Earlier I wrote about my first impressions as a chemist in chemical development blessed with a late stage project involving some azide chemistry. This project continued when I received a message that it was actually going into production. In the meantime all the use tests1 I had carried out bore fruit and we were able to define three suppliers for the benzyl alcohol and registered it as the starting material for the drug substance manufacture. So now I only had the synthesis of the benzyl chloride and its conversion to the azide to worry about.
Due to lack of capacity in chemical production it was decided to outsource these two steps to an external manufacturer and after a series of meetings with a particular company we went to their facilities to work with them in carrying out a pilot scale series, so that they could get to grips with the chemistry. So off we trundled (the production chemist and I) to spend ten days or so to familiarise them with the procedures.
We spent the first morning just finding the place, the taxi driver got lost! Ending up back at the hotel I decided to rent a car and eventually found the site. After discussing the chemistry in depth they decided to run the first 1 kMol reaction. Off they went and we disappeared back to the hotel to get some decent food and wait for a phone call. Five hours reaction time came and went, no call, so off we went back to the plant. The place looked suspiciously quiet, I thought “Oh dear” (to put it politely) something has gone wrong. We eventually found the responsible chemist wandering around in front of the building, shaking his head and mumbling to himself, and tears were streaming down his face. Obviously he had been exposed to the physiochemical properties of the benzyl halide. Indeed he had!
We followed him into the pilot plant, wearing gas masks and there it was, the reason for his mumbling. They had used an “old” reactor. The enamel was cracked (history does repeat itself) but this time the crack was larger and the 37% hydrochloric acid had merrily eaten its way through the reactor and into the cooling coils. Of course they shut the whole thing down, upon which the whole mess solidified. Scratch one reactor. Well we went home leaving them to it. I do not know how they disposed of the mess, and I don’t really want to.
Some weeks later we were back. They were ready to have a second go at it. This time it worked. During the five-hour reaction time I took a wander outside and started crying again. The room beneath the reactor was open to the world. I went in and saw an open bottle of ammonia directly under the valve in the bottom of the reactor and lots of white fumes. Looking up I saw a steady drip of the benzyl halide emerging from the valve. Pointing this out to the operator he was not bothered at all and said that ammonia was good for the lungs and anyway not a lot was being lost. But the reaction worked as advertised, as did the azide step. So we returned home and gave them the outsourcing work.
Some time later we were back at their state of the art facility. A rather large building containing safety boxes, each with two floors the upper one with two 10 m3 reactors and beneath centrifuges and pressure filters etc. There were four such boxes on each side of the building that was only about three years old. The control room was at the entrance end and everything was computer controlled, apart from adding the reagents manually which needed a human being! So this was quite a scale-up, about 15 times but far as I remember I think we did 12 kMol batches.
Before we started we had a look at the safety box we were going to use. For some reason this had all the glass tubes for the exhaust gases zig-zagging up the walls, presumably to save space. One of them was full of a liquid with a very light greenish tinge. After several enquiries I was told it was hydrochloric acid remnants from the cleaning process. Well as we were going to use hundreds of kilos of dusty sodium azide I insisted the liquid was removed. This put the schedule back and we actually began the next day with the chemistry.
Off we went with the synthesis of the benzyl chloride. All went well until the operator tried to transfer the product to the other reactor. The technician could not put more than 0.5 bar nitrogen pressure on the reactor to get the transfer going. He kept trying to increase the nitrogen pressure, no luck, so I suggested he should go and look. No response. So I ambled along and peered in the porthole. Wow, it looked like Old Faithful. The product was being pushed out of an inlet port of the reactor with the nitrogen pressure. The stream was about 2 meters high. So I sauntered back to the control room and told them this. Off went the nitrogen pressure on went the gas masks and safety suits and we went along to the box. It turned out that the operator who removed the pH electrode had not replaced the cap properly causing the geyser effect. I went back to the hotel so they could clean up the entire mess. The next day we started again. This time all went well and we were able to produce both the halide and the azide and obtained the required quantities and yield(s). That was it I thought, finally finished with this project.
Some months later It was all quiet on the western front when the phone rang again, yes the second time it rang. There was a nice voice on the other end and we were having a pleasant conversation until the word INSPECTION was uttered. A mock FDA inspection was planned for Monday next week and did I have time to attend? This was Wednesday: “Did I have a choice? No.” came back down the line. This was a three-day event run by our internal QA, who at some time and no doubt considerable expense had poached a FDA inspector. So I examined the SOP telling me what to do in the event and spent the rest of that week getting the house in order. The great day arrived at last and I packed all my folders and computer into the car and took off for the meeting site.
An enormous empty room greeted me. I was always early for meetings so I could drink all the coffee and eat the small sandwiches provided or snaffle all the ones I liked. People commenced arriving and the room started filling up with conversation, computers and grey A4 folders. The production chemist arrived accompanied by two mammoth shopping trolleys full of these folders, I don’t know where the trolleys came from but they were full to overflowing. Finally in rolled the tame ex FDA inspector followed by his minions and got the meeting started.
As it turned out they were interested in everything except the chemistry. The production validation protocols were examined in minute detail, as were the analytical methods and their results. Cleaning practices, labelling procedures, sop’s, calibration, you name it, it was examined, but not the chemistry. They probably even looked to see if the toilet paper conformed to cGMP, but still no chemistry. Finally on the last day chemistry reared its ugly head! “Did you examine your starting material for the existence of positional isomers and other elements (those related to the ones we had in the drug substance)?” Some kindly neighbour kicked the chair from under me where I had been happily dreaming about seaside holidays. “Yes and yes”, I said after picking myself up off the floor. “Good, good” said the fake inspector and turned back to giving the analytics guy round 15 of his version of the Spanish Inquisition. I went back to sleep. During a waking period I did observe that some of the questions they asked were silly to say the least, but I suppose he was only earning his big bucks. Still it was an interesting few days and I learnt what not to do.
Some months later I learnt that we (the company) had sold the compound lock stock and barrel to someone else. They of course visited us and carried out an inspection, more or less as above. I did not attend the meetings but was available should chemistry expertise have been required, but it wasn’t needed. That was that for me. I went back to the lab and cleared out all the samples and archived all the folders. Where they are now I don’t know, buried under a foot thick pile of dust I suppose. A happy end to my first large project
Notes:
1 A use test is employed in order to investigate material from potential starting material suppliers: The compound must meet the given specifications and perform in the series of reactions to deliver the desired product according to its specifications.
p.s.
After all this I have come to the conclusion that cGMP does not stand for “current Good Manufacturing Practice but Grosse Mengen Papier (translated from the German it means large quantities of Paper). This because I estimated we produced more paper than compound.
4,192 total views, no views today