What is a Reference Electrode Shunt and why would you use one?

In this video we will be talking about reference electrode shunts. We will cover what a reference electrode shunt is, why you would use it, examples of Nyquist and Bode plots with and without shunts, and when it might not be appropriate to use a shunt.
Timestamps
00:00 - Intro
00:58 - What is a reference electrode shunt?
02:27 - Why use a shunt? How does a shunt work?
04:39 - Example Bode and Nyquist plots with and without a shunt
06:41 - Why not to use a shunt

[@loganwilder7272]

Great video! I'm just starting to break into EIS and your videos are super helpful. Thank you for posting these.

[@Pineresearch]

Awesome! I'm glad you enjoyed the video. Stay tuned for more electrochemistry videos from us :)

[@TarikZahzah-ns2bu]

As a manufacturer of potentiostats, I would appreciate your insights regarding performing EIS in the high-frequency range, specifically in the kilohertz (kHz) or megahertz (MHz) range. In my research group, we recently had a discussion suggesting that measurements beyond 10 kHz may not provide reliable information and could be considered meaningless. However, I am aware that many researchers report using maximum frequency values that are ten times higher, such as 100 kHz. I'm curious to hear your perspective on this matter. If it is indeed advisable to limit the high-frequency range to 10 kHz, it is possible that the use of a reference electrode shunt may also be unnecessary.

[@Pineresearch]

Tarik, it depends on the system under study. It is true that a number of typical systems - especially those using a liquid-phase reference electrode with a frit - may have difficulty responding at frequencies above ~10-100 kHz, making those data somewhat "meaningless." However, that is why specifically the shunt can be useful because, in the case where information is desired at those frequencies, the shunt can be considered a way to obtain more accurate potentiostat response where the reference electrode is not limiting the EIS data at those frequencies. However, another reason why data at these frequencies are also sometimes not needed is because there are not many processes that typically occur in aqueous or non-aqueous systems that can be probed at those frequencies. Most often, all you may elucidate between ~10-100 kHz and 1 MHz is the uncompensated resistance and possibly cable inductive effects. The former is relevant but also easily probed at slightly lower frequencies as well, and the latter is largely irrelevant to the overall electrochemical response.

[@TarikZahzah-ns2bu]

@@Pineresearch Absolutely! The point you raise in the second paragraph aligns with what was emphasized in my group discussion. I can't recall any electrochemical process of significance that can be effectively examined at frequencies exceeding 10 kHz. Hence, I personally intend to avoid that frequency range in the future. I appreciate you sharing your valuable insights on this matter. Thank you!

[@TarikZahzah-ns2bu]

I have an additional practical question. Does the presence of a connected reference electrode shunt have any impact on the accuracy of your cyclic voltammetry data? Specifically, is there a possibility that during cyclic voltammetry, the shunt inadvertently and unintentionally functions as a pseudoreference instead of the intended reference electrode? One option would be to disconnect the reference electrode shunt as a precautionary measure during cyclic voltammetry. However, in situations where you are running a time-consuming sequence of both cyclic voltammetry and EIS measurements, it may not always be convenient or feasible to physically modify your electrode system for each measurement.

[@Pineresearch]

I highly doubt it. The best way to check would be to just run the experiment with and without a shunt yourself and see if you notice any impact (apart from normal variation between tests during normal electrochemistry testing). But CV is not fast enough for that to likely matter.

[@vidzajaknap7726]

Maybe stupid question, but when you start measurement you set the reference electrode potential according to the kind of electrode you use. What would you set there now since it changes from platinum to the other during the measurement.

[@Pineresearch]

Not a stupid question at all! In fact, I think you are pointing out an interesting quirk associated with using a shunt, which is that at the highest frequencies the "reference electrode" is technically just a Pt wire. This DOES mean, to your point, that there is uncertainty in the exact potential being applied at your working electrode during that time. Here is why most probably this is not an issue: the shunt is only really functional at high frequencies where the impedance of the series capacitor is less than the reference electrode, and at the highest frequencies the only phenomena that are normally being probed are things like uncompensated resistance, solution resistance, etc. These are (in theory) set impedances that are independent of applied potential, so basically the uncertainty in exact applied potential becomes irrelevant. Once the frequency drops to where you are measuring other phenomena like charge-transfer resistance or double-layer capacitance, the potentiostat has switched to sensing the actual reference electrode so the potential becomes as desired. I hope this is clear. I will also add this to talk about live this coming Friday, Sept 20 at 1pm EST in our next episode of Ask Us Anything About Electrochemistry, it will be episode #67 if you want to watch it afterwards. Link: kzbin.info/www/bejne/q3bOqZKuasSUask

[@vidzajaknap7726]

@@Pineresearch Thank you very much for quick reply

[@adammaraschky3845]

How can you verify that the shunt is getting rid of the error and not introducing a new error at high frequencies? I understand that the negative intercept with the real axis does not make sense. But in other situations I can imagine it would be useful to validate the impedance data using another means, such as current interrupt or one of the other techniques used to determine ohmic drop. I don't think the shunt would be needed for DC techniques. It might also be worth testing other capacitors to check if the data converge at a certain capacitance.

[@Pineresearch]

You are correct that the shunt is almost certainly not needed for DC techniques. And as it relates to other techniques for determining iR drop, you can do current interrupt or positive feedback (both DC methods) as a means of validation that your high frequency EIS data - either with or without a shunt - is perhaps roughly accurate. Personally, I find current interrupt to be a rather poor and very often inaccurate technique, but I like positive feedback and would use that for such verification. To your first point, it is certainly a possibility that the shunt could introduce some other errors. Perhaps the detection and use of the pseudo-reference Pt wire at those high frequencies, even with such quick speed, might be less accurate as there can be inaccuracies with the potential during that time, so possibly your working electrode may not be in the condition you think it is based on your applied setpoint. Additionally, choosing different capacitor values could be instructional, but mainly it would just be a way to adjust the point where the potentiostat switches from pseudo-reference to actual reference, and the larger the capacitor you pick the more frequencies you will be using the pseudo-reference, which could further cause errors. So this might be an interesting idea, but I would be cautious about using very large values of capacitor because at some point you're just using almost exclusively the pseudo-reference, particularly where it is not necessary to do so.

[@TarikZahzah-ns2bu]

Great video! I have both a question and a comment that I think could be valuable. First, the question. Let's consider a scenario where an electrode shunt is employed alongside a designated reference electrode. Based on your EIS data, is there a way to determine whether the potentiostat utilized the shunt as the path of least impedance instead of the reference electrode in the high-frequency range, without solely relying on a comparison between two experiments-one with the shunt and one without? Now, onto the comment. If possible, I highly recommend using a palladium wire specifically as the shunt or pseudoreference. The rationale behind this suggestion is that palladium exhibits a more consistent electrochemical potential in aqueous solutions due to a well-defined redox reaction known as the formation of palladium hydride. In contrast, using a platinum surface as the shunt introduces uncertainty regarding the precise redox reaction occurring. Employing a Pd wire can significantly impact the quality of data obtained in EIS experiments, particularly when employing small amplitudes (

[@Pineresearch]

Honestly, it's pretty tricky to answer your first question. Part of the reason is, to be honest, I think my explanation is slightly inaccurate regarding the way a shunt works. I just don't exactly know perfectly well how to explain the phenomenon from an electrical perspective. Here is what I mean: the shunt capacitor has an inverse relationship between applied frequency and impedance. The reference electrode has a roughly constant impedance. You can probably answer your question with some level of accuracy by finding the frequency where the capacitor exceeds the reference electrode. Frequencies above that are using the pseudo-reference, frequencies below that use the reference. The reason I say that explanation is slightly inaccurate is because the frequency applied and current being measured is with on the working electrode. I guess it is somewhat unclear to me whether it's accurate the reference electrode is "experiencing" the frequency during an EIS test the same way the working electrode is. Virtually no current passes through the reference electrode. So, maybe it is an overthinking here but I am not 100% certain my description above of that critical frequency where the impedances of capacity and reference electrode is ironclad, so to speak. Regarding your suggestion of Pd vs. Pt, that sounds reasonable to me but I cannot speak further to it. I don't know the difference between the two in this context and kind of experiment would be very severe or noticeable, especially at such fast frequencies and short time periods, but perhaps it would be useful to use Pd for a shunt over Pt if one has the option.

[@TarikZahzah-ns2bu]

@@Pineresearch Thank you for your detailed response. The reminder about the impedance formula for a capacitor was incredibly helpful in my understanding. Due to the inverse relationship between capacitor impedance and frequency, it is possible for the impedance at high frequencies to drop below a critical value, causing the impedance of the shunt electrode to become lower than that of the reference electrode. Consequently, the shunt behaves as a pseudoreference. This explanation makes perfect sense to me! Since I'm using the HydroFlex as my reference electrode, I should look up its constant impedance. It's funny that you mentioned contemplating what actually happens to the reference electrode during an EIS measurement, as that was precisely the question on my mind while watching the video. It seems like highly fundamental information. I assume that the reference electrode must somehow experience the frequency during EIS to enable the shunt to "short circuit" at high frequencies. I have an additional question. In the video, you mentioned that shunt capacitors are typically used in the picofarad to nanofarad range. Is there a general guideline for determining which order of magnitude to use within this broad range?

[@Pineresearch]

@@TarikZahzah-ns2bu First, you can look up your reference electrode impedance but probably better is just to measure it yourself. Use a Pt or Pd wire and the reference in some electrolyte, as a simple 2-electrode system, and do EIS mainly at high frequencies. The apparent iR drop from an experiment like that would be indicative of the reference electrode impedance (minus whatever small amount of Ω might be from the electrolyte, usually strong aqueous electrolyte only gives about 10-50 Ω at most). The reference electrode "experiencing" the frequency is exactly the point. It's just a little murky to me, and I'm not an electrical engineer so the precise details are a bit unclear to me. The potentiostat applies the sinusoidal potential perturbation between working and reference, so for sure it must be "experiencing" the sine waves I suppose. But the measurement of |Z| implies current and potential, and like I said, while a very tiny amount of current does pass through the reference (it is technically and electrically necessary for a potentiostat to function, despite the common assumption that no current passes at all) it is not the primary circuit path (current goes between working and counter). So that's why I don't totally know if the idea that the impedance of the reference (or the shunt capacitor, for that matter) is "felt" to the extent that the critical value like I mentioned is exactly where that transition occurs. I think it might be, I am just being honest that I may be slightly wrong for reasons I don't totally understand. Regarding the value of the capacitor, it follows from the previous paragraph and assumption I just laid out. Again, assuming this understanding is accurate despite my misgivings, you would simply need to look at a few things. At exactly which frequencies in your raw data are the features off, or you suspect may be artifacts? What are the measured impedances at those frequencies? Draw the Bode plot for a capacitor where you can change the value of C, and see where the intersection is, at which frequency and |Z|, depending on the value of C, with your raw data. This is hopefully where you might be able to narrow down what value of capacity will give you the most ideal intersection point for that transition between shunt and reference.

[@TarikZahzah-ns2bu]

@@Pineresearch In preparation for measuring the impedance of various reference electrodes, I would like to ensure I have the correct procedure for this experiment. If I understand correctly, I should connect the reference electrode I wish to measure as the working electrode (WE), while using a Pt wire as the counter electrode/reference electrode (CE/RE). Essentially, both the CE and RE cables are connected to the Pt wire. Subsequently, I would perform an EIS measurement, primarily focusing on high frequencies. What range of high frequencies would you recommend? Would a range between 100 kHz and 10 kHz be suitable? I do want to note that certain reference electrodes, such as the HydroFlex (RHE), may not be suitable for use as the WE, as they could be susceptible to damage.

[@Pineresearch]

@@TarikZahzah-ns2bu Exactly as you described it, yes. That is how you would measure the impedance of your reference electrode. Go from about 100 kHz to 1 kHz, and use a very small potential amplitude, like 5 mV or less. You could start with 1 mV and see if you get any data. Slowly increase the amplitude little by little until the EIS data is consistent and not too noisy. You will certainly damage the reference electrode if you try to push too much current through it, but that is why you want to limit the amplitude so minimal current passes during the EIS test. It is also a few fast frequencies (you can limit to like 4-5 points per decade) so the duration of current being passed should be extraordinarily short.

[@riya6688]

Very informative video; is it also possible to do one such detailed video on effect on EIS with a two electrode system where reference wire is shunted with the counter electrode.

[@Pineresearch]

Thank you. Well if you have a two electrode system, you probably won't need a shunt for a two electrode system with EIS.

[@TarikZahzah-ns2bu]

I have a highly practical question. When it comes to capacitors, would you suggest using a non-polarized capacitor or a polarized (electrolytic) capacitor? If a polarized capacitor is recommended, which terminal should be connected in series to the conductive metal wire (shunt), and which terminal should be connected in parallel to the reference electrode? I would greatly appreciate your advice since I lack expertise in electrical engineering.

[@Pineresearch]

To be honest, I am not an electrical engineer either and don't know for certain. My intuition is that non-polarized capacitors would be a better choice, in part because EIS fundamentally involves oscillating signals, and it's probably better to not have to worry about + and - leads.

[@TarikZahzah-ns2bu]

@@Pineresearch I agree. Based on my experience, polarized capacitors are commonly employed due to their compact size and affordability, even for larger capacitance values. On the other hand, non-polarized capacitors, typically made of ceramics, tend to be more expensive. Moreover, when a significant capacitance value is required, non-polarized capacitors tend to become physically larger, which can be impractical for your electrode system.

[@TarikZahzah-ns2bu]

I have a basic inquiry regarding the occurrence of artifacts in the high-frequency range specifically when utilizing reference electrodes with high impedance values, typically introduced by their frit. While a frit can be viewed as a frequency-independent pure resistor, I am curious as to why these artifacts are observed primarily in the high-frequency range. Could this phenomenon be attributed to the reaction kinetics of the reference electrode?

[@Pineresearch]

Hello Tarik, I'm getting some information from my colleagues about reference electrode shunts and as I dive deeper into the subject I'm finding it to get progressively more complicated. In general though, I think you are correct, the reference electrode frit acts as a resistor. However, it's reported that reference electrodes like Ag/AgCl and SCE have a capacitance as well. It's a very low capacitance ~ 7-10 pF, which makes me think it's the capacitance of the glass body of the reference electrode. At high frequencies, this capacitor adds phase delay and attributes to the high-frequency artifacts you observe. It might be possible that reference electrode kinetics also play a role, but I don't know about that. That would be a very interesting experiment to do.

[@TarikZahzah-ns2bu]

@@Pineresearch That's fascinating insight! Could you please provide me with a reference that delves further into this topic? However, I have a critical comment regarding this theory. Based on my understanding, since the impedance of a capacitor is inversely proportional to both capacitance and frequency, I would anticipate that the capacitance of the glass body primarily poses issues at mid- to low frequencies, rather than high frequencies. At high frequencies, the significant number of hertz would somewhat offset the relatively low capacitance value (around 7-10 picofarads) during multiplication, resulting in a moderate impedance for the reference electrode. Conversely, at mid- to low frequencies, when you multiply two small numbers together, the product becomes even smaller. As a result, due to the inverse relationship in the denominator, it leads to an incredibly large impedance value. I'm curious to hear your explanation for this phenomenon.

[@Pineresearch]

@@TarikZahzah-ns2bu You are fundamentally thinking of the impact of reference electrode capacitance backwards. Because its impedance increases as the frequency drops, it has LESS impact, not more. A large impedance means it represents a bypassed circuit, so effectively it is ignored, not present. I was out last week when Alex was handling some of these questions, and we've had internal discussions as he brought me up to speed. I am going to cover some of our conclusions a delve a little more deeply into these phenomena during today's Livestream (will be episode #14). I encourage you to watch it, or watch the replay.

[@TarikZahzah-ns2bu]

@@Pineresearch Ah, you're absolutely right! I didn't have the complete picture in mind. The electrical circuit of a glass reference electrode operates in parallel, which means the potentiostat can either probe via the frit (resistance) as intended or through the glass (capacitance). After rewatching the Livestream, everything became crystal clear! It's evident that at high frequencies, the impedance of the glass path is lower than that of the frit, resulting in erroneous data. However, at mid- to low frequencies, the impedance of the glass capacitor increases, effectively bypassing it. As a result, the potentiostat can probe through the frit as usual.

[@joseRodriguez-zi9cy]

Hello, Good morning, I appreciate the knowledge shared on this channel. Studying about the EIS technique I have a doubt that I have not yet been able to solve clearly, the question is What does it mean that a material has its phase angle at lower or higher frequencies than another material? In which of these cases would there be a higher resistance to corrosion?

[@Pineresearch]

Hello Jose, thanks for your question. First of all, I want to clarify that when you write "...material has its phase angle..." I assume you are referring to a kind of graphical dip in the phase angle on the Bode plot that is observed for features that are often fitted with a Randles circuit. If this is the case, then the simplest answer I can give you (neglecting all other processes that might be occurring on your materials of interest) is that the phase angle peak is related to the capacitance and the time constant. To elaborate just a little bit: if the capacitive effects being observed are very small, it means it can be charged/discharged very rapidly, meaning the peak will manifest at higher frequencies. Conversely, if the capacitive effects are very large, it will take longer, or be slower, for the charge/discharge phenomena to occur, meaning the phase peak will appear at lower frequencies. All of the above is also neglecting the resistance, which can also have an effect on where the phase angle peak appears; and on the time constant, which is a pseudo-measurement of how long this charge/discharge process takes. Finally, to address your question about corrosion resistance: the position of the peak *may* give insight into the corrosion resistance, but not necessarily. For example, following the previous discussion on capacitance: if the phase angle peak is at high frequency, it likely implies lower capacitance and a somewhat smaller time constant, which could also imply a smaller resistance. But it is not guaranteed that is the case. You can have a large resistance even with a small time constant if the capacitance is just extremely small. Conversely, a phase angle peak at low frequency might imply higher capacitance and a larger time constant, which could imply a larger resistance; but again, this is not guaranteed because you could still have a small resistance even with a large time constant. This is because the time constant is equal to R times C, so a proper circuit fitting analysis would likely be required to reveal whether your material is exhibiting high or low corrosion resistance.

[@nuomitang30]

Nice illustration 👍

[@Pineresearch]

Glad you liked it!

[@gangster-tz5xe]

Why does the impedance of the reference electrode affect the changes at all? After all, no current flows through the reference electrode

[@Pineresearch]

So first, to be clear, it is not accurate that 0 current flows through the reference electrode. Technically, a very small amount of current does flow through the reference electrode. In fact, if it didn't at all the potentiostat would not function properly. However, this is not likely the most precise answer to your question. I just want to point out this distinction. Reference electrode impedance impacts results primarily at high frequency. This is to say that many DC experiments, which typically run at low frequencies, are not as negatively impacted by reference electrode impedance. As shown in this video though, EIS at higher frequencies is where you typically observe this problem. With respect to the reference electrode, since impedance is fundamentally a function of potential, increased reference electrode impedance can affect the potentiostat's ability to sense the potential at the reference. Again, this is also a function of frequency, but keep in mind the potentiostat's ability to apply the potential you want it to at the working electrode is predicated on its ability to quickly and accurately sense at the reference, so if that is heavily impeded there can be measurement errors. If you want further discussion and clarification of this, you might consider joining us on Fridays at 1pm EST for a livestream, and I can talk about this live if you prefer.

[@gangster-tz5xe]

I'm sorry, but I'm not sure I'll be able to attend the live broadcast on Friday. I would appreciate it if you would answer this question in detail on the broadcast so that you can watch the recording. I understand that the operational amplifiers in the potentiostat have a finite resistance and, therefore, the current through the reference electrode is not zero. If I understand correctly, problems due to the reference electrode can occur when it does not keep the potential constant. This means that in this case, at a high frequency, the electrode does not keep the potential constant. It is known that a change in the potential of the electrode can occur with an increase in the current flowing through it. In this case, given that the impedance hodographs are distorted at high frequencies, it turns out that when working at high frequency, a larger current flows through the reference electrode, as a result of which its potential changes, and we get distorted data. As I said, I understand that a non-zero current flows through the reference electrode, but I do not understand why the magnitude of this current changes at a high frequency. I can assume that the electrometer (which measures the voltage between WE and RE) has a parasitic capacitance in the potentiostat, which leads to an increase in the current flowing through RE when the frequency of the voltage between WE and RE increases.

[@Pineresearch]

@@gangster-tz5xe I have added your question to the comments that I will address this afternoon during our Ask Us Anything About Electrochemistry livestream. For reference, you can watch later and it will be Episode #42. I also want to apologize for possibly confusing the point of your question with my initial response. Let me try to be clear about one thing, and then certainly I will try to elaborate during the livestream later. The point I made about the current through the reference electrode being non-zero is technically true. However, the bulk of your explanation in your last comment is not exactly accurate or related to the spirit of your initial question. I am sorry because I think by pointing out this non-zero current, I have opened a can of worms, so to speak. The reality is this is an extremely small current, about pA or smaller, and it is simply a technical necessity required for the instrument to measure properly. The tiny current that passes to the reference electrode in an experiment from the potentiostat almost certainly has zero impact on the reference potential. It does not affect this at all. Again, I think your thought process went down a road of assuming that the changes in impedance by frequency are dramatically impacting a current flowing through the reference electrode, thus impacting its set potential. This is simply not what is happening. Again, I will try to clarify this later live. Please join if you can, or feel free to watch the replay afterwards.

[@gangster-tz5xe]

Thanks for the future discussion on the broadcast. Mainly, I want to understand where the distortion comes from at a high frequency. What exactly does the high frequency do to cause distortion associated with the reference electrode. Why is there no distortion at a low frequency, but at a high one? What happens to the reference electrode at high frequency?

[@Pineresearch]

@@gangster-tz5xe Thanks for the follow up and again, I will draw and explain this with some more detail today. In short, the answer is actually related to capacitance through the reference electrode glass wall.

[@kanishksingh5463]

great fan of your videos..would you please make a video Diffrential pulse voltametry and square wave voltamertry...it would be highly useful

[@Pineresearch]

We've definitely been thinking about making some videos based on pulse voltammetry techniques. Although, I'm thinking it will be some time before I'll be able to make one. But please stay tuned for more.

[@himazafar4984]

can you also make a detailed video on EIS of the solution, working as a capacitor, and defusion model out of EIS measurement, like if we do some alteration in the solution, how our EIS measurement can help to indicate that, and how to make right fitting curve model for it? @pine

[@Pineresearch]

We do have a couple videos from our Advanced EIS webinar series kzbin.info/www/bejne/mpOYqa2rga-BarM and kzbin.info/www/bejne/rZrboWamgNt-g8U that go into more details on how to model the electrochemical system. In general, the solution itself is typically modeled as a resistor. Changes to the electrolyte simply change the resistance of the solution. Of course, when we talk about concentration gradients from the diffuse layer, that is a slightly different story. But I would imagine that if the electrolyte changes, it will change the capacitance at your electrode interface. I hope this was helpful.

[@TarikZahzah-ns2bu]

Apologies for the influx of questions regarding this specific topic, but several inquiries have recently arisen, both of a fundamental and practical nature. Here's another fundamental question I have: How does the potentiostat ensure a consistent application of the same (alternating) potential to the working electrode, given that the presence of the shunt alters the nature of the reference electrode when transitioning from high to mid/low frequencies? For instance, let's consider a system comprising an Ag/AgCl reference electrode and a Pd wire (+ capacitor) as the shunt. The desired potential for conducting EIS, set in the potentiostat software, is relative to Ag/AgCl. However, for impedance measurements at high frequencies, the applied potential now becomes relative to Pd/PdHx (palladium hydride) as the shunt acts as a pseudoreference. Subsequently, as the frequency shifts to mid- to low-range, the potentiostat reverts to applying the potential relative to Ag/AgCl as originally intended. Does this imply that the potentiostat briefly applies significantly different potentials to the working electrode in the high-frequency range compared to the mid- to low-frequency range? If not, how does the potentiostat address this apparent concern?

[@Pineresearch]

Hey Tarik, this is a very good question, but also fairly complex. I will try to address it during the livestream today because I know a little bit about how potentiostats work and while I might not be the EIS expert I might be able to share some insight about shunts and what a potentiostat is doing.