Network Analyzer versus Impedance Analyzer

[Cleber Borges]

Hello everyone,

Reading about the EIS area - Electrochemcical Impedance Spectroscopy - I found that currently commercial potentiostats implement Impedance analysis by the method:
FRA - Frequency Response Analyzer, also referred to as Transfer Function Analysis.

As far as I understand, this method would be the same as the one implemented by the Network Analyzer tool in AD2 (am I right? ? ), since in its description it consists of:
"The Network Analyzer is used to analyze transfer functions (the ratio between an output function and an input function)"

If I can have the Impedance Analysis (measurement) implemented by the two tools: (i) Network Analyzer and (ii) Impedance Analyzer;
with differences observed between the two tools, for example, in the frequency range:
Network Analyzer = 2 mHz up to 10 MHz - and no "open" and "short" compensation option
Impedance Analyzer = 200 uHz up to 25 MHz - and with "open" and "short" compensation option

Which made me have the following doubts:

[1] - What is the fundamental difference between them?
[2] - I saw that the Impedance Analyzer tool forces the current calculation while the Network Analyzer emphasizes the voltage attenuation. But what is the implication of this?

In the literature, I saw the defense of the FRA method based on measurement accuracy. The source signal is multiplied by the attenuated signal and the result is given from this combination.

[3] - Does the Network Analyzer tool algorithm follow this form?
[4] - In the WaveForm examples, there is the use of reference resistors for use of the Network Analyzer tool. Is it mandatory to use them or can you close the circuit by connecting to Ground-GND?

Many thanks for your attention and patience!
These details may be basic to those trained in electronics, but it is difficult to find them in referrals from other areas. And it certainly makes all the difference trying to do the measurement correctly.
Once again: thank you!

Cleber Borges

 

 

[attila]

Hi @Cleber Borges

1-2. The Network Analyzer is intended to characterize transfer function of amplifiers/filters (C2/C1) mainly in form of magnitude and phase, Bode magnitude/phase, Nichols, Nyquist plots. It has custom scriptable plots that can be used for impedance analysis too. This was added before implementing the IA interface.
The Impedance Analyzer goes further in calculations, it is intended to characterize elements (circuits, components, materials) in complex form like real and imaginary part, resistance and reactance, capacitance and dissipation, inductance and quality. The IA takes in account the scope probe impedance, the value of this can be comparable to the measured element at high impedance and high frequency; open and short compensation; DUT = like R*|C1/C2|- |RC2|- |Open|- |Short|

3. The NA performs FFT on C1, C2 and gives the magnitude for C1 as C1/Set_Amplitude, for the other channels as C#/C1 and phase C#-C1.

4. I'm not sure if understand correctly, but if you close the circuit without resistor basically you short the C2 to ground... The IA needs a resistor for current limitation and reference.

[Cleber Borges]

Hello @attila
Thank you for your consideration and for your response.
My question is more about the theory itself.
Since I have the Magnitude and Phase experimental measure, the rest of the properties are reasonably easy to calculate by a CAD (algebra software)
My concern is to know what are the real limitations involved in measuring when I ask about fundamental differences.

[attila]

Hi @Cleber Borges

The principle of measuring the gain/phase in NA and input gain/phase in IA is the same, except that the NA shows C2/C1 and IA C1/C2.

In the Network Analyzer FFT is used since it has option of using external signal generator and spectrum analysis (THD, THD+N, HD2, HD3... plots)

In the Impedance Analyzer Goertzel filter is used since the frequency is known, always generated by the AWG of the same device. The input gain and input phase are raw values (C1/C2) the further measurements (impedance, admittance, phase...) take in account the scope probe impedance and open/short compensation.

[Cleber Borges]

Hello @attila,  Good Morning,

I imagined that using the reference resistors (such as the impedance analyzer extension board) could have been used the I-V method, since it would be easy to calculate the current value (I) ...
But, given your explanation, I understand that the AI tool still uses the same method: Transfer Function Analysis (or the same as FRA - Frequency Response Analysis) ... [I needed this information to report]

Thank you very much for your response and dedicated attention to this issue.

Cleber Borges

[Andras]

Hi guys,

This is a fascinating topic!

@Cleber Borges, did you end up using AD2 as a FRA (Frequency Response Analyzer)?
Do I understand it correctly that you did experiments with fluids?
Did you get consistent results?
Could you share your settings with us?

Thanks,
Andras

[Cleber Borges]

@Andras I ordered the IA card and it took six months to arrive in Brazil. (Digilent-USA did not want to sell me direct for having representatives in Brazil) ...
So I left the project temporarily ?

I'm sorry you can not contribute information

Cleber

 

[Andras]

@Cleber Borges, I'm sorry to hear that! I had to wait a few weeks as well here in Hungary, but it wasn't as bad as in your case.
If you have no other option, let me know, I buy you one here and send it to you to Brazil.

If you had the Impedance Analyzer, how would you make the measurements?
I saw your other post:

 

The Impedance Analyzer has only two wires: + and -

 

So, would you put just those two wires into a liquid and run the Impedance Analyzer tool in WaveForms? Would you expect to get consistent result for the same type of liquids?

Thanks!
Andras

[Andras]

I did a quick experiment with different liquids. All of them are in a small ceramic container.

The blues (Ref1, Ref2 and Ref5) are tap water.
The red (Ref3) is salted water.
The green (Ref4) is carbon-filtered tap water.

I just put both the + and - wires of the Impedance Analyzer into the container and I made sure they are not directly connected.

Here are the results:

Cool stuff!

[Cleber Borges]

hello @Andras ...

My IA board (impedance analyzer) arrived a few days ago, but I did not even test it ... But now I have the AI board in my hands. Thanks for your attention.

The IA board does not allow a reference electrode (RE) such as figure-item B. Such a scheme would be to not pass current between the working electrode (WE)  and a reference electrode (RE). It would be to measure more accurately and discount several other effects in a more automated way. This scheme is not necessarily mandatory.

You can use the scheme of just two electrodes as you mention.
I am a chemist and I do not have much knowledge of the electrical / electronic part. I imagine that your measurements reflect well the salinity of the water tested.

 

Actually, I think this device - AD2 - could provide a lot of chemical information of Impedance measurement, but I just started reading about it.

It's great to know that more people are interested in AD2 and EIS ... so we can build a source of information.

Regards,

Cleber

 

[Andras]

Hey guys,

I've made some experiments that could be interesting for your as well. I put tap water into my ceramic container, I heated it to different temperatures and measured the impedance every 5 degrees.

You can see the values between 60 ° Celsius (140 ° Fahrenheit) and 5 ° Celsius (41 ° Fahrenheit). Red is 60 Celsius, blue is 5 Celsius and there are 10 steps between them.

[Andras]

Before I continued my tests with beer, I wanted to make sure of two things:
a, the amount of the sample fluid does not influence the impedance values
b, the measurement itself does not change the sample fluid so that its impedance is changed at every measurement

For case A, I made 3 references with 1 cup, 2 cups and 3 cups of tap water.  I used the same source and the same container for all the measurements.

 

Then I made 3 other measurements of the same 3 cups of water.

 

At this point I can conclude that the measurements are fairly harmless, there is only a small amount of change on the dataset after each consecutive measurement. I could probably use lower voltages, which could theoretically reduce the impact further.

The amount of the test fluid does change the dataset somewhat, so I will keep that in mind when I go ahead with my further tests and I will try to keep the volume of the fluid fixed.

As we could see from my previous post, the temperature is very important, so that's another parameter I will try to manage and keep constant between the tests.

[Cleber Borges]

Hello @Andras ...

Could you show how your electrode scheme looks? I think it is also geometry / electrode setup is a very important parameter.

Very cool the results of your experiments ...

In electrochemical impedance spectroscopy (EIS), I see that the voltage amplitudes are much smaller than you are using ... In general, the articles show that they are 100mV amplitudes due to problems with linearities in Nerst's law ... Have you tested at lower amplitudes?

Thank you very much.

Best regards,

Cleber

[Andras]

Hi @Cleber Borges and @attila,

Sure, my electrodes are simple wires fixed at a certain distance from each other, and they look like this:

Do you know what kind of electrodes researcher normally use when they do EIS?

 

It's interesting that you just mentioned the voltages, I tested them as well, because I feared that the high voltages would make changes in the target fluid especially if it's a living thing, like yeast.

So here is the test I had, comparing the different amplitudes/voltages:

Probably it isn't obvious on the screenshot, but the 5 mV - 50 mV charts are fuzzy, I think that amplitude would be not good for testing. 100 mV is the first value that looks ok to me, so I will start using that one from now on

I repeated a few 100 mV tests a few hours apart on my current fluid sample which was tap water + glucose + brewer's yeast. The result were not consistent:

I suspect that it is because there were some reactions going on with the sugar and yeast, so today I'm going to repeat the tests with a control sample of tap water, and the same glucose + yeast combination.

Edited July 24, 2019 by Andras

[Cleber Borges]

Hello @Andras ...

Actually, two parallel plates are enough as electrodes ... But always make sure that they are well fixed and constant as any change in geometry adjustment will cause a lot of variation in measurement.

The arrangement would only need more sophistication if the electrode itself developed a chemical potential at its interface! Which is not the case here ... So I think your scheme is satisfactory. [ 1 ]

I would like to know the results with your yeasts ... if you can post the results.
Very good to see a case study.
Thank you very much

Cleber

[ 1 ] - page 2 = *Two electrode setup*  (  www.ecochemie.nl/download/Applicationnotes/Autolab_Application_Note_EC08.pdf  )

((( The two-electrode configuration can therefore be used whenever precise control of the interfacial potential across the WE electrochemical interface is not critical and the behavior of the whole cell is under investigation. )))
 

[Andras]

Alright, so I made 9 consecutive measurements in 10 hours on both tap water and the yeast + sugar + water mix.

The results for the water are the following:

And results for the yeast:

 

The values are changing for both over time, which is expected, but there are a few strange things:

- The first tests for both are not in line with the changes happening later, they look off.
- The amount of changes for both samples are similar, does this mean that it doesn't matter if it's something living like the yeast or non-living like tap water?
- I'm not sure what the reason for the changes are: Could it be the little temperature fluctuation during the day? Could it be that the evaporation from both containers?
- It's not proven yet, but I got the suspicion that small bubbles form on the electrodes during the time between the tests and they change the measured values. Would I get different results if I shook them or didn't shake the electrodes before the measurements?

Also, I'm not convinced that the yeast started to work today. I can't really tell, I need to get more information about the brewer's yeast and its properties.

[Cleber Borges]

Hello @Andras ...

I think bubble formation is critical ...
1- Is CO2 forming?
2 - Just some other gas coming out?
3 - Yes, temperature is a critical parameter in measurements that require a lot of accuracy ... But I don't know how much temperature variation can affect a preliminary study that doesn't need as much precision and accuracy.

Cleaning the electrodes I think is a good idea ...
It would be nice to check how much influence on the results ...

 

[Andras]

Hello @Cleber Borges

Today I tested a few things: I made a measurement before and after cleaning the electrodes and I did the same before and after I whirled the mixture. Neither of them modified the values, which is good news.

I have found one interesting thing though: if I make the impedance measurement while the temperature sensor is submerged in the mixture, it distorts the values in the 20-25 Mhz range.
I use a multimeter to measure the temperature, so from now on I will not keep its probe in the mixture, and I will put it into the fluid only for a few seconds when I actually check its temperature.

I can't tell at the moment if there is any gas formation, it looks like there isn't any. It can be because I set one parameter or ratio wrong.

According to the literature my room temperature is higher than the ideal (it's 27-28 Celsius here) and the ideal would be somewhere below 20.

Today I added salt to the mix, because the minerals are said to be critical for the yeast as well.

Here are the results after 3 hours:

(tap water + glucose + brewer's yeast + Himalayan salt)

 

Since my goal would be to detect the phase of the yeast + sugar reaction with this impedance measurement, I will change the measured range to 100 Hz - 1 kHz, it looks like most of the changes are there.

According to the data the reaction starts right away, but it stop after about 2 hours. This, again, can be because I use wrong ratios or at a wrong temperature, I need to make more tests.

 

Andras

 

[Andras]

Alright, I finished the tests for today, here are the results. I used "traditional" yeast now, and not the brewer's yeast as before.

(A - blue - left)         27 Celsius water + 2.5g glucose + 2g yeast
(B - red - middle)     27 Celsius 2.8% milk + 2.5g glucose + 2g yeast
(C - green - right)    50 Celsius 2.8% milk + 2.5g glucose + 2g yeast
[please note that all the temperatures converged to our room temperature in the end to 28 Celsius]

The following two pictures were taken 5 minutes (light color) and 120 minutes (darker color) after mixing the ingredients together.

 

And the impedance values in the 100 Hz - 2 kHz range:

 

As a next step I will try to write a small script to make single impedance measurements every 30 seconds or so, and plot the changes over time, so that I could see the rate of the change over time.

[attila]

Szia @Andras

For this you don't necessarily need a script.
In the interface you can set constant frequency (Start = Stop), specify a long Settle time and press Single.
Like the following will run for 50 minutes, 100 samples at about 2/min rate, 30s + a few milliseconds due software processing.

[Andras]

Cool stuff @attila! Thanks, I will definitely try it tomorrow!

[Cleber Borges]

Hello @Andras ...

Very cool your results ...
This case study is very interesting!
Have you considered putting all this information on a blog too? (It can be an inspiration and starting point for others).

Edited July 31, 2019 by Cleber Borges

[Andras]

Hi guys,

Today I created a mixture with tap water and glucose just as before. I made a short (3.3 minute-long) and a long (2.7 hour-long) experiment.
I made a test with water + glucose mixture and then I did the test once again after I put the baker's yeast into the mix.
I focused on both the impedance and capacitance values.

Here are the results for the short test:

The blue is the tap water for reference.
The two reds are the water + glucose combination. I made the two measurements approximately 60 minutes apart, but I also made sure that they have the exactly same amount of glucose in them. I'm not 100% sure why the values are so different.
The yellow/green are the baker's yeast mixtures having approximately the same amount of yeast put into the water+glucose mix I used in the previous tests.
They showed a nice pattern, so I thought I repeat the experiment, and I let the yeast work for more time.

So this is the result of the long test:

Blue is tap water for reference.
Orange is the baker's yeast.
As you can see both the impedance and the capacitance show the activation of the yeast beautifully. I will probably use capacitance for my future tests, because that looks more persistent.

@Cleber Borges, thanks, I will probably put together a summary of these when I'm finished with them on my blog (blog.biobalancedetector.com). I hope people will find this forum too

@attila, thank you for the improvement!
I noticed one more thing: if I use the Impedance Analyzer as you described (Start=Stop) and I generate a chart in the time-domain, the x-axis is not labeled. Could you add labels on it, so that we could see how much time elapsed during the measurements?
The other thing is probably connected to this: the Vertical Quick Measurement tool seems to be not working in this mode, it doesn't move from the first column of the chart.

 

Edited August 1, 2019 by Andras

[attila]

Szia @Andras

The quick measure and cursor position used only frequency value which was useless in constant frequency mode, zero range.
The updated software uses percentage unit for x axis in this corner case:
https://forum.digilentinc.com/topic/8908-waveforms-beta-download/

 

[Andras]

@attila That's pretty cool!

Also, I'm not sure if this was there in the previous versions, but I really like the "Pick screen color" function, because I like to use RGB gradient tool (http://www.perbang.dk/rgbgradient/) to get my gradient colors. ?

 

Anyway, I have a question relating to my next phase of my experiments: is it possible to use the BNC Adapter with the Impedance Analyzer? I would need to stimulate something with a signal generator while doing the capacitance tests over and over.

As I quickly tested it, the Impedance Analyzer module seems to be not working if it's attached to the BNC module (which is attached to the AD2).

 

Thank you!

 

[attila]

It is not possible.

On the BNC adapter the Scope and AWG channels are available only through the BNC connectors. On the other side of the adapter the GND, Supplies, Trigger and Digital lines are available.

You could create impedance analyzer circuit with fixed reference resistor, like it is shown in the Help.