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

About to try this at my school. Any advice?
Setup so far:
-Hotplate w/ stirrer, stir bar
-Mineral oil, 70 mL as solvent
-~1mL t-BuOH
-reflux condenser w/ rubber joint
-100mL Erlenmeyer flask

The mix will be lightly stirred to promote K production, but can I get away with heavier stirring?
Also, due to the length of the club, it would be wonderful to get this done in an hour, instead of the usual 3-4. Stirring greatly speeds up reaction time, according to the Raging Nurd, but by how much?
If speeding up time really isn't an option, I could always set up the mix in the morning, and come back later as it's finishing. Possibly.

Any thoughts?

The stirring bit is controversial and I'm not convinced faster stirring works, as Nurdrage claims.

You can split the experiment into two parts, though: 2 hours for actual reaction, then cool to RT and stopper properly.

Then, next session, dilute flask contents slightly with fresh solvent, add a few drops of isopropanol (99 %) or ethanol (99%), transfer to a clean conical flask and heat to about 100 C (e.g. on steam bath). Swirling the flask regularly will make the droplets of K (assuming some formed in the first session, of course!) coalesce into larger and larger ones. This too will take 1 - 2 hours...

[Edited on 25-2-2013 by blogfast25]

elementcollector1 - 25-2-2013 at 09:43

Might be able to do that... Maybe. The 2-hr reaction will be the tricky part, because the time window is only an hour and 20 minutes long. Could set up at lunch...
Or I could try to extend Science Club for the day to fit the time window. Unlikely, but possible.
Could make the lab into 2 days, but we do have other stuff to do.
I think that I'll give this a try at home this week, and then show it to the club next week. That way I might get 'familiar' with the experiment, and might be able to do it faster.
Can you tell I'm nervous? ;D

MrHomeScientist - 25-2-2013 at 09:59

Definitely a good idea to familiarize yourself with an experiment before showing it off, especially if it involves explosive metals! I rarely record my first attempt at an experiment for posting to YouTube - I almost always run into problems that would make a video presentation of it pretty boring

I can't wait to try this one out myself - now that I have all the materials I just need to work out the logistics and set aside some time one of these weekends.

That's great that you have a Science Club; I feel like most schools nowadays don't have anything like that. Do you focus on chemistry or is it more broad?

elementcollector1 - 25-2-2013 at 10:04

It's much more broad than chemistry, and more often focuses on physics (our coach was a physicist or something, he's retired now). We just went to the Science Bowl, which we've been training for all year. It focuses on everything from Math to Energy to Earth and Space, with Chemistry, Physics and others thrown in there as well. (We took fifth and fourth place with our two teams, too.

)

Anyway, enough OT bragging about our awesome club.

Unfortunately, can't actually try out stirring - so I guess it's the 4-hour long haul for me. If it works, I'll bring in a sample of K metal to the meeting and compare results. If it doesn't... Ah well.

EDIT: I have 40 mesh Mg 'powder' as well as turnings, and the powder should speed up the rate of reaction significantly. Only problem is, I might not recognize the potassium for the magnesium!

Also, I'm attempting to make my 125 mL Erlenmeyer as dry as possible for the synth, because any water, no matter how small an amount, is going to impact yield. Any suggestions for getting the glass super-dry?

[Edited on 25-2-2013 by elementcollector1]

blogfast25 - 25-2-2013 at 13:12

Might be able to do that... Maybe. The 2-hr reaction will be the tricky part, because the time window is only an hour and 20 minutes long. Could set up at lunch...

Much higher up you'll find an entry of mine where I measured hydrogen evolution over time. I found the reaction was basically over after 1 hour, in my conditions. Most of the experiment time (again, in my conditions) is needed to actually grow the K-balls. And that is a lengthy process and difficult to speed up...

elementcollector1 - 25-2-2013 at 13:15

Hopefully, the magnesium powder will help.
Is there a way to powder the KOH too without it getting too deliquescent?

blogfast25 - 25-2-2013 at 13:17

Any suggestions for getting the glass super-dry?

[Edited on 25-2-2013 by elementcollector1]

Yes. DON'T bother! Normally dried glass is fine. Your KOH contains at least 10 % water. It's not a problem. My glassware, KOH, Mg, solvent and T-butanol came straight off the shelf.

elementcollector1 - 25-2-2013 at 13:20

Got it.
One thing I've always worried about is the reflux of the t-BuOH. According to Wiki, it boils at just 82 C, and the temperature of the reaction is around 200 C. Even if it is refluxing, shouldn't the alcohol just boil off instantly as soon as it hits the solvent again?

blogfast25 - 25-2-2013 at 13:20

Hopefully, the magnesium powder will help.
Is there a way to powder the KOH too without it getting too deliquescent?

Powdering isn't necessary: flakes will do just fine. It's what everyone else uses. This reaction requires certain phase transfers but it doesn't seem to be affected by system granulometry much.

Unfortunately we still don't seem to understand what is wrong when it doesn't work.

elementcollector1 - 25-2-2013 at 13:23

I thought it was solvent problems. Pretty sure every grade of magnesium imaginable was tested, so it couldn't have been that. Different solvents such as kerosene and vaseline were tested too, without success if I remember. I'm using pharmacy mineral oil, which claims "100% mineral oil", but I'm going to boil it just in case.

blogfast25 - 25-2-2013 at 13:23

Got it.
One thing I've always worried about is the reflux of the t-BuOH. According to Wiki, it boils at just 82 C, and the temperature of the reaction is around 200 C. Even if it is refluxing, shouldn't the alcohol just boil off instantly as soon as it hits the solvent again?

Nope. That's the definition of TOTAL REFLUX: everything condenses back into the reaction flask, inluding the t-butanol. It's also not necessary to cool your condenser like crazy: as long as it's has an area that's below the BP of t-butanol it WILL reflux it. Woelen for instance used a vertical Liebig condenser but with air pumped through the mantle instead of water. Worked perfectly!

[Edited on 25-2-2013 by blogfast25]

blogfast25 - 25-2-2013 at 13:28

I thought it was solvent problems.

No, we don't know that. Many solvents were found to work w/o problems: that 'special kerosene' (I forget the name now), normal kerosene, lamp oil, paraffin, candle wax, Tetralin, paraffinic baby oil, all tested positive. The patent suggest decane (for cesium).

Don't bother boiling your mineral oil, it's a waste of time. Instead spend some time prepping well, and fingers crossed! Don't forget a horse shoe!

[Edited on 25-2-2013 by blogfast25]

[Edited on 25-2-2013 by blogfast25]

elementcollector1 - 25-2-2013 at 13:31

Was it Shellsol D70?
'All tested positive' as in confirmed production of K, correct?

blogfast25 - 25-2-2013 at 13:33

Was it Shellsol D70?
'All tested positive' as in confirmed production of K, correct?

Yep, Shellsol D70, aka 'a waste of money'. All made K. The few instances (bar sheer incompetence) were it didn't work are poorly understood.

elementcollector1 - 25-2-2013 at 14:46

I don't understand. Does this mean the reaction is officially over at ~1 hr, then the rest of the time is just coalescing K?

blogfast25 - 26-2-2013 at 06:25

I don't understand. Does this mean the reaction is officially over at ~1 hr, then the rest of the time is just coalescing K?

In a nutshell: yes. I think you'll agree with me when you've witnessed your first succesful reaction.

Mailinmypocket - 26-2-2013 at 06:33

I'm happy to see somebody else trying this out! I really hope it works for you and then we can maybe get another hint as to why it sometimes doesn't work... After so many tries my own results never resulted in much more than K dots bobbing around in a fleeting existence

elementcollector1 - 26-2-2013 at 07:24

If I had to guess, you might have added too much t-BuOH, which dissolved right through your potassium. Or not enough magnesium was used. I don't really know without knowing exactly what you did first.

Mailinmypocket - 26-2-2013 at 07:39

If I had to guess, you might have added too much t-BuOH, which dissolved right through your potassium. Or not enough magnesium was used. I don't really know without knowing exactly what you did first.

I thought the excess t-BuOH was the culprit also, but given the 10+ runs I did I tried varying the amounts of it +\- and although the balls did grow bigger on some attempts than others, they never coalesced and always vanished after about 1h30m. I have also tried various types of Mg and KOH, and solvents... Ill have to go back through my notes and see which attempt was the most fruitful and maybe work up from that. Not until new flasks arrive though since my last 19/20 erlenmeyer bit the dust..

elementcollector1 - 26-2-2013 at 07:46

I'm just using rubber joints instead of ground glass. Are you sure you need ground-glass for this reaction?
Also, what size flask do you typically use? I'm thinking 125mL.

[Edited on 26-2-2013 by elementcollector1]

Mailinmypocket - 26-2-2013 at 08:27

I'm just using rubber joints instead of ground glass. Are you sure you need ground-glass for this reaction?
Also, what size flask do you typically use? I'm thinking 125mL.

[Edited on 26-2-2013 by elementcollector1]

Oh no, ground glass isn't necessary, I just prefer it. I made a setup with rubber stoppers and whatnot (see up the thread, not like it worked any better

). I was running it in a smaller 50ml flask, but the rubber stopper run was done in a 125.

elementcollector1 - 26-2-2013 at 08:35

Does it make a difference if a 50mL or a 125mL flask is used? For the 50 mL, the reactants would always be completely submerged, but couldn't grow as large without rendering removal impossible. For the 125mL, the reactants might not always be completely submerged in the oil, but the potassium could be allowed to grow significantly larger.
Hopefully going to give this experiment a try in a few days, if not today.

Mailinmypocket - 26-2-2013 at 08:44

Does it make a difference if a 50mL or a 125mL flask is used? For the 50 mL, the reactants would always be completely submerged, but couldn't grow as large without rendering removal impossible. For the 125mL, the reactants might not always be completely submerged in the oil, but the potassium could be allowed to grow significantly larger.
Hopefully going to give this experiment a try in a few days, if not today.

That's one of the reasons I used a small flask (the other being reagent wastage). If there is not enough room for the K to move, coalescence will not occur as readily. The times I did it in the 125ml I found that there was very little free space in the solvent, especially after the KOH dehydrates and increases in size.

elementcollector1 - 26-2-2013 at 08:47

Got it. The original versuchschemie.net entry used 50mL of liquid in a 125mL flask, which seemed to work well.
Unfortunately, the setup with a huge condenser on a tiny flask can be a bit wobbly... I'll probably need to clamp the condenser to a stand just to keep the whole thing from falling over.
What do you think about using a sand bath? My electric hotplate isn't all that hot, but a sandbath could be beneficial.

blogfast25 - 26-2-2013 at 09:58

Got it. The original versuchschemie.net entry used 50mL of liquid in a 125mL flask, which seemed to work well.
Unfortunately, the setup with a huge condenser on a tiny flask can be a bit wobbly... I'll probably need to clamp the condenser to a stand just to keep the whole thing from falling over.
What do you think about using a sand bath? My electric hotplate isn't all that hot, but a sandbath could be beneficial.

Sand bath is safer and tends to provide more even heat. In the unlikely event of flask breakage the sand provides the first barrier/absorber for the oil and any potassium.

A minimum of clamping is needed secure the assembly. One retort stand, one bosshead and one retort clamp will do it though. To occasionally swirl the flask, loosen the retort clamp a bit, swirl, then retighten till secure again.

[Edited on 26-2-2013 by blogfast25]

elementcollector1 - 26-2-2013 at 16:52

Well, I gave the experiment 2 tries today. One was a bit messy, and I ended up switching to ground-glass for better stirring and joint integrity. Here's my general setup for the second run:

Unfortunately, something strange happened: The mineral oil started to froth over, right up into the condenser! At this point I shut everything down. Is this normal?

One more note: I'm a little confused as to how to add the alcohol. If I inject it through the top of the condenser, it freezes on the sides and refuses to move. If I inject it into the flask, it instantly boils away. One solution I could see to this is injecting hotter alcohol in so as to give it a better chance of reaching the flask, but what do you people think?

Mailinmypocket - 26-2-2013 at 20:11

I have found that while the heat is ramping up there is always a sudden frothing that occurs. This is accompanied by dense white fumes. This always happens between ~170-190c (in my experiments anyways) but has never went into the condenser, that may just be a difference in set ups though? What kind of mineral oil are you using? I believe this is happening due to the KOH losing its water content and releasing steam, although I don't know for certain. It does stop frothing after less than a minutes time though as the heat increases!

As for adding the t-butanol, try mixing it up well with a couple milliliters of your mineral oil prior to adding it. Then add the oil mix in increments or however you were adding the alcohol... It's much easier than dealing with a fraction of a ml at a time.

Next time try and allow the mix to keep heating up even if it froths, provided we aren't talking about ridiculous amounts of it obviously. Once it calms down wait until the target temperature of about 220c and add the alcohol.

[Edited on 27-2-2013 by Mailinmypocket]

elementcollector1 - 26-2-2013 at 20:27

Well, can I re-use the reaction mix I currently have? It's cooled down to RT and is currently lying in my garage.
So, start adding the catalyst *after* the frothing ends. Got it...

blogfast25 - 27-2-2013 at 06:15

The freezing of the t-butanol shows you're cooling too hard. Try just gentle air cooling instead of brute force cold water.

Also, try my 'one pot method': add the catalyst (ALL of it) at the start, mixed in with the rest of the ingredients. Then assemble and start heating.

Do you have any temperature probe anywhere? Excessive foaming could point to too much heat. Start on a lower setting and gradually crank up the heat until you've got a gentle rolling boil. The catalyst will NOT instantly boil away, as you claim.

The frothing is almost certaily due to hydrogen evolution, in your case just a bit too fervently. But broadly it's a good sign.

I would keep the failing batch for a later date but start with fresh materials. You can always use the failed batch later on.

[Edited on 27-2-2013 by blogfast25]

elementcollector1 - 27-2-2013 at 08:40

For an air cooler, should I just leave the condenser ports open to the air? Or does air actually have to be blown through?

blogfast25 - 27-2-2013 at 13:16

For an air cooler, should I just leave the condenser ports open to the air? Or does air actually have to be blown through?

A little air, e.g, from a aquarium pump, would go along way. W/o any cooling there's no easy way of telling what temperature the cooler will end up at.

elementcollector1 - 27-2-2013 at 13:37

The original (@versuchschemie.net) used a wet towel around a glass tube. With a balloon over the top, this could work as a cooler, correct?

blogfast25 - 28-2-2013 at 14:03

Forget about the balloon. Neither use nor ornament. But I used exactly the same type of cooler. A modified graduated 25 ml pipette with some wet kitchen towel as coolant. Drizzle a bit of iced water on it from time to time.

elementcollector1 - 28-2-2013 at 18:15

Experiment 2: Failure. Initially used 5.1g of KOH, 2.4g of Mg, and eventually just added all of a 1mL t-BuOH / 2mL mineral oil solution. Nothing happened, so after 30 minutes, I put in more magnesium, bringing the total up to 3.3g. Still nothing. 1mL more catalyst. Still nothing. After an hour of nothing, I decided to add 0.9g more of KOH. Something. The solution before had a steady rate of gas, which I presumed to be t-BuOH boiling and recondensing, as well as possibly steam and hydrogen. When the KOH was added, now several large bubbles were appearing sporadically along with the small, steady ones. At this point dinner was calling, so I had to turn off the experiment.

I was worried the wet towel method might not have been strong enough, so I changed it out a few times with a condenser filled with water, and the two ports connected by one hose. Unfortunately, that warmed up as well...

All in all, I think I need a better condenser (one that has water flow, but the water isn't frigid) and possibly a longer reaction time (total time was 2 hrs.) I'm also worried about drizzling iced water on the towel, as it might drip down and sizzle on the hot sand (and scare the living daylights out of me).

Attempt #3: Use magnesium shavings from a camping firestarter instead of the 40 mesh powder I've been using. This should be technically slower, but given that I've seen no potassium after 2 tries, I think I'm going to start trying different stuff.

Hopefully I can re-distill t-BuOH from the reaction mixes, to prevent too much loss of catalyst...

[Edited on 1-3-2013 by elementcollector1]

blogfast25 - 1-3-2013 at 07:37

Experiment 2: Failure. Initially used 5.1g of KOH, 2.4g of Mg, and eventually just added all of a 1mL t-BuOH / 2mL mineral oil solution. Nothing happened, so after 30 minutes, I put in more magnesium, bringing the total up to 3.3g. Still nothing. 1mL more catalyst. Still nothing.

I was worried the wet towel method might not have been strong enough, so I changed it out a few times with a condenser filled with water, and the two ports connected by one hose. Unfortunately, that warmed up as well...

All in all, I think I need a better condenser (one that has water flow, but the water isn't frigid) and possibly a longer reaction time (total time was 2 hrs.) I'm also worried about drizzling iced water on the towel, as it might drip down and sizzle on the hot sand (and scare the living daylights out of me).

Hopefully I can re-distill t-BuOH from the reaction mixes, to prevent too much loss of catalyst...

[Edited on 1-3-2013 by elementcollector1]

30 minutes is a short time. Remember, first step in the reaction chain is to react the water from the KOH with the Mg: 2 H2O + Mg === > Mg(OH)2 + H2.

Using a wet towel does indeed cause some water to sizzle on the sand bath. Ain air cooled condenser is still the best solution, IMHO. Luke warm water would work but is a bit harder to engineer. Your condenser is just fine but your use of it isn't optimal.

Distilling t-butanol from half-finished mixes doesn’t guarantee what you’ll obtain. You’re setting yourself up for problems and uncertainties that way.

elementcollector1 - 1-3-2013 at 08:39

How about a low-speed fan or hairdryer of some sort, hooked up to a intake hose?
In the demo for the school next week, we'll be using water from inside (from a fume hood), so it likely won't be as frigid as the stuff I currently use.

blogfast25 - 1-3-2013 at 09:21

How about a low-speed fan or hairdryer of some sort, hooked up to a intake hose?
In the demo for the school next week, we'll be using water from inside (from a fume hood), so it likely won't be as frigid as the stuff I currently use.

Anything above 40 C should do it to avoid freezing the t-butanol.

elementcollector1 - 1-3-2013 at 10:29

Hairdryers can be pretty hot, so this could work. Although isn't the freezing point something like 28 C?

Anyway, did some stoichiometry calculations, and came across a major problem with the amounts used by both NurdRage and Pok in their respective syntheses: There's not enough KOH. Here are the assumed reactions between potassium hydroxide, tertiary butanol, and magnesium:
(CH3)3COH + KOH -> K(CH3)3COH + H2O
2K(CH3)3CO + Mg + 2H2O -> 2K + Mg(OH)2 + (CH3)3COH

Using 2.4g of Mg for the second reaction (as NurdRage did), I found that 22.16g of K(CH3)3CO was needed to completely react with the Mg. Using the first equation, this translates to roughly 11.1g KOH - over double the amount NurdRage used. I didn't calculate the ratios for Pok's synth, but based on these amounts alone I would say that a major problem for many is that there simply isn't enough KOH to react with all the magnesium and tertiary butanol. Also, where does the H2 come in? I tried balancing the reactions with MgO and H2 instead of Mg(OH)2, but couldn't do so.

Maybe I'm missing something, such as that it would actually benefit the reaction to have a severe lack of K.

Additional Note: Here is the brand of mineral oil that I have been using as solvent for my reactions:
http://www.scanavert.com/Mineral-Oil/Safeway/00321130780700/
The back of the label states "100% mineral oil", but this could become contaminated by water and air over the time I've kept it.

If this next test with different magnesium doesn't work (I'll assume it doesn't), what is a good solvent to use? I've seen baby oil, citronella oil (lemon-scented or some such?) and tetralin, as well as paraffin oil and candle wax (paraffin).

Any recommendations? From what I remember of reading the thread last night, the lemon-scented oil was proven to work (by blogfast25), tetralin by NurdRage, and Shellsol D70 / kerosene by other members (including Pok).

[Edited on 1-3-2013 by elementcollector1]

[Edited on 1-3-2013 by elementcollector1]

blogfast25 - 1-3-2013 at 11:10

Maybe I'm missing something, such as that it would actually benefit the reaction to have a severe lack of K.

[Edited on 1-3-2013 by elementcollector1]

Yes, you are. The quantities used by Pok, Nurdrage, woelen, me and several others are perfect.

Your error is that the t-butanol is only there as a catalyst. The molar ratio of t-butanol/KOH is approximately 0.1 (0.1 mol of t-butanol per 1 mol of KOH). And that is AS IT SHOULD BE!

In reality the amount of KOH is always slightly increased with respect to the 'theoretical' amount because ALL KOH contains about 10 % water.

As a catalyst, t-butanol does as catalysts do: it gets constantly regenerated so that you don't need much of it.

Trust me, the quantities prescribed are entirely correct and NOT the cause of any problems anyone might encounter.

For more info, look up the proposed reaction path higher up in the thread ( don't ask me where exactly...)

[Edited on 1-3-2013 by blogfast25]

elementcollector1 - 1-3-2013 at 11:23

I see. I also got the ratio of 0.01 mol t-BuOH per 0.1 mol KOH, so at the very least I'm doing something right. Anyway, I've decided to abandon the mineral oil as a solvent: To my knowledge, no one's reported a success with the likes of it.

Mailinmypocket - 1-3-2013 at 11:48

I see. I also got the ratio of 0.01 mol t-BuOH per 0.1 mol KOH, so at the very least I'm doing something right. Anyway, I've decided to abandon the mineral oil as a solvent: To my knowledge, no one's reported a success with the likes of it.

What kind of mineral oil are you using? I've tried quite a few types of solvents, and despite never getting any K that lasts, the best is lamp oil. At a camping or hardware store look for unscented paraffin -based lamp oil (4th bottle in the image below).

The other three, in order are mineral oil(laxative grade from the pharmacy

, baby oil(the identical one shown in nurdrages video) and odourless solvent sold from an art store. The mineral oil often got a waxy crust about 2 hours into the rxn. The baby oil was alright, but sometimes it got waxy as well. The unscented solvent supposedly was paraffin according to the MSDS, but boils violently above 180c.

Lamp oil would be best, odorless if possible, although I think I remember something about blogfast using citronella scented oil and getting K, so scented might work too!

blogfast25 - 1-3-2013 at 14:21

EC1:

With regards to the oil, the tocopherols quoted I don't like so much.

Try a clean, clear lampoil, preferably 'deodorised' or 'low odour'. Check boiling point (range) before using, it should be 200 or slightly more. If you want to be sure about water, dry with anhydrous soda or anhydrous MgSO4 (Epsom Salt). Mine came straight out of the bottle, though, the kind they sell for garden path oil lamps and such like.

[Edited on 1-3-2013 by blogfast25]

[Edited on 1-3-2013 by blogfast25]

blogfast25 - 1-3-2013 at 14:27

With regards to the citronella, that was me: I found cirtonella scented kerosene (slighly yellow) and because the citronella is only a minimal component I tried that kerosene. It worked but I didn't get any citronella scented potassium, just 'ordinary' K!

[Edited on 1-3-2013 by blogfast25]

elementcollector1 - 1-3-2013 at 14:47

EC1:

With regards to the oil, the tocopherols quoted I don't like so much.

Try a clean, clear lampoil, preferably 'deodorised' or 'low odour'. Check boiling point (range) before using, it should be 200 or slightly more. If you want to be sure about water, dry with anhydrous soda or anhydrous MgSO4 (Epsom Salt). Mine came straight out of the bottle, though, the kind they sell for garden path oil lamps and such like.

[Edited on 1-3-2013 by blogfast25]

[Edited on 1-3-2013 by blogfast25]

Guess I'll heat some Epsom salt to anhydrous, then. It goes to anhydrous around 200 C, so this is also a good test to see if my electric hot plate can reach that.

elementcollector1 - 1-3-2013 at 14:48

With regards to the citronella, that was me: I found cirtonella scented kerosene (slighly yellow) and because the citronella is only a minimal component I tried that kerosene. It worked but I didn't get any citronella scented potassium, just 'ordinary' K!

[Edited on 1-3-2013 by blogfast25]

Aww. Now that would have made an interesting story for the kids one day.
Parents are extremely nervous about kerosene and a hotplate, so I'm going to stick with lamp oil for now.

EDIT: Found the lamp oil! It's unscented, and perfectly clear (so much for the lemon flavor).

[Edited on 2-3-2013 by elementcollector1]

blogfast25 - 2-3-2013 at 04:17

Parents are extremely nervous about kerosene and a hotplate, so I'm going to stick with lamp oil for now.

EDIT: Found the lamp oil! It's unscented, and perfectly clear (so much for the lemon flavor).

[Edited on 2-3-2013 by elementcollector1]

Psst. Don't tell'm: lampoil is basically kerosene! Shhhh....

elementcollector1 - 2-3-2013 at 10:12

Psst. Don't tell'm: lampoil is basically kerosene! Shhhh....

I know... American education at its best.

elementcollector1 - 2-3-2013 at 17:42

(Warning: this post is picture-intensive)
Well, another failure. Although something did end up happening, no K was formed. Total reaction time was roughly four hours, starting from 12:30 and ending at about 4:15 (parents wanted to go on driving practice). Here are a bunch of pics:

The general setup (the bucket was swapped out for a larger one, and aluminum foil was later added to conserve heat, but the general idea stays the same).

Not pictured: The lamp oil was dried overnight with anhydrous MgSO4, with a cut balloon over the end of the flask to prevent too much air and water entering. I came back morning, and the white powder certainly seemed a little 'fluffier' (though it could have been just my imagination).

Ooh, shiny Mg turnings. These were drilled off a camping firestarter, and stored under oil for the night. The KOH is directly underneath, invisible under all the bright metal.

The catalyst mix: 1 mL t-BuOH / 2 mL lamp oil. Seemed to work perfectly, although there was always lots of bubbling whenever the catalyst was added. I accidentally spilled the remaining catalyst about halfway through, so I made a new one with half of what was used before. This might introduce some error, but then again, it is a catalyst.

And things start to get a little darker.

And darker.

The end of the reaction. The KOH/K-t-BuO / MgO / whatever has formed a crust on the bottom, while the 'burnt' Mg resides on top. No potassium globules were seen anywhere inside the flask, and the magnesium itself looks only half-reacted (not entirely consumed).

All in all, potential sources of error were:
-Magnesium metal may have been contaminated or otherwise rendered insufficient
-Solvent may have not worked (unlikely, as the t-BuOH dissolved fully).
-The catalyst may have somehow escaped the reaction (also unlikely, as reflux condenser was using outside water again.)Strangely, no frozen tert-butanol was seen, indicating either the cooling wasn't sufficient (yeah, right) or the t-BuOH was refluxing perfectly.

Thoughts? I thought I saw a similar problem a few pages up.

TL;DR - Failure number three, moving on to plan for failure number four!

[Edited on 3-3-2013 by elementcollector1]

blogfast25 - 3-3-2013 at 05:28

What temperature was your cooling water this time?
You're sure the catalyst ended up and stayed in the flask?

Mailinmypocket - 3-3-2013 at 06:12

All this talk of K has gotten me interested in trying again! Before that though, the last comments made by Pok further up thread are nagging in my mind. He said that my Mg (which was purchased as "Mg turnings for Grignard reactions") seems to be very coarse, and that the lack of surface area may be the cause of my failure. I have to admit that this Mg is coarser than most Grignard turnings I have seen...

Before deciding to try again, or call it a lost cause, what do you guys think of this Mg? I've attached an image with a ruler for scale. I'm inclined to think that Pok was correct, but has anyone used a similar size Mg with success?

elementcollector1 - 3-3-2013 at 10:40

What temperature was your cooling water this time?
You're sure the catalyst ended up and stayed in the flask?

It wasn't quite as frigid as the first time, but more than that I cannot say. The water flow was also a bit slower, so the water may have been 'warmed up' more by the vapors.

elementcollector1 - 3-3-2013 at 10:41

All this talk of K has gotten me interested in trying again! Before that though, the last comments made by Pok further up thread are nagging in my mind. He said that my Mg (which was purchased as "Mg turnings for Grignard reactions") seems to be very coarse, and that the lack of surface area may be the cause of my failure. I have to admit that this Mg is coarser than most Grignard turnings I have seen...

Before deciding to try again, or call it a lost cause, what do you guys think of this Mg? I've attached an image with a ruler for scale. I'm inclined to think that Pok was correct, but has anyone used a similar size Mg with success?

That does look pretty coarse. Are you sure you can't get some powder?
I used turnings in my last reaction, and they didn't work out so well, so I'd assume the more surface area the better.

Mailinmypocket - 3-3-2013 at 10:56

That does look pretty coarse. Are you sure you can't get some powder?
I used turnings in my last reaction, and they didn't work out so well, so I'd assume the more surface area the better.

The only other fine Mg I have is that intentionally oxide-coated shit that many experiments were done with previously. Sold to me as "fine magnesium" but when I received it the MSDS claimed a percentage of added oxides for flame retardence :mad:

I tried etching it with acid, didnt work. As you can see, it is quite fine, I'm sure it would work if it was not coated. I tried grinding it in a mortar and pestle to expose fresh surfaces also- nothing. I can get powder, I just expected turnings to be more like turnings and not pea-shooter sized bits of Mg.

[Edited on 3-3-2013 by Mailinmypocket]

elementcollector1 - 3-3-2013 at 11:02

...Put it in a coffee grinder?

elementcollector1 - 5-3-2013 at 10:17

Also, I have some magnesium powder just like that. It's very shiny, flammable (burns brightly), and doesn't state of any magnesium oxides / aluminum oxides on the label. It was obtained from Elemental-Scientific LLC. Due to the high surface area, it should be good for this reaction, but unfortunately the granules also look like tiny bits of potassium (making it hard to tell for me until the potassium gets bigger).

I'm fairly sure the catalyst managed to stay in the flask, as it still had that distinct aroma of tert-butanol when I turned off everything and opened the flask.

Still, I just remembered that I didn't remember to inject 0.4mL of catalyst at reflux, instead just going with the usual 0.1mL increments. This could contribute to loss of catalyst and potassium.

Additionally, I wonder if I could dilute the catalyst solution even further to prevent loss on the condenser (by virtue of the lamp oil's significantly lower freezing point). The typical ratio I use is 1mL catalyst:2mL solvent, as pure catalyst always freezes on the way down. This might not be a problem when I'm using water from a fume hood, but it certainly is when I'm using water from the outside.

blogfast25 - 5-3-2013 at 12:55

Additionally, I wonder if I could dilute the catalyst solution even further to prevent loss on the condenser (by virtue of the lamp oil's significantly lower freezing point). The typical ratio I use is 1mL catalyst:2mL solvent, as pure catalyst always freezes on the way down. This might not be a problem when I'm using water from a fume hood, but it certainly is when I'm using water from the outside.

There should be no problem with that.

elementcollector1 - 5-3-2013 at 16:18

Success?
I definitely saw something. Actually, quite a lot of somethings.
Basically, about halfway into adding catalyst (so, about half an hour after reflux) I noticed these absolutely tiny silvery bubbles that would float to the top, and then, without popping, quickly fall back down again. I have a few reasons why they just might be potassium:
1) As per the last experiment, the magnesium significantly darkened, even when the catalyst was just starting to be added. These things were shiny silver.
2) It was incredibly hard to tell, but they appeared to have a liquid, spherical nature - again, molten potassium.
3) They never popped, so they certainly weren't like the air bubbles that were also showing up regularly and quickly.

No pics yet, as the reaction is still going as I post this.

elementcollector1 - 5-3-2013 at 18:14

I hate to double post, but this is just frustrating. Took the flask and condenser off the heat plate, to find a cloudy solution, a small amount of unreacted magnesium, and a bunch of white fluff.
If the magnesium got consumed, which it did (there was significantly less by volume inside the flask), shouldn't this reaction be working?

EDIT: Going to continue reflux for another 2 hours. Now that I think about it, what if all those granules really were K? They easily could be, given their shiny appearance... but they could just as easily be unreacted Mg!

[Edited on 6-3-2013 by elementcollector1]

condennnsa - 5-3-2013 at 23:20

i bet your magnesium is the problem. to me all this evidence supports my 'magnesium poison' theory I posted farther up thread

blogfast25 - 6-3-2013 at 05:14

I hate to double post, but this is just frustrating. Took the flask and condenser off the heat plate, to find a cloudy solution, a small amount of unreacted magnesium, and a bunch of white fluff.
If the magnesium got consumed, which it did (there was significantly less by volume inside the flask), shouldn't this reaction be working?

EDIT: Going to continue reflux for another 2 hours. Now that I think about it, what if all those granules really were K? They easily could be, given their shiny appearance... but they could just as easily be unreacted Mg!

[Edited on 6-3-2013 by elementcollector1]

You should really try and let it run its full course, at least untill you've established with certainty you've got K. Oxygen could destroy a crop of embryonical K. Patience, EC1!

elementcollector1 - 6-3-2013 at 08:35

Quote: Originally posted by condennnsa

i bet your magnesium is the problem. to me all this evidence supports my 'magnesium poison' theory I posted farther up thread

I agree. The magnesium shavings at least reacted, darkening significantly in color. These things were untouched.

blogfast25 - 6-3-2013 at 10:54

I see a lot of people pointing the finger at this or that but precious little documented evidence for these assertions. It's bit like: 'I was eating pasta the other day and got a headache, ergo eating pasta causes headaches'. Not quite.

Unless someone makes a clear back-to-back comparison between an Mg that works and one that doesn't, this is all highly speculative and a potential wild goose chase to boot.

[Edited on 6-3-2013 by blogfast25]

elementcollector1 - 8-3-2013 at 13:01

I see a lot of people pointing the finger at this or that but precious little documented evidence for these assertions. It's bit like: 'I was eating pasta the other day and got a headache, ergo eating pasta causes headaches'. Not quite.

Unless someone makes a clear back-to-back comparison between an Mg that works and one that doesn't, this is all highly speculative and a potential wild goose chase to boot.

Well, if you say so. My reasoning is that under the same conditions, one Mg noticeably reacted and the other didn't. My hypothesis is that one grade of Mg is contaminated or otherwise rendered unusable. Data? 3 (5th trial as we speak) different occasions on which the Mg in question failed to react, 1 trial where the other Mg did react, darkening significantly in color.
I currently have a much-improved, ground-glass setup. Some things I added:
-A condenser with a much wider joint (due to the ground glass connection). There was an odd capillary effect with the old condenser, where liquid would get sucked up and then remain in the bottom of the condenser's outlet tube, dripping out only occasionally. The ground glass one has a much wider tube, removing this effect.
-Modified the catalyst mixture to be a 1:1 mix of lamp oil:alcohol, and added it all at once this time (although the sep funnel stopped working after a while, maybe due to pressure, I was able to pipette it in).
-It's a sunny day outside! This might help with heating, and ensuring t-BuOH doesn't freeze in the condenser (though it still might, I haven't had that problem since trial 1).
Going to run this reflux for 6 hours, and check back in on it.

My only question for this time around:
-What happens if you leave the reaction on 'too long'? Could the potassium get destroyed again by incoming water vapor or oxygen, or would it stay under the lamp oil? (I suppose that depends on density.)

Mailinmypocket - 8-3-2013 at 17:57

Okay, let's compare results and see what happens. Obviously we have different grades and types of Mg and solvents, but if we run the exact same quantities of reagents etc it might shed some light on what makes this reaction tick, or not. Are you still using the proportions of reagents you mentioned up-thread? I will run this tomorrow, following the original procedure as closely as possible, and adding all of the t-BuOH catalyst in the beginning as opposed to dosed additions.

I am thinking its either the Mg surface area or amount of catalyst. KOH is pretty invariable, so is the solvent... As long as it satisfies the same properties as Shelsol etc it should be okay. I have kept a sample of commercial K under lamp oil for months with no degradation so far, the oil obviously doesn't react with the K to an extent that is obvious. I may pass argon through the system, but then again, I don't think it is all that necessary given that nobody else has had to do it

elementcollector1 - 8-3-2013 at 19:01

No pictures, unfortunately. Rest assured I let this thing go for 6 hours.
This time, I added all the catalyst at once. Nothing.
Also, apparently that 1:1 ratio of catalyst to solvent wasn't good: some alcohol noticeably froze in the condenser.
Oddly, the Mg seemed to react in this one, but it simply disappeared, possibly into the white, sludgy mix. Or it's hidden inside.
If I do a next run (I'm about ready to give up after 5 tries in a row), I'll probably make sure everything is EXTRA dry, and do this inside.
Talked to my Science Club about this, and the coach said this might be due to the lamp oil boiling before 200 C.

I had an idea: What if one makes K-tert-butoxide, dries this, and reacts this with magnesium? Wouldn't this skip the intermediary KOH step?

blogfast25 - 9-3-2013 at 02:13

I had an idea: What if one makes K-tert-butoxide, dries this, and reacts this with magnesium? Wouldn't this skip the intermediary KOH step?

KO t-Bu (potassium t-butoxide) is hard to make from KOH. And you need a whole lot more t-butanol to pull it off. There's an example in the patent, using a t-butoxide.

I'm baffled by these fails...

[Edited on 9-3-2013 by blogfast25]

elementcollector1 - 9-3-2013 at 12:03

So am I. I'll have to do a check of everything to make sure everything is what it says it is. So far:
-KOH was proven to be KOH (precipitated Cr(OH)3 from K-alum upon boiling). Experiment by flame test, just to check?
-Mg powder was proven to be flammable and reactive, at least upon ignition (earlier experiment with MnO2 and Mg. Failed that thermite too.)
-t-BuOH was proven to be t-BuOH by freezing point (and I trust Mailinmypocket anyway).
-Lamp oil is still a bit unknown. Boiling point could be above or below 200 C, and parents refuse to give me thermometer. Ideas?
-Hot plate can reach 200 C, as evidenced by decomposing MgSO4.7H2O to the anhydrate.
-Glassware is secure, as is aluminum foil.
-Sand bath is working fine, as is condenser flow (although I think I might have overdid it on the last run...)

Here we go again!

Mailinmypocket - 9-3-2013 at 12:03

Alright- so I just started the reaction, using the same amounts of reagents as Pok, sandbath and damp filter paper user to cool the glass tube condenser. I decided to add all the t-BuOH at once before heating, just to switch things up a bit. About 20 minutes after heating there was the usual white fumes, and now there are about 30 very small K balls buzzing around. I'm not excited about it though since this has happened almost every time. Ill report back on the outcome in a few hours, we shall see!

If it doesn't work then it is either my Mg or the baby oil (the same one nurdrage shows in his video- did he actually try it!?) If this attempt fails I will go to work and get some tetralin, if that fails then the Mg is apparently the problem.

elementcollector1 - 9-3-2013 at 12:06

Alright- so I just started the reaction, using the same amounts of reagents as Pok, sandbath and damp filter paper user to cool the glass tube condenser. I decided to add all the t-BuOH at once before heating, just to switch things up a bit. About 20 minutes after heating there was the usual white fumes, and now there are about 30 very small K balls buzzing around. I'm not excited about it though since this has happened almost every time. Ill report back on the outcome in a few hours, we shall see!

If it doesn't work then it is either my Mg or the baby oil (the same one nurdrage shows in his video- did he actually try it!?) If this attempt fails I will go to work and get some tetralin, if that fails then the Mg is apparently the problem.

Hold the freakin phone. You added the alcohol before any heating? That's... actually a pretty good idea.
I always waited until reflux. Also, this might be too late (didn't check your post time), but swirl the flask a bit, maybe it'll coalesce all those K-balls into a larger, more-resistant-to-attack one.

Mailinmypocket - 9-3-2013 at 12:12

It's actually Blogfast that originally did the "one pot" reaction and had success, otherwise the times addition seems to be the norm, but I'm just playing with different variables right now to see what happens! The t-butanol I sent you is indeed not the problem, it's reagent grade from a supplier. I just swirled the flask a bit now, little micro balls are present but no major changes. Time to have a beer and visit the flask later and see how things are going

what lamp oil are you using? Maybe we can find an MSDS or something...

Edit- Okay, well, the odor of failure is strong in the air. Again. I'm faily convinced at this point that the Mg is the culprit. Despite it being reagent grade blah blah blah, it's too coarse, not enough surface area is present. This may explain why at first I always see K, and it then disappears. At first, enough fresh Mg is exposed to drive the system forward, however once the surfaces react things start going backwards due to an excess of catalyst that can't react with any fresh Mg. there is enough Mg, just not in the right form.

That's my crude theory anyways... Ill let it go 4 hours but I'm sure nothing will become of it.

[Edited on 9-3-2013 by Mailinmypocket]

blogfast25 - 9-3-2013 at 12:46

. I just swirled the flask a bit now, little micro balls are present but no major changes. Time to have a beer and visit the flask later and see how things are going

what lamp oil are you using? Maybe we can find an MSDS or something...

[Edited on 9-3-2013 by Mailinmypocket]

From micro balls, mighty balls of K grow. Initially, the K is indeed a fine dispersion in the rest. And it takes 2h or more to get reasonable size K globules.

Nurdrage also used 'one pot' a few times and it worked. He did claim to see more variability in the results. Personally, going by the assumed reaction path, I can't see why adding the catalyst all at once and before heating would make any difference at all (compared to the hocus pocus 'timed addition'). So I tried one pot and it worked just the same...

EC1: I really would invest in a temperature probe: temperature has a strong effect on most reaction speeds. It appears to be the one thing you're not controlling as most of us do or try...

[Edited on 9-3-2013 by blogfast25]

Mailinmypocket - 9-3-2013 at 12:48

Ah, you posted before I finished editing the last post! Actually... No K is visible anymore... I'm sure the Mg is the problem

blogfast25 - 9-3-2013 at 12:52

Ah, you posted before I finished editing the last post! Actually... No K is visible anymore... I'm sure the Mg is the problem

If K really formed, then disappeared how do you arrive at the conclusion that the Mg is at fault? Bear in mind that others have had trouble with powdered Mg too and that my own Mg powder is only as fine as sand.

[Edited on 9-3-2013 by blogfast25]

Mailinmypocket - 9-3-2013 at 13:01

Ah, you posted before I finished editing the last post! Actually... No K is visible anymore... I'm sure the Mg is the problem

Well, I'm just blaming the Mg since it is the only reactant that I have had problems with since the start. At first it was that Ca, Mg oxide coated shit, and then grignard turnings that are more like thick chips of Mg. I don't see why the solvent is the issue, I have had small bits of commercial K stored under both baby oil and lamp oil for almost 3 months now, no major changes in appearance. Unless something is happening at the solvent/catalyst level I am not aware of...

The KOH has been varied, from technical grade flakes to ACS pellets, nada difference. It appears as though ever since using the recent Mg I always get small K at first and it disappears, everything else has been the same as far as temperature, solvents(tried 3).

I can't say for sure that the Mg is the problem, I'm just leaning towards that conclusion really...

[Edited on 9-3-2013 by Mailinmypocket]

elementcollector1 - 9-3-2013 at 14:27

What about catalyst? What comes to mind is that the K forms, but when the Mg runs out the catalyst turns right around and starts consuming the K metal.

elementcollector1 - 9-3-2013 at 14:36

what lamp oil are you using? Maybe we can find an MSDS or something...

I've attempted using 2:
"Nowell's Lamp Oil":

"Sunnyside Crystal Lamp Oil":

Mailinmypocket - 9-3-2013 at 14:36