The luminescence of SrS:Cu is known to originate from the isolated Cu+ ion that substitutes for a Sr atom but locates at an off-center of the octahedral coordination in the host matrix.(37) - (39) The understanding of the SrS:Cu phosphor has been, however, rather insufficient to interpret many of the luminescence properties. For example, one or more broad bands vary from blue to green in SrS:Cu have been reported in the literature and the explanations were often contradictory.(40)- (43) Furthermore, the luminescence spectra and decay time of SrS:Cu are strongly temperature dependent but the reasons for this have not been given. On the other hand, most of the alkali halides have the same rock salt crystal structure as SrS, and the luminescence of the octahedrally coordinated Cu+ in these hosts are well understood,(44) -(46) they may thus well be used in understanding the luminescence properties of SrS:Cu. The studies carried out in this work were intended to clarify these issues, and possible luminescence mechanisms for SrS:Cu are proposed.
Fig. 8. PL emission spectra of ALE SrS:Cu thin films at RT.
PL studies on ALE SrS:Cu thin films at RT verified that both blue and green luminescence may exist. The emissions of an ALE SrS:Cu consist of a high energy (H) band and a low energy (L) band located at about 460 and 520 nm, respectively (Fig. 8). The ratio of the H and L bands thus determines the color gamut of the phosphor. However, the H band diminishes as the temperature decreases, as if the emission undergoes a red shift. Completely green luminescent SrS:Cu can be seen as the result of a "single" emission band located at about 530 nm, and this band persists at low temperature. EL of the SrS:Cu films is essentially the same as PL.
The L band is assigned to the isolated Cu+ ions at an off-center position in the octahedral coordination. Such an assignment is based on the excitation spectrum of SrS:Cu at low temperature, which can be studied only at L band since the H band has vanished. As can be seen from Fig 9, the 1A1g - 3Eg peak intensity is much higher than that of 1A1g - 1T2g. This is a typical off-center characteristic that is well-understood in Cu doped alkali halides.(45) In the case of NaF:Cu+ where Cu+ is on-center in the octahedral site, the 1A1g - 1T2g peak intensity is significantly enhanced.(46) It needs to be noted that the L band may also contain emission from aggregated Cu centers. Decay studies at the L band at 80 K reveal at least two components. One with the value of 80 - 110 ms is from the isolated Cu+ center, while the other 12 - 18 ms s is due to the presence of paired or aggregated Cu+ centers.
Fig. 9. Excitation spectrum of a SrS:Cu at 80 K.
The H band, which is in large part responsible for the blue emission at high temperature, is due to Cu+ at a site that experiences different crystal field than the center responsible for the L band. A possible site can be any one of the surrounding eight tetrahedral holes as shown in Fig. 10. The tetrahedral holes have a radius of rT = 0.77 Å, which makes them readily available to Cu+ ion (rCu+ = 0.77 Å). Also the three nearest S atoms that are shared by the octahedral and the neighboring tetrahedral holes form a relatively open channel (rc = 0.63 Å), allowing the Cu+ hop between the holes at high temperature. Furthermore, the weaker ligand-field splitting of the tetrahedral symmetry may cause the lowest excitation state to shift to higher energy as compared with that in an octahedral arrangement.
Fig. 10. Unit cell of the SrS structure. A neighboring tetrahedral hole is highlighted.
The large tetrahedral holes may enhance the possibility of forming paired and aggregated Cu centers which may then effectively block the formation of the H band. As a result, "single" green emission is possible. A recent paper reported that alkali metal codoping in MBE SrS:Cu,F results in a red shift of the emission from blue to green.(47) The authors attributed this color shift to reduced Cu coordination number due to the incorporation of alkali metal ions and formation of Cu-S vacancy complexes. The alkali ions together with Cu+ emission centers were assumed to be acceptors, while S vacancies acted as double donors. However, another likely interpretation is offered here that these alkali ions occupy the tetrahedral holes preventing Cu+ ions from entering so that results in only green emission.
The temperature dependence of the decay time constant in SrS:Cu leads to a further proposal that, regardless of the two emission bands, the luminescence is likely to follow the three-level mechanism commonly at work in Cu+ and Ag+ doped alkali halides.(44), (48) Figure 11a illustrates the energy level diagram for this luminescence mechanism. As described in paper [VII], the 3Eg energy level splits into two sublevels where emissions occur from T1g at low temperature but from T2g at high temperature. Because of the selection rule, the decay is much longer at T1g (highly populated at low temperature) than at T2g (highly populated at high temperature).
It is worth mentioning that the energies of the two sublevels are very close to each other (~200 cm-1) and cannot correspond to the large energy difference (~2500 cm-1) between the H and L bands in the emission spectra. [VII] In the case of NaF:Cu, a Jahn-Teller effect has been observed as a double-humped band,(46) but such effect is unlikely for SrS:Cu.(39)
Fig. 11. An illustration of (a) the three-level mechanism, and (b) the fitted decay values as function of temperature.
Some unresolved issues need to be addressed here. As can be seen from Fig. 11b, there are two fitted decay values in a SrS:Cu sample as a function of temperature. Presumably, they are the decay time constants of the isolated (slow component) and paired (fast component) centers. The assumption is nevertheless questionable for the temperature range above 200 K, since the SrS:Cu samples there exhibit extremely long afterglow and the fitted decay values may no longer correspond to the two types of emission centers. Although the decay time constants at temperatures below 80 K were not recorded, it is known that the decay time at 4 K does not differ significantly from that at 80 K.(38) The dotted line shown in Fig. 11b may therefore be a more realistic decay constant vs. temperature curve, which resembles the three-level luminescence.
The assumption that Cu+ ions locate at tetrahedral sites at high temperature raises another question about the influence of the crystal field on the energy levels of Cu+ centers and hence the probabilities for electron transition between the excited states and the ground state. A theoretical treatment is desirable to clarify this issue.
The luminescence of SrS:Ag,Cu,Ga films is clearly different from that of SrS:Cu films. As shown in Fig. 12, the EL emission has two broad bands at 80 K, the one located at about 520 nm corresponding to the Cu+ emission and the new band peaked at 430 nm. These bands coalesce into a single broad band with maximum at 460 nm when samples are at RT. Studies have indicated that the 430 nm peak observed at 80 K is most likely due to the emission of Ag pairs.(49)
Fig. 12. EL spectra of 1) a SrS:Ag,Cu,Ga, 2) a blue ALE SrS:Cu, and 3) a green ALE SrS:Cu thin films at 80 K.
Two types of luminescence mechanism have been proposed for the SrS:Ag,Cu, Ga phosphor. One suggests that energy transfer takes place from Cu to Ag and results in increased Ag emission.(50) The other, proposed in this work [VII], suggests that the luminescence is a result of the recombination of the excitation energy from Cu+ and Ag+ centers at ionized Ag+ - (Ag+) and Cu+- (Ag+) pair centers.
It is probably reasonable to regard these two proposals as pertaining to the same mechanism as is done by Jones et al.(51) Unfortunately, there are no direct experiments to confirm that this is so. On the other hand, decay studies (Fig. 13) show that the 520 nm emission at 80 K has a very similar decay value to that of SrS:Cu (about 15 and 95 ms), while much faster decay is recorded for the Ag pair emission at 430 nm (9 ms). This indicates that the luminescence events from Ag and Cu activators occur independently, at least at low temperature. More studies on these issues are certainly needed.
Fig. 13. Decays (at 80 K) of a SrS:Ag,Cu,Ga film studied at 430 and 520 nm emissions with excitations at 265, 295, and 320 nm. Two groups of decay originating from the Ag and Cu centers can be seen regardless of the excitation energy.
37. N. Yamashita, Jpn. J. Appl. Phys. 30 (1991) 3335.
38. T.C. Jones, W. Park, E. Mohammed, B.K. Wagner, C.J. Summers, and S.-S. Sun, in Flat Panel Display Materials, (ed. G. Parson, T.S. Fahlen, S. Morozumi, C. Seager, and C.C. Tsai) Materials Research Society, Vol. 508, Pittsburgh, 1998, pp. 281.
39. W. Lehmann, J. Electrochem. Soc. 117 (1970) 1389.
40. J. Kane, W.E. Harty, M. Ling, and P.N. Yocom, Conf. Rec. 1985 Inter. Display Res. Conf., pp. 163. 41. B. Moine and C. Pedrini, Phys. Rev. B 30 (1984) 992.
42. D.S. McClure and S.C. Weaver, J. Phys. Chem. Solids 52 (1991) 81.
43. P.D. Keir, J.F. Wager, B.L. Clark, D. Li, and D.A. Keszler, Appl. Phys. Lett. 75 (1999) 1398.
44. C. Pedrini, Solid State Comm. 38 (1981) 1237.
45. U. Troppenz, B. Hüttl, U. Storz, P. Kratzert, and K.-O. Veltaus, 4th Inter. Conf. Sci.
46. W. Park, T.C. Jones, E. Mohammed, C.J. Summers, and S.S. Sun, 4th Inter. Conf. Sci. Tech. Display Phosphors, Extended Abstracts, Bend, OR, 1998, pp. 215.
47. T.C. Jones, W. Park, and C.J. Summers, Appl. Phys. Lett. 75 (1999) 2398. 48. S.J. Yun, Y.S. Kim, J.S. Kang, S.H.K. Park, K.I. Cho, and D.S. Ma, SID 1999 Digest 30, pp. 1142.
49 U. Troppenz, B. Hüttl, U. Storz, P. Kratzert, and K.-O. Veltaus, 4th Inter. Conf. Sci. Tech. Display Phosphors, Extended Abstracts, Bend, OR, 1998, pp. 187.
50 W. Park, T.C. Jones, E. Mohammed, C.J. Summers, and S.S. Sun, 4th Inter. Conf. Sci. Tech. Display Phosphors, Extended Abstracts, Bend, OR, 1998, pp. 215.
51 T.C. Jones, W. Park, and C.J. Summers, Appl. Phys. Lett. 75 (1999) 2398.