Note: Descriptions are shown in the official language in which they were submitted.
2139839
SEALING STRUCTURE FOR LIGHT-EMITTING BULB ASSEMBLY
AND ~;THOD OF M~IUFACTUP~NG SAME
Technical Field
The present invention relates to a sealing struc-
ture for a light-emitting bulb assembly for use in a metal-
vapor discharge lamp such as a mercury-vapor lamp, a metal
halide lamp, or a sodium-vapor lamp, or a high-intensity dis-
charge lamp, and a method of manufacturing such a light-emit-
ting bulb assembly.
Background Art
Metal-vapor discharge lamps include a mercury-vapor
lamp, a metal halide lamp, and a sodium-vapor lamp. The mer-
cury-vapor lamp emits light excited from the mercury in a
positive column produced in a hot-cathode arc discharge. In
the metal halide lamp, a metal halide is evaporated into a
metal and a halogen by the heat of a mercury hot-cathode arc
discharge to emit light in a color inherent in the metal.
The sodium-vapor lamp emits light in yellowish orange at a D
line (589.0 nm, 589.9 nm) produced by a hot-cathode arc of a
sodium vapor. Heretofore, such metal-vapor discharge lamps
have been used as illuminating lamps for gymnasiums and fac-
tories, light sources for overhead projectors and color liq-
uid crystal projectors, fog lamps for automobiles, and so on.
The bulbs of metal-vapor discharge lamps were ini-
tially made of quartz glass. However, since the quartz glass
has poor fade resistance and a large thermal capacity, the
;~ 2139~39
-- 2 --
metal-vapor discharge lamps cannot be turned on quickly and
the individual bulbs have large dimensional variations.
Therefore, it has recently been proposed to make bulbs of
light-transmissive ceramic.
Generally, a light-emitting bulb assembly for a
discharge lamp comprises a bulb made of light-transmissive
ceramic in the form of fired alumina or the like, and a clo-
sure by which an electrode supported by an electrode support
is sealed and fixed in the bulb. To join the closure hermet-
ically to an open end of the bulb, a glass solder is filled
in a gap between end and inner surfaces of the open end of
the bulb and a confronting surface of the closure, heating
the glass solder to melt same, and then cooling and solidify-
ing the melted glass solder.
It is the general practice for the closure to have
the same coefficient of thermal expansion as and to be as
chemically stable against metal vapor and halogen vapor as
the bulb or the electrode support.
When the closure is joined to the bulb by the glass
solder, a starting rare gas and a discharging metal component
depending on the discharge lamp which incorporates the bulb
assembly, e.g., mercury if the discharge lamp is a high-pres-
sure mercury vapor lamp, or a metal halide if the discharge
lamp is a metal halide lamp, are sealed in the bulb.
The bulb assembly is turned on, its temperature mo-
mentarily increases from the atmospheric temperature to 900C
at which the bulb assembly remains energized stably. High
~139839
-- 3 --
thermal stresses are developed in the bulb assembly due to
such a large thermal change and a change in the internal
pressure.
When thermal stresses are produced, thermal strains
are developed in a portion having a different coefficient of
thermal expansion, specifically the closure that is inter-
posed between the bulb and the electrode support, tending to
cause the closure to be bro~en. More specifically, cracks
are produced in the closure itself and the glass solder which
has lower heat resistance than the light-transmissive ceramic
and the closure because of its composition, allowing the dis-
charging metal component to leak out of the bulb. As a re-
sult, the bulb assembly is not liable in producing stable
light emission, and the service life of the lamp is limited.
In a high-temperature, high-pressure environment in
which the temperature and the internal pressure of the bulb
assembly are increased, a metal halide (e.g., ~lI3, NaI, or
the like) sealed as a discharging metal component is liber-
ated as ions which erode the bulb assembly.
The liberated ions erode the glass solder more
quickly because the glass solder has lower erosion resistance
than the light-transmissive ceramic and the closure because
of its composition. The glass solder is liable to crack also
due to the low erosion resistance against the erosion caused
by the liberated ions.
Highly pure light-transmissive alumina which is
used in the bulb has poor wettability with respect to the
2139839
-- 4
glass solder. Therefore, the bonding strength at the bound-
ary between the glass and the bulb is low, tending to produce
cracks and a leakage of the sealed gas.
Various arrangements have heretofore been proposed
in order to solve the above problems.
Japanese lai~-open patent publication No. 1-143132
discloses a technique for brazing an insert having a coeffi-
cient of thermal expansion similar to that of alumina to a
sealed region of an outer circumferential element of alumina
which corresponds to a bulb. According to Japanese laid-open
patent publication No. 63-308861, a closure is composed of a
central body and an annular body disposed around the central
body, and a bulb is joined in solid phase to the closure (the
central body and the annular body). Japanese laid-open
patent publication No. 63-308861 particularly proposes spe-
cific dimensions and compositions of the central body and the
annular body which make up the closure. Specified dimensions
are also proposed in Japanese laid-open patent publication
No. 62-21306.
The disclosed proposals are effective in suppress-
ing a leakage of the discharqing metal component from the
bulb assembly for thereby keeping reliable light emission and
increasing the service life of the lamp
However, recent years have seen a demand for
brighter light emission to achieve higher added values of
light-emitting bulb assemblies, and it has been practiced to
increase the temperature of a liqht-emitting bulb assembly up
~ ~ 2139~83~
to about 1200C in excess of the conventional temperature of
900C in order to attain brighter light emission.
Since the higher bulb temperature leads to corre-
sponding thermal stresses in the bulb assembly, the conven-
tional light-emitting bulb assembly fails to keep suffi-
ciently reliable light emission and have a sufficiently long
service life. Specified dimensions of the closure and other
parts are not preferable as they pose limitations on the con-
figurations of the light-emitting bulb assembly and also the
configurations of the lamp which accommodates the light-emit-
ting bulb assembly.
The present invention has been made in order to
solve the above problems. It is an object of the present in-
vention to provide a light-emitting bulb assembly which is
highly reliable and has a long service life, and particularly
a novel sealing structure for such a light-emitting bulb as-
sembly and a simple method of manufacturing such a light-
emitting bulb assembly.
Disclosure of the Invention
Means and processes employed according to the pre-
sent invention for achieving the above object are as follows:
A sealing structure for a light-emitting bulb as-
sembly, includes a closure having a core which serves as an
electrode and sealing an open end of a bulb, the closure in-
cluding a bulb-side region disposed adjacent to the open end
of the bulb and made of a compositional ingredient having a
coefficient of thermal expansion which is substantially the
1 3~9 8 3 9 - ~
same as that of the bulb, a core-side region disposed adja-
cent to the core and made of a compositional ingredient hav-
ing a coefficient of thermal expansion which is substantially
the same as that of the core, and an intermediate region dis-
posed between the bulb-side region and the core-side region
and made of a compositional ingredient having compusitional
proportions adjusted such that a coefficient of thermal ex-
pansion thereof varies gradually from the coefficient of
thermal expansion of the bulb-side region toward the coeffi-
cient of thermal expansion of the core-side region.
Preferably, layers of the closure are progressively
thicker from the bulb-side region layer toward the core-side
region layer.
The bulb should preferably be made of light-trans-
missive ceramic, particularly highly pure alumina, and the
core should preferably be made primarily of tungsten.
The closure may be made of a gradient function ma-
terial.
The above sealing structure may be manufactured by
a method given below.
A method of manufacturing a light-emitting bulb as-
sembly including a closure having a core which serves as an
electrode and sealing an open end of a light-transmissive
bulb, comprises the steps of:
(a) preparing, from a fine powder of a light-trans-
missive bulb ingredient and a fine powder of a core ingredi-
ent, a bulb ingredient suspension in which the light-trans-
`- 2139839
.
missive bulb ingredient is greater than the core ingredient,
a core ingredient suspension in which the core ingredient is
greater than the light-transmissive bulb ingredient, and at
least-one intermediate suspension in which the light-trans-
missive bulb ingredient and the core ingredient have composi-
tional proportions lying between those of the bulb ingredient
suspension and the core ingredient suspension;
(b) forming an unfired laminated body composed of
an unfired bulb-side region layer to be disposed adjacent to
the light-transmissive bulb and formed from the bulb ingredi-
ent suspension, an unfired core-side region layer to be dis-
posed adjacent to the core and formed from the core ingredi-
ent suspension, and at least one unfired intermediate region
layer disposed between the unfired bulb-side region layer and
the unfired core-side region layer and formed from the at
least one intermediate suspension; and
(c) firing the unfired laminated body.
The step (b) may comprise the steps of:
(d) pouring the bulb ingredient suspension into a
cavity defined in a mold assembly composed of a plurality of
joined molds each made of a porous material, causing a sol-
vent of the bulb ingredient suspension to penetrate into the
mold assembly, and thereafter discharging an excessive amount
of the bulb ingredient suspension fro~ the mold assembly,
thereby forming the bulb-side region layer on an inner sur-
face of the cavity;
(e) thereafter, successively pouring the at least
2 1 ~ 3 8 3 ~
-- 8 --
one intermediate suspension and the core ingredient suspen-
sion onto an inner surface of the bulb-side region layer, al-
lowing solvents of the at least one intermediate suspension
and the core ingredient suspension to penetrate into the mold
assembly, and thereafter discharging excessive amounts of the
at least one intermediate suspension and the core ingredient
suspension from the mold assembly, thereby forming a molded
laminated body; and
(f) separating the molds from each other, thereby
releasing the molded laminated body as the unfired laminated
body.
Alternatively, the step (b) may comprise the steps
of producing green sheets respectively from the core ingredi-
ent suspension, the at least one intermediate suspension, and
the bulb ingredient suspension, and successively winding the
green sheets around the core, thereby for~ing the unfired
laminated body.
In the above sealing structure, the core comprises
a conductive core made of tungsten or the like, and the clo-
sure hermetically joined in solid phase to the opening of the
bulb comprises a fired laminated body composed of a core-side
region layer, at least one intermediate region layer, and a
bulb-side region layer which are successively arranged from
the conductive core toward the bulb. The core-side region
layer includes at least 50 % by volume of an ingredient of
the conductive core, and the bulb-side region layer includes
at least 80 % by volume of an ingredient of light-transmis-
2139~39
g
sive ceramic. The intermediate region layer between thecore-side region layer and the bulb-side region layer in-
cludes light-transmissive ceramic having a volume ratio which
is progressively closer to the volume ratio of the light-
transmissive ceramic of the bulb-side region in a direction
toward the bulb-side region, and also includes the ingredient
of the core having a volume ratio which is progressively
closer to the volume ratio of the ingredient of the core in
the core-side region layer in a direction toward the core-
side region layer.
In each of the layers of the closure, a network
structure of crystals is formed between common ingredients by
firing, thereby integrally joining the ingredients. A firing
process for reducing surface energy is applied to the joining
of the core and the opening of the bulb to each other.
Impurities such as of glass are often added in a small amount
in an effort to accelerate the firing process.
More specifically, each of the layers traps the
powder of the ingredient of the conductive core, and the in-
gredient of the light-transmissive ceramic forms a solid so-
lution and is crystallized. Adjacent layers are integrally
joined to each other in solid phase as the ingredient of the
light-transmissive ceramic in the layers forms a solid solu-
tion and is crystallized at the mating surfaces of the lay-
ers. The conductive core and the core-side region layer are
also integr~lly joined to each other in solid phase because
the ingredient of the light-transmissive ceramic in the core-
- 2139839
-- 10 --
side region layer is crystallized in contact with the core,
forming a glassy substance which fills in its grain bound-
aries, and also because the ingredient of the conductive core
is contained in both the core and the core-side region layer.
Furthermore, the bulb-side region layer and the bulb are also
integrally joined to each other in solid phase because the
ingredient of the light-transmissive ceramic in the bulb-side
region layer is crystallized in contact with the bulb, form-
ing a glassy substance which fills in its grain boundaries,
and also because the ingredient of the light-transmissive ce-
ramic is contained in both the bulb-side region layer and the
bulb.
Therefore, the closure after it has been fired ls
firmly bonded to the conductive core, making it possible to
seal a main electrode. Additionally, the closure after it
has been fired makes it possible to hermetically seal the
opening of the bulb through the formation of a glass phase in
the grain boundaries of the ingredient of the light-transmis-
sive ceramic in the bulb-side region layer and the bulb.
In addition, the distribution of coefficients of
thermal expansion from the conductive core through the core-
side region layer, the intermediate region layer, and the
bulb-side region layer to the bulb is a gradient distribution
ranging from the coefficient of thermal expansion of the con-
ductive core to the coefficient of thermal expansion of the
bulb.
In the method of manufacturing the sealing struc-
2139839
ture, when the closure to be hermetically joined in solidphase to the opening of the bulb which is made of light-
transmissive ceramic is to be fired, an unfired core-side re-
gion layer, an unfired intermediate region layer, and an un-
fired bulb-side region layer are successively stacked on a
core made of a conductive material, thereby forming an un-
fired l~m; n~ted body.
The unfired core-side region layer, the unfired in-
termediate region layer, and the unfired bulb-side region
layer which are successively stacked are formed from a core
ingredient suspension including a powder of a conductive ma-
terial ingredient or a core ingredient and a powder of a
light-transmissive ceramic ingredient or a bulb ingredient,
with at least 50 % by volume of the conductive material in-
gredient, a bulb ingredient suspension including both powders
with at least 80 ~ by volume of the light-transmissive ce-
ramic ingredient, and a plurality of intermediate suspensions
including both powders with the volume ratio of the light-
transmissive ceramic ingredient being progressively increased
to a value close to 100 % and the volume ratio of the conduc-
tive material ingredient being progressively reduced from 100
% .
To successively deposit the unfired core-side re-
gion layer, the unfired intermediate region layer, and the
unfired bulb-side region layer on an outer surface of the
core, they are deposited in a descending order of volume ra-
tios of the-conductive material ingredient, thereby forming
2 1 3 9 8 3 ~ -
- 12 -
the unfired laminated body. Thereafter, the unfired lami-
nated body is disposed at the opening of the bulb so as to
position the main electrode connected to the core in the
bulb, and then fired.
After the laminated body has been fired, since the
light-transmissive ceramic ingredient forms a solid solution
and is crystallized, trapping the powder of the core ingredi-
ent, in each of the layers, the fired closure is of an inte-
gral structure achieved by the formation of a solid solution
of and crystallization of the light-transmissive ceramic in-
gredient between adjacent ones of the layers. The fired clo-
sure is firmly bonded to the core, making it possible to seal
the main electrode, through the formation of a glass phase in
the grain boundaries of the light-transmissive ceramic ingre-
dient in the core-side region layer while it is being held in
contact with the core, and also through the coexistence of
the conductive core ingredient. The fired closure also makes
it possible to hermetically seal the opening of the bulb
through the formation of a glass phase in the grain bound-
aries of the light-transmissive ceramic ingredient in the
bulb-side region layer and the bulb.
Moreover, the distribution of coefficients of ther-
mal expansion from the core through the core-side region
layer, the intermediate region layer, and the bulb-side re-
gion layer to the bulb is a gradient distribution ranging
from the coefficient of thermal expansion of the core to the
coefficient of thermal expansion of the bulb.
213g839
- 13 -
Brief Description of the Invention
Fig. ~ is a cross-sectional view of a light-emit-
ting bulb asse~bly according to a first embodiment of the
present invention;
Fig. 2 is a graph showing a particle diameter dis-
tribution in light-transmissive alumina used to produce a
bulb and a closure of the light-emitting bulb assembly;
Fig. 3 is a diagram showing a process of manufac-
turing the closure of the light-emitting bulb assembly;
Fig. 4 is a perspective view of the closure;
Figs. 5(a) through 5(c) are a cross-sectional view
showing the structure of the closure and diagrams showing
composition distributions of the closure;
Fig. 6 is a diagram showing a process of manufac-
turing a closure of a light-emitting bulb assembly according
to a second embodiment of the present invention;
Fig. 7 is a perspective view of an unfired molded
body which will be fired into the closure;
Figs. 8(a) and 8(b) are perspective views of a mat-
ing mold assembly used to produce the closure;
Fig. 9 is a perspective view of the mating mold as-
sembly with an auxiliary member attached thereto;
Figs. lO(a) and lO(b) are views illustrative of a
process of manufacturing the closure;
Fig. 11 is a cross-sectional view of the closure
which is molded in the mating mold assembly;
Figs. 12(a) and 12(b) are diagrams showing composi-
2139~3g
- 14 -
tion distributions of the closure;
Fig. 13 is a cross-sectional view of the unfired
closure with an electrode attached thereto;
Fig. 14 is a cross-sectional view of the closure as
it is mounted in a bulb;
Fig. 15 is a cross-sectional view of a light-emit-
ting bulb assembly according to a modification of the first
embodiment;
Fig. 16 is a cross-sectional view of a light-emit-
ting bulb assembly according to a third embodiment of the
present invention;
Fig. 17 is a diagram showing a process of preparing
a slip for a closure of the light-emitting bulb assembly;
Figs. 18(a) through 18(e) are diagram showing a
slip-casting process;
Fig. 19 is a cross-sectional view of a light-emit-
ting bulb assembly according to a modification of the third
embodiment;
Fig. 20 is a cross-sectional view of a light-emit-
ting bulb assembly according to a fourth embodiment of the
present invention;
Fig. 21 is a diagram showing materials used to man-
ufacture the light-emitting bulb assembly;
Fig. 22 is a diagram showing respective slips used
to manufacture the light-emitting bulb assembly;
Figs. 23(a) through 23(f) are views showing a pro-
cess of manufacturing the light-emitting bulb assembly;
2139~39
- 15 -
Fig. 24 is a cross-sectional view of a light-emi~-
ting bulb assembly according to a fifth embodiment of the
present invention;
Fig. 25 is a diagram showing respective slips used
to manufacture the light-emitting bulb assembly;
Fig. 26 is a perspective view of a tubular pipe
used to manufacture the light-emitting bulb assembly;
~ igs. 27~a) and 27(b) are views showing a process
of manufacturing the light-emittin~ bulb assembly;
Fig. 28 is a cross-sectional view of a light-emit-
ting bulb assembly according to a sixth e~bodiment of the
present invention;
Figs. 29(a) through 29(e) are views showing slips
used to manufacture a closure of the light-emitting bulb as-
sembly and a process of manufacturing the closure; and
Figs. 30(a) through ~O(d) are views showin~ a modi-
fication of the process of manufacturing the closure.
Best Mode For Carrying Out the Invention
Preferred embodiments of light-emitting bulb assem-
blies according to the present invention will be described
below with reference to the drawings.
As shown in Fig. 1, a light-emitting bulb assembly
according to a first embodiment of the present invention com-
prises a tubular bulb lF, a closure 2 fixedly mounted in an
electrode holding hole la defined in a larger-diameter open
end of the bulb lF, a closure 2A fixedly mounted in an elec-
trode holding hole lb defined in a smaller-diameter open end
2133839
- 16 -
of the bulb lF, and a pair of main electrodes 3 disposed in
the bulb lF. The main electrodes ~ are in the form of tung-
sten coils, respectively, which are supported by respective
support shafts 4 of tungsten which extend through the clo-
sures 2, 2A. The closures 2, 2A differ from each other only
with respect to their diameters, and are produced by a manu-
facturing process which will be described later on.
The end of the bulb lF with the electrode holding
hole lb has a slender introduction tube lc for entering a
starting rare gas metal and various discharging material
amalgams. The slender introduction tube lc has an open end
sealed by a sealant ld of a cermet of alumina or a metal such
as nickel or the like.
A process of manufacturing the light-emitting bulb
assembly 1, including a process of manufacturing the bulb lF
and the closure 2, and the manner of supporting the main
electrodes 3 with the support shafts 4 will successively be
described below.
Synthesis of a fine powder of alumina which will be
used as a material of the bulb and the closure will first be
described below.
To synthesize a fine powder of alumina, an aluminum
salt which will become alumina having a purity of 99.98 mol %
or more when thermally decomposed is used as a starting mate-
rial,
An aluminum salt for synthesizing such highly pure
alumina may be am~onium alum or aluminum ammonium carbonite
~13~39
- 17 -
hydroxysite ~NH4AlC03(OH)2).
The aluminum salt is then weighed, dissolved to-
gether with a dispersing agent in distilled water, thus pro-
ducing a suspended aqueous solution, and then dried by a
spray drying process. The dried aluminum salt is thereafter
thermally decomposed, thereby producing a fine powder of alu-
mina only. The dried aluminum salt is thermally decomposed
at 900 - 2000C, e.g., 1050GC, in the atmosphere for 2 hours.
The fine powder of alumina produced by the spray drying pro-
cess and the thermal decomposition has an average particle
diameter ranging from 0.2 to 0.3 ~m and a purity of 99.99 mol
% or higher. The fine powder of alumina is thus prepared.
The synthesized fine powder of alumina is obtained as a sec-
ondary a~gregate of fine powder of alumina having the above
particle diameter, the secondary aggregate being of a size
greater than the above particle diameter.
As another material of the closure than alumina, a
fine powder of tungsten is prepared which has a purity of 99
mol % or higher and an average particle diameter of about 0.5
The bulb lF and the closure 2 are fabricated of the
above materials, respectively.
The bulb lF is manufactured as follows:
To the synthesized fine powder of alumina
(secondary aggregate), there is added an organic binder which
is composed primarily of an ac~ylic thermoplastic resin. The
fine powder of alumina and the added organic binder are mixed
2139~39
- 18 -
with each other in a wet manner using an organic solvent such
as of alcohol, benzene, or the like by a plastic (nylon) ball
mill for about 24 hours, so that the fine powder of alumina
and the organic binder are sufficiently wetted. The mixture
is then distilled and dried, thereby removing the solvent,
and kneaded into a compound having a desired viscosity rang-
ing from 50,000 to 150,000 cps.
The organic binder is a mixture of an acrylic ther-
moplastic resin, paraffin wax, and atactic polypropylene.
The total amount of the organic binder with respect to 100 g
of the fine powder of alumina is 25 g.
The ingredients of the organic binder are of the
following proportions, and adds up to the total amount (25 g)
of the organic binder:
Acrylic thermoplastic resin 20 ~ 23 g
(preferably, 21.5 g)
Paraffin wax 3 g or less
(preferably, 2.0 g)
Atactic polypropylene 2 g or less
(preferably, l.S g)
The mixture is distilled and dried at 130C for 24
hours, and thereafter kneaded at 1~0C by a roll mill of alu-
mina into the compound having the desired viscosity.
Subsequently, the compound is injection-~olded into
a molded body shaped as shown in Fig. 1 by a mold assembly
(not shown). The molded body is heated in a nitrogen atmo-
sphere up to a temperature at which the organic binder of the
2 1 3 9 8 3 9
- 19 -
acrylic thermoplastic resin, etc is thermally decomposed and
fully carbonized, so that the molded body is degreased. The
specific upper limit temperature up to which the molded body
is to be heated in this initial heat treatment may be deter-
mined depending on the capability of a heat treatment furnace
used and the temperature at which the organic binder is ther-
mally decomposed. In this embodiment, the molded body is
heated from room temperature (20C) to 450C in 72 hours.
Other processing conditions are given below. While the
molded body is being heated up to 450~C, it is kept under a
constant pressure.
Processing pressure 1 ~ 8 kg/cm2
(optimum pressure: a kg/cm2)
Time required to heat the molded body from 20C
to 450C 72 hours or shorter
In the initial heat treatment, the added organic
binder composed of an acrylic thermoplastic resin, paraffin
wax, and atactic polypropylene is thermally decomposed and
carbonized, so that the molded body is degreased.
Then, the molded body (degreased body) is fired in
the ~.mosphere by subsequent heat treatment under conditions
given below, thereby producing a fired body. The molded body
is heated at a rate of 100C/hour.
Processing temperature 1200 - 1300~C
(optimum temperature: 1235~C)
Time during which the molded body is kept at the
processing temperature 0 4 hours
- 2139~83-9~
. ~ ,. ..
- 20 -
(optimum time: 2 hours).
The molded body is fired by the subsequent heat
treatment in the temperature range of from 1200 to 1300C for
the reasons that the density of the fired molded body will be
9S % or more of the theoretical density for being subject to
su~sequent hot isostatic pressing, and large crystals will
not be produced in the fired body. If the molded body were
fired at a temperature lower than 1200C, then the density of
the fired molded body would be less than 95 % of the theoret-
ical density and the molded body would not be subject to hot
isostatic pressing. If the molded body were fired at a tem-
perature higher than 1300C, then the fired body would have
large crystals at a greater frequency, and would not be suf-
ficiently strong.
The molded body is thus fired after it is degreased
by the initial heat treatment and the subsequent heat treat-
ment. The volume of the molded body thus fired is reduced
such that the volume of the molded body is ~2.5 % of the vol-
ume of the molded body before it is fired. The packing ratio
of the fired body is about 100 % (bulk density: 3.976).
Until the subsequent heat treatment is completed, the car-
bonized material which has been modified in the initial heat
treatment is completely burned away.
Thereafter, the fired body is subjected to hot iso-
static pressing in an argon atmosphere or an argon atmosphere
which contains 20 vol. % or less of oxygen under conditions
given below. At this time, the fired body is heated at a
~: 213~3~
~ J ~
- 21 -
rate of 200~C/hour. The fired body thus pressed exhibits a
light-transmitting ability.
Processing temperature 1200 - 1250C
(optimum temperature: 1230C)
Processing pressure lO00 ~ 2000 atm
(optimum pressure: 1000 atm)
Processing time 1 ~ 4 hours
(optimum processing time: 2 hours)
The fired body is subjected to hot isostatic press-
ing in the above temperature range and pressure range in or-
der to achieve a desired high light-transmitting ability and
i~prove its mechanical strength to avoid damage during the
hot isostatic pressing. If the hot isostatic pressing were
carried out at a temperature lower than 1200C or under a
pressure lower than 1000 atm, then though the fired body
would be rendered light-transmissive, but the obtained light-
transmitting ability would be low. If the hot isostatic
pressing were carried out at a temperature in excess of
1250C, then abnormal grain growth would be accelerated,
inviting a reduction in the mechanical strength and the
light-transmitting ability. If the hot isostatic pressing
were carried out under a pressure in excess of 2000 atm, then
stresses would concentrate in regions where bores and flaws,
even if extremely small, are located in the fired body, tend-
ing to cause the fired body to crack in those regions.
Thereafter, the ends of the fired body are ground
by a diamond grinding wheel (not shown) to remove edges,
2133839
- 22 -
thereby completing the light-transmissive bulb lF of alumina.
Specifically, as shown in Fig. 1, the light-transmissive bulb
lF with the electrode holding holes la, lb defined in its re-
spective opposite ends is fabricated.
The inner and outer surfaces of the bulb lF thus
produced are then ground by a brush with a diamond grinding
grain having a particle diameter of 0.5 ~m until the bulb lF
will have a wall thickness of 0.2 mm or less. When the inner
and outer surfaces of the bulb lF are thus ground, surface
irregularities are removed from the surfaces of the bulb lF
to prevent light from being scattered by the surfaces of the
bulb lF and improve a liner transmittance thereof.
The bulb lF includes a light-emitting region having
an inside diameter of about 4.0 mm, a wall thickness of about
O.3 mm, an entire length of about 40 mm, and has properties
given below. As a result of a structural observation using a
transmission electron microscope (TEM), no gaps and lattice
defects in the grain boundary phase and crystal grains which
would be responsible for scattering light were found. The
diameter of the electrode holding hole lb is abou~ 1 mm or
less.
Linear transmittance with respect to visible light
having wavelengths ranging from 380 to 760 nm:
70 % or higher
Linear transmittance with respect to light having
having a wavelength of 500 nm:
82 ~ or higher (at a wall
~ 2139~39
- 23 -
thickness: 0.5 mm)
Average particle diameter of crystal grains:
about 0.7 ~m (maximum
- particle diameter:
1.4 ~m)
Mechanical strength (JIS R1601):
Bending strength St
(roo~ temperature) = 98 kg/cm2
(900C) = 81 kg/cm2
Weibull coefficient
(room temperature) = 9.3
(900C) = 8.1
In the measurement of the particle diameter and the
mechanical strength, there was used a specimen (whose shape,
thickness, etc. were according to JIS R1601) fabricated as a
substitute for the bulb lF according to the above embodiment
The specimen was fabricated under the conditions in the above
process.
The particle diameter was calculated by lapping,
with a diamond grinding grain, the surfaces of the specimen
fabricated so that its shape, thickness, etc. were according
to JIS Rl601, subjecting the specimen to grain boundary etch-
ing with dissolved potassium hydroxide, observing the sur-
faces of the specimen with a scanning electron microscope,
and analyzing the image of profiles of crystal grains. In
the image analysis, the crystal grains were assumed to be
spherical or polygonal in shape, and their diameters and the
2139839
, .
- 24 -
maximum value of inter-vertex distances were used to calcu-
late particle diameters.
The linear transmittance was measured by lapping
the opposite surfaces of the fabricated specimen, 0.5 mm
thick, and thereafter determining the linear transmittance
with a double-beam spectrophotometer.
The completed bulb lF made of light-transmissive
alumina has smaller crystal grain diameters than general
light-transmissive ceramics which are produced by firing alu-
mina with a sintering additive of MgO or the like for greater
crystal grains (see Fig. 2).
The bulb lF fabricated from highly-pure alumina has
a light-transmitting ability while having small crystal grain
diameters different from those of general light-transmissive
ceramics for the following reasons:
Since only a small amount of oxide such as MgO or
the like mixed as an impurity (a total of 0.01 mol % or less
at maximum) is contained in the powder of alumina, the impu-
rity forms in its entirety a solid solution with alumina,
producing almost no grain boundary phase. Therefore, the ef-
fect of a grain boundary phase which is responsible for dif-
fusing light in general light-transmissive alumina is elimi-
nated, resulting in an increase in the linear transmittance
with respect to visible light.
Furthermore, the following considerations are taken
into account:
If it is assumed that all the crystal grains and
213983~
crystallites have a circular cross section, then a crystal
grain having a diameter D and made up of n crystallites each
having a diameter d satisfies the following equation ~:
n = (D/d)2
The value of n calculated according to the above
equation can be converted into crystallite boundaries con-
tained in the cross section of one crystal grain.
The lattice constants of various light-transmissive
aluminas obtained from highly pure alumina ~having average
particle diameters of 0.72, 0.85, 0.99, 1.16, 1.35, 1.52 ~)
were determined using an X-ray diffraction apparatus, and the
diameters d of the crystallites of the light-transmissive
aluminas having the above average particle diameters were
calculated from diffraction peaks (012) according to the
Scherrer's equation which relates the diameter d of a crys-
tallite to the width of a diffraction line. As a result, it
was found that the diameters d of the crystallites were con-
stant irrespective of the sizes of the crystal grains. The
Scherrer's equation is given in P. Gallezot, "Catalysis,
Science and Technology", vol. 5 p. 221, Springer-Verlag
(198~), and P. Scherrer, I'Gottinger Nachrichen", 2, 98
t1918).
It can therefore be seen from the above equation
(1) that the smaller the diameters D (average particle diame-
ter) of the crystal grains, the fewer the crystallite bound-
aries in one crystal grain.
Generally, it is considered that when light is ap-
2139839
- 26 -
plied to a polycrystalline material such as of ceramic, the
light is diffused by surfaces where refractive indexes are
not continuous, i.e., regions where the arrangement of atoms
is discontinuous. Since a crystallite boundary in a crystal
grain is nothing but such a region where the arrangement of
atoms is discontinuous, it causes a diffusion of light.
Consequently, the fewer the crystallite boundaries in a crys-
tal grain, i.e., the smaller the diameter D of a crystal
grain, the smaller the effect of the crystallite boundaries
which are responsible for diffusing light, giving rise to a-n
increase in the linear transmittance with respect to visible
light.
The closures 1, lA are manufactured as described
below. A process of manufacturing the closures will be de-
scribed below with reference to ~ig. 3.
First, a vehicle to be used to suspend therein the
fine powder of alumina (secondary aggregate) synthesized as
described above and the fine powder of tungsten is prepared
from various organic materials given in Table 1 below (step
1). To prepare the vehicle, the organic materials are
weighed and uniformly mixed by a mixer.
Table 1
Ingredients Volume ratio
a-terpineol 50
butyl acetate carbitol 20
ethyl cellulose 3
2139839 ~
. . .
polyvinyl butyral 7
ethanol 10
The fine powder of alumina, the prepared vehicle,
an organic solvent (butyl diphthalate), and a dispersing
agent ta~monium carboxylic acid) are mixed at volume ratios
glven in Table 2, below, and kneaded into an alumina slurry
by three rolls (step 2~.
Table 2
Ingredients Volume ratio
fine powder of alumina 64
vehicle 32
butyl diphthalate 3.5
ammonium carboxyl acid 0.5
The fine powder of tungsten, the prepared vehicle,
an organic solvent ~butyl diphthalate), and a dispersing
agent (am~onium carboxylic acid) are mixed at volume ratios
given in Table 3, below, and kneaded into a tungsten slurry
by three rolls (step 2).
T~hle 3
Ingredients Volume ratio
fine powder of tungsten 82
vehicle 15
butyl diphthalate 2.6
ammonium carboxyl acid O.4
Using the alumina slurry prepared at the volume ra-
tios given in Table 2 and the tungsten slurry prepared at the
, .
2` 1 3 ~3 ~ 3 ~
- 28 -
volume ratios given in Table 3, eight slurries composed of
tungsten and alumina mixed at volume ratios (tungsten/ alu-
mina) given in Table 4, below, are prepared (step 3).
Table ~
SlurriesVolume ratio (tungsten/alumina)
1st layer slurry 80~20
2nd layer slurry 60/40
3rd layer slurry 40J60
4th layer slurry 30/70
5th layer slurry 20/80
6th layer slurry 10/90
7th layer slurry S/95
8th layer slurry 3/g7
Each of the mixed slurries thus prepared is suffi-
ciently mixed such that alumina and tungsten are uniformly
dispersed, and thereafter debubbled (step 4). ~ore specifi-
cally, eac~ of the mixed slurries is put in a resin container
in a vacuum desiccator, and air in the vacuum desiccator ls
drawn out by a vacuum pump for a few tens of minutes (e.g.,
about 20 minutes) while the slurry in the resin container is
being stirred by a magnet stirrer or the like. While the
slurry is being debubbled in vacuum, the organic solvent is
partly volatilized to achieve a slurry viscosity of 30,000
cP .
Then, the mixed slurries shown in Table 4 are con-
centrically deposited to a predetermined thickness on the
2133~39
- 29 -
outer circumferential surface of each of the support shafts 4
supporting the main electrodes 3, which serves as cores of
the closures. The mixed slurries shown in Table 4 are ap-
plied in a descending order of volume ratios of tungsten,
i.e , from the first layer slurry to the eighth layer slurry.
A laminated body 20 as a precursor of each of the closures 2,
2A is thus formed around the support shafts 4 as shown in
~ig. 4 (step 5) The mixed slurries are applied to and de-
posited on the outer circumferential surface of each of the
support shafts 4 in the order from the first layer slurry to
the eighth layer slurry by coating and drying each of the
slurries su~cessively from the first layer slurry.
In this manner, an innermost layer composed of the
first layer slurry is formed in a core-side region of the
closure which is located adjacent to the core, a plurality of
intermediate layers composed of the second through seventh
layer slurries are formed in an intermediate region of the
closure, and an outermost layer composed of the eighth layer
slurry is formed in a bulb-side region of the closure which
is located adjacent to the open end of the bulb.
Figs. 5~a), S(b~, and 5(c) are a cross-sectional
view showing the structure of the closure and diagrams show-
ing the relationship between volume ratios of tungsten and
alumina in each of the layer slurries of the closure. As
shown in Figs. S(a) through 5(c), the closure 20 is of such
composition distributions that the volume ratio of alumina
increases up to about 100 % outwardly from the support shaft
- ~13983~
- 30 -
4 as shown in Fig. 5(c), and the volume ratio of tungsten de-
creases from 80 ~ outwardly from the support shaft 4 as shown
in Fig. 5~
Then, the laminated body 20 is heated to 600~C for
10 hours in a moisture-containing hydrogen reducing at~o-
sphere, so that the laminated body 20 is degreased (step 6).
Specifically, when the laminated body 20 is heated, the or-
ganic materials and organic solvent which are contained in
the vehicle that were added when the slurries were prepared
are thermally decomposed and carbonized, thereby degreasing
the formed body.
The degreased laminated body 20 is subsequently
heated to 1800C for 2 hours in a vacuum atmosphere, so that
the laminated body 20 (degreased body) is fired (step 7).
Each of the closures 2, 2A is now obtained as the fired la~i-
nated body 20. Until this subsequent heat treatment is com-
pleted, the carbonized materials modified in the above ini-
tial heat treatment are fully burned away.
In each of the layers of the closures 2, 2A, a net-
work structure of crystals is formed between common ingredi-
ents by firing, thereby integrally joining the ingredients.
A firing process for reducing surface energy is applied to
the joining of the support shafts 4 and the surfaces of the
electrode holding holes la, lb of the bulb lF to each other.
Impurities such as of glass are often added in a small amount
in an effort to accelerate the firing process.
More specifically, in the firing process, the alu-
~ ~139839
mina forms a solid solution and is crystallized, trapping thepowder of tungsten, in each layer of the laminated body 20.
Ad~acent layers of the laminated body 20 are integrally
joined to each other in solid phase as the alumina in the
layers forms a solid solution and is ~rystallized at the mat-
ing surfaces of the layers. The support shaft 4 and the in-
nermost layer composed of the first layer slurry are also in-
tegrally joined to each other in solid phase because alumina
in the innermost layer is crystallized in contact with the
support shaft 4, forming a glassy substance in its grain
boundaries, and also because tungsten is contained in both
the support shaft 4 and the innermost layer. As a result,
the fired ~losures 2, 2A are strongly bonded to the support
shafts 4 which support the main electrodes 3, hermetically
sealing and securing the support shafts 4 and hence the main
electrodes 3 in the bulb 1.
The distribution of coefficients of thermal expan-
sion from the support shaft 4 through the innermost layer and
the intermediate layers to the outermost layer is a gradient
distribution ranging from the coefficient of thermal expan-
sion of the support shaft 4 (the coefficient of thermal ex-
pansion of tungsten) to a coefficient of thermal expansion
which is close to the coefficient of thermal expansion of the
bulb lF (the coefficient of thermal expansion of alumina),
based on the composition distributions thereof.
After the support shafts 4 have been sealed and se-
cured, the outer circumferential surfaces of the outermost
- 2139~33
layers of the closures 2, 2A are cut or ground so as to fit
in the electrode holding holes la, lb in the bulb lF (step
8). The closures are now completed, and the manufacturing
process is ended.
Assembling the completed closures 2, 2A into the
bulb lF and fabrication of the liaht-emitting bulb assembly 1
will be described below.
First, as shown in Fig. 1, the closure 2A
(identical to that shown in Figs. 4 and 5) which has been
fired and machined on its outer circu~ferential surface is
fitted in the electrode holding hole lb in the bulb lF,
bringing the outer circumferential surface of the closure 2A
into contact with the inner circumferential surface of the
electrode holding hole lb. Thereafter, an infrared radiation
or high-output laser beam is locally applied to the contact-
ing surfaces to heat them.
The localized heating causes the alumina in the
outermost layer composed of the eighth layer slurry of the
closure 2A and the alumina in the bulb lF to be fired and
crystallized, and also causes grain boundaries in the joined
surfaces to be embedded by a glass phase that is primarily of
a structure of spinel, garnet or the like. The closure 2A
and the bulb lF are therefore joined in solid phase to each
other. As a consequence, the closure 2A and the bulb lF are
hermetically secured to each other by the formation of a
glass phase in the grain boundaries of alumina in the outer-
most layer and the bulb lF.
'~139839
Similarly, the closure 2 (see Figs. 4 and 5) which
has been fired and machined on its outer circumferential sur-
face is fitted in the electrode holding hole la in the bulb
lF, and an infrared radiation or high-output laser beam is
local}y applied to the contacting surfaces to heat them. The
closure 2A a~d the bulb lF are integrally joined in solid
phase to each other. The bulb lF is now ready for being
filled with a starting rare gas metal and a discharging mate-
rial.
Then, an amalgam of a given starting rare gas metal
and a discharging material (an alloy of Sn, Na-T1-In, Se-Na,
Dy-T1, or a halide of each of the metals) is introduced
through the slender introduction tube lc into the bulb lF
whose ends have been sealed, and thereafter the slender in-
troduction tube lc is sealed by the sealant ld.
Since the closures 2, 2A and the bulb lF are inte-
grally joined in solid phase to each other without use of
soldering glass which has heretofore been relied upon, the
materials which have been sealed in the bulb lF are reliably
prevented from leaking out.
The bulb lF with the main electrodes mounted
therein are generally incorporated in an outer tube of a
high-pressure discharge lamp such as a metal halide lamp or
the like.
Light-emitting bulb assemblies (inventive examples)
in which the volume ratios of tungsten in the innermost layer
or the volume ratios of alumina in the outermost layer of the
~ 1 3 9 ~ 3 ~
- 34 -
closure 2 according to the first embodiment are of various
values which fall in the range according to the present in-
vention, light-emitting bulb assemblies (comparative exam-
ples) in which these volume ratios are of values which fall
out of the range according to the present invention, and
light-emitting bulb assemblies (conventional examples) in
which the closure of alumina is fixed to the bulb by alumina
cermet will be compared with each other. Results of the com-
parison are given in Tables 5 and 6 below. ~ach of the
light-emitting bulb assemblies has a bulb which is identical
to the bulb according to the first embodiment of the present
invention. The closures have various numbers of layers in-
cluding innermost, outermost, and intermediate layers. The
volume ratios of alumina and tungsten from the innermost
layer through the intermediate layers to the outermost layer
are of distributions having increasing and decreasing gradi-
ents.
The durability of the light-emitting bulb assem-
blies was evaluated according to an accumulation of energiza-
tion periods (energization service life) by applying repeat-
edly turning them on for 5 hours and turning them off for 0.5
hour for thereby developing thermal stresses in the light-
emitting bulb assemblies. Each of the light-emitting bulb
assemblies was turned on by a voltage of 100 V (100 W) ap-
plied between the main electrodes 3 across a discharging ma-
terial of Hg - TlI3 ~0.11 g) sealed in the bulb. Since the
stably energized state becomes greatly unstable in the event
- 2 1 3 9 8 3 9 :
-- 35 --
of a leakage of the sealed materials, the accumulation of en-
ergization periods was interrupted at the time the energized
state became unstable.
Table 5
Speci- Type Tungsten/alumina Nwl-ber of Energiza-
men volume ratio layers tion ser-
No.
Innermost Outermost vice life
layer layer
Inventive 55/45 3/97 7 3500
2 Inventive 65/35 3/97 8 4300
3 Inventive 75/25 3/97 9 5200
4 Inventive 85/15 3/97 10 8000
Comparative~5/65 3/97 4 *l
6 Comparative45/55 35/65 3 3000
7 Conventional - - - 3000
*l ~Unable to measure due to a conduction failure.
Similarly, the light-emitting bulbs in which a dis-
charging material of Hg - TlI - NaI - InI3 (0.13g) was sealed
were also compared. Results of the comparison are given in
Table 6 below.
Table 6
Speci- Type Tungsten/alumina Number of Energiza-
men volume ratio layers tion ser-
No.
21~839
- 3~ -
Innermost Outermost vice life
layer layer
1 Inventive 55/95 3/97 7 3400
2 Inventive 65/35 ~/97 8 3800
3 Inventive 75/25 ~/97 9 4300
4 Inventive 85/15 3/97 10 5000
Comparative35/65 3/97 4 ~2
6 Comparative45/55 35~65 3 3000
7 Conventional - - - 3000
*2 Unable to measure due to a conduction failure.
It can be seen from the above test results that the
light-emitting bulb assembly according to the present inven-
tion has very high durabllity even when repeatedly turned on
and off. The light-emitting bulb assembly according to the
present invention has increased resistance against thermal
stresses because the closures 2, 2A are joined in solid phase
which have a gradient coefficient of thermal expansion that
is closer to the coefficient of thermal expansion of either
the support shafts 4 with the main electrodes 3 on their dis-
tal ends or the bulb lF toward the support shafts 4 and the
~ulb lF. Because of such increased resistance against ther-
mal stresses, the light-emitting bulb assembly is capable of
highly reliable light emission and has a long service life.
The light-emitting bulb assembly can also be made available
with ease.
The light-emitting bulb assemblies according to the
2 1 :~ 3 8 3 ~
- 37 -
inventive examples with the discharging material of Hg - TlI3
(0.11 g) sealed in the bulb had a luminance of 183,000 nt,
and the light-emitting bulb assemblies according to the in-
ventive examples with the discharging material of ~g - TlI -
NaI - InI3 (0.13g) sealed in the bulb had a luminance of
290,000 nt.
Since the bulb lF according to this embodiment is
made of light-transmissive alumina composed of small crystal
grains having an average particle diameter of about 0.7 ~m
and a maximum particle diameter of about 1.4 ~m and does not
form any grain boundary phase, the mechanical strength
(bending strength, Weibull coefficient) in a range from room
temperature to a temperature upon di~chargin~ is higher than
a general bulb assembly of light-transmissive ceramics which
are produced by firing alumina with a sintering additive of
MgO or the like for greater crystal grains. As a result, the
light-emitting bulb assembly with the bulb lF according to
the present embodiment has a reduced wall thickness as well
as an increased service life. Inasmuch as the reduced wall
thickness lowers the thermal capacity of the light-emi~ting
bulb assembly, allowing the light-emitting bulb assembly to
be heated quickly to a desired temperature, the starting time
required for the discharging metal component to be evaporated
up to a saturated vapor pressure until energization of the
bulb assembly becomes stable is shortened.
Inasmuch as no grain boundary phase is formed and
crystallite boundaries in crystal grains which are responsi-
213g839
- 38 -
ble for diffused light are reduced based on small grain diam-
eters, the diffusion of liqht caused while the light passes
through the wall of the bulb lF is suppressed, and the bulb
lF has high linear transmittance of 70 % or more with respect
to light (visible light) having a wavelength ranging from 380
to 760 nm (linear transmittance with respect to liqht having
a wavelength of 500 nm: 82 %, thickness: 0.5 mm). Therefore,
a high-pressure discharge lamp having the light-emitting bulb
assembly 1 with the bulb lF has increased luminance.
In addition, since there exists no grain boundary
phase unlike the conventional bulb, any erosion of grain
boundaries with discharging metal vapor components (ions) is
suppressed, thereby preventing the discharging metal vapor
components from leaking out of the bulb even though the bulb
has a reduced wall thickness. Therefore, the highly luminous
discharge lamp can have an increased service life as the dis-
charging metal vapor components are prevented from leaking
out of the bulb wall even though the bulb wall has a reduced
wall thickness. With the light-emitting bulb assembly 1 ac-
cording to this embodiment, the electrode holding hole lb is
of a small diameter to reduce the amount of the sealant used
for thereby suppressing any erosion of the sealant with the
discharging metal vapor components (ions), 50 that any leak-
age of the discharging metal vapor components is avoided more
reliably.
A second e~bodiment of the present invention will
be described below. Closures of a light-emitting bulb assem-
213~839
- 39 -
bly according to the second embodiment are different as to a
process of manufacturing them and their structure from the
closures of the light-emitting bulb assembly according to the
second embodiment. The different process and structure will
be described below. Components according to the second em-
bodiment are denoted by re~erence numerals which are identi-
cal to those of the components according to the first embodi-
ment, with a suffix "a".
The materials of the closure 2a (see Fig. 14) ac-
cording to the second embodiment are also a fine powder of
highly pure alumina synthesized by drying an aqueous solution
of suspended aluminum salt according to a spray drying pro-
cess and then thermally decomposing the aluminum salt, and a
fine powder of highly pure tungsten.
A process of manufacturing the closure 2a according
to the second embodiment will be described below with refer-
ence to ~ig. 6.
As shown in Fig. 6, eleven slurries with the fol-
lowing volume ratios of tungsten and alumina (tungsten/ alu-
mina) are prepared from a fine powder of alumina and a fine
powder of tungsten (step 1):
1st slurry: tungsten/alumina = lOOJ0
2nd slurry: tungsten/alumina = 90/10
3rd slurry: tungsten/alumina = 80/20
4th slurry: tungsten/alumina = 70/30
5th slurry: tungsten/alumina = 60/40
6th slurry: tungsten/alumina = 50/50
2139839~
- 40 -
7th slurry: tungsten/alumina = 40/60
8th slurry: tungsten~alumina = 30/70
9th slurry: tungsten/alumina = 20/80
10th slurry: tungstenJalumina = 10/90
11th slurry: tungsten/alumina = 0/100
The above slurries are prepared as follows: First,
the fine powder of alumina and the fine powder of tungsten
are weighed such ~hat their volume ratios are of the above
~numerical values, and a dispersing agent of ammonium car-
boxylic acid and distilled water are added to the weighed
powders. They are then mixed with each other in a wet manner
by a ceramic (alumina) ball mill for about 24 hours, so that
the fine powders of alumina and tungsten are uniformly pre-
sent in the solvent while breaking up excessive aggregates.
The ratio (volume ratio) at which the dispersing
agent of ammonium carboxylic acid is added to the fine pow-
ders in each of the slurries is 2 g with respect to 100 g of
the total fine powders in each of the slurries.
Then, each of the slurries is debubbled ~step 2).
Specifically, each of the slurries taken from the ball mill
is put in a resin container in a vacuum desiccator, and air
in the vacuum desiccator is drawn out by a vacuum pump for a
few tens of minutes (e.g., about 20 minutes) while the slurry
in the resin container is being stirred by a magnet stirrer
or the like.
Thereafter, a desired molded body 20a shown in Fig.
7 is produced using a mating mold assembly 10 shown in Fig.
~ 2 1 ~ 9 8 3 9
- 41 -
8(a) according to a process described below. The ratio of
vertical and horizontal dimensions of the molded body 20a and
the closure 2a shown in Figs. 7 and lO(a), lO(b~ is not 1 : 1
for illustrative purpose.
The mating mold assembly 10 comprises a pair of
symmetric molds lla, llb each made of a porous inorganic ma-
terial such as plaster or the like or a porous resin with
minute pores which has substantially the same function as
plaster. The molds lla, llb are joined to each other, defin-
ing a slurry pouring space 13 between mating surfaces of the
molds lla, llb as shown in Fig. 8ta).
As shown in Fig. 8(b), the molds lla, llb have re-
spective grooves (cavities) 13a, 13b defined in the respec-
tive mating surfaces 15a, 15b and curved in the vicinity of
lower mold ends. The grooves 13a, 13b are cut in the respec^
tive mating surfaces 15a, 15b by an end mill having a spheri-
cal cutter on its distal end. Alternatively, the grooves
13a, 13b may initially be formed in the respective mating
surfaces l5a, 15b.
Then, the debubbled slurries are poured in a de-
scending order of contents of alumina, i.e., from the
eleventh slurry to the first layer slurry, into the slurry
pouring space 13 of the mating mold assembly 10 (step 3).
Specifically, as shown in Fig. 9, a cylindrical
member 17 is placed on the upper surface of the mating mold
assembly 10, and the eleventh slurry, which is of an amount
greater than the volume of the slurry pouring space 13, is
- - 2139839
- 92 -
poured into the cylindrical mem~er 17. An annular piece of
clay 19 is applied to the lower end of the cylindrical member
17 to provide a seal between the lower surface of the cylin-
drical member 17 and the upper surface of the mating mold as-
sembly 10. The clay may be replaced with rubber.
~ fter the eleventh slurry has ~een poured into the
slurry pouring space 13, the poured eleventh slurry is left
for a predetermined period of time. During this time, the
solvent (distilled water) of the eleventh slurry is drawn
into the pores of the porous molds lla, llb by capillary ac-
tion. Accordingly, a powder ~alumina powder in the eleventh
slurry) bounded by the dispersing agent of ammonium car-
boxylic acid is uniformly deposited on the wall surface of
the slurry pouring surface 13, forming a thin layer llS
thereon as shown in Figs. lO(a) and lO(b).
The period of time during which the poured eleventh
slurry is left after the eleventh slurry has been poured into
the slurry pouring space 13 governs the thickness of the thin
layer llS. The period of time during which the poured
eleventh slurry is left is experimentally determined so that
the formed thin layer llS has a predetermined value. The pe-
riod of time durlng whlch the poured eleventh slurry is left
and t~e slurry pouring space 13 are determined also in view
of volume shrinkage after firing. The period of tlme during
which the poured eleventh slurry is left according to this
embodiment is adjusted so that the formed thin layer llS has
a predetermined value
21~9839
, .
.
- 43 -
While the poured eleventh slurry is ~eing left, a
negative pressure may be maintained outside of the mo~ds for
forcibly drawing the solvent of the slurry out of the molds.
This allows the poured eleventh slurry to be left for a
shorter period of time, permits the slurry to be directly de-
bubbled through the molds, and also makes it possible to in-
crease the filling ratio by strongly drawing the solvent.
~ fter the poured eleventh slurry has ~een left for
the predetermined period of time, the eleventh slurry remain-
ing inside the cylindrical member 17 and on the inner surface
of the thin layer llS is discharged. Then, the tenth slurry
is poured, left for a predetermined period of time, and dis-
charged. Thereafter, the ninth through ~irst slurries are
also poured, left for a predetermined period of time, and
discharged. After the eleventh through first slurries are
repeatedly poured, left for a predetermined ~eriod of time,
and discharged, the powders in the slurries (the power of
alumina alone, the powder of mixed alumina and tungsten, and
the powder of tunqsten alone) are uniformly deposited in lay-
ers, forming thin layers llS, lOS, 9S, , lS successively
on the wall surface of the slurry pouring space 13. These
thin layers llS, lOS, 9S, , lS jointly form a molded body
20a as a precursor of the closure 2a.
Figs. 12(a) and 12(b) are diagrams showing the re-
lationship between volume ratios of tungsten and alumina in
each of the thin layers. As shown in Figs. 12(a) and 12(b),
the molded body 20a is of such composition distributions that
2 1 3 9 8 3 9
- 44 -
the volume ratio of alumina increases from O % up to 100 %
from the central thin layer lS toward the outer thin layers
as shown in Fig. 12(b), and the volume ratio of tungsten de-
creases from 100 % to O % from the central thin layer lS to-
ward the outer thin layers as shown in ~ig. 12(a). The thin
layer 2S in the molded body 20a corresponds to the innermost
layer (or the core-side layer) of the laminated body 20 ac-
cording to the preceding embodiment, the thin layer llS cor-
responds to the outermost layer (or the bulb-side layer) of
the laminated body 20 according to the preceding embodiment,
and the thin layers 3S ~^lOS correspond to the intermediate
layers of the laminated body 20 according to the preceding
embodiment. The thin-layers 2S - lOS are disposed around and
covers the central layer lS.
When the cycles of pouring, leaving for a predeter-
mined period of time, and discharging the eleventh through
first slurries are completed, the mating mold assembly 10 is
separated, releasing the molded body 20a shaped as shown in
Fig. 7. The molded body 20a is dried until the solvent is
thoroughly removed therefrom (step q)
Thereafter, the molded body 20a is heated to 600C
for 10 hours in a moisture-containing hydrogen reducing atmo-
sphere, so that the molded body 20a is degreased and tem-
porarily fired tstép 5). Specifically, when the molded body
20a is heated, the dispersing agent which was added when the
slurries were prepared is thermally decomposed, thereby de-
greasing the molded body 20a.
2139839
- ~5 -
Then, as shown in Fig. 13, support holding holes
21a, 21b are defined respectively in the opposite ends of the
molded body 20a, and a support shaft 4 which supports a main
electrode 3 is fitted in the support holding hole 21a that is
defined in the distal end of the central layer lS, and a
shaft 5 of tungsten is fitted in the support holding hole
21b, thereby setting the main electrode 3 (step 6).
The molded body 20a with the main electrode 3 set
is subsequently heated to 1500C for 2 hours in a vacuum at-
mospher~, so that the molded body 20a is fired (step 7). The
closure 2A is now obtained as the fired molded body 20a.
Until this subsequent heat treatment is completed, the car-
bonized materials modified when the molded body is degreased
are fully burned away.
In the firing process, the thin layers of the
molded body 20a are integrally joined in solid phase as with
the laminated body 20 according to the preceding embodiment.
The support shaft 4, the shaft 5, and the thin layer lS are
also integrally joined in solid phase by volume shrinkage
upon firing and coexistence of tungsten. As a result, the
fired closure 2a is firmly bonded to the support shaft 4
which supports the main electrode 3 and the shaft 5, hermeti-
cally sealing and securing the support shaft 4 and the main
electrode 3. The closure 2a is now completed, and the pro-
cess of manufacturing same is completed in its entirety.
The outside diameter of the fired closure 2a is de-
termined by the diameter of the slurry pouring space 13 which
~ 1 3 9 8 3 9
- 46 -
takes into account volume shrinkage upon firing Therefore,
the fired closure 2a is not required to be machined at its
outer circumferential surface.
The distribution of coefficients of thermal expan-
sion from the support shaft 4 through the thin layers 2S
through 9S to the thin layer lOS is a gradient distribution
ranging from the coefficient of thermal expansion of the sup-
port shaft 4 (the coefficient of thermal expansion of tung-
sten) to the coefficient of thermal expansion of the bulb lF
(the coeficient of thermal expansion of alumina), based on
the composition distributions thereof.
~ s shown in Fig. 14, the completed closure 2a is
fitted in the electrode holding hole la in the bulb lF, and
then an infrared radiation or high-output laser beam is lo-
cally applied to the contacting surfaces of the closure 2a
and the bulb lF to heat them.
The localized heating causes the alumina in the
thin layer lOS of the closure 2a and the alumina in the bulb
lF to form a glass phase in the grain boundaries in the
joined surfaces. The closure 2a and the bulb lF are there-
fore joined in solid phase to each other. As a consequence,
the closure 2a and the bulb lF are hermetically secured to
each other. Then, a starting rare gas metal and a discharq-
ing material are filled in the bulb lF. The light-emitting
bulb assemb~y shown in Fig. 14 is now completed.
The light-emitting bulb assembly with the closure
2a was also measured for its energization service llfe when
2 1 3 3 8 3 5 -`
- 47 -
repeatedly turned on and off. As a result, it was found that
the light-emitting bulb assembly with the closure 2a also had
very high durability as with the light-emitting bulb assembly
with the closure 2. The light-emitting bulb assembly with
the closure 2a has increased resistance against thermal
~tresses because the closure 2a has a gradient coefficient of
thermal expansion that is closer to the coefficient of ther-
mal expansion of either the support shaft 4 having the main
electrode 3 or the bulb lF toward the support shaft 4 and the
bulb lF. Because of such increased resistance against ther-
mal stresses, the light-emitting bulb assembly is capable of
highly reliable light emission and has a long service life.
The light-emitting bulb assembly can also be made available
with ease.
The light-emitting bulb assembly with the closure
2a also offers the following advantages:
Since the volume ratio of alumina is 100 % in the
thin layer llS which is exposed in the bulb lF in supporting
the main electrode 3 in the bulb lF, i.e., the thin layer llS
is an insulation, back arcs from the main electrode 3 can be
avoided for more stable energization of the light-emitting
bulb assembly.
Because the main electrode 3 and the shaft 5 which
serves as an external terminal are hermetically sealed by the
thin layer ~central layer) lS whose volume rat~o of tung~ten
is 100 %, a desired voltage can be applied to the main elec-
trode 3 without fail.
- 2139839
.
- i8 -
In additlon, a~ the thln layers are formed by pour-
ing ~lurries, it is possible to uniformize the thic~nesses of
the thin layers for reliably maintaining compos$tion distri-
butions in the layer~ and a gradient distribution of coeffi-
cient~ of thermal expansion.
While the two embodiments of the present invention
have been described above, the present invention i~ not lim-
ited to the~e embodiments, but various changes and modif~ca-
tions may be made therein without departing from the scope of
the present invention.
- The materials of the bulb lF, the closure 2, and
the closure 2a include a fine powder of alumina whoce purity
is 99.99 mol % or higher in the above embodiments. However,
insofar as the bulb lF has practlcal llnear transmittance
~linear transmittance with respect to light having a wave-
le~gth ranging from 380 to 760 nm), the material ls not lim-
ited to such a fine powder of alumina.
For example, the bulb lF may be in the form of a
fired body composed primarily of an oxide such as ~lumina,
magnesia, zirconia, yttria or silica and a nitride such as
aluminum nitride, with a compound (sintering additlve) added
for suppressing abnormal grain growth and accelerating fir-
ing. The closures 2, 2a may be fabricated using the same
fine powder of ceramic a~ ~he bulb lF thus produced.
Specifically, the bul~ lF may be made of a fine powder of
Alllmin~ having a purity of 99.2 m~l % and an av~erage particle ~ ter
rAn~in~ frQ~ 0.3 to 1.0 ~m, and the ~los~lr~s 2, 2a may be made of
2~1 3~g~8~3-9~
such a fine powder of alumina and a fine powder of tungsten.
While the materials of the closures 2, 2a i~clude a
fine powder of tungsten in the above embodiment~, the ~ateri-
als of the closures 2, 2a may ~e modified depending on the
material of the support shaft 4 whlch serves as a core. For
example, if the support shaft 4 i~ made of niobium or
molybdenum, then the materials of the closures 2, 2a may
include a fine powder of niobium or molybdenum.
The bulb may be of any of various shapes. ~or ex-
ample, rather than having the larger-diametèr electrode hold-
ing hole ~a and the smaller-diameter electrode holding hole
lb which are defined respectively in the oppo~ite ends of the
bulb lF, the bulb may be of a cylindr~cal shape with its both
ends being simply open or may be a curved bulb.
In the fabrication process according to the first
embodiment, each of the mixed slurries i~ coated and dried in
forming the la~lnated body 20 around the support shaf t 4 of
tungsten which supports the main electrode 3. However, green
sheets may be produced from the respective mixed slurries,
and successively wou~d around the support shaft 4 in a de-
scending order of volume ratios of tungsten. In this case,
it is-preferable to stack the green sheets such that the
joined surfaces of the green sheets are alternately staggered
180 around the support ~haft.
In ~oining the closures 2, 2a and the bulb lF to
each other in solid p~ase, the contacting ~urfaces are lo-
cally heated. However, they may be heated in the vicinity of
- ~ ~ 21~3~98~9
- 50 -
the support shaft 4. Even when they are heated in the vicin-
ity of the support shaft 4, since the applied thermal energy
is transmitted to the outermost layers of the closures 2, 2a,
the closures 2, 2a and the bulb lF can be joined to each
other in solid phase. The closures 2, 2a may be fired while
the degreased closures 2, 2a are being assembled in the bulb
lF.
The closure 2 is assembled in the bulb lF by being
fitted in the electrode holding hole la. Instead, as shown
in Fig. 15, the closure 2 ~ay be held against an open end of
the bulb lF to bring the end of the bulb lF into contact with
the side of the outermost layer of the closure 2, and the
contacting surfaces may be locally heated to join the closure
2 and the bulb lF to each other in solid phase at their ends.
The gradient of the volume ratios of alumina and
tungsten in the mixed slurries is not limited to the values
indicated in the above embodiments, but may be of any of var-
ious other values.
The closure 2 may be made of a gradient function
material whose compositional proportions vary linearly from
the core toward the bulb.
The first and second embodiments described above
offer the following advantages:
In the light-emitting bulb assemblies according to
the first and second embodiments, the closure joined in solid
phase to the opening of the bulb which is made of light-
transmissive ceramic comprises a multilayer laminated body,
~ ~ 2 1 3 9 ~ 3 g
' -- .
- 51 -
and the distribution of coefficients of thermal expansion
from the innermost layer near the central conductive core to-
ward the outermost layer near the bulb is a gradient distri-
bution ranging from the coefficient of thermal expansion of
the conductive core toward the coefficient of thermal expan-
sion of the bulb based on the gradient of composition ratios
of the layers.
Therefore, the compositions of the layers may be of
a gradient pattern, and the layers, and the closure and the
bulb may be firmly hermetically joined to each other in solid
phase.
Based on the gradient distribution of the coeffi-
cients of thermal expansion, the concentration of thermal
stresses produced upon energization of the bulb assembly can
be reduced to avoid cracks in the solid-phase joints. As a
result, the materials sealed in the bulb assembly are pre-
vented from leaking out, so that the bulb assembly is capable
of highly reliable light emission and has a prolonged service
life.
The light-emitting bulb assemblies according to the
above embodiments have a bulb made of light-transmissive alu-
mina having an average particle diameter of 1 ~m or less and
a m~x;mll~ particle diameter of 2 ~m or less. Consequently,
the mechanical strength of the light-emitting bulb assemblies
ranging from normal temperature to a discharging temperature
is higher than that of the conventional light-emitting bulb
assemblies. Therefore, the wall thickness of the light-emit-
21398~9
- 52 -
ting bulb assemblies can be reduced to 0.2 mm or smaller,
which is about 1/3 of that of the conventional light-emitting
bulb assemblies.
Since almost no grain boundary phase such as a
spinel phase is formed and crystallite boundaries in crystal
grains which are responsible for diffusing light are reduced
based on the small particle diameter, diffusion of light
while the light is passing through the wall of the bulb is
suppressed, thus providing high linear transmittance with re-
spect to light (visible light) having a wavelength ranging
from 380 to 760 nm. Consequently, the amount of ~ight trans-
mitted from a high-luminance discharge light-emitting bulb
assembly is made greater than that from a conventional light-
emitting bulb assembly, and hence the luminance of a high-
pressure discharge lamp which employs a high-luminance dis-
charge light-emitting bulb assembly is increased. That is,
the amount of light transmitted from a high-luminance dis-
charge light-emitting bulb assembly at the time light is ap-
plied to the high-luminance discharge light-emitting bulb as-
sembly is made substantially equal to the amG-~nt of light ap-
plied to the high-l1lmin~nce discharge light-emitting bulb as-
sembly by suppressing diffusion of light. The luminance can
further be increased by thinning out the wall of the bulb.
Inasmuch as the closure is fired and fabricated of
highly pure alumina, the mechanical strength of ~he closure
is increased, and the durability of the light-emitting bulb
assembly as a whole is also increased.
~13~839
~ ~ .
- 53 -
According to the processes of manufacturing the
light-emitting bulb assemblies according to the first and
second embodiments, a plurality of suspensions with different
volume ratios are prepared, a laminated closure having a gra-
dient distribution of coefficients of thermal expansion is
fabricated using the ~repared suspensions, and the closure
and a bulb are firmly hermetically joined in solid phase to
each other. Thus, a light-emitting bulb assembly which is
highly reliable and has a long service life can easily be
manufactured. A laminated closure having a gradient distri-
bution of coefficients of thermal expansion may separately be
fired and fabricated, and joined in solid phase to a bulb.
According to the process of manufacturing the
light-emitting bulb assembly according to the first embodi-
ment, layers are successively stacked in a descending order
of volume ratios of a conductive component by a simple pro-
cess of coating the layers or the like, for thereby easily
producing an unfired laminated body which is a precursor of a
laminated closure having a gradient distribution of coeffi-
cients of thermal expansion.
The suspensions with different volume ratios are
formed into respective green sheets, and layers are succes-
sively stacked in a descending order of volume ratios of a
conductive component (or a core) by a simple process of wind-
ing the green sheets, for thereby easily producing an unfired
laminated body which is a precursor of a laminated closure
having a gradient distribution of coefficients of thermal ex-
- 2139839
- 54 -
pansion.
According to the process of manufacturing the
light-emitting bulb assembly according to the second embodi-
ment, thin layers are successively stacked in an order of
volume ratios of a conductive component (or a core) by re-
peating a simple process of pouring a suspension into a
porous mold assembly, causing the solvent to penetrate into
the mold assembly, and discharging the excessive suspension,
for therehy easily producing an unfired laminated body which
is a precursor of a laminated closure having a gradient dis-
tribution of coefficients of thermal expansion. The thick-
nesses of the thin layers can be uniformized for reliably
maintaining composition distributions in the layers and a
gradient distribution of coefficients of thermal expansion.
The central layer capable of being connected to an
external source is formed of the conductive component within
the innermost layer of the closure, and a given voltage can
be applied without fail through the central layer to the main
electrode.
A sealing structure of a light-emitting bulb assem-
bly according to a third embodiment of the present invention
and a method of manufacturing such a sealing structure will
be described below with reference to Figs. 16 through 19.
Fig. 16 is a cross-sectional view of a light-emit-
ting bulb assembly according to the third embodiment of the
present invention, particularly showing in detail a sealing
structure of a bulb incorporated in an outer tube of a ~etal
~13383g
- 55 -
vapor discharge lamp.
A bulb 301 has openings 302 defined respectively in
its opposite ends. End caps 303 as closures are integrally
attached to the respective open ends 302, and electrode rods
304 as cores of the closures extend through and are held by
the end caps 303, respectively.
The bulb 301 is made of light-transmissive poly-
crystalline alumina, and the electrode rods 304 are made of a
tungsten-base material of W/Th or the like which is highly
resistant to light-emitting substances. Each of the elec-
trode rods 304 has an externally threaded portion 305
threaded in the corresponding end cap 303 and a flange 306
held against an outer end surface of the end cap 303. The
flange 306 has an outer surface sealed by a sealant 307 such
as of platinum solder or glass, and one of the electrode rods
30~ has a hole 308 defined therein for introducing amalgam.
Each of the end caps 303 is of a multilayer struc-
ture as with the above embodiments. More specifically, each
of the end caps 303 is composed of a plurality of layers
303l, 3032, , 303n arranged along the axial direction of
the bulb 1. The layer 3031 (the bulb-side region layer)
joined to the open end 302 of the bulb 301 has a coefficient
of thermal expansion which is substantially the same as that
of the light-transmissive alumina of which the bulb 301 is
made. The outermost layer 303n (the core-side region layer)
has an internally threaded surface 309 in which the exter-
nally threaded portion 305 of the electrode rod 304 is
- ~ 2 1 3 9 8 3 ~
.
- 56 -
threaded. The outermost layer 30~n has a coefficient of
thermal expansion which is substantially the same as that of
the electrode rod 304. The compositional proportions of the
intermediate layers 3032, -, 303n_1 (intermediate region
layers) interposed between the layers 3031, ~03n are adjusted
such that the intermediate layers 3032, , 303n_l have re-
spective coefficients of thermal expansion varying gradually
from that of the innermost layer 303l toward that of the out-
ermost layer 303n.
The thicknesses of the respective layers increase
progressively from the innermost layer 3031 toward the outer-
most layer 303n~ This is effective to reducing stresses that
are developed when the layers are thermally expanded.
A tapered gap 310 is defined between the electrode
rod 304 and the layers 3031, , 303n_l except the outermost
layer 303n~ The tapered gap 310 prevents the layers 3031,
, 303n_l from contacting the electrode rod 304 when the
lamp is assembled.
A process of manufacturing the light-emitting bulb
assembly of the above structure for a metal vapor discharge
lamp will be described below with reference to Figs. 17 and
18(a) through 18(e).
First, slips for fabricating the end caps 303 are
prepared. To prepare such slips, as many containers C1 Cn
as the number (n) of layers of each of the end caps 303 are
employed as shown in Fig. 17. Material powders are weighed
for obtaining desired coefficients of therma~ expansion, and
2 1 3 9 8 3 9
- 57 -
distilled water, a commercially available dispersing agent
and a binder are added to the weighed material powders. They
are then uniformly mixed for 24 hours by a ball mill, thereby
producing slips Sl -- Sn respectively in the containers C
Cn
Table 7, given below, shows compositional propor-
tions of material powders of respective slips for an end cap
303 which is composed of a total of eleven layers. In Table
7, the compositional proportions are represented by weight %,
and the slip No. corresponds to the number of a layer of the
end cap 303.
Tab e 7
Slip No. Al23 W Ni
100 0 0
2 90 9
3 80 18 2
4 70 27 3
36 4
6 50 45 5
7 40 54 6
8 30 63 7
9 20 72 8
81 9
11 0 90 10
Then! as shown in Fig. 18(a), a tubular mold 312 is
set on a porous plate or plaster board 311, and the slips Sl
2139833
- 58 -
Sn prepared as described above are successively poured
into the mold 312, thereby molding a laminated body. When
each of the slips Sl ~ Sn is to be poured, it is poured af-
ter the previously poured slip has lost its water content to
a certain extent so that they will not be mixed with each
other and the solvent of the previously poured slip will pen-
etrate into the board 311.
As shown in Fig. 18(b), a mold bar 313 may be set
either before or after the slips are poured. When the lami-
nated body is partly dried, the laminated body is removed
from the mold 312. The removed laminated body serves as an
end cap 303 with a tapered through hole 314 defined therein
as shown in Fig. 18(c~. The through hole 314 may be of a
stepped shape as shown in Fig. 18(d).
A bulb 301 molded of a pure alumina slip is pre-
pared, and the end cap 303 which is made wet is joined to an
end of the bulb 301 as shown in Fig. 18(e), after which the
bulb 301 and the end cap ~03 are dried. At this time, the
bulb 301 and the end cap 303 are unfired, and the bulb 301 is
not light-transmissive.
Then, the bulb 301 and the end cap 303 are de-
greased at 600C for 5 hours in a moisture-containing hydro-
gen reducing atmosphere, and then fired at 1300~C for 5 hours
in a dry hydrogen reducing atmosphere. Thereafter, the pro-
duced fired body is subjected to HIP in an argon atmosphere,
and then annealed at 1150C in a dry hydrogen reducing atmo-
sphere, thereby producing an integral body of the light-
213~983~9
. ~ `
- 59 -
transmissive bulb 301 and the end cap 303.
The hole 314 defined in the end cap 303 is tapped
to produce an internally threaded surface 309, and then an
electrode rod 304 is inserted and an externally threaded por-
tion 305 of the electrode rod 304 is threaded in the inter-
nally threaded surface 309. Finally, the electrode rod 304
is fixed and sealed by a platinum solder 307, and an amalgam
introduced into the bulb 301 through a hole 308 defined in
the electrode rod 304 by a jig in the form of a platinum
pipe. In this manner, the lamp is completed.
While the bulb and the end cap are simultaneously
fired in the illustrated embodiment, they may be separately
fired and then joined to each other. According to such a
modification, the bulb of alumina may be degreased and fired
in the atmosphere, then subjected to HIP, and thereafter an-
nealed into a light-transmissive bulb of alumina. The end
cap which is fired in the same manner as described above may
not be subjected to HIP and annealed. The bulb and the end
cap may be joined to each other by laser heating in vacuum or
at 2000C or higher, or glass having the same coefficient of
thermal expansion as alumina. ~he glass should preferably be
melted high-melting-point glass of a high softening point of
900C or higher.
The end cap may be formed by a doctor blade process
or an injection molding process as well as the slip casting
process.
In the doctor blade process, prepared slurries are
213983~
-
- 60 -
formed into tapes of desired thicknesses, and the tapes are
integrally joined together into an end cap having a gradient
function by thermal compression. The same slurries may be
used to cast the bulb or poured into a mold and then solidi-
fied into the bulb.
In ~.e injection molding process, sheets of desired
thicknesses are formed and bonded together with heat, thus
producing an end cap which will be joined to a previously
molded bulb by thermal compression.
According to the third embodiment, each of the end
caps which seal the open ends of a metal vapor discharge lamp
is of a multilayer structure, and the coefficients of thermal
expansion of the layers vary gradually from the open end of
the bulb toward the core which holds an electrode, so that
the end caps have a gradient function. Consequently, the end
caps are effective to prevent damage due to different thermal
expansions and leakage of the metal vapor sealed in the bulb.
Fig. 19 shows a modification of the third embodi-
ment. According to the modification, a bulb 301l differs
from the bulb ~01 shown in Fig. 16 in that the opposite ends
of the bulb are not fully open, but have respective end sur-
faces 301a. The end surfaces 301a have respective small
openings as large as a larger-diameter portion of the tapered
through hole 314 for allowing the electrode rods 304 to be
inserted therethrough into the bulb.
Light-emitting bulb assemblies according to fourth
and fifth embodiments will be described below with reference
213983~
- 61 -
to Figs. ~0 through 27(a) and 27(b).
~ ig. 20 shows in cross section a light-emitting
bulb assembly according to the fourth embodiment of the pre-
sent invention, for being incorporated in an outer tube of a
metal vapor discharge lamp. A tubular bulb 401 shown in Fig.
20 is made of light-transmissive polycrystalline alumina hav-
ing a high purity of 99.99 ~ = 4N, and electrode sealing mem-
bers 403 are disposed as closures against inner walls of op-
posite end openings 402 of the bulb 401.
The electrode sealing members 403 are made of an
alumina material having a lower purity of 93 - 97 %, for ex-
ample, than the bulb 401 which serves as a light-emitting
body. Each of the electrode sealing members 403 is of a mul-
tilayer structure which comprises a first layer 403a as a
bulb-side region and a second layer 403b as a core-side re-
gion (the multilayer structure may be composed of three lay-
ers or more including an intermediate layer or layers). The
first layer 403a held against the inner wall surface of the
bulb 401 is made of alumina having a purity of 96 %, for ex-
ample, and the second layer 403b inward of the first layer
403a is made of alumina having a purity of 93 ~, for example.
Electrode rods 404 as cores are inserted in the re-
spective electrode sealing members 403, and caps 405 through
which the electrode rods 404 extend are disposed against the
open ends of the bulb 401. Sealing glass 406 produced by
melting and cooling a glass solder is positioned to provide a
seal between the electrode sealing members 403 and the elec-
- 62 -
trode rods 404, between the electrode rods 404 and the caps
405, and between the ends of the bulb 401 and the electrode
sealing members 403 and the caps 405.
The purity of the caps 405 is preferably an average
of the purities of the bulb 401 and the electrode sealing
members 403. The caps 405 may be dispensed with as required.
Since a glass component is present in grain bound-
aries of alumina ceramics in the inner walls of the electrode
sealing ~embers 403 which are made of an alumina material
having a lower purity than the bulb 401 and which are dis-
posed in the openings of the bulb 401, the electrode sealing
members 403 adhere well to the sealing glass solder, thereby
improving a sealing capability. A composition gradient
structure made of aluminas having different purities serves
to suppress the generation of thermal stresses.
A process of manufacturing the above ceramic light-
emitting bulb assembly will be described below with reference
to Figs. 21 through 23(a) ~ 23(f).
First, as shown in Fig. 21, a fine powder of alu-
mina having a high purity of 4N or more for producing light-
transmissive alumina is prepared in a container C41, and a
fine powder of alumina having a lower purity (93 % in this
embodiment) is prepared in a container C42. The fine powder
of low purity contains impurities of silica, magnesia, and so
on. The fine powders of alumina should preferably be se-
lected which have similar firing behaviors.
To the powders which have been weighed, there are
`3 9 ~ 3 g ~
. ~ . . - . .
- 63 -
added predetermined amounts of distilled water, a commer-
cially available dispersing agent, and a binder. The materi-
als are then mixed for 24 hours by a ball mill, producing
slips for being cast. Suitable amounts of these slips are
mixed into several kinds of slips having different purities.
The slips are mixed for about 1 hour by a stirrer. In this
manner, as shown in Fig. 22, an alumina slip S41 having a high
purity (4N) is prepared in a container C43, an alumina slip
S42 having a purity of 96 % is prepared in a container C44,
and an alumina slip S43 having a purity of 93 % is prepared in
a container C4s.
Thereafter, as shown in Figs. 23(a) and 23(b),
while masking, with masks 412, peripheral portions of slip
inlet/outlet ports of a porous mold assembly or plaster mold
assembly 411 that can be divided into two molds (only one
mold is shown in the cross-sectional and plan views of Figs.
23(a) and 23(b)), the alumina slip S41 having a high purity is
poured from the container C43 into the plaster mold assembly
411, and left for a predetermined period of time. After a
highly pure alumina layer 413 has been deposited on an inner
circumferential surface of the plaster mold assembly 411, the
alumina slip S41 is discharged.
Then, as shown in Fig. 23(d), one end of the plas-
ter mold assembly 411 is dipped in the alumina slip S42 having
a purity of 96 % to deposit an alumina layer on only a seal-
ing portion for thereby forming a 96%-alumina layer 414 on an
inner circumferential surface of the highly pure alumina
213983~ -
- .
- 64 -
layer 413 as shown in Fig. 23(e). Likewise, a 96%-alumina
layer 414 is also deposited on an inner circumferential sur-
face of the highly pure alumina layer 413 at the other end of
the plaster mold assembly 411. Then, one end of the plaster
mold assembly 411 is dipped in the alumina slip S43 having a
purity of 93 ~ to deposit an alumina layer on only a sealing
portion for thereby forming a 93%-alumina layer 415 on an in-
ner circumferential surface of the 96%-alumina layer 414 as
shown in Fig. 23(f). Likewise, a 93%-alumina layer 415 is
also deposited on an inner circumferential surface of the
96%-alumina layer 414 at the other end of the plaster mold
assembly 411.
The formed body thus produced is fired at laO0C
for 6 hours in a hydrogen reducing atmosphere, thus producing
a bulb 401 having a light-emitting portion composed of the
light-transmissive alumina layer and sealing portions com-
posed of the electrode sealing members 403 which comprise
alumina layers of low purity.
By selecting powders, the bulb may be fired at
1350~C for 6 hours in the air and thereafter heated at 1350C
for 2 hours under 1000 atmospheric pressures in an argon at-
mosphere by way of hot isostatic pressing. In this case,
however, since almost no alumina of low purity is generally
sintered at this temperature, the purity of alumina in the
innermost layer in the sealed portions have to be 97 % or
higher.
The bulb 401 and the electrode sealing members 403
2139~33
.
- 65 -
thus produced are then machined at their inner surfaces and
the light-emitting portion is machined at its outer circum-
ferential surface, and then a metal vapor discharge lamp is
assembled~
A fifth embodiment which is a modification of the
fourth embodiment will be described below with reference to
Figs. 24 through 27(a) and 27~b). In the fifth embodiment, a
tubular bulb 521 is made of light-transmissive polycrys-
talline alumina having a high purity of 99.99 % = 4N, and
electrode sealing members 523 of a laminated structure made
of alumina of low purity are disposed on respective opposite
ends 522 of the bulb 521. Electrode rods 524 as cores are
inserted respectively in the electrode sealing members 523.
Caps 525 of alumina through which the electrode rods 524 ex-
tend are disposed outside of the electrode sealing members
523, and the electrode sealing members 523, the electrode
rods 524, and the caps 525 are sealed by sealing glass 526.
The electrode sealing members 523 is made of an
alumina material which has a lower purity ~e.g., 99 97 %)
than the bulb 521 which serves as a light-emitting portion.
Each of the electrode sealing members 523 is of a laminated
structure including a first layer 523a, a second layer 523b,
and a third layer 523c ~the laminated structure may include
four or more layers) arranged along the axial direction of
the bulb 521 or the electrode rods 524. The first layer
523a, the second layer 523b, and the third layer 523c are
progressively thicker in the direction from the first layer
9 8 3 ~
- 66 -
523a toward the third layer S23c. As a result, the third
layer 523c and the second layer 523b have a greater area held
against the electrode rods 524 than the first layer 523a.
The caps 525 are made of alumina having the same purity as
that of the third layer 52~c.
The caps 525 may be dispensed with as required.
A process of manufacturing the above ceramic light-
emitting bulb assembly will be described below with reference
to Figs. 25 through 27(a) and 27(b).
First, as with the fourth embodiment, a fine powder
of alumina having a high purity of 4N or more for producing
light-transmissive alumina, and a fine powder of alumina hav-
ing a lower purity (93 ~ in this embodiment) are prepared.
To the powders which have been weighed, there are added pre-
determined amounts of distilled water, a commercially avail-
able dispersing agent, and a binder. The materials are then
mixed for 24 hours by a ball mill, thereby producing, as
shown in Fig. 25, an alumina slip Ssl having a high purity
~4N) is prepared in a container Cs1, an alumina slip S5~ hav-
ing a purity of 97 % in a container Cs2, an alumina slip Ss3
having a purity of 95 ~ in a container Cs3, and an alumina
slip Ss~ having a purity of 93 % in a container Cs4.
Thereafter, as shown in Fig. 27(a), a tubular mold
532 having a size matching the outside diameter of a bulb is
set on a porous mold assembly or plaster mold assembly 531,
and a mold bar 533 is vertically placed centrally in the mold
532. Then, the alumina slip Ss4 having a purity of 93 ~, the
2 1 3 9 8~ 9
. .
- 67 -
alumina slip Ss3 having a purity of 95 ~, the alumina slip S52
having a purity of g7 %, and the alumina slip Ss1 having a
high purity are successively poured into a space defined by
the mold 532 and the mold bar 533, thereby molding a lami-
nated body. When each of the alumina slips is to be poured,
it is poured after the previously poured slip has lost its
water content to a certain extent so that they will not be
mixed with each other.
A pipe 534 shown in Fig. 26 which will serve as the
bulb 521 is formed of the highly pure alumina slip S5l. The
pipe 534 is then inserted into the mold 532 while the highly
pure alumina slip Ssl for producing an end 522a of the bulb
521 is not being dried, and integrally joined to the lami-
nated body, thereby producing a molded body as shown in Fig.
27~b). Thereafter, as with the above embodiment, the molded
body is fired, machined, and assembled.
With the present invention, as described above,
electrode sealing members made of an alumina material having
a lower purity than a light-emitting portion are disposed on
respective opposite ends of a bulb, and a glass solder or
sealing glass is held in contact with the electrode sealing
members to keep them out of contact with the bulb.
Therefore, the sealing capability is made highly reliable for
allowing the lamp to have an increased service life.
As with the above embodiment, since the composition
of the electrode sealing members is of a gradient nature, the
sealing capability of the sealing regions is further in-
2 1 3 ~ 8 3 ~
- 68 -
creased.
A sealing structure of a light-emitting bulb assem-
bly for a metal vapor discharge lamp according to a sixth em-
bodiment of the present invention and a method of manufactur-
ing such a sealing structure will ~e described below with
reference to Figs. 28 through 30(a) ~ 30(d).
Fig. 28 shows a bulb 601 made of light-transmissive
polycrystalline alumina to be incorporated in an outer tube
of a metal vapor discharge lamp. Caps 604 of alumina as clo-
sures are fitted in respective end openings 602 of the bulb
601 through sealing glass 603.
Each of the caps 604 comprises a high-purity alu-
mina portion 604a, a gradient-composition portion 609b, and a
low-purity alumina portion 604c. The high-purity alumina
portion 604a as a bulb-side region is made of Al2O3 having a
purity of 99.99 % and exposed to the interior of the bulb
601. The low-purity alumina portion 604c as a core-side re-
gion is made of Al2O3 having a purity of 93.0 ~ and exposed to
the exterior of the bulb 601. The gradient-composition por-
tion 604b as an intermediate region has a section held
against the high-purity alumina portion 604a and having a pu-
rity of 99.99 %, is progressively lower in purity toward the
low-purity alumina portion 604c, and has a section held
against the low-purity alumina portion 604c and having a pu-
rity of 93.0 ~. The gradient-composition portion 604b with
such a continuous gradient composition has a greatly in-
creased peeling strength. The low-purity alumina portion
- - ` `- Z~1 3 9 ~ 3 9
.
.
- 69 -
604c has a greater width along the axial direction of the
bulb than the width of the high-purity alumina portion 604a.
As shown in Fig. 29(d), each of the caps 604 has
axial holes 605, 606 defined therein. An internal electrode
rod 607 is press in the hole 605, and an external electrode
rod (lead) 608 is pressed in the hole 606. The holes 605,
606 are such diameters which will be about 200 ~m larger than
electrode rods 607, 608 after being fired. This prevents the
caps from being obstructed and cracked by the electrodes when
fired.
The low-purity alumina portion 60~c has a radial
hole 609 defined from its side toward the inside thereof in
communication with the axial hole 605. A conductive film 610
of tungsten ~W) or the like is disposed in the radial hole
609 and on the outer surface of the low-purity alumina por-
tion 604c. The conductive film 610 serves to provide a good
electric connection between the internal electrode rod 607
and the external electrode rod 608. The conductive film 610
may be made of Nb, Ta, Mo, Ni, or the like.
A process of manufacturing each of the caps 604
will be described below with reference to Figs. 29(a) through
29(e). First, as shown in Fig 29(a), 10~ g of A12O3 of a high
purity (99.99 %), 100 g of A12O3 of a low purity (93.0 %), 50
g of water, and a deflocculant are mixed for 24 hours by a
ball mill, thereby producing a slip S61 of A12O3 of a high pu-
rity and a slip S62 of A12O3 of a low purity.
Then, as shown in Fig. 29(b), the slips S61, S62 are
- ~ 21398~
- 70 -
mixed with each other to produce a plurality of slips S63 hav-
ing purities ranging between 99.99 ~ and 93.0 %. Thereafter,
as shown in Fig. 29(c), the slips are successively poured,
from the highly pure slip S61, into a mold 615 set on a porous
body or a plaster body 614, producing a molded body 616 prior
to being fired by way of one-sided deposition.
The molded body 616 is then temporarily fired at
1100C for 2 hours, so that the molded body 616 has a hard-
ness that allows the molded body 616 to be handled.
Thereafter, the molded body 616 is machined to form axial
holes 605, 606 and a radial hole 609, as shown in Fig. 19(d),
and shaped into a cap. A conductive paste 610 is then intro-
duced into the radial hole 609 and applied to the outer sur-
face of the low-purity alumina portion 604c. With the inter-
nal electrode rod 607 and the external electrode rod 608 be-
ing inserted, the assembly is fired at 1570C for 3 hours in
an atmosphere of N2 and H2 (N2 : H2 = 80 : 20), thereby pro-
ducing a cap 604 as shown in Fig. 29(e). The cap 609 is in-
serted in one of the openings 602 of the bulb 601, and sealed
by glass 603 or an alloy of a low melting point.
Figs. 30(a) through 3~(d) show a modification of
the process of manufacturing the light-emitting body accord-
ing to the sixth embodiment. According to the modified pro-
cess, as shown in Fig. 30(a), using two porous bodies or
plaster bodies 614a, 614b and two molds 615a, 615b, a molded
body 616a serving as a high-purity alumina portion and a gra-
dient-composition portion, and a molded body 616b serving as
213983~
_ -- 71 --
a low-purity alumina portion are produced as shown in Fig.
30(b~.
Then, as shown in Fig. 30(c), the surface of the
molded body 616b is coated with a conductive paste, and the
molded body 616a is bonded integrally to the molded body 616b
by the conductive paste. Subsequently, an internal electrode
rod 607 and an external electrode rod 608 are inserted, and
the assembly is fired into a cap 604 as shown in Fig. 30(d) .
Since the conductive paste which interconnects the molded
bodies 616a, 616b provides an electric connection between the
internal electrode rod 607 and the external electrode rod
608, the radial hole 609 as shown in Fig. 29(d) is not re-
quired.
According to the sixth embodlment as described
above, each of caps which close respective openings of a
light-emitting bulb assembly for a metal vapor discharge lamp
and support internal and external electrodes separately from
each other is composed of a high-purity alumina portion ex-
posed to the interior of the bulb assemblyr a low-purity alu-
mina portion exposed to the exterior of the bulb assembly,
and a gradient-composition portion interconnecting the high-
purity alumina portion and the low-purity alumina portion,
and a conductive film which provides an electric connection
between the internal and external electrodes is disposed on
the surface of the low-purity alumina portion. The conduc-
tive film has a peeling strength increased to lO kg/cm2 from
a conventional value ranging from l to 4 kg/cm2.
~13983~
- 72 -
Since the high-purity alumina portion is exposed to
the interior of the bulb assembly, the lamp characteristics
are prevented from being degraded due to a corrosive compo-
nent such as Na. As no conductive film is disposed on the
high-purity alumina portion and the gradient-composition por-
tion, no back arcs are produced. Metals such as Nb, Ta, Mo,
Ti, and so on may also be used as the conductive film
(metallized film).
Industrial Applicability
A sealing structure ~or a light-emitting bulb as-
sembly allows a discharge light-emitting bulb assembly to be
highly reliable and have a long service life. The light-
emitting bulb assembly can be used in a metal-vapor discharge
lamp such as a mercury-vapor lamp, a metal halide lamp, or a
sodium-vapor lamp, or a high-intensity discharge lamp.
