Author Topic: Why the nitrite wacker "doesn't work"  (Read 3253 times)

0 Members and 1 Guest are viewing this topic.

lab_bitch

  • Guest
Why the nitrite wacker "doesn't work"
« on: April 16, 2002, 10:58:00 AM »
After about four months of fucking around with it, I think I finally figured out why no one can get even half the yield that the patent claims for this method (+90%).  For those of you who aren't that familiar with this method, here's a brief overview of it.

Basically, there are three different reactions going on here at the same time.  The first one is the oxidation of safrole to the acetal and reduction of Pd+2 to Pd0.  Now, the expensive Pd must be reoxidized in order to be catalytic, so Cu+2 is added.  Two Cu+2 are both reduced to Cu+1 and Pd0 is reoxidized to Pd+2.  Since a molar amount of Cu+2 is not very feasible, a secondary reoxidant is added to reoxidize the Cu+1 to Cu+2.  The first substance used for this was molecular oxygen.  While this method should work fine, there is another substance that is easier for the ghetto chemist to use.  These are the alkyl nitrites.  Alkyl nitrites are esters of nitrous acid and an aliphatic alcohol.  Due to a very weak N-O bond, these molecules easily oxidize Cu+1 to Cu+2 by homolytic splitting, releasing NO gas and the aliphatic alcohol that the nitrite was made from.

As long as there is alkyl nitrite present, this cycle will continue until all of the safrole has been oxidized.  If, however, the alkyl nitrite is depleted before all of the safrole is converted, the Cu+1 cannot be reoxidized.  Since Cu+1 has very low solubility in the reaction mixture, it then precipitates and cannot be reoxidized, even if more alkyl nitrite is added.  After all of the Cu+2 becomes depleted, there is nothing to reoxidize the Pd0.  It does not necessarily precipitate, but does become inactive.  To make the story short, the reaction is over.

Now, one would think that if he were to add 2.5 mol alkyl nitrite per mol safrole (2 mol is the stoichiometric amount) that the reaction would go to completion just fine.  Well, your wrong.  I did this about five times and never got more than 50% conversion.  Supprisingly, upon hydrolyzing the reaction mixture at the end and heating to 50C for an hour, no ethyl nitrite (that is the nitrite that I was using) was evolved.  I happen to know that if fresh ethyl nitrite is added to the reaction right before hydrolysis, it boils off during the heating.  In other words, there was no ethyl nitrite in the post reaction mixture.

Now, I was getting really confused about this whole thing.  If the ethyl nitrite wasn't in the post rxn mixture and it didn't all react, then where the hell did it go?  Did it evaporate?  Did it decompose?  Well, last night, I finally discovered the answer to my question.  Before starting the rxn, I weighed the entire flask.  It weighed about 276 g.  After the rxn was finished (before hydrolysis), I again weighed it.  It was now about 260 g.  Sixteen grams of something had left the flask in the exhaust gas.  Doing some quick calculations, I figured that if all the safrole had been converted and 2 mol of NO gas was given off for every mol of safrole (I was using 20 g), then between 7 and 8 grams of NO gas would be given off.  So what the fuck were the other 8 grams???  I had been running an ice cold gas condenser between the rxn flask and the exhaust hose, so no solvent or ethyl nitrite could have escaped.  Or could it have?  After consulting my thermodynamics book, I came to the following conclusion.  Suppose that the vapor pressure of the ethyl nitrite is 200 mm Hg in the reaction mixture.  That means that for every mol of NO produced, 0.36 mol of ethyl nitrite also comes off.  That's assuming ideal gas/solution behavior.  By no means are these numbers accurate; I'm just trying to prove a point.  This goes to show that a very good amount of the original ethyl nitrite could end up being carried off with the NO gas stream.  Of course some of the ethanol solvent could also be carried off, and that must be considered as well.

Putting all of this aside, the assumption that a lot of the ethyl nitrite escapes from the reaction mixture supports my observations very well.  Here is an overview of them.

1) After about an hour, the copper co-catalyst precipitated and the reaction stopped, even though not half of the safrole had been converted.

2) While hydrolyzing, no ethyl nitrite escaped, even though not all of it had reacted.

3) The weight loss of the flask did not correspond to the amount of NO gas that would have been generated to produce the conversion that I obtained.

So why are none of us ghetto chemists able to pull this method off, yet over in Japan, they are getting +90% yields?  Simple.  They do their reaction in a closed vessel.  All of the contents stay in the reaction vessel, and pressure builds up as the rxn progresses.  They even hint at this, saying that after completion, the rxn contents can be distilled to obatain *NO*, Alcohol, Safrole, and Ketone.

Now that I've told you why it doesn't work, I guess I have to tell you how to make it work.  Right now, I don't really know for sure, but if my theory is correct, this is how I would proceed.  Start the reaction like you normally would, adding 2.5 mol EtONO per mol safrole.  Since the EtONO seems to run out after about an hour, add 2.5 more mol equivalents after say, 45 minutes.  You should have to do this again before the reaction is complete, after another 45 minutes (these are just guesses).  DO NOT LET THE RXN RUN OUT OF NITRITE!!!  Once the CuCl precipitates, it is OVER!  Adding a kg of EtONO won't do a damn bit of good at that point.  You will know when the rxn is done when it stops giving off NO gas.

I'll try this procedure soon and post the results.  And another tip - PRE-STIR PdCl2 AND CuCl2 for AT LEAST SIX HOURS before adding anything else to the reaction.  The catalyst solution should be transparent dark amber.  If it is opaque or cloudy, or if you can see undissolved catalyst, stir it until it is completely dissolved.  Otherwise, you can expect shit for yields.

For more information, check out my other two posts on the wacker, and the method that I am trying to follow.

Post 293619

(lab_bitch: "VERY interesting literature on the wacker", Chemistry Discourse)
     Literature on the wacker

Post 296918

(lab_bitch: "Wacker Mechanism", Chemistry Discourse)
     Wacker Mechanism

Post 270818

(potisgood: "Fuck Benzo!  Ethyl Nitrite is the Shit", Methods Discourse)
     Procedure - This method doesn't work

foxy2

  • Guest
Pressurize
« Reply #1 on: April 16, 2002, 11:19:00 AM »
Pressurize with O2 ala the O2 wacker, that might bee a good answer.  Or how about just trying a reaction with 4x ethylnitrite OR isopropylnitrite as I think Ritter was saying.  This will prove/disprove your hypothesis. 

If you are correct then you could try pressurized O2 to save on the amount of reactants used or just to provide "another" way.

Those who give up essential liberties for temporary safety deserve neither liberty nor safety

Sunlight

  • Guest
My experience
« Reply #2 on: April 16, 2002, 12:49:00 PM »
I thought something similar when working with methyl nitrite, and I made several small runs in closed vessels, and yield was not significatively different that when making it in an opened vessel. In other hand the procedure proposed by Uncle Fester, that I tried as well, solves this possible problem because the methyl nitrite is bubbled continously in the rxn, but it doesn't work at all.
Anyway I don't want to discourage you, if you get a 50 % of conversion using a 0.7 % of PdCl2 and CuCl2, it is by far better than the results I got with methyl nitrite, with a yield around a 10 % or less with that amount of catalyst and cocatalyst.
I'm glad to find a serious experiencer like you, try your procedure and tell us, please. I think it could be even better to make a small test (3 gr or so) in a bottle that can be hermetically closed and supports the required pressure like I made, then make a TLC and/or recover the ketone via bisulphite. You won't waste your precursors and expensive catalyst.

lab_bitch

  • Guest
Did you prestir?
« Reply #3 on: April 16, 2002, 02:53:00 PM »
I read in your procedure that you started the reaction before all of the catalyst was added.  Did you completely dissolve the palladium?  This is absolutely necessary, as without doing so, it might as well not even be in there.

Sunlight

  • Guest
Procedure
« Reply #4 on: April 16, 2002, 05:36:00 PM »
You've read the final procedures. And not, I think it's impossible to dissolve that amount pf PdCl2 in the solvent used (with my proprtions), what happened is that when it was added to the solution of methyl nitrite and safrol in methanol or the methyl nitrite bubbled in the stirred but undissolved PdCl2 with safrol and methanol, PdCl2 went into the solution reacting vigorously, heating the solution. When the PdCl2 was added at once to methyl nitrite, safrol and methanol the rxn went up to 45 C or so.

lab_bitch

  • Guest
60% conversion
« Reply #5 on: April 17, 2002, 11:44:00 AM »
I just distilled the rxn I was talking about earlier and got 60% conversion (not yield).  Yield was around 35% (molar), but was probably due in the most part to mechanical losses and during distillation.  Here's a breakdown of the post rxn contents:

8 g safrole/isosafrole
8 g ketone

Starting amount of safrole was 20 g

This was the reaction where the CuCl precipitated (i.e. the ethyl nitrite was depleted b/f the rxn was finished).  Next time, I'll add twice the EtONO and see if I get 100% conversion.

Sunlight

  • Guest
degradation
« Reply #6 on: April 17, 2002, 04:18:00 PM »
I hope you have stored the rxn in the freezer during this time, the mixture makes degrade ketone relatively fast. Do you have TLC plates ? I made a micro rxn with ethyl nitrite and isopropyl nitrite a few months ago, and TLC showed not too much conversion of safrol and 4 or 5 spots more, one of them in the place of ketone... Sad but true.

Osmium

  • Guest
I've posted this before several times, but I ...
« Reply #7 on: April 22, 2002, 04:14:00 AM »
I've posted this before several times, but I think this might be useful to this discussion.
Os. tried this reaction several years ago, and when he tried to distill the 200g or so of tar produced (with a heating mantle - bad idea!) a violent explosion took place which filled the room with smoke and completely shattered the distilling apparatus. No material was left over for analysis, all that os was able to find was black brittle burned charcoal-like residues sticking to the broken glass and some tarmac like material sticking to the ceiling.
I still don't know for sure what happened, I can only guess that somehow the NOx produced in this reaction reacted with the alkene producing something resembling pseudonitrosites.

I'm not fat just horizontally disproportionate.

lab_bitch

  • Guest
Never had an explosion
« Reply #8 on: April 22, 2002, 07:44:00 AM »
I'm not doubting you, but I've distilled this rxn over 10 times and never had anything violent happen.

lab_bitch

  • Guest
Forget everything I said
« Reply #9 on: April 29, 2002, 10:20:00 AM »
OK, I guess I'm full of shit.  I just ran a wacker with 1.4 L EtOH, 230 g safrole, 360 g EtONO, 1.5 g PdCl2, and 5.8 g CuCl2.  I only got back 44 g ketone.  There was definitely excess EtONO in there b/c while hydrolyzing, 100 g of gas came off.  I also ran an identical reaction with 50 g and got back 38 g ketone.  Both times, the rxn stopped when the unknown solid precipitated.  This solid is red in acid and green in base.  I think its some sort of copper salt.  It's not really soluble in anything.  I had to use conc. H2SO4 to remove it from my paperless filter funnel.

Now I'm thinking that this precipitate forms when the ketone reaches a certain concentration, since the same amount of ketone was formed in the two rxns.  I'm about to run the following rxns to get more information

All solvent volumes are 1.12 L EtOH

1) 100 g safrole
   1.6 g PdCl2
   6.2 g CuCl2
   200 g EtONO

2) 100 g safrole
   0.8 g PdCl2
   3.1 g CuCl2
   200 g EtONO

3) 50 g safrole
   1.6 g PdCl2
   6.2 g CuCl2
   200 g EtONO

4) 100 g safrole
   1.6 g PdCl2
   6.2 g CuCl2
   100 g EtONO

5) 100 g safrole
   1.6 g PdCl2
   3.1 g CuCl2
   200 g EtONO

Hopefully, these will give me some more insight into this reaction.  I'll post when I'm done.

I need help identifying the precipitate.  Does anyone know any procedures for identifying substances like this?

Osmium

  • Guest
That's a lot of chems you are wasting there.
« Reply #10 on: April 29, 2002, 10:54:00 AM »
That's a lot of chems you are wasting there. If i were you I'd do the reactions half the size, or even smaller. Depending on what workup you use (e.g. column chromatography) you could perform those tests with 5 or 10g safrole or even less.

To identify the precipitate you have to first isolate and purify it. Yes that sounds obvious, but I say it anyway. Then perform some standard cation and anion analytics with it.

I'm not fat just horizontally disproportionate.

Sunlight

  • Guest
I agree
« Reply #11 on: April 29, 2002, 09:28:00 PM »
I agree absolutely with Osmium. I lost a few hundred grams of safrol when I tried that way, then I started making 16 gr test, and recovering the ketone with bisulphite. That precipitated you say seems to be common with the nitrites, I saw it any time I made the rxn. It has a similar color to PdCl2 right ?

lab_bitch

  • Guest
It only wastes Pd
« Reply #12 on: April 30, 2002, 02:35:00 PM »
Most of the unreacted safrole is recovered, so it only wastes Pd, which ain't a problem for me.

Osmium

  • Guest
Smaller tests can be done much faster too.
« Reply #13 on: April 30, 2002, 04:23:00 PM »
Smaller tests can be done much faster too. Workup will be a breeze compared to big sized reactions.

I'm not fat just horizontally disproportionate.

Sunlight

  • Guest
True, true
« Reply #14 on: May 01, 2002, 01:30:00 PM »
I doscovered the TLC one year ago or so, and Madmax is absolutely right, that's the way to go.

ClearLight

  • Guest
tlc...
« Reply #15 on: May 02, 2002, 12:54:00 PM »
Whats your solute, visualization reagent and rF on your plates??



Infinite Radiant Light - THKRA

Sunlight

  • Guest
Standard work
« Reply #16 on: May 02, 2002, 03:26:00 PM »
You need to have a bit of your pure target product, or know the rf. You make a TLC of the system before rxn, and after rxn together with your target product, and then you can see what have happened. When you have done it the first time, you don't need to make the previous TLC or add your target product because you can interprete the results. You need a good solvent system where the formed products give different spots. I believe this is the standard procedure. It is not very reliable in the quantitative aspect, but it gives you a good reference, for example you can see if the initial product has been consumed or not.

ClearLight

  • Guest
oh..
« Reply #17 on: May 02, 2002, 07:00:00 PM »

  I thought you had an analytical procedure.... I'm starting a new thread...

Infinite Radiant Light - THKRA

lab_bitch

  • Guest
New Results
« Reply #18 on: May 14, 2002, 08:55:00 PM »
Ok, I just finished a new round of experiments and came up with some interesting obersvations.  Mainly, I found that the yield of ketone is probably based on solvent volume.  For the following experiments, I ended up with product in the range of 26g - 35g.  In all experiments, 1.12 L of denatured alcohol was used as the solvent, and excess EtONO was left at the end of the rxn.

1) 100 g safrole
   1.6 g PdCl2
   6.2 g CuCl2
   35 g ketone

2) 100 g safrole
   0.8 g PdCl2
   3.1 g CuCl2
   26 g ketone

3) 100 g safrole
   0.8 g PdCl2
   3.1 g CuCl2
   32 g ketone

4) 50 g safrole
   0.8 g PdCl2
   3.1 g CuCl2
   26 g ketone

I think I didn't strip off enough EtOH for (2), so I didn't extract all of the oil.  That's why (3) has a much better yield.  I came to this conclusion b/c the ketone + safrole + crap extracted from (3) was 100g, while it was only 95g for (2).  Since ketone is the most polar compound in there, it probably got left in the rxn mixture.  On (1), I also didn't evaporate all of the ethanol (only about half), so the yields for it could actually be higher, but probably by less than 5g.  I'm about to run one with more solvent and see what happens.

In all runs, the rxn stopped after a green precipitate fell out of solution.  After hydrolysis, this precipitate turns yellow-brown.  After evaporating half the ethanol, it turns red-brown.  After evaporating all of the ethanol,  much less precipitate remains, and it is dark green.  It seems that more ppt. is collected when more catalyst is used, so I am pretty sure that it contains either Cu or Pd or both.

Also, on (4), I only got back 45g of organics, when I should have gotten 53g.  On (3), I got back 100g when I should have gotten 104g.  Both times, I evaporated all of the ethanol.  Could some of the product be carried off with the ethanol during stripping?  It seems pretty unlikely.  The only explaination would be that one of the oils is part of the precipitate.

Osmium, do you tell me where to find info on the ppt. analysis?

Well, I'll post more after I run a few more batches.

Osmium

  • Guest
Remove the precipitate by filtration.
« Reply #19 on: May 15, 2002, 02:19:00 AM »
Remove the precipitate by filtration. Wash it with all kinds of solvents. See if it dissolves in some of them. Check if it dissolves in acids. Report back how much precipitate you got (after drying!). Is it slimy or a solid? Hard inorganic or more organic-like, waxy?
Pyrolyse some of the precipitate in a test tube. Are any volatile organics given off? Do you have access to a good scale? Lots of chemicals without having to buy each of them from a lab supplier?

Testing for chloride: mix a few milligrams of the precipitate with several times its weight of Na2CO3 in a small test tube. Add some water, boil it for a few minutes. Neutralise carefully with a chloride free acid, filter. Add AgNO3-solution dropwise. White precipitate is AgCl, which will redissolve when conc. NH3 is added. Repeat whole procedure without adding the precipitate in the beginning and compare them both to check for chloride contamination in your reagents.

Don't remember any methods to separate Cu and Pd right now, but I guess you will have to somehow dissolve the precipitate, e.g. with aqua regia, and then try to selectively reduce first the Pd and then the Cu.



I'm not fat just horizontally disproportionate.