The Importance of Understanding Process Validation for Pharmaceutical Process Chemists
kilomentor | 17 October, 2011 05:15
Process validation is a legal requirement applied to active pharmaceutical ingredient manufacturers by governments and it is administered by their regulators. Process development chemists contribute to the documentation of process validation.
Philosophical Differences between Process Chemists and Regulators
Process chemists when faced with a synthesis problem ask themselves the question, “How might this be made to work?” We then seek the optimal operating parameters for a scheme that achieves the desired physical reality. Regulators look at what we have achieved and ask themselves, “How could this fail?” Instead of looking at the optimal conditions that have been found, they are looking at the limits for failure of every critical parameter of the system we have optimized and are asking how this could upset the output quality of the system? We, as creators, struggle to succeed at jigging the system to give the result we need. We struggle to understand the coded messages of our experiments to deduce what a practical process would be. The regulators analyze what we have created to see at what points it could break down. Since any process can be made to fail by deviating enough from its teaching, this thinking is alien and frustrating to us.
Drug Substance vs Drug Product
As Dr. Oljan Repic has pointed out in his book,[ Principles of Process Research and Chemical Development in the Pharmaceutical Industry, John Wiley & Sons, Inc. p. 179-194], a chemical process for making an active pharmaceutical ingredient (API) from commercial bulk chemicals is very, very different from making a final dosage form from its API. This is because, in the former, chemical steps result in the creation of new substances by first combining a great many undesired materials (substrates, substrate impurities, reagents, catalysts, solvents etc) and then, after exposure to some reacting conditions, separating these and other undesirable materials (co-products, by-products, unanticipated and unidentified impurities) from the desired reaction product. Contrastingly, making a drug product from an API does not involve chemical reactions or purifications (other than drying). The drug product is some kind of mechanical mixture or micro-engineered assemblage of the inputs and so it is a simple combination of what is put into the process modulated by the manipulations. This is not to say that how an active ingredient is formulated into a drug product doesn’t affect the medicine’s effectiveness or even safety. It does. What I (and I think, Dr. Repic) are maintaining is that formulating is an inherently simpler change than synthesis and consequently failing to control the process of the former within the limits established is much more certain to cause the product to fail its specifications. Since there is no purification aspect in formulating, if it is garbage in, it will most certainly be garbage out. There is no possibility for purification in formulation so there is no capacity to rectify any serious deviations.
The synthesis of a pure chemical substance such as an API is nowhere near so dependent upon the starting material purity and the nitty-gritty of the process. In fact in the limit (100% purity), the properties of a pure chemical substance are completely independent both of its route of synthesis and the details of the most immediate transformation by which it is prepared. Even in the real world of actual APIs, the only evidence of the route of synthesis that remains in ‘pure’ product is the faint fingerprint of trace impurities that are still carried in it. Effective purification decreases this below any level of practical significance related to efficacy or safety and even, in many cases, below the level of analytical detection.
Even if it might come as a surprise to beginning formulators that how an API and the other excipients in a drug product recipe are combined can significantly alter the bioavailability of a tablet; it would not surprise even one tyro chemist that deviation from the details of a chemical process would cause the quality and yield of such product to suffer.
It is the Purification that Must Work
In a synthesis it is really only the final purification that needs to be effective to deliver a drug substance of the required purity, with the proviso of course that the analytical methods used to assess that purity need be validated. The process by which the final product is produced needs to be known essentially only so that the analytical methods can be tested for their capacity to detect and quantify the authentic likely/expected impurities.
This however is not the law. Regulators and politicians err on the side of caution and the validation rules they approve are more rigid than they need to be, because the rules derive historically from, and still closely resemble, those for formulated drug products. Process validation was originally conceived and implemented as a method to assure that each unit of a drug product could be guaranteed safe even when it could not be analyzed, either because the analysis would destroy the product or because the proper functioning of the product could not be analyzed. Drug security by validating the process was initially directed to drugs formulated for injection into the blood stream. If these were contaminated by bacteria the patient would likely quickly die. Even if one took a proper statistical representative sample of the product there still could be some fatally toxic units of product undiscovered. The solution was to control the process used to make the units so strictly that it would be extremely unlikely to make even one contaminated sample. This was the only course of action that could be effective because the fatal contamination came from the process (from the bacteria carried in the air) not from any starting materials.
Process validation also turned out to be valuable for assuring the efficacy of drug dosage forms where the delivery of the active ingredient from the dosage form ( ie tablet, capsule) into the blood stream of the patient was challenging. In these cases, analysis of the dosage form was not showing reliably whether a particular lot of drug product could deliver the clinically required exposure to the medicine. Drug product prepared by a validated process was found much more likely to produce the desired clinical outcome.
Process validation differs significantly from chemical analysis in its ability to provide protection from different types of contamination problems. A drug substance is almost always a pure chemical compound. It is a single molecular entity and can be analyzed; precisely, to confirm its constitution, and sensitively, to show the presence of homogeneously dispersed contaminants. Drug substances are not mechanical mixtures like flakes of this and that stirred together as is often the case for tablets and capsules. Such mixtures are statistically homogeneous only at the scale of a dose but heterogeneous at the particle scale. This is what one finds in a mixture that is pressed into a pill. Process validation is better than analysis to guarantee that heterogeneous, most often particulate, contamination is not present. The reason for this is that chemical analysis of an active pharmaceutical ingredient (API) requires taking a representative sample from the entire batch to perform the analysis upon. It is quite likely that even with an approved statistically sound sampling procedure a heterogeneously contaminated batch would not get any contaminant into the sample being analyzed. Take for example the situation where glass particles from a chipped piece of equipment contaminated a batch. The dense glass would migrate to the bottom of the container and might escape the sampling. Thus it would not be detected. A validated process, however, could make sure that there was no conceivable condition where any glass could remain in a final product. The final filtration before the final crystallization of an API is part of a process for just such a validation purpose.
Process Validation: A Development Scientist’s Most Important Activity
Drug discovery chemists do not need to learn about process validation; process development chemists do. It is the latter’s work that is being examined. Nevertheless, scientists in general always like to know whatever they can about the synthetic routes they devise and discovery chemists may eventually transition into career development chemists. The additional experimentation that process validation requires does increase scientific knowledge about that process. It answers questions about the practical limits of those process conditions that most significantly impact final product quality. But aside from the scientific merits, it is important that pharmaceutical chemists, particularly process development chemists, understand process validation, since, because the pharmaceutical business is a highly government regulated activity, any API they might make cannot be included in a medicine for sale unless it is made by a validated process. This makes validation the most economically significant test to which a scaled up industrial process can be subjected and so all the chemist’s activities should be viewed from the perspective of its contribution towards success in validation.
Many and perhaps most companies that manufacture APIs exclude scientists below the managerial level from any contact with the government regulators who assess the quality and completeness of the validation process. Likewise, it is only senior technical people who sign off on or assemble the validation report, which is the core document of the validation exercise. Yet exposing employees, who have very specialized tasks, to the all encompassing validation program lets them see their precise role in it. Specifically, chemists, who are focused on process development, are likely to do much better in organizing their data, writing reports, etc. if they understand the larger purpose their data and conclusions will serve when process validation protocols are assembled and executed and the Development Report is folded into the Validation Report. Indeed it is highly recommended that general training in the concepts of process validation should be given to all individuals who are part of the development and manufacturing program not just to those who require such training as part of Good Manufacturing Practice (GMP).
The Corporate Validation Policy
Validation is so important to API producing companies that very often it is heralded in a written Corporate Validation Policy. Where this document exists it is the overarching document pertaining to the validation methodology. As such it sets down general principles that employees should adhere to, rther than the nitty-gritty details; nevertheless, it is a good place to start in examining that most important activity called process validation.
A lifecycle policy approach to the validation concept is regarded by many experts as preferable or even essential. Validation activities begin with the inception of the development program and logically should not end until the product is retired. A corporate validation policy should provide validation expectations over the lifetime of the process. It should state the company’s expectations for both the organization’s scientific and its documentation goals. Both of these classes of goals need to be kept in mind and adhered to from the first day of process development to the last. In other words the Product Development Report (PDR) should be a ‘living’ progress report that is continuously updated through the product lifecycle. Many firms make the mistake of creating a static, concrete, dated, PDR paper which is archived once that first regulatory approval is obtained. Because improvements are being made throughout the lifetime of the process, these may require additional validation work and regulatory filings. These changes should set in motion an update of the PDR.
Since every organization has somewhat different internal structure, any policy needs to allocate responsibilities for maintaining the product/process documentation including the mature product responsibilities relating to monitoring, trending, and change control.
The Validation Master Plan
The Validation Master Plan is the most senior planning document that pertains to a specific validation. It makes concrete and project specific what is set down in the Validation Policy only in a general way. It takes into account even the most recent changes in the organization’s structure, responsibilities and reporting chains. The Validation Master Plan assigns actual project milestones to named persons and groups. It sets dates for meeting those milestones. It sets out what activities are to be performed when and it decides the science that will take place at small and at large scales. If the validation is for a new drug substance the Validation Master Plan needs to be revisited and modified as the clinical phases advance.
All aspects of the validation documentation are important. The regulators look not for what the company considers most important; what is scientifically most significant; what has been presented most clearly and completely or where the maximum inherent risk is, but what has been overlooked because that is from their perspective where that process deviation leading to dangerous medicine could most likely arise.
Some Aspects of Process Validation that are not the Development Chemists’ Responsibility
Validation must demonstrate that the process being examined is performed under good manufacturing practice (GMP). Mercifully there are many aspects of GMP that are not the responsibility of development chemists. Every organization that manufactures product for sale, makes batches for clinical (human) tests, or to support regulatory filings must have a separate and independent quality control unit to test, approve, and reject products; set and approve specifications for raw materials and products; and write and approve these and other written procedures. Another requirement of GMP is that all personnel associated with production must be trained concerning GMP and the company must be able to prove that training with the signatures of those trained. You may have been asked to submit your resume to a company representative responsible for meeting GMP standards. The company is required to show that its scientists too are adequately credentialed for their work.
The facilities and equipment to be included in a process need to be qualified before they can be used in an acceptable validation exercise. These qualifications are usually unfamiliar to practicing chemists. They are matters that are implicit in our everyday activities and are completely taken for granted by us.
Facility qualification relates to the suitability of the physical site and its infrastructure and personnel to carry out the process that is to be validated. Most of these matters are strictly the domain of architects, contractors and engineers.
Drug substances in general have requirements for the facilities in which they are manufactured. The FDA will not allow someone to make a GMP drug substance in a bathtub in a garage! It insists upon written assurance that the manufacturing facility has adequate space, heating, ventilation, plumbing and maintenance.
The Development Report
The core document that development chemists do contribute for API validation is the Development Report. Process chemists recognize the Development Report as the history of the entire process development but it is at the same time a regulatory validation document of enormous importance. Validation would need to be done, if for nothing else, at least to convince regulators that the product is reproducibly safe, but API product safety is not the only goal of process development. Optimization aims to minimize cost, maximize throughput, achieve appropriate safety, reduce environmental burden and deliver consistent, convenient operation of the steps. Consequently some IPCs and some specifications are not critical in the regulatory sense of controlling final product quality. Deviations from these do not put the API quality at risk. Some of these aforementioned standard ranges alert batch sheet reviewers, when they are changing non-randomly, to potential creeping deviations (trends) that can have economic, safety, environmental or simply consistency consequences, without there being quality ones. This means that an API could fail a noncritical IPC and still be safe and effective to use. Failing a specification or IPC that correlates with a critical process parameter (CPP) and in turn a critical quality attribute (CQA), on the other hand, would mark a failed batch.
A synthesis can in strict theory be traced back to the most elementary of chemical building blocks- even atoms themselves. What needs to be decided for validation is how many of the steps in the retrosynthetic direction need to be validated to insure the required safety. This ‘depth’ of the validation study establishes what chemical transformations or treatments are included within the validation and what steps are accepted simply as the making of starting materials to be controlled by their material specifications independent of their particular synthesis.
Today, it is a good bet that most synthetic API processes are multi-step with at least one complex chemical transformation. Similarly it is unlikely to be disputed that the final process step that creates an API, and the subsequently entailed purification, needs to be validated. Beyond this there is no consensus answer as to when to apply process validation for intermediate steps in a multi-step process. Each process it seems must be evaluated as a separate case. So many factors are in play. This does not mean that a coherent policy or procedure cannot be established for a company describing how it should do that evaluation. In fact, putting such a decision tree in place can lead to better, faster, even less expensive, choices.
The process validation for an API is at least the sum of separate validations; one for each intermediate step that contains a critical parameter. Thus determination of what steps are to be included in the full validation amounts to identifying what steps contain at least one critical parameter. This is much the same as asking what the starting materials are for the steps that will be validated because steps for making an API starting material do not need validation. Terminology can be the source of as much disagreement as the chemistry in these discussions so definitions are important. A ‘step’ leads either to an isolated or non-isolated intermediate. Other important determinations to get straight are: what is(are) the ultimate intermediate(s) or final intermediate(s), and what overall scheme of reactions and purifications will be required to produce the final API.
Determining what constitutes a starting material is debatable. According to Roger W. Koops, [Process Validation of Synthetic Chemical Processes for the Production of Active Pharmaceutical Ingredients (APIs), Journal of Validation Technology, Volume 8 Number 2 January 2002], “a starting material should be a readily available item that has a standard grade (or grades) associated with it, and has been well characterized. If produced under contract by a vendor, it should be produced by a known and established process, and the end product should be, again, capable of achieving a standard grade. Second, the vendor should be qualified (under a vendor qualification program)to to ensure that a material meets consistent quality standards. A rudimentary quality agreement should be established to outline change notification and quality requirements. What must be avoided are unshared, unannounced process changes that can change the impurity profile of the downstream drug. If a material is manufactured by a contract party and the process was supplied by the innovator, the material is simply a third-party manufactured intermediate. The contractor now becomes an extension of the innovator, and the transferred process must be included in the full validation evaluation”.
Starting materials, generally speaking, are inputs to a synthesis whose atoms are partially incorporated into the drug substance. They do not need to have their own syntheses validated. Their quality is accepted based on their certificates of analysis. In contrast, intermediates are created directly or indirectly from starting materials. They arise downstream in the process with respect to some starting materials. Most intermediates arise in steps that need validation, but at least in principle there can be exceptions. A step starting with an intermediate and resulting in another intermediate does not need to be validated if the interconversion process step has no critical parameters. That is, the step is so rugged that no matter how it is executed it does not affect the purity of the final drug product. This would be a rare case indeed. The reaction would need to be ultra tolerant of reaction parameter deviations and/or the isolation would need to be exceptionally powerful at excluding impurities.
As the overall chemical development proceeds an impurity profile for the final product should emerge The validation policy should emphasize the importance of characterizing process impurities as early in the development process as possible. From this, each intermediate process step can be evaluated as to its influence on the purity of the end product. If an intermediate needs to be controlled because it could contribute to an adverse quality profile of the API, process validation should be applied to its related process step.
Each intermediate needs to be evaluated for how it contributes to the final API, in regard to the impurity profile and the specific process for that intermediate. Besides isolated intermediates, non-isolated intermediates should also be evaluated, since steps comprising these may be ones where even stricter controls may be required. For example, an intermediate may not be isolated due to structural instability when not in solution. In this case, concentration and/or potency of the intermediate may be a critical attribute that needs to be controlled.
Either the corporate validation policy or the particular validation plan needs to provide guidance whether the validation exercise at plant scale should be designed so as to “stress” the accepted parameters by operating at the maximum or minimum accepted level (the edge of failure), or to aim for the set point target of each parameter. Some companies have attempted to matrix the parameter limits in an effort to test the extreme limits of process parameters.
This author holds the view that the laboratory development phase is the place to establish the boundaries of the operating ranges, and how combinations between extremes of parameters affect the final outcome of the process. Process validation, by definition, is used to demonstrate the limits of consistency of the process as designed and thus should aim for the set points of operating parameters. From a manufacturing standpoint, it is advantageous to operate based upon the targeted parameter values for batch consistency.
Because the Process Development Report (PDR) is such a key element of validation documentation, it is important for the corporate validation policy to set out at what points interim reports need to be submitted during process development because interim reports make the compilation of the PDR easier to manage.
The Corporate Validation Policy may need to provide some guidance concerning choosing critical parameters since their number strongly affects the breadth and hence the expense of the experimentation.
Some organizations choose to call critical only parameters, that affect the impurity profile and the ability of a material to pass specifications. However, the ability of an output product to meet specification is not the sole indicator that its process is running smoothly. There are sometimes parameters that, for example, effect yield, without affecting the final product’s properties. Nonetheless, variation of yield beyond normal experimental error points towards a process out of control. Whatever the inclusion/exclusion rules, the basis of decision can be either a general policy prescription or something considered only on a case-by-case basis at the level of the Validation Master Plan (VMP).
Validation Protocols
Validation Master Plan amalgamates validation protocols. The hallmark of a process step that is ‘under control’ is that is predictable. The sequence of operations that constitute the step are unequivocally established in advance, the acceptable operating ranges of the critical parameters are set down in advance, and the critical specifications of the output material are set down in advance. So what sets a validation batch apart from any other pilot experiment is that there can be no recourse to any procedural or corrective activity that is not set out in advance in the protocol. The organization cannot redefine success in a process step being validated after the requisite batch is complete and so eliminate deviations between what was predicted and what transpired. The regulators are the referees who decide whether, when the organization executes the protocol, it follows its own procedures and meets its own standards.
Although some experts urge that the protocol should be drafted to be ‘as concise as possible’, protocols are normally known for their verbosity. A process validation protocol will describe in excruciating detail the ‘how to’ for the validation exercise. The protocol resembles a master record in many ways but provides a greater level of descriptive detail regarding the execution of the process, the monitoring of the steps containing critical parameters and sampling instructions for the in-process checks, particularly critical in-process checks. In every case, either a procedure should be described in detail in the protocol or a reference should be provided in the protocol to an already established procedure for handling the task (ie a standard operating procedure). The instructions should not just be sufficiently unambiguous that it would direct a trained operator to choose the correct action. It should be such that a regulatory reviewer can only conceive a single proper performance of the instruction. Certain styles of drafting can condense the presentation without losing explicitness. Where applicable, data should be entered directly into the protocol. If an exercise is particularly complex, a series of protocols may be preferable. Protocols should have places in them for reviewers’ initials or signatures.
One way of looking at the components of a validation protocol is to see that they reproduce many of the parts of a summary report about a set of batches, but unlike the report the protocol is written in advance (remember after developing a process you are supposed to be able to predict everything that is crucial to product safety in advance).
The protocol should include many sections that process chemists would likely consider report padding. There should be an introductory section defining the scope of the protocol relating back to the overall Validation Master Plan. The question where does this particular protocol fit into the overall validations should also be answered. The responsibilities for execution, testing, review and approval for the overall validation exercise need to be reproduced here even if they also appear elsewhere. Also, for each particular protocol the department and job responsibilities need to be restated and the entire protocol needs to be approved by the top technical officer and the top quality officer. The purpose of the sign-off is to make unequivocally clear that top management has seen and bought into the exact protocol exercise that is to be executed.
Because a description in words alone may be more difficult to follow than an illustration, a process flow diagram (PFD) for the particular step or portion of a process that is being validated should be provided annotated with starting materials, reagents, solvents, process conditions, transfers, by-product concentrates, waste streams, and purifications etc. If there are any recycle loops or patches prepared in advance these need to be presented here in advance. The operations should be assigned an identification code so that further elaborations can be related unequivocally to the correct portion of the PFD. If seed crystals are used where they come from, what quantity they are what specifications they meet need to be set down. It is generally important and advantageous to identify critical process steps containing one or more critical process parameters on the PFD document. The idea is to give the operators no ‘wiggle room’ as to how the execution of the exercise is to proceed. This in turn gives regulators confidence that you have confidence in your process understanding and control.
A complete protocol should provide a bill of materials for every consumable item used during the process. Particular attention should be given to be sure that even materials that are not part of the chemical transformations also be there, such as filters, filter aids, packaging for chemicals before and after the main chemical transformation, seed crystals etc..Some materials have both a name and an identification code. A reference to the specification or COA should be in the listing. If the PFD has the operations in the sequence labelled, this label should be here as a cross reference.
All the equipment used in the process needs to be referenced. The actual documentation does not need to be reproduced in the report but the IQ, OQ and PQ report numbers need to be given as well as the cleaning procedures and the cleaning validation reports and instructions. The ranges of parameters, particularly any critical parameters that will be used should be noted and highlighted. Of course, the required operating range needs to be within the validated range for each equipment. If there is any special procedures that involve the equipment during the process these procedures must be detailed here or referenced accurately.
Facility documentation only needs to be referenced along with the DQ, IQ, OQ and PQ materials. The only recapitulation of the facility validation would refer to the particular importance of any unusual facility capability for the protocol being validated. One example of such a special competence would be the inclusion in the facility of Class C clean rooms.
Critical Parameters
Only at this point do we get to elements that process chemist would consider the essence of the scaled up inventiveness. Specific page references should be made to the Development Report where the critical parameters are established and the data used to establish them is organized. Various ranges of the critical parameter need to be discussed. The master procedure range is the range that will be set out in the master batch sheet. Within that range will be the set point which is the numerical setting used on the regulating equipment. Under regular operating conditions the set point can only be transiently met because of the equipment’s inherent limitations to overshoot and undershoot. The master procedure range of course cannot be less than the range set by the equipment’s inherent limitations. The next wider range is the verified range which is the range actually experienced in the plant. The Development Report may make predictions of a wider verifiable range from laboratory experiments. Using development data, a prediction will also be made of the process limits for the parameter. This is the edge of failure exceeding which the API is expect to suffer quality repercussions. Another limit that may be mentioned is the practical range of the equipment with respect to the critic parameter. For example in a steam-heated reactor the practical limit for external heating would be 150 C. For each critical parameter the controls that will be applied to maintain it need to be listed.
In-Process Tests
Every batch sheet comprises in-process tests that are used to assess the progress of the processing. A list is required of all the in process tests within the protocol showing at what point the sample for testing is taken, what the test method will be, who will perform the test and how the results will be reported. If the test is a go/no go test which is repeated until the go signal is obtained this should be explained. Achieving a particular result may be a critical parameter. If this is the case that needs to be made clear. For example neither the reaction temperature nor the time may be critical but the outcome of an in-process test for reaction completion may be critical. Thus a lower temperature and longer time or higher temperature and shorter time may each achieve the completeness required to pass a critical IPC.
Sampling Plan
Process development chemists know that sampling heterogeneous reaction mixtures does not give results representative of the mixture in its entirety. Even testing of the separated phases individually is fraught with difficulties. Solids that are to be sampled are rarely homogeneous. Even homogeneous solid mixtures can demix with mild shaking. As a consequence if one is going to have a significant analytic result, the sampling must achieve a representative sample. Within the protocol, the plan should indicate the sampling location, the amount of sample to be taken, how it is to be collected, how it is to be stabilized if necessary, how stored, and how transported to the testing facility. Both operators and regulators need to know before the exercise exactly how the sampling is planned. Of course at what point the test will be done must be indicated unequivocally. If there is a sampling deviation, it is helpful if it is reported in the batch sheet before any result for that sample is available.
Acceptance Criteria
All criteria that must be within predefined ranges must be listed in advance both critical parameters and IPC results and testing on the final product.
Deviations and Investigations
The most desirable outcome from executing a validation protocol is that all critical parameters are held within the operating range and the resultant intermediate has test measurements within the range of intermediates that have provided safe and effective IPA. That is what we aim for. What if some critical parameters are measured outside the master procedure range but not outside the verified range? What if some critical parameters have strayed outside the verified range but the intermediate seems unchanged in properties from intermediate prepared without deviations? What if the migration out of bounds was brief or very small in magnitude? What is very important is that as much as possible the assessment of the impact of a possible deviation should be made before the validation is actually executed. Identifying a parameter as critical when it is not critical is unlikely after a thorough development but getting the limits of failure somewhat incorrect is rather likely. One does not want to abandon or fail a validation batch when in fact it will result in pure safe final IPA. Falling outside of most critical parameter ranges is easily avoided but sometimes the desired operating range falls close to a limit of failure because the manufacturing needs to achieve some objective not related to final product safety or effectiveness but to minimizing cost, improving safety or reducing waste. It is in these situations where preliminary considerations of less desirable outcomes can be invaluable.
Equipment and Facility Qualification
Equipment Design qualification (DQ) relates to equipment/machines. Design qualification is normally the responsibility of the equipment vendors. Successful Installation qualification (IQ) demonstrates that the equipment has been properly installed. Operational qualification (OQ) demonstrates that the equipment can achieve the operating parameters intended when exposed to typical substrates. An OQ does not not need to specifically test the conditions of a particular process. That is the function of performance qualification (PQ). In general, it is best to have all facility and equipment qualifications completed prior to process validation (certainly all IQ/OQ should be completed). Performance Qualification (PQ) can be performed during process validation, but it is generally advisable to have completed this testing prior to process validation since it would add additional risk to the exercise. However, there may be some instances where it is useful to concurrently perform an equipment or facility PQ during process validation. In those cases, a separate PQ protocol should be drafted and referenced in the process validation protocol.
Analytical Validation
The validation of the required analytical methods that are used to follow the trends of the critical parameters during the process experimentation may be regarded as a sub-program to the overall validation project. Secure conclusions can only be derived from reliable data collected using validated methods, so methods validation needs to proceed hand-in-hand with the examination of the chemical transformation processes themselves.
Cleaning Validation
Equipment cleaning methods need to be included as part of the Validation Report. Cleaning validation is also sufficiently complex that it is better treated as a sub-discipline with its own report sub-section. Some API producers prefer to have standard cleaning procedures for their equipment which are always executed irrespective of the process that is being cleaned up. These methods are part of the SOPs for operations. Where these methods alone cannot be expected to satisfy the cleaning requirements additional cleaning may be set in place as part of the Master Batch Record following a particular exposure. Another organization might specify the cleaning treatment based on what the chemical step has exposed equipment to. That decision can be part of the Corporate Validation strategy. Cleaning methods can be developed in parallel with the chemical process development and cleaning validation can conveniently occur either prior to or concurrent with process validation. These decisions can also be policy decisions or part of a Validation Master Plan for a particular validation.
Cleaning of reactors in which any broadly insoluble, polymeric, or tarry by-product is formed can expend valuable reactor time and penalize throughput. When process chemists include such steps in a process going to scale up they should take a special responsibility to work with production personnel to develop efficient cleaning or at least pass along a warning of the difficulty and whatever they already know about overcoming it.
Conclusion
In conclusion, although process chemists need not become validation experts and although they may not be personally enmeshed in the validation exercise, some overall familiarity with the subject, from their own perspective and emphasizing their unique role, can enhance their work and their enjoyment of it.
General |
Permalink |
Add comment (SPAM IS DELETED INSTANTLY) |
Trackbacks (0)
Is Glycerin a General Green Solution to Solvent Recovery and Replacement for Process Chemistry?
kilomentor | 13 June, 2011 07:15
General |
Permalink |
Comments (6) |
Trackbacks (0)
More Information about using Inorganic Salt Complexes to simply Purify Alcohols at Scale
kilomentor | 16 May, 2011 08:19
Barry Sharpless et al. were the first to teach the separation of alcohol mixtures and alcohols from non-alcohols using calcium and manganese salt complexes. The original J. Org. Chem. paper [J. Org. Chem. 40(9) 1252=-1257 (1975)] unluckily recommended as a most preferred modality, using one mole of anhydrous calcium chloride per two moles of alcohol. This turns out to have been a good pressumption based on theory, but not such a good practical recipe. Soon after Sharpless’s publication, Cook et al. [Biochemical and Biophysical Research Commun. 68(1) 143-148 (1976)] showed in a crystal structure that 2:1 alcohol:calcium would be theoretically expected. Mentally extrapolating from what the Cook paper says about the probable structures, strings of calcium chloride ions are interspersed by layers of hydrophobic organic substructures where the thickness of these hydrophobic layers is somewhat related to the length and shape of the hydrocarbon substructures. How fast and how purely these clay-like layers form probably depends upon how rapidly the alcohols get into the layers and how quickly they sort themselves out between layers. This is however conjecture. In any case, in the light of more experience, this 2:1 stoichiometry recommendation turned out to be poor advice, Professor Sharpless has told me. Large excesses of calcium chloride, generally, work best.
It is not just some alcohols that complex inorganic salts such as calcium chloride, calcium bromide, manganese chloride, lithium bromide etc. ; all alcohols do but there are differences in affinities and separation is only possible where precipitate/crystallize occurs. This formation of solid complexes by stirring the substrate mixture, and the inorganic salts together with an appropriate organic liquid (most often hexane) depending upon the specic situation may be driven by relative rates or thermodynamics.
Alcohols are not the only functional groups that form complexes with such inorganic salts, many other functional groups do, but not as strongly and they don’t often precipitate. The interaction between these other functional groups do cause the inorganic salts to dissolve as some sort of soluble complexes or miscelles but precipitation/crystallization is not typically observed.
Whether an insoluble complex forms depends also on the character of the salt chosen to combine with it. L-menthol for example formed complexes with calcium bromide and manganese chloride but apparently not with calcium chloride.
Lower alcohols seem to enhance the rates of complexation perhaps by getting the inorganic salts dissolved as complexes in the organic liquid although there must be something more than this because ethanol catalyst was used in the exchanges of Table 1 in the J. Org. Chem. paper even though the starting complex would be somewhat soluble in the hexane liquid. Yet still catalysis is found to be useful.
Based on this preliminary understanding, certain questions that affect the ruggedness and the practical breadth of the method come to mind.
Are 1,2- or 1,3-diols very strongly and dependably complexed and precipitated? It would be useful to have an almost generally useful method to drop out such compounds from a reaction mixture.
Will the presence of other complexed functional groups in the compounds to be separated tend to screw up such a complexing/ precipitating scheme?
For example, the geraniol/citronellol mixture works since the other functional groupt here (alkene) does not form complexes; but what if both molecules contained alcohols and say ketones? Ketones do form complexes but the complexes don’t seem to precipitate (although cholestenone does form an isolable complex with calcium bromide). Actually, not all ketones appear to form complexes. The solvents methyl isobutyl ketone (MIBK) and methyl amyl ketone can be used as the liquid phases for forming complexes and so must participate very poorly (or it would swamp the effect).
Other examples of complexes between inorganic salts and organic compounds containing an alcohol group were known. In BE1555968 reference is made to isolating sterols by heating in hydrocarbon solvents with 14-16 molar equivalents of anhydrous zinc chloride and isolating the solid. This is referenced to in DE827199. Also in BE1164769, there is described a process for isolating sterols from mixtures by dissolving the mixture in a hydrocarbon, adding to the solution an aqueous solution of a metal salt suitable for complex formation, removing the water to a great extent by azeotropic distillation, and after cooling the mixture, isolating and splitting up the precipitated adduct in the usual way.
In the patent BR1555968, 3-hydroxy steroids and 3 oxo-steroids in various natural oil mixtures were separated by treatment with calcium bromide dehydrate in MIBK or n-amyl methyl ketone. It is unclear why the 3-oxo-steroids gave complexes under these conditions since all other literature suggests that ketones form weak complexes and/or these complexes are soluble and do not precipitate.
Kilomentor thanks Professor Sharpless for answering questions about this technique and for directing me to the Cook et al. paper.
General |
Permalink |
Comments (1) |
Trackbacks (0)
Saturated Calcium Chloride Brine for Separations?
kilomentor | 11 May, 2011 15:54
Calcium chloride is very soluble in water. It is a slow, inefficient but ‘thirsty’ drying agent. In moist air the anhydrous solid is deliquescent. Quite a few solvents can be expected to be immiscible with its aqueous brine solution and these could serve as an immiscible partner phase for liquid-liquid extractions. Saturated aq. calcium chloride can be expected to behave similarly to saturated sodium chloride solution; it forms a separate phase with isopropanol or pyridine, liquids with which water alone is miscible. But what molecular structural types would interact and pass into a calcium chloride brine that would not as easily or more easily dissolve in plain water?
This question seems appropriate because it would be logical to think that the complexing power of calcium ion would be largely erased when comingled with so much of the Lewis base, water. Yet this is apparently not so, if expired patent US 3225113 is to be believed. According to this patent, calcium chloride brine forms complexes with hydrocarbon ethers, when the ratio of the number of carbons oxygens is <8:1. The best complexes are when the ratio is 2:1 as in polyethylene glycol ethers.
This complexing could be useful in work-ups that require a solvent switch where an ether or glyme solvent needs to be removed in favour of another solvent. Thus, when using soluble polyethylene glycol polymer as solvent/diluent in synthesis there may be an option besides the prior art procedure of precipitating these polymer with hexanes or diethyl ether. Aqueous saturated calcium chloride is very likely to precipitate the glyme as a filterable solid.
Patent US3225113 shows that a solution of calcium chloride in water can precipitate quantitatively glyme solvents in the presence of another hydrocarbon-like substance. Filtration and washing the solid with the non-polar hydrocarbon cleans up the glyme calcium chloride complex. The aqueous and hydrocarbon phases can be themselves separated since they are immiscible. This complexing also occurs between aqueous calcium chloride and 1,4-dioxane suggesting a way to solvent switch from dioxane.
It is thus also not a foregone conclusion that no organic species would occupy the calcium chloride/ water layer in a two phase combination with a hydrocarbon or halogenated solvent. A substance that complexes with calcium ion might enter such an aqueous salt layer if its complex is stronger than the water one. Glymes apparently meet the criterion. 1,2- or 1,3-diols or cyclic polyethers might also prefer aqueous calcium chloride more than water alone or might precipitate out when so treated. This would be a particularly propitious means to recover expensive cryptate cyclic ether ligands.
Whether a particular polyether is likely to be strongly complexed by an aqueous calcium chloride solution may be judged by a simple test described and applied in the aforementioned patent. In it, 5 cc. of a saturated calcium chloride solution was added to a test-tube containing the ether. No heat was added to the reaction zones. The temperature rise at the end of two minutes is measured. Symmetrical dioxane gave a rise from 26 to 42 C. The increase for bis[2(2methoxyethoxy)ethyl] ether was from 28 to 71 C with solid complex separation.
Aqueous calcium chloride stirred with a potentially complexing compound in an immiscible organic solvent that does not complex calcium chloride, such as a hydrocarbon or MIBK, could it seems partially precipitate some material, dissolve some in the aqueous –salt phase and dissolve some in the hydrocarbon or MIBK phase. Filtration would separate the solid and the two liquid layers could be themselves separated giving three distinct phases. Uncomplexed organic will most likely be concentrated in the hydrocarbon or MIBK solution.
This method might be applicable to the frequent problem changing from a high boiling polar aprotic solvent such as DMF, DMA, N-methylpyrrolidone or DMSO, to a low boiling hydrocarbon solvent by distilling away the dipolar aprotic solvent using a small volume of a higher boiling glyme solvent to chase it, adding the low boiling hydrocarbon phase and then precipitating the glyme complex with an aqueous calcium chloride solution. Filtration would give two phases: the low oiling hydrocarbon liquid containing the reaction mixture and the residual saturated calcium chloride brine.
General |
Permalink |
Comments (1) |
Trackbacks (0)
Use of free radical inhibitors or antioxidants to increase the overall yield of organic synthesis steps.
kilomentor | 25 March, 2011 13:30
General |
Permalink |
Comments (2) |
Trackbacks (0)
The Pharmaceutical Salt called Pamoate or Embonate
kilomentor | 26 February, 2011 14:47
A solution of 3- [2- [4-(6-fluoro- 1,2-benzisoxazol-3-yl)- I-piperidinyl) ethyl] -6,7,8,9-tetrahydro-2-methyl-4H-Pyrido[1,2-a]pyrimidin-4-one,19.70 g (0. 048mol) in ethanol (600ml) was added to a solution 18.64 g of pamoic acid (0. 048mol) in N,N-dimethylformamide (400ml). (1g/22 ml )
The mixture was stirred for 3 hours. The resulting precipitate was filtered off by suction, washed with ethanol and dried, yielding 3 1 g (8.1 %) of 3-[2-[4-(6-fluoro- 1,2benzisoxazol-3-yl)- I -piperidinyl)ethyl) -6,7,8,9-tetrahydro-2-methyl-4H-pyfido[ 1, 2ajpyrimidin-4-one 4,4'-methylenebis[3-hydroxy-2-naphthalenecarboxylate) (1: 1); mp. 269.2'C. This is a very poor yield of salt; just 8.1%. Pamoic acid apparently is soluble in dimethylformamide. This is useful information. The risperidone was dissolved in the usual ethanol. Perhaps the experimentalist did not wait long enough for the solid to all precipitate. They filtered after 3 hours. WO05016261A2 Example 1:
The pamoate salt of haloperidol (mw 375.86) can be prepared by treatment of haloperidol with pamoic acid or pamoate salt in solvent. Haloperidol pamoate can be prepared by adding a solution of haloperidol in an appropriate solvent, ie. ethanol with acetic acid, to a solution of disodium pamoate (mw 432.36) , pamoic acid (mw 388.36) or other pamoate salt and leaving undisturbed for 1-3 or more days until precipitation. Alternatively, other methods such as evaporation, slow or fast cooling or stirring solutions can also be used to precipitate salt.
Specifically, 2.5 ml of a 0.1M solution of haloperidol (94 mg) in an acidified ethanol (5% acetic acid) was added to 2.5 ml of a 0.1M solution of disodium pamoate)(108 mg) in ethanol/water (50/50). The mixture was allowed to sit at room temperature for 1-3 days. The resulting precipitate was filtered off by suction, washed with ethanol and dried in a vacuum oven at 60°C, yielding 240mg (1`26% of theoretical) of 1:1 haloperidol pamoate salt.
Example 2:
2.5 ml of a 0.25M solution of haloperidol (0.625 mmoles, 235 mg) in an acidified ethanol (5% acetic acid) was added to 12.5 ml of a 0.05M solution of disodium pamoate (0.625 mmoles, 270 mg) in ethanol/water (75/25). The mixture was allowed to sit at room temperature for 1-3 days. The resulting precipitate was filtered off by suction, washed with ethanol and dried in a vacuum oven at 60°C, yielding 206mg of 2:1 haloperidol pamoate salt. The stoichiometry is for a 1:1 adduct and haloperidol only has one basic nitrogen that can be neutralized. If the salt is 1:1 the yield would be 47%.If the dibasic pamoic acid neutralizes two haloperidols thern the yield is 59%since half the pamoic acid must remain without a reactive partner.
Example 3:
2.5 ml of a 0.25M solution of haloperidol in an acidified ethanol (5% acetic acid) was added to 6.25 ml of a O.1M solution of disodium pamoate in ethanol/water (50/50). The mixture was allowed to sit at room temperature for 1-3 days. The resulting precipitate was filtered off by suction, washed with ethanol and dried in a vacuum oven at 60°C, yielding 264mg of 2:1 haloperidol pamoate salt. - 1 1
Example 4:
ml of a 0.05M solution of haloperidol in an acidified ethanol (5% acetic acid) was added to 1 ml of a 0.25M solution of disodium pamoate in ethanol/water (50/50). The mixture was allowed to sit at room temperature for 1-3 days. The resulting precipitate was filtered off by suction, washed with ethanol and dried in a vacuum oven at 60°C, yielding 107 mg of 1:1 haloperidol pamoate salt.
Example 5:
5.ml of a 0.05M solution of haloperidol in an acidified ethanol (5% acetic acid) was added to 2.5 ml of a O.1M solution of disodium pamoate in ethanol/water (50/50). The mixture was allowed to sit at room temperature for 1-3 days. The resulting precipitate was filtered off by suction, washed with ethanol and dried in a vacuum oven at 60°C, yielding 119 mg of 1:1 haloperidol pamoate salt.
Example 6:
A (0.05 - 0.5M) solution of aripiprazole in an acidified ethanol is added to a (0.05 - 0.5M) disodium pamoate solution in a mixture of water/ethanol (100/0 0/100). The mixture is allowed to sit at room temperature for 1-3 days. The resulting precipitate is filtered off by suction, washed with solvent and dried in a vacuum oven at 60°C. These methods teach the method of adding the base acidified with 5% acetic acid in ethanol to the disodium pamoate in ethanol/water. The disodium salt is more soluble and so this method depends upon the acidification of sodium pamoate with acetic acid to create the pamoic acid in situ where it can interact with the amine in the presence of acetic acid. The more insoluble amine pamoate crystallizes. These examples illustrate the fact that pamoates often must be allowed to change form from a gel like form to crystalline over some time. Heating sometimes accelerates this change. WO04017970A1 (C) Preparation of 3-(3-methoxyphenyl)-3-(3- dimethylaminopropyl]-4,4-dimethyl-piperidine-2,6-dione pamoate salt (anhydrous) A solution of AGN-2979 bisulphate salt obtained in Step B (1 mmole, 430 mg) in 10 ml of water was mixed with methylene chloride (20 ml) and basified with aqueous ammonium hydroxide (29% w/w). After separation of the layers, the aqueous phase was extracted twice with methylene chloride. The combined organic phases were dried over anhydrous magnesium sulphate and the solvent was evaporated under reduced pressure. The residue was dissolved in ethanol (10 ml) and mixed with a hot solution of pamoic acid (embonic acid, 390 mg,1 mmole) in hot ethanol (30 ml) and the mixture was heated to reflux. After cooling, the pamoate salt crystallised and the salt was recrystallised in hot ethanol to give a pale yellow powder (melting point = 146°-150°C. The procedure separates free base, evaporates to an oil and dissolves it in ethanol. It is mixed with a hot solution of pamoic acid dissolved in hot ethanol. The embonate came out in crystalline form on cooling. WO05075454A2 FORMS OF 4-(4-METHYLPIPERAZIN-1-YLMETHYL)-n-[4-METHYL-3-(4-PYRIDIN-3-YL)PYRIMIDIN-2-YLAMINO)PHENYL]-BENZAMIDE - IMATINIB Example 10
4.l(4-Methyl-1 -piperazinyl)methyl]-N-[4-methyl-3-[ [4-(3-pyridinyl)-2- pyrimidinyl]amino]phenyl]- benzamide, pamoate A mixture of 4-[(4-methyl-1- piperazinyl) methyl]-N-[4-methyl-3-[[4-(3-pyridinyl)-2- pyrimidinyl]amino] phenyl]-benzamide (4.94 g, 10 mmol) and 4,4'-methylenebis[3-hydroxy-2- naphthoic acid (Fluke, Buchs, Switzerland; 3.88 g, 10 mmol) in ethanol (50 mL) is heated.
Water (25 mL) is then added. Upon cooling, the product crystallizes and is filtered-off and dried to afford 4-[(4-methyl-1- piperazinyl)methyl]-N- [4-methyl-3-[[4-(3-pyridinyl)- 2- pyrimidinyl]amino]phenyl]-benzemide, pamoate as a pale- yellow solid, having the following analytical properties: Analysis found: C, 69.12; H. 5.62; N. 10.88%; H2O, 2.50%. Calculated for C52H47N7O7- 1.26 H2O: C, 69.04; H. 5.52; N. 10.84%; H2O, 2. 51%. Heating pamoic acid in ethanol will create some solubility. The solids must have dissolved since the addition of water is usually done to the point of turbidity and then the crystals allowed to come out as the solution cools. WO05012233A1 MELDONIUM SALTS, METHOD OF THEIR PREPARATION AND PHARMACEUTICAL COMPOSITION ON THEIR BASIS EXAMPLE 10
Meldonium pamoate (1:1; x H20). Meldoniurn (5.46 g, 30 mmol) and pamoic acid (5.82 g, 15 mmol) are mixed with water and acetone (15 ml), the formed suspension is evaporated, 30-40 ml toluene is added to the residual viscous mass, it is grated, and evaporation is repeated. If the residue is insufficiently dry, treatment with toluene is repeated. Mp. 128-133°C (decomp.). H NMR spectrum (DMSO-d6), 6, ppm: 2.41 (2H, t, CH2COO-); 3.14 (2H, t, CH2N); 3.25 (9H, s, Me3N+); 4.75 (2H, s, -CH=(pam)) , 7.12 (2H, t, Harom); 7.26 (2H, td, Harom); 7.77 (2H, d, Harom); 8.18 (2H, d, Harom); 8.35 (2H, s, Harom). Found, %: C 62,90; H 5,83; N 4,98. Calculated, %: C 63,07; H S,84; N 5,07. Initially H:O content in the sample was 1.71%; after 24 hours maintenance at 100% humidity sample mass increased by 9% due to absorbed water. Pamoic acid is not particularly soluble in either water or acetone. Evaporation would readily remove the acetone. The water would only be grudgingly removed as an azeotrope with toluene. WO0008016A1 PAROXETINE SALTS Example 32 : Preparation of paroxetine pamoate 1: 1 salt.
A solution of paroxetine base in toluene (5 ml, 2. 10 g) was added to a solution of pamoic acid (2.48 g) in pyridine (40 ml), and the mixture was stirred at ambient temperature for 30 minutes. The solvent was then removed by distillation at reduced pressure, the residual oil diluted with toluene (30 ml) and the solvent again removed by distillation at reduced pressure. This procedure was repeated two more times. The solid product was washed with hot diethyl ether (c. 100 ml x 3) , and filtered under nitrogen to give a pale yellow solid. The product was washed twice more with diethyl ether (2 x 100 n- A), and then with methanol (30 ml), and finally dried under vacuum. Yield = 3.27 g,
IR nujol mull:
Bands at 1636, 1558, 1508, 1459, 1377, 1183, 1036, 830, 722 CM-1.
Example 33 : Preparation of paroxetine pamoate 2:1 salt.
A solution of paroxetine base in toluene (10 ml, 4.2 g) was added to a solution of pamoic acid (2.48 g) in pyridine (40 ml). The mixture was stirred at ambient temperature for 30 minutes. The solvent was then removed by distillation at reduced pressure, the residual oil diluted with toluene (30 ml) and the solvent again removed by distillation at reduced pressure. This procedure was repeated two more times. The solid product was washed with diethyl ether (c. 50 ml), and filtered under nitrogen to give a white solid. This solid was washed twice more with diethyl ether (2 x 10 ml), and then dried under vacuum. Yield 6.7 g.
IR nujol mull:
Bands at 1641, 1461, 13 77, 1181, 1035, 829, 757 cm- 1.
Pamoic acid is soluble in pyridine presumably as a pyridinium salt. It can be recrystallized from dilute aqueous pyridine. It is also soluble in nitrobenzene.
General |
Permalink |
Trackbacks (0)
Simple, Rapid Optimization of a Chemical Process Step
kilomentor | 10 February, 2011 14:53
| Experiment # | Yield | Rank Yield | Rank | (100-Rank)/(N+1) |
| 1 | 41 | 11 | 1 | 7.1 |
| 2 | 50 | 14 | 2 | 14.3 |
| 3 | 11 | 25 | 3 | 21.4 |
| 4 | 14 | 34 | 4 | 28.6 |
| 5 | 80 | 41 | 5 | 35.7 |
| 6 | 34 | 59 | 6 | 42.9 |
| 7 | 84 | 61 | 7 | 50.0 |
| 8 | 25 | 70 | 8 | 57.1 |
| 9 | 85 | 73 | 9 | 64.3 |
| 10 | 73 | 80 | 10 | 71.5 |
| 11 | 70 | 84 | 11 | 78.5 |
| 12 | 61 | 85 | 12 | 86.7 |
| 13 | 88 | 88 | 13 | 92.8 |
General |
Permalink |
Comments (1) |
Trackbacks (0)
Optimization continued
kilomentor | 10 February, 2011 14:50
| Experiment # | Yield | Rank Yield | Rank | (100-Rank)/(N+1) |
| 1 | 30 | 15 | 1 | 7.1 |
| 2 | 63 | 23 | 2 | 14.3 |
| 3 | 65 | 26 | 3 | 21.4 |
| 4 | 26 | 30 | 4 | 28.6 |
| 5 | 37 | 37 | 5 | 35.7 |
| 6 | 15 | 42 | 6 | 42.9 |
| 7 | 50 | 50 | 7 | 50.0 |
| 8 | 23 | 63 | 8 | 57.1 |
| 9 | 75 | 65 | 9 | 64.3 |
| 10 | 94 | 75 | 10 | 71.5 |
| 11 | 42 | 76 | 11 | 78.5 |
| 12 | 91 | 91 | 12 | 86.7 |
| 13 | 78 | 94 | 13 | 92.8 |
| Experiment # | Yield | Rank Yield | Rank | (100-Rank)/(N+1) |
| 1 | 0 | 0 | 1 | 7.1 |
| 2 | 45 | 0 | 2 | 14.3 |
| 3 | 0 | 11 | 3 | 21.4 |
| 4 | 11 | 15 | 4 | 28.6 |
| 5 | 15 | 23 | 5 | 35.7 |
| 6 | 32 | 24 | 6 | 42.9 |
| 7 | 53 | 32 | 7 | 50.0 |
| 8 | 33 | 33 | 8 | 57.1 |
| 9 | 44 | 42 | 9 | 64.3 |
| 10 | 24 | 44 | 10 | 71.5 |
| 11 | 60 | 45 | 11 | 78.5 |
| 12 | 42 | 53 | 12 | 86.7 |
| 13 | 23 | 60 | 13 | 92.8 |
General |
Permalink |
Comments (2) |
Trackbacks (0)
The Kilomentor Approach-More Easily Isolated Intermediates-Fewer Technically Demanding Physical Separations.
kilomentor | 01 February, 2011 09:19
The Kilomentor approach to process development is geared towards the simple, rugged and dependable process step and avoids the technically demanding step, which uses sophisticated and expensive equipment. This keeps with the tenor of the times wherein so much process work is being moved to the developing economies where at least for now labour is plentiful, so long as the work process is rugged.
The Kilomentor approach to every process step is to divide it into the reaction phase, the success of which is measured by:
(i) the assay yield (the quantity of desired product in the reaction mixture as a percentage of theory)
(ii) the isolation yield which assesses the success of the effort to separate a product of practical purity out of the reaction soup (again expressed as % of the theoretical possible
The standard overall yield That is all that is normally reported is the product of these (as fractions) expressed as a percentage.
Because of our historic addiction to the assessment of synthetic elegance by counting reaction steps-with the fewer the better, there has been a bias against isolating intermediates as functional group derivatives and then decomposing them back to the original product intermediate. Because we so often only looked at the overall yield, we could not specifically see that we were leaving good product behind just because we were using less effective isolation means in order to minimize the reaction steps.
At the same time, we often convinced ourselves that even this advantage forming a functional group derivative to improve the isolated yield, would be given back in the step of decomposing the derivative to recover the original functionality. For example, the isolation of a pure ketone intermediate might be done by a painstaking vacuum fractional distillation without adding any additional reaction steps to the’ academic’ step count. The same ketone might readily and quantitatively form a solid oxime or hydrazone. Yet we would make arguments to ourselves that the backwards hydrolysis of clean oxime to ketone could be difficult and inefficient and messy when really actual hydrolysis technology made the cleavage trivial and efficient.
Another bad reason that we prefer challenging physical separations (fractional distillation or chromatography) to chemical derivatization is that the reagents for chemical derivatization would add to the chemical cost, while the challenging separation uses labour, time, and equipment which are not considered in the early, small-scale, chemicals-only costing.
General |
Permalink |
Comments (2) |
Trackbacks (0)
Chemistry related Poetry
kilomentor | 10 January, 2011 13:17
Chemists manipulate atoms and molecules even though we never see them. We do this working with symbolic pictures. Sometimes we almost forget that these representations are only a type of crude model for a part of unfathomably complex mysterious nature. I received the following poem from the late Dann Sargent when he taught with me in the Harvard Summer School. It keeps me always alert to this. I came upon this copy when I was tidying up to move house.
In Broken Images
Robert Graves
He is quick, thinking in clear images;
I am slow, thinking in broken images.
He becomes dull, trusting to his clear images;
I become sharp, mistrusting my broken images.
Trusting his images, he assumes their relevance;
Mistrusting my images, I question their relevance.
Assuming their relevance, he assumes the fact;
Questioning their relevance, I question the fact.
When the fact fails him, he questions his senses;
When the fact fails me, I approve my senses.
He continues quick and dull in his clear images;
I continue slow and sharp in my broken images.
He in a new confusion of his understanding;
I in a new understanding of my confusion.
General |
Permalink |
Add comment (SPAM IS DELETED INSTANTLY) |
Trackbacks (0)
Reducing Reaction Time by Adjusting Concentration during the Process in Chemical Process Development
kilomentor | 29 December, 2010 05:48
Most synthetic reactions are second order or higher kinetically. Once initiated, they proceed most rapidly in the initial stage and then slow down as the starting materials are consumed and their concentration declines in a constant liquid reactor volume. As a consequence, the larger part of reaction time is spent waiting for a smaller part of the reacting to finish.
When such a reaction is exothermic, the largest part of the exotherm occurs in this early stage. It is for this reason that process chemists strenuously avoid having the full stoichiometric quantities of all the reactants together and then initiating the reaction (say by heating). This is a recipe for a disastrous runaway exothermic event. Instead, in the preferred approach, one reactant is added gradually to a mixture of the other essential chemicals at the reaction temperature. Operating this way, any unwanted exothermicity in excess of what can be balanced by cooling, can be choked off by slowing or stopping the addition.
The question being here considered is whether after the faster part of the reaction has passed can anything be done to accelerate the reaction in the later stages when the concentration of the rate controlling species have fallen? If the reaction is being conducted at the reflux temperature of a solvent, the reaction can in principle be accelerated by distilling out of the reaction part of the reaction solvent. Because the reaction temperature is the boiling temperature of the solvent, this distillation removes only solvent and does not change in any way the reaction conditions. Since removing solvent increases the concentration of starting materials, the rate of their consumption will increase and the point of effective disappearance of starting materials will arrive more quickly. If the volume for a bimolecular reaction is reduce in half, the rate of reaction will be increased by a factor of four.
There is a limit to how low the volume can be taken in a standard reactor. The volume should not be reduced below that which can be effectively stirred (the minimum stirrable volume).
Another advantage of concentrating the reaction mixture is that the volume at the point of maximum volume in the process may be lowered. This will allow a higher rate of production to be obtained. If your total volume at the point of maximum volume can be reduced in half (for the sake of simplicity of explanation) you would only need half as many repeats of the process step to transform the same amount of starting material.
A potential flaw with such a concentration procedure occurs if some essential element of the process is actually volatile with the solvent distillate and is removed. This would slow down or stop the reaction. Although some reaction ingredient may not be blown out of a reaction mixture distilling in the lab, distilling in the plant can have substantially different characteristics and one needs to be aware of the possible loss of even quite non-volatile materials in an aerosol.
General |
Permalink |
Comments (1) |
Trackbacks (0)
Dichloroacetic Acid Solvent
kilomentor | 24 December, 2010 07:30
Dichloroacetic acid is miscible with both water and organic solvents and is an acid that can be added into refluxing organic solvent using a continuous dilution head so that a local excess of strong acid is never experienced and the mixture remains completely anhydrous? An 80:20 v/v dichloroacetic acid/methylene chloride mixture can dissolve many otherwise poorly soluble materials, including polymers. Nevertheless it is hardly ever considered as a reaction solvent or for neutralizations or salt formation isolations.
Dichloroacetic acid is much cheaper than trifluoroacetic acid although for some applications the greater volatility of the latter makes it more advantageous.
Dichloroacetic acid is a liquid at ambient temperature and typical reaction temperatures. The bp is 193-194ºC and the melting point of the higher melting of two forms is +9.7ºC. The pKa is 1.48, meaning that one half of the molecules are ionized at a pH of 1.48. The density is 1.563. This is 12.12 molar acid.
Dichloroacetates are pharmaceutically acceptable salts. Dichloroacetic acid therefore is acceptable at low levels as an impurity in pharmaceutical products and so is attractive as a reagent in drug synthesis. The Safety; Toxicology data are:
Acid: LD50 rat oral, 2.82 g/kg; rabbit skin, 510 μl/kg; TDL0; rat oral, 3.195 g/kg/90 days; mouse oral, 7.1g/kg/10 weeks.
Sodium salt: LD50: rat oral 5.281 g/kg; mouse, 4.845 g/kg.; i.p.,; TDLo rat oral, 2.45 g/kg./7 weeks, 30.425 g/kg/12 weeks; dog oral, 4.55 g/kg/13 weeks.
Because dichloroacetic acid is a liquid, it can be present in excess with precipitating dichloracetate salts to contribute to their insolubility by the common ion effect.
General |
Permalink |
Add comment (SPAM IS DELETED INSTANTLY) |
Trackbacks (0)
The Potential Practical Use of Cyclodextrins in aChemical Process Separation at Scale.
kilomentor | 01 December, 2010 11:45
I had occasion to read a review, Past, Present and Future of Cyclodextrin Research, Jozsef Szejtii, Pure Appl. Chem. 76(10) 1825-1845 (2004).
I did not realize how practical use of cyclodextrins for separation might be. Beta cyclodextrin only costs $2-5 dollars per kilogram. Beta cyclodextrin has a molecular weight of 1,134 and if it forms a 1:1 complex with a drug substance mw. about 500, a kilogram would complex about 500 gm of API. That is the cyclodextrin to complex a kilogram of such API would be between 4 and 10 dollars. Nor did I realize that cyclodextrin can form very strong complexes which will take an insoluble drug into aqueous solution. Beta cyclodextrin forms a strong complex with cholesterol which then crystallizes.
A cyclodextrin complex may be quite stable while in water solution but when it is dried completely the stability of the complex can completely disappear because it is the removal of the hydrophobic interactions in bulk water that accounts for the stability of the complex. When the water is removed the basis for the stabilization is removed and one gets just an intimate powder mixture.
One would anticipate therefore that slurrying a mixture of two essentially water insoluble compounds, one of which forms a stabilized inclusion complex and another that does not, would partition the former into aqueous solution and leave the latter undissolved.
Solvent switching to an organic solvent should cause the organict hat had been complexed to move into the organic layer and a complex of the new organic solvent (if that solvent forms a complex) and cyclodextrin to precipitate.
General |
Permalink |
Comments (3) |
Trackbacks (0)
Organocatalytic Relay Cascade Reactions
kilomentor | 25 November, 2010 11:15
The other day, I was reading journals in the University of Toronto chemistry library. A hot topic for synthetic chemists is asymmetric organocatalytic cascade/domino reactions. An example of such a publication is Angew. Chem. Int. Ed. 2009, 48, 9834-9838 which in turn cites many pieces of similar prior art.
It is indisputable that these reactions are promising approaches for rapid assembly of complex asymmetric structures but they do not pretend to be, and they should not be mistaken to be, methods that could be used for scaled up processes. Looking at the cited paper, although atom economy is identified as an advantage of the methodology, two catalysts are used, each in 15% molar ratio. These particular catalysts had molecular weights of 582 and 519. No procedure for recovery and reuse of these ‘catalysts’ in reusable purities was reported. The atom economy when these losses are taken into account is not so attractive.
Another point is that for a process that is run at scale it is the isolated yield of pure product that goes into calculating the overall yield but no method is provided for separating the enantioselectively and diastereoselectively formed principal products from the rest of the reaction soup.
This is not to say that what has been achieved is not admirable but process chemist readers should not make the mistake of thinking for one moment that such technology is ready for process adoption.
General |
Permalink |
Comments (1) |
Trackbacks (0)
Identifying Chemical Process Stopping Points for Working in the Kilolab or Pilot Plant
kilomentor | 25 November, 2010 11:01
It is not as if there is no planning in the laboratory. If a synthetic lab procedure is so long that the reaction and work-up cannot be completed in a single day, chemists can use their experience to extrapolate from similar procedures and guess at what points manipulations can be stopped and under what conditions intermediate solutions or crude solids can be stored without damage. Occasionally there are misjudgments and surprises and a product will be prepared in lower than expected yield or poorer purity. but then even in the worst situation what is lost is no more than a couple of man-days of labor and the price of the starting materials consumed. Also, in the laboratory because the capacities of refrigerators, freezers and evaporators are so much greater than the quantities of material being transformed, there are do-able fixes for the situation where a stoppage is forced at almost any stage.
There is no room for such risk taking on-scale. For advanced intermediates that are themselves the product of a series of sequential steps, one misstep can be economically disabling. The more points in the process that have been verified as safe to stop, by actual test results, the more confidently the process team can be. Moreover, to be a safe stopping point it must be proven safe not just for the quantity and quality of the product but also for the protection of the processing equipment.
As a general rule once a reaction has been initiated the kinetics must be allowed to run undisturbed to the proper end-point according to the batch sheet. The dynamic transformations cannot be expected to respond to any speed up, slow down without some quantity or quality deviation. After the end point condition has been reached and the reaction quenched then the mixture is likely more stable and various stopping points during the work-up can be tested by holding portions of a process mixture for given periods under controlled conditions and examining the mixtures and isolating the product to see whether an unacceptable deviation has occurred or not.
It is more difficult to demonstrate a good stopping point where the mixture in the process equipment is heterogeneous. The difficulty is that it is difficult to take a representative sample out of a heterogeneous mixture and it is difficult to analyze that sample to show that no change affecting quantity, quality or the protection of the reactor has occurred. Since one cannot easily take a precise fraction of a heterogeneous mixture, working up that fraction after a pause will not accurately tell you whether the yield would have been different.
Finally to fairly test the stability of an aliquot at a proposed stopping point the aliquot must be left in contact with a sample of the reactor material. In my experience this is rarely ever done. At the very least it should be kept in mind where an aliquot might be corrosive to the reactor material.
General |
Permalink |
Add comment (SPAM IS DELETED INSTANTLY) |
Trackbacks (0)
Navigation
Menu
Search
Calendar
| « | May 2013 | » | ||||
|---|---|---|---|---|---|---|
| Mo | Tu | We | Th | Fr | Sa | Su |
| 1 | 2 | 3 | 4 | 5 | ||
| 6 | 7 | 8 | 9 | 10 | 11 | 12 |
| 13 | 14 | 15 | 16 | 17 | 18 | 19 |
| 20 | 21 | 22 | 23 | 24 | 25 | 26 |
| 27 | 28 | 29 | 30 | 31 | ||
Recently...
- Continuous Chemical Flow Reactors that Scale so Well are not New
- Using the Bisulfite Adduct for Purification at Scale: Practical Purity and Purging of Impurities by the Process Steps
- The Potential Use of Acetic Acid and Acetic Anhydride for performing Solvent Switches during Work-Ups
- A Question about Genotoxic Impurities
- Improving the Purity of Product from a Good-Enough Process
- The Problem of Oiling Out in Chemical Process Development
- Christmas and New Years Greetings
- Selective Silyl Group Protection: A Possible Preparation for Scalable Extractive Separations using Acetonitrile and Hexane Phases
- Friedel Craft Work Up of Aluminum Chloride catalyzed Reaction At Scale
- Removal of Hydroxyl Impurities from a Solid Product at Scale
- Separating Sulphur-containing from Sulphur-free Compounds both in the Lab and At Scale
- The Relationship between the Risk of Catastrophic Failure and the Size of the Scale Up Steps in Chemical Process Development
- The Advantages of Transfer Hydrogenation for Chemical Process Scale Up
- Distillation with Reduced Pressure using Carbon Dioxide Atmosphere to Create Enhanced Vacuum
- Reactions on Immobilized Enzymes can Telescope Process Steps, Eliminate Isolations and Reduce Costs at All Stages of Pharmaceutical Development