Sciencemadness Discussion Board - Make Potassium (from versuchschemie.de)

I'd like to report that the "crust" i got from the paraffin solvent run has a gravel like consistency. It forms large chunks that break up when poked but the smaller pieces themselves are rock-hard.

So there is some condition that favors the formation of crusts vs. Sand.

[Edited on 20-12-2010 by NurdRage]

In both my tests I obtained extremely fine ‘sand’ (clearly the majority of the reaction by-product) but also some hard, white chunks. I thought they might be residual KOH but I think now I may be wrong on that. I’ll see if I can investigate…

woelen - 20-12-2010 at 11:42

@len1: Your remarks have been incorporated into the webpage and I have modified the equations. When MgO is formed instead of Mg(OH)2, then there is no excess of magnesium anymore, just have a look at my analysis. Taking into account the loss of K through the addition of the t-butanol, I find a yield of isolated K-metal between 75% and 80%, which is acceptable. Please have a look at the end of the web page (for your convenience I post the same link here again).

http://woelen.homescience.net/science/chem/exps/synthesis_K/...

The reaction is somewhat inefficient, because half of the magnesium is used simply for making H2, which escapes from the system and also a considerable part of the magnesium is used up in the initial phase where water from the KOH reacts with the magnesium. But even with these inefficiencies it remains a VERY NICE find. One can almost have a gram for gram conversion of magnesium to potassium and for the home chemists that of course is great fun!

@Jor: I now cleaned up the erlenmeyer and all the grey gunk. The glass of the erlenmeyer is not etched at all! I first added the Mg, then the KOH. The KOH hence was lying in a bed of Mg-powder and did not touch the glass.

[Edited on 20-12-10 by woelen]

Sedit - 20-12-2010 at 12:35

develop a work up that can coalesce the crude, formed metal really quickly, less than 1/2 hour or so... I rinsed the latest batch with kerosene and it's all potassium, once the byproduct has been washed away.

[Edited on 19-12-2010 by blogfast25]

Wasn't this very thing mentioned in the patent on how to purify the resulting Potassium? They mention the use of Dioxane to clean the Potassium and cause it to rise from the reaction mixture IIRC. I feel this was employed due to its density and if my theory is correct then any non reactive solvent with a density greater then that of Potassium and boiling point above Potassiums melting point may work.

[Edited on 20-12-2010 by Sedit]

vulture - 20-12-2010 at 13:22

Apparently purification of potassium metal should be described in Brauer. I wonder, if Mg(OtBu)2 really is the crux of this reaction if it would be worthwile to synthesize and purify it first. It might also have other uses in organic chemistry.
It's even commercially available from Aldrich, but it's rather expensive.

[Edited on 20-12-2010 by vulture]

blogfast25 - 20-12-2010 at 13:49

Since as I was a little impatient and ended up with two batches of 1 – 2 mm K balls instead of a few larger ones, I tried Garage Chemist’s isopropanol (IPA) method (see above), to see if I can get some coalescence in kerosene at about 100C.

I have no IPA on my shelves and used salted out windscreen de-icer (aerosol grade): it may still contain some water and was slightly cloudy (salt?) I’ll distil it tomorrow (or get some rubbing alcohol).

I pipetted some 50 K globules into a test tube and replaced the supernatant solvent with a few charges of kerosene so it was clear enough to observe and then heated the tube au bain marie. Nothing happened up to that point and then I dropped a couple of drops of the IPA into it: the balls started ‘dancing’ up and down and up and down while also becoming very silvery. And there was coalescence, albeit slowly. Perhaps there was still some water in the IPA: some dross also formed (KOH or K 2-propanoxyde?)

After about half an hour I stopped the test and noticed something unusual: while cooling the tube and swirling it in cold water much coalescence took place all AT ONCE, with nearly all the balls coming together in one rather oddly shaped ball. This must have happened when the K was in the process of solidifying.

Beneath (another really bad shot – focus on background, not object! How do I do it??) after replacing the cloudy kerosene and remelting/resolidifying:

I wonder if this may be the basis for a shorter procedure: cook for about 1 ½ hour, replace dirty Shellsol with another clean solvent and coalesce with IPA?

[Edited on 20-12-2010 by blogfast25]

len1 - 20-12-2010 at 17:30

I had thought Mg would still be in excess on an MgO basis given I found it was in more than 2-fold excess on an Mg(OH)2 basis further up this thread. Ill have to recheck that. Thanks Wilco.

Yes I am sorry, on an MgO basis there is a 0.3gm KOH deficit in the way I mixed the reagents - or almost stoichiometric amounts, so the yield is about the same with respect to both Mg and KOH

6.4 Mg to 12.6 KOH

Mg + KOH -> MgO + K + 1/2 H2

24 39

Mg + 2H2O

24 36

0.84 1.26

(6.4-0.84)*56/24 = 12.97

[Edited on 21-12-2010 by len1]

condennnsa - 21-12-2010 at 01:10

I've been going over the patent and example 3 caught my eye:

In the reactor described in Example 1, a mixture of 74.8 g potassium tertiary amylate (0.59 mole), 7.3 magnesium chips (0.3 mole) and 400 ml decalin is heated with stirring at 150° C. for 3 hours. A regulus consisting of 15.1 g (0.38 mole) potassium is recovered from the cold reaction product mixture. Yield: 65.5%.

In this example I notice they use 0.3 moles of mg and end up with 0.38 moles of Potassium. How's that possible?
And also, where does the resulting t-amyl alcohol go from this reaction? Does it end up as magnesium tertiary-amylate?, thus the products of this being mg t-amylate rather than MgO?

woelen - 21-12-2010 at 02:37

What they did in the experiment is add 1 mole of Mg for each 2 moles of potassium tert amylate. Theoretically the reaction should go like this:

Mg + 2 t-PeOK --> 2K + (t-PeO)2Mg (here t-Pe stands for the tert-pentyl group)

Probably the patent writers had the above reaction in mind.

In this situation, I also cannot see any other possible reaction. There is no KOH in the mix, nor water, so there is no option of producing hydrogen, as opposed to the reaction we have done with t-BuOH and KOH.

So, indeed, I see no mechanism of formation of MgO, simply because there is no hydroxide in the system. There also is no water in the system, which indirectly could lead to formation of hydroxide and oxide. Decalin is just the solvent, it is an inert cycloalkane with a boiling point just below 200 C.

So, in this reaction, 1 mole of Mg can form 2 moles of potassium (theoretically). In the original reaction from Pok, one mole of Mg forms one mole of potassium and one mole of hydrogen atoms (hence half a mole of hydrogen molecules H2).

Still, Pok's original reaction is more interesting for the home chemist, because tert-butanol (or tert-amylalcohol) plus KOH is easier to obtain than the potassium salt of these alcohols.

For people who cannot find Shellsol D70, the use of decalin may indeed be an option. It's boiling point may be a tad low, but it comes quite close to the lower part of the boiling range of Shellsol D70.

len1 - 21-12-2010 at 03:20

@woelen If I read your site correctly, you stirred only when adding butanol and still got 85% yield - is that correct?

Another interesting observation I made today while trying out various ideas. You can simmer the mixture all you like,but until the addition of alcohol you only see flurries of Mg flakes thrown up to the surface of the D70 by random H2 bubbles coming out of the reagents on the bottom of the flask. But within 5mins of adding the alcohol tiny balls of potassium start darting on the surface. There seems to be some force attracting them to each other as they never separate by far, and bump into each other many times until they coalesce.

I wonder if tiliting the flask slightly will help coalesce the potassium by graviating it all towards one part of the flask.

[Edited on 21-12-2010 by len1]

woelen - 21-12-2010 at 05:52

Yes, I only swirled the flask somewhat when adding the tert-butanol, I would not even call it stirring, just gently rocking it forward and backward (remember, the flask had the big cooler on it, the cooler was loosely attached to a clamp and it was very hot, so real stirring and swirling is not an option).

I doubt whether tilting really helps. The crud also moves to the lower part of the flask and forms a horizontal bed again. One thing which might help is the use of a spherical flask or even a pear-shaped with the 'sharp' side pointing downwards. This, however, might introduce another problem, because it might be difficult for the tert-butanol to reach the lowest parts of the pear-shaped flask.

Actually, having the K-metal coalesce into a single big ball is not really advantageous. This may cause difficulties taking it out of the flask. I already had some difficulty getting the bigger balls out, because there hardly was room for the balls and the thickness of the spatula in which the balls was laying while being moved out of the flask.

My preferred size would be balls of 4 to 5 mm diameter. A single ball then is ideal for experiments, no further cutting needed.

blogfast25 - 21-12-2010 at 06:42

Quote: Originally posted by len1

I wonder if tiliting the flask slightly will help coalesce the potassium by graviating it all towards one part of the flask.

[Edited on 21-12-2010 by len1]

That's why I mentioned much earlier to maybe use a round flask. Or even one of these with a conical bottom... I really think it might help just a little to keep the reagents together and promote coalescence.

In the mean time I'm still surprised that coalecsence of liquid K is so hard to achieve (takes so long): mercury it ain't!

[Edited on 21-12-2010 by blogfast25]

Jor - 21-12-2010 at 06:43

I will do the experiment in January. I have been thinking about stirring it slowly with a Teflon stir bar on the magnetic stirrer, but I am afraid that at the temperatures involved, the Teflon might decompose or even dissolve in the very hot oil. These decomposition products are very harmful, so I'd like to avoid it. How significant is decomposition at 220C for 3 hours?

[Edited on 21-12-2010 by Jor]

blogfast25 - 21-12-2010 at 06:45

Quote: Originally posted by woelen

Actually, having the K-metal coalesce into a single big ball is not really advantageous. This may cause difficulties taking it out of the flask. I already had some difficulty getting the bigger balls out, because there hardly was room for the balls and the thickness of the spatula in which the balls was laying while being moved out of the flask.

Why not use a wooden or clean SS skewer (a lomg needle): prick into the ball and lift it out?

bbartlog - 21-12-2010 at 06:56

Quote:

I have been thinking about stirring it slowly with a Teflon stir bar

Is Teflon compatible with molten alkali metal? I know that at some temperature there is a violent reaction, as the fluorine in the PTFE combines with the alkali. I think in this case you would be better off with occasional manual swirling.

woelen - 21-12-2010 at 06:58

Quote: Originally posted by Jor

I will do the experiment in January. I have been thinking about stirring it slowly with a Teflon stir bar on the magnetic stirrer, but I am afraid that at the temperatures involved, the Teflon might decompose or even dissolve in the very hot oil. These decomposition products are very harmful, so I'd like to avoid it. How significant is decomposition at 220C for 3 hours?

Don't make it too fancy and don't risk your equipment. If you want to introduce a new variable, then try indeed with a round-bottom flask or a pear-shaped flask and see whether that works.

I personally think that stirring is not good, just some simmering assures that there is circulation of liquid and the simmering causes globules of metal to move around and eventually coalesce, but the process is slow.

I also think that it is not that interesting to investigate the results with all kinds of mechanical agitation and methods of stirring. Whether it coalesces in 2 hours, 3 hours or 4 hours is not that relevant, as long as it does coalesce.

More interesting is to find out whether other chemicals also can be used, such as amyl alcohols or other solvents like decalin. The fact that up to now only Shellsol D70 seems to give good results makes me feel still somewhat uncomfortable. Shellsol D70 is not a precisely described chemical entity, it just is a brand of solvent. If Shell for whatever reason decides to stop selling this stuff, then there is a problem. It may even be that 5 years from now the material marketed under the name 'Shellsol D70' has very different properties from what it has now. So, a better thing would be if we find a well-defined chemical entity which can be used as solvent and which is readily available for the general public.

S.C. Wack - 21-12-2010 at 07:12

Quote: Originally posted by Jor

I have been thinking about stirring it slowly with a Teflon stir bar

I also recall warnings about Teflon (or any other halocarbon) with K or the Na alloy.

NurdRage - 21-12-2010 at 07:22

Quote: Originally posted by woelen

More interesting is to find out whether other chemicals also can be used, such as amyl alcohols or other solvents like decalin. The fact that up to now only Shellsol D70 seems to give good results makes me feel still somewhat uncomfortable. Shellsol D70 is not a precisely described chemical entity, it just is a brand of solvent. If Shell for whatever reason decides to stop selling this stuff, then there is a problem. It may even be that 5 years from now the material marketed under the name 'Shellsol D70' has very different properties from what it has now. So, a better thing would be if we find a well-defined chemical entity which can be used as solvent and which is readily available for the general public.

I have already reported t-amyl alcohol is usable as well as candle wax and paraffin oils.

[Edited on 21-12-2010 by NurdRage]

woelen - 21-12-2010 at 08:36

Yes, you're right. I missed the report of 5 mm sized potassium balls, I only read about the 1 ... 3 mm globules, which to my taste is too much of a hassle to isolate. But if you get most of the metal in 5 mm balls using other mineral oils than Shellsol D70, then indeed that is good news.

I think we should try to summarize all things up about this process and see whether we can make a compact thread with all kinds of recommendations.

Our experiences also should be coalesced into a single ball of useful information

. I'll think about this and carefully read all of this thread again and see if I can make a start with this one of the next upcoming days.

Jor - 21-12-2010 at 08:52

But why do you really want to have small pieces? They will be harder to store for a longer time as they will more quickly oxidise.
And if it really is impossible to get it out of the flask, you can always try to cut it with a spatula inside the flask.
But I agree it is a very convenient size to use in experiments, and you won't have to cut it wich may be somewhat dangerous when the material gets oxidised.

But it seems the solvent is not really critical, as long as it contains alkanes/cycloalkanes and has a high boiling point. I think the viscosity determines the size of the balls.

As quite some people are about to have potassium, it may be a good idea to note that it somewhat problematic to store, because as seen int he link by Formatik, it forms a crust of KO2 and KOH.H2O, wich can lead to a fire when the material is cut. I would personally never cut it under a flammable solvent, because the damage done when something goes wrong is much larger than when something goes wrong when you cut K in the air, only leading to ignition of the K and not the solvent.
The KOH.H2O is problematic as when you melt or heat the K (also when you cut?) it reacts with the potassium, and ignites. The KO2 is less of a problem I think. So I will store it in my dessicator, maybe inside another container, containing a sand wetted with a solution of hydroquinone in dilute NaOH, to keep away oxygen.

a_bab - 21-12-2010 at 09:51

I have mine stored "classically" in parafin oil (boiled before to eliminate the water). It stores very well. I have some pieces that must be 20-30 years old and they only have a very thin dark blue, almost black crust. I never saw the red oxides some would report. It is dangerous to handle however since a 1 gram cleaned piece left in air it will eventually catch fire. Na never does this; it just slowly "dies away".

I personally find sodium is more troblesome to store and Li even worse (for Li the best way is an airtight container, with no oil - ideally flushed with Ar as it happily reacts with N).

Coalescence experiment…

blogfast25 - 21-12-2010 at 13:55

Based again on garage chemist’s coalescence of K with the help of some isopropanol (IPA) I tested this idea again. I gathered all my small balls of K (including the one larger one) of both previous experiments on a suitable SS sieve, thereby getting rid of most of the MgO ‘mud’ (but not the crusty bits) and losing some of the fine K and transferred them into a 100 ml round flask and rinsed them with a few more washes of clean kerosene (lamp oil).

The round flask was then mounted as follows on a water bath:

(I know: ‘hot water + potassium + kerosene + open flame = not a great idea’ but armed with the knowledge of mild danger and a CO2 extinguisher I took my chances).

On reaching 70C inside the flask a few drops of (the now distilled IPA) were added and coalescence started but really quite slowly, I doubt if it coalesced much faster than the long reflux times needed in pok’s, len1’s and woelen’s experiments. So I quit after about 20 mins…

As indicated above, I had noticed considerable agglomeration with the test tube test when it went through the melting point of K and decided to test this again. So I dismantled the set up and allowed the round flask to cool down, swirling gently and monitoring the temperature carefully. And yes! Going past the 68C MP of K, probably about 25 – 50 % percent of the K coalesced into 4 – 5 mm balls all at once. And then again and again: by toggling the flask temperature between 70 and 65 C, each time on cooling the balls grew and more large ones appeared. It fully confirms the earlier observation. Do the balls somehow get ‘sticky’ during solidification? Are ‘seed crystals’ on the surface of the globules promoting ‘sticky collisions’? It appears thus…

This potentially paves the way to a much reduced total time: use a round flask and cook (1 ½ - 2 hours) until you’ve got 1 – 2 mm balls or larger. Allow to cool and toggle between 65 and 70C for coalescence. Alternatively, decant off dirty solvent first, wash with a few aliquots of clean solvent and toggle between 65 and 70C. IPA may be optional but won’t harm either…

Pix tomorrow, my balls are now sleeping...

[Edited on 21-12-2010 by blogfast25]

blogfast25 - 21-12-2010 at 13:59

PTFE + hot K = danger:

(CF2)n + 2n K === > n C + 2n KF + much heat!

No need for stirrer bars. Or for stirring...

len1 - 21-12-2010 at 14:52

Yes, thats what I was saying up the thread, that you can coalesce potassium a few degrees within its mp, above that it is just too fluid. But I dont need isopropanol for that, it just happens with straight potassium in D70.

Wilco, Ive also tried your method with no stirring except when adding the butanol, but this time I got a load of potassium shot, much like blogfast was getting (a very dangerous mix). Maybe coalescence in this case is a random event, or perhaps its a function of the particle size of magnesium.

[Edited on 21-12-2010 by len1]

mr.crow - 21-12-2010 at 15:33

Holy Cow blogfast25, you're lucky your "pyrex" measuring cup didn't explode.

regarding teflon, K metal seems like one of the only things that can pry Fluorine away from its carbonaceous slumber

rrkss - 22-12-2010 at 00:00

Quote: Originally posted by Nicodem

Higher t-alkoxides (t-amyl alcohol and above) might speed up the reaction if the limited solubility of t-butoxides is found to be an issue.

On this point i'd like to add my purely qualitative observation that t-amyl alcohol did seem to give me greater rate. the hydrogen bubbles seem to go twice as fast than my trials with t-butanol.

I only tried t-amyl alcohol once though and thus i can't say definitively that it's better. But from what i did see, i think it's worth exploring longer chain alcohols.

Nurdrage this is awesome news as t-amyl alcohol is definately easier to get than t-butyl alcohol for me. It can easily be made using ethyl bromide and acetone as reagents in a grignard reaction.

len1 - 22-12-2010 at 00:14

According to the theory above, my experiments with magnesium turnings 3 years ago failed because they lacked sufficient fresh surface after the reaction with water to initiate the main reaction.

I hence repeated the experiment using 1.6gm Mg powder as before, and 5gm Mg turning produced by 12mm drilling into an Mg block. The latter are much cheaper, economizing on the more expensive powder normally reserved for Grignards.

An initial load of 1.0gm Mg powder was added to the mixture and heated to 150C to dehydrate the KOH. The theoretical amount further up the thread is 0.84gm so a slight excess was used. The dehydration reaction does not occur in the D70 bulk, but in the bottom layer where the Mg contacts the molten KOH, hence the use of the powder for this reaction.

After H2 evolution slowed, the remaining 0.6gm Mg powder and 5gm Mg turnings were added to the flask, so the reagents now looked exactly like in the photo 3 years ago, with the turning protruding slightly above the D70 surface. t-BuOH was added, and initiated the reaction, so potassium globules soon appeared on the surface.

The H2 evolution now and the magnesium turning started to slowly disappear. After 5hrs they were completely absent, and this time the reaction produced little sand. The yield appears to be as before (will know exactly tomorrow).

The formation of MgO crust or sand seems to be a function of the reaction rate, wth slow reactions favouring crust. This in turn is governed by Mg particle size.

[Edited on 22-12-2010 by len1]

Deactivated magnesium test

NurdRage - 22-12-2010 at 01:44

I know it might seem obvious that "deactivated" magnesium shouldn't work, but i thought i'd try it anyway to make certain we're not missing anything.

Abstract: Deactivated magnesium doesn't work to make potassium

So i ran another run using all the same reagents i used in my original paraffin run (25mL paraffin oil, 5g KOH, 2g Mg, 1mL t-amyl alcohol).

But this time I "deactivated" the grignard-quality magnesium turnings by first soaking them in tap water for two days, allowing to dry, then soaking in water for one more day, and then drying in an oven at 130celsius. The magnesium came out very dull brownish with visible powdery appearance.

I ran the reaction using the same conditions and while the first dehydration/hydrogen evolution step worked fine, after adding the alcohol the hydrogen evolution was extremely slow, despite using the more active t-amyl alcohol.

after 2 hours of heating no potassium formation was observed. The reaction was allowed to cool and i did a hard-test by pouring water directly into the mixture. A bit of bubbling accord due to the exothermic solvation of KOH. But no flaming characteristic of potassium metal. Magnesium turnings were still present and by eye it appeared mostly unchanged.

-----> Bottom line the quality of magnesium does play a role. While we've all proven with our different sources that it's not critical, I've shown here that it can be sole cause of a bad run. So when troubleshooting a particular run, care must be taken to use reasonable quality magnesium.

blogfast25 - 22-12-2010 at 06:42

Quote: Originally posted by len1

Wilco, Ive also tried your method with no stirring except when adding the butanol, but this time I got a load of potassium shot, much like blogfast was getting (a very dangerous mix). Maybe coalescence in this case is a random event, or perhaps its a function of the particle size of magnesium.

[Edited on 21-12-2010 by len1]

len1:

How long did you cook for this time? My experiments were resp. cookimg times of 2 h and 1 1/2 h, so much shorter than your initial ones... and wilco's (and pok's).

And why do you refer to the smaller globules as 'a dangerous mix'?

[Edited on 22-12-2010 by blogfast25]

blogfast25 - 22-12-2010 at 06:43

Quote: Originally posted by mr.crow

Holy Cow blogfast25, you're lucky your "pyrex" measuring cup didn't explode.

Calm down, I've used some of my pyrex jugs for years on end: yes, with direct flame...

blogfast25 - 22-12-2010 at 06:51

I ran the reaction using the same conditions and while the first dehydration/hydrogen evolution step worked fine, after adding the alcohol the hydrogen evolution was extremely slow, despite using the more active t-amyl alcohol.

Nurd: you now seem very sure of your case for the 'more active t-amyl alcohol', do you have more data? And are we definitely talking about 2-methyl-2-butanol as 't-amyl alcohol'?

watson.fawkes - 22-12-2010 at 06:57

Quote: Originally posted by mr.crow

Holy Cow blogfast25, you're lucky your "pyrex" measuring cup didn't explode.

Calm down, I've used some of my pyrex jugs for years on end: yes, with direct flame...

In the US, the brand name "Pyrex" is no longer made from borosilicate glass. In Europe (including the UK), it still is. For as many times as this has come up, it's apparently not yet common knowledge.

watson.fawkes - 22-12-2010 at 06:58

Quote: Originally posted by len1

According to the theory above, my experiments with magnesium turnings 3 years ago failed because they lacked sufficient fresh surface after the reaction with water to initiate the main reaction.

I hence repeated the experiment [...]

What was the stirring regime you used here? I infer from the description that you weren't stirring in the dehydration phase. You've reported up-thread that stirring seems to inhibit the initiation phase. What did you do here?

blogfast25 - 22-12-2010 at 07:01

Quote: Originally posted by watson.fawkes

Quote: Originally posted by mr.crow

Holy Cow blogfast25, you're lucky your "pyrex" measuring cup didn't explode.

Calm down, I've used some of my pyrex jugs for years on end: yes, with direct flame...

I have confidence when the label of 'kitchenware pyrex' states that it can be used directly for cooking that that is the case. I have had one or two of these babies crash when I cooled them down too quickly, mainly due to the thickness of the glass which causes big temp. differentials and hence large stresses. Other than that they're very sturdy and much cheaper than lab beakers. But they don't do then in 'conical' or 'round'...

[Edited on 22-12-2010 by blogfast25]

More Experimenting

Alchemist - 22-12-2010 at 07:06

Hello all,

Well this is probably one of the best experiments ever posted on “Science Madness”. Just wish I had some tertiary alcohol, and Shellsol D 70. I could use Paraffin oil, but with the holidays, and etc. I have no funds to buy more chemicals for awhile and still no tert’s.
So hoping someone else will carry out more experiments, the patent says, preferred reaction accelerators are alcohols having 3 to 8 carbon atoms, such as propanol, butanol, pentanol, particularly tertiary butanol or tertiary pentanol. It does NOT say that we have to use Tert’s, but only that they are better. Experiment with some other 3 to 8 carbon atom alcohols. Maybe n-butanol or sec-butanol, and etc.. Maybe with longer heating times or more alcohol?
Also even though the patent does not state the use of Aluminum, why not try it and also Al/Mg or Al/Hg. Maybe longer times would be better here also?
Hope my funds are better in the new year. I could really use some Potassium.
Keep up the great work everyone and have a great New Year, the Alchemist…..

[Edited on 22-12-2010 by Alchemist]

bbartlog - 22-12-2010 at 07:30

Quote:

I know it might seem obvious that "deactivated" magnesium shouldn't work, but i thought i'd try it anyway to make certain we're not missing anything.

Doing trials to verify negative results? That's hard core... It's almost like you're a scientist or something :cool:

Jor - 22-12-2010 at 07:53

So NurdRage, are you making a video for your channel about making the K?

I would be quite interested !

blogfast25 - 22-12-2010 at 10:21

Some K globules made with pok’s reaction, coalesced from very small ones in kerosene at temperatures close to the MP of K (63.6C). The largest one is about 1 cm diameter. they're in Shellsol D70:

More globules and more commentary to follow...

[Edited on 22-12-2010 by blogfast25]

Eclectic - 22-12-2010 at 11:20

I still think the commonly available glycol ethers may work. (Brake Fluid)
Has anyone tried Barbier reaction of acetone, MEK, or cyclohexanone with zinc and alkyl bromides using the ketone itself as solvent to make tert alcohols?

[Edited on 12-22-2010 by Eclectic]

blogfast25 - 22-12-2010 at 13:52

Well, well. I tried to coalesce the rest of my fines, this time using Shellsol D70 but it seems to be slower than in kerosene (subjective).

Soon I’ll try with a very low viscosity solvent like heptane or octane: lower viscosity means all other things being equal that the layer of solvent between two adjacent globules is thinner and that must be beneficial for coalescence…

Sedit - 22-12-2010 at 15:47

In response to nurdrages post about deactivated Mg, I took a chunk of Mg that was sitting in water for a couple days and for the most part has stopped producing hydrogen so I decided to run a quick and simple test.

I took that large piece of Mg and dipped it into HCl holding it there for just a little while as it cleaned and pitted itself. I then washed it under a steady stream of water and then dried it, I once again washed it to ensure all acid was removed.

I compaired this side by side in plain old water with a fresh turning which came from a block of Mg.

The activity of the HCl washed Magnesium was roughtly 2x that of the fresh turning judging from the hydrogen production.

=================================================================================

Is there anything I can add to molten Mg to increase surface area such as sand or possibly carbon? I ask because the slag that formed while I was cleaning the dirty Mg was in the order of 5-10x more active then the pure Mg itself.

If some sort of inert filler could be added then it may very well step this reaction up a notch and make the entire process faster and simpler.

PS: Anyone here have the balls to Ballmill Mg/KOH and an inert solvent to see if any Potassium is produced? I think it may work at a very low temperature but I don't have the time to baby sit it and its not something I wish to leave unattended for even a few seconds.

..... Im also sensing a low temperature Phosphorus production variant stemming from this reaction something in the future.

Arthur Dent - 22-12-2010 at 16:05

Very cool stuff! I've been thinking about trying out the experiment soon, and have gathered many components.

Since it seems that even dry KOH still contains enough water to affect the reaction, would boiling the straight potassium hydroxide in the solvent for a little while before adding the other reagents permit to "drive off" the remaining water and permiting a relatively anhydrous mix of solvent and KOH?

For the reaction apparatus, would a small graham condenser topped-off with a bubbler be a good method of preventing air from entering the vessel? Or would this be overkill?

Given the relative rarity of the "magic ingredient", namely the Shellsol solvent, would it be possible to make our own by mixing 60% kerosene with 40% cyclohexane? According to the Shell MSDS, it might be a relatively similar "formula" ...

See the quickly-drawn (sorry) sketch below:

in this setup, the condenser would be hooked-up to a cool water source, secured with a neoprene stopper to a 3-neck boiling flask, on top would be a small plastic bubbler filled with solvent, secured with a stopper, and the two remaining neck of the flask would be for temperature monitoring and t-bu addition.

My next task will be to find a proper source of KOH, because the reagent-grade stuff is too expensive.

Thanks for any suggestions and recommendations.

Robert

Pok - 22-12-2010 at 17:54

Since it seems that even dry KOH still contains enough water to affect the reaction, would boiling the straight potassium hydroxide in the solvent for a little while before adding the other reagents permit to "drive off" the remaining water and permiting a relatively anhydrous mix of solvent and KOH?

I don't think so. If this would be possible, the patent would probably recommend this because this would decrease Mg consumption. I could imagine that the water can't evolve out of the hydrophobic solvent and would rather stay at the KOH even at high temperatures. But I don't know exactly.

For the reaction apparatus, would a small graham condenser topped-off with a bubbler be a good method of preventing air from entering the vessel? Or would this be overkill?

I think this would be a good idea. condenser = reflux, bubbler = air exclusion. If you carefully control the temperature you can use a simple (but sufficiently long) glas pipe as a reflux condenser. But if you have a graham condenser why not using it? And the idea with a bubbler is a great idea, I think. I will use it

the next time instead of a stupid balloon. Of course, this will only work as long as you don't change the temperature of your mix up and down very often. If you do so: air would bubble from the outside despite (or because of) using such a bubbler because of increasing and decreasing pressure in your reactor. The same effect may occure at occasionally swirling the reactor (cooling down while taking and swirling it -> decrease of pressure -> influx of air through the bubbler) Hydrogen evolution within the 4 hours of reaction can be very slow and will not guarantee a constant efflux! But if you keep the temperature quite constant it will work very well, I think.

Given the relative rarity of the "magic ingredient", namely the Shellsol solvent, would it be possible to make our own by mixing 60% kerosene with 40% cyclohexane? According to the Shell MSDS, it might be a relatively similar "formula" ...

I think this wouldn't make sense. The cyclohexane would vaporize at 81°C and would be lost (or only present in the vapour above your reaction mix)!! If you can find a 200-250°C boiling kerosene free from aromatics (like shellsol D70) and mainly C11-C14 paraffines (like shellsol D70) it should work.

See the quickly-drawn (sorry) sketch below:

The sketch looks fine

My next task will be to find a proper source of KOH, because the reagent-grade stuff is too expensive.

Ebay KOH should be good enough. Maybe you can find a store which sells raw materials for home-made soap production (internet or elsewhere). They sell NaOH, KOH and other stuff in sufficient quality (I got some of my KOH from such a store and used it successfully).

[Edited on 23-12-2010 by Pok]

MagicJigPipe - 22-12-2010 at 18:01

Does anyone have any idea as to the purity of Mg in Coleman firestarters from Wal-Mart? I suppose regardless of the purity, if it works this will be a great source of Mg/K for those that, for some reason, cannot obtain it over the internet or do not wish to.

That's all I have at the moment. My attempt will be done with OTC Mg and inert reaction medium.

Crossing my fingers...

EDIT:

I almost feel bad about posting a source on here because things like that seem to disappear from shelves. Should I take it off?

Also, Pok, I don't think the boiling point of the solvent is extremely critical as long as it is higher than ~200*C or so, right? If it is critical, why? The only thing I can think of is agitation of the reaction mixture.

And are we absolutely certain that low aromatics is a necessity? I see we have some evidence that supports that hypothesis but we can't be quite certain yet. I suppose alkali metals are more likely to react with aromatics... That makes sense then... Could it just reduce yield instead of somehow poisoning the reaction?

[Edited on 12-23-2010 by MagicJigPipe]

Possibly a Substitute for Shellsol

rrkss - 22-12-2010 at 23:15

http://www.homedepot.com/catalog/pdfImages/ab/ab478e2b-db4c-...

Looks like an OTC substitute for shellsol. Its a high boiling point aliphatic petroleum distillate according to the MSDS.

MagicJigPipe - 22-12-2010 at 23:42

But it's boiling point is "greater than" 316*F which is only 158*C. It would have to be MUCH greater to be on par with Shellsol.

I suggested using Naptha earlier. Then I realized that the boiling points I was seeing were in fahrenheit rather than celsius (I have become so accustomed to reading temperatures in C). So, when I saw a boiling points of 200-300 degrees, I thought, "Hey, this should work!". Wrong...

len2 - 23-12-2010 at 00:16

Blogfast - I always let the reaction go 4-5hrs. In this case I used woelens's sand bath on a hot plate with sand temperature controlled. It did not work under my conditions, although we can clearly see it worked for Wilco. I had to drain the entire solid mass from D70, the added the dioxan. Its slighlty wet so the potassium fizzed. The on heating to 80C on an oil bath it rose to the top and coalesced with slight shaking into a blob about 4cm diameter. This is quite a good method, because the slight amount of moisture in the dioxan generates a protective atmosphere inside the reagent bottle. I used schott plastic capped bottles, tightly shut after initial hydrogen release, immersed in an oil bath at 80C. The K rises to the top and you can leave even small shot there to coalesce over 1-2hrs. The final aggregation can be done by gentle rocking for the last few inutes. It seems theres some force attracting potassium balls to each other - because they seem to 'know' where each other is.

The dioxan is obv mixing with a bit of D70 - no problem, I can distill it off and use again.

The potassium shot is dangerous outside D70, where it can spontaneously burst into flame, especially in solvent such as xylene and toluene.

I cant unfortunately change - I have no edit privelages on what I posted ages ago. I would however like to apologize to Pok (but he should learn how liquid density changes with T in almost all cases, esp far from melting point)

[Edited on 23-12-2010 by len2]

[Edited on 23-12-2010 by len2]

condennnsa - 23-12-2010 at 00:23

len, I'd like to suggest that you change what you said about this patent in the Sodium metal - illustrated practical guide , now we know the patent is genuine, and the procedure works, just so we keep the sciencemadness prepublication section accurate.

Video of heating step

NurdRage - 23-12-2010 at 00:25

Here it is:

http://www.youtube.com/watch?v=8azkO_7iBl4

The video starts right after i add the last shot of t-amyl alcohol. it then time lapses 4 hours into 2 minutes and you can see the conversion of magnesium turnings into potassium balls. Finally i show the molten potassium ball after it fully coalesces.

The final crust was rock-hard.

The parameters are: 30mL paraffin oil (0.86g/mL density), 6g KOH, 2.2g Mg, 2mL t-amyl alcohol (i used excess because my first shot didn't work).

If you guys want me to post a version with a different time-lapse factor let me know.

=================================

Quote: Originally posted by Sedit

The activity of the HCl washed Magnesium was roughtly 2x that of the fresh turning judging from the hydrogen production.

Looks like we have our activation means, a hydrochloric acid wash.

len2 - 23-12-2010 at 00:28

Nudrage do you shake your mixture - or you also let it be like Wilco? What size particles are your Mg?

[Edited on 23-12-2010 by len2]

condennnsa - 23-12-2010 at 00:29

Nurdrage , what was the paraffin oil you used? Is it the viscous, clear, odorless stuff sold in drugstores?

Sedit - 23-12-2010 at 00:34

Nurdrage I must stress that yes the HCl washing does indeed help but the Slag, Oh the slag.... that is so active it rivals Aluminum in HCl if you know what I mean. I want to reproduce that in a controled fashion.

My guess is the crude, more then likely oxides of Mg and Fe, are not allowing the molten Mg to form a nice melt but instead a hard crusty slag. This slag must contain very fine Mg with an activation level well beyond anything I have seen yet.

IF someone has pure sand please mix it in with a small amount of melted Mg till its a solid crust. Once broken compair that to the way that normal magnesium shavings react. Its shocking.

NurdRage - 23-12-2010 at 00:37

@Len2

No shaking or stirring during the video.

I did some initial grinding when i mixed the reagents, then after the "dehydration" step i stuck my spatula in and ground it up some more since it hardened to cement. I popped a shot of alcohol, replaced the condenser, and then just let it simmer on its own, no extra stirring.

@Condennsa

I used this oil: http://www.sigmaaldrich.com/catalog/ProductDetail.do?lang=en...

@Sedit

Sand and molten magnesium? Did you make magnesium silicide?

----> interesting implications

len2 - 23-12-2010 at 00:46

It seems you can get coalescence w/o stirring because liquid potassium rises in paraffin - it doesnt in D70. The other question I have is what is you hotplate temp?

Thanks Nurdrage. Well Id say then your paraffin oil method is superior, because it eliminates the 1hrs worth of purification and the dioxane. Ill try it, I just hope I dont have to use the spectro grade - Ill use the stuff we have round for oil baths and see if it works.

[Edited on 23-12-2010 by len2]

NurdRage - 23-12-2010 at 00:50

@len2

Hot plate temp set to 300 celsius. but when i stuck a thermocouple directly into the oil it read 180 celsius.

The actual temperature of the crust bed is probably somewhere in between (probably to the lower end).

If you're going to go with the paraffin oil route to save on coalescence work-up, then i think an important feature is the density of the oil. Mine was 0.86g which is more dense than the liquid density of K at 0.82g. A nice advantage is that when everything cools, solid K is denser and sinks to the bottom, making it safer and less susceptible to air oxidation.

[Edited on 23-12-2010 by NurdRage]

Sedit - 23-12-2010 at 01:03

No nurdrage im just trying to think of an inert agent to suspend the very fine, and very active Mg particals that are forming simular to whats going on with my slag as I melt down crappy magnesium from an old anode.

Perhaps MgO would be the best agent to use in this case I suppose. Carbon on the other hand keeps nagging at the back of my mind as a possible solution.

The main goal is to turn the metal into a material that contains very high surface area simular to how many catalyst are precipitated onto things like silica.

After many runs over and over the clean metal reacts with water and 5% AcOH very slowly where as the crap thats left over reacts vigorously with each and in AcOH it appears like someone tossed Zinc into HCl. This showes me that the slag is WAY more reactive then the pure metal, and as an added bonus its brittle meaning I could break it into finer pieces with realative ease.

I feel if I can figure out how to make this "slag" in a controlled fashion the reaction time will drop dramaticly and quite possibly so will the activation temperature.

NurdRage - 23-12-2010 at 01:14

@Sedit

The highly active magnesium would be great.

I'm also interested in your attempts to reactivate "deactivated" magnesium. Being able to use a wide range of easily obtainable, cheap, low grade magnesium would be just as useful, if not more so, than producing super-grade stuff. A little extra time is alright if it still works.

woelen - 23-12-2010 at 01:53

Removed some spam and an angry reaction.

len2 - 23-12-2010 at 02:55

Another interesting aspect is that the KOH does not react with the glass surface - which remains crystal clear - so that good flasks can be used. Without adding the magnesium, you can see a milk-like fluid - which the KOH becomes at 160C floating at the bottom, and above it clear D70. I presume there is no etching because the surface tension is such that the D70 forms a light coat on the bottom of the glass preventing the KOH from touching it.

Pok - 23-12-2010 at 04:29

@MagicJigPipe

Quote: Originally posted by MagicJigPipe

Also, Pok, I don't think the boiling point of the solvent is extremely critical as long as it is higher than ~200*C or so, right? If it is critical, why? The only thing I can think of is agitation of the reaction mixture.

I only used Shellsol D70 and can't give you infos about other solvents. But I also think agitation is neccessary and 200-250°C boiling point will do this job.

Quote: Originally posted by MagicJigPipe

And are we absolutely certain that low aromatics is a necessity?

The patent says "yes, its important" afaik. Aromatics could react with K.

Quote: Originally posted by MagicJigPipe

I see we have some evidence that supports that hypothesis but we can't be quite certain yet. I suppose alkali metals are more likely to react with aromatics... That makes sense then... Could it just reduce yield instead of somehow poisoning the reaction?

I think it would be yield reduction and not poisoning the reaction, but remember: even a low % aromatics content in the solvent could destroy a large amount of the K, because so much solvent is needed.

@len2

Quote: Originally posted by len2

I cant unfortunately change - I have no edit privelages on what I posted ages ago. I would however like to apologize to Pok (but he should learn how liquid density changes with T in almost all cases, esp far from melting point)

It's ok. We both were a bit nervous.

I know about density change with T. But the MSDS said: density increase from 15->20°C - I also was suprised, maybe a mistake in the MSDS.

blogfast25 - 23-12-2010 at 04:38

Len:

Very interesting point about the dioxane: strange that the coalescence occurs when the K floats, I'd have expected quite the opposite... But nurdrage's results confirm it too...

In my 'medium kerosene' the K doesn't float, must work only with a heavier grade.

Another scheme I dreamt up (w/o realistic prospects of testing) is centrifuging the solvent-molten K mixture to apply force on the K globules and force them into each other...

[Edited on 23-12-2010 by blogfast25]

oxide-free magnesium

Arthur Dent - 23-12-2010 at 05:24

Quote: Originally posted by Sedit

...im just trying to think of an inert agent to suspend the very fine, and very active Mg particals ...

I thought of an idea last night.... what if we took a solid piece of magnesium and grinded it with a medium coarse file (as per Pok's technique) but under the shellsol?

This way, the Mg couldn't react with the air because it would already be in the solvent as the fresh Mg particules would be exposed...

I can realise that filing metallic magnesium immersed in a flammable liquid is something to be done VERY carefully but I can imagine this would provide "ideal" conditions where the magnesium would be mostly devoid of oxides and what not.

Just a thought.

Robert

blogfast25 - 23-12-2010 at 05:34

Probably Robert, but as established the Mg really isn't as critical as we once thought... Neither is probably the solvent, within reason...

[Edited on 23-12-2010 by blogfast25]

blogfast25 - 23-12-2010 at 05:42

Here it is:

http://www.youtube.com/watch?v=8azkO_7iBl4

The video starts right after i add the last shot of t-amyl alcohol. it then time lapses 4 hours into 2 minutes and you can see the conversion of magnesium turnings into potassium balls. Finally i show the molten potassium ball after it fully coalesces.

The final crust was rock-hard.

The parameters are: 30mL paraffin oil (0.86g/mL density), 6g KOH, 2.2g Mg, 2mL t-amyl alcohol (i used excess because my first shot didn't work).

If you guys want me to post a version with a different time-lapse factor let me know.

=================================

Quote: Originally posted by Sedit

The activity of the HCl washed Magnesium was roughtly 2x that of the fresh turning judging from the hydrogen production.

Looks like we have our activation means, a hydrochloric acid wash.

Fantastic: now that's what I call coalescence!

Can you tell us again the grade of 0.86 d parrafin oil you used or where you got it?

Smart also to reduce the amount of solvent a little bit...

Edit: Oh, I see: IR grade. Wiki states: 'Liquid paraffin, or mineral oil, is a mixture of heavier alkanes, and has a number of names, including nujol, adepsine oil, alboline, glymol, medicinal paraffin, or saxol. It has a density of around 0.8 g/cm3.[3] Liquid paraffin (medicinal) is used to aid bowel movement in persons suffering chronic constipation; it passes through the gastrointestinal tract without itself being taken into the body, but it limits the amount of water removed from the stool. In the food industry, where it may be called "wax"'

That might just about do for potassium at about 200C (d = 0.828 at MP)...

Shebang! 27 quid for 500 ml from the extortionists at Aldrich... :mad:

[Edited on 23-12-2010 by blogfast25]

condennnsa - 23-12-2010 at 06:17

Well I finally found a source for t-butanol (Merck), at 25 euro/1L, but it'll be some 3-4 weeks till I get it. Also ordered 1 kg of 63 um magnesium from czort. God I hate the holidays, I have to wait so much time to try this out.

Indeed nurdrage, the fact that you were able to coalesce all potassium in that huge ball is amazing. Your results make me believe that paraffin oil might be a better solvent for this procedure than shellsol, and it's also completely odorless.

[Edited on 23-12-2010 by condennnsa]

watson.fawkes - 23-12-2010 at 08:08

Quote: Originally posted by rrkss

http://www.homedepot.com/catalog/pdfImages/ab/ab478e2b-db4c-...

Looks like an OTC substitute for shellsol. Its a high boiling point aliphatic petroleum distillate according to the MSDS.

CAS 8052-41-3 is also known as Stoddard solvent. There are two kinds of refinement used to get Shellsol D70 and its relatives: fractional distillation with a relative narrow boiling point range and hydrogen treatment to remove aromatics and alkenes. Stoddard solvent is refined only with the first of these, not the second. While the jury is out on the exact effect of aromatic and alkene impurities, it seems that they don't poison the reaction unless the quantities are particularly high. They may (although this has not been verified in the present process) reduce yields by reacting with metallic potassium.

I might point out that it's certainly not out of the question to do your own hydrogenation treatment, although the pressure vessel and other equipment to do so isn't nearly as common or inexpensive as glassware.

Eclectic - 23-12-2010 at 10:04

K-1 water white kerosine made for unvented heaters and lamps should be fully hydrogenated. Aromatics cause smoke and odors, and would totally freak out EPA and consumer protection agencies.

Other Alcohols

Alchemist - 23-12-2010 at 11:27

Hello again,

Has anyone as yet tried any other alcohols other than t-butanol or t-pentanol and not just tert alcohols. There must be more c3 to c8 carbon alcohols that would work than just the above two!

Keep up the great work everyone, the Alchemist.....

NurdRage - 23-12-2010 at 14:25

It's amazing how we've all coalesced (pun very much intended) into an unofficial research team.

We're all exploring various aspects that we find interesting: I think Sedit is investigating magnesium quality and activation, Woelen and Len1/2 are exploring mechanisms and coalescence methods, Blogfast25 seems to be looking at solvents and coalescence methods, and i'm exploring solvents and activation conditions.

We need a theme song

Anyway, as we compile more data i agree with Woelen that we should start assembling it together in a nice concise and comprehensive form.

@alchemist
I think someone tried isopropanol but that didn't work. t-butanol and t-pentanol (t-amyl alcohol) were explored first because those were readily available and perhaps the easiest/cheapest to get for the amateur. They can even be made by a very motivated amateur. More exotic alcohols i think are less accessible.

If you want to explore other alcohols go for it, or if you can propose an accessible synthesis for them using OTC materials i'd love to hear it.

I might try them when i can find the time.

watson.fawkes - 23-12-2010 at 15:58

If you want to explore other alcohols go for it, or if you can propose an accessible synthesis for them using OTC materials i'd love to hear it.

I might try them when i can find the time.

Funny you mention this; I was just looking into this possibility this morning.

Amyl alcohols are found as a major component in fusel alcohols, ordinarily considered by-products of fermentation, rather than targets. There's an interesting paper on the Wikipedia page for fusel alcohol that addresses their origin. The Ehrlich pathway for fusel alcohol production: a century of research on Saccharomyces cerevisiae metabolism; full text is available at the link. Here's the central topic:

Quote:

Fusel alcohols are derived from amino acid catabolism via a pathway that was first proposed a century ago by Ehrlich.

I haven't dug into the paper very far, but it looks like leucine is a decent candidate for fermentation with some yield as the primary alcohol isoamyl alcohol (3-methyl-1-butanol). Leucine is readily found in soy protein, available in powdered form at any health food store. Now I don't know what the activity of a heavier, primary alcohol is going to be for making potassium, but the patent text hints that they might work. (It might also be that they don't work well, but well enough to disclose in the patent so that they can claim them.) I also don't know what the yield from soy protein is likely to be, but the Wikipedia page on leucine states that soy protein is about 5% leucine. It's unlikely to be worth doing if all you want is this particular fraction, but it's certainly possible that the fusel oil from such a fermentation might mediate the Mg / KOH reaction in relatively impure form.

So if someone else wants to brew up a really smelly wort and tell us all about it here, I'm all ears. I am not all noses, no no no.

[Edited on 24-12-2010 by watson.fawkes]

My Mg is better then yours

Sedit - 23-12-2010 at 23:27

It's amazing how we've all coalesced (pun very much intended) into an unofficial research team.

LMAO seriously though when you first suggested about me looking for a super activated Magnesium I thought eh ok, cool. But then as I started to play with some of the products of my Mg cleaning experiance I really think I may be on to something pretty big.

Using damn near freezing cold water(have you been outside lately

) I added a couple chunks of Mg from an ingot I made. Then I added a few chunks of the stuff I removed from the dirty Magnesium. The results are amazing to me. The clean Mg did almost nothing for atlest 5-10 minutes or so, I dunno my perception of time is messed up when im playing, Yet the instant the other material which is like a sponge almost hit the water a rapid steady stream of bubbles came off. It took over 20 minutes I believe before the normal Mg was even producing half of the bubbles as this other material.

I believe there will be a level point somewhere in between as the clean magnesium gets going and the old starts to level out but I may very well be wrong.

I plan on setting up two inverted test tubes filled with distilled water and a sample of magnesium fillings under one and my material under the other. Judging from the rate if it stays steady I would expect my material to fill the tube in well under one tenth the time if that but only experiment will tell.

I have a theory and it may not help this potassium synthesis if its how its working. I feel the oxidation that is contained in the slag is some water soluble material that is giving a huge surface area to the active amount of Magnesium.

In the case of the current potassium experiment I would suggest that someone take magnesium pieces and ball mill them with wax and something to grind them like stainless steel shot. The resulting material should be very active once the wax melts under the current reaction conditions and provide a great surface area causing the reaction to run exponetialy faster as does the hydrogen production in my water based experiments.

Hey, can I patent this

blogfast25 - 24-12-2010 at 06:44

@Nurdrage:

Well, my ‘Coalescence Department’ is now closed of course, with your ‘floating coalescence’ method putting it firmly out of business! Thanks for nothing!

But thankfully there remains much to researched, IMHO:

• Type of alcohol
• Quantity of alcohol
• KOH/Mg ratios
• Reagents/solvent ratios
• Magnesium

WE may want to develop a test tube procedure for the evaluation of different treatments all at once, in order speed up testing and increase resolution…

But first I need to get my hands on a suitably heavy mineral oil… Yesterday I found an ‘outdoor lamp oil’ that’s clearly heavier than the kerosene I’ve got but it contains Citronella oil (which itself is made up mainly of Citronellal [an aldehyde], Citronellol [a primary alcohol] and Geraniol [another primary alcohol]) so maybe I’ll be making ‘citrus flavoured’ potassium, definitely patentable, that one!

Today I’m gonna see if the pharmacy has got anything…

Nicodem - 24-12-2010 at 09:49

I think it would be more interesting to do a scientific research on this reaction to figure out the mechanism of the reaction. Perhaps science is not as popular as madness, but I still think this is a great opportunity for this forum to contribute to science, given the complete obscurity of this reaction in the scientific literature (BTW, did anybody actually made a thorough literature search?).
I could contribute by doing the basic experiment, thus testing the reaction of t-BuOK with Mg to figure out if the currently proposed mechanism has any validity. If the equilibrium of t-BuOK + Mg <=> t-(BuO)2Mg + K is truly directed toward the left side, then there should be no formation of potassium until a pellet of KOH is added (this addition representing a crude validation test, since the actual absence of a reaction would mean nothing given the stochastic nature of the Mg surface activation process). Unfortunately I have no shellsol or undecane. Perhaps after my vacations are over I will give this a try with paraffin wax. Even though a negative reaction outcome would not give any totally conclusive answers, a positive outcome would. It would tell us the equilibrium direction of the reaction and disprove the hypothesis about the driving force of the overall KOH + Mg reaction.

Using a longer chain t-alcohol as catalyst might prove useful in the synthesis of sodium where t-BuOH is unlikely to work given the practical insolubility of t-BuONa in alkanes. A crude way to "fatty t-alcohols" can be the reaction of a large excess of MeMgI or EtMgI(Br) with pure margarine in refluxing THF or other ethers. The preparation of grignard reagents require Mg and dry solvents. Purifying the products would require some knowledge of organic chemistry, but already with extractions it might be possible to get them pure enough for this reaction. I doubt it requires a very pure catalyst, as all the impurities would be destroyed by the reaction conditions (hot KOH!).

PS: It would be nice if one of the members with experience about this reaction writes a general warning and an explanation about the hazards involved. For example, what to do if the flask containing potassium balls heated at 200 °C suddenly breaks and the metal and oil catch fire? I think there will be many who will now try to reproduce the experiment, and many probably don't have a clue of the possible consequences if things go bad. Personally, heating alkali metals scares me terribly. Glassware is not immune to heat shocks! My memories of burning sodium metal are not pleasant and somehow to me potassium sounds worse. After a consensus is reached about the best warning disclaimer and instructions on how to extinguish the fire, we could then link to it from the first post in the thread.

aonomus - 24-12-2010 at 10:29

I would argue having a large amount of available dry sodium chloride or fine free flowing sand would be the best way to deal with an alkali metal fire. One has to be careful to ensure that the NaCl or sand is dry and free flowing, as I've seen sand buckets in labs which have never been maintained, and have material akin to cement inside that would do no good for a fire. Another idea is sodium bicarbonate, it is readily available, and soaks up a ton of heat (converting from bicarbonate to carbonate, and finally to oxide if heated enough).

UnintentionalChaos - 24-12-2010 at 10:40

Quote: Originally posted by aonomus

Another idea is sodium bicarbonate, it is readily available, and soaks up a ton of heat (converting from bicarbonate to carbonate, and finally to oxide if heated enough).

Except that alkali metals burn quite happily in CO2 and water vapor, as would be let off by the bicarbonate. You might even make the fire worse. I'd go with sand or NaCl, personally.

woelen - 24-12-2010 at 11:07

One of the things which I have seen is beakers placed directly on a hot plate. That is not something I would like to do with this. Use a sand bath, it is the safest option there is. The samd sucks the oil such that it cannot pour/run off the hotplate.

I still have some Shellsol D70 left, with a lot of t-BuOK dissolved in it. After christmas, I'll try putting some Mg in this solvent and heating it. No KOH involved, no other compounds. This is close to the situation, mentioned in one of the patents in which t-amylate is reacted with magnesium.

blogfast25 - 24-12-2010 at 12:36

To circumnavigate the extortionists at Aldrich, I’ve just bought 200 g of low melting paraffin wax for microscopy purposes, MP= 56C (eBay, £1.04, not bad!). Going by the MPs of the n-alkanes (a simplification of course) this must be a C20 – C30 fraction with a density over 0.8.

Assuming that my kerosene is around C12 (going by its BP), then a mix of 1 mol wax and 1.6 mol of kerosene should give an average of around C17, with an MP of around 20C and a density of just over 0.8. About 350 g wax mixed with 270 g kerosene…

Other blends of kerosene with unblended paraffinic candle wax (no stearic acid, just alkanes - no aromatics) to increasee the density of the solvent mixture should also lead to the right balance of MP and solvent density and to achieving Nurdrage's ecellent coalescence result...

[Edited on 24-12-2010 by blogfast25]

[Edited on 24-12-2010 by blogfast25]

blogfast25 - 24-12-2010 at 12:43

@nicodem:

Most of us here are better equiped for applied research than for pure research. For that reason I think that most work here will concentrate on practicalities, how to shorten procedure time and such like...

From a safety perspective, sand bath as heating and dry, free flowing sand at the ready is probably the best we can do. Note also that the glass here is not stressed thermally a lot: all heating and cooling is done quite gradually. A sudden catastrophic failure of a glass flask is therefore not very likely... But a big RED safety blub at the start of this thread would probably make a lot of sense... :cool:

[Edited on 24-12-2010 by blogfast25]

NurdRage - 24-12-2010 at 14:38

To circumnavigate the extortionists at Aldrich, I’ve just bought 200 g of low melting paraffin wax for microscopy purposes, MP= 56C (eBay, £1.04, not bad!). Going by the MPs of the n-alkanes (a simplification of course) this must be a C20 – C30 fraction with a density over 0.8.

[Edited on 24-12-2010 by blogfast25]

[Edited on 24-12-2010 by blogfast25]

I wish i got to you sooner before you ordered:

I strongly urge NOT TO USE PARAFFINS WITH MELTING POINTS ABOVE ROOM TEMPERATURE!!!

This comes from my experience with candle wax as a solvent (essentially a high MP paraffin). While it works great, the problem is safety.

When it solidifies it traps tiny suspended balls of potassium making it extremely dangerous to handle. Worst of all it doesn't get destroyed when you apply alcohol, only when it's deformed or cracked do you expose the potassium.

So it seems safe when you toss it in isopropanol, it doesn't bubble or react. But then when you pick it up and it cracks or breaks the exposed potassium reacts with whatever moisture or water may accidentally be near it, then it bursts into flames. The heat melts the remaining paraffin and exposes more potassium.

Another issue is when you pour the dead mixture a thin layer of it is stuck in your flask and solidifies.

I've had several pieces of glassware bursts into flames on me due to residual wax with suspended K. I suffered no injuries but unexpected fire is definitely above my "comfort zone"

That being said, with very high diligence, high MP paraffin can be used safely if you carefully keep everything above the melting point and destroy the potassium residues, wastes, and bits with Isopropyl alcohol before labeling them "safe".

In my opinion though it's not a novice operation and not something that should be used when safer liquid alternatives are available.

If you can blend your paraffin to make it have a melting point below room temperature then go for it. Otherwise i strongly urge not to perform the experiments with solid paraffins.

NurdRage - 24-12-2010 at 14:40

Quote: Originally posted by Nicodem

I think it would be more interesting to do a scientific research on this reaction to figure out the mechanism of the reaction. Perhaps science is not as popular as madness, but I still think this is a great opportunity for this forum to contribute to science, given the complete obscurity of this reaction in the scientific literature (BTW, did anybody actually made a thorough literature search?).

I'm only going to contribute to pure research if i'm Co-author on the resulting paper.

Sedit - 24-12-2010 at 17:05

I finally found a Magnesium firestarter block as a source of Magnesium to compair to the crap I got from the used anode, Im pretty sure that the anode is either some alloy or just to old and used up to much. It melts which is something I can't do with my new Mg and its activity as a fresh metal is much less.

However my slag compairs to the new fresh Mg and even a little more active so I can only imagine the power a Magnesium spong made from good fresh Magnesium would have.

Im going to focus some attention on creating some sort of magnesium spong because delving into the unknown is what keeps me interested in science else my mind starts to wounder.

aonomus - 24-12-2010 at 17:50

I think that after the christmas break I will attempt this synthesis, I'll make sure to grab a 1H NMR to try to determine aromatics presence (though I'm sure the low ppm shift peaks will be a complete mess). The paraffin/kerosene idea sounds the most appealing since the potassium will separate by density by itself, though the flammability is the worrying part...

[Edited on 25-12-2010 by aonomus]

len1 - 24-12-2010 at 19:27

I think its OK to heat pyrex directky on a hot plate. Around here oil baths to 250C are frequently used which is basically the same operation - I have never seen one crack

blogfast25 - 25-12-2010 at 06:23

If you can blend your paraffin to make it have a melting point below room temperature then go for it. Otherwise i strongly urge not to perform the experiments with solid paraffins.

That's THE WHOLE IDEA, nurdy, until I've found a cheaper alternative to your IR paraffin, to modify the density of lighter kerosene with heavier melted/dissolved paraffin wax, hoping that at RT such mixtures remain essentially liquid or are very low melting. Tomorrow I'll try with petroleum jelly (unadulterated Vaseline - MP = 37C) as density modifier...

For my 'left over' bits of Shellsol/kerosene that contain small amounts of K and fines, I find that treating with copious amounts of water kills it without safety issues (better than ethanol IMHO because EtOH is also flammable - see len1's minor incident). The large amount of water acts as a heat sink, thus keeping temperature close to RT and the risk of the solvent catching fire very low...

[Edited on 25-12-2010 by blogfast25]

blogfast25 - 25-12-2010 at 08:25

Quote: Originally posted by aonomus

The paraffin/kerosene idea sounds the most appealing since the potassium will separate by density by itself, though the flammability is the worrying part...

[Edited on 25-12-2010 by aonomus]

It's not a great worry if you use decent refluxing and stay away from open flames (use sand bath or electrical hot plate)...

NurdRage - 25-12-2010 at 10:49

That's THE WHOLE IDEA, nurdy,

Sorry, I didn't mean to come off sounding all condescending. This is one of those experiments where the "safety zone" isn't obvious so i thought i'd explain it again to make sure nothing was missed.

If you're aware of it all then go for it and ignore my ramblings.

good luck, i look forward to your work, an OTC blend of solvents with exactly the right properties for maximum safety, reaction rate and coalescence properties would be awesome.

Eclectic - 25-12-2010 at 13:45

Take a look at AW 32 hydraulic oil, it's available as high purity food grade also.

rrkss - 25-12-2010 at 14:18

I've been looking at smokeless parrafin for tiki torches as a possible solvent. According to the MSDS it is 99% blend of C13 and C14 straight chain alkanes. With the right boiling range of over 225 degrees Celsius. If it is smokeless the aromatic content must be incredibly low.

http://www.buytikitorches.com/category/tiki-torches-oil

MSDS for said product
http://www.buytikitorches.com/MSDS.pdf

Eclectic - 25-12-2010 at 18:17

Even better!

condennnsa - 26-12-2010 at 05:56

Well johnson's baby oil is actually the same as paraffin oil, right?

who'd have thought we be making K in baby oil...

blogfast25 - 26-12-2010 at 07:54

Quote: Originally posted by rrkss

I've been looking at smokeless parrafin for tiki torches as a possible solvent. According to the MSDS it is 99% blend of C13 and C14 straight chain alkanes. With the right boiling range of over 225 degrees Celsius. If it is smokeless the aromatic content must be incredibly low.

http://www.buytikitorches.com/category/tiki-torches-oil

MSDS for said product
http://www.buytikitorches.com/MSDS.pdf

According to Wiki, at C14 the density is still only 0.76, too low for liquid K to float in (d = 0.828 at MP). I think you need at least C17 for floating K...

blogfast25 - 26-12-2010 at 07:56

good luck, i look forward to your work, an OTC blend of solvents with exactly the right properties for maximum safety, reaction rate and coalescence properties would be awesome.

No condescention was detected at my end, Nurdrage, I thought you'd slightly misread my point, that's all...

blogfast25 - 26-12-2010 at 08:04

Quote: Originally posted by condennnsa

Well johnson's baby oil is actually the same as paraffin oil, right?

who'd have thought we be making K in baby oil...

I think that might work, especially as they now do a 'heavier' grade, 'bay oil gel':

http://www.boots.com/en/Johnsons-Baby-Oil-Gel-200ml_6079/

It does contain some camomile (for 'extra relaxed K'?) and some other stuff we don't much care for:

Hydrogenated Butylene/Ethylene/Styrene Copolymer, Hydrogenated Ethylene/Propylene/Styrene [Copolymer, my edit], Fragrance (Chamomile Oil), Tocopheryl (Vitamin E) Acetate, Retinyl (Vitamin A) Palmitate, Corn (Zea Mays) Oils, Cholecalciferol (Vitamin D3)

So there's a bit of K's favourite food in there: OXYGEN!

And with the hydrogenated (thus inert) polymeric additives this stuff may have some weird rheological properties (non-newtonian fluid)...

[Edited on 26-12-2010 by blogfast25]

blogfast25 - 26-12-2010 at 10:09

Well, well, well...

My first attempt at ‘creating’ a solvent in which molten K would float actually failed.

Instead of carrying out an actual reaction I used some leftover K balls, mostly some 2 mm ones, some smaller ones, some fines and some MgO. I decanted off the supernatant kerosene and replaced it with a mixture of 8 g of clean medium kerosene and 20 g of petroleum gel (Vaseline ®, MP just above body temperature). This was an estimate to get about C17 from about C12 (kerosene) and about C20 (Vaseline). No floatation was observed, not even at 200 C! I added more Vaseline to try and increase the heavy fraction and thus the density but to no avail.

Finally I ran a test tube test with pure molten Vaseline and a 3 mm ball of K and the damn thing just didn’t want to float! But it did when I added a drop of IPA, because (a small amount of) hydrogen evolved and presumably small amounts of it stuck to the (by now) ball(s), providing buoyancy…

So now I’m wondering if we’re slightly barking up the wrong tree when it comes to density: is it possible that nurdrage’s ‘floating coalescence’ isn’t just caused by high solvent density of the ‘IR paraffin’ but also by reaction hydrogen providing some lift for the nascent potassium?

Food for thought…

Another attempt will now be made with the 56C MP paraffin wax when it finally gets here (this time definitely no ‘thanks to the L*rd!’) and possibly baby oil (the unadulterated one).

Eclectic - 26-12-2010 at 10:09

Baby oil is fairly high viscosity? It looks like "white mineral oil", ISO VG 10-15, viscosity 70 would have high enough specific gravity and low viscosity. They are sold as low temperature hydraulic oils (ISO 15-22) and high speed machine spindle lubricants.

http://www.seversonoil.com/pdfs/FamilyOfBrands/FO_White_Oil....

http://www.farwestoil.com/crossreference.pdf

http://www.amazon.com/Albatross-SO/dp/B002BGZZXI
http://albatross-usa.com/lilywhitesewingmachineoiliso-22.asp...

http://www.drillspot.com/products/275646/Mobil_VELOCITE_NO_6...

[Edited on 12-26-2010 by Eclectic]

blogfast25 - 26-12-2010 at 10:14

Quote: Originally posted by Eclectic

Baby oil is fairly high viscosity? It looks like "white mineral oil", ISO VG 10-15, viscosity 70 would have high enough specific gravity and low viscosity. They are sold as low temperature hydraulic oils (ISO 15-22) and high speed machine spindle lubricants.

http://www.seversonoil.com/pdfs/FamilyOfBrands/FO_White_Oil....

http://www.farwestoil.com/crossreference.pdf

http://www.amazon.com/Albatross-SO/dp/B002BGZZXI

Nice links but no mention of densities though... But presence of naphtenics would indeed help keep viscosity low. And naphtenics tend to be a little denser: n-hexane = 0.6548, cyclohexane = 0.779, n-octane = 0.703, cyclooctane = 0.834

[Edited on 26-12-2010 by blogfast25]

NurdRage - 26-12-2010 at 14:04