Sciencemadness Discussion Board - Nitroethane from ethanol

Nitroethane from ethanol

Grizli7 - 12-7-2023 at 08:46

Hi all. I was bored and I accidentally found out about neural networks and decided to ask a question, how to get nitroethane from ethanol and now what was the answer.

Quote:

Text from AI
The synthesis of nitroethane from ethanol involves a two-step reaction process: oxidation of ethanol to acetaldehyde, followed by the nitration of acetaldehyde to nitroethane.

Step 1: Oxidation of Ethanol to Acetaldehyde

1.In a round-bottom flask, add 100 mL of ethanol and a few boiling stones.
2.Attach a reflux condenser to the flask and heat the mixture gently using a heating mantle or oil bath.
3.Slowly add a 10% sodium dichromate solution dropwise to the flask while stirring continuously.
4.Continue heating the mixture under reflux for about 4-5 hours.
5.After the reaction is complete, cool the flask and transfer the reaction mixture to a separating funnel.

Step 2: Nitration of Acetaldehyde to Nitroethane

1.In a separate round-bottom flask, add 100 mL of acetaldehyde obtained from the previous step.
2.Place the flask in an ice bath to maintain a low temperature during the reaction.
3.Slowly add a mixture of 40 mL concentrated nitric acid and 20 mL concentrated sulfuric acid dropwise to the flask while stirring continuously.
4.Continue stirring the mixture for about 2-3 hours, ensuring that the temperature remains below 10°C.
5.After the reaction is complete, pour the reaction mixture onto crushed ice in a beaker.
6.Allow the mixture to cool and collect the resulting precipitate of nitroethane.
7.Wash the nitroethane with cold water to remove any impurities.
8.Dry the nitroethane using anhydrous sodium sulfate or a desiccant to remove any remaining water.

Note: The synthesis of nitroethane involves the use of hazardous chemicals and should only be attempted by those with proper training, adequate safety precautions, and appropriate laboratory facilities.

I began to look for more information and if at the first step you can really get acetaldehyde, perhaps not quite as it offers a neural network, then in the second step I have questions:
1. Does this work?
2. The concentrated nitric acid is indicated, but the concentration is not indicated? Is it possible to limit yourself to 70% although as far as I remember, if a mixture of sulfur and nitric acids is used to introduce nitro -group as a nitrous agent, then there is a concentration of nitric acid above 90%
3. Maybe it is better to use the equimolar draining of acetaldehyde and the nitrating mixture? How will the nitrous mixture act on the already formed nitroethane? I mean which process will be more likely -oxidation or further nitrating with the formation of dinitroethane or even trinitroethane?
4. May be 10 celcius is to high? let 5 be better?

[Edit by mod]
Edited to showcase possible convention for referencing AI text.

[Edited on 13-7-2023 by j_sum1]

Schleimsäure - 12-7-2023 at 09:34

My impression is, that many of the given procedures by Chat GPT don't work in practice.

Ususally nitration of acetaldehyde should give Glyoxal.

Texium - 12-7-2023 at 09:56

This entire procedure is obviously a load of crap. Don’t waste your time with Chat GPT or other AI. They don’t “know” anything about chemistry. They just string together sentences that sound good on the surface.

I think we need a new rule on this forum of NO posting AI generated procedure, unless it’s for the sole purpose of picking it apart and making fun of how stupid it is.

unionised - 12-7-2023 at 10:52

Philosophically speaking does
"Note: The synthesis of nitroethane involves the use of hazardous chemicals and should only be attempted by those with proper training, adequate safety precautions, and appropriate laboratory facilities."

rule out anyone who had to use AI to find a route from ethanol to nitroethane?

That proposed scheme is just wrong.
Laughably so.

One important thing to mention
You ask- quite sensibly "May be 10 celcius is to high? let 5 be better?".

It's possible to get problems by doing a reaction too cold.

You cool everything down, then add a bit of A to B but, because it's cold, not much happens.
So you add more A and so on.
Then you let the mixture warm up and it starts to react.
But that reaction generates heat- so the reaction gets faster.

And you get what's called "thermal runaway".

If you have to follow a recipe, use one from something reputable like Org Syn and actually follow it.

Come to think of it, read some random bits of Org Syn; it's very informative.
http://www.orgsyn.org/

Grizli7 - 12-7-2023 at 13:02

Thank for all answers. I thought something like that -that the GPT chat can hardly offer the right path, so thanks to everyone who confirmed my fears. As for the applicability of neural networks to syntheses, I know how scientists have developed a highly specialized neural network for chemistry. There, Mendeleev’s numbers were used, I vaguely imagine what it is, but that is not the meaning of that. The point is that the developed neural network with the participation of professional chemical scientists issued a working technique at best 1 out of 5. So it is possible that a working chemical artificial intelligence will appear in the not-so-distant future.
So the question remains open- is there a way to obtain nitroethane from ethanol?

clearly_not_atara - 12-7-2023 at 13:46

Quote:

So it is possible that a working chemical artificial intelligence will appear in the not-so-distant future.

I doubt it.

Quote:

So the question remains open- is there a way to obtain nitroethane from ethanol?

No spoonfeeding!

Lionel Spanner - 12-7-2023 at 14:03

Quote: Originally posted by Grizli7

So the question remains open- is there a way to obtain nitroethane from ethanol?

Not directly, but J. Chem. Soc. (1944), p.24 (DOI 10.1039/JR9440000024) has a method for making it from ethyl sulphate.

arkoma - 12-7-2023 at 16:41

Quote: Originally posted by Grizli7

So the question remains open- is there a way to obtain nitroethane from ethanol?

Go forth and dig here:

https://www.erowid.org/archive/rhodium/

fusso - 12-7-2023 at 20:51

Quote: Originally posted by Grizli7

As for the applicability of neural networks to syntheses, I know how scientists have developed a highly specialized neural network for chemistry. There, Mendeleev’s numbers were used, I vaguely imagine what it is, but that is not the meaning of that. The point is that the developed neural network with the participation of professional chemical scientists issued a working technique at best 1 out of 5. So it is possible that a working chemical artificial intelligence will appear in the not-so-distant future.

Tools like reaxys already exists.

[Edited on 230713 by fusso]

Schleimsäure - 12-7-2023 at 21:00

Quote: Originally posted by Grizli7

Thank for all answers. I thought something like that -that the GPT chat can hardly offer the right path, so thanks to everyone who confirmed my fears. As for the applicability of neural networks to syntheses, I know how scientists have developed a highly specialized neural network for chemistry. There, Mendeleev’s numbers were used, I vaguely imagine what it is, but that is not the meaning of that. The point is that the developed neural network with the participation of professional chemical scientists issued a working technique at best 1 out of 5. So it is possible that a working chemical artificial intelligence will appear in the not-so-distant future.
So the question remains open- is there a way to obtain nitroethane from ethanol?

I doubt it too. As a fellow member mentioned there is a route from sodium ethyl sulfate from 1944.
Other than that. You really think not one of 1000s of skilled German, US etc. Chemists never tried to get the relatively expensive Nitroethan out of Ethanol/Adectaldehyde with just "mixed acid"? And a a poor trained "AI" is more capable?

Try it. I have acetaldehyde and fuming HNO3, even oleum 30%. I know, the reaction will be highly exothermic, even with proper cooling.

Lot's of crap will evolve possibly with fireworks, but no good yield of Nitroethane.

[Edited on 13-7-2023 by Schleimsäure]

Grizli7 - 15-7-2023 at 03:14

I am familiar with the work with Erowid and also with the ethyl sulfate method. I am interested in, it is possible new methods of information about which are not widespread. For example from beta alanine

Grizli7 - 15-7-2023 at 03:54

Quote: Originally posted by Schleimsäure

Quote: Originally posted by Grizli7

I can not share your skepticism regarding new ways of cognition. Behind AI the future is that a lot of money is invested in them, and if now the attitude of errors to a positive result is permissible, then in the future, a very near future, it will inevitably change and not in favor of a people.
As for your statement about thousands of scientists, I consider it wrong and that's why. At one time, I was very interested in the topic of obtaining nitroalcans, but I did not find serious work about the selectivity of final products, maybe I am mistaken, but it seems to that the existing industrial technology simply arranged all participants from manufacturers to consumers.
In addition, you do not take into account -that the cost of gas does not matter which one: ethane, propane, butane and nitric acid in some regions can be much lower than the cost of ethanol, acetaldehyde and acids too.
You also did not take into account that science is developing and that previous generations of scientists could not be able to do the future. Also think about what, if you already have a production plant, it doesn’t matter what, and the new technology has approximately the same profitability, then you will not throw out the equipment and rush to a new one.
Well, finally, I’ll say about this method of cognition as an contingency, remember how Penicilin or nylon was discovered, I’m sure you can remember a lot of examples, if you reason from this point of view, it means that it doesn’t matter how many scientists work on the problem, this means that solution can be found absolutely contingency in any area with a non-targeted research and by the one person.

Texium - 15-7-2023 at 06:04

Well let’s put it this way. If Chat GPT were to take a college organic chemistry test (one with open-ended questions not pulled directly from a textbook, that is) it would bomb it miserably. Asking it for advice is like asking the most clueless student in a class of undergraduates who doesn’t know what they’re talking about, but thinks that they can somehow pull one over on the professor/TA by answering everything very confidently.

Call me a troglodyte if you wish, but I fear a world in which AI becomes smarter and more reliable. Not because I think the machines are going to take over and physically kill us or anything like that, but because it will dumb people down. Just as we now have an obesity epidemic thanks in part to technology that allows people to avoid using their muscles (among other factors), I foresee a stupidity epidemic if people are able to stop using their brains.

Schleimsäure - 15-7-2023 at 11:34

Quote: Originally posted by Texium

Well let’s put it this way. If Chat GPT were to take a college organic chemistry test (one with open-ended questions not pulled directly from a textbook, that is) it would bomb it miserably. Asking it for advice is like asking the most clueless student in a class of undergraduates who doesn’t know what they’re talking about, but thinks that they can somehow pull one over on the professor/TA by answering everything very confidently.

Call me a troglodyte if you wish, but I fear a world in which AI becomes smarter and more reliable. Not because I think the machines are going to take over and physically kill us or anything like that, but because it will dumb people down. Just as we now have an obesity epidemic thanks in part to technology that allows people to avoid using their muscles (among other factors), I foresee a stupidity epidemic if people are able to stop using their brains.

Totally agree. Nothing to add.

@Grizli: Chat GPT "knows" a lot of basics. That's it. In many fields. It can not "replace" a skilled chemist or engineer though.
Absoluetely not at his point, and probably not in the forseeable future.
Even if it is "trained" to the max, it just derives everything from human brains in the end.
And one thing it clearly misses is intution for example.

It has just one big advantage like all computing, it is fast.

And as Texium noted, it will dumb people even more down.