I totally agree, 4-points method is absolutely required, otherwise there is a high risk that the wires for doing the measurement are more resistive than the sample it self (and then you'll measure the wires !). Carefully check that the response is ohmic, by using increasing but still extremely low currents. Higher current will provide higher accuracy, provided that no Joule heating occurs. Hence the need to build the I-V curve, which should be linear. Otherwise, temperature control is necessary.

Alain

I am using 4 point method. However, my real problem comes from the resistance of the connection which gives me an error too big for the resistivities I am measuring (I am using two copper plates to trap the wire in between the two, however this causes some capacitance issues as well as some variations from measurement to measurement). In total there are 4 copper plates (two for each side of the wire) and I connect the power source to the 4 plates for doing conductivity measurements. But one of the problems, is that if I press the plates together too hard, the morphology of the wire changes, changing resistance values. Or if a dust particle stays in between the plates, the connection with the wire is worse, so resistance is also affected. 

I don't know if I'm explaining myself correctly: What I would like to know is how to do a proper connection between my microwire (10 um in diameter) and an outside power source without affecting the morphology of the wire, or altering the properties in any way. By the way, I've tried silver paint connections but it is complicated to have any degree of accuracy when measuring the distance in between two silver paint connections.

Thanks very much again for your help. 

Guillermo

And answering your questions: The wire isn't a defined length, although the space in between the plates is always 1mm to have a constant length of wire "conducting". 

I use small currents as being the wire so thin (10 um- 20 um) the heating causes some issues. I do I-V curves and they are always linear. However, the conductivity which I calculate from one measurement to the next (with the same wire) might have a difference of 2 microOhms between them. And I am trying to see if there is any way to improve this error. 

Thanks

For your connections, you definitely have to use silver paint for glueing the contacts. Just be aware that it takes hours before it becomes very dry, inspite of the very dry aspect. And if not dry, the resistance will be higher than it should. So better putting the silver paint on day d, and doing the measurements on day d+1. With such system, no problem at all.

Alain

Thanks a lot Alain. That is exactly what I am going to do. Let's see if it works!

Guillermo