OpenFOAM precipitation/crystallisation reaction solver

Question

This question is probably related to chemical engineers that are around this forum.

I am looking into writing a solver for a precipitation reaction (struvite to be precise): A+B+C+D <=> E+F A, B, C, D, F - liquids; E - solid (for now I ignore the 'solid' part and focus only on mass transfer)

So far I have tried to avoid using combustion-thermo-chemistry libs/classes (which as far as I understand are highly interlinked and dependent of each other) and wrote a simple multispecies transfer solver that performs a forwards reaction.

For mass transport source term currently my option is on using 'Laminar finite rate model' as in Fluent, though - I was not able to find out in the literature whether this is appropriate for precipitation reactions (If anyone could answer that then I'd be grateful).

Unfortunately, this is where I end up at and in order to move the development forwards (get backwards transfer working) I either need to rewrite some classes from chemistry/thermo/combustion libs, or go with modifying reacting foam, which again is using mentioned libs/classes and I'm not fully familiar with the models used there.

I'd be glad if anyone could give me some pointers on:

a) How to use existing combustion/thermo/chemistry libs/classes for my reactions, that as far as my understanding of chemistry goes are not combustive?

b) Any literature pointers on what models are used on such reactions?

c) If anyone worked on similar stuff maybe you could share your ideas on what would be the best way to approach these type of problems?

Cheers,

Ben

Answer 1

I'd be glad if anyone could give me some pointers on:

a) How to use existing combustion/thermo/chemistry libs/classes for my reactions, that as far as my understanding of chemistry goes are not combustive?

Your best bet is to use Cantera, which include models for liquid-phase thermodynamics. You will probably need to add some functionality yourself, but Cantera is open-source, and Ray Speth, who was at one point the lead developer of Cantera, is a great guy, and you can contact him about what you want to do. (Disclaimer: Ray is an old colleague.)

The main issue you're going to run into with other libraries is that is that combustion thermo & kinetics libraries mostly focus on gas-phase chemistry, and for the temperatures and pressures under consideration (excluding engines, most of the time, less than 10 bar, and aside from possibly an initial transient, 700-4000 K) the ideal gas law is a good approximation. Consequently, there just isn't much in the way of non-ideality or fugacity coefficient effects, and you're really going to have to hack these solvers to include that functionality, plus activity coefficient models, if you need them. A colleague of mine several years ago "faked" liquid-phase chemistry for hydrocarbons by treating them as very high-pressure gases, but I don't think that's a good solution for your problem, or for the general case. This largely rules out closed-source libraries (e.g., CHEMKIN-PRO and its competitors). It also rules out many open-source libraries (e.g., TChem, PoKiTT) that focus on gas-phase chemistry exclusively.

b) Any literature pointers on what models are used on such reactions?

It mostly depends on the specific chemical system, and what you're trying to actually model. I think the main difficulty with precipitation can be if you want to understand particle sizes in the precipitate, and where the solid is actually going to go (is it entrained in the liquid and swept away by the main flow? does it just fall to the bottom of your control volume?).