CAPE-OPEN

Dear all,

I would like to model Boudouard reaction, i.e. a disproportionation of carbon monoxide into carbon dioxide and elemental carbon (C + CO2 = 2CO) in COCO using Gibbs reactor. The model should not be too sophisticated, using ideal gas law is sufficient, applying Peng-Robinson EOS would be a nice-to-have.

Obviously, this is a very basic problem known from elementary chemical engineering and chemistry textbooks, but as long as I do not know how to solve it properly in this software environment, I cannot move to more complex problems (e.g. chlorination of metals and metal chloride mixtures of iron, manganese, vanadium, etc.. A simple example on the horizon is the reversible chlorination of iron(II)chloride to iron(III) chloride: FeCl2 + 1/2 Cl2 = FeCl3 with subsequent (flash) separation. Both inorganic chlorides have very special physical properties like sublimation in the temperature range of interest.

The complete sets of physical properties for CO and CO2 are available in the ChemSep's PCD manager and I must admit that I did not thoroughly check those data for consistency. I have observed some deviations against reference values in heat capacities of ideal gas but I value them as insignificant.

Implementation of elemental carbon as "solid only" was discussed in this forum already http://www.cape-open-forum.org/viewtopi ... tal+carbon. Using a conversion reactor seems not to be option to me, because knowing the conversion at certain temperature is actually what I want to get as a result - especially for multicomponent mixtures. Providing explicit equation for temperature dependent equilibrium constant would be an option, if I had to deal only with one reaction. Hence, I have compiled physical properties data for elemental carbon [Kohlenstoff] as graphite (including - very, very low! - vapor pressure and critical data) myself using various sources and estimation methods.

In theory, following Le Chatelier's principle, Boudouard reaction is pressure dependent. Because the reaction only takes place in the vapor phase, the activity of the elemental carbon aC in the equilibrium constant defining equation
[ Keq=(fCO)^2/(aC*(fCO2)) ]
is 1 (unity). To my surprise, I was able to generate reasonable results (at the pressure of 10^5 Pa) close to those from literature and commercial software in the temperature range from ambient to 900K. Unfortunately however, the results in the temperature range from 900K to 1500K were not close enough for "a workable engineering solution". Another surprise observed was that the reaction equilibrium is also dependent of the concentration of the carbon! This is understandable for the cases where carbon is provided under its necessary stoichiometric amount. However, as long the carbon is present as a (solid or liquid) phase, the equilibrium in the gas phase should not be influenced.

In the attached file, the reaction is defined by its compounds but I have also provided a reaction package, where I have used fugacity as equilibrium basis. Assigning reaction package to the reactor delivers the same results. All calculations are valid for the pressure of 10^5 Pa. I am aware that the Gibbs reactor in COCO performs equilibrium reactions for a single phase only. However, I can not treat carbon as gas, because then I would compromise the Le Chatelier's principle.

I need help. I appreciate any suggestion or hint to come out with a workable solution.

Thank you very much.

Best regards,
Mitja Medved


P.S.: Just in case - here the temperature dependent equilibrium constants:
* log10(Keq)=((9141/T)+0,000224*T-9,595) from https://en.wikipedia.org/wiki/Boudouard_reaction
* log10(Keq)=-((8817/T)-9,071) from René Kelling's PhD Thesis (2016)

For this particular case, the equations are reasonably straight forward if you formulate it as an equilibrium reaction (best to do so in ln space). The equilibrium reactor can perform reactions in a multi-phase system. but the reaction itself is still constrained to a single phase. The reason for this is the version 1.0 reaction standard in CAPE-OPEN, which does not make it easy to define multi-phase reactions. A new standard specification for this is in the making, and then the reactors can be updated to support that.

But, as you have the equations, this seems to me to be a fairly small task to set up as a custom reactor, e.g. in the Scilab, Matlab or Excel CAPE-OPEN unit operation. You can use the equation solvers in these applications to solve for the extent of reaction that closes the equilibrium constraint.

Addressed to All:

Thanks to the Posting of mmedved for bringing the Boudouard Reaction to mind again. I read about it years ago regarding NASA planning for a manned mission to Mars.

I am seeking advice on my plan to simulate the Reverse Boudouard Reaction because of its metallurgical and environmental significance. The reaction can be expressed as shown.
I think that the simultaneous solution of these two equation will provide the unknowns, f, and kf. Any advice or comments are appreciated. Thanks.

I would think also in this case a custom solution (Matlab, Scilab, Excel) would work fine.

Addressed to All:

Referencing the Posting of 01/01/18, the last equation, Equation 3, an expression for a specific rate of reaction, can be combined with the performance equation for a mixed flow reactor to provide an expression which represents the mass balance to be solved.
Advice or comments are welcomed. Thanks.

Addressed to All:

This presents an update to the last Posting. More work remains. Advice or comments are welcomed.