LOP. Cerium - Doped Orthophosphates: New Promising Scintillators. Akmpicki, E.Berman, A. J.Wojtowicz and M.Balcerzyk. Boston University, Chemistry Dept.
IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 40, NO. 4,AUGUST 1993
Cerium Doped Orthophosphates: New Promising Scintillators A k m p i c k i , E.Berman, A.J.Wojtowicz and M.Balcerzyk
Boston University, Chemistry Dept. 590 Commonwealth Ave. Boston MA 02215 and L. A. Boatner Solid State Division, Oak Ridge National Laboratory P.O. Box 2008, Oak Ridge, Tennessee 37831-6056 We report initial results for a new class of scintillating materials, cerium doped rare earth orthophosphates. The most promising is Ce:LuP04,which under y-excitation shows fast, single-exponential decay, of 25 ns at room temperature, and a light output of 17,200 photons per 1 MeV. These features combined with a reasonable density, (6.53 g/cm3),make it a strong contender for many applications.
is concerned. A comparison of the basic characteristics of these scintillatorsis given in Table I.
Lutetium orthophosphate (LuPO,) doped with cerium and other rare earths is a material that has Table 1 COMPARISON OF Ce DOPED LUTETIUM undergone an extensive spectroscopic investigations, COMPOUNDS WITH BGO concerned mostly with Raman scattering and various nonlinear effects [1,2]. The luminescence of LuP0,:Ce BGO LOP LSO has not been investigated in any detail. It does figure, Relative light output 100 217 500 however, in the tables of potential scintillator materials, compiled by Derenzo et al. [ 3 ] .The effective Z number of Emission wavelength this material is 63.7. The present paper is devoted to the 480 360 420 (nm) photoemission and scintillation aspects of this remarkable Decayconstant(nsec) 300 24 40 compound. To our knowledge, it is the second lutetium compound, doped with cerium, which exhibits superior Density (g/cm3) 7.13 6.53 7.4 scintillation properties. The first compound being Ce: Lu,(SiO,)O, otherwise known as "LSO", discovered by Melcher and Schweitzer, .The two compounds share the properties of a high density, a very fast decay time of Index of refraction a few tens of nanoseconds and an efficiency higher than BGO. This makes these materials prime contenders for medical applications such as positron emission The issue of why the concentrated compounds tomography, (PET). Since it is both fashionable and useful to abbreviate the names of these compounds, we show a reduced light output has been addressed by the present authors in several publications, [6, 7, 81. The shall refer to the Lu-orthophosphateas LOP. There are at the present time two generic types of basic observation is that thus far the stoichiometric scintillators, which use the Ce3+ion as the light emitting compounds, (and this includes another prototype material species. They fall in the category of concentrated or CeP,O,,, see ref. 8) have shown no indication of any stoichiometric compounds, in which Ce is a chemical transfer of primary electron-hole excitations to the active constituent, (e.g. CeF3) and diluted or doped compounds, ion which, in turn, limits their eficiency. On the other such as LSO and GSO, the later being an analog of LSO hand the doped materials, in which the active Ce ion is but incorporating Gd instead of Lu, [ 5 ] . There is no present in much smaller concentrations, must rely on question that, given the present state of knowledge, the efficient transfer in order to reach their reported light doped compounds hold the lead as far as the light output output. This subject is of extreme importance for the
0018-9499/93$03.00 0 1993 IEEE
future development of scintillators and is currently being investigated. 2. MATERIAL
The pure lanthanide orthophosphates are structurally divided into two classes: the first half of the series (LaPO, to GdF'O,), has the monoclinic monazite structure, while the second half, (TbPO, to LuPO, plus YPO, and ScPO,), is characterized by the tetragonal zircon structure, . One of the advantages of the orthophosphates is that there is only one lanthanide ion site. The Cedoped LuPO, crystals investigated here were prepared by first reacting the Lu, and Ce oxides with lead hydrogen phosphate at 1360 "C,[ 101. Excess PbHPO, decomposed to form a Pb2P,0, flux, and the crystal-growth process was carried out by cooling the melt, which was contained in a platinum crucible, at a linear rate of approx. 1.0 "C per hour. The resulting flux grown crystals formed as rectangular parallelepipeds up to 4.0x0.5xO.2 cm3. While specimens of this size are adequate for research purposes, commercial detector applications will clearly require larger crystals. Two methods namely top-seeded high-temperature solution growth and hydrothermal are presently under investigation. LuPO, is an extremely durable material that, at normal temperatures, is not attacked by water, water vapor, organic solvents, or most of the common concentrated acids (e.g. boiling nitric or hydrochloric acids). The material has a melting point in excess of 2000 "C . Additionally it is extremely resistant to radiation
Fig. 1 Luminescencespectra of LOP:O.898%Ce Optical excitation at 250 nm.
Time, ns Fig. 2 Scintillationdecay of L0P:O. 178%Ce damage. It has a Mohr hardness of 5.5. and is readily polished using conventional techniques. Although Lutetium is generally considered to be an expensive material, its price is related to the cost of separation from other lanthanides and lack of demand. It should be noted that it is more abundant than Hg, I or Cd. The nominal Ce concentrations ranged from a fraction of a percent to 40 percent in the growth charge. However, mass spectrometric analysis showed that over the whole range, the actual Ce concentrationis 2.22 % of nominal. In what follows we shall use the actual concentrations. On account of the large difference in ionic radii of Ce and Lu, it is not presently known if significantly larger Ce concentrations can be achieved.
The details of the measurement setup have been described previously, .The luminescence of Ce:LOP represents a textbook case of what one would expect of the Ce3+ ion. Fig.1 gives the emission spectrum of a sample doped with 0.898% Ce at three different temperatures. This emission results from the transition between the lowest level of the 5d excited state and the 2F,,2 and 2F,n spin-orbit-split ground state of the 4f configuration. The line widths are typically broadened by the large electron-lattice coupling of the 5d states. In more dilute samples, (0.05%Ce), the splitting of the two components is approximately 2200 cm-l, in good agreement with Ref.1. Fig. 2 gives the decay time of the 0.178 % Ce sample excited by the ionizing radiation of a Ru/Rh source. It has been fitted with a single exponential at 23.1 ns with the background subtracted. The decay is strongly
using a BGO crystal as a standard. This result is shown in Fig.3. The photopeak for BGO occurs at channel 410 and for the 0.178% Ce :LOP at channel 890, indicating a more than twofold increase of the light output. Fig.4 shows the trend of light output as a 5000 function of concentration. It is quite evident that the light output saturates for concentrations around 0.1 mol 96. This contrasts with the behavior of stoichiometric scintillators, ( CeP50,, - see Ref. 8) , in which the L relationship is linear and signifies absence of lattice-toactivator energy transfer. The solid lines were calculated 0 using a simple statistical model (details to be published). Lines (a), (b), and (c) correspond to volumes controlled by cerium ions of about 500,1500, and 3000 units cells Some preliminary data were obtained on Ce:YPO,, (YOP). These crystals were not analyzed for true Ce concentration and we therefore can only give nominal concentrations. The likelihood is that the actual 0 500 1000 concentration is quite a bit lower, just as in the case of Channel number LOP. In the case of YOP we have observed a very important difference, to be discussed more fully in the Fig. 3 Energy spectra under 207Biexcitation. Amplifier next section. While the emission and excitation spectra shaping time 0.5 ps. are not significantly different from those of LOP, we observe a radically different mode of decay (Fig. 5),which temperature dependent, increasing to some 36 ns at 550 is no longer monotonic but shows a definite rise time (of K. This temperature dependence is characteristic of some 9 ns) followed by a decay of about the same materials showing self absorption, as has been magnitude as in LOP. The presence of the rise time documented by us in the case of Ce pentaphosphates, 181. signifies a slowdown of the lattice-to-activator energy Now we come to the all important issue of the transfer, which is also observed in GSO crystals, [ 5 ] .The light output of Ce:LOP. Using the 0.51 and 1.05 MeV inferior density of YOP, (4.31 g/cm3), makes it less radiation of 207Bi,we obtain a series of energy spectra attractive than LOP. Furthermore the light output is significantly lower, being at best equal to BGO. 10000
........ ...... -. ....
0 0.10 Ce concentration, %
Fig. 4 Light output of LOP as a function of concentration.
100 Time, ns
Fig. 5 YOP scintillationdecay curve. Nominal Ce concentration 4.3 mol%.
The results presented in this paper constitute another example that a host containing cerium as a dopant, but with lutetium, lanthanum, gadolinium or yttrium as the stoichiometric constituent, leads to remarkable scintillator performance . Among these four ions, lutetium is clearly superior and that is why we refer to this phenomenon as the "promise of lutetium". Because of the closed shell configuration, three of these four (trivalent) ions, Lu, Y and La are optically inactive. Gd is somewhat active because, having a half filled f-shell, it shows some weak f-ftransitions in the W. The common characteristic is, therefore, an inability to compete for holes and electrons and efficiently scintillate on their own. As a consequence, all are ideal constituents of the scintillator material, since they do not compete for holes with anions and Ce ions. Additionally their high atomic number, (especially of Lu), contribute to the stopping power. This observation is not entirely sufficient to explain the differences between them and particularly the "promise of lutetium". A deeper understanding is clearly needed. In general, the light output of a scintillator is given by the product of three quantities, CxSxQ, where C refers to the conversion of primary (gamma) energy to electron-hole pairs [ l l ] , S to the lattice-to-activator energy transfer, and Q the quantum efficiency of the activator. While enough data exist to determine C and Q in these materials, the key factor is the quantity S, or the ability of Lu-containing lattices to more efficiently and rapidly transfer energy to the Ce ion, [121. A comparison of Figs. 2 and 5 indicates that while transfer in LOP is much faster than the decay, (in fact unobservable), it is considerably slower in YOP. Details of this energy transfer and reasons for this difference are presently under investigation.
The authors are grateful to D.Oblas of the GTE Laboratories to have provided the glow discharge mass spectroscopic analysis of the samples. The support of the US. Department of Energy under Grant DE-FG02-90ER61033 and contract no. DE-AC05-840R21400 with Martin Marietta Energy Systems, Inc. is acknowledged. One of the authors (M.B.) is supported by the Kosciuszko Foundation.
REFERENCES 1. G.M.Williams, N .Edelstein, L.A.Boatner and M.M.Abraham, Anomalously small 4 f ~ 5 doscillator strengths and 4 f 4 f electronic Raman scattering cross sections for Ce3+in crystals of LuPO,, Phys. Rev.B 40,4143,(1989). P.C.Becker, J.G.Conway, 2. G.M.Williams, N .Edelstein, L.A.Boatner and M.M.Abraham, Intensities of electronic Raman scattering between crystal - field levels of Ce3+ in LuP0,:Nonresonant and near-resonant excitation; Phys.Rev.B 40,4132,(1989). 3. S.E.Derenzo, W.W.Moses, J.L.Cahoon, T.A. .DeVol, and L.A.Boatner, X-ray fluorescence measurements of 412 inorganic compounds; IEEE Nuclear Science Symposium, Conference Record, p 143, Santa Fe, NM 1991. 4. C.L.Melcher and J.S.Schweitzer, Ceriumdoped Lutetium Oxyorthosilicate: A fast, efficient new scintillator; IEEE Trans.Nucl. Sci. NS-39, pp 502505 (1992). J. S.Schweitzer, T.Utsu and 5. C.L.Melcher, S.Akiyama, Scintillation properties of GSO; IEEE Trans.NucZ.Sci.,37,161,( 1990). 6. A.J Wojtowicz, E.Berman and A.Lempicki, Cerium compounds as scintillators;Conference Record, IEEE Nuclear Science Symposium, Santa Fe, p 153, (1991). and A.Lempicki, 7. A.J.Wojtowicz, E.Berman Stoichiometric cerium compounds as scintillators Part I: CeF,, IEEE Trans. NucLSci., 39,494, (1992) A.Lempicki, 8. A.J.Wojtowicz, E.Berman and Stoichiometric Cerium compounds as scintillatorsPart 11: CeP5OI4,IEEE Trans. Nucl.Sci. NS-39, pp 1542-1548 (1992). 9 W.O.Milligan, D.F.Mullica, G.W.Beal1 and L,A.Boatner, Structural investigations of P O , , ScPO, and LuPO,; Inorg. Chim.Acta, 60,39,(1982). 10. T.Hayurst, G.Shalimoff , N.Edelstein, L.A.Boatner and M.M.Abraham, Optical spectra and Zeeman effect for Er3+in LuPO, and HfSiO,; J. Chem. Phys., 74,5449,(1981). 11. D.J.Robbins, On predicting the maximum efficiency of phosphor systems excited by ionizing radiation; J.Electrochem.Soc., 127,2694,(1980) 12. A.Lempicki, A.J.Wojtowicz, E.Berman, Fundamental limits of scintillator performance, accepted for publication in Nucl. Instr. and Meth. A