Ceramic envelope, lamp made said envelope, as well as method of manufacture of metal ceramic plug

HUP9801470A3Inactive Publication Date: 2001-02-28OSRAM SYLVANIA INC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
HU1998001470
Authority / Receiving Office
HU · HU
Patent Type
Applications
Current Assignee / Owner
Priority Date
1997-06-27
Filing Date
1998-06-26
Publication Date
2001-02-28
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Existing metal halide high-intensity discharge lamps face limitations in longevity due to the reaction between metal halides and conventional frits, and the mismatch in thermal expansion coefficients between the discharge tube materials and metal-ceramic plugs, leading to cracks and leaks.

Method used

A multi-part metal-ceramic plug with axially overlapping layers of varying thermal expansion coefficients, sintered directly to the discharge tube and feedthrough, minimizing thermal stress by matching the expansion rates of adjacent components and eliminating the need for frits.

Benefits of technology

The solution provides a gas-tight seal that lasts for a long time, maintaining luminous flux and extending the lamp's life by minimizing thermal stresses and cracks, while allowing easy pumping and filling without the use of glass frits.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to a ceramic bulb, a lamp made of the same, and a method for producing a metal-ceramic plug. The bulb comprises a transparent ceramic discharge tube (5) having a first end portion (6a) and a second end portion (6b), the discharge tube (5) defining a discharge space and having a longitudinal axis of symmetry; a non-conductive, metal-ceramic first end plug (8a) with which the first end portion (6a) of the discharge tube (5) is closed; a non-conductive, metal-ceramic second end plug (8b) with which the second end portion (6b) of the discharge tube (5) is closed; a first and a second metal feedthrough (7a, 7b) passing through the first and second end plugs (8a, 8b), respectively; an electrode (9) at the inner end of the first and second feedthroughs (7a, 7b), respectively. According to the invention, at least the second end plug (8b) has a multi-part structure of at least three parts, each lead (7a, 7b) having an inner end and an outer end;the leads (7a, 7b) are made of one of the metals or alloys in the group consisting of tungsten, molybdenum and rhenium, and alloys of at least two of these metals; the thermal expansion coefficient of at least one part of the multi-part end plug (8b) is between the thermal expansion coefficient of the discharge tube (5) and the thermal expansion coefficient of the lead (7b); the multi-part structure comprises at least four axially overlapping parts with different thermal expansion coefficients, including a first and a last part; the first part is located innermost with respect to the discharge space, and the last part is located outermost with respect to the discharge space; the multi-part end plug (8b) is sintered directly to the discharge tube (5) and the feedthrough (7b) in such a way that the innermost part of the end plug (8b) is sintered directly to the discharge tube (5) and the outermost part of the end plug (8b) is sintered directly to the corresponding feedthrough (7b). à;
Need to check novelty before this filing date? Find Prior Art

Description

The description is 26 pages long (including 12 pages of illustrations). HU 221 366 Β1 comprises at least four axially overlapping parts with different thermal expansion coefficients, including a first and a last part; the first part is the innermost part with respect to the discharge space, the last part is the outermost part with respect to the discharge space; the multi-part end plug (8b) is sintered directly to the discharge tube (5) and the lead-through (7b) in such a way that the innermost part of the end plug (8b) is sintered directly to the discharge tube (5) and the outermost part of the end plug (8b) is sintered directly to the corresponding lead-through (7b). The invention relates to a ceramic envelope, a lamp made therefrom, and more preferably to a metal halide lamp with a polycrystalline alumina envelope, the ends of which are closed by ceramic-like plugs. The invention relates in particular to an envelope with at least one metal-ceramic plug in which there are parts or zones or layers with a gradually varying coefficient of thermal expansion. The invention also relates to a metal-ceramic plug and a method for producing the same. Metal halide high-intensity discharge lamps are desirable to operate at high wall temperatures to improve efficiency, change color temperature, and / or increase the color rendering index of the light source. Metal halide lamps are typically filled with a fill containing one or more metals, such as sodium halides (primarily its iodides and bromides). Sodium is often combined with scandium. Additional dopants include thorium, thallium, indium, and lithium. Other types of fills contain rare earth metals, such as thulium, holmium, and dysprosium. Lamps filled with such fills have highly desirable spectral properties: luminous efficacy in excess of 100 lm / W, color temperatures around 3700 K, and color rendering indices around 85. The low vapor pressure of some metal halide dopants requires the fused quartz lamp envelope to be operated at higher than normal temperatures.At wall temperatures above 900-1000 °C, the lifetime of lamps is limited by the interaction between the metal halides and the quartz glass wall. The use of discharge tube materials that can operate at higher temperatures than quartz glass and are more chemically resistant than quartz glass is an effective way to extend the lifetime of lamps containing these metal halides. Polycrystalline alumina forms a sodium-resistant envelope for high-pressure sodium lamps. Polycrystalline alumina can operate at higher temperatures than quartz glass and is believed to be more chemically resistant than quartz glass. The ends of the polycrystalline alumina vessel are sealed with alumina plugs. The gas-tight seal is provided by a sealing glass, often called a fusible ceramic or frit. However, chemical investigation of the metal halide in polycrystalline alumina envelopes has shown that the reaction between the metal halides and conventional frits, even supposedly “halide-resistant” frits, significantly limits the service life. Examples of such frits include frits based on CaO, Al2O3, BaO, MgO and B2O3. Consequently, a fritless sealing process is highly desirable. In polycrystalline alumina lamps, niobium transitions are usually used because of their similar coefficient of thermal expansion. One problem, especially when the fillings contain rare earth halides, is the reaction between the niobium transitions and the filling. This problem has been somewhat alleviated by the use of special arrangements in which the plug and transition are simultaneously replaced by a plug made of conductive metal ceramics. These metal ceramics are composite, sintered bodies, usually containing alumina (the material of the discharge tube) and molybdenum or tungsten. Both are halide-resistant materials. Hing et al., U.S. Patent No. 4,354,964, describes a conductive alumina / metal (e.g., tungsten or molybdenum) metal ceramic containing 4 to 20% by volume of metal as a material for plugs or bushings in polycrystalline alumina envelopes of metal halide high intensity discharge lamps. The metal ceramic contains refractory metal rods (as electrodes or current conductors). These are embedded in the metal ceramic body in a green or pre-fired state and then fired together during the final sintering of the metal ceramic to a high density. The cited patent does not describe the process by which such metal ceramics are connected to polycrystalline alumina tubes. The difference in thermal expansion coefficient between the metal ceramic and the polycrystalline alumina, or between the metal ceramic and the tungsten or molybdenum electrode, cannot be eliminated simultaneously.This difference in thermal expansion coefficients can cause cracking and leakage in either the polycrystalline alumina tubes or the metal ceramic, or both, when the lamp is turned on and off. Izumiya et al., U.S. Patent No. 4,731,561, discloses a polycrystalline alumina tube having one end sealed with a sintered, conductive alumina / molybdenum or alumina / tungsten metal ceramic. The other end of the polycrystalline alumina tube is sealed with a frit-sealed metal ceramic. Each metal ceramic is coated with an insulating layer to prevent arc flashback. In addition to conductive cermet plugs, Kajihara et al. US Patent No. 4,687,969 describes non-conductive cermets with outward and inward passing and protruding leads. One end of the polycrystalline alumina tube is a sintered cermet and the other end is a frit-sealed cermet. However, cracking in the cermet cannot be prevented because the composition of the plug is constant and not directional. All of these one-piece plugs have the disadvantage that their coefficient of thermal expansion does not match that of the surrounding component (e.g., the discharge vessel). One solution is proposed, for example, by Partlow et al. in US 4,602,956. HU 221 366 BI patent description. This patent describes a cermet plug having a core and one or more layers of cermet of a different composition surrounding the core and substantially coaxial therewith. The core consists essentially of 10-30 vol.% tungsten or molybdenum, the remainder being aluminum oxide. The layers consist essentially of 5-10 vol.% tungsten or molybdenum, the remainder being aluminum oxide. This cermet plug is hermetically sealed to the end wall of the discharge tube by using “halide-resistant” frits. However, conductive cermet plugs are not sufficiently gas-tight over a long period of time due to their fine structure. Another solution is a denser, non-conductive cermet plug. Consequently, a separate metal lead-through is required. Bünk et al. US 5,404,07 describes a high-pressure discharge lamp with a ceramic discharge vessel, the ends of which are closed by non-conductive cermet plugs, which consist of, for example, alumina and tungsten or molybdenum. In one specific embodiment (Fig. 9), the cermet plug consists of concentric sections containing tungsten in different proportions. These sections implement a gradually varying coefficient of thermal expansion. Nagayama EP 650 184 describes a discharge tube whose end plugs consist of a non-conductive cermet. The cermet plug is made of axially overlapping layers of different compositions (axially stepped sealing, see Figs. 16 et seq.). The first layer of the plug is integrally connected to the open end of the discharge vessel. The metal lead-through is formed by a tungsten-based rod. The sealing between the lead-through and the last axially overlapping layer of the plug is achieved by a rather complicated method. In this case: - a threaded part of the lead-through is in direct contact with the last layer of the plug, - an external metal disc (“flange”) contacts the outer surface of the last layer, - a gasket, such as platinum or glass solder, covers the flange and the outer surface of the last layer. One of the rods, which acts as a conduit, has an axial opening through which the charge is introduced into the discharge vessel. Evans et al. US Patent No. 4,155,758, in connection with Figure 14, also discloses an axially stepped plug. However, this one is made of three layers of conductive cermets. The object of our invention is a ceramic lampshade for a high-pressure discharge lamp, especially for a metal halide lamp, with a very long-lasting gas-tight seal. A further object is a lamp provided with such a lampshade. A further object is a method for producing such a lampshade. This task is solved according to the invention in such a way that - briefly summarized - the characteristics of the ceramic bulb are as follows: - a transparent ceramic tube has a first tube end and a second tube end, the tube has a longitudinal axis, and the tube defines a discharge space, - comprises a first, substantially non-conductive cermet end plug, said first end plug closing the first tube end of the ceramic tube, - comprises a second, substantially non-conductive cermet end plug, said second end plug closing the second tube end of the ceramic tube, - at least the second plug has a multi-part structure and consists of at least three parts - a first and a second metal via pass through the first and second plugs, respectively; each via has an inner end and an outer end, and the vias are made of a metal or alloy from the group consisting of tungsten, molybdenum and rhenium and alloys of at least two of these metals, - an electrode is placed at each inner end of the first and second passage, - the coefficient of thermal expansion of at least one part of the multi-part plug is between the coefficient of thermal expansion of the discharge tube and the lead-through, - the structure of the multi-part plug comprises at least four axially overlapping parts with different coefficients of thermal expansion, including a first and a last part; the first part is the innermost part relative to the discharge space, the last part is the outermost part relative to the discharge space, - the multi-part plug is sintered directly to the discharge tube and the feedthrough in such a way that the innermost part of the plug is sintered directly to the discharge tube and the outermost part of the plug is sintered directly to the feedthrough. The difference in thermal expansion coefficients for all adjacent components (including the pipe and associated feedthrough) is preferably less than 1.0·106 / K. This minimizes thermal stresses and cracks. The second passage is usually a tube and contacts the multi-part structure. However, another embodiment of the second passage - preferably when using a separate filling hole - is a pin or rod. The first lead may be a rod combined with a one-piece plug (as is well known) or may be similar to the second lead. Accordingly, the first plug may be formed by a one-piece body or a multi-piece structure. The function of the features outlined above is as follows. The stepped cermet end plug comprises sections or zones or layers with slightly different coefficients of thermal expansion. The coefficients of thermal expansion decrease from the outermost part of the plug to the innermost part of the plug. The outermost part is defined as the part axially furthest from the discharge volume. The innermost part is defined as the part axially closest to the discharge volume. The innermost zone has an outer surface (as viewed in the radial direction) which is in contact with either the inner wall of the end of the alumina discharge tube or with a separate alumina insert. Its coefficient of thermal expansion is well matched to that of the alumina discharge tube or insert. On the other hand, the thermal expansion behavior of the outermost zone matches that of its passage. The inner surface of the outermost zone (as viewed in the radial direction) is in contact with the passage. The intermediate part of the plug is HU 221 366 Β1 zones serve as transition zones that gradually bridge the difference between the coefficient of thermal expansion of the innermost and outermost zone or part. Advantageously, not all intermediate parts are in contact with the feedthrough. This can be done in two different ways. According to the first, the inner diameter of the intermediate parts is larger than that of the outermost parts. According to a more elegant and easier to manufacture solution, all parts have the same inner (and even the same outer) diameter. However, the feedthrough only enters some (up to three) outer parts. It does not have to enter the inner parts, which are not thermally adapted. Another important feature is that the total length of the multi-part structure is as small as possible (preferably less than 5 mm), because only then can a homogeneous and uniform density of the structure be achieved. The different characteristics of the different zones can be achieved by mixing different amounts of metal powder (preferably tungsten or molybdenum) with the alumina powder at the beginning of the metal ceramic preparation. Surprisingly, the most promising is the combination of a tungsten-containing plug with a molybdenum via. There are several options for producing plugs with different coefficients of thermal expansion. One way is that the composition of the different parts comprises alumina as a first component and a metal, preferably tungsten or molybdenum, as a second component. The composition of the parts differs from each other in the proportion of metal added to the alumina. Another way to achieve the above objective is to have different components in the composition of different parts, for example aluminum nitride and aluminum oxynitride. While aluminum nitride has a given coefficient of thermal expansion (see, for example, US Patent No. 5,075,587), the coefficient of thermal expansion of aluminum oxynitride depends on the ratio of its components, i.e., aluminum oxide to aluminum nitride. The situation is similar to that of cermets, which are composed of aluminum oxide and a metal (tungsten or molybdenum). In one preferred embodiment, the plug is designed as a cylindrical disc and consists of concentric parts with the same outer diameter (possibly except for the innermost part) and with axially stepped thermal expansion coefficients. Instead of the thermal properties of the parts of the plug changing stepwise, it is also possible to have a smooth axial variation of the thermal expansion coefficient of the plug. This embodiment can also be understood as having an infinite number of parts. In another preferred embodiment, the plug is formed by a layered structure, cylindrical in shape and provided with a central bore. The bore diameter may be constant or variable. Only the outermost layer, adjacent to the passage, is in gas-tight contact with this passage. The other layers are remote from the passage. The outer surface of the innermost layer, viewed in the radial direction, is in contact with the end of the discharge vessel. In this embodiment, to avoid capillary effects, it is advantageous if the distance between the layers of the feedthrough and the plug is at least 1 mm, except for the outermost layer which is in contact with the feedthrough. This distance may be the same for all layers. The distance between the innermost layer of the plug and the lead-through is particularly important. This distance is preferably at least 3 mm. This allows the electrode to be inserted into this space. A preferred structure is a telescopic multi-part plug in which the distance between the parts or layers and the lead-through decreases stepwise from the innermost layer to the outermost layer. In a particularly preferred embodiment, the multi-part plug is formed by a layered, cylindrical structure having a constant inner and outer diameter. This structure consists of four or five zones. The passage is formed by a tube which enters the outermost part and possibly the adjacent intermediate parts, but does not enter the inner parts adjacent to the discharge. The innermost part contacts the discharge vessel or a ceramic insert member which is typically annular and has a composition similar to or identical to that of the discharge vessel. The multi-part structure is preferably recessed into the insert member. A typical value is 0.5 mm. The advantage of an axially stepped closure is that the temperature load on the closure is minimized and the gas tightness is optimized when only a small portion of the plug, preferably the innermost layer, is located at one end of the discharge tube. The innermost layer is optionally completely enclosed in the end of the discharge tube or only partially enclosed therein. The "sealing" length between the innermost layer(s) and the end of the discharge vessel is at least 0.8 mm. Typical values ​​are between 1 and 2 mm. Preferably, there is a similar sealing length between the outermost layer(s) and the lead-through. The metal ceramic of the invention consists of an alumina embedding material (matrix) in which metal (preferably molybdenum or tungsten) particles are embedded. These particles are at least approximately spherical in shape. It has been found that the different thermal behavior of the alumina embedding material and the metal particles is a critical characteristic. The average coefficient of thermal expansion of alumina-tungsten cermets is known as a function of the amount of tungsten, see for example P. Hing, “The Relationship between Physical Properties and Microstructures of Dense Sintered Cermet Materials”, pp. 135-142, Science of Ceramics, ed. KJ de Vries, Vol. 9, Nederlandse Keramische Verenigung (1977). Accordingly, the proportion of tungsten required for a given thermal expansion can be determined. It has been found that microscopic stresses are generated at the interface of tungsten particles in the alumina embedding material. These stresses decrease as the size of the minority component decreases. The minority component is often called the dispersoid or dispersed phase. In some regions, this minority component is alumina, in other regions it is a metal (tungsten). Therefore, we prefer a very fine particle size of tungsten powder, at least less than 50% by volume. HU 221 366 Β1 in the case of tungsten-containing alumina / tungsten metal ceramics. The typical value of the average particle size is 0.6-0.9 pm. In practice, tungsten precursors, such as water-soluble ammonium tungstate, can be used to produce very fine tungsten particles in an alumina embedding medium. Tungsten precursors are water-soluble, mixed with alumina powder, and calcined to convert them into fine tungsten particles. A similar method has been used to produce nanophase tungsten carbide / cobalt composite powder, see LE Me Candlish, Β. K. Kim, and Β. H. Kear, “Characterization and Properties of Chemically Processed Nanophase WCCo Composites,” pp. 227-237. In “High Performance Composites for the 1990s,” eds. S. Das, C. Ballard, and F. Marikar, TMS Warrendale, Pennsylvania, USA, 1991. Conversely, for alumina / tungsten cermets containing at least 50 vol.% tungsten, alumina (water-soluble) precursors such as aluminum nitrate can be used to produce very fine alumina particle sizes. The average particle size is typically 0.6-0.9 pm. It is important to select suitable starting materials for the production of metal ceramics to create: 1. uniform distribution of the dispersed phase, 2. fine particle size of the dispersed phase, 3. compatible raw density and firing shrinkage with adjacent layers to produce stepped metal ceramics free from cracks or distortion, 4. such a raw density and shrinkage behavior during firing that a direct bond is formed between the metal lead-through and the metal-ceramic plug, or between the metal-ceramic plug and the polycrystalline alumina discharge tube. Typical size ranges for such cermet plugs: - outer diameter 3.0-4.0 mm (with the condition that the first part should have a larger diameter), - the total length of the axially stepped plug is at most 10 mm, preferably less than 5 mm. The axial thickness of the innermost zone is preferably between 1.0 and 3.0 mm. The axial thickness of each intermediate zone and the outermost zone is preferably between 0.3 and 1.5 mm. The conduits are preferably tubular. The dimensions of these tubes are in the following typical ranges: - the outer diameter is between 0.9 and 1.6 mm, - the inner diameter is between 0.6 and 1.2 mm, - total length between 10 and 15 mm, - the wall thickness is at most 0.25, preferably about 0.1 mm. It is preferable if the outermost part or zone or layer contains more than 50% by volume of metal. This high metal content allows this part to be welded to the associated lead-through. This complements the direct sintering between the two bodies. The bond between the two bodies is thus improved by the additional welding. This is a safety measure in case a part of the direct sintering becomes leaky. In a particularly preferred embodiment with reduced temperature loading, the multi-part structure is positioned at a distance from the hot discharge space and an additional hollow cylindrical member (preferably an alumina capillary tube) is positioned between the end of the discharge vessel and the multi-part structure. This arrangement can reduce the operating temperature of the multi-part structure by approximately 200°C. The gas-tight connection between the hollow member (capillary tube) and the multi-part structure is preferably provided by a sleeve element surrounding the contact zone of the two members.In a preferred embodiment, the axially stepped plug concept allows a special filling method. For this purpose, a separate filling hole in the second multi-part plug for evacuating and filling the discharge vessel is used. In this embodiment, the diameter of the filling opening or filling hole is not limited by the diameter of the tubular passage. The hole is axially aligned with the geometric axis, but is arranged eccentrically. After filling, the hole is closed with a suitable rod, hereinafter referred to as a hole stopper. The discharge vessel is thus resistant to corrosion and temperature changes. Lamps with such plugs are gas-tight for a very long time and retain their luminous flux value excellently throughout their service life. This is explained by the fact that not only is the plug connected to the end of the discharge vessel and the passage without any glass frit or ceramic sealing material, but the hole stopper also closes the filling hole without these materials.This is possible through a very clever arrangement: the important feature is that the outermost cermet layer or part of the plug has a composition that allows it to be welded. To meet this requirement, the proportion of metal in the outermost layer must exceed 50% by volume. This layer can be conductive, but does not necessarily have to be. When manufacturing the discharge tube, the first part of the plug is connected to the discharge tube by co-firing as described previously. After the metal-ceramic plug has been fired together with the end of the discharge tube, the discharge space is pumped out, rinsed and filled through the filling opening. The hole stopper is then inserted into the filling opening and welded to the metal-ceramic plug on the outer surface of the outermost part. This creates a hermetic seal. The rod or hole stopper may be made of metal (preferably molybdenum or tungsten) or a cermet. Preferably, it is made of the same material as the outermost layer of the plug. Any conventional welding method can be used, such as resistance welding, laser welding, electron beam welding or AVI welding. The plugs under investigation are therefore very often strictly non-conductive. In one advantageous, special embodiment, a plug can be used in which the outermost layer has a high proportion of metal. This layer can optionally be made of a conductive metal ceramic. Up to HU 221 366 Β1 the closest layer (the last intermediate layer) is also conductive, unlike all other layers which are closer to the discharge space and are not conductive. This arrangement is referred to here as an “essentially non-conductive plug”. The main advantages of this invention, which represents a breakthrough in sealing technology, are as follows: - There is no frit at all in the sealing, and yet a well-proven and very reliable sealing method, sintering, can be used. - The filling opening should be large enough to allow for easy pumping and filling. - This type of closure works for all lamp powers and all discharge vessel sizes. It is worth mentioning that one of the preferred compositions of the discharge vessel is polycrystalline alumina doped with magnesia and possibly yttria or zirconia. This composition is also preferred for the hollow sleeve element mentioned above. In contrast, for the multi-part structure, the preferred composition of the aluminum powder is either pure alumina or magnesia-doped alumina. Pure alumina is preferred for the outer zones with a high tungsten content, and doped alumina for the inner zones with a low tungsten content. Our invention is described in more detail with reference to its exemplary embodiments with the help of our drawings, of which the Figure 1 is a highly schematic partial view-section of a lamp containing a ceramic discharge tube, the Figure 2 is a detailed view of the first end of the discharge tube, as a first embodiment of the invention, Figure 3 shows the expansion as a function of temperature for different metal ceramic parts, Figure 4 shows the expansion value at different temperatures for different proportions of tungsten in the metal ceramic part, Figure 5 is a detailed view of the second end of the discharge tube, as a second embodiment of the invention, Figure 6 is a detailed view of the second end of the discharge tube, as a third embodiment of the invention, Figure 7 is a detailed view of the second end of the discharge tube, as a fourth embodiment of the invention, Figure 8 is a detailed view of the second end of the discharge tube, as a fifth embodiment of the invention, in Figure 9. Schematic of the manufacturing steps of axially stepped metal ceramics using pressing technology, Figure 10 is a detailed view of the second end of the discharge tube, as a sixth embodiment of the invention, Figure 11 shows a further seventh embodiment of the invention, Figure 12 shows a further eighth embodiment of the invention, the Figure 13. Thermal expansion coefficient (K1) as a function of temperature (°C). In Figure 1, a metal halide discharge lamp 1 with a power of 150 W is shown very schematically. The lamp has an essentially cylindrical outer envelope 2 made of quartz glass, which is closed by flattening at the ends of the envelope 3 and is provided with tube heads 4. A ceramic envelope serves as a discharge vessel or discharge tube 5, which is enclosed in the outer envelope 2. The discharge tube 5 has a central longitudinal axis A, the two ends of which are made of aluminum oxide. The discharge tube 5 can be, for example, a cylindrical tube (not shown) or can be convex in the middle as shown in the figure. It is provided with cylindrical end parts 6a and 6b at the two ends. Two current leads 7a, 7b are inserted into ceramic-like (metal-ceramic) end plugs 8a and 8b in the end parts 6a and 6b, respectively. The first current lead 7a is formed by a molybdenum pin which is sintered directly into the first end plug 8a in the first end part 6a. The end plug 8a is a one-piece body made of a ceramic-like composite material (aluminum oxide and tungsten), which is known, for example, from EP-A 609 477. The second current lead 7b is a molybdenum tube which is placed directly in the end plug 8b in the second end part 6b. The end plug 8b is a multi-part plug. The internal tip of the current lead 7a, 7b has electrodes 9. It is preferable to apply an insulating coating 10, such as pure alumina, to the inner surface of the cermet end plugs 8a and 8b to prevent arcing between the plasma column of the arc discharge and the cermet end plugs 8a, 8b, which can cause darkening and leakage. The discharge tube 5 contains a charge containing an inert ignition gas, such as argon, and mercury and metal halide additives, such as iodides of rare earth metals. During the manufacture of the lamp, the tubular current passage 7b acts as a pump and a filling opening. The current passage 7b is thus used to vent and then fill the discharge tube 5. This technique is well known (see the references above). Only then is the current passage 7b closed. Figure 2 is a detailed view of the other end part 6b of the discharge tube 5. This figure shows that the metal ceramic end plug 8b consists of seven ring-like plug parts or zones 11a—11g. These are arranged axially overlapping and one behind the other. The inner surface 12 of the first, innermost, zone 11a borders the discharge space. Its outer surface 13 is in contact with the inner surface of the adjacent first, intermediate zone 11b. The innermost, zone 11a is made of pure alumina. The adjacent first, intermediate zone 11b contains 15 vol.% tungsten, the rest being alumina. The composition of the further zones follows the principles outlined above. The proportion of tungsten increases towards the outermost zone. There is 22% tungsten in the 11c zone, 27% tungsten in the 1ld zone, 32% tungsten in the alle zone, 37% tungsten in the llf zone, and 40% tungsten in the 1lg zone. In general, for seven zones, the preferred ranges of zone composition are as follows: - innermost annular zone 11a (first layer): 100% by volume of alumina, HU 221 366 Bl -the adjacent intermediate, zone 11b: 10-20% W, the rest alumina, - the second intermediate, zone 11c: 20-25% W, the rest alumina, - the third intermediate, 1 ld zone: 25-30% W, the rest aluminum oxide, - the fourth intermediate, He zone: 30-35% W, the rest aluminum oxide, - the fifth intermediate, llf zone: 35-40% W, the rest aluminum oxide, - the outermost annular, 1 lg zone (last layer): 40-43% W, the rest aluminum oxide. The thermal behavior of the outermost annular zone 11g matches that of the molybdenum tube 7b, which acts as a current conductor. The annular zone 11g is sintered directly to the current conductor 7b (molybdenum tube). In contrast, the other zones 11a-11f do not touch the molybdenum tube. A small gap 14, approximately 50 pm wide, remains between the current conductor 7b and the zones 11a-11f of the plug. In Figure 3, the absolute thermal expansion (in percent relative to 0 °C) is plotted as a function of the temperature of the tubular current conductor 7b (molybdenum, curve A), the outermost annular zone 1 lg (alumina, curve B) and two intermediate layers (alumina with 30% tungsten, curve C and alumina with 20% tungsten, curve D). A special solution is to use a metal ceramic containing tungsten as the metal component in combination with a conductor made of molybdenum. Tungsten has a much lower coefficient of thermal expansion than molybdenum. Therefore, adapting the desired characteristics of the annular zones 1 la—1 lg becomes easier if tungsten is added to the alumina, since a smaller amount of tungsten than molybdenum is sufficient to achieve the desired coefficient of thermal expansion of a given zone. In Figure 4, the absolute thermal expansion (in percent relative to 0 °C) is plotted at different temperatures T as a function of the tungsten content for different cermet end-plug zones. It can be seen that at high temperatures, with a tungsten content of about 40% (the rest being alumina), the thermal characteristics are similar to those of pure molybdenum (arrows). The difference between the absolute expansions of adjacent annular zones 1 la-1 lg is very small. The seven zones (11a-1 lg) are indicated by arrows. A second embodiment of the axially stepped closure is shown in Figure 5. The end plug or end closure member 25 consists of six plug parts 25a-25f. The outermost plug part 25f of the end plug 25 is here also directly connected to the tubular molybdenum current lead 26, while the innermost plug part 25a is sintered directly to the end part 6b of the polycrystalline alumina discharge tube. The innermost plug part 25a has a cylindrical structure. This means that it is inserted into the end part 6b of the vessel, but a radially extending rim 27 rests on the outer surface of the end part 6b. The distance between the inner radial surface 24 of the plug 25a facing the current lead 26 and the current lead 26 itself is approximately 5 mm. This annular space 28 surrounds the electrode 29 within the first zone of the end plug 25. Between the intermediate plug 25b-25e and the current lead 26 there is only a small annular capillary or gap of about 100 pm. The cylindrical shape used as the innermost annular plug 25a is bonded as follows. First, the metal ceramic end plug 25 and the current lead 26 are pre-fired together to form a subassembly. This is then mounted on the second open end 6b of a polycrystalline alumina tube (pre-fired or already sintered to transparency) and the entire assembly is heated to a high temperature to simultaneously bond the outermost annular plug (layer) 25f and the metal (tungsten or molybdenum) current lead 26, and the innermost annular plug (layer) 25a and the end 6b of the polycrystalline alumina tube. Generally, the metal ceramic end cap or end closure member 25 is formed by a layered, cylindrical structure having a central hole filled with a tungsten or molybdenum, tubular (or in another embodiment rod-like) current lead 26. The current lead 26 is connected to an axially disposed tungsten or molybdenum electrode 29 inside the discharge tube 5 and to a current lead outside the discharge tube 5. The metal ceramic hollow cylinder is formed by a multilayer metal ceramic in which the volume ratio of alumina to metal increases axially inward. The concentration of the metal phase increases from a low level in the first, innermost (lower) plug 25a (adjacent to the discharge space) to nearly 100% in the last, outermost (upper) plug 25f (farthest from the discharge space).The upper layer of the cermet, in which the level of the metal phase is high, is bonded directly to the current lead 26 (by direct sintering), while the first, lower, plug (layer) 25a of the cermet, which is essentially alumina (containing a very low level of metal phase), is bonded directly to the polycrystalline alumina discharge tube, which is preferably either elliptical or straight cylindrical. These two sinter bonds (direct bonds) create a hermetic seal and a nearly perfect thermal expansion match with both the metal lead and the polycrystalline alumina tube. The specific embodiment of Figure 5 has a six-layer structure. The thermal expansion coefficients of the ceramic metal plug or layer 25f-25a (from top to bottom) are designed to be 5.0; 5.5; 6.0; 6.5; 7.0; 7.5 -10~6 / °C, respectively. The thermal expansion coefficient of the upper plug (layer) 25f is almost exactly the same as the thermal expansion coefficient of the pure tungsten current guide 26 (4.8 10~6 / °C), and the thermal expansion coefficient of the lower plug (layer) 25a is quite close to the thermal expansion coefficient of the end portion 6b of the polycrystalline alumina tube (8 106 / °C). The axial thickness of the plugs (layers) 25b-25e in the sintered state can be as thin as 0.2 mm when using the layer-by-layer stacking technique. When using the sputtering technique, the layer thickness can be reduced to 0.01 mm, see R. Watanabe and A. Kawasaki “Recent Development of Functionally Gradient Materials for Special Application to Space Planes” HU 221 366 Β1 was published in Composite Materials, edited by AT Di Benedetto, L. Nicolais and R. Watanabe, pp. 197-208 (Elsevir Science, 1992). The axial thickness of the upper plug 25f and the lower plug 25a (layer) should be approximately the same as the wall thickness of the discharge tube 5 (0.5-0.8 mm) so that the zone in contact with the end part or the lead-through is long enough. This is favorable for achieving a permanent frit-free bond. The designed thermal expansion coefficients of the layers correspond to the following tungsten proportions (from the upper layer to the lower layer): 70, 52, 38, 24, 15 and 6 vol%. The corresponding proportion of tungsten in mass % is 92, 84, 75, 60.45 and 25 wt%. In other embodiments, the plug 18 is divided into even more sections, zones, or layers. In this way, the difference in thermal expansion behavior between adjacent sections is even smaller. The number of sections can be increased to ten, twelve, or even more layers. In a further preferred embodiment (FIG. 6), the layers or zones of the plug 18 are arranged telescopically. This means that the distance between the individual zones and the current path 26 decreases in steps from the innermost zone 18a to the last zone 18d. The outermost zone 18e is again sintered directly to the current path 26. In this embodiment, the current path 26 is made of molybdenum. The outermost zone (layer) 18e is an aluminum nitride layer (with a thermal expansion coefficient of 5.7 10~6 / °C, close to the thermal expansion coefficient of molybdenum of 5.0 106 / °C), which is adjacent to the molybdenum current path 26. The innermost zone 18a, as well as the intermediate or transitional zone (layer) 18b—18d between the aluminum nitride zone 18e and the end portion of the polycrystalline aluminum oxide tube 6b, are made of aluminum oxynitride with different ratios of aluminum oxide and aluminum nitride. The thermal expansion coefficient of aluminum oxynitride depends on the nitrogen content and is known to be 5 AlN. 9 in the case of Al2O3 7.8 10~6 / °C. A more promising embodiment utilizes the fact that aluminum nitride is known to be compatible with molybdenum, and an aluminum nitride / molybdenum metal ceramic is described (M. Tanaka, A. Kawasaki and R. Watanabe, “Thermomechanical Properties of SiCAlN-Mo Functionally Gradient Composites”, Funtai Oyobi Funmatsu Yakin, Vol. 39, No.4, pp. 309-313, 1992). Accordingly, the outermost layer in contact with the transition is made of aluminum nitride-molybdenum metal ceramic instead of pure aluminum nitride. The first intermediate layer adjacent to the outermost layer is made of pure aluminum nitride or of a metal ceramic containing aluminum nitride and molybdenum in different proportions. In another embodiment, the cermet zones are composed of alumina and non-metallic components such as metal carbides and metal borides. Examples of such components include tungsten carbide and tungsten boride, see Izumiya et al., U.S. Patent No. 4,825,126. In another embodiment for a 35 W lamp (Fig. 7), the arrangement is similar to that of Fig. 2. The second plug 32 consists of four non-conductive zones 32a-32d, which are arranged one behind the other in the axial direction. Since the amount of tungsten (60%) in the outermost zone 32d is large enough for welding, a weld seam 33 is made on the outer surface of the last layer, which connects the molybdenum tube 34 to the last zone 32d. For a 35 W metal halide lamp, the typical dimensions of the axially stepped closure are as follows: The length of the discharge tube 5 is 14 mm. Each end is closed with a plug 32, the total length of which is 5 mm. The plug 32 consists of four axially overlapping zones 32a-32d, in which the proportion of tungsten is 70% by weight, 50% by weight, 30% by weight and 10% by weight, respectively. The lower part or zone is partially inserted into the end of the tube by 2 mm. The first end of the tube has a molybdenum rod with a diameter of 0.3 mm and a total length of 16 mm as a lead-through, and the second end of the tube has a molybdenum tube with an inner diameter of 1.0 mm and an outer diameter of 0.8 mm as a lead-through. Only the second end of the tube is provided with a stepped plug, while the first end of the tube has a homogeneous plug 32, the components of which are identical to the components of the lower part of the stepped plug. This lower part contains 10% tungsten by weight, the rest is aluminum oxide. The hermeticity of the bond between metal and metal ceramic is based on the formation of a layer consisting of a solid solution. In a particularly preferred embodiment (Figs. 8a and 8b), the second plug 35 consists of four axially stepped layers. The innermost layer 35a contains 10 vol.% (typically 5-15 vol.%) molybdenum, the rest being aluminum oxide. The first layer 35a is inserted into the second end part 6b of the discharge vessel and sintered directly thereto. The first intermediate layer 35b contains 30 vol.% (typically 25-35 vol.%) molybdenum, the rest being aluminum oxide. The second intermediate layer 35c contains 45 vol.% (typically 40-50 vol.%) molybdenum, the rest being aluminum oxide. The outermost layer 35d contains 65 vol.% (typically more than 60 vol.%) molybdenum (or tungsten), the rest being aluminum oxide. The axially disposed passage 36 is a molybdenum rod with a diameter of 300 pm. The filling hole located laterally in the plug 35 is parallel to the passage 36. The diameter of the filling hole is 650 pm. 8a-8b. shows the situation after the discharge space has been evacuated and the filling components have been introduced. The rod-like hole stopper, whose length is approximately equal to the entire axial length of the plug 35, is ready to be inserted into the filling hole 37. The hole stopper 38 is preferably made of molybdenum or a metal ceramic containing a large amount of molybdenum or tungsten. The composition of the hole stopper 38 is most preferably the same as the composition of the outermost layer 35d of the plug 35. After the hole stopper 38 has been inserted into the filling hole 37 (Fig. 8b), a welded joint 39a is made between the outer end of the hole stopper 38 and the outer surface 40 of the outermost layer 35d of the plug. Additionally, a similar welded joint 39b is made between the outer end of the lead-through 36 HU 221 366 Bl and the outer surface 40 of the outermost layer 35d of the plug 35. The manufacture of the plug 35 begins by preparing the powder mixtures for each layer. For example, tungsten precursors such as ammonium tungstate or ammonium molybdate can be dissolved in water and mixed with alumina powder in a predetermined ratio (e.g., Baikowski CR 30, 15, 6, 1 powders of different average particle sizes) together with binders such as polyvinyl alcohol and / or polyethylene glycol. Materials that promote the sintering of alumina, such as MgO (derived from water-soluble magnesium nitrate), can be used. Alternatively, fine tungsten or molybdenum powder (e.g., OSRAM SYLVANIA, Towanda, Pennsylvania, USA, M-10 tungsten powder with an average particle size of 0.8 pm, or other types such as M20 (1.3 pm), M-37 (3 pm), M-55 (5.2 pm), and M-65 (12 pm)) can be mixed with water-dispersed alumina powder and ball milled (e.g., with alumina balls) to produce a uniform mixture.The resulting mixture can be spray-dried or pan-dried. The dried mixture is deagglomerated in a mill, such as a vibratory mill, to break up soft agglomerates. In the case of metal precursors, the mixture is heated (e.g. to 1000 °C in hydrogen, or in vacuum, or inert gas), which causes the precursor metal to break into particles. The mixed powder is then loaded into a mold containing a core rod (designed to fit the diameter of the tungsten or molybdenum tube or rod) and compacted to a given bulk density (e.g., at a pressure of 280 kPa). Powders for successive layers are prepared individually and fed into the mold, compacted again, and then a final layer of high tungsten content is added. The entire assembly is compacted at a pressure of 70-315 kPa and then removed from the mold. (The core rod can be designed to be stepped for each layer, so that the dimensional shrinkage of each layer is compatible with the subsequent processes, i.e., direct bonding of the tungsten tube forming the top layer and direct bonding of the polycrystalline alumina tube forming the bottom layer). The hollow cylindrical green body is then pre-fired at a relatively low temperature in hydrogen, or in a vacuum, or in an inert gas to remove the binding materials.This will prevent shrinkage and make the body somewhat firmer for handling. FIG. 10 shows a further embodiment similar to that of FIG. 7. Here again the second end of the receptacle of a 35 W metal halide lamp is shown. The second multi-part plug 32' here also consists of four axially overlapping zones 32'a-32'd, the composition of which is the same as that given in connection with FIG. 7. However, the second molybdenum tube 34', which acts as a lead-in, is recessed and only passes through the three outer zones 32'b-32'd. It is sintered directly to these three zones 32'b-32'd. The dimensions of this embodiment are as follows. The sintered thickness of the four zones 32'a-32'd is: the innermost zone 32'a is 1.7 mm, the adjacent intermediate zone 32'b is 0.5 mm, the second intermediate zone 32'c is 0.4 mm, and the outermost zone 32'd is 0.7 mm. In the embodiment shown in Fig. All., the ends of a polycrystalline alumina discharge vessel 41 are closed by disc-like insert members 42 also made of polycrystalline alumina. A multi-part structure 43 is disposed in the central bore of the insert member 42, which consists of five zones 43a-43e of different compositions. The cermet powders provide the following sintered thickness of the cermet: about 1.5 mm in the innermost zone 43a, which contains 10 wt.% tungsten, the rest of the alumina with 800 ppm magnesia; about 0.6 mm in the adjacent intermediate zone 43b, which contains 30 wt.% tungsten, the rest of the alumina with 800 ppm magnesia; 0.5 mm in the second intermediate zone, 43c, which contains 50 wt% tungsten, the rest being alumina with 800 ppm magnesia; 0.8 mm in the third intermediate zone, 43d, which contains 70 wt% tungsten, the rest being pure alumina; and 0.7 mm in the outermost zone, 43e, which contains 90 wt% tungsten, the rest being pure alumina.The stepped cermet structure 42 was assembled and connected to a molybdenum tube acting as a lead-in 44 by firing it in dry hydrogen for 1-2 hours at 1500-1600 °C. The lead-in 44 penetrated the three outer zones 43c-43e, but did not contact the two inner zones 43a and 43b. For the first closure, a first stepped metal-ceramic lead-through system was fired together with the discharge vessel 41 and the insert member 42, which had an outer diameter of 6.5 mm, an inner diameter of 2.5 mm, and a length of 2.5 mm. The remaining parts were formed by firing in wet hydrogen at about 1300-1400 °C for about one hour. The sealing length between the multi-part structure 43 and the insert member 42 was about 1-1.3 mm. The multi-part structure 43 was recessed about 0.8 mm into the insert member 42. This first closing firing produced a closed end structure. The other end was closed by inserting a second metal-ceramic lead-through system into this end and a second closing was performed. The entire assembly was then finally sintered in wet hydrogen at approximately 1900 °C for several hours. In Figure 13, the coefficient of thermal expansion of the various parts of the multi-part structure 43, the polycrystalline alumina material of the insert member and the discharge vessel, and the molybdenum boat, is plotted. Taking the typical operating temperature of the cermet multi-part structure 43 as 700°C, it can be seen that the difference in the coefficient of thermal expansion of adjacent parts is approximately 1.0.10-6 / K. Figure 12 shows another preferred embodiment, the temperature load of which is reduced. Here too, disc-shaped insert members 42 are provided at the ends of the discharge vessel 41. Both insert members 42 are made of polycrystalline alumina. The feed-through system consists of three members. A homogeneous capillary tube 45 is inserted into the central bore of the insert member 42. The capillary tube 45 is elongated. HU 221 366 BI is provided with a multi-part structure 46 which rests on it. The contact zone between them is surrounded by a polycrystalline alumina sleeve 47. The lead-through 48 is a molybdenum tube. The multi-part structure 46 is a multi-part cermet consisting of five (or four) layers. The innermost layer, 46a, contains 10 wt% tungsten and is 1.7 mm long. The first adjacent intermediate layer, 46b, contains 30 wt% tungsten and is 0.7 mm long. The second intermediate layer, 46c, contains 50 wt% tungsten and is 0.5 mm long. The rest are all alumina with 800 ppm magnesia. The third intermediate layer, 46d, contains 70 wt% tungsten and is 0.8 mm long. The outermost layer, 46e, contains 90 wt% tungsten and is 0.7 mm long. The lead-through (tube) 48 passes only through the three outer layers 46c-46e, but does not contact the inner layers 46a and 46b. An electrode system, not shown, is connected to the inner end of the lead-through 48. The lead-through 48 is sealed according to well-known methods. These are included in the cited prior art. The manufacturing process for this embodiment is as follows. During the first closing firing, the stepped metal ceramic is fired together with the lead-through with the pre-fired sleeve. The sleeve has an outer diameter of 5.3 mm, an inner diameter of 3 mm, and a length of approximately 5 mm. The sleeve is pre-fired for several hours at approximately 800-900 °C. The closing firing was carried out for a maximum of one hour in wet hydrogen at 1100-1200 °C. The length of the seal between the stepped metal ceramic and the sleeve (in the sintered state) was approximately 1.5 mm. The stepped metal ceramic was embedded in the sleeve by approximately 2.5 mm. The first closing firing produced one of the end structures. The firing temperature for the closure of the metal ceramic and the sleeve was chosen so that after co-firing the inner diameter of the sleeve fits into the outer diameter of the capillary tube of 2.8 mm. The capillary tube and the discharge vessel and the insert member are now permanently sintered together into one unit. Two previously closed parts were assembled with the capillaries at both ends of the discharge vessel. The entire unit was finally sintered in wet hydrogen at high temperature (approximately 1800-1950 °C) for a maximum of 30 minutes, and thus a hermetic bond was formed between the capillary tube and the metal ceramic through the sleeve. Figure 9 shows the pressing technology for the production of axially stepped metal ceramics. In the first step (Fig. 9a), a cylindrical mold 20 is filled with a pure alumina suspension 21a, which is made with an organic binder of the poly(vinyl alcohol) type. After the piston 22 has been retracted to a certain extent, the next suspension 21b, which consists of, for example, 90% alumina and 10% tungsten (Fig. 9b), is filled into the mold 20. This operation is repeated several times until the last suspension (sixth layer 21g in Fig. 9c) is filled. The latter consists of, for example, 60% alumina and 40% tungsten and has the same thermal behavior as the tube. During filling, the piston 22 is moved downwards step by step. Then (Fig. 9d) the metal ceramic plug is pressed with an additional piston (arrow). Then a hole 23 is drilled in the “raw” metal ceramic. This hole has a diameter that is sufficient to ensure that the metal ceramic has an optimal shrinkage ratio relative to the molybdenum tube to be inserted into the hole 23 (Fig. 9e). The plug is then pre-fired. In another embodiment, a mold containing a core rod is filled with metal ceramic powders in the following order: 70 wt% tungsten, 50 wt% tungsten, 30 wt% tungsten, 20 wt% tungsten, and 10 wt% tungsten. After filling, each powder is roughly mixed in the mold in turn. The upper and lower dies are operated after all layers have been filled, applying a uniaxial pressure of 280 kPa. The dies are then removed and the compacted metal ceramic is separated from the core rod. The inner diameter of the metal ceramic disk can then be reamed so that the inner layer, 21a-21f, is slightly larger than the inner diameter of the outer layer, 21g. In embodiments in which the via passes through all zones of the stepped cermet, an additional step is necessary: ​​to prevent close contact between the molybdenum tube and the layers 21a-21f of the multi-part structure or plug, which behave differently from the thermal behavior of the metal tube (compared to the layer 21g), these five zones are drilled a second time with a drill of slightly larger diameter than the drill used the first time (Fig. 9f). The resulting enlarged hole 24x forms a gap after the via is inserted. This gap should be as small as possible (typically 50 pm) so that the charge does not condense in the gap. Only then does the plug pre-fire. The tungsten or molybdenum tube or rod is inserted into the hole in the pre-fired, multi-layered, hollow cylindrical metal ceramic. The finished unit consisting of the plug and feedthrough with the gap 14 is shown, for example, in Figure 2. The plug and lead assembly is pre-fired (1200-1500 °C) or pre-fired and sintered at a relatively high temperature (e.g. 1800-2000 °C) to form a predetermined tight fit bond (e.g. between 4 and 18%) between the high tungsten or molybdenum content top layer and the metal lead. During firing, the top layer shrinks relative to the tungsten tube or molybdenum rod, thus creating a frit-free, hermetic bond. It is important to design the dimensional shrinkage of each layer relative to the gap between the tungsten or molybdenum part and the raw or pre-fired metal ceramic (by optimizing the particle size of the metal and the alumina phase and the compaction pressure) so that the other layers do not prevent the formation of a tight fit bond between the top layer and the tungsten or molybdenum part. The pre-fired and sintered unit can optionally be hot isostatically pressed at high temperatures (e.g. 1800 °C) to produce fully dense bodies. The sintered or hot isostatically pressed stepped plug with tungsten or molybdenum vias is then pressed into a pre-fired polycrystalline aluminum10 HU 221 366 B1 is placed in a alumina tube or inside the stem of a pre-fired, elliptical, polycrystalline alumina tube. Polycrystalline alumina can be produced by pre-firing (1000-1500 °C) an alumina raw body doped with sintering aids, MgO, MgO plus zirconia or MgO plus erbium oxide. Stepped metal-ceramic bodies with densified leads are placed at a predetermined distance at both ends of the pre-firing polycrystalline alumina shell. During sintering of the entire unit in hydrogen or nitrogen and hydrogen at 1800-2000 °C, the polycrystalline alumina tube densifies to transparency and shrinks in size, so that on the one hand a tight fit bond is created between the lower layer of the multi-part plug containing a low level of the metal phase and the polycrystalline alumina tube, and on the other hand a defined cavity length is created between the tips of the opposing electrodes.If the tungsten or molybdenum transition at one end of the polycrystalline alumina is a rod, this sintering process produces a closed, ready-to-dispense dome at one end. The degree of direct bond strength between the underlying metal ceramic layer and the polycrystalline alumina is determined by the gap between them, the pre-firing temperature used, and the sintering shrinkage. Lamp fillings containing various metal halides and filling gas can be introduced through a tubular tungsten or molybdenum via at one end of the lead-through-metal-ceramic closure. To complete the production of a complete discharge tube made of polycrystalline alumina, sealed with stepped metal ceramics, halide-resistant, with tungsten or molybdenum vias, the tungsten or molybdenum tubes can finally be sealed by laser (NdYAG or CO2) welding. In a preferred embodiment, the lower layer is cylindrical in shape. The pre-fired, cermet-metal-lead assembly can then be mounted on one open end of a pre-fired or already sintered to transparency polycrystalline alumina tube, and the entire assembly is heated to a high temperature to simultaneously create a shrink bond between the upper layer and the tungsten or molybdenum, and between the lower layer and the polycrystalline alumina. It is apparent that an insulating coating, such as pure alumina, can be applied to the cermet enclosure to prevent arcing between the plasma column and the cermet, which could cause darkening and leakage. To further improve the gas tightness of the seal, a frit can be applied to the outer surface (farthest from the discharge) of the top layer in the case of an axially stepped seal, or of the outermost layer in the case of a radially stepped seal. A preferred polycrystalline alumina discharge tube is made of alumina doped with 500 ppm MgO and optionally additionally with about 350 ppm Y2O3. The grain size of this ceramic is preferably as small as possible (below 1 pm) to improve mechanical strength. The lead-through, especially if tubular, is either flush with or preferably recessed into the inner (discharge-facing) surface of the plug. It is preferable to shorten the length of the bond between the innermost or lower layer and the polycrystalline alumina tube as much as possible. It is a good value to choose the bond length so that it is as small as the wall thickness of the polycrystalline alumina discharge tube. The principles of the invention are of course also applicable to the use of a tube made of other types of ceramics (e.g. Y2O3) in combination with other cermet materials. Instead of using the integrated end part of the discharge tube, a separate ring-shaped, ceramic end part can of course also be used. Preferably, only the innermost, lower region of the multi-part plug is inserted into the end portion of the discharge tube. This lower region requires a sufficiently large axial length. The inventive design creates a smooth gradient of thermal expansion of the metal ceramic, thus bridging the polycrystalline alumina discharge tube and the metal via. This is necessary to minimize thermal stresses during the cooling phase of the manufacturing cycle of the plug via units and during the on and off cycles of the lamp. The radially stepped cermet end plug can be produced by various methods, including pressing and spraying. A multi-layer structure can be formed by pressing in the axial direction. A powder mixture of alumina and metal (tungsten or molybdenum) can be prepared by ball milling an aqueous solution of alumina and metal powders and organic binders such as polyvinyl alcohol and / or polyethylene glycol. Metal precursors such as ammonium tungstate can be dissolved in water and alumina powder added. The ball milled slurry can be spray-dried or pan-dried. When a metal precursor is used, the mixture requires high temperature pyrolysis (e.g. 1000 °C) to produce metal particles. When a metal powder is used, the dried mixture for the innermost layer can be placed in a mold containing a core rod. The core rod is then removed and replaced with a smaller core rod. The powder mixture for the next layer is placed in the cavity between the core rod and the mold.The above loading operation is repeated with the following powder mixtures, followed by final compaction to obtain the final green body, which consists of several axially arranged layers. The green structure can then be removed and pre-fired at a relatively low temperature (1000-1500 °C), in vacuum, or in hydrogen or argon, to remove the binders. During pre-firing, the internal diameter of the metal ceramic can shrink by 0-10%, depending on the pre-firing temperature. It is important to choose the starting alumina powder and metal powder of sufficient particle size, as well as the solids loaded into the dense suspension, so that the several layers shrink evenly. HU 221 366 Bl Another way to form an axially multilayered structure is by spraying. A mixture of alumina powder and metal powder (tungsten or molybdenum) can be prepared by ball milling an aqueous solution of alumina and metal powders and organic binders such as polyvinyl alcohol, polyethylene glycol or polyox. Metal precursors such as ammonium tungstate can be dissolved in water and alumina powder added to it. The slurry ground in the ball mill can be sprayed onto a rotating, porous, slightly oversized polymer mandrel that is heated. Spraying can be done using a dual-jet, ultrasonic or electrostatic atomizer. The binder content and solids content of the slurry are selected so that the aqueous mixture adheres to and is deposited on the tungsten or molybdenum tube or rod. This is very similar to spraying a phosphor suspension onto the inside of a fluorescent tube.Slight heating of the mandrel during the spraying process can be beneficial in terms of better adhesion of the powder mixture to the metal and can improve the cohesion of the powder mixture itself. The spraying and application of successive layers is done with dense suspensions of decreasing metal content (as the mandrel moves axially) to create an axial gradient. The thickness of the layers can be as thin as 0.1 mm, according to the cited article by R. Watanabe and A. Kawasaki. The blank can be cold isostatically pressed and then pre-fired at relatively low temperatures in hydrogen, or nitrogen and hydrogen, or in vacuum to burn out the mandrel, remove the binders, and produce an axially stepped metal-ceramic. During pre-firing, the internal diameter of the metal-ceramic can shrink by 0-10%, depending on the pre-firing temperature. It is important to select the starting alumina powder and metal powder of sufficient particle size, the solids loaded into the dense suspension, and the temperature of the cold isostatic pressing step so that the multiple layers shrink coherently. The tungsten or molybdenum tube or rod is then placed in the central hole of the pre-fired, axially stepped metal-ceramic. The entire assembly is heated in hydrogen or nitrogen and hydrogen to high temperatures (1800-2000 °C) to sinter the metal-ceramic and to form a strong bond between the metal-ceramic and the metal-ceramic. The degree of strong bond is typically between 4 and 18%, depending on the dimensional shrinkage during sintering and the gap between the inner diameter of the pre-fired metal-ceramic and the outer diameter of the metal-ceramic. The sintered metal-ceramic and metal-ceramic assembly can optionally be hot isostatically pressed at high temperatures to further reduce the remaining pores. The sintered metal ceramic and lead assembly is placed inside a pre-fired straight polycrystalline alumina tube or inside the straight portion of a pre-fired elliptical polycrystalline alumina bulb. The polycrystalline alumina consists of alumina, preferably doped with MgO or MgO plus zirconia. The entire assembly is sintered in hydrogen or hydrogen and nitrogen to densify the polycrystalline alumina to transparency. During sintering, the polycrystalline alumina shrinks to the outer diameter of the metal ceramic and forms a tight fit. The degree of tight fit depends on the shrinkage of the polycrystalline alumina and the gap between the metal ceramic and the pre-fired polycrystalline alumina.The pre-fired polycrystalline alumina must have a sintered metal ceramic and via assembly at both ends so that when the polycrystalline alumina is sintered, the gap between the electrode tips is reduced to a defined cavity length for the lamp. If the via of the sintered end structure at one end of the polycrystalline alumina is a rod, the sintering step produces a closed envelope at one end containing hermetically sealed vias ready for dispensing. It is possible to simultaneously create a strong fit bond between the outermost layer and the tungsten or molybdenum tube and the innermost layer and the polycrystalline alumina by one-step sintering, during which the pre-fired stepped metal ceramic solidifies to near full density and the polycrystalline alumina is sintered to transparency. Lamp fills containing various metal halides, mercury and fill gases can then be introduced into the bulb through a tubular tungsten or molybdenum lead at one end of the lead and metal ceramic closure. The tungsten or molybdenum tubes can then be sealed using a laser (NdYAG or CO2) welding method to form a complete bulb made of polycrystalline alumina, sealed with stepped metal ceramics and halide-resistant tungsten or molybdenum leads (see Figure 1). This technique is well known. If the last layer of the second plug is weldable, it can be conductive or non-conductive.

Claims

PATENT CLAIMS 1. Ceramic lampshade for a high-pressure discharge lamp, comprising: - a transparent ceramic discharge tube (5) having a first end portion (6a) and a second end portion (6b), the discharge tube (5) defining a discharge space and having a longitudinal axis of symmetry; - a first end plug (8a) made of a non-conductive, metal-ceramic material, with which the first end portion (6a) of the discharge tube (5) is closed; - a second end plug (8b) made of a non-conductive, metal-ceramic material, with which the second end portion (6b) of the discharge tube (5) is closed; - a first and a second metal feedthrough (7a, 7b) passing through the first and second end plugs (8a, 8b), respectively; - an electrode (9) at the inner end of the first and second feedthrough (7a, 7b), respectively; characterised in that - at least the second end plug (8b) has a multi-part structure (43, 46) consisting of at least three parts;HU 221 366 Bl - each lead-through (7a, 7b) has an inner end and an outer end, and the leads (7a, 7b) are made of one of the metals or alloys in the group consisting of tungsten, molybdenum and rhenium and alloys of at least two of these metals; - the thermal expansion coefficient of at least one part of the multi-part end plug (8b) is between the thermal expansion coefficient of the discharge tube (5) and the thermal expansion coefficient of the lead-through (7b); - the multi-part structure (43, 46) comprises at least four axially overlapping parts with different thermal expansion coefficients, including a first and a last part; the first part is the innermost part relative to the discharge space, the last part is the outermost part relative to the discharge space; - the multi-part end plug (8b) is sintered directly to the discharge tube (5) and the feedthrough (7b) in such a way that the innermost part of the end plug (8b) is sintered directly to the discharge tube (5) and the outermost part of the end plug (8b) is sintered directly to the corresponding feedthrough (7b); 2. Ceramic envelope according to claim 1, characterized in that the composition of the parts differs from each other in the proportion of metal added to the alumina.

3. Ceramic bulb according to claim 1, characterized in that the different parts have different components in their composition.

4. Ceramic envelope according to claim 1, characterized in that the end plug (8b) is formed by a layered, cylindrical structure with a central bore, in which at least the outermost, last layer adjacent to the second passage (7b) is in gas-tight contact with this passage (7b).

5. Ceramic envelope according to claim 4, characterized in that only the outermost layer adjacent to the passage (7b) is in gas-tight contact with this passage, and the distance between the layers of the passage (7b) and the end plug (8b) is at least 1 mm, excluding the outermost layer.

6. Ceramic bulb according to claim 4, characterized in that the lead-through (7b) is recessed into the end plug (8b) and, starting from the outermost layer, passes not through all but only a few zones or layers.

7. Ceramic envelope according to claim 5, characterized in that the distance between the layers and the second via (7b) decreases telescopically or is continuously curved as the distance between the layer and the discharge space increases.

8. Ceramic envelope according to claim 1, characterized in that only the innermost layer is at least partially present in the end of the discharge tube (5).

9. Ceramic bulb according to claim 1, characterized in that the second end plug (8b) consists of at least five axially arranged parts.

10. The ceramic bulb according to claim 1, characterized in that the first, innermost plug portion (25a) of the second end plug (8b) has a cylindrical structure.

11. Ceramic bulb according to claim 1, characterized in that the second passage (7b) is tubular.

12. Ceramic bulb according to claim 1, characterized in that the end of the discharge vessel (5) is closed by a disc-like insert member (42) having a central bore for the multi-part structure (43, 46) which is recessed in the insert member (42).

13. Ceramic cap according to claim 1, characterized in that the amount of metal in the last, outermost plug portion of the second end plug (8b) is at least 50% by volume.

14. Ceramic cap according to claim 13, characterized in that the last, outermost plug part of the second end plug (8b) is weldable.

15. Ceramic bulb according to claim 14, characterized in that the second lead-through (7b) is welded to the last, outermost plug part of the second end plug (8b).

16. Ceramic cap according to claim 12, characterized in that the second plug part has a separate filling hole (37) or bore.

17. A ceramic cap according to claim 16, characterized in that a hole stopper (38) made of weldable material fits into the filling hole (37) or bore.

18. Ceramic cap according to claim 17, characterized in that the hole stopper (38) is welded to the outer surface of the last plug part of the second end plug (8b).

19. Ceramic envelope according to claim 1, characterized in that the ceramic material of the discharge tube (5) consists of aluminum oxide doped with magnesia and possibly additionally with yttrium oxide.

20. The ceramic envelope of claim 1, wherein the stepped cermet body is made of pure alumina at least in the outermost zone and magnesia-doped alumina at least in the inner zone.

21. A ceramic envelope according to claim 1, characterized in that the multi-part structure (43, 46) is bonded to a hollow member on the discharge side, and the bonding zone is surrounded by a sleeve (47).

22. The ceramic cap according to claim 1, characterized in that the first end plug (8a) is a one-piece body or a multi-piece body similar to the multi-piece end plug (8b).

23. Ceramic envelope according to claim 1, characterized in that the difference in thermal expansion coefficients of adjacent parts of the multi-part structure (43, 46), including the discharge tube (5) and the associated lead-through (7b), is preferably less than 1.0 10-6 / K.

24. A lamp, characterized in that it is provided with a ceramic lampshade according to any one of claims 1-23.

25. A method for producing a ceramic bulb with a metal-ceramic plug, characterized in that a ceramic bulb according to any one of claims 1-23 is produced by producing the multi-part structure of the multi-part end plug by layer-by-layer material extrusion or spraying.