Wednesday, July 24, 2013

Process Wednesday: It's not just the organic chemists who have issues scaling up

Over at It's The Rheo Thing, a while back, John Spevacek talked about the issues with scaling up polymer manufacture:
Working with polymers at the lab bench is usually quite different than working with them at a larger scale (pilot plant, production, etc.). Benchtop versions of blown film lines, injection molders, extrusion lines...don't exist. Small extruders exist, but they don't have mass-loss feeders, custom screw elements and a whole host of other features that their big-boy versions have. 
And so at the benchtop we cheat. We use solvents to dissolve the polymers, which then allows us to easily stir in additives or other materials. We eventually have to get the solvent out, but that is usually pretty easy. Depending on the volatility of the solvent, a hood can work just fine. In other cases, an oven or a hot plate will do nicely. In short, the solvents make our lives much easier. 
In contrast, solvents at a large scale are generally avoided. Not only is flammability a concern (which then leads to higher capital equipment costs for the electrical equipment to make it non-sparking) but also the solvents need to be either recovered (more capital equipment) or oxidized (more capital equipment) which leads to CO2 emissions... [snip] 
...This difference in processing options with the size of the operation can lead to real nightmares in scaling up products. You can dissolve a newly-developed polyolefin in hot decalin and add in all the antioxidants you want, but you can't find out if it will make a nice blown film until you go to the pilot line. Even a cast film of the polyolefin will not provide meaningful samples as they would be completely lacking in the strength-inducing orientation that the blown-film line provides.
That's an interesting problem with the kind of chemistry that John participates in -- that the equipment itself provides the desired physical properties to be tested. (Any errors here are mine, not John's.)

For the most part, this doesn't seem to be the issue with classic old-skool chemical manufacturing; the physical properties of a compound are the same at bench scale or at plant scale. What we tend to worry about (and just like John, have difficulty simulating) is large quantities of your compound/reaction will interact with your equipment -- whether or not your agitator will stir, or your filtration will be effective, etc. Even then, the best-resourced among us have access to miniaturized versions of plant-scale equipment. Lucky us, I suppose. 

6 comments:

  1. I like this article. I've been a strong advocate of moving "structure-property" relationships to "structure-processing-property" relationships with polymers. The example I typically use is polystyrene disposable cutlery vs styrofoam cups. Both are the same materials, processed in different ways providing different properties.

    This doesn't get away from the scaling issue, but there are a lot of cool polymers that are synthetically generated and characterized, but never put in any sort of (semi-)realistic processing scenario that can enhance (or degrade) the interesting properties.

    ReplyDelete
  2. If you have done much crystallization and isolation's or transfer of viscous reagents at large scale not sure could say readily agree "classic old-skool chemical manufacturing; the physical properties of a compound are the same at bench scale or at plant scale" (unless that is included in your qualifier "for the most part").

    ReplyDelete
    Replies
    1. Perhaps it was written unclearly. Yes, there's a qualifier in there.

      (But even on plant scale, the physical properties are the same (i.e. this reagent is really thick) but the transfer problems are much worse, yes?)

      Delete
    2. Well certainly the physical properties are "technically" always inherent and largely invariant (and would make same statement for polymers) but as suggested from "practical" perspective it matters greatly and therefore from that angle equipment and handling always is an issue in chemical scale-up and not exclusive to polymers (which could by their nature include less simple or obvious resolution options).

      Delete
    3. I have to disagree. Polymers are large enough that they are inherently nonisotropic at any scale where small molecules are isotropic. Therefore the mechanical properties do depend on the processing, and the processing depends on the scale of the equipment.

      Delete
  3. We used to have a large benchtop twin screw extruder in my grad lab, although the group's focused had changed so almost no one used it.


    A scale up danger in the synthesis of polymers is the autoacceleration effect possible in free radical polymerizations.

    http://en.wikipedia.org/wiki/Autoacceleration

    I've had this happen on the scale of a few grams of acrylate monomer during my time working on RAFT polymerizations. My supervisor used to tell us that when you get up to larger scales you can possibly blow up a fumehood.

    ReplyDelete