Automated Syntheses

[Herro]

Didn't see anything in search about this; thought it was interesting.

I work in a engineering lab where the researchers autonomously monitor and control bioreactors and I thought, "Why can't you automate some HE syntheses?"

Electronics are not my forte, but I can assume that with a home computer and some sort of interface to connect to (serial port?), you could automate simply synthesis. For starters, I was thinking AP.

You'd need a mass (or weight) sensor tared with a beaker to determine the mass of liquids to be added, along with a top-stirring motor to be controlled through the computer. A program could be written to determine the flow rate of the pumps for the h2o2, acetone and acid, the density of the liquids being added; calculating the total volume. Another line could monitor the temperature with a simple probe and direct a pump to circulate colder salt/ice water through the system.

A webcam could even be used if you want to document and/or monitor the process. The drying/storage process would be another issue to automate and probably much more difficult.

What do you guys think? Feasible? What other procedures could be done a-human, for safety or convenience?

[plutobound]

1) ROR - Rate of return or $/qty. If you spent $500 in electronics to make 3 kgs, it becomes very expensive product (and you could have made a lot more product for your money). If you spend $50,000 to make 100,000 metric tons, then the cost per kg is noise.

2) Automation is good for doing the same thing over and over again. It works well for industrial processes. Experimentation/Research by it's very nature means constant changes. I work in an R&D lab and the only thing automated is the hand dryer in the bathroom.

[Shalashaska]

This could probably work for someone that just won the lottery or someone that doesn't pay bills, but I doubt this is feasible for anyone unwilling to spend more than $2-3,000 for equipment. Also, with nitrations, it gets much more complicated with lots of variables like optimum yield and such things. And then again it all comes back to, why? Why would you want such large amounts of explosives without a terroristic motive? If not, you'd have to either sell it or stockpile it, and I can't quite decide which is more dangerous.

Or is this question merely hypothetical? (Of course it is since no one on this board would THINK of making explosives or doing anything against the fine laws of our government.)

[Herro]

Of course this is purely theoretical.

The cost for such a design would be somewhat steep, but I've sunk much more money into useless classes.

I suppose the goal would be more having it run remotely for safety or just while you're doing something else. I'm not saying build a backyard demolitions factory, just a small setup; a proof of concept on the hobby scale.

We invented the remote control, we're a lazy species. Who wouldn't want to push a button and a day later have 30g of some HE in solution?

[TheBlob]

I like the Idea but I don't think I would like an automated synthesis of high explosives, especially when dealing with unstable explosives like acetone peroxyde.

I wouldn't trust myself enough to set-up an automated synthesis and then go away while it happily mixes acetone, peroxyde and concentrated HCl. As Murphy said, if something can go wrong, it will, and when it will go wrong, you won't be there to fix it up before it's too late.

There's also the topic of safety, of course you can't get hurt if you're not there, but I would prefer being there myself and doing it to prevent my lab from being destroyed by a robo-chemist.

[ProdigyChild]

I like the idea, too. If your reaction involves a step using hazardous chemicals (carcinogenic?) and results in more or less harmless final products then an automated process is really desirable.

An different approach can be the question: where do electronics outperform the human experimentor?

Temperature control comes to my mind immediately. Electronic temperature sensors cheap and precise and heat can easily be released by a resistor.

Very slow reactions are also preferably controlled by electronics because a human is too impatient to watch a reaction for 3 days adding a droplet of a chemical every now and then :(

[megalomania]

There are several companies that use automated explosive synthesis technology for safety reasons. I think all lead azide is made remotely by machine. One potential use for the independent experimenter is to automatically synthesize small quantities of chemicals, for safety reasons, but to do so continuously. Steady and slow is the way to go, and over time you can build up large amounts of material.

I can already see how certain reactions could use automation with existing technology. Metering pumps can be set to deliver precise quantities of chemicals over set periods of time. A series of timer controlled metering pumps could mix chemicals, temperature controllers deliver heat or cold as desired, another pump removes the reaction product for distillation/purification...

I designed a simple setup to automatically make iced tea for me everyday. I hate making tea, so I make 2.5 gallons at a time to give me 3-4 days. I don't much care for 4 day old tea, it is best when fresh brewed. My system dispenses tap water through a water filter into a reservoir based on a timer, the timer also activates a tea maker (like a coffee pot) that directs the brewed tea into a chilled container. All I have to do is dump in the sugar using a scooper modified to contain exactly the amount of sugar I like, and stick in the tea bags.

This is the same principle as a chemical reaction. I add the ingredients, and the final product makes itself. Anything a human can do that does not require you to be there every second can be duplicated by a machine. Many chemical reactions are of the "mix A with B, heat for x minutes, and distill" variety, and these could be automated without much cost.

[TheBlob]

Yeah I agree with the practical use of such a system, but what I mean is: If something goes wrong, for example the raction starts bubbling or the temperature raises dramatically, you wouldn't be there to react accordingly. I don't think you could think of everything that could happen and program your system to react to each of these situations. I wouldn't trust a system I made myself enough to handle flammable liquids or, even worse, explosives.

There's a difference between an industrial system, which was designed by a whole team of engineers, and an homemade system. By the way, I am also pretty sure that even industrial automated systems always have someone assigned to watch them, should something go wrong.

[megalomania]

There are many examples of reactions that run "overnight" or for several hours at a stretch. Sometimes it can be very bad to be around if a reaction does go dreadfully wrong. If certain reactions did start bubbling and the temp spikes the only thing you will be able to do is stop some of the flying glass with your face... The point of automated reactions for some explosives is there is no one around to get killed if something does go wrong. Many times there is nothing you can do for a reaction that is out of control except clear the area.

Not every reaction is suitable to automation, as I already said. It's a matter of routine. If you do the reaction over and over with little variation in batches, nothing ever happens that needs user interaction, and it is pretty much a routine, then automation can come into play.

Automation need not mean you are nowhere to be found either. Some of the more prolific chemists conduct different areas of research while they are waiting for some reaction to complete. They multitask in the lab to get more done in less time. This is what I do when I cook food, I hate to sit in the kitchen and waste my time preparing food when I could be doing something more productive. I stick something in the oven or crockpot and I walk away. Yes, sometimes I end up with burned food, but that's only because I forget I am cooking something. This is why I have a tea maker now instead of using the kettle, 3 burned dry pots was enough. I can be disturbingly absentminded sometimes.

[Nitronium Soup]

I personally think the best approach to making the best of your time in the laboratory is not with complex automatons and electronics, but rather just 'MacGuyver-ing' simple, time saving methods.

Like setting up a magnetic stirrer and using a separatory funnel to slowly add reagent over the course of a half hour, you'd be surprised how many dolts in my organic class will sit there with a dropper in their hand and swirling the flask with the other. Meanwhile, I'm grabbing a soda :D.

Burned food pales in comparison with a runaway nitration, which is why the one system I would want is a temperature sensor wired to a trapdoor and ice bath quench. However, in my opinion it's still far better to react to the rise in temp yourself and press the big red ABORT button, than rely on electronics.

If I were REALLY paranoid about a given reaction, I'd build myself a sturdy steel containment hood with 1/2" Lexan windows, a webcam for remote viewing, and controls for fire suppressant, quench, and whatever else.

[asoix]

It's actually not that expensive to make simple electronics/computer interfaces. I hacked together a really simple serial port stepper motor controller in about an hour using a few Darlington Arrays to avoid frying my motherboard, and it works fine.

I think it would probably be very easy to control a distillation electronically. A digital thermometer, stepper-motor or digital potentiometer could control the hotplate's temp. I have a rotating microwave oven platter that could easily be set up to rotate in a fresh beaker to catch 3 or 4 different fractions at preset temperatures.

Some longer processes, like electrolytic production of oxidizers could be automated pretty simply and even the addition and pumping off of different phases of the reaction could be automated without too much hassle... As long as you have a setup that involves switches and sensors in the first place

I think lab automation is a really cool thing.