ABSTRACT
Introduction Ten years ago, the first Ceramic Metal Halide Discharge Lamps were introduced into the market. It was the combination of advanced metal halide fill systems with ceramic arc tubes known from HPS lamp technology that lead to the high performance of today's HID lamps. The main problem which had to be solved was the development of a metal halide resistant lamp seal. First attempts to seal metal halide ceramic lamps with a seal known from HPS lamps failed because of strong chemical attack of the Iodides and free Iodine on the seal components. The tubular lead through, consisting of niobium, is corroded according to the chemical reaction (1):
Nb + 2.5 J2 NbJ5 (l)
According to (l) the lead through is penetrated within several hours and the lamp slowly leaks. Therefore, lead through systems based on those used for HPS lamps are not suitable for MH ceramic lamps. Additionally, the frit material used for HPS lamps cannot be used for MH ceramic lamps. The frit contains earth alkaline oxides (CaO, SrO, BaO) besides other oxides. Within the first 100 hours of operation strong chemical reactions of these oxides with the metal iodides of the fill are observed. The rare earth iodides (i.e., DyJ3) in the fill react strongly according to (2)
3 CaO + 2 DyJ3 Dy203 + 3 CaJ2 (2)
Reaction (2) suggests that it is useful to eliminate earth alkaline oxides (CaO, SrO, BaO) and to add rare earth oxides such as Dy203. The presence of Dy203 shifts the equilibrium for chemical reactions given in (2) to the left side suppressing such undesired reactions. A metal halide compatible frit was found on basis A1203-Dy203-Si02. Si02 is added to lower the frit melting temperature and to form a glassy matrix. It is not suitable with metallic Na in HPS lamps but it is resistant in a halide atmosphere up to a critical temperature. Temperature of the frit sealed portion may not be raised above 750°C. This has recently been confirmed in a detailed examination by additional heating of the frit portion [1]. All attempts to replace niobium as lead through by a morc halide resistant material such as molybdenum or tungsten failed. Several approaches were made [2,3]:
figure 1: attempts to join Molybdenum as tube (left side) or rod (right side) to PCA The requirements of lamp operation such as alternating thermal stress during the quick run up and cool down are very severe for a lead through solution with only molybdenum or tungsten. A critical gap between the thermal expansion coefficient of PCA (8 x 104 ICI ) and either Molybdenum (5,3 x 10'6 ICI) or Tungsten (4,6 x 10-6 IC-1) has to be bridged. Up to now no joining technique is found which allows reproducible sealing results under production conditions.
