Phase stabilization in transparent Lu2O3:Eu ceramics by lattice expansion

Phase stabilization in transparent Lu2O3:Eu ceramics by lattice expansion

Optical Materials 35 (2012) 74–78 Contents lists available at SciVerse ScienceDirect Optical Materials journal homepage: www.elsevier.com/locate/opt...

891KB Sizes 11 Downloads 52 Views

Optical Materials 35 (2012) 74–78

Contents lists available at SciVerse ScienceDirect

Optical Materials journal homepage: www.elsevier.com/locate/optmat

Phase stabilization in transparent Lu2O3:Eu ceramics by lattice expansion Z.M. Seeley ⇑, Z.R. Dai, J.D. Kuntz, N.J. Cherepy, S.A. Payne Chemical Sciences Division, Lawrence Livermore National Laboratory, United States

a r t i c l e

i n f o

Article history: Received 6 June 2012 Received in revised form 18 July 2012 Accepted 24 July 2012 Available online 15 August 2012 Keywords: Transparent ceramic Lutetium oxide Gadolinium oxide Densification Sinter-HIP

a b s t r a c t Gadolinium lutetium oxide transparent ceramics doped with europium (Gd,Lu)2O3:Eu were fabricated via vacuum sintering and hot isostatic pressing (HIP). Nano-scale starting powder with the composition GdxLu1.9xEu0.1O3 (x = 0, 0.3, 0.6, 0.9, 1.0, and 1.1) were uniaxially pressed and sintered under high vacuum at 1625 °C to obtain 97% dense structures with closed porosity. Sintered compacts were then subjected to 200 MPa argon gas at temperatures between 1750 and 1900 °C to reach full density. It was observed that a small portion of the Eu3+ ions were exsolved from the Lu2O3 cubic crystal lattice and concentrated at the grain boundaries, where they precipitated into a secondary monoclinic phase creating optical scattering defects. Addition of Gd3+ ions into the Lu2O3 cubic lattice formed the solid solution (Gd,Lu)2O3:Eu and stretched the lattice parameter allowing the larger Eu3+ ions to stay in solid solution, reducing the secondary phase and improving the transparency of the ceramics. Excess gadolinium, however, resulted in a complete phase transformation to monoclinic at pressures and temperatures sufficient for densification. Light yield performance was measured and all samples show equal amounts of the characteristic Eu3+ luminescence, indicating gadolinium addition had no adverse effect. This material has potential to improve the performance of high energy radiography devices. Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction The field of transparent ceramics is emerging with applications in many areas such as solid state laser host materials [1], optical lenses [2], transparent armor [3], radiation detectors [4], and radiography scintillator screens [5]. Polycrystalline ceramics possess improved mechanical properties, offer wide range of compositions, lower cost of fabrication, less wasted material, and near net shape forming methods as compared to single crystal oxides. However, there are some limitations when achieving transparency with a polycrystalline microstructure [6]. A cubic crystal structure is necessary to avoid light scattering at grain boundaries and birefringence. Also, even a trace amount of secondary phase or residual porosity is detrimental to transparency. To accommodate these limitations, care has been taken to reach extremely high purities and density with close control over the stoichiometry of the ceramics. Rare earth sesquioxides are a unique class of materials for their catalytic, magnetic, electronic, and optical properties, and their existence in various polymorphic forms is well characterized [7,8]. In the cubic bixbyite structure, these oxides are capable of being formed into transparent ceramics. Lutetium oxide doped with europium (Lu2O3:Eu) has become a material of interest as a dense, transparent scintillating radiography screen [9]. In our ⇑ Corresponding author. E-mail address: [email protected] (Z.M. Seeley). 0925-3467/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.optmat.2012.07.005

previous work, we demonstrated that vacuum sintering followed by hot isostatic pressing was a successful method to make highly transparent Lu2O3:Eu [10]. Further investigation of the transparency of this material revealed a small amount of optical scatter, on the order of a = 0.1 cm1. We are working to develop an X-ray CT screen optimized for the MeV range, with lens-coupled optical imaging readout, which requires a scintillator screen in the form of a fully transparent sheet, about 3 mm thick. In order to fully realize the potential of Lu-based ceramics for high resolution MeV-range CT screens, the optical scatter must be reduced by a factor of 10. Incorporation of gadolinium into the Lu2O3 lattice has recently been employed to enhance the phase stability and energy migration within the material [11–14]. Transparent ceramics of (Gd,Lu)2O3:Eu have been reported to show improved scintillation time response, with lower afterglow than Lu2O3:Eu, and potential for higher light yield, however little information is given on the transparency of the processed (Gd,Lu)2O3:Eu ceramics. In the present study, we have found that incorporation of gadolinium into the Lu2O3:Eu lattice reduces scattering defects, resulting in highly transparent ceramics, by preventing the europium from segregating into a secondary phase. 2. Experimental Nanoparticles with the composition GdxLu1.9xEu0.1O3 (x = 0, 0.3, 0.6, 0.9, 1.0, and 1.1) were synthesized via the flame spray

Z.M. Seeley et al. / Optical Materials 35 (2012) 74–78

pyrolysis (FSP) method by Nanocerox™ (Ann Arbor, MI). A Philips CM300-FEG high resolution transmission electron microscopy (HRTEM) operated at 300 kV was used to characterize the nanoparticles for structure and crystallinity. Nanoparticles were suspended in an aqueous solution containing polyethylene glycol (PEG) and ammonium polymethacrylate (Darvan C-N) using an ultrasonic probe (Cole Parmer, Vernon Hills, Il) and a high shear mixer (Thinky, Japan). This suspension was spray-dried (Buchi, New Castle, DE) at 210 °C into flowing nitrogen to protect the organics. The dried powder was then sieved (<50 lm) resulting in uniform agglomerates of nanoparticles with an even distribution of organic additives. Formulated nanoparticles were then uniaxially pressed at 50 MPa to form green compacts approximately 35% dense, followed by a heat treatment at 900 °C in air to burn out the organics. Calcined compacts were then loaded into a tungsten element vacuum furnace (Thermal Technologies, Santa Rosa, CA) and sintered under a vacuum of <2  106 Torr at 1625 °C for 2 h to reach closed porosity and densities of approximately 97%. The sintered samples were then hot isostatically pressed (HIP’ed) under 200 MPa of inert argon gas pressure at temperatures ranging between 1750 and 1900 °C for 4 h in a tungsten element HIP (American Isostatic Presses, Columbus, OH). Since the samples were closed porosity after vacuum sintering, no canning was necessary during the HIP step. Ceramic surfaces were ground flat and parallel, given an inspection polish, and wipe-cleaned with acetone and methanol. Samples were analyzed for defects with visible microscopy (Nikon). A focused ion beam (FIB) was used to extract a slice from within the Lu2O3:Eu sample and analyzed under HRTEM in order to determine the chemical composition and crystalline structure of the scattering defects. X-ray diffraction (XRD) was used to determine the crystallinity and lattice parameter of HIP’ed samples. The cubic lattice parameter for each composition was calculated from the four most intense cubic peaks using Bragg’s law and averaged. Scatter loss measurements using a 633 nm He–Ne laser and a 400 inch integrating sphere were performed on HIP’ed ceramic samples. Scatter coefficients, including both bulk and surface scatter, were determined using Beer’s law and were normalized by the sample thickness. Beta radioluminescence employed a 90Sr/90Ysource (1 MeV average beta energy). Radioluminescence spectra were collected with a Princeton Instruments/Acton Spec 10 spectrograph coupled to a thermoelectrically cooled CCD camera. Further details are explained elsewhere [15]. Light yields are obtained by integrating the spectra and comparing with several known scintillators (IQI-301 glass and crystals of BGO and LuAG:Ce) in the same geometry.

75

Fig. 1. Bright field TEM micrograph and selected area electron diffraction pattern of (A) Lu1.9Eu0.1O3, and (B) Gd0.9Lu1.0Eu0.1O3 FSP particles.

3. Results and discussion Flame spray pyrolysis (FSP) powder synthesis provides a high quality starting powder for forming transparent ceramics. As can be seen in Fig. 1, this method produces uniform particle size and composition in equiaxed particles with very little agglomeration. The average particle size of Lu1.9Eu0.1O3 is 20 nm with a specific surface area (SSA) of 22 m2/g, and the selected area electron diffraction pattern indicates a cubic crystal structure. With the addition of gadolinium into the precursor, the ‘‘as-synthesized’’ particle size increases in average size to about 30 nm with an SSA of 18 m2/ g, and displays monoclinic crystal structure. This phase transformation is expected due to the larger atomic radius of gadolinium and quenching from high temperature during FSP synthesis [7,16]. However, this monoclinic phase transforms back to the cubic phase during subsequent processing which is evident by XRD and the formation of bulk transparent ceramics described below.

Fig. 2 shows the transparency of the HIP’ed samples after polishing. For clarity, it is specified that this image is not backlit and no post-annealing (oxygen bleaching) step was required. Transparent ceramic samples were obtained on the Lu-rich end of the composition range due to the tendency of these compositions to form a fully dense homogeneous cubic bixbyite structure. Optical transparency on the Gd-rich end of the compositional range began to degrade as these compositions are more likely to form a monoclinic phase at high temperatures due to the larger average ionic radius [16]. As the HIP temperature increases, less gadolinium is required to trigger this phase transformation from cubic to monoclinic. Initial monoclinic phase precipitation causes light scattering and reduced transparency, but with sufficient gadolinium and temperature the entire sample becomes monoclinic and ultimately results in complete opacity. In a previous study we found that optimization of sinter and HIP processing led to transparent Lu2O3:Eu [10,17]. However, in the

76

Z.M. Seeley et al. / Optical Materials 35 (2012) 74–78

Fig. 2. Photograph of HIP’ed and polished samples as a function of HIP temperature and gadolinium concentration. Sample thickness is 4 mm and samples are not backlit.

Fig. 3. Optical microscope images of the interior of the transparent ceramics depicted in Fig. 2. Illumination is in transmission and images are focused into the center of the sample. Dark regions are sources of light scatter.

present study closer examination using optical microscopy reveals a small amount of secondary phase remaining in the Lu2O3:Eu resulting in some residual scattering defects. Upon careful inspection of the transparent samples, a slight cloudy appearance was observed in samples containing no gadolinium, and those that were HIP’ed at lower temperatures (1750 °C). Fig. 3 shows optical micrographs taken of the sample interior. Illumination for these images is in transmission mode, and images are focused below the surface into the center of the sample. Therefore, dark spots and regions are sources of light scatter in the bulk of the sample. In most of the samples HIP’ed at 1750 °C, pores are visible between grains. This indicates that at 1750 °C, diffusion was not sufficient to HIP these samples to full density. Residual porosity remains and is the source of scatter causing these samples to look cloudy. As the HIP temperature increases to 1850 °C, the pores are fully closed but a small amount of secondary phase remains at the grain boundaries as scattering defects. The amount of this secondary phase appears to decrease with increasing gadolinium content and correlates with the relative transparency of the samples in Fig. 2. The samples which appear most transparent have sufficient gadolinium to minimize this secondary phase at the grain boundaries, but not enough to cause the phase transformation to the monoclinic phase, i.e. Gd0.9Lu1.0Eu0.1O3 HIP’ed at 1850 °C and Gd0.6Lu1.3Eu0.1O3 HIP’ed at 1900 °C. Fig. 4 shows optical scatter of the HIP’ed samples measured using a 633 nm He–Ne laser and a 4 inch diameter integrating sphere. This data confirms the visual appearance provided by the photograph of the samples, shown in Fig. 2. Samples HIP’ed at

Fig. 4. Optical scatterometry (633 nm) of compacts after hot isostatic pressing at 1750, 1850, and 1900 °C as a function of gadolinium concentration.

1750 °C show the highest amount of scatter, with the Gd0.9 composition being the most transparent. Due to the fact that pores are the principal source of scatter in samples sintered at 1750 °C, this indicates that the Gd0.9 composition HIP’ed to higher density than samples with less gadolinium. For this composition, this HIP condition is close to the phase transformation from cubic to monoclinic improving the ionic mobility which is the likely cause for improved sintering. Samples HIP’ed at 1850 °C display less scatter due to the absence of residual porosity and again, the Gd0.9 composition has the best transparency. Samples HIP’ed at 1900 °C show even less scatter due to enhanced pore removal, however at this temperature, the Gd0.9 composition forms significant secondary phase and therefore scatter is minimized at the Gd0.6 and Gd0.3 compositions.

Z.M. Seeley et al. / Optical Materials 35 (2012) 74–78

77

Fig. 6. Cubic lattice parameter of Gdxlu1.9xEu0.1O3 as a function of gadolinium concentration calculated from the four most intense XRD peaks.

Fig. 7. Beta radioluminescence spectra and integrated light yield for transparent Lu2O3:5%Eu with several concentrations of gadolinium addition. Fig. 5. HR-TEM image of the secondary phase (SP) at the triple point between three cubic phases of Lu2O3:Eu (G1, G2, and G3). Selected-area electron diffraction patterns of the Lu2O3:Eu grain (G3) and the secondary phase (SP) are also shown.

A common trend between Figs. 2–4 is that scatter increases as the Gd content is reduced to zero for all HIP temperatures. From the optical microscopy images in Fig. 3, the small scattering defects appear to be located along grain boundaries and at triple points between several grains. Use of a focused ion beam to locate, cross-section, and create a TEM slice of one of these scattering defects allowed a closer examination. Fig. 5 shows a micrograph of a scattering defect located at the triple point between 3 grains. Along with a small pore, this micrograph reveals a secondary phase approximately 500 nm in size. Selected area electron diffraction patterns confirm that the 3 surrounding grains are cubic with a lattice parameter of 10.55 Å, closely matching that of Lu2O3:Eu, however the secondary phase exhibits a monoclinic structure. Furthermore, X-ray EDS analysis (not shown) indicates that the secondary phase shown in Fig. 5 has 3 times higher europium concentration than the surrounding Lu2O3:Eu grains. Because this secondary phase is present in significantly larger proportion, compared to that present in our starting powder, we know that it must have formed during the processing stages. This indicates that during the grain growth of sintering and HIP’ing, some of the europium is being exsolved to the Lu2O3 grain boundaries. This process is likely due to the mismatch in ionic radii between Lu3+ (86 pm) and Eu3+ (95 pm), where the larger europium ions are unable to easily fit into the cubic Lu2O3 lattice. As the grains continue to grow some of the europium preferentially segregates to the grain boundaries where defects in the lattice structure are more accommodating for the larger ion. Europium can then travel along the grain boundaries to the triple points, where eventually

the local concentration increases to the point of precipitation into a stable Eu-rich monoclinic phase. As seen in Figs. 2–4, addition of gadolinium into the Lu2O3:Eu system reduces the amount of scattering secondary phase present at the grain boundaries. In order to understand the influence of gadolinium, X-ray diffraction on several of the HIP’ed samples was used to study the crystal structure and lattice parameter. XRD patterns reveal that all samples are composed of a single cubic phase; however a systematic shift in the 2-theta value with increasing gadolinium concentration was noticed. Fig. 6 shows the effect of gadolinium concentration on the average lattice parameter calculated from the four most intense cubic peaks with (h k l) values of (2 2 2), (4 0 0), (4 4 0), and (6 2 2). At low gadolinium concentrations, it is clearly seen that the lattice parameter follows Vegard’s rule and increases in a linear fashion as the gadolinium concentration increases, consistent with previous findings [11]. As the gadolinium concentration increases (above x = 0.8), the lattice parameter is unable to maintain the same linear trend suggesting that at these concentrations the gadolinium is no longer going into complete solid solution with the lutetium, and is starting to segregate to the grain boundaries. It is hypothesized that this increase in lattice parameter allows the large europium ion (only slightly larger than the gadolinium ion) to fit easier into the cubic lattice and stay in solid solution during the sintering and densification processes, thus reducing the formation of secondary phase scattering defects. In order to determine the effect of gadolinium addition on light yield, radioluminescence spectra were taken for several samples after HIP’ing, as shown in Fig. 7. For clarification, no post treatment anneals of any kind were necessary. All samples show the characteristic Eu3+ emission with little variation in total integrated light

78

Z.M. Seeley et al. / Optical Materials 35 (2012) 74–78

yield indicating that gadolinium has no adverse affect on the luminescence mechanism. 4. Conclusion Highly transparent (Gd,Lu)2O3:Eu ceramics were formed by the sinter-HIP method. Transparency of the samples was dependent on the HIP temperature as well as the concentration of gadolinium. Either too much or too little gadolinium resulted in formation of a monoclinic secondary phase. Our results suggest that ionic size mismatch causes europium ions to exsolve from the cubic Lu2O3 lattice during ceramic processing and densification. An Eu-rich monoclinic secondary phase precipitated at the grain boundaries and caused optical scattering defects in the ceramic. Incorporation of gadolinium into Lu2O3:Eu promoted a solid solution of ions in the cubic bixbyite structure by stretching the lattice parameter to accommodate the larger europium ions. The formation of secondary phase scattering defects was reduced and the composition Gd0.6Lu1.3Eu0.1O3 resulted in a highly transparent ceramic with excellent luminescence, well-matched to the requirements for MeV X-ray CT imaging radiography. Incorporation of excessive gadolinium resulted in complete phase transformation to monoclinic. Acknowledgements Thanks to Todd Stefanik of Nanocerox Inc., Jeff Roberts for flame spray synthesis, Scott Fisher for mechanical fabrication, Marcia Kellam for optical scatter measurements, and the Confined Large Optical Scintillator Screen and Imaging System (CoLOSSIS) team including Patrick Allen, James Trebes, Daniel Schneberk, and Gary Stone. This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344 and funded by the US DOE, Office of NNSA, Enhanced Surveillance Subprogram. LLNL-JRNL-559892.

References [1] A. Ikesue, Y. Aung, T. Yoda, S. Nakayama, T. Kamimura, Fabrication and laser performance of polycrystal and single crystal Nd:YAG by advanced ceramic processing, Opt. Mater. 29 (2007) 1289. [2] U. Peuchert, Y. Okano, Y. Menke, S. Reichel, A. Ikesue, Transparent cubic-ZrO2 ceramics for application as optical lenses, J. Eur. Ceram. Soc. 29 (2009) 283. [3] R. Klement, S. Rolc, R. Mikulikova, J. Krestan, Transparent armour materials, J. Eur. Ceram. Soc. 28 (2008) 1091–1095. [4] N. Cherepy, S. Payne, S. Asztalos, G. Hull, J. Kuntz, T. Niedermayr, S. Pimputkar, J. Roberts, R. Sanner, T. Tillotson, E. Loef, C. Wilson, K. Shah, U. Roy, R. Hawrami, A. Burger, L. Boatner, W. Choong, W. Moses, Scintillators with potential to supersede lanthanum bromide, IEEE Tran. Nucl. Sci. 56 (2009) 873. [5] C.W.E. van Eijk, Inorganic scintillators in medical imaging detectors, Nucl. Instrum. Methods Phys. Res. A 509 (2003) 17–25. [6] A. Ikesue, Y.L. Aung, T. Taira, T. Kamimura, K. Yoshida, G. Messing, Progress in ceramic lasers, Annu. Rev. Mater. Res. 36 (2006) 397–429. [7] G. Adachi, N. Imanaka, The binary rare earth oxides, Chem. Rev. 98 (1998) 1479–1514. [8] M. Zinkevich, Thermodynamics of rare earth sesquioxides, Prog. Mater. Sci. 52 (2007) 597–647. [9] T. Martin, P. Douissard, Z. Seeley, N. Cherepy, S. Payne, E. Mathieu, J. Schuladen, New high stopping power thin scintillators based on Lu2O3 and Lu3Ga5xInxO12 for high resolution X-ray imaging, IEEE Trans. Nucl. Sci. (2012) 1–6. [10] Z. Seeley, J. Kuntz, N. Cherepy, S. Payne, Transparent Lu2O3:Eu ceramics by sinter and HIP optimization, Opt. Mater. 33 (2011) 1721–1726. [11] L. Qin, Y. Wu, H. Shi, W. Chai, K. Shu, G. Ren, X. Chen, Effects of doping Lu2O3 into Gd2O3 on phase transformation and luminescence, IEEE Trans. Nucl. Sci. 56 (2009) 2979–2982. [12] Y. Wu, L. Qin, H. Shi, Y. Zhang, Y. Yang, K. Shu, Preparation of europium-doped gadolinium lutetium oxide solid solution transparent ceramics and its optical properties, IEEE Trans. Nucl. Sci. 57 (2010) 1343–1347. [13] H. Retot, S. Blahuta, A. Bessiere, B. Viana, B. LaCourse, E. Mattmann, Improved scintillation time response in (Lu0.5Gd0.5)2O3:Eu3+ compared with Lu2O3:Eu3+ transparent ceramics, J. Phys. D: Appl. Phys. 44 (2011) 235101. [14] H. Retot, A. Bessiere, B. Viana, B. LaCourse, E. Mattmann, 5d–4f and 4f–4f emissions in Ln-doped sesquioxide ceramics, Opt. Mater. 33 (2011) 1008– 1011. [15] N.J. Cherepy, J.D. Kuntz, J.J. Roberts, T.A. Hurst, O.B. Drury, R.D. Sanner, T.M. Tillotson, S.A. Payne, Transparent ceramic scintillator fabrication, properties and applications, Proc. SPIE 7079 (2008) 70790X. [16] S. Schneider, R. Roth, Phase equilibria in systems involving the rare-earth oxides. Part II. Solid state reactions in trivalent rare-earth oxide systems, J. Res. National Bureau of Standards A. Phys. Chem. 64 (A) (1960) 317. [17] Z. Seeley, J. Kuntz, N. Cherepy, S. Payne, Transparent Lu2O3:Eu ceramics, Mater. Res. Soc. Symp. Proc. 1341 (2011) u01.