Making an electrode for a conductivity meter

Hello,

Definition : overvoltage is the difference between the cell voltage at a given current and the cell voltage at zero current after taking off the ohmic drop. i.e. : Ucell (I) = Ucell (I=0) + summ(R*I) + overvoltage. Overvoltage comes from two sources :

• a kinetic source : it then corresponds to the additional energy (i.e. an additional voltage i.e. an overvoltage) needed to activate the electrochemical reactions.

• a mass transfer limitation source. In solid, current is carried by electrons but they can't "swim". So, in solutions, current is carried by the ions. At the interfaces between solutions and solid electrodes, the current continuity is provided by the electrochemical reactions. Closed to the interfaces, the ions mass transfer can only be provided by diffusion process. When applied current is low enough, the flow rate of ions diffusion near the interface can sustain the rythm and there is no mass transfer limitations. When applied current is high, then diffusion flow rate becomes limiting. Since it is directly linked to the concentration gradient between bulk and electrode/solution interface, the reactant concentration at the interface decreases, until the diffusion flow rate fits with the applied current. This phenomena manifests by an additional voltage (i.e. an overvoltage), which remains low as far as the reactant concentration at the electrode/solution interface remains sufficiently high. When the applied current is very high, the latter concentration becomes closed to 0, the overvoltage raise up quickly to tend towards infinity (actually, until you reach the potential where an another electrochemical reaction occurs, at most solvent decomposition potential) because the electrochemical reaction can't go faster than the diffusion flow rate of the reactant. This means that the diffusion (mass transfer) limits the current you can put on your system.

To complete the explanation, the evolution of overvoltage vs current is not linear. It follows the Butler-Volmer relation, which involves the ratio of reactant concentrations between bulk and interface, and which has an exponential relation of the current vs overvoltage.

Best regards.

--
Enlevez ".nospam" de mon adresse e-mail pour me répondre.
--------------------------------------------------------------------------------------------------------
Remove ".nospam" from my email address to reply me.

"John Woodgate" schreef in bericht news:BZuFQ0A$ snipped-for-privacy@jmwa.demon.co.uk...

If you wind your wire around the toroid in the obvious way - starting at 0 degrees and adding turns until you get around to 360 degrees - you've made a loop around the toroid.

Rayner and Kibble's "Coaxial AC Bridges" (ISBN 0-85274-389-0 and available from NPL

discusses this at length in section 4.2.1.

They list a couple of ways of making the winding non-progressive or "astatic", starting with taking the end of the coil back around the toroid to the start, which roughly cancels the loop, proceeding through the Ayrton-Perry or "bootlace" technique where you put on half the turns going clock-wise around the toroid, and the remainder back-tracking anti-clockwise over the top of the first lot, to a more complicated variant where you wind the first 25% of the turns over 180 degrees of the toroid, going around the toroid clockwise, then wind the next 50% over the full 360 degrees, going back around the toroid anti-clock wise, then reverse again to wind the last

25% over 180 degrees, going around the toroid clockwise back to your starting point.

The last variant gives you a coil with less self-capacitance.

I posted most of this on s.e.d. on Mar 12 2003, at 1:32 am, in the thread "Conductivity meter probe questions".

-------- Bill Sloman, Nijmegen

I read in sci.electronics.design that Bill Sloman wrote (in ) about 'Making an electrode for a conductivity meter', on Thu, 30 Dec 2004:

Thank you for the information.

--
Regards, John Woodgate, OOO - Own Opinions Only. 
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

Which one will be the positive, and which one negative?

John

Um, the overvoltage of platinized platinum is VERY LOW, not zero. Much lower than anything else common.

--
Many thanks,

Don Lancaster
Synergetics   3860 West First Street  Box 809  Thatcher, AZ 85552
voice: (928)428-4073 email: don@tinaja.com   

Please visit my GURU's LAIR web site at http://www.tinaja.com

A gross simplification: Platinum can't be used to make a battery.

Copper, tin, zinc when in (ordinary) water will. Stick two pieces of copper into wet ground and you will have a voltage between them. Stick one zinc and one copper and you can light a light bulb.

--
Nicholas O. Lindan, Cleveland, Ohio
Consulting Engineer:  Electronics; Informatics; Photonics.
Remove spaces etc. to reply: n o lindan at net com dot com
psst.. want to buy an f-stop timer? nolindan.com/da/fstop/

I read in sci.electronics.design that John Larkin wrote (in ) about 'Making an electrode for a conductivity meter', on Thu,

30 Dec 2004:

The one in the patch of ground with the lower pH is positive.

--
Regards, John Woodgate, OOO - Own Opinions Only. 
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

I read in sci.electronics.design that Rich Grise wrote (in ) about 'Making an electrode for a conductivity meter', on Fri, 31 Dec 2004:

Roughly the same, in a cell, as zinc-carbon, but far too costly of course.

The problem with using carbon electrodes in a conductivity cell, AIUI, is that carbon is very prone to gain a thin layer of gas (hydrogen and/or oxygen in water) on its surface, creating significant over- voltages. Remember 'polarization' in the zinc-carbon cell and the use of manganese dioxide depolarizer to remove the gas?

--
Regards, John Woodgate, OOO - Own Opinions Only. 
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

Is this the same thing as the Warburg impedance?

Thanks, Nick J.

No, this is dc current. The Warburg impedance is the mix of diffusion and electron transfer under ac signal conditions.

As for this conductivity meter, the electrochemical reaction plays no role at all, so all the talk about Pt having a low overvoltage etc means nothing. Pt is used because it does not dissolve. I would not use Au, having seen it peel off as oxide layers under a large ac signal. The platinisation increases surface area, and thus capacitance. It is the double layer capacity that dominates here, being of the order of 20-200 microFarad per cm^2. At sufficiently high frequency (probably a few kHz are used in commercial meters), the impedance of the Pt/solution interface is negligible compared to the solution resistance. Any electrochemical reaction taking place at the same time will have an impedance much higher than the capacitive one, and not come into play. E.g. at about 1600 Hz and a capacity of 100 uF, you have an impedance of 1 ohm, which is much less than the solution resistance you are measuring. Any electrochemical reaction taking place would for one thing have a much larger impedance, and for another in any case, reduce that figure even more because it is in parallel with the capacity.

I would not, however, build the meter myself. OK, you can get the Pt foil from a chemical supplier, but you then have to mount the two foils firmly so that the cell geometry is absolutely fixed. Even then, you still need to set up the electronics, calibration etc. Try to borrow a commercial conductivity meter. I say all this as an old hand at building my own gear, years ago. That was a time when you couldn't buy the stuff, like amplifiers, oscillators and potentiostats. I even built a multiplier once, using light- sensitive resistors and LED's. I had great fun doing all that, but today, it's easier to buy it all. Well, almost all. I do put potentiostats together from (bought) power op-amps.

--
Dieter Britz,   Kemisk Institut, Aarhus Universitet, Danmark.

Hello everyone and Happy New Year,

Roughly yes. Actually Warburg impedance is what you get in an impedance measurement when system is mass transfer limited : it's the mass transfer limited system's effect on impedance measurement. Best regards.

Enlevez ".nospam" de mon adresse e-mail pour me répondre.

-------------------------------------------------------------------------------------------------------- Remove ".nospam" from my email address to reply me.

Hello everyone,

Impedance spectroscopy segregates phenomenas according to their respective time constants. Usually, charge transfer process is much faster than mass transfer process, and the the former will appear on the impedance spectra at higher frequencies than the latter. Warbug impedance is characteristic of the mass transfer-limited system. It does not appear otherwise. This not means that the system is not a mix between both electron and mass transfer but each phenomenas are segregated...

Let's say it MUST NOT play any role. When you pass a current through an electrochemical cell, you always have the following cell voltage decomposition : Ecell = Ecell(I=0) + summ(R.I) + overvoltages. The purpose of a conductivity cell is to determine summ(R). And, for that, you must have : Ecell = summ(R) * Imeas. Ecell(I=0) = Eqa - Eqc. When both electrodes are identitical, then Eqa = Eqc. Now, to have a significant measurement of your summ(R), you must have : summ(R) >> overvoltages. And that's why Pt is used. If the purpose of using Pt in conductivity cells was only to avoid its dissolution, the use of back carbon would provide, for most applications convenient and much cheaper solution.

The increase in surface area also has an effect to decrease the charge transfer overvoltage, which is linked to the current density through Butler-Volmer equation.

I disagree... First because the charge transfer resistance is a partly consequence of the Helmoltz double layer capacitance build-up at the electrode/solution interface. Second because, when you make an EIS sepctra of an electrochemical system, the first loop (at high frequencies) corresponds to the charge transfer impedance : the capacitance corresponds to the Helmoltz double layer capacitance and the loop's diameter correspond to the charge transfer resistance. To continue the comparison between EIS and conductivity cell measurement, you can either work :

• at frequencies so high that the first capacitive loop is not yet formed. That's why high frequencies AC current is used for the conductivity measurement.

• with first capacitive loop as low as possible. That's why material with very low charge transfer resistance, such as platinized platinum is used. In commercial cells, both are used.

Best regards.

--
Enlevez ".nospam" de mon adresse e-mail pour me répondre.
--------------------------------------------------------------------------------------------------------
Remove ".nospam" from my email address to reply me.

No, both the voltage and frequency can vary over a wide range (as can the conductivity of water from very contaminated to very cure). One industrial- strength instrument for large scale dialysis installations which I did instrumentation design for has both parameters programmable. There's a bit of art to this stuff, and I suggest finding a genuine expert in this kind of instrumentation, locally in your own country, who is willing to help you pro bono.

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

Hi

I know industrial conductivity units use AC rather than DC, is there a standard frequency? If there is not then dependant on the dielectric constant of the material measured then there will be some repeatability issues. Even if the "cell" small, and has a single reactive entity, you can get significant roll-off say 3dB per decade. But most electrochem cells have at least 2 or if not 3 reactive entities thus having say 6 or even 9db roll-off. Dieter you are right, about not build my own electrodes but I will build my own bridge and connect it to a low cost PC data logger. On that note, can any one donate any conductivity electrodes to a poor student?

Cheers

Wayne

--------------------------------------------------------------------------------------------------------

at

That helps. I couldn't find "Warburg" listed in B&F's index (good book, poor index) but its described in the section on EIS (chap. 10). Fig. 10.1.14 shows Rct (charge transfer R) in series with Zw (Warburg Z, defined as a "resistance to mass transfer").

I can see how Pt would be especially good for EIS where very small amplitude signals are used (e.g. the MacDonald EIS book suggests < 20mV, in order to stay well below most redox potentials). For simple conductivity measurements done with older, analog-based instruments, I've read that platinum black helped avoid signal response attenuation due to polarization at higher conductivities. Some digital instruments have other ways of dealing with polarization effects (e.g. dynamically changing the drive freq to better match the conductivity range).

Of course if you've got EIS equipment then you've got the best conductivity meter of all (paraphrasing from MacDonald et al) so you can check your results. You should get the same bulk resistivity measurement at different frequencies if you're properly compensating for all the other effects that can cause inaccuracies in simple conductivity meter measurements.

That might be overkill for whatever Wayne had in mind. He might get good enough accuracy by simply using an analog output on a DAQ box to drive a 500 mVp (low enough to avoid reducing any dissolved oxygen) square wave (say

3Khz) across the cell and use a 1-chip I/V transimpedance amp and high sample rate analog input to integrate the signal's trailing 1/4 (thus rejecting the initial charging of the cell capacitance) and do full-wave rectification on it (to make any DC potentials cancel out) in software. Of course this would have to be calibrated against two standard solutions of different conductivities or by comparing to a commercial conductivity meter (maybe he could borrow one just to do a calibration). If he wants to get really fancy he could even make the software compensate for capacitive "droop" as described in the Beckman patent (US 4808930).

Depending on whats in his solution, stainless steel or titanim electrodes may give reasonable longevity between calibrations. Commercial probes often use these materials. A concentric arrangement is preferable, e.g. two tubes with the smaller one inside the bigger one so small relative movement between them tends to cancel out as far as affecting the cell constant.

Nick J.

a

No. The commercial conductiy meter we used as a reference could be switched between a couple of hundred Hz (for low conductivity solutions) to a couple of kHz for high conductivity solution (up to

200mS/cm, where it gave up).

The system we sold worked as conductivity to frequency converter, with a lower limits of 2200Hz for zero conductivity. The frequency rose linearly with conductivity up to about 20kHz at 2mS/cm, then went on to about 100kHz at 200kHz/cm - the response was non-linear, but reproducible.

material

Since the idea is to keep the capacitative impedance at the electrode surfaces negligible in comparison with the resistive impedance through the liquid, repeatability ought not to be a problem. We had more trouble with the effective surface area of our platinum-black covered electrodes, which initially tended to get beaten flat by the spray heads they were exposed to, though we eventually worked out how to prevent that.

get significant roll-off

not 3

Sure, but if the reactive impedances are kept negligibly small, this isn't an issue. If you go for a four terminal (Kelvin) bridge, the surface impedances can be even less of a problem, and for high conductivity solutions, this might be the way to go, if you can't use an electrode-free inductive system.

build my

Sensible, but the bridge is the trickier project ...

------- Bill Sloman, Nijmegen

Agreed. Conductivity meters are normally designed specifically to prevent (or to minimize to insignificance) faradaic processes at the electrodes. Thats why chemically inert electrodes (like SS or titanium which form a "self-healing" protective oxide layer) are often used. Thats not to say you can't measure conductivity in an electrochemical cell where faradaic processes are occurring but its a simpler problem if you don't have to worry about the electrodes reacting with the solution under test. Possibly Pt tends to foul less and last longer than other metals under a broader range of conditions? I can see Pt being good for EIS methods where very low voltages must be used in order to keep the electrochemical system operating in the linear region but its an expensive choice of conductivity probe material for a student on a limited budget.

Nick J.

Why would anyone design a conductivity meter with that a signal that large? Gold works just fine from a corrosion standpoint.

The commercial meter I designed used a 1 Hz square wave.

I used two octoganal plates of gold-plated stainless steel with PTFE spacers at all eight edges. If worked fine.

I am one of the fellows who designs the gear you buy. This thread is posted to sci.electronics.design, where there are electronics engineers, and to ci.chem.electrochem.battery, sci.chem.electrochem and sci.engr.chem, where there are chemists.

I know that my gold-plated electrodes worked well because I tested my design against three different brands of commercial conductivity meters under a wide range of conditions. I am still wondering if it would have worked better if I had specified platinum plating.

The one I designed ran DC one way for a half second or so, then DC the other way for a half second or so. This was for a low-cost, low-performace conductivity meter measuring deionized water only.