We ran our first experiments in the reaction calorimeter today. They were very elementary, involving only the metering one reactant into another at constant reaction temperature. The results suggested strongly that the reactants were consumed promptly and that good control of temperature is obtained by adjustment of the feed rate. Halt the feed and Qr falls off promptly. This is a very desirable attribute in semi-batch processing.
From the data workup we determined the adiabatic ΔT, or ΔTad, and were able to follow the heat evolution measured in several ways, but most interestingly in watts per liter. It is desirable to know how many watts of heat evolution your reactor is capable of removing. Engineers think in terms of heat evolution as watts per unit volume. Calibration of a reaction vessel can tell you how many watts of heat the reactor can remove at a defined level of fill and agitator speed. RC1 data can tell engineering how many watts per liter the reaction mass is capable of.
Next we ran the reactor in adiabatic mode where the jacket temperature is programmed to follow the reaction mass temperature. The idea is to exert dynamic temperature control via the jacket to make the vessel behave as a Dewar. We predicted the maximum reaction temperature by simply adding the ΔTad to the initial temperature. We allowed the reaction enthalpy to ramp the temperature.
The actual endpoint temperature was within two degrees of the predicted temperature and below the bp of THF. This is a measure of the potential for runaway. If the Maximum Temperature of the Synthetic Reaction (MTSR) is below the solvent bp, then you are in a lesser hazard zone. This temperature would be achieved in an adiabatic system.
A few observations- the heat capacity, Cp, is not constant over the course of a reaction. A little reflection should suggest this. But it does not automatically follow that the Cp increases in magnitude over the reaction progress, which would offer some thermal buffering capacity.
Lamentations on Chemistry 
Chemical Engineering! Reactor Design! I love it when you talk dirty.
Seriously, this does look like a nice little bit of fun that I’ve only been exposed to in the class room.It seems like you’re not having any significant viscosity changes in the reaction, as that can make heat transfer to the kettle wall much more difficult.
The RC1 will monitor a metric related to viscosity. The power needed to maintain an rpm setpoint is recorded. Viscosity buildup effects can be seen in the Qr (reaction enthalpy) curve. Poor mixing may result in irregular heat flow to the thermal sensors, giving Qr flow that is not smooth. Reactant addition to a viscous mass leads to short term accumulation of reactant at the surface. This is revealed as irregular Qr as parcels of poorly mixed reaction mass pass by the sensors.
Hey Gaussling,
Just curious to know:
Can the maximum temp be calculated with any accuracy from first principles (reaction enthalpies, addition rates, reaction rates, heat capacity and the like) ?
Or is it generally easier (and safer) to do the hands on calorimetry?
You can calculate reaction enthalpy by hand or with CHETAH and come up with a number. This assumes that you know the reaction mechanism and have accounted for all the pathways. The formation of intermediates will complicate this. Solvent effects will add lots of uncertainty as well. Not only do solvent effects affect the rxn kinetics, the solvation enthalpy of reactants and products must figure in the calculation. Solvent affects may not be constant over the course of the reaction.
As you proceed with a reaction, the composition of the reaction mixture changes constantly. This affects the polarity of the reaction mass and the heat capacity. The magnitude of these changes may or may not cause a significant change in the total enthalpy.
Once you have the delta H, you divide it by the Cp to get delta T. How you would calculate the heat capacity is unknown to me. I’m a synthetic organikker who is new to RC, so there is a lot about this that I do not understand yet.
Given that sensitive instrumentation exists, it is easier for most of us to measure thermal properties than calculate them.