Note: Descriptions are shown in the official language in which they were submitted.
~Z86373
TANTALUM CAPACITOR LEAD wIRE
8ackground of the Invent;on
The present invent;on relates to the tantalum
capacitor art and more particularly to the production of
tantalum wire for use as leads to tantalum powder
capacitors. In the production of tantalum capacitors,
tantalum powder is compressed to a pellet, the pellet
including a tantalum lead wire, the resultant green
pellet with the associated lead w1re is then subjected
to a s;ntering operation, normally under a vacuum, to
create a metallurgical and electr;cal bond between the
;nd;v;dual powder gra;ns and to the lead w;re.
Thereafter, the resultant s;ntered body ;s anod;zed and
;mpregnated w;th an electrolyte, preferably sol;d, and
encapsulated to form the f;nished capacitor. It ;s
essential that there be a good electrical and
metallurgical bond between the tantalum lead wire and
the capacitor pellet. It is also essential that the
tantalum lead have sufficient mechanical strength and
flexibility to withstand the rigors of further
fabrication and attachment (which is often done on
automatic machinery) to other circuit elements.
As circu;t m;n;atur;zat;on has advanced, the need
for smaller capac;tors has also developed. Capac;tor
m;n;atur;zat;on has progressed to the po;nt where many
are of a d;ameter smaller than 2.5 mm and the capac;tor
leads are as small as .25 mm ;n diameter. This has been
due to the;mprovement in obta;n;ng higher capacitance
Per unit we;ght of tantalum powder. As capac;tors
become smaller, the percentage of value in the lead wire
becomes larger so that with the smallest capacitors the
lead wire is almost SOX of the capacitor value.
.
' . ' ,' - :,
',,~ '
- : :
lZ86373
-- 2
Another problem with capacitor miniaturization is that
the necessary small diameter of the capacitor lead wire is
such that at the elevated temperatures employed in the
sintering operation, grain growth in the tantalum lead wire
can be sufficiently great for the grain size to equal the
tantalum wire diameter. This makes a very brittle wire.
In the past this problem of grain growth has been addressed
by various means to inhibit the grain growth in the wire.
Such inhibitors are oxides, nitrides, and various alloying
lo constituents. This has a disadvantage that these
inhibitors are difficult to control and expensive to add
and may also interfere with the metallurgical and
electrical characteristics of the tantalum lead wire as
well as perhaps contamination of the tantalum powder
itself. contamination introduced by the use of finder
powder sizes which contain higher values of oxygen would
also tend to further embrittle the lead wire.
Summary of the Invention
The present invention is directed to overcoming the
objections of the prior art by providing a tantalum
capacitor lead which can be subjected to high temperature
sintering without deleterious grain growth and at
substantially lower cost. This objective is accomplished
by providing a tantalum capacitor lead wire formed of a
core and more than one discrete surface layers of tantalum.
The interface between each tantalum layer acts as a grain
growth inhibiting boundary when the capacitor is sintered
at an elevated temperature. In a preferred embodiment,
when the wire has a diameter of .25 mm, the individual
tantalum layers have a radial thickness of .01 mm or less.
When such a wire is subjected to the high temperature of
tantalum capacitor sintering, grain growth in the layer and
between layers is severely limited. Some grain growth
between layers would be expected when very high sintering
temperatures are used, on the order of greater than
1800 C, but even this growth would be inhibited such that
~ '
:
.
'
:
1~6373
the extent of grain growth is substantially less than if a
solid single Ta layer is used. It is also apparent that
more layers and thinner layers would aid in producing a
finer grain structure after sintering. Accordingly, such
a wire can be subjected to 1950 C for 30 minutes to
provide appreciably less grain growth between layers.
Single crystal growth around the circumference of the
layer will not occur, no matter how pure the tantalum may
be. Grain growth will occur in the thickness dimension
first and will then stop. Growth in the circumferential
direction becomes increasingly more difficult because the
axis of maximum growth depends on crystal orientation.
Each grain tends to stabilize when one dimension equals 4
times any other dimension, i.e. crystal growth is
encouraged to have minimum surface area. With prior art
tantalum wires, unless grain stabilized, the grain growth
can extend completely through the solid Ta wire to give a
bamboo effect.
The product of the present invention is preferably
formed by wrapping a tantalum foil around a metal billet to
provide more than one layer of tantalum around the billet.
In a preferred form of the invention, at least three
tantalum layers are used and are compacted. The compacted
body is inserted into an extrusion billet. The resultant
composite is then extruded and the extruded composite is
further reduced by rolling and/or
., - ' ' . .
... .
'': ' . ~ ' ' ' ' . :
,:.: .. , . . . - .
.'. - - : -
:
.: - :
1~86373
-- 4 --
draw;ng to a w;re of the requis;te small dimens;on for
use as the f7nal tantalum lead.
The core around which the tanta~um sheet is
initially wrapped may be niobium or Nb 1X Zr. If
niobium, (or Nb 1X Zr) it remains in the center of the
wire and is embodied in the final capacitor. Since
niobium is cheaper and has approximately one-half the
density of tantalum, the composite wire of a given size
made according to the present invention, in addition to
its other advantages, will be for the same voLume
significantly less expensive than one formed of solid
tantalum. The core material can also be made of a Ta
alloy such as Ta-Nb where the density of this alLoy is
substantially less than that of sol;d Ta. The actual
15 alloy chosen would be determ;ned on the spec;f;c
electrical and mechanical propert;es des;red for the
lowest cost (less dense) appl;cat;on. If a Ta-Nb alloy
is used for the core there should be at least 2ûX Nb
present to compensate for the lower scrap value of the
20 w;re trimm;ngS.
The total thickness, ;.e., the number of layers and
thickness of the Ta fo;l used, would be the min;mum
amount necessary to provide the requ;red electr;cal and
mechanical properties. During high temperature
25 sintering, a certain amount of interdiffus;on between
niob;um and tantalum will occur. A thicker tantalum
layer ~about .û254 mm) w;ll show l;ttle or no n;ob;um
present at the surface of the lead w;re (Table I). It
is also possible to use a thin layer of h;gh melt;ng
30 point material like molybdenum or tungsten as a
diffusion barrier to further reduce this alloying
tendency. Obviously, a solid molybdenum or tungsten
core can also be used.
If copper is used as the core material, it can be
35removed to form a hollow wire which is then used as a
- :
', , ' ' ' . ' ' ' -
- .
,
1286373
- 5 -
capacitor lead and thus eliminate the alloying problem.
Detailed Description of the Invention
In order to more fully comprehend the invention,
reference should be had to the following detailed
descript;on taken ;n connect;on w;th the attached
draw;ngs where;n:
F;g. 1 ;s a schematic diagrammatic cross-sect;onal
view of a preferred starting billet for making a w;re of
the present invent;on;
Fig. 2 is a flow sheet of the various process steps
employed in one preferred embod;ment of the ;nvent;on.
Fig. 3 is a schematic diagrammatic cross-sectional
view of a preferred starting billet for making a wire of
the present invention according to another embod;ment
15 thereof;
Fig. 4 is a schemat;c d;agrammat;c sect;onal view of
an end of a final capacitor wire made accord;ng to one
peferred form of the ;nvent;on.
Referring now to Fig. 1, the start;ng billet for
20 making the tantalum wire is shown at 1û with a core 12
having a number of tantalum layers 14 surrounding the
core. As can be seen, there are a number of interfaces
13 between the various layers, in the illustrated case,
five layers being shown. An outer layer of Copper 16 is
25 used for extrusion-
Fig. 2 should be considered in connection with the
follo~ing example which shows one preferred method of
practicing the invention.
EXAMPLE I
A niobium ~99% Nb 1X Zr) rod 50 mm long and 38 mm
diameter is cleaned in acetone and wrapped with 12
l~yers of tantalum foil derived from powder metallurgy
stock, the foil being .15 mm thick. Such a powder
derived foil is inherently more grain stabilized than
electrom beam tantalum due to the relatively large
" ....... .
.
.
.
:: :
~' ',','''' ' ' ' ~ '.'
'-. ' '
~ .. - -
1286373
- 6 -
amount of i~purities resulting from the powder process.
This composite is then assembled to produce a structure
having a final diameter of 47 mm. The compacted
tantalum foil niob;um compos;te is then ;nserted ;nto a
copper b;llet hav;ng an interior diameter of 48mm and an
exterior diameter of 51 mm. This is sealed, heated to a
temperature of 87ûC and then extruded under a
pressure of 250 Tons at a rate of 65 inches/minute to an
extrusion diameter of 12.8 mm. The resultant extrusion
product is then further drawn through a number of
dra~ing dies to a final diameter of .38 mm. The final
product is then etched in acid to remove the outer layer
of copper. Further drawing (after annealing) of the
bare wire was done to obtain superior surface
qualities. The wire is then cleaned, cut to appropriate
capacttor lead length, and assembled into the capacitor
compacts to make the "green" capacitor pellets to be
vacuum sintered.
"Tantalum" and "niobium" includes alloys of tantalum
and/or niobium suitable for use as capacitor leads.
The wire used in Example I ~as vacuum sintered at
1950C for 30 minutes and then subjected to scanning
electron microscope examination of its surface to detect
the presence of Nb which had diffused from the core to
25 the wire surface. This test was done at a number of wire
thicknesses, using the same ratio of Ta surface layer to
core since all samples were drawn from the same starting
material.
TABLE I
Diameter
inches Ta thickness (mm) XNb (Atomic)
.040" .0406 0
.025" .0254 0
.020" .0191 0
.015" .0145 3-5X
.0113" .0111 25-30Z
. .
' ~ ~. ' ' '
: : :
.. . . .
: ' .
12t~6373
-- 7 --
The same wire samples were then subjected to a
standard bend test where each wire was bent 90,
straightened and rebent 90 in a plane removed 120
from the first bend (Table II) .
TA~LE II
Diameter Results
.040" Not satisfactory
.025" passed
.020" passed
.015" passed
.0115" passed
It is believed that the failure of the .040" wire
was due to the larger diameter which, on bending,
l5 generates high surface strain due to its distance from
the neutral axis of the w;re.
While one preferred embodiment of the invention has
been described above, numerous modifications may be made
without depart;ng from the spirit of the invention. For
20 example, the Nb core can be replaced by a copper core
which is leached out of the wire after the leads have
been cut to length and before the leads are inserted in
the green compact. (see dotted line steps in Fig. 2)
This creates a hollow tube of Tantalum having a surface
25 comprising many layers of Ta which inhibit grain
growth. Similarly the Nb core can have a Cu center.
Other core materials can be used so long as adequate
provisions are made, such as the use of diffusion
barriers, to prevent undesirable constituents of the
30core from diffusing to the surface of the Ta wire.
where absolute prevention of diffusion of the Nb to
the surface is to be prevented, a layer of tungsten or
molybdenum may be provided between the tantalum and
niobium.
.
. .
.
. .
,
128637~
In another embodiment of the invention which, is
part;cularly useful for manufacturing a product with
minimum cost rather than optimum performance as the the
principal criteria to be considered, a single layer of
tantaLum is prov;ded on a refractory metal base, such as
niobium. An example of this embodiment of the invention
is shown in Example II set forth below.
EXAMPLE II
The following components are assembled: copper tube
15.9 mm O.D. with 12.7 mm I.D.; tantalum tube 12.7 mm
O.D. with 0.5 mm wall; niobium solid rod 11.4 mm
diameter. These components were cleaned and etched with
appropriate acid, assembled and swaged to 7 mm O.D.
This composite assembly is shown in Fig. 3. The
15 composite was then drawn to 0.5 mm O.D. and the copper
was removed by etching. This gave a final product
having a tantalum outer layer thickness of about
û.015 mm with a final total wire diameter of 0.4 mm,
this being the same diameter as in the product made by
initial extrusion in Example I.
It appears that, while the extrusion step and the
multi-layer wrapping are desirable when max;mum
performance is necessary, these are not necessary to
provide an inexpensive product which is sat;sfactory for
25 many uses.
In Example II, pure tantalum was used rather than
tantalum sheet formed by powder metallurgy techniques.
Accordingly, there was grain growth through the outer
layer of tantalum during subsequent heating to
30 1800C. However, grain growth did not extend
completely around the circumference of the tantalum,
since, as po;nted out above, after any d;mens;on of the
crystal becomes greater than four t;mes any other
d;mens;on of the crystal, further grain growth ;s
35 inh;bited. It should also be noted that the f;nal
:
- -.
~,
' .
12863*3
tantalum layer is thick enough to prevent diffusion of
the niobium to the surface of the tantalum wirè, even
when sintering temperatures on the order of 1950C are
employed with times of 30 minutes.
In the specific embodiment of Example II, a drawn
tube was utilized. A seam welded tantalum tube (which
is inherently less expensive) may be equally used.
It is also poss;ble to use one or more layers of
tantalum sheet or foils wrapped around a n;obium core or
10 to use several concentr;c tubes of tantalum. In both of
these cases, grain growth between layers would be
limited.
While swaging has been mentioned as a first step in
Example II, it can be used for taking the product down
15 to even finer diameter than in Example II, or it can be
for lesser reduction. The principal objective of the
swaginy is to perm;t subsequent further reduct;on of the
wire as a unitary product, although ;t is not necessary
that a compLete metallurg;cal bond be obta;ned between
20 the copper, tantalum and niobium.
As mentioned ;n connect;on w;th var;ous
modifications of Example I, diffus;on barr;ers can be
employed 1n the system of Example II and the tantalum
and the niobium can compr;se var;ous alloys.
An important characterist;c of both Examples I and
II is that the final capac;tor iead have a uniform
tantalum surface w;th a radial thickness less than 0.3
mm. In Example II, the final tantalum thickness ;s
0.015 mm.
~hen the capac;tor lead wires are finally cut, ;t is
preferred that the tantalum outer surface be pinched and
closed on itself, thus exposing a minimum amount of the
core material. This cutting operation embod;es the use
of rounded cutt;ng surfaces and, in effect, draws the
35tantalum coating down onto itself and thus over the
. . . , . :
' ' ~ ' ' '. : : . - '
- -. : -
: : . .
.'
~, "
. . . . . .
12863'~3
-- ~o --
niobium core so that little if any niobium is exposed in
the final capacitor lead for interfering with the
anodized tantalulm dielectric surface. An example of
this type of preferred lead end is shown in exaggerated
cross-section in Fig. 4. The wire should be in the
annealed or stress relieved condition for the best
results.
In either Example I or Example II, the'core can be
made of alternate wrapped layers of niobium and tantalum
sheets or foils. After sintering, the grain size of the
core would be fine grained due to the limited diffusion
of the niobium into the tantalum layers and should
result in greater ductility in the wire after
sintering.
.. . . . . .
, . - .
.,: - . ' ; - . - i
~' ' .
- . ~ .
.. . .
