Hi,

For a four-point measurement (V, I), it is preferable to sputter the thin films on the four contact-point of copper. Also, you can sputter the four contact-points on the top of the thin films.

Hi Jayant,

as M.R. Britel already mentioned using four-point probe measurement technique is neccessary especially for thin films to get rid of contact resistances. (Depending on your sample caracter, it might be possible to induce also Schottky contacts what you have to consider as well and however, another possibility for measuring low resistances is the van-der-Pauw measurement technique.)

Further on it is recommended to have a very stable current source and use voltmeters with high accuracy because current should be sometimes quite low, to do not destroy your sample. I have built a circuit board with soldered spring contacts (for example see: https://www.distrelec.de/ishopWebFront/search/luceneSearch.do?dispatch=find&keywordPhrase=5099-D-2.0N-AU-1.0C ) onto which I press my sample using additional non magnetic springs. It is important that the sample can not tilt or move during measurement, which would lead to drift... Also think about current flow direction with respect to magnetic field or magnetization of your sample (keywords: AMR, Hall...).

If you plan to measure GMR also think about CPP measurements because the GMR ratio is larger for CPP than for CIP.

Cheers

Dear Britel and Patrick,

Many thanks for your reply. I am more concern about the magnetoelectric voltage coefficient . When mulitferroic or magnetoelectric material kept in magnetic field it generate some voltage. We measure this induced voltage using lock in amplifier with two probe method only. Since the induced voltage signal is small (in the range of nanovolt or microvoltt) . With only copper wire does it produce noise in the sample or Do I need BNC/ triaxial cable for the measurement ? Also do I need to to shield the sample with conducting metal box for grounding?

Hello Jayant,

the 4 probe arrangement is usually required when you measure a voltage generated by an (external) current. As was pointed out before, contact resistances otherwise spoil your measurement.

In your experiment it seems to me that you (have to) consider the samples you study as an intrinsic voltage source. Since you don't drive a current through the sample, The consideration above probably is less useful here.

Rather you should be concerned about the current load imposed by your voltage measurement arrangement on the voltage source. You may cause a voltage drop on the "internal resistance" of the material. (Similar to a battery: its voltage decreases when you pull more current, and you cannot avoid this, even by a 4 point probe, which therefore doesn't really make sense here).

So your focus must be on high impedance voltage metering.

Actually how do you generate the alternating voltage you measure with the lock-in?

To what is written by Kai I would like to add several practical points:

 If the sample will be conductive, no stationary charge induced by magneto-electric effect can be detected at all. Instead, one can observe a transitional charge upon applying the external field which will decay in the time with the respective RC constant. For this case one additional contribution will arise (and very often it will dominate in the measurements) : the conducting sample with the wires connected to it is a loop,  which will generate an electromotive force on each change of the magnetic induction (Faraday's law). One should always keep in mind that. 

I can propose to read an experimental section in our paper: 

Direct and converse magnetoelectric effects in Metglas/LiNbO3/Metglas trilayers

JAP

DOI: 10.1063/1.4816400