Sciencemadness Discussion Board - Allyl alcohol in 70% yield

Allyl alcohol in 70% yield

bleckster - 10-12-2013 at 22:34

The published methods for allyl alcohol are well known as poorly yielding, with oxalic acid and glycerin yielding 20-30% and formic acid and glycerin reportedly yielding 40%. I tried the oxalic method first because I had little formic left and got basically nothing in the very end. Upon seeking an improvement, I came across a patent application which reported to give 80% yields using formic acid. The basic change is that you do the reaction in an inert atmosphere. Doing it in air causes excessive loss through conversion to carbon dioxide, as well as causing the glycerin to turn brown. The following synthesis is my first try, by which I got a 70% yield (0.42 mol), calculated on the total amount of glycerin used in the reaction (0.6 mol). It never turned brown.

55.25 g (600 mmol) glycerin and 16.39 g (356 mmol) 90% formic acid were charged in a 2-neck 100 ml flask fitted with a thermometer and a claisen adapter with an addition funnel fitted on one arm and a distillation apparatus on the other. No stir bar was used. An argon filled baloon was fitted to the top the funnel so that its stopcock could control the addition rate of the gas. The vacuum port of the distillation apparatus was run to a bubbler tube submerged in a 0.1M NaOH solution. The thermometer adapter was temporarily removed from the second neck of the flask and a glass tube was inserted into the reactants and argon was bubbled through them slowly for about one minute and then the thermometer replaced. The argon from the balloon was then used to slowly replace the air in the system. Heat was applied and boiling began at 120C internal temp with the distillate coming over at 98C. The argon stopcock was closed as the natural gas evolution from the reaction was enough to keep positive pressure in the system. At 220C rapid carbon dioxide evolution from the system was observed. The heat was carefully controlled to bring the internal temperature to 235C where it was held for 25 minutes until the drops were 10 seconds apart and CO2 evolution had started to slow. Heat was discontinued and the contents cooled to room temp. At all times the reactants and products remained completely colorless. The distillate weighed 25.15 g and was saved in a tightly stoppered flask. A second portion of 11.69 g (254 mmol) 90% formic acid was charged into the reaction flask and another distillation cycle repeated as before, this time with heating on high. A first fraction of ~5 mL distillate was collected between internal temp 120-180C, a second starting around 220C accompanied by continous evolution of CO2 from the bubbler at a rate of 4-5 bubbles/sec as before, at which point the heat was reduced to keep the temp from rising any higher. When the temp reached 240C, distillation was stopped and everything was cooled. The distillate weighed 18.9 g. Another 11.69 g (254 mmol) of formic acid was charged and the cycle repeated producing another 16.45 g. The small glycerin residue in the flask had a slight yellow tint at this point and was discarded.

The combined distillate, weighing 60.4 g, and containing small amounts of allyl formate, was refluxed with 500 mL 20% NaOH solution for 1 hour to hydrolyze the formate and then fractionally distilled through 2 vigreaux columns (20 plates) with the distillate coming over at 87-90C. CaCl2 was added in ~5g portions with the contents turning cloudy, being warmed on a low hotplate, until the CaCl2 had been dissolved, stopping when two distinct layers had formed. The bottom aqueous layer was separated in a funnel and the drying process repeated with the top layer, this time not forming two layers and the CaCl2 not dissolving completely but the solution still remaining cloudy. A fraction weighing 24.62 was distilled at 97C.

Some notes and comments:

Original source: http://www.google.com/patents/WO2008092115A1

With the first oxalic acid method I tried, I made every mistake possible, including letting the temperature get too high because I used a partial immersion thermometer to monitor the reaction temp and it wasn't inserted to the right depth and read about 20-30C too low, so in the end the glycerin broke down into a charred black mess. So I was much more careful for this reaction. The original source is clear about 235C being the ideal temp, so on the first distill I brought the temperature up very, very slowly and held it at 235C for as long as possible. On the second and third, since the cycles didn't last as long, it was harder to prevent it raising above 235C. I think I could have gone to 240 before stopping that first one and gotten more product. The full immersion thermometer I used for reading the internal temperature turned out to be 2 degrees low, which I discovered a few days later, so I think I probably could have taken all of the distill cycles a tad further. It's worth noting that the increase in reaction temperature happens very slowly as more and more glycerin is reacted and converted, so each degree of increase means that much more product. It was only a first try, so I played it safe, but I do think it's entirely possible to get to 80% yield or probably more if 95% formic acid is used.

I forgot to weigh the remaining glycerin post reaction to assess how much had been reacted, so my yields are based on the theoretical amount from starting material. I need to break that bad habit. I do remember that it wasn't much, about ~5 mL.

The distillate before the workup contains a small amount of allyl formate which is responsible for the unbelieveably overwhelming lacrymatory effect. If you pour this into an open beaker and leave it on your bench, you will have to clear the room a minute later and even after you wash your face it still lingers. You'll have to take a shower to really get it off. As long as you keep all the flasks stoppered when weighing and storing inbetween cycles and work in a well ventilated area, you'll be ok. I actually did the main reaction outside of my hood, due to the closed system, but I did the workup in a fume hood. Basically, before you do the workup and hydrolyze the allyl formate, you should be extra careful about exposure (cleaning the flasks is fun). Once the formate has been converted to alcohol, it's ok. The watery alcohol actually doesn't have much of a lacrymatory effect when you smell it, however once you dehydrate it the effect is more intense and a good whiff will send you to wash your face at least.

I used CaCl2 to dry the alcohol because that's what I saw in the published literature. I was actually not familar with using CaCl2, as I always use MgSO4, which is very quick acting and quick to pass through a coffee filter, but I was unsure if MgSO4 would react with the alcohol. When I added the CaCl2 the solution turned cloudy and I realized I would have to use thicker qualitative paper, which would have taken a long time. It said to distill from CaCl2, which I did, however by the end it had puffed up quite a bit and I possibly lost some product in there. I think I'll try MgSO4 next time.

I don't have a schlenk line or a bubbler, or even a gas tank (just a can of compressed argon marketed for saving your opened paint cans and bottles of red wine), so I had to rig up an alternative. I thought someone out there might find it helpful, but just in case my description was unclear, here is what it looked like:

<img src="http://i750.photobucket.com/albums/xx142/thebleckster/IMG_0136_zps49d1f68e.jpg" width="800" />

[<a href="u2u.php?action=send&username=bfesser">bfesser</a>: reduced image size(s)]

[Edited on 11.12.13 by bfesser]

DJF90 - 11-12-2013 at 00:49

Nice work Bleckster. I have seen a similar method used in a journal, which employed a flow of nitrogen gas through the reaction mixture to acheive the same result. I had been meaning to give it a go under static argon, as you have, but I haven't had the time so I'm glad you've done some of the leg work for me. Kudos!

EDIT: Be careful how you store your product. Some people aren't aware that allyl alcohol forms peroxides...

[Edited on 11-12-2013 by DJF90]

Dr.Bob - 11-12-2013 at 07:21

That is a pretty amazing job given the simple setup. Allyl alcohol is a very useful chemical, and glycerol is very easy to come by, especially now that so many people make biodiesel, which generates a lot of it, and they are always looking for something to do with it. If you make the chloride or bromide from that, it would be another very useful chemical, just used some a few weeks ago. Very well done.

bleckster - 11-12-2013 at 08:55

Thank you very much. Yes, I did make the bromide using the procedure in Vogel's. Question though, I notice that Sigma-Aldrich sells allyl bromide with propylene oxide as a stabilizer:

http://www.sigmaaldrich.com/catalog/product/aldrich/a29585?l...

Does anybody know WHY they put a stabilizer in? I've noticed that most things that do need stabilizers react with the oxygen in the air inside the container, like ether, which forms peroxides. Also catechol, which turns brown or sodium ethoxide, which turns yellow. I've used the compressed argon to void the containers of those last two and capped them tightly and so far they have not discolored (last few weeks). Can the same be done with the bromide if I store it? I can't find any information on what happens to it and why.

Thanks.

Nicodem - 11-12-2013 at 09:54

Thank you for the excellent report.

Epoxides are commonly used as HBr scavengers to inhibit the decomposition of alkyl halides. Without any stabilizer, allyl bromide, just like many other very reactive alkyl halides, undergoes decomposition. This can be relatively rapid at room temperature and what's worse it is autocatalytic. Once it starts, and it can start any time, it tends to go to the end.
Other typical stabilizers are anhydrous Na2CO3, K2CO3, amylene, silver foil, and few others, depending on the compound. I suggest you to store allyl bromide over powdered potassium carbonate.

Magpie - 11-12-2013 at 11:44

Yes, thanks for this detailed report. I made allyl alcohol a few years back using oxalic acid and glycerine. My yield was a very disappointing 18.7%.

Paddywhacker - 11-12-2013 at 18:45

Glyceryl oxalate is something that I have been meaning to try making by Fischer–Speier esterification, using an ion exchange resin as catalyst and a Dean-Stark trap to remove the water. I wonder if this ester would be a better starting material than the glycerol/oxalic acid mixture that was mentioned.

bleckster - 12-12-2013 at 11:18

Excellent Nicodem, thank you for that--very helpful as I wouldn't have known where to get propylene oxide very easily!

Also, to everyone else, I think I will try this reaction using oxalic acid and see what the results are, seeing as I have a bunch of it to spare. I will report the results soon.

[Edited on 12-12-2013 by bleckster]

88% yield of allyl alcohol from 95% formic acid

bleckster - 29-12-2013 at 21:23

So I tried using oxalic acid for this reaction, however I got the same results as doing it in air--basically nothing. Avoid oxalic acid for this in general.

I did however do another run of the first reaction I posted, this time with 95% formic acid, scaled up x4, and got 88% yield! Very cool.

Also, regarding the workup, I did some solubility tests of different desiccating agents for the purposes of salting out the water. MgSO4 isn't very helpful, nor is sodium carbonate. Calcium chloride does work, however it is very slow to react and it makes for a milky white water layer. Potassium carbonate, however, is the clear winner. It is immediately reactive and dissolves cleanly with immediate precipitation of two clear layers even after adding just 10g. Of course, you need to completely saturate and salt out the water layer (which is on the bottom, btw), so just keep adding more in 10g increments until no more dissolves with shaking.

After reacting 3.1 moles of glycerin and hydrolyzing the distillate overnight and distilling, I got about ~400 mL of the 72% allyl alcohol/water azeotrope between 87-97C. It took ~110g K2CO3 to completely salt out the water layer. The top alcohol layer was separated, a final 10g K2CO3 added, and again fractionally distilled, yielding 158.7g pure allyl alcohol.

Be careful when cleaning your flasks!

Heuteufel - 30-12-2013 at 04:03

I am very impressed by the yields you obtained, however I am not sure, if your assumption, that the final product you obtained is nearly 100 % allyl alcohol (anhydrous) is correct. A titration of your product with a bromine solution would easily show how much allyl alcohol the liquid you obtained contains. Here a quote from the Organic Syntheses website (Org. Synth. 1921, 1, 15):

Quote:

The 195–260° fractions of the distillates are treated with potassium carbonate to salt out the allyl alcohol and to neutralize the little formic acid present. This allyl alcohol is then distilled and the fraction boiling up to about 103° is collected, or if a column is used, up to 98°. In this way, 845 g. of an allyl alcohol is obtained, which by bromine titration shows a purity of about 68–70 per cent (Note 4). This is equivalent to 570–590 g. of pure allyl alcohol (45–47 per cent of the theoretical amount) (Note 5). The alcohol may be made practically anhydrous by refluxing with successive portions of fused potassium carbonate until no further action is observed. The carbonate will remain finely divided and will not become sticky when water is absent. A considerable amount of allyl alcohol is lost mechanically during the drying in this way, so that the potassium carbonate which is used here should be employed for the salting out of fresh portions of allyl alcohol in the first part of subsequent preparations. The allyl alcohol thus produced is dry enough for all practical purposes (98–99 per cent), and it is unnecessary to dry with lime or barium oxide as advised in the literature in order to remove all the water. The allyl alcohol obtained by this process boils at 94–97°.

If your product is only e.g. a 70 % solution of allyl alcohol, your yield will of course be lower (but still pretty high).

For the bromine titration:

Quote:

4. To determine the purity of any sample of allyl alcohol, 1 cc. is run into 15–25 cc. of carbon tetrachloride and this solution is then treated in the cold with a carbon tetrachloride solution of bromine (standardized with potassium iodide and sodium thiosulfate) until a permanent bromine coloration is obtained. The amount of allyl alcohol present in any solution may also be determined roughly by conversion to allyl bromide. From several experiments it was found that the allyl bromide obtained was equivalent to the amount of allyl alcohol as determined by bromine titration.

Chloroform is also fine, if you don't have any carbon tetrachloride.

[Edited on 30-12-2013 by Heuteufel]

DJF90 - 30-12-2013 at 06:21

An initial improvement in allyl alcohol content can be made by conducting the initial distillation at reduced pressure after adding the potassium carbonate. In this way the aqueous distillate contains as much as 90 % allyl alcohol, viz. the usual 70 %. This is mentioned in the notes/discussion of the OrgSyn article.

What I do find interesting is they make no mention of the allyl formate byproduct. I've seen elsewhere that the ratio of allyl alcohol to allyl formate is 100:13.

bleckster - 31-12-2013 at 01:12

Quote: Originally posted by Heuteufel

I am very impressed by the yields you obtained, however I am not sure, if your assumption, that the final product you obtained is nearly 100 % allyl alcohol (anhydrous) is correct. A titration of your product with a bromine solution would easily show how much allyl alcohol the liquid you obtained contains.

Hmm, you are probably right in that it's not completely, absolutely anhydrous, as I did not *reflux* with potassium carbonate to make sure it was completely unreactive, I just shook it until it would not dissolve anymore. Bromine titration sounds like way too much trouble to bother with.

Maybe we can use some simple logic here:

http://eweb.chemeng.ed.ac.uk/jack/newWork/Chemeng/azeotrope/...

The azeotrope of allyl alcohol/water contains 72.9% allyl alcohol and boils at 88C, so if I'm collecting everything during a fractional distillation up to the BP of allyl alcohol at 97C, then I should actually theoretically have more than 72.9% in the distillate, depending on what percentage of the distillate comes over at 97 (I didn't determine this).

So if I'm starting the salting out process with 73% and I salt out quite a big water layer (that actually ends up looking like *half* the original contents) and then separate and discard that water/potassium carbonate (I also did not weigh this aqueous layer, however I will do that next time and determine how much water there is in addition to the weight of potassium carbonate added), this should get me pretty darn close to 98-99%, considering the fact that, at this point, I added potassium carbonate and shook hard and it didn't clump or get sticky.

Or is there something I am misunderstanding...? I actually do want to know when I have something wrong--I'm interested in knowledge, not praise! ;-)

The one thing I don't really understand is how the authors of the original workups from 70 years ago determined how much water to add the initial mixture to in order to hydrolyze the formate. I've just been going off of their amounts of alcohol/formate/water to water/NaOH. In one case I reduced the water in half while keeping the NaOH the same and it still worked. Does anybody have any information on this?

Scr0t - 31-12-2013 at 01:59

Take a density reading, for example take a small (50ml) volumetric flask of the sample and weigh it at 20°C with a scale accurate to +/-0.01g. If I remember correctly the density will be accurate to about +/-0.002g/ml.

Compare to the density table given here.

[Edited on 31-12-2013 by Scr0t]