blogfast25
International Hazard
Posts: 10562
Registered: 3-2-2008
Location: Neverland
Member Is Offline
Mood: No Mood
|
|
Titanium metal from TiO2 thermite reaction
It's with some pride that I announce the fact that I've produced titanium metal from a homemade chlorate-boosted Anatase thermite...
Earlier, I had developed a sulfur-free chlorate boosted formulation for SiO<sub>2</sub> themites. You can find the rationale for this development on this forum, here (scroll down a little).
A couple of weeks ago I had adjusted the SiO<sub>2</sub>/Al/KClO<sub>3</sub>/CaF<sub>2</sub> = 100/96/81/55
formulation to TiO<sub>2</sub>/Al/KClO<sub>3</sub>/CaF<sub>2</sub> = 100/72/61/47 and tried it with my home brewed
'emulsion paint' TiO<sub>2</sub> and it burned but somewhat sputteringly and of course no metal. The homemade TiO<sub>2</sub>
undoubtedly contained residues of paint resin and possibly other white pigments. Without concentrated sulfuric acid, it is difficult to purify it.
But I recently bought a good grade of pure TiO<sub>2</sub>, Anatase "99 %", so I decided to stick that in there too, as well as in an
adjusted formulation with magnalium. The total charge in both cases was 20 g, contained in an egg cup and embedded in a sand-filled steel bucket.
Materials used: very fine Anatase 99 %, 400 mesh Al powder, potassium chlorate reagent grade, ground Fluorite for CaF<sub>2</sub>.
Ignition with a stoichiometric mic of Al/KClO<sub>3</sub> mix and an Mg ribbon fuse.
The Anatase formulation burns like hell: very fast and furious, very regular: it's completely on a par with its SiO<sub>2</sub> analog.
Immediately after the reaction had finished I went over to look and saw three quite large, darker regions at the bottom of the white/yellow hot
crucible (eggcup). Darker regions in the hot slag are an indication of materials with higher heat conductivity (such as metals), I've seen this happen
many times before.
But at that point I didn't even dare to hope too much and decided to light the magnalium version
(TiO<sub>2</sub>/MgAl/KClO<sub>3</sub>/CaF<sub>2</sub> = 100/82/61/49). That basically exploded! I knew it was
going to be fast, possibly too fast, so I ran away after lighting the Mg ribbon and the whole thing went "poof!" while I still had my back to it,
throwing content as far as a meter away from the test point!
But after the straight Al test had cooled down it became clear that the darker regions were indeed titanium metal: oxidised (only slightly) on the
outside on the top and lightly covered in dark slag at the bottom, a little buffing up with rough sanding paper revealed the hard, shiny metal. It has
a slight golden tinge to it.
Interestingly, slag/metal separation was simply exemplary: most of the metal neatly at the bottom, easily separable from the slag mix. It appears to
me a lot of the slag had been blown off, covering the globules only in a slight oxide/slag coating. The metal had also passivated, like Al does. Three
4 -6 mm blobs, total weight about 1.9 g.
Tonight I'll be running a larger test to confirm.
It may also prove difficult to get chemical confirmation of the metal's identity: Ti only really dissolves in concentrated acids like sulfuric or
nitric, also apparently HF. Nice one... I haven't got any of those... 
|
|
|
ScienceGeek
Hazard to Others
Posts: 151
Registered: 22-1-2008
Location: Norway
Member Is Offline
|
|
You can try burning the metal! It doesn't take more than ca 600 degrees Celsius, and you can easily light a thin strip of the metal with a torch.
Also, try burning it in pure Nitrogen! Titanium is one of the few metals that will burn in pure Nitrogen (if I remember correctly)
|
|
|
microcosmicus
Hazard to Others
Posts: 287
Registered: 31-12-2007
Member Is Offline
Mood: spin up
|
|
HCl will also dissolve Ti if you are patient. To test fot Ti, put some filings of your
metal in HCl, give them time to dissolve, then add H2O2 --- if your solution
turns from purple to orange, that indicates the presence of Ti. The peroxide
complex is considered a good test for Ti. For details and pictures of the
reaction, see W. Oelen's website:
http://woelen.homescience.net/science/chem/exps/titanium/ind...
Edit(woelen): Made link to website work again.
[Edited on 30-7-16 by woelen]
|
|
|
blogfast25
International Hazard
Posts: 10562
Registered: 3-2-2008
Location: Neverland
Member Is Offline
Mood: No Mood
|
|
Yes to SG and MC: I'm doing that now.
Another test late this afternoon using the same formulation but this time with a larger batch size (92 g) also yielded good metal, including one
blob of 11.5 g, the thickness of a pound coin but larger in surface. Some smaller globules were spewed out of the crucible, so it's probably
running a little too hot. I may have to cut down on the booster mix a little. The slag was also easy to break down. 
[Edited on 14-3-2008 by blogfast25]
|
|
|
mrjeffy321
Hazard to Others
Posts: 149
Registered: 11-6-2005
Member Is Offline
|
|
I was successful in my attempt to recreate your experiment.
I performed two reactions, one with 20 grams of thermite and one of 100 grams of thermite. The 20 gram reaction yielded a lot of slag material, but I
did find a few small globules of metal in the mixture. The 100 gram thermite yielded a much larger amount of metal which did separate out along the
bottom of the reaction vessel. I tested the metal using “microcosmicus”s suggestion with the HCl and H2O2 and it did yield a positive result for
Titanium (orange solution).
|
|
|
microcosmicus
Hazard to Others
Posts: 287
Registered: 31-12-2007
Member Is Offline
Mood: spin up
|
|
Duh, there is an even more obvious physical test --- melting point. Since Ti melts at 1688C
but Al at 660 C, it should be easy enough to tell them apart. If you have a decent amount
of Ti in your metal, it should stay solid when you heat it to orange or yellow heat. (Heating a
blob as opposed to a strip since the point of this exercise is not to burn the metal, as opposed
to the test ScienceGeek suggested.)
Also, it would be interesting to know how much aluminum is alloyed with your titanium. I
am guessing that there is not going to be calcim or potassium in there because their
reduction potentials are much higher than aluminum or even magnesium. You could
perhaps assay this in the Archimedean fashion --- Al has a specific gravity of 2.7 whilst
Ti has a specific gravity of 4.5.
By the way, aluminum in there sounds desirable ---- many titanium alloys contain a few
percent aluminum. Wikipedia has nice list of allloy compositions:
http://en.wikipedia.org/wiki/Titanium_alloy
Now that you have figured out a way of making titatnium via a thermite reaction (congratulations !)
it would be interesting to see if one could produce different alloys by
adjusting the
proportions of ingredients and adding alloying agents either in their
reduced or oxidized forms.
Given that I have most of the ingredients, I will have to try my had at
this once I run down to the
local neighborhood ceramic store to pick up some fluorospar.
[Edited on 14-3-2008 by microcosmicus]
|
|
|
12AX7
Post Harlot
Posts: 4803
Registered: 8-3-2005
Location: oscillating
Member Is Offline
Mood: informative
|
|
Ductile titanium has less than 0.1% total O, C, N, B, etc. I don't see this method ever producing viable material, but it is remarkable it makes
something metallic!
Tim
|
|
|
mrjeffy321
Hazard to Others
Posts: 149
Registered: 11-6-2005
Member Is Offline
|
|
I might be able to run a few analyses on the Ti metal I produced using some fairly sophisticated instruments. If all goes according to plan, I can
get a much better estimate of the ratio of Ti to Al than what could be gained by measuring specific gravities.
I will try to do a scaled up reaction this weekend and get a big piece of TiO2 to do the measurements on.
|
|
|
microcosmicus
Hazard to Others
Posts: 287
Registered: 31-12-2007
Member Is Offline
Mood: spin up
|
|
To be ductile, any metal is going top have to be relatively free of impurities which
could mess up the crystal lattice. However, lack of ductility is not necessarily a
showstopper --- just think of cast iron, which is chock full of C, Si, etc, but still a
useful engineering material.
As for possible uses, thermite welding of titanium is one possibility. Also, it
might work for casting. To be sure, this may not be of interest to industry,
quite likely won't be relevant for aerospace use, but it may be of interest
to the hobbyist. For instance, it might be of use for someone who does not have
a furnace capable of reaching 1700C as a means of melting titanium to
cast small parts.
|
|
|
chloric1
International Hazard
Posts: 1070
Registered: 8-10-2003
Location: GroupVII of the periodic table
Member Is Offline
Mood: Stoichiometrically Balanced
|
|
This is something from Theodore Greys column in the How2.0 section of popular science (January 2008?). He chemically identifies titanium by the color
and character of its sparks! By simple holding a grinding wheel to your product, you will observe abundant long white hot sparks. I was dazzled
first time I tried it with my little Dremel tool. I have never seen anything like it! I call in "on demand 4th of July".
Fellow molecular manipulator
|
|
|
blogfast25
International Hazard
Posts: 10562
Registered: 3-2-2008
Location: Neverland
Member Is Offline
Mood: No Mood
|
|
My metal tests positive for Ti too, using Woelen's peroxide test.
I'm convinced it's possible to make quite high purity metal with thermites: they do so with V2O5/Ca reductions for Vanadium.
++++++
The metal does react slowly with 32 w% HCl: it would take forever to make some reasonable quantity of Ti [+III] or Ti [+IV] salts this way but for
simple testing it works.
It also reacts slowly with hot 50 w% NaOH, with bubble formation, again too slow to be useful but it confirms its amphoteric character.
[Edited on 15-3-2008 by blogfast25]
|
|
|
microcosmicus
Hazard to Others
Posts: 287
Registered: 31-12-2007
Member Is Offline
Mood: spin up
|
|
Also note that V is a common alloying agent for TI along with Al, so maybe one
could consider adding some V2O5 to the mix.
As for speeding up the reaction with HCl, maybe apply electricity. To make titanium
salts, one could either try electrolysis or the old method of fusing the TiO2 with
an alkali to make the titanium equivalent of waterglass, which can then be
dissolved and used in aqueous chemistry.
As for impurities, we can start with conservation --- only stuff in the initial mix
or its immediate surroundings (crucible and atmosphere) can wind up in
the final product. Assuming that one starts with pure material and uses
a suitalbe crucible, that rules out most of the periodic table. In particular,
the B and C impurities mentioned above should not be a problem and
one can keep the N down by not having air blowing all over the place ---
for example, have the reaction going on in an enclosed retort in which the
air is leaving as the mixture heats up and no new air is coming in--- at 1000C,
we have 1E-5 moles N2 per cc as opposed to 0.1 moles for a cc of Ti.
Thus, any N contamination should around or below the part-per-thousand level.
especially since I would expect most of the TiN top be on the outside of the
fused mass, where it would be removed with the slag.
Assuming a reasonable choice of crucible (such as alumina),
the only contaminants left to consider are Al, Ca, K,. O, Cl, F. As mentioned
earlier, one can choose to regard Al as an alloying agent than a contaminant.
Ca, K are not going to be reduced by Al so I would only
expect them to appear in the form of compounds which got frozen into
the metal before they could make it into the slag.
Ditto for O, which I would only expect as part of an oxide.
Thus, the upshot seems to be that the only contamination to expect would
be in the form of crud which did not make it to the slag layer or
halogens making trouble. If this turns out not to be the case, then
this method should be capable of making a rather good metal.
At least the oxides stuck in the metal should be easy enough to test
for --- digest the stuff with acid and see if there is any residue of
undissolved dust left over. Alternatively, as suggested above,
use the metal as an electrode in the acid solution and electroplate
away the titanium, then see if insoluble residue falls out.
[Edited on 15-3-2008 by microcosmicus]
|
|
|
blogfast25
International Hazard
Posts: 10562
Registered: 3-2-2008
Location: Neverland
Member Is Offline
Mood: No Mood
|
|
The yield of the 92 g reaction was about 60 %, for the 20 g one, 47 %.
Cleaning up the metal is proving a little hard: sand paper does it of course but that isn't practical. The fine slag coating (mainly alumina) proves
to resist better to dissolving than the metal itself.
37 w% nitric doesn't seem to attack the metal at all.
There's a quite simple titrometric method to determine Ti (Ti<sup>3+</sup> titrated with Fe<sup>3+</sup>,
SCN<sup>-</sup> as indicator) but I need to dissolve the metal first...
I might try the electrolytic method. Otherwise, it's a search for conc. sulfuric acid...
|
|
|
blogfast25
International Hazard
Posts: 10562
Registered: 3-2-2008
Location: Neverland
Member Is Offline
Mood: No Mood
|
|
Another test with 120 g of mix and and less booster: TiO<sub>2</sub>/Al/KClO<sub>3</sub>/CaF<sub>2</sub> =
100/63/40.5/40.7
This burned well too and metal was formed but the metal/slag separation is poorer and the slag is extremely hard to get off the metal. Clearly this
reaction ran a few degrees cooler.
I'm pretty sure much of the smaller metal globules have been re-oxidised: the slag is full of protrusions of a whitish/grey material. Ideally you'd
run this reaction under argon to protect the metal against O<sub>2</sub> or N<sub>2</sub>. Small parts of the metal also have
a golden tinge.
For now, I'll have to play around with the flux (CaF<sub>2</sub>
quantity, to try and protect the metal as best as possible.
Later on I'll be running a mixed oxide thermite CuO/TiO<sub>2</sub>, no booster, CuO should be energetic enough.
++++++++
The CuO/TiO<sub>2</sub> mixed thermite burned well too but slower than a straight CuO. Target composition of the alloy Cu/Ti = 60/40,
formulation CuO/TiO<sub>2</sub>/Al = 75.1/66.7/47, plus 10 w% CaF<sub>2</sub> added.
Plenty of metal, although metal/slag separation was quite poor, this would definitely need 10 - 20 % booster mix to increase temperature.
The metal very much looks like a bronze: a pale golden yellow. It's not attacked significantly by 37 w% HNO<sub>3</sub> at RT, but it does
start dissolving on heating. But on cooling it stops dissolving again, unlike straight copper which just dissolves all the way. Will try and test for
Ti but with this poor a slag separation, this might not be all that indicative because the Ti could come from the slag, not the presumed alloy.
And then I ran a straight TiO<sub>2</sub> with magnalium, w/o boost (TiO<sub>2</sub>/MgAl = 100/52). That burned through all
the way but formed a strange slag and no metal.
The outside of the slag is white and looks like fine magnesia, then, digging deeper it became dark grey - black, and rusty brown in the centre. The
slag was porous and clearly no or very little melting had taken place. Definitely needs boosting.
[Edited on 16-3-2008 by blogfast25]
|
|
|
mrjeffy321
Hazard to Others
Posts: 149
Registered: 11-6-2005
Member Is Offline
|
|
I performed a 530 gram KClO3-boosted TiO2 thermite reaction using a ratio between TiO2 : Al : KClO3 : CaF2 of 100 : 75 : 50 : 40 by mass
respectively. The thermite mixture was placed inside a small flower pot which was, in turn, placed inside a larger flow pot and completely insulated
all around (except top) with sand. The reaction was easy to ignite and proceeded quickly (only ~30 seconds of actual ‘burn time’).
After the remnants of the reaction had cooled and hardened I began extracting out the Titanium metal chunks which were produced. The Titanium metal
was not found all in a single lump, but rather and multiple smaller lumps scattered in the Aluminum Oxide slag. The Ti-metal lumps were a decent size
(not small little balls like I had found in previous reactions) and varied in diameter from roughly that of a U.S. dime to quarter, although larger
and smaller ones did exist. The Titanium metal was difficult to separate completely from the Al2O3 slag with just a pair of pliers and a hammer, so
most of the Titanium I extract still has some Aluminum Oxide still on it making it difficult to measure the mass of the Titanium produced and the
thermite’s yield.
I used a wire brush and sanding pads mounted on an electric drill to clean up the Titanium chunks (remove oxide coating and expose the shiny metal
underneath, as well as try to remove some Al2O3). I was able to clean several of the Titanium pieces up nicely and put a shiny finish on a few of
them.
In the near future, I will try to analyze the metal content percentages of the Titanium produced to see how much Aluminum (or other impurities) might
be alloyed with it.
I took numerous pictures (and a video of the reaction). Here is a sample to give you an idea of the Titanium chunks produced,
http://i62.photobucket.com/albums/h104/mrjeffy321/ScienceMad...
http://i62.photobucket.com/albums/h104/mrjeffy321/ScienceMad...
|
|
|
blogfast25
International Hazard
Posts: 10562
Registered: 3-2-2008
Location: Neverland
Member Is Offline
Mood: No Mood
|
|
mrjeffy321:
Thanks for confirming my results.
Did you use Anatase or Rutile?
Look forward to the results of analysis! 
|
|
|
mrjeffy321
Hazard to Others
Posts: 149
Registered: 11-6-2005
Member Is Offline
|
|
| Quote: | Originally posted by blogfast25
Did you use Anatase or Rutile?
|
I am not certain, but I think it was probably rutile since I used standard pottery-grade TiO2.
|
|
|
blogfast25
International Hazard
Posts: 10562
Registered: 3-2-2008
Location: Neverland
Member Is Offline
Mood: No Mood
|
|
Yep, that's Rutile, 99 % certain. It's also the cheaper of the two...
|
|
|
![[*]](images/xpblue/default_icon.gif){kind=icon}
