Most of my industrial life has been spent in what can be called a semi-batch processing world where products tend to be high value, low volume. Fine chemical products sold at the scale of 1 ton/yr or less can be produced in a campaign of less than a dozen runs in 200 gal to 1000 gallon reactors in a batch or semi-batch mode. Depending on the space yield, of course.
In my polylactic acid (PLA) days many years ago, we found ourselves necessarily in the monomer business. If you hope to introduce a new polymer to the market- a very difficult proposition by the way- you must be firmly in control of monomer supply and costs. Especially if the new polymer uses new monomers. New to the market in bulk, that is.
Our task in the scale up of polylactic acid was to come up with a dirt cheap supply of lactide, the cyclodimer of lactic acid. The monomer world is one of high volume, low unit cost. By the time I had my tour of duty with PLA, a short tour in fact, much of the lactide art was tied up in patents. Luckily, my company had purchased a technology package that allowed us to practice.
Our method for producing lactide was a continuous process called continuous reactive distillation. Basically, a stream of lactic acid and strong acid catalyst was injected onto a middle plate of a 25 plate distillation column which stood outdoors. The column was atop a small bottoms reservoir containing heated xylenes.
The solvent xylene was heated in a reboiler which was located 20 ft away from the column assembly. Hot solvent was circulated in a loop between the bottoms reservoir and the reboiler. After startup the solvent built up an equilibrium concentration of lactide and a dogs lunch of oligomers.
At the injection point in the column, 85 % lactic acid and catalyst entered the middle of a multiple plate column that was charged with refluxing xylene vapor and condensate. While in the column the lactic acid esterified first as L2, the open chain dimer, then some fraction of it cyclodimerized to lactide.
The water that was extruded by the esterification process was vaporized by the hot xylene and equilibrated up the column to the overhead stream and out of the column to a condenser. When condensed it phase separated in a receiver called a “boot” that had a cylindrical bottom protuberance that collected the water. The upper xylene phase was returned to the process.
Meanwhile, the xylene loop accumulated lactide and oligomers. The loop had a draw-off point where some predetermined percentage of the bottoms loop was tapped for continuous lactide isolation. This is where the fun began.
When cooled even just a little, the xylene phase emulsified. Badly. So, the trick was to induce a phase separation by forcing the emulsion through a ceramic filter. Here, the water phase and most of the oligomeric species were partitioned into a separate mass flow while the xylene phase was sent to a sieve bed for drying.
After passing through the sieve bed, the xylene phase was sent to the continuous crystallizer where it was chilled a bit to precipitate the lactide. A slurry of lactide solids called magma was then sent to a continuous centrifuge where the solids were isolated and the supernatant was returned to the bottoms loop.
The Achilles heel of the process was residual acid. Since the monomer and the oligomers are all acidic species, and the catalyst is nearly as strong as sulfuric acid, pulling the neutral lactide cleanly and cheaply from this acidic hell broth was a problem. So big in fact, that it eventually was the straw that broke the camels back. This shut our fledgling company down.
Residual acid in the monomer has a disastrous effect on the quality of PLA. It gives low MW product that is amber in color. The winning technology was the back-biting process for lactide production. It was applied by our competitors who won the battle and they (Dow-Cargill) went to market.
Lamentations on Chemistry 