Chemical Forums
Chemistry Forums for Students => Inorganic Chemistry Forum => Topic started by: curiouscat on April 23, 2015, 01:40:02 AM
-
Is it practically feasible to process NH4Cl industrial scale to yield NH3 & HCl. Is this reaction quantitative? To get reasonably fast rates how high do I need to heat it.
Context: I have a waste stream containing NH4Cl dissolved in water. Trying to think how to reprocess it.
I could send it to a multi effect evaporator. Then get solid NH4Cl & heat it. Upstream process has use for both NH3 (aq.) & HCl (32%).
Or could I somehow just heat the aq. solution & get NH3 + HCl without bothering to ever go to the solid NH4Cl?
Once I get the mixed gases how do I separately dissolve them in water?
Comments?
-
I think this will not work, ammonia and hydrochloric gas immediately form ammonium chloride by re-sublimation. Experiment: open a bottle of each component and stay them together you will see white smoke development.
-
Instead of heating, could you displace NH3 from NH4Cl brine? With a strong nonvolatile base like CaO maybe? That would be half of the work.
Then, reprocess CaCl2 by molten electrolysis? I'm not quite sure...
Maybe HCl can be displaced from CaCl2 by H2SO4, and then you dump CaSO4, but this may not make economical sense.
What would the electrolysis of aqueous NH3Cl lead to?
-
Instead of heating, could you displace NH3 from NH4Cl brine? With a strong nonvolatile base like CaO maybe? That would be half of the work.
Then, reprocess CaCl2 by molten electrolysis? I'm not quite sure...
Maybe HCl can be displaced from CaCl2 by H2SO4, and then you dump CaSO4, but this may not make economical sense.
What would the electrolysis of aqueous NH3Cl lead to?
Thanks @Enthalpy. Those ideas would work in principle.
But at that point I might as well use NaOH / Ca(OH)2 in the upstream process directly.
Basically, I have an upstream reaction that can be carried out with either NaOH or NH3 or Ca(OH)2. More prefrably either of the first two.
Using NaOH is the default but that generates a large amount of NaCl (aq.) getting rid of which is a nuisance.
I was wondering if using NH3 might let us reprocess & recycle. The key challenge is how to do this:
NH4Cl :rarrow: NH3 + HCl
-
I think this will not work, ammonia and hydrochloric gas immediately form ammonium chloride by re-sublimation. Experiment: open a bottle of each component and stay them together you will see white smoke development.
Yes, but that's at room Temp. I'm suspecting (hoping?) that at elevated T the equilibrium will favor the gases.
Then it'd be a question of separating them while still hot. Not easy by any means, but is there a fundamental chemistry problem?
-
I have misued CPropepShell (meant to compute chemical equilibria in a rocket engine), mixing ionized helium (to bring neutral heat) with NH4Cl at 1atm, and:
- The lowest temperature to get an answer is 792K.
- Then, all Cl makes HCl, but N and H make N2 and H2, only 1/1000 make NH3.
- At 2375K, atomic Cl climbs to 0.01* of HCl, Cl2 is trace, the rest is essentially the same.
- At 2830K, atomic Cl climbs to 0.1* of HCl, the rest is the same.
- At 3400K, atomic Cl is 0.5* HCl, NH3 stays as traces
Tried again at 500bar: the smallest T for an answer is 912K, then
- NH3 moles are half as abundent as N2 moles.
These are strict equilibria, which make sense in a hot flame, not necessarily at lower temperatures where some reaction paths aren't allowed. He is in small mount here, no big equilibrium shift.
Do I remember that NH3 is synthesized at a lower temperature with a catalyst (and big pressure) to favour it over N2 and H2? At least CPropepShell predicts very little of it. Logically, NH3 is destroyed at a lower temperature than the salt.
I still don't grasp how you could separate HCl from NH3 even if this latter did form. It would require some form of chemical separation, wouldn't it?
But what looks credible is to recycle HCl and maybe H2, release N2. Just heat at 1atm does that. Upon cooling, all the Cl2 fraction will form HCl, N2 and the remaining H2 won't react. You plant would buy NH3 and destroy it cleanly, cycle HCl.
Or make ammonium chlorophosphate, sell it as a fertilizer?
-
Or make ammonium chlorophosphate, sell it as a fertilizer?
If it comes to that then might as well sell Ammonium Chloride itself. :)
-
I still don't grasp how you could separate HCl from NH3 even if this latter did form. It would require some form of chemical separation, wouldn't it?
Agreed. I don't know of a good way yet. I'm just brainstorming for now.
-
In the Solvay process, NH4Cl is treated with Ca(OH)2 to recover NH3. CaCl2 is a waste, it is either dumped into the sea or used in winter road clearence.
-
To my opinion, and unless someone has an astounding idea, the proper way is to decompose NH4Cl at enough heat and atmospheric pressure. Dissolution in water separates HCl, the N2 and H2 left over can be burnt together if no better use. Clean, cheap.
Use successive water-gas contacts so the gas meets clean water in the last step, leaving well washed N2 and H2.
-
To my opinion, and unless someone has an astounding idea, the proper way is to decompose NH4Cl at enough heat and atmospheric pressure. Dissolution in water separates HCl, the N2 and H2 left over can be burnt together if no better use. Clean, cheap.
It will be cheaper for me to just sell the NH4Cl I think. Unless I can recycle the NH3 the idea is an economic dud.
-
I have misued CPropepShell (meant to compute chemical equilibria in a rocket engine), mixing ionized helium (to bring neutral heat) with NH4Cl at 1atm, and:
- The lowest temperature to get an answer is 792K.
I don't understand what you wrote here. What is the equilibrium prediction of your code at say, 200 C or 300 C.
To me that seems the interesting region.
-
CPropepShell failed to compute an equilibrium below 500°C. Equilibria of solid+gas products are difficult to compute, and 300C isn't the normal use of this software.
-
More in detail: Propep's aim is to compute equilibria in rocket combustion chambers (and nozzles if supposing they shift during the expansion). Because of 3000-4000K, it neglects all chemical structure, all reaction details, and needs only to know the atomic composition and heat of formation of the propellants. Then it checks which of its known combustion products can apply (including OH, gaseous MgO, Cl*, even C2) and solves all equilibria.
As a sidenote, it was programmed in Fortran for I dont' know what computer, ported much later to Ms-dos and never serviced since. Several graphical user interfaces were added, none is perfect nor really easy, one package being CPropepShell. Rpa (Rocket Propellant Analysis) was written recently and is far better than Propep plus interface, except that it lacks flexibility for my horrible tinkering.
You get it: what you ask at 200 or 300°C is cute lace, what Propep does is sheet stamping. Propep answers the best thermochemical equilibrium even if no reaction path leads to it. Sometimes it also fails to converge.
With ammonium chloride at 1atm and +200°C I get no answer, at +300°C I get the molar fractions
HCL 3.3650e-001
H2 4.9050e-001
NH3 9.5031e-003
N2 1.6350e-001
and at +274°C, NH3 climbs to 1.54e-2,
but I suspect Propep ignores NH4Cl as a combustion products.
Maybe an other equilibrium software suits the needs of chemistry better.
-
Maybe an other equilibrium software suits the needs of chemistry better.
Indeed. I'd love recommendations.
I guess I could always do it the old school way. With NIST data & by hand.
-
I too would enjoy recommendations about an equilibrium software. No idea what already exists. Did I see one for aqueous electrolytes?
Problem is: at moderate temperature, >>billions of different compounds are possible, and the final one depends on the synthesis path, not just the reactants' elements, so a software can't decide autonomously what product shall appear. Many different approaches are possible:
- The user types every compound, inputs their thermodynamic properties. The software gives an equilibrium at a given T and P. About the complexity of a spreadsheet.
- Slightly better: the software equilibrates the reactions and converts the units.
- The soft knows hundred compounds: CO, CO2, H2O, NO, H3O+... including their change of state maybe. The user types the others. Still accessible to a university project.
- The soft knows all compounds in the Crc Handbook or a database, possibly a set of solvents.
-
Your standards are too high. ;D
Personally I'd be happy enough with a code that reliably computed equilibria for species & reactions I chose.
-
Problem is: at moderate temperature, >>billions of different compounds are possible, and the final one depends on the synthesis path, not just the reactants' elements, so a software can't decide autonomously what product shall appear.
Can you clarify what you mean by this? I always thought equilibrium was independent of the nature / number / difficulty of connecting synthetic paths.
-
Can you clarify what you mean by this? [Enthalpy's: "the final one depends on the synthesis path"] I always thought equilibrium was independent of the nature / number / difficulty of connecting synthetic paths.
Often, the user wants to compute an equilibrium among a limited set of possible species, which may not include the all-best equilibriium species because these are out of reach for the conditions. For instance, to evaluate the acidity of aqueous NaClO3, you suppose it doesn't decompose into the more favourable NaCl and O2. I could have put it clearer from the beginning.
For these useful purposes, the user has to specify what products must be equilibrated by the software. It is a difference with Propep, which knows a set (200? 500?) of potential products and choses amont them according to the propellants' elements and according to the combustion conditions, without user intervention.
Though, it doesn't look more complicated at all for the software, rather easier, changed my mind. Just different programming.
-
Your standards are too high. ;D
Personally I'd be happy enough with a code that reliably computed equilibria for species & reactions I chose.
For each possible species, you would have to type the formation enthalpy, entropy or Gibbs, heat capacity (possibly a polynom of the temperature). Then, such a software is but more than a spreadsheet. I bet it exists for Mathcad/Maple/Mathematica/Matlab/etc.
But...
Wouldn't you expect the software to equilibrate all the reactions? This would avoid mistakes and save time. Between just two reactants and three products, there are already many reactions and error sources.
You'd get tired very soon (or at least I would) to type thermo data again and again at every use, every equilibrium. The very minimum I expect from such a software is that is knows the thermo data for all usual small molecules like H20, CO, CO2... there must be 100 to 1000 of them.
As well, the software better knows the phase changes for common compounds, in order to avoid elementary mistakes.
And so on and so forth... My bet is that, as soon as someone has a little software that works in one particular case (say, only aqueous ions) he will immediately want to make it more general.
The best strategy looks like:
(1) Find a software that already exists, or
(2) Let a university develop it for free, then put your company's name on it and sell;
(3) Make the soft very general from the beginning, but start small. It looks like a many-years project.
-
I've just tried a short Google search, and no obvious equilibrium software popped up.
A dozen exist for flames. A few ones for aqueous ions.
Well, if no software exists outside these two narrow cases, it's time to make one. Take several years at a chemistry university, helped by many students.
Make the software very general and flexible right from the beginning, as it will evolve and expand a lot. Use abstract programming techniques, maybe a Lisp programme or a symbol-manipulation overlay on a procedural language; just object-oriented looks too restrictive and clumsy. The flexibility to add data and methods suggests data-driven programming, expert systems, A-lists and the like.
-
Well, if no software exists outside these two narrow cases, it's time to make one. Take several years at a chemistry university, helped by many students.
Who's gonna bell the cat? :)
-
Who's gonna bell the cat? :)
I won't. I thought you would. ;D
Look, I've already drawn a mockup of the software. I helps figure out the users' needs and the project's complexity. Not as complicated and abstract as I had feared, it can be programmed in standard style, but from that simple example:
- The algorithm knows phase changes and runs stably nevertheless, hum. Includes dissolution by reaction.
- The algorithm must live with incomplete data like solubilities.
- Automatic balance of reactions already seems necessary.
- I wouldn't bear more than one unknown compound per reaction, whose data is input manually. One-shot data can mean just H, S, Cp but reusable data includes Mp, Antoine, Van der Waals or better, Hm, Hv, plus the ionic and non-ionic species with dH and S in varied solvents. Ouch.
- When dissolution is a reaction, the compounds data base must link to the solvent and solute, seemingly.
- The compounds data base must have a simple format (txt, maybe xml) so users can edit it and share species, possibly centralized at the software's site.
- Some nasty details should be foreseen even if not treated immediately. Say, the change in P and mp when a solvent absorbs a compound.
- Drawing curves as a function of P, T, composition seems necessary.
My feeling is that this software can't start much simpler. If phase changes are excluded, it's limited to reactions strictly within a solvent or completely in gas phase, and then you miss the dissolution of produced NH3 in H2O or the precipitation of NaCl in acetone. Adding such features later would be more complicated than at least foreseeing them from the beginning.
The algorithm and the user's interface are more or less the maximum achievable by a single person (though I've seen bigger ones); it's slightly worse than Rpa for instance (Rocket Propellant Analysis). The data base is beyond one person's reach, but if users find the soft good, they will contribute.
-
That mockup looks very nice. What did you use to make it?
-
I won't. I thought you would. ;D
Waaaay to difficult for my skills. :)
-
Glad that you like the mockup... I used a stone-old bitmap drawing soft to adapt a screenshot, but a programming environment to design Gui software would have been better.
A picture has much convincing power, and in this case, it shows what the software should do and how to use it. Since putting it on Chemicalforums, I feel many people get interested, so maybe the software will exist some day.
Chemical equilibria are difficult to program properly, with an algorithm running smoothly despite the liquid <-> solid transitions. It takes a workable understanding of both thermodynamics and numerical analysis. If Alexander Ponomarenko programmed himself the solver of his Rocket Propulsion Analysis
http://www.propulsion-analysis.com/
then he'd be able to do this one as well.
A big difficulty is the size of the compounds base. 1000 compounds with all properties for gases, solids, liquids, solutions in water, heptane... are already too much for a single author, but the users would soon bang against the limits. This base will obviously grow very much over a long time. Two methods then:
- Make it a free cooperative project. Maybe one PhD thesis to write the solver and some graphical interface. Then, users will happily contribute the database.
- Make it commercial from the beginning - but then, pay enough people and data sources, initially and over time, to provide the compounds base to the customers. It does demand capital.
-
A picture has much convincing power, and in this case, it shows what the software should do and how to use it. Since putting it on Chemicalforums, I feel many people get interested, so maybe the software will exist some day.
Here's to hoping! It would be very nice to have it for sure.
Make it commercial from the beginning - but then, pay enough people and data sources, initially and over time, to provide the compounds base to the customers. It does demand capital.
Have you looked at professional Process Simulators? e.g. Aspen / Hysis / Pro2 etc.? My impression is that they can already do a fair bit of equilibrium analysis.