From a terribly interesting new text called Right First Time in Fine-Chemical Process Scale-up by the wonderfully named Lumbertus A. Hulshof, a comment on commercial scale chemistry and optimizing for speed or yield:
From an economical viewpoint this parameter [CJ's note: productivity (i.e. kilograms/hour-meter3] is usually predominant, but not much in the mind of the chemist who is more focused on a maximal chemical yield. The chemist may ignore the fact that for processes with a higher contribution of the variable costs (raw materials) compared to the fixed costs (labour, equipment, etc.) it is better to optimise on yield, whereas for processes with a lower contribution of the variable costs compared to the fixed costs to the total costs the plant residence time is a crucial factor to optimise and not the yield.
As an example, the condensation of glyoxylic acid and phenol in water gives p-hydroxymandelic acid in 75% yield after several hours, while a 55% chemical yield is obtained already after a period of 2 hours. Thus, the productivity of the process can be considerably higher, while the chemical yield has still not reached its highest value... the chemical yield is after a half-life usually 50%, after two half-lives 75%, after three half-lives 87.5% and after four half-lives 93.8%. So, the gain in yield after 4 half-lives is a poor 19% compared to 75% after two half-lives, whereas the residence time in the plant is twice as long!It seems to me that, for the manufacturing-oriented chemist, it's much easier to focus on understanding the economics of the variable costs (amount of solvent, equivalents of reagent, choice of raw materials) than it is to understand the fixed costs of a chemical plant (labor, utilities, etc.) I've also found that the "cost to run the plant per day" may not be a number that management is interested in divulging, and it may be inaccurately calculated or not really well-understood. Ultimately, though, you have to keep both numbers in mind.
You may be correct that appreciation of variable costs in more ingrained in chemists, partly because in grad school rarely is the people time a concern and overhead nebulous factor taken out of grants. However this is at the core of most ChemE training that hopefully can be used to educate process chemists. Although management may not want to divulge burden rates externally all plants I dealt with had defined cost per day and we always worked to reduce time because often had the biggest impact of production costs. Activity does take longer in the plant naturally however the more one studies and know the process can often find ways to shorten. No more customary overnight or 72 hours reaction times allowed without proof.
ReplyDeleteI love Process Wednesday. I always learn something interesting. :)
ReplyDeleteThis time vs yield analysis is also critical in early stage discovery.
ReplyDeleteI thought that part of the reason to optimize yield is to minimize purification - for example, in recrystallization, impurities prevent some of the desired product from crystallizing and so for every bit of impurity or starting material left, you lose some of your product. You also have to go through more manipulations presumably to get the desired material out. Is this at the stage when the workup and purification steps are consistent enough or known enough to make good correlations between yield in situ and isolated product, or is this a legitimate issue?
ReplyDeleteI was also wondering about that issue. Lower conversion means more impurities being present, means a more troublesome workup/purification procedure.
ReplyDeleteProcess Chemistry is not always that simple as in most cases it is likely best to maximize at reaction stage to minimize purification requirements there are times that secondary impurities that are difficult to remove if run too long so may halt before achieving full conversion. Comes back to understanding the mechanisms and by-products involved, which sometimes early on is more guess work or semi-empirical that complete knowledge. If have a inexpensive reagent that is easy to remove, say by extraction or wash, then may employ excess to push to the right side. When one knows what crystallization or other methods can handle or not handle certain impurities will influence details of operations. Unlike medchem in process chem the reaction and purification are approached more as a package to balance out because typically do not have realistic chromo options (and purity profile demands less critical)
ReplyDeleteI think I'd like to get this book.
ReplyDelete