How much sample of volcanic rock is required to undertake a K-Ar date measurement, and does it matter if the rock contains vesicles? Are there tricks of the trade to get a good sample?
I run an argon lab which does also K-Ar measurements. The sample amount depends on the age because you need enough signal strength to measure the radiogenic argon component precisely. Young rocks have very low 40Ar* and to get enough volts on the detector you need more sample. Having the incorrect amount is akin to trying to measure micrometers with a yardstick.
For K measurements 30 mg in our lab is routine. For the argon measurement we usually use about 2 mg; for very young rocks up to 200mg for rocks with ca. 2% K. We can date volcanic rocks of less than 20000 years using this technique and have dated historic eruptions successfully in our lab.
Usually customers can say if the rocks are very young, so I usually prepare the sample amount accordingly. I had a couple of completely blind samples where I had no idea of the age of the sample a priori, I ran a test with a ca 4 mg which yielded almost no gas, so I re-ran with more material.
Regarding sampling technique itself- look for minerals with more potassium, such as muscovite, biotite, amphiboles, or more felsic rocks. If the rocks are obviously weathered, they are not suitable.
I also have to say that Ar/Ar is a more accurate tool for most rocks and minerals, and then just a few mg is enough if the samples are Mesozoic or older. If they are very young, higher sample amounts are required as with the K-Ar technique.