Look and see if this is something that should be addressed using a different approach.
For example; is this an application that could justify online analysers?
At a refinery in Siberia the client had two reasons for using a continuous online analyser. First because it gave them better quality control and reduced off spec products. Second, health and safety. Sending people out in 30degree below to gather samples of hot heavy fuel oil and residuum was considered unsafe.
In this case there was only a couple of key control parameters that needed measuring. If there is a more detailed multi-parameter analysis required then consider if an automatic sampler is justified.
I have seen some pretty shoddy sample collecting and handling in various refineries, oilfields and in a variety of countries (not necessarily a reflection on the staff who are usually operators and not engineers, but also because they often take samples for control where there are no online instruments. It doesn't pay to be fussy because by the time the sample has been collected and analysed the process has moved on and the results are an hour or so out of date.
In one refinery they collected hot heavy fuel oil in open top 2 litre plastic Coke bottles with the tops cut off and a wire carrying handle added. Fine except that by the time the sample had been walked the 500meters to the lab a fair amount of the volatiles would have departed.
On a site in the UK they were sampling the product stream from the de-asphalter. They collected it in a bottle with a screw cap so no loss of volatiles. However, when they got it to the lab, even though it had obviously stratified in the bottle, the just drew of a sample from the top and analysed that.
If you are having problems with spills, the chances are your whole sample collection handling and processing needs looking at.
Most probably the procedure is old and whiskery and no one has given it much thought.
I would suggest talking with companies that produce sample collection vessels that will not cause spillage. Try Jiskoot for a start (http://www.jiskoot.com/
) and a few others.
Chances are you will, with the help of the manufacturer who has been there and done that on multiple sites, put together a justification on environmental and health and safety grounds for some sort of modern purpose designed sampling system.
The clincher ought to be that the system will probably save money and help deliver a better product.
So, if you are taking samples of tar is this for a cone and plate viscosity measurement in the lab? Or is it something more sophisticated?