Hello Khalil Hamam ,

I am not sure of the exact design of your home-made fixture, but, from your brief description, I would say the reason why the impedances measured with the Keysight 16451B fixture are higher than what you measured with your home-made fixture is that the Keysight fixture uses guarded electrodes. In other words, simply employing two flat electrodes will not work because such a configuration will allow surface currents to be included in the impedance measurement. The use of guarded electrodes eliminates the surface currents and only allows volume (or bulk) currents to be measured. In addition, the guard electrode eliminates the influence of any fringing electric fields due to the sample having finite extent.

I have attached a diagram depicting the use of a guard electrode to eliminate the fringing electric field and any surface currents that will be invariably present. The top electrode consists of the measurement electrode surrounded by the guard electrode. Notice that the current measuring meter is only connected to the measuring electrode. In this way, only the volume (bulk) currents are registered by the meter, the surface currents simply flow back through the battery. Without the guard electrode, the bulk DC resistance (impedance) would be R=V/(Isurface+Ivolume), but with the guard electrode the bulk DC resistance would be R=V/Isurface, which is larger - the result you get.

I hope this terse explanation helps, but if you still have questions, then don't hesitate to ask.

Regards,

Tom Cuff

Dear Thomas Cuff Thank you for you reply! I 'm a little bit confused about how should i connect my wires to make a similar four wires connection as in 16451B fixture. Your diagram is not clear ...can you look at the attached image and compare it to your figure. Would you please show me how does your figure is related to the attached figure?... Thank you in advance

Hello Khalil Hamam ,

Let me make sure I understand your question. I believe you are asking me how to connect your home-made fixture so that it mimicks the Keysight fixture. If that is, indeed, your question, then my answer is as follows.

The diagram I presented, with my initial answer, is for a fixture used to measure impedance (resistance) at DC (0 Hz). The LCR bridge you are using is an AC bridge, specifically designed to make LCR measurment from 4 Hz to about 8 to 10 MHz. Your home-made fixture would need to have a guarded electrode, the electrodes would have to be connected to the LCR bridge using coaxial cable, and the coaxial cable would have to be the same ones (same type) used in the Keysight fixture.

Does your home-made fixture have a guarded electrode and are you using the proper type of coaxial cables terminated with the proper type of connectors?

You will note from the diagram of the Keysight fixture that it uses the 4-wire measurement technique wherein the voltage (potential) is applied to the fixture through the center conductors of one pair of coaxial cables (HIghpotential & LOwpotential), while the current generated across the sample are measured by the center conductors of a second pair of coaxial cables (HIghcurrent & LOwcurrent). Note, the 4-wire approach means that you do not need to worry about the resistance of the center conductors of the coaxial cables dropping the forcing voltages being applied to the sample. I believe the shield (outer conductor) of the coaxial cables are all connected to the same potential as the HIghpotential, but I am not positive about this, yet. I will need to read the manuals of the Keysight fixture and the Hioki LCR bridge. I will get back to you in a day or two.

It would help me if you would include a photograph or two of your home-made fixture clearly showing the electrodes and cables.

Regards,

Tom Cuff

Dear Thomas Cuff , Exactly that what i'm trying to do. Mimic the Keysight fixture to perform high temperature measurements (the 16451B fixture is expensive and can't handle more than 55C).

I have this cable (BNC-BNC coaxial shielded cable) from ietlabs

https://www.ietlabs.com/1689-9602-bnc-bnc-extender-cable-1816.html

and i made electrodes similar to the one provided by 16451B fixture manufacturer. My home-made electrodes are copper based material and the gap used in the guarded electrode was made from teflon (Electrode-B as shown on page 73, attached file)

**Could you please double check your answer regarding how do they connected " You will note from the diagram of the Keysight fixture that it uses the 4-wire measurement technique wherein the voltage (potential) is applied to the fixture through the center conductors of one pair of coaxial cables (HIghpotential & LOwpotential), while the current generated across the sample are measured by the center conductors of a second pair of coaxial cables (HIghcurrent & LOwcurrent). "

I can see that both HIghcurrent &HIghpotential are connected to the unguarded electrode while, the other LOwpotential&LOwcurrent are connected to the guarded electrodes!.

Hello Khalil Hamam ,

It sounds as if you have built something very similar to the Keysight fixture. The use of copper electrodes is a possible problem, especially at high temperature as they oxidize very easily, even at room temperature. Are you sure the Keysight electrodes have a bare copper surface? I would have expected them to be plated with nickel or platinum. I already downloaded the PDF document you attached, and I have been looking through it for the last hour but I have not been able to find if they describe the material from which their electrodes are constructed and whether or not they are plated. Have you found this information anywhere?

Regards,

Tom Cuff

Their electrodes are not copper they seems (most likely) to be platinum or at least nickel due to the color of the electrodes.

Hello Khalil Hamam ,

When you configure the Keysight and your home-made fixtures to have the same electrodes and electrode spacing with air as the dielectric, do you get very different results for the measured impedance at room temperature?

Clarify something for me, please. When you say the electrode material of your home-made fixture are "copper based material", do you mean they are made of some type of brass alloy or do you mean something else? If, indeed, you are using a brass alloy, then you might possibly have problems with the oxide layers that will form on the brass surface, especially after heating the fixture. The oxide layer will increase in thickness each time you cycle the temperature, and the oxide layer acts like it is in series, electrically, with the dielectric sample under test.

I apologize for my pedantic exposition on the 4-wire measurement scheme. I did not realize from your original question that you had essentially cloned the Keysight fixture and why you had done this (for high temperature measurements). Obviously, the diagram in my initial response was a 2-wire guarded electrode measurement (Sorry, I initially said 3-wire, but the guard electrode wire does not count. It is only the wires on the measuring electrodes that count.), and you correctly copied the 4-wire wiring system shown in the Keysight manual. Everything you have done looks adequate to me. I can understand your frustration at having the comparison measurements diverge after all this hard work.

One last question: you said your home-made fixture was to operate above 55C, but are your coaxial cables rated for use significantly above 55C?

Regards,

Tom Cuff

Hi Thomas Cuff Yes my electrodes are made of brass (copper), so may be the oxides has an effect on the readings! due to the formation of oxides layers. What do i have to do! plate them with silver layer ! I have silver paste available in my lab do you thin it will work!?. Regarding the wires, i don't insert them inside the oven! (the end is at the oven gate), then, heat-isolated wires were used.

Hello Khalil Hamam ,

Silver plating, like copper or brass, also tarnishes, except that silver reacts with sulphur compounds in the air. I am sure you have seen silver utensils turn dark after a few months. That is why people are forever polishing their silverware, unless they are covered with a thin layer of laquer to exclude the surrounding air. Electroplating the electrodes with platinum would probably be the best idea; you could, of course, plate them with gold, which does not tarnish, but unalloyed gold is extremely soft. Note, before performing the electroplating operation, the brass parts have to have the coat of tarnish removed by a chemical dip in an alkaline solution of KCN or something equivalent. WARNING, DO NOT do this dipping process yourself, hire a professional. If the pH of the solution goes below 7 for any reason, HCN (hydrogen cyanide) gas will be released which will kill you almost instantly.

Clarify something else for me. You say that the coaxial cables do not enter the furnace, but that "heat-isolated" wires connect the fixture with the coaxial cables. Are these "heat-isolated" wires simply wires covered with an electrically insulating layer of teflon or something similar? The reason I am asking is because you must use coaxial cables from the LCR bridge all the way to the fixture or you are going to get incorrect impedance results. The reason is that the wires without an outer conductor (shield) will electrically couple to external objects and even yourself due to stray capacitive coupling.

Regards,

Tom

Hi Khalil Hamam ,

Sorry, in my last message (above), I forgot to completely answer your question about your silver paste. When I talked about plating the brass parts of your home-made fixture, I meant electroplating not simply rubbing the parts with silver paste. In other words, after performing the chemical dip to remove the tarnish film, the parts would be placed in an electrolytic cell as the cathode with the anode being a piece of platinum. In this way, the platinum would be deposited on the brass pieces.

Regards,

Tom

Thank you Thomas Cuff , I appreciate your nice response to my questions. I'll try to apply your recommendation soon.