-
Carbide Surface Treatment
Graph shows that the ECP surface
treatment doubled joint strength on the
average and dramatically increased
process reliability.
The Problem
Tungsten carbide
is made to
be wear and corrosion resistant.
Yet,
in order to
braze it successfully, you
must create a physical and chemical joining. More
information on Tungsten carbide and
other tool materials can be found in
our
carbide tooling
index.
In
addition carbide made with a shortage of carbon is
a poor quality product.
If
there is a bit too much carbon then the
carbide is good and the excess carbon
deposits on the surface.
This free carbon deposits as graphite
and interferes
with brazing.
Time in a sintering oven is
expensive and there is a strong motivation to
remove the parts before they are dully
cool.
The figure of 1,000 F is given
as
the point for the beginning of
significant oxide formation on tungsten carbide in
air.
However the question
arises as to what level of formation is
significant.
Tools are steadily becoming
thinner
.
The
material in the width of the saw cut becomes
sawdust which is less
valuable than
lumber. Thus saws have been getting constantly
thinner.
Now
there are saw
blades in mills that maybe three feet across and
yet somewhere
between 0.060”
and
0.090” thick.
State of the art seems to be about
.060” kerf
on a 12” saw blade with a
blade change every sixteen hours.
The thinnest we
have gone
with carbide is 0.025” in a butt braze.
We know a filer in a mill who
runs .053” kerf.
The carbide surface must be properly
prepared in order to get he necessary
strength from a braze join over a very
small area.
This is a reciprocating knife.
To make it we butt brazed a piece of
carbide 0.5”
by 0.025” to a piece of
spring steel.
Two of the
knives are shown next to a
dime.
On the right is a knife in vise.
The vise was then tightened to get a
bend and test for strength.
You can see by the bend how much
pressure is on
it.
I
think it would have taken more bending but I don’t
know how much
more.
In any case this shows what
can be done with good surface preparation and
brazing techniques.
Intermetallic
A good braze joint creates a chemical
bond and an intermetallic compound.
A
bad braze creates three distinct
layers.
Think of plywood with and
without
glue.
Bad
Microphotograph;
surfaces
(Three
bad
distinct layers)
Good surfaces
(Three
interlocked
layers)
Microphotograph;
Good
Pretinning
Pretinning
or tinning is the
application of the brazing alloy to one part
first.
It is
very important
here because it clearly reveals the surface
condition of the
carbide.
Below are two views of curved tungsten
carbide parts.
These were
designed for a machine that would
duplicate the look of an adze on
wood.
The parts need to be pretinned.
This is the customer’s attempt on an
untreated surface using a torch and a braze
alloy rod.
This is the same part surface treated
and pretinned.
This was pretinned in
an
oven.
The cut alloy was
laid on the part and then it was heated.
Remember we
are looking down
into the curve.
You can see where the
molten alloy puddled in
the center
however the capillary attraction was so strong
that the alloy largely
stayed in place
on the curves in spite of gravity.
Braze Treating Options
Sandblasting
Washing
Plating
Toney
salt bath Process
ARS
ECP
Tuffco
Sandblasting
Sandblasting works somewhat.
Carbide is typically sintered in sand
and then
the parts are sandblasted to
clean them up.
However you have the
usual
sandblasting problems.
The sand (or whatever media) can become
contaminated with carbon.
In addition the sandblasting process
can smear
the carbon in an even layer
rather than removing it.
There is a study that shows that using
alumina for sandblasting carbide results
in the alumina shattering and becoming
embedded into the carbide.
This
happens but I have never seen where it
is a significant problem.
Washing
Carbide
parts are often washed after sandblasting however
this is rarely meant
to do more than
remove loose dust.
This process
provides an excellent
opportunity for
simple chemical wash that will remove oxides and
free carbon
however this is almost
never done.
Plating
This is
standard plating technology.
Sometimes
there is a layer of cobalt or
nickel
plated directly to the part and sometimes it is
plated over an
intermediate layer of
copper or something similar.
Copper
plates well to
tungsten carbide.
Nickel and cobalt both plate well to
copper.
The problem is
that
the coating or plating is only as good as the bond
to the underlying
surface.
If the surface is not prepared properly
the plating will peel off.
Toney Salt Bath Process
L.B. Toney was granted a patent to
treat tungsten carbide in a salt bath.
#
2,979,811
Patented Apr. 18, 1961.
The
illustrations on the front page of the patent show
the kind of wetting one would expect
after a salt bath cleaning. My guess is
that L.B. Toney was at a company that
had a salt bath process and decided to
put carbide tips into the molten salt
to see what would happen.
It looks
like
the 2200 F salt thoroughly cleaned
the carbide.
This press was once described as making
the cobalt migrate to the surface.
There have been other explanations.
Currently I believe the explanation is
that they remove material from the
surface leaving it Cobalt rich.
Maybe a decade ago Ajax Electric sold
equipment that would allow people to
do
the Toney process.
We looked at it
but, in our opinion, the process was 40
years old, expensive and hard to
control.
It involves molten salt at
2200
degrees F.
Besides
heating carbide to 2200 F isn’t particularly good
for
it. Cobalt is unusual in that it
goes through a phase change form cobalt II to
cobalt III.
I know of two
companies that did buy and install the salt bath
equipment and then later abandoned it.
The Toney process by any name
leaves
holes in the surface of the material.
This is generally explained as
being good because the braze alloy
flows in
to the holes and that’s how it
bonds.
The holes are
microscopic in size and much too small for the
described process to be effective.
The following photographs
were all taken at 1,000x magnification.
The
left-hand photos are
standard photos from a Scanning Electron
Microscope (SEM).
The right
hand photos were taken using Backscatter
Electron Imagining (BSE) which
identifies the elements by shades of
gray.
Cobalt is darker and
tungsten carbide is lighter.
Untreated Tungsten carbide.
Fairly flat, impervious surface in the
SEM on
the left.
Scattered
cobalt and tungsten carbide in the BSE on the
right.
Cobalt is the light
material.