BRIDGMAN GROWTH AND PROPERTIES OF LuAlO3-Nd3+ LASER CRYSTALS

The solidification behavior of LuA103 melts in Bridgman geometry is presentedThe contribution of different growth parameters to crystallization of single phase LuA10, are studied. Growth scheme for L U A ~ O ~ N ~ ~ + crystals possessing improved laser properties is described. LuA103 (space group ~:E-~bnm) is the end member of the rhombic rareearth group (LnA103, where Ln = Y or Ln ion between Gd and Lu). Five compounds from this group have been so far reported as laser host crystals, i.e. YA103 , GdA103, ErA103, (Er,Lu)AlO, and LuA103/1/, with lasing ions pr3+, Nd3+, Dy3+, HO~+, IZr3+ and ~m~+/l.~/Growth of LuA103 Nd3* single crystals was first reported in 3 A rod 18mm long and 5 mm in diameter cut from Czochralski grown material lased at 1-0832um (Ethr.= 8J) and 1-3437pm (Ethr.=15J) at 300K /4/ thus offering generation wavelength shift to long-wave part of the spectrum as compared with YA103~d~+/l/Owing to this potential usefulness of LuA10,k4d3+ can be extended if large size high-quality material is availableIn view of results of M-Leduc, L-Shearer on optical polarization of helium atoms, L U A ~ O ~ N ~ ~ + based lasers have prospects of application in this fieldLuAIOS was prepared also by Bridgman technique /5/. More recently Nd3+doped optical quality material was obtained by this technique which reveal substantially better laser perArticle published online by EDP Sciences and available at http://dx.doi.org/10.1051/jp4:1991799 C7-380 JOURNAL DE PHYSIQUE IV formance /6/ than that reported by Czochralski grown material. This paper summarizes various aspects of Bridgman single crystal growth of ~ u A 1 0 ~ ~ d ~ * for laser applications. 2--Experimental DetailsVertical Bridgman technique has been used in the present work to study spontaneous solidification behaviour of LuA103 melts and growth conditions of laser quality LuA103-~d~~ crystals. The main experimental conditions were the same as described earlier in /5/Additional variable parameters included cyclic melt overheating then solidification on cooling processes to study statistical characteristics of freezing end extra crystal. annealing at the last stages of crystallization to study LuA103 stability to high temperaturas on cooling from the melting point. 3.-Results and Discussion. The previous results of investigations of LuA103 /3, 5, 7 9/ show that the main unclearness concerns nucleation and growth mechanism of this phase (it is formed assumingly when crystallization follows some metastable path on cooling of the melts in which the garnet and lutetium oxide phases are the most stable, while LuA103 has even no stability field in subsolidus region of the phase diagram) and high temperature stability of LuA103 upon cooling from the melting point (on heating from the room temperature crystalline LuA103 readily decomposes at L 1300C, separating garnet and lutetium oxide phases)Our experiments with melts lcm3 in volume have shown that by analogy with some metals the ability of melts to become supercooled before solidification occurs, depends on the overheating degree and it's length. The supercooling degree increases with increase of the overheating temperatureThe overheating length acts in the same direction as the overheating temperatureNote that LuA103 phase originate only when solidification occures at high ( 2 100 K) melt supercoolings with respect to the melting point /9/. For comparison, YA103 structure is readily formed on cooling of 1:l melts irrespective to the freezing temperature / 5 / The observed dependence together with results obtained for the melts with garnet composition /5/ show that probably no change in coordination number of cations occurs in the melts at high overheating temperaturesMore evidently nucleus-size particles of stable phases which are present in initial melts disappear completely during overheating process giving way to simplicity principle /lo/ to determine solid phase structure. In Bridguian growth the melt is cooled from the bottom so it is here where LuAlCb may originate first acting further as a seed crystal. Experiments have shown that after LuAIOS phase is formed the melt undercooling is being removed due to rapid solidification of the part of the initial melt and replacement of the interface up to the level corresponding to the equilibrium melting point of Ld103. This means that solid Ld1O3 which is nucleated at temperatures some 2100 K below the melting point does not undergo any phase decomposition on heating from nucleation temperature to that close to the melting point. Growth runs which included additional crystal annealing at the cooling stage of the grown material have shown that LuA103 conserves transparency with no signs of decomposition up to temperatures some 200-300 K below the melting point. At lower. temperatures of annealing it undergoes complete phase decomposition as it is the case when LuA103 is heated from the room temperatureHowever the observed behaviour of solid LuA103 below the melting point cannot be considered as demonstration of thermodynamic stability of this phase in the mentioned temperature regionMore evidently, the phase decomposition rate has a complex dependence on the temperatureIt may be said that the rate of Lu,A150i2 nucleus formation in solid LuA103 has a maximum at some temperature between 1700 and 1300C while at higher and lower temperatures it is going downThe amount of new phase is dependent on the rate of nucleus formation, their growth velocity and time interval, so at low temperatures the growth velocity, which is diffusion limited, is decreased due to decrease of atoms mobilityThe observed stability of solid LuA103 at higher temperatures needs additional investigations. The properties of LuA103, revealed in this investigation, show that crystal growth can be carried out by imploying a scheme presented in Fig.1It is based on the observed phenomena of conservation of the nucleated crystallizat.ion centers of LuA103 at temperature increase, i.eformation of these centers in supercooled melts and their conser-C 7 3 8 2 JOURNAL DE PHYSIQUE IV vation during temperature increase close to the melting point. Fig-1-Growth scheme of LuA10, single crystalsTm -melting point, AT+ -melt overheating, A T -melt supercoolingArrows indicate the direction of temperature change at each stage of crystallization. On this base conditions are found which offer crystallization to occur near the equilibrium, which in practice permit to obtain large size optical quality crystals of rhombic LuA103, as well as other metastable aluminats of the lanthanide familyThe next point shows that Bridgman technique is more preferable than Czochralski in view of application to LuA103 growth. In Bridgman geometry the melt supercooling necessary to nucleate the metastable phase is self removed, while in Czochralski arrangement an empirical selection of thermal conditions is needed. Spectral and generation experiments with Bridgman grown L U A ~ O ~ N ~ ~ + /6/ have shown that these crystals have substantially better laser performance than those reported earlier for Czochralski grown materialIn the pulse mode a rod 5mm in diameter and 40mm long lased at the wavelengths of the main and additional channels of t?d3+ ions at 1.0677 (qhr.=O.8 J, 77 K), 1-0675(2 J, 300 K), 1-3429 (1-5 J, 100 K) and 1-34371qu (3 J, 300 K). The melt solidification properties of metastable LuA103 are studied. Basing on obtained experimental results a growth scheme of LuALo,-N~~+ crystals using Bridgman technique is proposed, which permits crystallization to occur near to the equilibrium and to obtain large size material of improved optical and laser quality. The author wishes to thank Professor A-A-Kaminskii who initiated this investigation and Professor A-A-Chernov for discussions around unusual properties of this crystalline phase. The X-ray phase identification has been performed by DrG-0-Shirinyan.