1 reply to this topic

[Momo_212]

Hello everyone,

I am currently building a conceptual Aspen Plus model for a MOF synthesis process and I would appreciate technical feedback on the modeling strategy.

Process objective

The process I am trying to represent is based on:

1. acid leaching of stainless-steel waste to recover dissolved metal species,

2. oxidation/speciation adjustment of iron in solution,

3. future integration of a PET hydrolysis section to recover a terephthalate-based ligand,

4. simplified MOF formation from the metal-rich solution and the ligand feed.

At the moment, my flowsheet includes:

• stainless steel + HCl mixing,

• heating and leaching,

• Fe oxidation/speciation adjustment,

• ligand entry point,

• simplified MOF formation block,

• product outlet section.

Current modeling choices

• I am using ELECNRTL for the leaching / ionic aqueous chemistry section.

• The leachate contains ionic species such as Fe3+, Ni2+, Cr3+, Cl-, etc.

• I initially tried to represent MOF formation using RSTOIC, but I had mass-balance and custom-component issues.

• I am currently using RYIELD as a temporary approximation for MOF formation, mainly to keep the process framework running.

Main challenges encountered

1. Thermodynamic conflict
   I am not sure what the best approach is for linking an ionic aqueous stream to a MOF precipitation/self-assembly step.

2. pH sensitivity
   pH seems critical for the synthesis, but I am still trying to determine the best Aspen strategy for handling it in a meaningful way.

3. Custom solid MOF
   Representing the MOF as a user-defined solid led to several property issues (missing Cp, solid volume, enthalpy-related parameters, etc.).

4. Ligand section
   I plan to integrate PET hydrolysis later to generate the ligand upstream, but this part is not fully built yet.

What I would really like feedback on

I would appreciate advice on the following points:

• Is it reasonable to keep ELECNRTL up to the ligand/mixing stage and still use a simplified MOF formation step downstream?

• For an early conceptual model, is RYIELD a reasonable placeholder for MOF formation, or would you strongly recommend another Aspen block strategy?

• What is the best practice for handling a user-defined solid MOF in Aspen Plus?

• If the feed to the synthesis section is still ionic, would you keep the same thermo package, or split the process into different sections?

• Any recommendations for representing PET hydrolysis to disodium terephthalate / ligand recovery in Aspen Plus would also be very helpful.

I am not trying to claim this is a validated predictive MOF model yet. My goal is to build a physically organized and progressively improvable process framework.

I have attached / can share an annotated flowsheet if useful.

Thanks in advance for any guidance.

[Pilesar]

You have a very ambitious goal. This is the type of model that you can spend months on, get 95% of the way, and still not have anything useful. The only way I expect you to have success is if you have many sets of real process data that you can use to tune the model. My guess is that custom models will be needed to represent process behavior. If this were my task, I would do the following:
1) Complete literature search including academic graduate theses. Look for similar process systems and review how they are modelled. If found, contact the thesis author directly for advice. Look for the individual process steps if you do not find the entire process train.
2) Contact AspenPlus tech support. There may be limitations on support for student licenses, but contact them anyway. Asking cannot hurt you and they may have knowledge to share.
3) If this is a commercial process, initiate discussion with engineers at the company doing this type of work. They may want to partner with you.
4) Consider some form of neural net model in areas where rigorous first-principles modeling is inadequate.