I am not very clear about your sample and the substrate on which film is deposited..... For making ohmic contact, since ZnO is an n-type semiconductor with work function ~ 5.3 eV. Hence for your contact to be ohmic, you should choose a material with a lesser work function than that of ZnO (5.3 eV). In this case Au with a work function of 5.2 eV is the best choice than Ag (4.7 eV)

You can check whether your contact is ohmic or not by recording an I-V at a particular voltage. If I-V is linear and symmetrical in the +ve and -ve cycle, then the contact is ohmic.

In your case since you are using Ag paste it will result in a larger potential barrier. It also depends to some extent on the quality of silver paste. There is a probability that the solvent of your Ag paste may be reacting with ZnO resulting in non-repeatable characteristics. I am not sure whether you are doing your measurements at room or elevated temperatures.

In my opinion the best will be if you first deposit Au electrodes on your sample and then over them you take Ag paste contact and first record an I-V to check whether the contact is ohmic.

If i am wrong somewhere, i request the fellow researchers to correct me so that we don't pass wrong information..............

I built and worked on a similar system with same instrumentation. I must ask, have you done troubleshooting to the system? Please verify that the system is indeed working properly by measuring something easy such as some doped Si with a known resistivity to ensure there are no problems with your Van der Pauw setup before you proceed.

To test for an ohmic contact simply measure an IV curve between two of your contact points. You will be able to infer from the graph if it is ohmic or not.

What type of substrate is your thin film on? If it is conductive it may alter your measurements considerably. Recall, Van der Pauw measurement assumes homogeneous film.

I have had trouble with silver paste before. I found that melting a very small dot of indium onto the sample where the probes make contact has worked better than silver paste.

I was successful in measuring mobility with same equipment using 1000 Gauss magnet. You should be able to easily do a quick calculation based on the precision of your instrumentation and the expected Hall voltage for your sample under B-field.

I agree with Dr. Requena's comment that an indium dot can be a easier way to make an ohmic comment than silver paste. At least it can be a good check on whether your silver paste contact is the problem.

From my experience, varying resistance is indeed usually the result of a contact issue. Since you haven't cured your silver paste, there may be some liquid phase present, and you might be seeing the result of ionic conductivity.

Curing conditions for your paste depend on the manufacturer, but for the stuff I've used, usually 120 degrees C for an hour on a hot plate is sufficient.

I am not very clear about your sample and the substrate on which film is deposited..... For making ohmic contact, since ZnO is an n-type semiconductor with work function ~ 5.3 eV. Hence for your contact to be ohmic, you should choose a material with a lesser work function than that of ZnO (5.3 eV). In this case Au with a work function of 5.2 eV is the best choice than Ag (4.7 eV)

You can check whether your contact is ohmic or not by recording an I-V at a particular voltage. If I-V is linear and symmetrical in the +ve and -ve cycle, then the contact is ohmic.

In your case since you are using Ag paste it will result in a larger potential barrier. It also depends to some extent on the quality of silver paste. There is a probability that the solvent of your Ag paste may be reacting with ZnO resulting in non-repeatable characteristics. I am not sure whether you are doing your measurements at room or elevated temperatures.

In my opinion the best will be if you first deposit Au electrodes on your sample and then over them you take Ag paste contact and first record an I-V to check whether the contact is ohmic.

If i am wrong somewhere, i request the fellow researchers to correct me so that we don't pass wrong information..............

I had a similar issue with my doped ZnO films while doing an R vs T measurement. I was using soldered indium contacts and a Keithley 2400 sourcemeter. Everytime I started the measurement, it was giving me increased resistivity. Could not find out the reason.

I have seen this before as well. My opinion is that it is a time constant issue involved with charging of the film like a capacitor. It depends on the thickness of your film and its resistivity. Thicker films that are more resistive are worse. It also depends on your instrumentation setup and whether the wires are shielded (twisted shielded pair is best). Contacts are always an issue with ZnO and should be checked first. We developed a system specifically to deal with high resistivity ZnO and similar wide bandgap materials using a buffered electrometer setup. (http://dx.doi.org/10.1063/1.3436648). For Hall effect on resistive material especially, you need as high a magnetic field as you can get, since the hall voltage is quite small and there are competing effects which will skew your results if you are not careful (see book by D. C. Look).

I am using Glass substrate for deposition. Hall measurements are carried at room temperature.

I want to support the comment from Vishal Balouria above. I have been measuring resistance and Hall voltage on oxide thin films on insulating substrates recently as well, and whereas the room T resistance is only ~ 100-10,000 Ohm, making contacts directly with Ag paste or epoxy directly on the films will generally give me poor contacts with noise in the measured voltages. However, first depositing contact pads of Ag (or Ti/Au or some other common combination) using thermal or electron beam evaporation and then contacting these pads to wires using Ag epoxy works a charm, with stable resistances over measurement temperatures of 2.5-400 K, and lasting for at least several weeks between measurement times.

Any better option than Indium dots, for ZnO films to make contacts for Hall measurements ?

The problem with Indium dots is they don't solidify much and tend to spread over the film.

Thank You.

Making Al contacts with shadow mask is a better option for contacts on ZnO films. Aluminium can be easily evaporated by PVD.

Have you observed exponential variations of resistivity with time (covering periods of minutes)? I am not an expert, but they might be related to a curious effect called (I don't know why) "persistent photoresistance". Here some examples:

- Solid State Communications 149 (2009) 1347

- Solid State Communications 151 (2011) 1700

- J Vac Sci Technol A 31 (2013) 01A146

- Appl. Phys. Lett. 99 (2011) 112503

I hope it helps...