Lecturer's Notes
Author: Case History, Dr. P. J. Crean;
First indented text represents questions for the class. The
second level of indented text consists of notes for the lecturer.
The overheads for the class (the un-indented text) can be
downloaded in printable form. Overhead
Transparencies (PDF format; requires Acrobat Reader.
A polyimide (PI) is made by reacting an aromatic dianhydride
(DAN) with an aromatic diamino ether (DA).
DAN + DA ----> PI
(Structures)
This is a process for a new polymer which has novel and useful
properties. The polymer was commercialized eight months ago, and
is very successful. The process, however, suffers from several
problems, including variability in the quality of the monomers,
particularly the DA.
The DA process included a crystallization from an alcohol solvent
using a decolorizing carbon. Last month, the supplier of the
carbon decolorizing agent went out of business, and an alternate
source was found. Polymer made since the changeover has not had
consistently acceptable properties.
The problem was presented to you via a written memo from the
technical director of the DA plant. She asked you to find the
impurity in the new, ineffective decolorizing carbon. A sample of
the offending carbon and a retained sample of the previous supply
was attached.
You are mass spectroscopist who has been with the company for 3
years. Your familiarity with polymer production is minimal. The
memo was directed to you because the Technical Director read an
article in ChemTech a few months ago discussing the great power
of newer mass spectrometry techniques.
What is your response to the memo?
Do you know enough to analyze the problem? What should be your first step?
You have received a sample in the mail with very little information. If you are going to help solve the problem, you must understand it.
The telephone is an analytical tool. You should call the Technical Director and ask for more background concerning the nature of the problem.
The director tells you that the production facility has been shut down, 200 staff members are furloughed, and the future of a new profitable business is in jeopardy.
Assume your technician had a three week backlog of samples to run, and you were already working 10 hours/day. Discuss the level of importance you should apply to this request.
One must be able to differentiate between levels of "top priority". A useful criterion is to ask if the company is losing money because of the problem. A plant shut down, which affects the company's income, the well being of the plants employees, and the plans, and possibly financial well being of customers depending on the product, is cause to drop everything and spend long hours to solve the problem.
What are the potential pitfalls in the chain of communication followed?
Was the technical director's reason for requesting mass spectrometry valid?
Are you confident that her analysis of the problem is correct?
What issues should you be alert for?Did you consider calling a scientist or engineer closer to the problem than the Director? Analyze the problem before the sample! Higher levels of management may be less familiar with the details of a problem than the those working on it directly.
The client is usually not familiar with analytical methadology. The requested method is often not the appropriate one when a poorly defined production problem occurs. You must learn enough about the problem to analyze it yourself, and select an approach to a solution.
Since mass spectrometry is your specialty would you be correct to apply it immediately, hoping for a quick answer? Resist the temptation to use your favorite method. Consider the cost of your time in answering. A scientists time is very expensive, especially when benefits and overhead (the laboratory, technician, an allocated share of the cost of management, etc.) are added in. Your time at this stage is better spend learning enough to select the most appropriate technique.
You talk to the engineer running the process, and are assured
that the only change has been the source of the carbon, and
analyze the impurities in the carbon. Extraction of all organic
material and mass spec analysis demonstrates no differences
between the carbons.
The Technical Director of the polymerization plant has been
informed by the staff of the DA plant that the DA they are making
frequently fails a color test. (In a standardized procedure, the
transmission in solution of the DA made since the change in
carbon source is as much as 6% off that of a pure white standard.
Any more than a 1% deviation results in the failure of a
"polymerization test" run at the polymer plant.
The polymerization facility is 200 miles from the location of the
monomer plant.
The polymerization test used by both plants is a controlled lab
polymerization which is considered successful if a goal viscosity
is reached under the test conditions without the formation of
gel. The Technical Director tells you that the color test is a
reliable predictor of polymerization success.
Can you therefor assume that it is so?
Do you know the staff of the plant well enough to know who might give you another viewpoint of the problem? Do not assume that any one individual has all the facts.
How could you have developed the necessary contacts?
Building your own network of contacts throughout the organization is a long term operation. You should start as soon as you get on the job. Ideally your lab has formalized procedures for maintaining contact between your lab an all potential clients in the company.
You call the DA plant, speak to a chemical engineer who oversees
the process, and request samples of "white" and
defective DA.
You separate the colored component from the 'bad' DA by
chromatography and determine by IR that it is N,N-dimethyl
paraphenylene diamine.
You are not a polymer chemist, and are uncertain of the meaning of the result. What do you do to take full advantage of the talents of the people around you.
The most effective approach to any problem is to apply all the talents of a team. You must make it your job to learn the capabilities of your co-workers, so that you know who to draft into the team to supply the competence the problem demands. Get to know everyone around when you start a new job. They will be happy to tell you what the they are doing in their field of expertise.
A discussion with a polymer chemist down the hall convinces
you that there is no obvious route by which the color former can
lead to gel formation in the polymerization test. It, however,
can cause chain termination. The concentration of this
contaminant proved unrelated to the polymerization test.
Further work demonstrates a second impurity in all samples of DA.
Mass spectrometry identifies it as a trifunctional derivative of
DA. . No pure C24 was available for calibration, but an IR
absorption band was ratioed against an internal band to form a
specification for usable DA. However, there is a poor correlation
between the trifunctional impurity and the color of the sample.
You convince the DA plant to run the polymerization test with
controlled amounts of the dimer in spite of the lack of a
correlation with color. They demonstrate that this trifunctional
material results in failure of the polymerization test.
Why? You may need help from your polymer chemist colleague.
What other variables could be looked for?Is it sufficient to say that the concentration of the delaterious component should be kept as low as possible. Consider the cost of the process changes needed for minimal concentration. A goal range is preferable and will ensure a more consistent process.
The problem seems not to be related to the color as had been assumed for years.
How does this kind of problem develop? Was the analytical lab an active partner in developing the plant tests?
An effective analytical department will help plants develop suitable analyses so that such problems do not arise. This requires ongoing communication between Analytical and clients.
Small amounts of trifunctional impurity, C24, such as are
found in the satisfactory DA, do not interfere with the
polymerization.
You call the Technical Director at the polymer plant and inform
her that if the C24 level is kept low, the polymerization test
would be passed.
Will she be satisfied with this answer? What was the real question first asked?
What people ask for is not always what they have in mind. Plant people rarely want the analytical result as a goal. They really want the plant to keep running.
She points out that her question wasn't really " what was
causing the failure of the test" but "how do we ensure
consistent output from the plant."
You return to a review of the chemistry. Dimer formation clearly
depends on some critical reaction parameter which changes when
the new source of carbon is used. You suspect a catalytic effect,
re-investigate the carbon. No differences were found in the types
of inorganic impurities. It had already been shown that there
were no differences in extractable organic impurities. It was
eventually determined that the original carbon was unusually
basic. You do not do a mechanistic study, but assume that the
basicity suppresses C24 formation.
Working with the DA plant you work out conditions for modifying
the carbon and making DA which passes the polymerization test.
Good, you remember that you job is to be a member of the problem solving team, not necessarily an analytical researcher.
Knowledge that the new carbon was not itself introducing
deleterious materials allowed other process changes to duplicate
the original polymerization conditions. The polymerization test
could be consistently passed using DA made under the new
conditions provided C24 was maintained below a certain limit.
You proudly write a 15 page description of the entire 'detective
story', giving the plant the details of your thinking and
describing the theory of the analytical methods. Your memo leads
up in an orderly fashion to your conclusions.
Was your method of reporting results well suited to a busy plant engineer or plant technical director?
Consider putting your conclusions at the top of a one page memo.
Your relationship to the plant is aggravated when you are
informed the next day that the DA plant finds your answer
incorrect. They have run the polymerization test with DA that has
less than the critical amount of C24 monomer, and it still
failed. The technical director tells you they intend to bring in
an outside consultant, rather than trusting to Analytical.
You have assumed that the plant is dependent on your lab. Is that a safe assumption? If this work is outsourced, how safe is your job?
Have you been in touch with both plants throughout the work? Are you confident they are communicating well between themselves? What can go wrong if all players do not have the same set of facts?
The consistent results fell apart even worse when attempted by the polymerization plant (a few hundred miles away). The client (polymerization plant) reasonably wanted the monomers to pass it successfully before a plant run was made.
Why would the plant be unwilling to try a plant run on the basis of your conclusions?
Consider the scale of the "experiment" and the expense of cleaning out a batch of insoluble gel.
The polymerization lab immediately informed analytical that they had solved nothing because the DA make under the new conditions still failed the lab test. It must be remembered that the test had been developed by the client's lab, and the plant had depended on it for several years.
In correcting other's errors, remember you are part of a team. Making a plant chemist look bad may make you look good (for the moment) but will close off communication. Good supporting data is needed
The test had proven not to be robust when Analytical was applying
it, and they suspected that it had some deficiencies. Rather than
discuss the problem directly with the client, Analytical reviewed
all aspects of the polymerization test.
Consider all parts of the reaction. Solvent, temp. etc. Reagents may not be precisely what the users think them to be. Storage conditions can cause reagents to change with time.
The solvent used was dry dimethylacetamide (DMAC). The known
sensitivity of polymerization reactions to trace moisture led to
analysis of the water content of the solvent. The DMAC was dried
over molecular sieve that was efficient but slow in removing
trace moisture. (Operators, to save time, were cutting back on
the contact time with the drying agent.) However, good
reproducibility could be obtained at a specified water level. The
conclusion was that excess water would mitigate the cross linking
effect of the trifunctional C24. Holding the water constant made
the C24 the only variable in the 'end use' test.
With this knowledge Analytical could explain to the client what
was causing the problem. The client then had only to correlate
the result of the 'end use' polymerization test with the actual
plant polymerization conditions, and to determine the C24 limits
which could be tolerated.
Discuss the apparently round-about route to a solution. Could it have been improved?
In retrospect it all seems obvious and makes sense, It is tempting to be critical of those involved for not being quicker to put it all together. No doubt many quarterbacks become aware on Monday mornings, after viewing game films, of opportunities and options missed due to the charge of a 300+ pound defensive lineman.
It is important to note the role of analyzing all aspects of the problem in supporting the production operation. By gaining an understanding of the chemistry of the process and the details of the polymerization test, a solution was found which saved the business.
The polymerization test has been shown to be sensitive to many parameters. What would you recommend as a quality control test?
It is now 5 years later, and the new polymer is proving to be
a big seller. The monomer plant is too small to satisfy the
demand and an external source of monomer is found. The second
company that manufactures the monomer also has a limited
capacity. As the polymer finds more uses, your company increases
its order from the outside supplier. The first shipment at the
larger quantity results in the same problem the plant had
experienced years ago.
Long discussions with the supplier bring out the fact that to
meet the demand, the supplier raised the temperature of a drying
process to increase productivity.
What might have happened?
The increased temperature could only have improved the drying step.
Consider that the lab test required that the water content be held constant.
Successful processing turns out to require that the water level in the solvent be precisely on specification, neither above or below. This case illustrates the concept of "on spec" production. Optimum is not necessarily lowest.
