I thought this was interesting ...
The following replies came from Giacinto Scoles, Roman Schmied and Kevin Lehmann. Many thanks to them for their helpful suggestion of a needle valve that can be made inexpensively. There are links to pictures below taken by Roman of one of their needle valves
The best needle valve I know is made by squeezing a hardened copper capillary ( 1 mm id and 3 mm od ) between a brass plate and a brass wedge. The capillary needs to be well cleaned before use. This valve controls the flow down to very low values and can be changed in setting many times before the capillary gets "tired". It is the best valve I know vis a vis contamination and dead volume but it cannot, of course, be closed completely and needs a secondary valve set up to empty the upstream volume separately. It is, I believe , described in a book on vacuum thechniques edited by John Yates. G.S.
What we use to control very small fluxes is a "pinch valve": a piece of copper tubing (1/8 inch diameter) squeezed between two brass wedges. Two screws hold the wedges together, and the amount of tension on these screws, through the elasticity of the copper tubing, determines the flow. (notice that the elasticity of brand new copper is quite low; it needs to be bent a few times to work-harden).
Very small flows can be achieved thus, with surprising stability. One of the main advantages is that there is no contamination at all. Also, it is practically for free.
We use this to control the flow of various gases into pickup cells for helium nanodroplet isolation spectroscopy. Roman
Greetings: We make our own "pinch values" to produce such low flows. They are not "reproducible" but very stable and cost almost nothing. They involve a controlled crushing of a small Copper tube. Kevin
The procedure to clean the tube is as follows ( if I remeber correctly of course! ): First you flow nitric acid through the tube and then wash, first with water and then with alcohol to make it dry. The hardening of the tube ( also necessary to give it some elasticity ) is done by heating it up and then quenching it, dumping it, while still hot, in clean cold water. Sincerely I do not remember if the quenching was done before or after the cleaning. Probably it can be done either way.
When I used first in the sixties we called it the KISTEMAKER valve, from the name of Jaap Kistemaker who was ( for many years ) the director of the FOM lab in Amsterdam whom we belived to be the inventor. I was recently told that Prof. Kistemaker , now approx 90 years old, still goes regularly to the laboratory, so , if you do not feel that this thing is being blown a bit out of proportions , you could try writing to him for a proper quotation of the real source. By the way, one of the main reasons of existance of this valve is that does not allow for fractionation ( isotopic or otherwise ) to arise because of the ( molecular ) flow out of the narrow orifice like a normal needle valve would instead allow for. This is because the flow in the capillary just before the orifice does not allow the back diffusion of any concentration gradient that may form because of the fractionation. Giacinto.