MSMPR Crystallizer

//One universal basic required here to get things going once loaded
window.onload = function () {
    //restoreDefaultValues(); //Un-comment this if you want to start with defaults
    Main();
};
//Any global variables go here


//Main is hard wired as THE place to start calculating when input changes
//It does no calculations itself, it merely sets them up, sends off variables, gets results and, if necessary, plots them.
function Main() {
    saveSettings();

    //Send all the inputs as a structured object
    //If you need to convert to, say, SI units, do it here!
    const inputs = {
        g: sliders.Slideg.value,
        b: sliders.Slideb.value,
        j: sliders.Slidej.value,
        dS: sliders.SlidedS.value,
        dM: sliders.SlidedM.value,
        dQ: sliders.SlidedQ.value,
    }

    //Send inputs off to CalcIt where the names are instantly available
    //Get all the resonses as an object, result
    const result = CalcIt(inputs)

    //Set all the text box outputs
    document.getElementById('Comments').value = result.Comments
    //Do all relevant plots by calling plotIt - if there's no plot, nothing happens
    //plotIt is part of the app infrastructure in app.new.js
    if (result.plots) {
        for (let i = 0; i 

The Mixed-suspension, Mixed-Product-Removal (MSMPR) crystallizer is an amazing device. Saturated solution enters a chamber already full of crystals and an equal flow exits, containing whatever crystals exist at that end of the flow chamber. The image is of a DTB, Draft Tube Baffle configuration as one example. There is perfect mixing, zero breakage of crystals, no build-ups on walls or stirrers. Yes, it's an ideal, but especially for inorganic crystallizers, plenty of MSMPR systems work astonishingly well.

Our interest here is that unlike every other crystallizer, the equations are simple and the outcomes clear. Our reference system, with subscript 0, produces a population (per unit volume) n₀=1E14 of seeds which grow at a rate G₀ of 1e-7 m/s. This means that the nucleation rate B₀=n₀.G₀ = 1e7/m³s. This is all happening with a saturation of S₀, a mass of crystals M₀ and a flow rate Q₀.

The basic MSMPR equation tells us that after a time t (here assumed to be 1000s), for a crystal length L, the number of crystals with that size, n_L is given by:

`n_L=n_0exp(-L/(Gt))`

So a plot of ln(n_L) versus L is a nice straight line with slope -1/Gt and intercept n₀. Of course in the real world the line isn't so perfect, but it is surprisingly common. The other core equations say that measures of "average" size are simple functions of Gt. In this app we get L_median which is just 3.67Gt, in this case 367μm.

Because this is a log plot we see that the system is dominated by small crystals. But as is often the case with particle sizes, this is something of an illusion. From the cumulative number distribution 70% of the crystals are (move your mouse over the line to see this) less than 110μm. But we are usually more interested in the crystal as a whole, so it's the volume or mass distribution that interests us. When that's plotted we see that there are an equal number below ~360μm as above, i.e. our calculated L_median value.

Changes to the parameters

The specific reference values will be of no interest. Making them a user option would just complicate the interface. The point of the app is to show the general MSMPR equation then to look at the effects of process variables - saturation, mass of crystals ("magma") in the reactor and the flow rate. You set up a second set of conditions, with subscript 1, by changing ΔS, ΔM and ΔQ, relative changes (by a factor of 5 more or less) in those parameters. You then see the second straight MSMPR line along with the number and volume cumulative distributions. The change with ΔQ is straightforward, it's just changing t, the time spent in the reactor.

The other reason for the app is that different crystallizations show different power law dependencies for ΔS and ΔM on the the growth rate G and the nucleation rate B.

`G_1=G_0(ΔS)^g`

`B_1=B_0(ΔS)^b(ΔM)^j`

In the book chapter example, g=1.58, b=2.09 and j=1. Any process with a power law dependence much above 1 is likely to be challenging. If you set these three parameters to 1, then changes to saturation or mass lead to orderly changes in the crystals. But once one or more is above 2 then small changes get amplified - and that's a key problem with real-world crystallization. Here we are merely modelling an ideal single stage. In practice the crystallizers will be multi-stage, and their mixing (and therefore degree of saturation) isn't perfect, removal of crystals isn't size-independent, and models don't (can't) cope with issues such as size-dependent or crystal-dependent growth rates.

Real world data

If you were doing this for real, where would you get the reference data and how would you work out the various power laws? The answer, as with everything in crystallization is that it isn't easy. A lab-scale MSMPR reactor is surprisingly hard to construct so you have to try to get representitive values from simplified experiments. Then you have to do your best to fit your process data to get as good an idea as possible about what's happening. It's not perfect, but it's better than twiddling buttons with the hope that things will get better.