In principle, ion beam techniques are capable of detecting most of the elements in the periodic table [IV]. Even though the initial purpose of the analyses carried out in this work was to quantify all impurities in SrS:Ce based thin films, it was neither possible nor necessary to list every element in precision. A manageable solution was to isolate the problem from the complicated reality, therefore, to investigate only those impurities that are particularly interested in [III].
Impurity contents were compared for ALE SrS:Ce and reactively evaporated SrS:Ce,Mn,Cl thin films. In general, all samples studied were of high purity. The most noticeable impurities found were H, C, and O, and occasionally, weak signals corresponding to Ti, Cu, Mn, and Zn were detected. Figure 4 illustrates a case of Zn contamination in a reactive evaporated SrS:Ce/Si sample that was deposited after evaporating ZnS in the same chamber.
Fig. 4. PIXE spectrum of a SrS:Ce,Mn,Cl/Si sample prepared after deposition of ZnS in the same equipment. Signals from Zn are clearly visible.
The concentrations of H, C, and O impurities in various SrS:Ce based films made by ALE and reactive evaporation are summarized in Table II. The EL performances of TFEL devices made from the same batches as those used for analysis are listed for comparison.
Table II. Ion beam analysis of H, C, and O concentrations (at.%) in various ALE SrS:Ce and reactively evaporated SrS:Ce,Mn,Cl samples, and a comparison of their EL performances.
In ALE SrS:Ce covered in situ with Al2O3, the H and C concentrations are of the same magnitude as the concentration of Ce dopant (0.1-0.4 at.%) while the O concentration is slightly higher (0.2-0.9 at.%). It is known that the top Al2O3 layer introduces non-negligible interference signal for the oxygen measurement by TOF-ERDA. The oxygen content was significantly reduced (<0.1 at.%) when Al2O3 was replaced with ZnS as top layer. Comparison of the Ce doped and undoped ZnS/SrS films shows that each doped Ce atom brings with it less than one C and H atom. The ALE deposition process using Sr(thd)2, Ce(thd)4, and H2S as precursors therefore produces SrS and SrS:Ce films with rather high purity.
Study on a set of ALE SrS:Ce based samples having a large variety of EL performances showed that higher concentrations of H, C, and O impurity reduce the EL luminance. Furthermore, a carbon profile across the substrate was observed in several ALE SrS:Ce samples, and examination of the corresponding EL devices showed that EL brightness increases with the decrease in carbon content along the flow direction of the reactants. Most likely, the thd ligands of the precursors undergo slight decomposition at the deposition temperature employed.
Another important finding was the relationship between sodium content and the EL performance. Sodium is used as codopant in SrS:Ce powder because of the charge mismatch of trivalent Ce with the host.(17) Codoping improves the Ce incorporation and accordingly also the PL properties. In the case of ALE SrS:Ce, however, codoping with Na has not made significant improvement in EL.(18) On the other hand, an early study showed that the Na contamination in ALE SrS:Ce thin films may occur when soda lime glass is used as substrate.(19) Analysis of Na content was performed on a set of ALE SrS:Ce and SrS:Ce,Na films with different EL luminance. In Fig. 5, Na concentrations in SrS bulk and at the interface between the SrS and upper insulator were plotted against the EL luminances of the corresponding EL devices. As can be seen, Na contents were about the same in the SrS:Ce bulk (0.1-0.3 at.%) and in the SrS:Ce,Na bulk (0.2 - 0.4 at.%). These values are close to the detection limit of TOF-ERDA (0.1 at.%) and they are quite comparable to each other even though the EL performances of the samples vary a great deal. On the other hand, high Na content at the interface in these samples clearly indicates a weakening of the EL performance of SrS:Ce and SrS:Ce,Na thin films.
Fig. 5. Na concentration of ALE SrS:Ce and SrS:Ce,Na thin films vs the EL luminance of the corresponding EL device.
Compared with ALE SrS:Ce, higher contents of H, C, and O impurities were found in all reactively evaporated SrS:Ce,Mn,Cl films, suggesting that the H, C, and O impurities originate from the deposition process itself. (Table II) Possible contamination from air and moisture can not be completely ruled out, however, since Al2O3 layer is deposited ex situ after the SrS film has been exposed to air. Also the in situ deposited ZnS film is known to be highly polycrystalline and probably does not cat as a good passivation layer. Nevertheless, the reactively evaporated SrS:Ce,Mn,Cl samples did show better EL performances than the ALE SrS:Ce samples despite their higher impurity contents.
In relating the impurity contents to the EL properties of SrS:Ce thin films, however, one should be aware of the capability of and the information provided by the ion beam analysis techniques. Besides the stoichiometry of the phosphor bulk and its impurity contents, factors such as codopants, defects, crystallinity, and film morphology should be taken into consideration.
One limitation of the ion beam techniques is that they are unable to provide information on the location of the impurities in the polycrystalline films. The fact that reactively evaporated SrS:Ce,Mn,Cl gave better EL performances than ALE SrS:Ce samples, despite higher H, C, and O contents, hinted that impurities may influence the EL performance differently at different locations, for example in the crystal structure or at grain boundaries. There is no report on probing the local impurities in an EL phosphor. Apparently, this issue is also important for understanding the origin of space charges, which are crucial for EL performance. Besides the various vacancies, space charges have also often been associated with impurities, in both phosphor bulk and at phosphor/insulator interfaces. In this respect, alternative techniques such as transmission electron microscopy (TEM) combined with energy dispersive spectroscopy (EDS) may be considered. A TEM-EDS study on the SrS:Ce,Mn,Cl sample did show a higher concentration of Cl at the grain boundary than inside the grain.(20) The drawback of TEM-EDS is that the elements of low atomic number (Z < 11) cannot be detected.
Though combined ion beam techniques are capable of detecting the impurities at ppm levels, the depth profiling of low concentrations of impurities such as H, C, and O is difficult due to the low signal to noise ratio and strong signal disturbance from the top Al2O3 layer and surface contamination. A more consummate analyses may certainly be achieved by incorporate other analytical tools with ion beam techniques.(21) Nevertheless, routine ion beam analysis of particularly interesting impurities in the SrS:Ce thin films provided accurate enough information for the purpose of this study.
17. C. Fouassier and A. Garcia, in Inorganic and Organic Electroluminescence, EL 96 Berlin, (ed. R.H. Mauch and H.-E. Gumlich), Wissenschaft und Technik, Berlin, 1996, pp. 313.
18. P. Soininen, E. Nykänen, L. Niinistö, and M. Leskelä, in Inorganic and Organic Electroluminescence, EL 96 Berlin, (ed. R.H. Mauch and H.-E. Gumlich), Wissenschaft und Technik, Berlin, 1996, pp. 149.
19. H. Antson, M. Grasserbauer, M. Hiltunen, T. Koskinen, M. Leskelä, L. Niinistö, G. Stingeder, and M. Tammenmaa, Fresenius J. Anal. Chem. 322 (1985) 175.
20. Unpublished result.
21. L. Niinistö, Ann. Chim. 87 (1997) 221.