5 Steps to Improve Your Efficiency in the Development of Chemical Processes

Govern the process. That is the real purpose of development of a chemical process for API or drug substance manufacturing. But what does govern the process mean? It means to handle the process parameters in order to get the desired results in a robust manner; in other words, to lead the process where you want, like a helmsman sails his watercraft.

For that purpose, it is mandatory to gain enough understanding of the chemical transformations involved as well as the behavior of a specific molecule and to get the know-how which allows, not only to set up reaction conditions, but to predict what happens outside of the defined working ranges.

Process development is a permanent exercise of learning from two fundamental sources: previous knowledge and experimentation.

To learn from the previously established knowledge in the state of the technique and in the foundations of science (chemistry and physics, in this case) is critical to improve success possibilities exponentially. The lack of experience or poor basic knowledge increase the risk of underestimating this stage provoking a permanent (insane) need of reinventing the wheel and discover already discovered facts. Just the contrary of efficiency definition. But that is not the matter at hand.

The second source of knowledge is to learn from research and experimentation, which is the core activity of any chemical process development. It must be (or should be) planned to consider the previous intellectual work and its execution will be the most time and resources consuming task. An efficient management of experimentation leads to a shorter development times and costs.

Next, you can find the 5 steps that could be followed to increase the efficiency of your experimentation.

Step 1: Define a purpose of the trial. Ask something to your trial. Ask a question.

Any project of any type should have defined its scope and goals from the very beginning. When project objectives are deployed in a work breakdown structure, it is critical to define and set out as clearly as possible the purpose of all the activities planned. The achievement of every single of those purposes will bring us closer to the project success. In this regard, trials intended to gain process understanding are not an exception. They are essential activities in a chemical process development; do not run any trial without a clearly defined goal to aim. A very common mistake is to run trials with an ill-defined purposes due to invest less time that necessary planning them. In most of the cases, you will not save time.

But, how to define this objective? My recommendation is to formulate a question to be answered by the data obtained from the trial (or set of trials). Make the simplest question which give you the answer, but not simpler. Yes, you’re right. It is an Einstein’s quote.

How impurity profile changes with an excess of reagent X? Does temperature promote degradation ? Which is the mass balance of this work-up? Is reaction rate pH depending?

Could you make several questions to the same trial (or set of trials)? In other words, could a trial have several objectives? Sure, it could. It will depend on the trial, on the objectives, on the design, on your experience to manage it …. On many things. But, please, do not bite more than you can chew. Keep a main goal in mind and avoid interfering with a secondary objective. We enjoy ambition. One answer per trial can seem no so much, however I can ensure you that one is usually more than the average when a process development is finished. Any case, it is up to you.

Step 2: Design your trial. Use your brain before your hands.

Once the objective is defined and before going to the bench, the trials must be carefully designed. It would be a kind of visualization exercise of how things are going to be done. For sure, it is necessary for reaction parameters (temperature, stirring, addition rates…) but also for sampling procedures, samples treatment, analytical methods, timing for analysis, equipment availability and large number of commonly overlooked details.

In my experience, the main reason for repeating trials is a poor design of them. Do not regret to spend time in trials design but keep this in mind: the simplest the best. Split trials if necessary. Complex trials are more likely to fail by execution reasons and what is worse, to give fake conclusion because some errors are hidden.

Lack of design in your trials probably would not be responsible for project failure, on the contrary a good design of your experimentation will result on an outstanding increase of efficiency: same results in shorter times. Let me know if I am wrong: Is not time the biggest concern in your process development? You are always late, even the first day. Aren’t you? Me too.

Step 3: Execution. Doing things well.

With an appropriate training, only few months are necessary to be able to run trials in an autonomous way. Some trials are more complex to carry out than others. Experience will always help to improve trials execution. At the same time, some people will be more skilled than others. In summary, there are several factors impacting the performance of trials execution. However, some trials throughout the project will have deficiencies, independently of experience or realization difficulty. It is something to assume and to deal with.

In my experience, there are three possible origin these deficiencies.:

• Errors. Human mistakes, equipment failure, calculation error, wrong communication between people involved, … An error is always possible, and it could be the most likely explanation for an outlier result. Do not forget that, even when someone says “Impossible to be wrong. Everything was done as described in the lab notebook.”
• Wrong procedures used. Sometimes procedures are not executed correctly for different reasons: lack of experience of the operator, standardized procedures do not fit to trial aims, … To choose the right procedure to do some operations is not always trivial, even in later stages of the development. If doubt choose one at your best criteria and keep it in all the trials until you decide to change for specific reasons. Avoid randomized use of different procedures and try to harmonized it among all the team members.
• Disregarding meaningful parameters. To underestimate or to forget some relevant parameter usually leads to a complex mixture of incongruous result (deficient reproducibility), many hours of unproductive work and some headache. Sometimes it is not easy to know from the beginning what the critical parameters are. Have you ever felt like a genius finding the most critical parameter ever seen in a reaction? Did you have problems to reproduce it? I guess you know what I am talking about.

Some of the deficiencies are quite easy to detect and repeating the trial can solve them. Yes, efficiency is affected, but the impact would be quite low. However, some other can be harder to detect and they can lead us to fake conclusions which is one of the main root causes of unproductive and time-consuming tasks during process development.

Fake conclusions: the worst enemy of efficiency during process development.

Step 4: Analysis of data. Be conservative drawing conclusions.

Data obtained from experimentation must be carefully analyzed. Usually, information of every single trial has low value if it is not compared with other experiments or corroborated with earlier gained knowledge. A usual mistake is to analyze the information superficially as data available increase. It is senseless to run additional trials if there is no enough time or resources to examine it properly.

We enjoy drawing fast conclusions, we really like to find “the solution”, but be cautious and try to confirm if your conclusion is consistent with previous data, with available information or with the state of the art. If it is already robust, try to demonstrate that is false by running additional experiments. The earliest you realize a conclusion is wrong, the best. Each wrong conclusion assumed as correct will turn into a delay in the project.

Step 5: Closing. Has your trial answer your initial question? Do you remember it yet?

Finally, each single trial (or set of trials) must be ideally closed by answering the question done in the first part of this article. In other words, achieving the objective of the trial. If so, probably some conclusion can be drawn. If not, it would be necessary to figure out why the trial is not providing the desired information. I bet that someone of the previous stages has failed.

• Is the objective well defined?
• Has been the trial designed suitably for the goal?
• Is there any deficiency during execution?
• Have data been correctly analyzed?

Find the root cause of trial failure and you will avoid further blind attempts.

Besides the result, the experimental work must be ended with a reasoned decision of which will be the next activity to do or the next plan of action to carry out, including to give up some line of research or even the whole project, if needed. Even when there are doubts about what to do next, take your time to decide but you should not procrastinate the decision.

Leaving out any of these stages or not carrying them out correctly can lead to different kind of problems but in the end, it will always result in a lack of efficiency in the development of the processes that will directly impact the time and cost of the overall project.

This is not an undisputed truth, just my point based on my experience. Please, feel free to disagree. I will appreciate it.

Facundo Andina

Make the time work for you.