Karl Fischer Reagent & Sample Prep

[Liza]

Hello:


Please help w/ the following questions regarding KF:


(1) Hydranal Composite 1, 2, and 5 differ in their "strength" of titration (e.g. 1 mg H2O/ml for Composite 1; 2 mg H2O/ml for Composite 2). Since a smaller amount of the "stronger" reagent (e.g. Composite 5) would be required to titrate the same amount of H2O as compared with a "weaker" reagent, why not just use Composite 5 all the time?


(2) For ketones & aldehydes, it's best to use Composite 5K. For all others, can we use Composite 1, 2, and 5? In other words, are there any special cases I need to watch out for?


(3) For testing uncoated and coated pellet samples, is grinding needed? I'm not even sure if these are soluble in CH3OH...


Thanks. I appreciate it.


Liz

[Bruce Hamilton]

Hi Liz,

Firstly, There is an excellent free booklet by Eugen Scholz, entitled Hydranal Manual, and also there is now a Hydranal Multimedia Guide available free from Sigma Aldrich. I highly recommend getting one or both of the above.

In answer to your questions.

Question 1.

You chose the Composite 5 reagent according to the amount of water you wish to titrate, and considering the volume/accuracy of your titration device ( often an automatic burette ). Some manufacturers used to provide recommendations, relating reagent to system, so check the instrument manual as well.

The quantity of water in the added sample can also be limited by either sample size or sample solubility in your titration solvent, so you may have to use the 1 or 2 mg/ml reagents, because you only have 1 or 2 mg water present. If possible, ensure the sample dissolves completely and quickly, otherwise titration times are a real drag - 10 - 20 mins or more.

Generally, the smaller the quantity of water you wish to titrate, the more precautions you have to take to prevent water contamination, so the 5 mg/ml, and 2 mg/ml are preferred by lazy analysts like me. You can calibrate the 5 mg/ml using water in a syringe ( eg 1 - 5 ul syringe ).

Question 2

Yes, some organic acids and bases have to be neutralised prior to the titration, KF solutions are weakly acidic and can also react with inorganic oxides, some samples can cause precipitation, you have to have sufficient alcohol present to ensure a rapid reaction, etc. etc..

The above booklet identifies many of the common issues, but you should check new methods, with the easiest way being to add a known quantity of water to repeat samples and check the recovery. Some items don't release all the water unless heated, and special ovens are added to instruments.

Question 3

Consider the answer to question 1. How much water is in your pellets, and how soluble is your pellet in the selected titration solvent?. Assuming that you can titrate all the water in one pellet ( say maximum of 25 mg for an instrument with 10 ml burette and 5 mg/ml titration reagent ), do you want to know the individual pellet water content, or do you want to grind up 20 and obtain an average. You can titrate more water, but the titration often takes so long that you could watch a football match before the result appears.

I prefer to grind samples, and take a weighed sub-sample, unless they are really, really, really , soluble in the titration solvent, and the ground material is also not strongly hygroscopic. Insoluble material delays the endpoint, and is very likely to decrease the accuracy.

I hope that helps,

Bruce Hamilton

[Consumer Products Guy]

Bruce has provided some pretty good responses. For intact pellet or tablet samples, you may want to consider a specialized unit. Brinkmann has an option of a Polytron homogenizer in its the titration beaker to pulverize stuff like this in situ. Orion also had (has?) a Turbo2 KF titrator to pulverize stuff, but I don't know if that is still available.

[gbalock]

Liz,
In response to Q2, Comp 5K is for aldehydes and ketones. The regular Comp 5 replaces the alcohol with another solvent. In the KF environment the alcohols react with carbonyls to from acetals and ketals and one of the end products is water. So, if you are trying to measure water, this won't work.

[Markus Laeubli, Metrohm]

Dear Liza

in respect to question 3) there are serveral approaches possible depending on the result you are expecting. You may want to determine surface water, bound water, extractable water, free water, total water ....
Metrohm (represented by Brinkmann in USA and Canada) has a broad range of KF tools.

[Liza]

Hello all:


Thanks for your help with the previously posted questions. I still need help with question no. 2 though, regarding ketones & aldehydes.


Question: Most of the pharmaceutical actives have multiple function groups. The product I'm analyzing has a ketone group out of many functional groups, will it react with methanol to form water?


Liza

[Bruce Hamilton]

Hello all:

Question: Most of the pharmaceutical actives have multiple function groups. The product I'm analyzing has a ketone group out of many functional groups, will it react with methanol to form water?
Liza

Yes. However the rate can be quite slow. In general, as chain length of aliphatic ketones increases, the reactivity decreases, and aromatic ketones are less reactive than aliphatic ketones. It's often possible to quickly titrate the water before the ketone is affected.

If you are concerned, purchase some of the reagent and solvent designed for aldehydes and ketones and test some samples using both systems. The alddyde/ketone reagent can be used for other samples as well, but it's slower.

There is also the possibility of using just the aldehyde/ketone solvent component with the conventional titrant for samples with ketone functionality ( aldehydes are too reactive ).

Once again, it's worth getting some of the free literature, it's very good and also discusses such issues.

Bruce Hamilton

[DR]

If a moisture balance is available, it's also a good idea to compare KF values with LOD values for "new" compounds, although LOD is nonspecific (meaning that if there is any solvent in your sample, LOD will show that along with water - knowing where your sample came from is good).