of volcanic ash [
35,
36].
In all cases, there is no evidence of any forms of large-scale crystals, but only a mix of irregular shards with a large range of dimensions, which is an indication that the good crystal structure, revealed using XRD, is only at the nanoscale level and that these nanocrystals are attached among themselves in a macroscopic phase.
This aspect is allegedly due to the loose structure of the YTO crystal unit cell, which allows for deformations and other inter-nanocrystalline transition phases, which could also be related to the variate compositions revealed using the X-ray diffractograms (
Table S1). Also, it can be observed that the higher the total dopant concentration, the smaller the ceramic grains, while the aspect of the powders remains the same (
Figure S5).
3.6. Upconversion Light on a Wide Area
Prior to the upconversion emission spectroscopic investigation, the pellets were tested by illuminating them with a grazing IR laser light with two wavelengths, 973.5 nm and 975.5 nm, over an area of about 3 mm × 3.5 mm with a power of 175 mW. The pictures are presented in
Figure 7 for YTO 1-2 (for the other cases, see
Figure S6). The microcrystals shine the upconverted radiation differently, revealing a mosaic-like local variation in brightness (best seen in the YTO 1-2 case,
Figure 7A,B) and bright spots in the other cases (
Figure S6).
This brightness variability is due to the size of the crystals and not caused by dopant concentration nonuniformity, because the EDX analysis (
Figure S7) revealed a homogeneous distribution of the dopants. This is notable, especially in the cases with higher concentrations of Er
3+, YTO 2-4, YTO 4-4, and YTO 3-6, with spots having slightly green or red hues, yet the spectral shapes being the same. The stripe variation in brightness is due to the unevenness of the laser field emitted by the diode.
How the dimension of the microcrystal influences the upconversion mechanism remains to be established, but the images strongly suggest that the sizes of the microcrystals are influencing the phononic energy transfers and/or nanodot-like energy confinement.
Reorienting the pellets relative to the polarization of the laser light shows no variability in the brightness of the spots, underscoring the idea that the energy transfers are isotropic, not depending on some preferred light–crystal relative orientations.
3.7. Upconversion Spectra
The specifics of the UC spectra measurement we used (pinpoint focusing the laser light) allowed us to probe the individual spots on the pellets and check the variability of the response spectra across their surface.
The brightest spots on the surfaces of the pellets were targeted with 150 mW illumination power (continuous wave) and the spectra were measured for two close wavelengths, 973.5 nm and 975.5 nm.
Figure 8 and
Figure 9 present the upconversion spectra for the visible part for both values of the excitation radiation wavelength.
The difference between these excitation energy values is about 20 cm−1, but the effects on the upconversion response are important. This fact is an indication that the upconversion mechanism can be linked to some detuning of some yet-unspecified resonant energy transfers between Er3+ and Yb3+ ions. Not only are the upconversion spectral compositions different for the two incident wavelengths, but also the emitted power curves are altered.
This high sensitivity with the incident radiation wavelength of Er,Yb:YTO ceramics is different from the other ceramics (Y2O3, BaY2ZnO5, BaGd2ZnO5, CaGd2ZnO5, BaY2O4, and BaY2MgO5, for the same dopant concentration) that we tested. In the 973.5–975.5 nm excitation wavelength range, these ceramics display a similar upconversion spectra with a linear dependence of the emitted power vs. incident power, unlike Y2TiO5.
Figure 10 presents the peaks (between 840 nm and 882 nm) of the emission of the transition
4S
3/2→
4I
13/2 of Er
3+ and the peaks between 900 nm and 940 nm, which are from the anti-Stokes sideband generated by the emission of Yb
3+ when decaying from
2F
5/2 to
2F
7/2. Their widths and intensities are indications that the phononic interactions are significant. Also, they help to estimate the phonon energy for Y
2TiO
5 at about 610 cm
−1.
It can be seen that, in the case when the excitation was performed at 973.5 nm radiation, the anti-Stokes peaks for Yb3+ are almost double those in the case with 975.5 nm. The YTO 1-8 case is an exception, revealing a diminished interaction between the phonons and Yb3+ ions.
One can observe in
Figure 10 the variability in the
4S
3/2→
4I
13/2 transition of Er
3+ across the dopant concentrations and also in the wavelength of the exciting radiation. For the 975.5 nm excitation, it is observed that a higher Yb
3+ concentration promotes the decay to the ground of Er
3+, instead of being trapped to
4I
13/2. At the 973.5 nm excitation, it is seen that Er
3+ has a greater tendency to reach
4I
13/2, only 8% of Yb
3+ strongly hindering that. The mechanism by which the Yb
3+ ions are influencing the Er
3+ energy level transitions, when embedded in Y
2TiO
5 crystal matrix, will be the subject of further research.
5. Conclusions
Erbium and ytterbium-doped diyttrium titanate ceramic samples were obtained at 1250 °C from oxide powders, which were synthesized using the sol–gel method. The characterization using XRD and SEM showed that the doped Y2TiO5 ceramics’ microstructure and phase composition were very sensitive to the dopant concentrations. When the Yb3+ concentration is increased, additional hexagonal and fluorite phases appear, which fixes the maximum threshold for the Yb3+ concentration at 10%, beyond which the Y2TiO5 orthorhombic phase becomes heavily distorted.
The samples were illuminated with 973.5 and 975.5 nm from a thermally tuned laser diode, and the resulting UC visible spectra were measured and characterized. In the case of Y2TiO5:1%Er3+:2%Yb3+, we observed that the upper Stark level of 4S3/2 of Er3+ becomes more populated when illuminated with 973.5 vs. 975.5 nm, this fact inducing a notable change in the hue and peak intensities in the UC response. Also, the emitted green–red intensities, both relative and absolute, for the other samples were observed and compared for the 973.5 and 975.5 nm irradiations, showing notable differences that were not observed in other crystalline lattices that we studied or found in the literature. The curves relating the emitted intensity versus the excitation power of the incident radiation were not linear, with a saturation tendency beyond 100 mW for green (515–575 nm), in the case of 973.5 nm irradiation, and for red (640–700 nm), in the case of 975.5 nm illumination.
Simulations were performed for the distributions of the interionic distances for Er3+↔Er3+ and Er3+↔Yb3+, and the samples were ordered by the distance of the maximum for each case of the dopant concentrations. It is seen that Er3+↔Er3+ quenching takes place when Er3+ is 4%, regardless of the Yb3+ concentration, showing that Er3+↔Er3+ energy transfers are more prevalent than those between Er3+ and Yb3+. The fact that the upper Stark level of 4S3/2 is more populated when the illumination is performed with 973.5 nm vs. 975.5 nm is puzzling, since the distance between the two levels of 4S3/2, split by the crystal field of Y2TiO5, is more than 20 cm−1, which is the energy difference between 973.5 nm and 975.5 nm.
The source of this energy difference will be the subject of further investigations; nevertheless, the sensitivity of the effect calls for practical engineering applications with a high degree of precision. Some of these are laser diode wavelength tuning, cheap laser diode quality control, laser wavelength indicators, crystalline structural modification indicators when YTO is incorporated in other bulk ceramics, cheap doppler sensors with under 2 nm precision detection for non-inertial optical fiber gyroscopes, and last but not least, forgery prevention by including certain YTO nanodots in the protected samples.