Quantify variance of pH meas, collaboration

[shoneill]

I don't wish to overly-labour the point, but I still see weak points in your experiment.

On a side-note I have studies electrochemical sensors in college and spent six years developing glass pH electrodes in industry in Switzerland.  I know a thing or two.

Errors in pH measurement are not at all similar to errors in "conventional" measurements.  Noise is almost always no clue whatsoever as to the accuracy of the measurement.

Things like jnction potential, stirring error, non-linearity of probes, blocked diaphragmas and so on can seriously deteriorate the measurement accuracy.  Also bear in mind that several pH electrodes can deliver very similar "False" measurements.  Without deeper knowledge of which version is right, it's difficult to seperate the good measurements from the bad measurements.

Please try to get your hands on the book I mentioned, it's the absolute number 1 reference for pH electrodes and pH measurements.  It explains why reproducibility, noise and accuracy have almost nothing to do with each other regarding pH measurements.

Shane.

[ShaunDustin]

Shane,

I appreciate you taking the time to respond.  A lot of us make pH measurements and trust the instrument without understanding all the issues.  I am certainly guilty of that, but in defense of all those of us that measure w/o understanding the tool all that well, there's not a lot of accessible information; that's what led me to this study.  It doesn't help much when electrodes are sold with no documentation other than a check box on the packaging that says  _x  pH  ___ ORP. I asked the Cole-Parmer rep if the 12 ORP electrodes I bought a few months ago were multifunction pH/ORP like the advertising inmplied; they said they were. When I got them they were ORP electrodes. My bad for being dumb and not knowing what I was asking for, but I think it shows how little understanding there is about this on a basic level that neither of us knew enough to know what we didn't know.

I will keep looking for that book.  Are there any other references you might suggest?  Would you mind enumerating some of the flaws you see so that I can make adjustments? 

As far as electrode type, I was planning to use an industrial electrode from Sensorex:

http://www.sensorex.com/products/ph_electrodes/process/flat/in_line.html

I chose it because it will be scalable and it's an accepted solution for full-scale process monitoring in the industry. I am open to suggestions on other types to check, but single junction and glass electrodes may not last in my application. I assume that this means I will be giving up some short term accuracy for long-term reliability.

My application will be steady state and constant temperature.  This limits my variables but it also may limit the applicability of the tool.

I would appreciate any further insights you might have on this.

Shaun

[shoneill]

Shaun,

no bother, I like the feeling that my time developing pH electrodes wasn't totally wasted.....

Yes, Sensorex.  I remember testing them in the lab a few years ago.  They work on an ISFET basis.  Nice idea in itself, not requiring a glass membrane, but their reference electrode junctions are really really bad. They have solidified gel electrolytes which cet diluted rather quickly, and ceramic pin diaphragms which get blocked really quickly.  All in all, the reference electrode simply isn't very good.

If I may do some advertising for an electrode I helped develop, look here. I would recommend the pH+. It also has a gellified reference electrolyte, but it has plenty of KCl crystallized within to help replenish anything which would otherwise dilute the electrolyte. In addition, it has a ground sleeve ring diaphragm which is simply miles ahead of ceramic pin diaphragms in any kind of bio application. It won't clog (well sometimes it will, but not until LONG afgter ceramic pin ones already have).

Errors you need to look out for are
1) Alkali error : non-linearity of the pH glass above pH 12. If autoclaving the electrode is neccessary, this can become a problem at pH 10 eventually as the glass deteriorates.

2) Diffusion potential : The liquid junction of the reference electrode should theoretically have zero potential drop. This is, of course, never actually the case but gellified reference systems (especially ceramic pin ones) can have junction potentials of up to 0.5 pH units! Ring type liquid junctions are much superior in this regard (0.2 pH max IIRC).

3) Stirring error : Often confused with the Diffusion potential, the difference in measured value when stirring or not stirring. In case of doubt, the value WITHOUT stirring is correct. Ring diaphragms are here again far superior. Ceramic - 0.2pH, Ring - 0.05 pH.

4) Increasing membrane impedance : If autoclaving glass pH electrodes, the impedance of the glass gradually rises. This can eventually become a problem above 1-2 Gigaohm membrane resistance. ISFETS don't have this problem, but they have lots of others (Long-time drift being one big one).

I think you'll find that the robustness difference between your linked Sensorex electrode and a proper Bio pH-electrode (as the pH+ linked above) is surprisingly small. Sensorex argues that HF has no affect on their electrodes. We tried this out once and it died after the same exposure to HF as the pH+ electrode. The fact that the body is made of glass-reinforced plastic doesn't help.

All in all, my experience has been that ISFETs are a nice idea, but the people developing them are electronics people who simply have no clue about the physics of pH measurement. The electrodes are designed accordingly. I think you'll be better off with a glass electrode, assuming you can sort out the impedance problem...... That is the only real advantage of an ISFET - low impedance - in the range of Kiloohms.

Shane.

[ShaunDustin]

What if I use an amplifier like this one:

http://www.omega.com/ppt/pptsc.asp?ref=PHTX21&Nav=gredd05

with the glass electrode.  Would that help to get a more useful signal with the DAQ limitations??

Maybe my application will not be that helpful to others; I am running an anaerobic process that is seeded and fed with wastewater--by definition not sterile, so autoclaving is not a concern.  The feedstock (labscale) is a defined medium, so I need to monitor it, but a variation from 6-8 is tollerable, so it doesn't need to be incredibly accurate.  I WOULD like higher resolution on the back side of the process--someone else wants to use my effluent in an algae reactor--so that might be  the control on my process. It sounds like I need to be very careful in stating the limitations of the study so that people don't get the wrong idea about what the data means.  This may not be a general purpose tool like I had hoped.

My process pH needs to stay between 7-8; if it drops below about 6.5 my process is on its way to pickling itself.  At 4 its all over but scrubbing out the reactor so I can start over again. To address the stirring question, since it's a continuous process, I was planning on pumping the standards through at the same flow rate the electrodes will see in the process.

Shaun

 

[shoneill]

Shaun,

That amplifier looks good.  If the specs are as good as they say, that could be a pretty good way of going about a pH measurement with Glass pH-Electrodes.  I'm glad you sent me that link, I may use it myself in the future.....

OTOH, More and more manufacturers are developing electrodes with built-in amplifiers.  I know E+H and Metrohm have them at least, even though Metrohm is for Lab use, not industry.

Maybe you might find others too.  The impedance is the only REAL disadvantage of the glass electrode versus ISFETs.

Shane.