Team MMC Hammer

Composite PartsLab Notebook

(see physical notebook for a more complete description, including diagrams)

Thursday, February 5th, 2004

Project for the day: observe and learn process of making metal matrix composites, as performed by Jess Wannasin.

Materials:

Setup and Procedure:

Tuesday, February 10th, 2004

Method used today for removing MMC from mold:
  1. The mold tube was clamped in the vice so as to remove the crimp.
  2. The Swagelok was removed by holding the mold tube in a vice, clamping the Swagelok with a pair of vice grips pliers, and beating the pliers with a hammer.
  3. The mold tube was cut in a plane perpendicular to its cross-section and tangent to the punch.  This cut was made until it contacted the remaining part of the Swagelok.  The slice of tube without the MMC was then cut off.
  4. A cut was made throug the Swagelok and it was pried off using screwdrivers and a hammer.
  5. The cut made in (3) was continued to the end of the tube.
  6. The punch was pried out using a small screwdriver and a hammer.
  7. The tube was laid with the sliced side up and pounded flat with a hammer, thus releasing the MMC.  Some of the MMC broke out before the rest, so the part of the tube where this MMC had been was also pounded in order to get the rest out.

Tuesday, February 24th, 2004

Entry today is for experiments conducted on 2-16-04 (outside of class) and 2-19-04.
We made two MMCs: one with alumina packed on top of titanium carbide (henceforth referred to as "#1") and the other with silicon carbide on top of titanium carbide ("#2").  The particle sizes are as follows:
The procedure outlined in the entry for 2-5-04 was followed for each.  We removed the MMCs following the procedure outlined on 2-10-04, although exact steps were altered as we saw fit.

Thursday, February 26th, 2004

The following was performed on 2-24-04:
We cut and polished the samples removed on 2-24-04, although we could not polish the silicon carbide part because it broke out of the mold and was thus impossible to polish properly.  The titanium carbide sample was stuck in the tube so we used a slow oil cutter to cut it smoothly so that one end was silicon carbide/tin-lead MMC and the other was pure tin-lead.  We used a 4000 grit (5 micron grain size) polisher to polish it.  We then performed Rockwell B hardness tests on the silicon carbide/tin-lead and pure tin-lead pieces.  The material was too soft for a Rockwell A test.  The data obtained follows:
Material: SiC/tin-lead
Rockwell B hardness values:
87
25.5
97
Material: tin-lead
Rockwell B hardness values:
97
26.5
59
We are skeptical of our results, as there is such a wide range.
We also discussed our project with Professor Chris Schuh.
-----
The following was performed on 2-26-04:
The alumina/titanium carbide sample was removed from the mold and was cut using the slow oil saw, then polished.  Rockwell B hardness tests were performed and the data is as follows:
Material: alumina/tin-lead
Rockwell B hardness:
81
(sample cross-section too small to obtain more measurements)
Material: titanium carbide/tin-lead
Rockwell B hardness:
42
32
87 (not valid - head of indenter pushed in too far due to cracking)
Again, we are somewhat skeptical of our results.  Larger parts should be able to be more accurately analyzed.
We also made several new mold tubes and worked with Toby Bashaw to set up a mold-machining session as well as obtain better tools for cutting open the molds.

Tuesday March 9, 2004



Began Production of third set of parts following Jess's pellet procedure.

Ceramic pellets were packed and put into pipes, but drilling took too
long resulting in a decision to postpone casting.

Two Parts were prepared:
I. SiC 36 Grit (600-850 microns)
II. SiC 120 Grit

One of the swagelok connectors is jammed on the wrong side of the
pipe.  We will have to push it out next time.


Tuesday March 16, 2004


 
Third set of pellet parts prepared on March 9th were casted:

Part I: SiC 36 Grit = 317.3 grams
Part II: SiC 120 Grit = 309.1 grams (foil was added to balance weight)

parts were spun at 2500 rpm setting (we measured 2070 rpm).  More
relfective tape was added to the middle of the armature for better
readings.

Fourth set of pellet parts were prepared and casted:

Part I: Boron Carbide 100 grit - 296.7 grams (marked with one notch)
Part II: SiC 36 grit - 296. 2 grams (no notch)

parts were spun at 2500 rpm setting (we measuring 2070 rpm).

Additional pellets were packed for future casting:
2 pellets of SiC 120 grit
2 pellets of Boron Carbide 100 grit


Meeting with Instructors:

-Continue citing works
-Possible contact: John Centorino in Prof Cima's group
-Investigate Silicon Rubber mold for preform fabrication
-Clarify goals of our project --- redefine success!!!
-Possible project focus: High Packing Fraction of Ceramic
    -Try 1 in pressing dye
    -As particles get smaller, they are dominated by surface pressures
    -Tap density measurements
        - fine powders can't reach as high a tap density
    -possible procedure: measure tap density, then look at a
    possible bimodal distribution of particle sizes and perhaps apply
    isostatic pressure
-Mention other more useful applications
    -What properties are unique in our material?
    -What are these properties good for? (e.g. specific modulus
    for golf club)
    -Ask Roylance for articles about impact materials
    (e.g. sporting goods)
    -What properties can we control?
-Do SEM after spring break


Thursday March 18, 2004



Prepared and casted 5th and 6th sets of pellet parts out of packed
powder sets from March 16th.

Set #5:
Two parts of SiC 120 grit weighing 291.7 grams and 293.4 grams. Foil
was added to balance weight.

Parts were spun at set point 4000 rpm (measured 2860 rpm). This is the
highest rpm we have tested this far!

Set #6:
Two parts of Boron Carbide 100 grit weighing 291.5 g and 293.35
grams. Foil was added to lighted part to balance weight.

Parts were spun at set point of 4000 rpm (measured 2860 rpm).


Tuesday March 30, 2004



Rockwell Hardness Test B (100 kilos) conducted:

SiC 36 grit:
91
87

These values are probably inaccurated because the needle on tester was
revolving too much.  Our material is much too soft for Rockwell
Hardness B!

Rockwell Hardness Test H (60 kilos) conducted:

SiC 36 Grit:
51
35

Tin Lead:
Off the charts soft!

Although we did not get any reliable Rockwell Hardness results, we did
determine from the H test that SiC 36 grit is relatively harder than
pure tin lead.


Thursday April 1, 2004



SEM Imaging: Building 4

Our files are located in MMC Hammer subdirectory of SEM Computer

Sample 1: SiC 36 Grit (sample with visible particles)

We should etch the tin-lead because otherwise we aren't looking at
cloear boundaries becuase the metal can smear when cut!

Next presentation should include Tin-Lead phase diagram.  Are finer
particles pushed away by moving dendrites when the metal solidifies?

Sample 2: SiC 120 grit (non swage-lock section with less shiny side up)
lighter sections are tin lead
Al is present again
Na is present (could be sweat)

BSE:  dark sections are low atomic # or low porosity

SiC120-001 (BSE: identifies SiC particle corrolating with EDS data)
SiC120-1-Se (Secondary picture of SiC120-001 BSE)
SiC120-2-Se (secondary picture of SiC120-2-BSE)
SiC120-2-BSE (SiC particle)

We are seeing SiC particles on the surface of this sample that could
be from the grinding paper!

SEM EDS machine: EDS2000


Goals and Progress: Meeting with Prof. Chiang

issues we have investigated or hope to investigate:
-relative hardness
-images and EDS
-mold for preform
-particle size
-commercial feasibility of process
-binder and epoxies
-new mold
-infiltration speed
-strength of mold at elevated temp
-phase diagram of tin-lead
-etchers
-issues with boron nitride mold release

Possible Goals of Our Project:

1. Materials Route:
How would you predict the properties knowing volume fraction, loading,
etc?

-does it follow any established mixing rule?
-try bimodal distribution of ceramic particles
-project what would happen if we could use Al as matrix

2. Demonstrating capability of the Process
Mold can turn out different shapes, be scaled up, for to different
temperatures

Metrics of success via this route:
-higher infiltration speed
-high pressures can be achieved
-different shapes can be infiltrated - shape complexity
-partial reinforcement can be achieved

To do this route:
-plot of pressure/diameter/length showing limits of process
-machine an insert for the mold to show back-flow
-show you can selectively reinforce a part


Saturday April 3, 2004



Attaching pipes to new mold and cutting runners:

There are 47.3cm between the 2 bolts of our centrifuge

We need more 1/2" piping!
We need to order swageloks!

2.2cm of piping goes into swagelok.
Space between the swagelok is 15.5cm
--> total tubing needed is 19.9cm

Because we were worried about the pipe being too long, we tried
19.5cm, but this ended up actually being too short (by about .75cm).
So, on the next try, we used 20.25cm and this worked perfectly.


Creating mold for preform:

Platsil 71-10 RTV Silicon Rubber from Polytek Development Corp (given
to us by Toby Bashaw).

Part B is pink
Part A is clear

Directions (as described on containers):
"Mix ratio by weight is 10A to 100B.  Accurate weighing is essential
to obtain optimum physical properties from cured rubber.  Combine
proper amounts of A and B in a clear mixing container.  Mix well,
scraping sides and bottom repeatedly.  Pour over a properly prepared
mold or model as soon after mixing as possible.  Demold after 30
minutes are room temperature."

We used clear, plastic 8 oz. drinking cups for mold container.

The following amounts were used:

Part B: 163.5 g (optimal amoung of A = 16.35)
Part A: 18.35 g
--> our mix was 11.3 part A to 100 part B

Next time, we must make a stronger effort to ensure a more accurate
ratio.


First Prototype attempt (with no ceramic particles):

-Block of tin lead was heat in furnace to 350 C. 
-Tin was poured into one side of part. 

Because part was not preheated, it was cold and metal seemed to
solidify about the small hole and not go into the mold very much. The
amount of metal used was not weighed.  We should do theoretical metal
weight calculations.

-Two sides were connected and put into furnace at 350C for 15 minutes
-Casting was attempted, but set-up was put onto bolts crooked and it
wasn't working.


Sunday April 4, 2004



Ed and Kevin made first preform
    -used 120 grit SiC
    -used 16.5 grams prehydrated ethyl silicate binder
    -used 11.0 grams ethyl alcohol
    -used 1.5 grams sodium acetate

mixed binder and alcohol, then added sodium acetate.  Powder was
poured into the mold and solution was poured on top.


Tuesday April 6, 2004



Ed and Jenny made an improved silicone rubber mold by melting wax to
stick polyethylene part to the bottom of an ~10 oz plastic cup.  Wax
was melted by dripping wax from a candle.

This resulted in a more centered mold, though it was still somewhat
crooked.

Part A: 181 grams
Part B: 25 grams (should have been 18 grams)

Kevin and Lizzie made more preforms using same proportions and process
used by Ed and Kevin on 4/4/04.

3 preforms were produced:

Preform 1 (preform two assumed to be same and same amounts added):
Original Mass = 187.5 grams,
Mass of SiC = 113 grams,
binder added = 16.5 grams,
ethyl alcohol added = 11 grams,
Ammonium carbonated added = 1.7 grams

Preform 3:
Original Mass (plus cup) = 295 grams
Mass of SiC = 98 grams
binder added = 15 grams
ethyl alcohol added = 10 grams
Ammonium Carbonate added = 1.5 grams

For #3, we cut a hole out of the bottom of the cup and put mold in and
turned the cup upside down onto pyrex glass beaker (to prevent
leakage).


Thursday April 8, 2004



Creating more Silicon Rubber molds:

For the two half molds:
-we connect screws to polyethylene half mols
-we used hot candle wax to melt red wax to bottom of plastic cup
-we connect the screw/half mold to the red wax

for this pair of molds we used:
~450 grams of part A
~45 grams of Part B



Second Prototype Attempt:

Materials:
Non-sintered SiC 120g preforms, 15 % binder
Tin-lead (22% lead)

-Metal was poured into cold mold halves
-No metal was poured into runner
-1/2 size preforms were used on each side
-After metal was poured, parts were screwed together
-Whole set-up was put into furnace for ~20 minutes
-molds were sprun at 500 rpm for 5 minutes

Parts were difficult to remove! Also, metal leaked into centrifuge! No
permanent damage was caused to centrifuge and metal came out easily.

Hammer tapping was tried to remove preform, but this was unsuccessful.

The parts were NOT infiltrated!  Preform floated.

Friday April 9, 2004



Email from Kevin Mccomber:

"Today Jenny and I got in around 9:30 and did work on the MMC stuff.
Jenny opened the other half of the molding apparatus (the second mold)
and a lot of powder fell out - obviously not infiltrated, as it had
floated to the top.  This means the preform broke in the mold.  We did
get the part out (as well as the one you couldn't get out) and there
appears to be a little infiltration, although the particles are quite
scattered in the metal.

After Jenny left, Jess and I talked and we decided not to make more
preforms but instead to focus on improving the preforms we already
have since making more unsuccessful runs wouldn't be worth our time.
We went to Yin-Lin's llab and put two half-forms in the over to
partially sinter them (for strength).  The I took one of their blocks
of high-temperature porous alumina and brought it to the lab, where I
cut and filed it down to the shape of the half mold (although some
sticks out of the top of the half mold).  Jess figured we could use
this and the sintered piece in a run to make some parts in which the
preform will definitely hold its shape.

This leads to the next problem - when to do the casting.  Jess already
boron-nitrided the inside of the mold so that's all set to go.  We
just need to polish the mold interfaces, melt the metal and do the
spinning."


Email from Jessada Wannasin:

"There are four problems to deal with:

1. The preform.  I think we can use sintered preforms for now.  So,
its ok.

2. Balancing.  I checked your setup and doung that it might not be
balanced.  So, after filling, we will need to add weight and balance
the setup.  This should not be a big problem.

3. Holding bolts in the centrifuge.  As you are aware already, we need
to grind off the bolts to get better clearance.  this must be solved
before the next run.

4.  Leakage.  This is the most important problem.  I suggest lowering
the casting temperature to 250 C.  Then we can use silicon or teflon
o-ring and sealant grease.  Then we should be fine."

Email from Jessada Wannasin:

"I took apart the molds and coated all parts.  I also polished the
surfaces, so please be careful not to introduce any scratches on the
surfaces!"


Sunday April 11, 2004



Email from Kevin Mccomber:

"I just spent some time cutting and filinf down the bolts that hold
the mold on the centrifuge arm.  Now they habe no threads and are
considerably narrower (as well as square) so we shouldn't have a
problem putting the molds on them.  Hopefully, they'll still be strong
enough, but since they're narrower, the Swageloks and the runner will
now be taking most of the centrifugal force."


Monday April 12, 2004



Successful!  Basis of our initial prototype presentation!!!

Solved leakage problems by
-polishing
-using graphite o-rings (more important factor!)

Used:
-SiC 120 grit 1/2 size preform sintered (1500C for 1 hr) with 15%
binder
-Alumina brick 1/2 size preform carved to the correct shape sintered
at unkown condition with no binder

Matrix was 22% lead

Weughed the parts
-Differed by ~50 grams (alumina was lighter)
-Metal rings were used to equilibrate weights and sent the center of
gravity to center of runner.

Parts were filled by putting Alumina and SiC in their respective
molds, screwing the molds together and then pouring metal into small
holes at the top.  The runner had previously been filled with metal.

Molds were connected to runner and heated for ~20 minutes in furnace

Setup was speun at 700rpm for ~5 minutes.

No leakage!  But centrifuge was off balance.

Four hours later, we tried to take apart the molds.  It was difficult
because the metal partially infiltrated the graphite o-ring.  Next
time we must boron nitride the mold mating surfaces as well.  We pried
the halved apart, then used torch to remove excess metal from around
edge.  Then, we hammer-tapped the parts out.

On one mold, the two halves twisted significantly before we could pry
them apart.  This showed that our design, with the small inlet, worked
well.



Tuesday April 13, 2004



No lab session - presentations took up entire period.


Thursday April 14, 2004

Email from Jessada Wannasin:

"A real application ---
Non leaded projectiles can be a real application that you guys can
explore. There are several publications on that such as:
US patent 5,760,331. "Non-lead, environmentally safe projectiles and
method of making the same."
https://www.ms.ornl.gov/researchgroups/SPM/pubs/powdbull/pbullets.html

I know that we only have 3 weeks left.  So, making any changes now
will be difficult.

If you read these articles and find this application to be
interesting, we can think about using pure tin and tungsten powders
soon. Important testings will be hardness and compression tests.

If we choose this application, a few interesting sections that you
guys should add are patent search, cost modeling, market size
analysis.  I think it we find out that nobody has made projectiles
using an infiltration process, this may be patentable.  If it is
economically better than other processes, it could also be
commercialized.  And this will make this project even more
interesting!!!

You guys can discuss among yourself and we can discuss more with the
other staff members tomorrow."

Thurday April 15, 2004

Meeting with instructors:

What will we have demonstrated:

Possible project extension:
Match properties to lead for non-toxic projectile.  Use tungsten
(which wouldn't need to be sinted because its heavy) and pure tin.

-Look at dynamic balancing to help with our unbalanced centrifuge.
-Look for tin/lead etchant
-measure density to get volume fraction ceramic in our parts

Cutting initial prototype:

The regular diamond saw couldn't cut the sample because it was too big
(we wanted to cut it lengthwise).  The band saw in the shop could not
cut through the ceramic.  Kevin will meet with Yinlin to cut the part.

Preform fabrication:

We also made more ammonium carbonate catalyst


.

Tuesday April 20, 2004

Sintered performs made on Thur. 4/15/04
Al-2O3 and _ size SiC 220 grit perform: sintered for ~2 hrs at 1500 deg. C
SiC 120 grit and full size SiC 220 grit perform: sintered for ~ 1h at 1500 deg. C
Note: we meant to sinter everything for the same length of time but there were problems with one of the furnaces

4th prototype run:
metal: Tin 22% lead
performs: _ size SiC 120 grit; one was sintered at 1500 deg. C for 1 hr and other was not sintered

metal was poured in using our new method of pouring metal into one half of each mold and then bolting mold halves together

setup was spun in centrifuge at ~700 rpm for 5 mins

Thursday April 22, 2004

Imaged samples using microscope in Jess's other lab

Kevin cut 4th run prototypes (done on Tues. 4/20/04) with diamond cutter in Yinlin's lab

Lizzie polished samples

Lizzie used 1% natal solution for approximately 30 seconds to etch 3rd run Al2O3 part (SiC 120g and Al2O3 _ size performs)

 Imaging: etched (Al2O3) sample was examined using optical microscopy and revealed grain structure of tin lead

Method for sanding and polishing mold halves:
To allow for easy separation of mold halves, a procedure was developed and implemented to sand and polish mold halves after a run

Tuesday April 27, 2004

5th run of prototypes was done

metal: Tin 22% lead
performs: _ size SiC 120 grit and SiC 220 grit; both had been sintered at 1500 deg. C for one hour

metal was poured by pouring metal into one half of each mold and then bolting mold halves together

setup was spun in centrifuge at ~700-750 rpm for 5 mins

Lizzie attempted to image parts on SEM to see effects of etching but there were problems because SEM thought there was a leak because of porosity in sample and variable pressure mode did not work

 Kevin discussed water jet cutting with Yinlin and found out where it could be done on campus; he made an appointment to cut our samples on Friday

Jenny worked with Yinlin to correct order with Swagelok to get our parts to make our increased metal design Ed put together the new assembly that would allow more metal (enough metal so that whole MMC parts could be produced)

Thursday April 29, 2004:

A perform mold was made to comply with the ASTM specifications for 3 point beam bending tests on metals:

sample: span >= 2.0 inches, span/thickness > 15, width / thickness > 10

to create the object around which the mold was formed:
 strips of black plastic in Jess's small parts cabinet 0.05 inches (50 mils) in thickness were cut to approximately 3 inches long, 1 inch wide

three of these plastic strips were taped together with black tape

note: all dimensions were made to exceed the minimums defined by ASTM so they could be sanded down later as necessary

to create the silicone rubber mold:
 the strip was secured to the bottom of a plastic cup with red wax

part A of silicone rubber: 33 g
part B of silicone rubber: 296 g

note: theoretical ratio is 1:10 but I used a higher ratio because it has been shown to work better in the past

New metal design:
The amount of metal volume in the new design: 5.30 cubic inches
Amount of metal needed:
Vp of particles = 40% Volmetal = 5.16 cubic inches
Vp of particles = 70% Volmetal = 2.60 cubic inches

Thus we should have enough metal

Jenny (and Jess) made new performs in a new method which proved superior

Instead of pouring ceramic powder in mold and then pouring solution of binder in, ceramic powder was stirred with solution and this mixture was then poured into mold

350 g of 220 grit SiC was combined with correct amounts of binder, ethyl alcohol, and ammonium carbonate as specified previously

it was assured that mixture was mixed very well mixture was poured into 2 full size perform molds as well as mold for part for 3 point beam bending test

Friday April 30, 2004

Kevin used a water jet cutter to cut the 5th run prototypes; this took some time to set up but the actual run was only 10 mins and is thus much better than using the diamond cutter which takes more than 2 hours and must be continually monitered

Saturday May 1, 2004

Removed the performs made on Thur. 4/29/04
- seemed stronger than others made with old method
- beam bending perform broke, however

Kevin and Lizzie made new performs using method developed by Jess and Jenny on Thur. 4/29/04

Combined 300 g SiC 220 grit with 45g binder, 30g ethyl alcohol, 4.5 g ammonium carbonate

Poured mixture (well stirred) into 2 whole molds and beam bending mold

Kevin and Lizzie ran 6th set of prototypes - in new assembly
We had tried to run this setup previously a couple times, but it kept freezing up

This time assembly (which already had metal) was preheated at ~330 deg. C for more than half an hour before spinning

Assembly was secured in centrifuge with a metal pipe that came up through hole (cut for this purpose) in lid

Spun at ~750 rpm for 3 mins

Mold half 1: SiC 120 grit
Mold half 2: SiC 220 grit

Tuesday May 4, 2004

Planned characterization of the parts

Setup the backflow run
- bolted the copper disk between the mold halves
- had to use alumina o-rings because we didnt have any graphite left

Researched Brinell Hardness measurement techniques

Estimated the Brinell Hardness values

Polished previously cut samples

Cleaved samples (#5) so that we had fracture surfaces to image

Wednesday May 5, 2004

Ran the backflow
- metal leakage, likely due to alumina o-rings being less effective than traditional graphite ones

Thursday May 6, 2004

Took apart parts designed to show backflow
- this was difficult due to metal leakage and bolts catching on the copper plate

Used SEM to image samples that we cleaved on Tuesday

Designed Brinell Hardness measurements

Took more optical images for sample (#5 with SiC 220 grit) so that we could determine the ceramic volume fraction

Friday May 7, 2004, Monday May 10, 2004

Attempted Brinell Hardness

Worked on and completed poster

Tuesday May 11, 2004

Did Brinell Hardness measurements
  • used Intstron compression tester
  • made 3 indentations on each of two MMC samples
  • also made an indentation on pure tin-lead
  • measured the size of the indentations using optical microscopy

    • Wrote up final presentation slides