Spectroscopic properties of Fe3+ in GGG and the effect of co-doping with rare-earth ions

The spectroscopic properties of Fe3+ in tetrahedral d-sites of GGG and the effect of co-doping with ~ m 3 + are investigated. It is thus shown that a very efficient energy transfer from ~ e 3 + (d) to ~ m ~ + takes place due to a good superposition of Fe emission and Tm 3 ~ 6 > 3 ~ 4 absorption spectra and due to a favorable packing of the d-sites around the dodecahedral site occupied by Tm. It was also found that the Fe-sensitized luminescence of Tm has different spectral and temporal characteristics from the normal emission of this ion in GGG. Trivalent thullium in garnets, either as activator (A) or sensitizer (S) for H O ~ + offers interesting prospects for two-micron laser emission. Unfortunately the absorption bands of Tm are weak and thus the pump efficiency is low. A common way to improve the pump efficiency is the sensitization with transition ion metals which show strong and broad absorption matching better with the existing pump sources. A very important condition for sensitization is a fast S-A energy transfer and this implies a good match of S emission and A absorption spectra, short S-A distances and a good packing of A ions around the sensitizer. The garnet crystals offer two sites for the transition ions, the octahedral a-site (of local C3i symmetry) and the tetrahedral d-site (S4), the preference for substitution being determined by the mismatch of ionic radii of the dopant and host cation and on the electronic structure of the former. The d-sites show a better packing around the dodecahedra1 c-site (D2 symmetry) occupied by the rare-earth ions and the minimal d-c distance is shorter than the octahedral a-c distance. ~ r 3 f , the usual sensitizer for ~ m 3 + , occupies only the octahedral sites in garnets, i.c. not the most favorable sites for sensitization. This work investigates the possibilities of Tm sensitization by Fe3+ which could occupy both the aand d-sites in these lattices. We used in this investigation GGG crystals grown by Czochralski technique and doped by ~ e 3 + (0.5 at %) or ~e3+/0.5 at %) Tm3+(5 at %). The lowest spectral term of the ground electronic onfiguration 3d5 of ~ e 3 + is a spin sextet ( 6 ~ ) while the excited states are spin quatets or doublets. This leads to the forbiddeness of transitions between the ground and excited states. Hoever the spinorbit mixing of the sextet and quartet spin states lessons to a given extent this interdiction for sectet->quartet transitions. The corresponding terms are split by the cubic component of the crystal field according to the rules 6s->6A1, 4 ~ > 4 ~ ~ , 4 ~ > 4 ~ + 4 ~ 2 , ~ F > ~ A ~ + ~ T ~ + % ~ , and the Tanabe-Sugano diagrams which show the position of the crystal field levels as function of the crystal field strength are similar for octahedral and tetrahedral coordination. The crystal field states are further split by the spin-orbit interaction.6~~->r7+r8; 4~2->Tg ;4 ~ > T 6 + n+ Tg ; 4~2->r6+r /+2~8 and the lower symmetry component of the crystal field could split further the T8 quartets into doublets. Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jp4:1994478 JOURNAL DE PHYSIQUE IV The tetrahedral site lacks inversion and thus electric dipole transitions between states of the ground 3d5 electronic configuration are allowed owing to the mixing from a add-parity states. However, in the octahedral sites, which have inversion, these transitions are forbidden, unless lower-symmetry perturbations are present. Thus in case of the isolated Fe3+ centers in diluted garnets we could expect that the optical spectra will be dominated by the tetrahedral centers. Indeed, weak and broad absorption bands covering all the visible spectrum have been observed, in our samples except for a strong and sharp line at 421.26 nm (accompanied by a structural phonon sideband), which was attributed to the ~ A I > ~ T ~ ( ~ D ~ ) transition of Fe3+ (d) center. We also remark the beginning of a broad band at about 535 nm (attributed to 6 ~ 1 > 4 ~ 2 (4G) transition) and several non-resolved lines in the region 470-490nm, connected with ~ A I > ~ E , 4A1(4G) transitions. A very intense absorption band is seen in the region of 250nm and, as in other ~ e ~ + systems [1,2] and this could be attributed to charge transfer transitions inside the octahedral and tetrahedral centers. As remarked for other systems, this band could lend intensity to the intra 3d5 transitions close to it. A strong emission was observed by pumping in any of the lines observed in absorption and it consists of three relativity sharp lines at 753.3nm, 754.8nm and 756.8nm, companied by a broad phonon sideband with several shoulders (figure 1). The origin of this triplet is not clear but its presence at very low temperatures shows that it could not be connected with a structure of the emitting state. At the same time this could not be due to the crystal field splitting of the ground state, which usually is of the order of 10-2 cm-1. The static excitation spectrum for this emission (figure 2) evidences clearly the transitions seen also in absorption. The wavelength of the sharp lines of emission correspond to a 4 ~ 1 ( 4 ~ ) > 6 ~ 1 transition for a ratio (D2/B) in the TanabeSugano diagram around 1, i.e. a value characteristic for tetrahedral Fe3+ in octahedral sites (D2/B,2). With increasing temperature the sharp emission Fe3+ lines broaden while the phonon sideband broadens and gains strength so as at 300K it dominates completely the emission and extends over all the range from 760 to 850nm. The luminescence decay is exponential with a lifetime of about 4.5ms at 300K. The emission band of Fe3+ (d) shows a very good superposition with Tm 3 ~ 6 > 3 ~ 4 absorption, especially a t the room temperature ; this suggests the use of Fe3+ in tetrahedral positions as sensitizer for Tm. Co-doping GGG with Fe3+ (0.5 at %) and ~ m 3 + (5 %) has a very strong effect upon Fe3+ and ~ m 3 + absorption and emission. Thus a well resolved satellite (T) of the sharp absorption M line 6A1->4~2 (4D) of Fe3+ was observed at 419.56 nm ; similar satellites could be expected in other Fe3+ transitions but they cannot be resolved because the large linewidth. Excitation in the transition ~ A ~ > ~ T Z (4D) of Fe at 421.26 nm leads to an emission similar to that observed in Feonly doped samples, but the lines are broadened and dips at the wavelengths corresponding to Tm 3 ~ 6 > 3 ~ 4 absorption have been observed in the broad emission sideband. Besides the radiative Fe->Tm transfer testified by these dips, a strong non-radiative transfer, manifested by a marked modification of Fe3+ luminescence decay takes place, the effect being stronger at higher temperatures. The observed decay could be fitted with a Forster 8t1/2 law, corresponding to an electric dipole interaction between donor and acceptor only at long time (longer than about 500ps) after the beginning of decay, which is much faster than predicted by this law. By contrary, no Fe3+ emission was observed by exciting in the T satellite at 419.56nm. Excitation in Fe absorption lines leads also to ~ m 3 + 3 H 4 > 3 ~ 6 emission (no emission was observed from higher energy levels of Tm) but this is dominated by the presence of three new centers (observed also in absorption), whose transition Wl->Zl is shifted from the normal line N (794.98nm) to 795.33nm (center Fl), 795.29 nm (center