In the first post I presented an overview of the proposed method of obtaining a kinetic model for certain systems. In today's post, the experimental setup is described in more detail, starting with the drawing below.
The hydrogen heater (tube furnace) is used to provide a constant heat input. This is used to heat the liquid solvent and reactants in the bubble column to a temperature above the desired reaction temperature. At that point in time, a PID temperature controller (not shown) is activated and the boiling temperature is adjusted by controlling the hydrogen flow rate.
When the reactor temperature is under control, the catalyst solution or slurry is charged to the reactor by opening the ball valve under the catalyst vessel. The heat released by reaction(s) generates additional solvent boilup which requires more hydrogen flow to maintain the partial pressure and the temperature of the boiling solution.
Boilup and excess hydrogen flow to a single tube-in-tube metal condenser. The condensate is returned to the reactor. The total pressure of the system is maintained by a pressure controller on the offgas. This pressure is set based on the desired hydrogen concentration in the liquid. A pressure of 1000 psi might be a typical value. For systems that don't require a reactant gas, nitrogen may be used and the total pressure may be reduced.
Note that the reactor has no direct heating. This greatly reduces the problem of plate out of solids on the hot surfaces which would change the heat transfer rate. The hot hydrogen sparger or steel frit may still present a problem with flow restriction, but heat transfer will not be affected.
In the work thus far, the hydrogen flow rate is the only variable that requires digital recording. However, this places extra demands upon the accuracy of the system model. In the future, I will explore the benefit of recording more variables. I suspect this modification would allow the method to be applicable to more systems.
With the present method, the added sensors are needed to determine parameters such as the heat transfer coefficients in the condenser and the heat input by the heater. Observations of these variables during the temperature controlled period prior to the addition of the catalyst should be sufficient.
Another modification to the setup would be to eliminate the hydrogen flow sensor and instead to record the PID controller output signal. Again, this places added demands on the model because it may require more knowledge of the actuator and valve characteristics.
Determining kinetics for solid catalyzed gas/liquid reactions is a difficult task. The task is made even more difficult by the series nature of the example reaction system. Studying individual reactions would require varying the concentration of the intermediates. In some cases, those intermediates may not be available. Even if they are available, the number of experiments can get quite large. In today's fast paced development environment, the proposed method appears very attractive.