It is generally a very bad idea to draw any current through a reference electrode. Any current flow will result in a change in the environment around the metal/controlled electrolyte interface. This will result in a change in the half cell potential. How much change occurs depends upon the amount of current and the type of reference. Silver/silver chloride-seawater electrodes are notorious for being highly susceptible to damage due to current. I have heard of instances where frequent use of a low impedance analog voltmeter resulted in damage. Copper/copper sulfate electrodes can undergo a little bit of copper corrosion, but if the current flow is such that chloride ions are drawn into the electrode from the outside environment, the half cell potential will drift away from the documented values. Zinc reference electrodes are generally the most tolerant of current flow, but increasing the zinc content of the environment around the electrode and drawing contaminating ions from outside the electrode will eventually do damage.

So, the best approach is don't draw any current through the reference - period. If you find a reference that has had current flow for any reason, you need to recalibrate the reference with another reference that is known to be good. If the difference between the two electrodes is more than a millivolt or two, write the "damaged" electrode off as garbage unless you can physically retrieve it and send it back to be remanufactured.

Thank you Stephen N Smith Sir. Actually I needed to build a set up wherein a voltage has to be sensed through a 'Voltage Sensing Relay' that has a lower limit of a few volts, the 'sensed' voltage is actually below this limit, so I thought of using a transformer for stepping up the 'sensed' voltage. after your valuable input I don't think that my idea would work. Could you please suggest an alternative?

Your best bet is probably going to be to use an operational amplifier (OpAmp) to boost the sensed voltage up to the range for your relay. If you do this, be sure to select an OpAmp with an input impedance of at least 1 million ohms. This is pretty easy if you build the circuit using a prepackaged chip that uses CMOS, or similar, technology. The beauty of an OpAmp is that you can then adjust the gain of the amplifier to match the input voltage requirements of your voltage sensing relay, yet you maintain the low current input (high impedance) required by the reference electrode. If you search through the CP meter suppliers, you may see products referred to as "impedance amplifiers". These are simply OpAmps that have the gain hard wired at 1.0 so that you get the same voltage out as goes in. You could modify one of these commercial products, but it would be just as easy to build your own.

thank you Sir