In the Oct. 16, 2014 post, I discussed process options for the trimethyl pentene hydrogenation process. In that post, I discussed an option of recycling the CSTR effluent. I said that this would involve a difficult separation between the unreacted pentenes and the isooctane product. I have discovered that the separation will be even more difficult than I imagined.
An incorrect assumption
Previously, I assumed that isooctane was less volatile than the pentenes. That assumption was based on the slightly higher molecular weight of isooctane compared to the pentenes.
With my assumed order of volatility, a single column would have been needed. The overhead would include hydrogen, pentane solvent, and the unreacted pentenes. All of these overhead components can be recycled to the reactor. The bottom product would be isooctane. The distillation would require a large number of trays and possibly a high reflux ratio.
The actual volatility
When I started simulation of the distillation column for the chosen process configuration (with a CSTR- fixed bed reactor combination), I discovered that isooctane is slightly more volatile than the pentenes. The figure below shows the K values for the components. The K value for isooctane is between the values for TMP1 and pentane. As a check on the K values, I also obtained the normal boiling temperatures of the components. They gave the same order of volatility.
The effect on the separation scheme
For the process option with recycle of the CSTR output, the separation scheme would require two columns, not just one as originally assumed. In one configuration, the first column would split between pentane solvent and isooctane. The bottom product would then be split between isooctane and the pentenes. Regardless of the separation configuration, two columns would be needed and one column would involve the difficult separation between isooctane and pentenes.
the process configuration involving the fixed bed reactor now looks like a clear winner because it requires only a single, small column.
On my part, I should not have assumed the volatility ranking of the components when I was considering process schemes.
This example also shows that reactor selection and design can have a major impact on the process configuration. Reactor conversion is often cited as affecting the recycle rate and associated processing costs. During the conceptual design phase, we also should look at how reactors may be used to affect the process configuration.
Reactors also can create byproducts which affect process configuration as well as yield loss. Thus, reactor selection and design should include consideration of byproduct production. In my ideal system being developed as an illustration, I have not included byproduct reactions or impurities that may exist in the pentene feed stream.