The electronics involved with the coulometric titration is extremely precise and has little influence on the overall precision of the measurement.  However, the two major issues in dealing with improving precision in coulometric Karl-Fischer titrations are background rates and precision of sample introduction.

The background rate has a much bigger influence when the total number of micrograms of water titrated are very low, (5-40ug), whereas the precision of sample introduction influences the precision directly regardless of the amount of micrograms of water titrated.

Introduction: Improving the Precision of Sample

First off is the nature of the bulk sample.  It should be uniform with regards to the container from which it was taken.  All samples should be in glass containers with tightly sealed caps.  Have as little airspace above the sample to minimize the addition of water from the air to the sample.  Gently mix the sample before withdrawing the sample with the syringe to insure uniformity.

For liquid samples, a syringe should be used with a needle that is 2.5” to 3.5” long.   The needle should be long enough to reach down into the vessel solution when injecting the sample.  This allows the final drop at the end of the needle to be washed into the vessel solution.  If the needle does not touch the vessel solution, then the final drop may cling to the end of the tip and be wiped off by the underside of the septum as the needle is withdrawn from the septum.  This results in more variability is terms of how much sample has been added to the solution as well as introducing liquid on the underside of the septum that contains some water and will ultimately add to the background rate.  For a 1ml or 2ml injection, this is not quite as critical since the volume of the drop left on the needle is very small percentage of the 1 or 2ml sample volume.  However, for a 0.100ml injection or less it could constitute a significant percent variability.

Let’s first address using the weight method (weighing the syringe after filling with sample and after injection).  For all samples, be sure to wipe the needle tip with a tissue to eliminate any fluid that on the outside of the needle.  In this manner you will eliminate any fluid that would have been part of the sample weight, but would not have made it into the vessel because it would have been wiped off by the septum as the sample is injected.  For volatile samples, use the rubber squares to cap the end of the syringe needle before weighing to minimize sample evaporation and water introduction.  Remember to put on the rubber square again after injection and before the tare weighing.  Just as in the previous paragraph these tips are more important for low volume injections (<0.100ml).

For improving precision with sample addition using the volumetric method, one needs a very precise method of obtaining a consistent volume.  Using the markings on the syringe barrel and sighting the end of the plunger by eye one can sometimes achieve adequate precision (2-5%).  If you are using this method, it is best to utilize at least 75%-100% of the total volume of the syringe for better precision.  However, this method can be dramatically improved by using a Chaney Adapter modified syringe that has a hard stop to which the plunger is positioned before sample addition.  Another alternative is the digital syringe adapter that shows precisely the volume that is being delivered.  In addition to being very precise, sample addition using the Chaney adapter or digital syringe is very fast compared to the weight method although one must know the density or specific gravity of the sample.

Improving the Precision via Background Reduction

Background rates influence the precision because it is subtracted from the total number of micrograms calculated from the titration of the sample.  In the following example let us suppose that the sample being measured has 10ug of water in it.  Let us suppose that the background rate is 18ug/minute (0.3ug/sec) and that the time for titration of the sample is 30 seconds.  In titrating the sample the instrument would titrate the 10ug in the sample and 9ug in background rate (for ½ minute) thus giving a total of 19ug.  For the final answer, the 9ug would be subtracted giving 10ug or the correct answer.  The problem with a high background for low moisture samples is that the background represents a significant percentage of the sample and any variation in the background during the titration would alter what should be subtracted from the final value.

In the previous example let’s suppose that the background value was 18ug/minute just before the titration, but change to 12ug/minute as the titration started (a 33% change).  The total would now be 10ug + 6ug or 16ug, but the background value that is subtracted would be 9ug since the background value is determined in the 15 seconds before the titration starts.  This would give a final answer of 7ug that would be 30% lower than it should be.  The same example but with a 100ug of water sample rather than 10ug would show a lowering of the value by only 6% rather than 30%.  Similarly the 1000ug sample would show less than a percent difference.

What happens now if we are able to lower our background rate significantly for the sample containing 10ug of water?  If we run through a scenario where the background rate has been brought down to 5ug/minute (0.08ug/sec), then we would have 2.5ug background being added.  If the background rate changed to 33% lower during titration as in the initial example, then it would yield 1.65ug of background during the 30 seconds of titration.  The total would then be 11.65ug and subtracting the 2.5ug from the initial background rate would give 9.15ug or only 8.5% lower than the 10ug that it should have been.

In summary, if you are measuring low-end samples and want to improve the precision, you must lower the background rate as much as you can.  If you are measuring higher water content samples, then the background is not so much an issue.

Methods to Lower Background Rate:

• Use fresh reagents
• Make sure that the level of the solution in the vessel is higher than the level of the solution in the generator (very important)
• Make sure all the fittings on the glassware are tight and greased (if necessary)
• Have fresh silica gel in the drying tube.
• Make sure the septum does not have open holes where moisture can enter.
• Eliminate moisture above the vessel solution by doing the following.  Turn the instrument on to standby, lift the vessel slightly and swirl the vessel solution so that the solution washes the walls of the vessel and captures any moisture.  Turn the instrument off standby to titrate to set point.  Repeat this for a number of times until the background rate decreases to an acceptable value.