It all depends how you define "high purity water" In the absence of air contamination, the ion concentrations should all be below 10-6 M, and the pH will be between 6 and 8, and their should be no reason to know it exactly, because once it comes in contact with anything (metal surfaces) the pH will change as it becomes contaminated. For pure water open to the atmosphere, there is enough carbonate in the water that an accurate pH measurement can be made with a pH electrode, but you need a good one. For this purpose the key is to use an combination electrode (or reference electrode) that has a more porous frit. The typical glass frit (electrolyte junction between the outer filling solution and the aqueous sample) has a flow rate of 8 microliters per hour. Good pH probes (and it really is the reference probe), like Thermo-Orion's Ross electrodes (about $400 USD), have glass fiber junctions that have flow rates of 120 microliters per hour. Hence the total electochemical cell has better electrical properties, and a stable reading is achieved quickly. These combination pH probes have double junction reference probes. I think, even for systems with very low ion concentrations, getting one of these probes is the best way to go. Because flow of the outer filling solution into our sample occurs (and the rate mentioned above, it is best to use a large sample size, and discard the sample once the reading has been made, since it now has some contamination from the filling solution.

The measurement of pH in high purity water is tricky because its extremely low conductivity. Since pH measurement with a classic electrode is an electrical measurement, the presence of ions that can enable electron transfer between the measuring and reference electrode is crucial.

As explained in the attached document, the conductivity of the high purity water (ideally 0.055 µS.cm–1) can be measured more practically and ensures that the pH is close to 7.

You can also try a non-electrochemical approach, using indicato and absorbance measurements. You will have to prepare a series of solutions with certain pH at fixed concentration of the indicator, make analytical curve and then measure the absorbance at the analytical wavelength of the water with indicator (i.e. methyl organe).

It will help You too avoid issues linked to the low conductivity of high purity, deionized water, but requires a lot of work, unfortunately.

I think that you can try Spectrophotometric Determination.

Here are two references about concept and measurements. In fact, it is more accurate in sea water analysis than most pH meter.

http://nsm1.nsm.iup.edu/jford/courses/CHEM323/Readings/SpectroDeterm_pH.pdf

http://www.ocean-sci.net/7/597/2011/os-7-597-2011.pdf

Also try an optical pH probe. unlike some of the methods noted above, this can give you real time info just like a standard pH probe, but because it is optical, it can be used at much lower ionic strength. see for example: http://www.oceanoptics.com/Products/phbcgtrans.asp

It all depends how you define "high purity water" In the absence of air contamination, the ion concentrations should all be below 10-6 M, and the pH will be between 6 and 8, and their should be no reason to know it exactly, because once it comes in contact with anything (metal surfaces) the pH will change as it becomes contaminated. For pure water open to the atmosphere, there is enough carbonate in the water that an accurate pH measurement can be made with a pH electrode, but you need a good one. For this purpose the key is to use an combination electrode (or reference electrode) that has a more porous frit. The typical glass frit (electrolyte junction between the outer filling solution and the aqueous sample) has a flow rate of 8 microliters per hour. Good pH probes (and it really is the reference probe), like Thermo-Orion's Ross electrodes (about $400 USD), have glass fiber junctions that have flow rates of 120 microliters per hour. Hence the total electochemical cell has better electrical properties, and a stable reading is achieved quickly. These combination pH probes have double junction reference probes. I think, even for systems with very low ion concentrations, getting one of these probes is the best way to go. Because flow of the outer filling solution into our sample occurs (and the rate mentioned above, it is best to use a large sample size, and discard the sample once the reading has been made, since it now has some contamination from the filling solution.

Dear Wojciech and Xiaolan

Spectrophotometric determination of pH using indicators is no option in this case.

As indicated by Chad, the ion concentrations in ultra-pure water are all below 10^-6 moles per liter. A pH sensitive colour indicator is a weak acid or a weak base that changes its colour with its protonation state. Just by adding it you are dramatically altering the pH, unless you would add a concentration lower than 10^-8 moles per liter, which would not make sense, because you would be on top or below the detection limit of your spectrophotometer.

Maybe I'm right or wrong, but if I had such a problem I would try such an approach. There is a great variety of indicators so one can adjust the molecule to Your apprioriate pH. I would check it, it wouldn't take ages.

Titrate with a weak acid or a weak base

As I understand, pH is always trouble indicators. It is easily measured, but not acculate unless in some buffer system. In most cases, Co2 play some role for water, espeacilly in the "ultra-pure water " system. For the spectrophicatric method, the indicator conc should not be any problem, you can use the 10 cm , even 50 or 100 cm -path-length spectrophotometer cuvette. The point should the choice of indicator. Good luck!

Dear Breynaert,

If the concentration will be approx about 10^-8 and You put i.e. 10^-6 of the indicator which has two forms with two different analytical wavelengths, the concentration at proper wavelength will be a linear function of absorption (A=ecl, Lambert - Beer and so on). The only issue, as You said, is detection threeshold. If Parashuram would adjust the indicator molecule with very strong absorption coeffictient (e) 10^-8 seems not to be a problem. He can also use other organic molecules which tautomers will differ according to the protonation / deprotonation at certain pH, what will strongly influence the electronic structure of the molecule and their absorbance UV-Vis spectra as the results. There is a lot of data about spectroscopic estimations of the concentrations and I'm not an expert in this field, however it should be checked empirically :-) Nobody knows everything. I'd be so convinced about estimation of pH of ultra-pure water, despite I'm electrochemist.

We can make a lot of discussion, but without experimental testing we won't be sure whose approach is better in this particular case. Only experiment can judge this not "likes" and "dislikes" on RG :-)

The pH for pure deionised water will drift mostly due to the amount of carbon dioxide that is dissolves into it. The longer it is left exposed to the lab environment the closer it will get to be somewhere near that of pH 5.4; the pH of unpolluted rain water.

I hope the suggestions given by Dr. Chard may solve the purpose. Titration approach as suggested by Dr. Mohamad is fully ruled out because at such low ionic solution where pH is very near to 7 titration by weak acid or weak base is fully ruled out.

I think you should try a conductometric method since ionic conductivity of ultrapure water is practically zero

Kind attention to Dr. Bayram: Will you please elaborate how to proceed with conductometric method for pH determination? Do you mean the use of plot conductivity and pH or any thing else?

We use pH sensors specifically designed for ultrapure water (UPW). They use a flowing reference electrode along with the pH probe and temperature probe to measure the pH of 18 MOhm.cm UPW. The manufacturer of our sensors is Yokogawa but there should be other manufacturers that make UPW pH sensors. The problem with the average pH sensor is the extremely low conductivity of UPW that doesn't allow them to work properly, likely the problem you are having.

Kind attention to Dr. David: Thank you for sharing your experience. You mentioned that Yokogawa made electrode solved the problem. Will you kindly provide the product specification and source of availability of such electrodes? I tried with a specific brand of Metrohm electrode which is specilly meant for high purity water. I am not satisfied with the product. Will you please mention what should be the design criteria for pH electrode suitable for high purity water?

High purity water has pH = 7 (18.2 MOhm (25°C)). As soon as you start to measure pH your water will not be high purity anymore. Low conductivity water should be measured in a flow trough mode. The water to be measured must be prevented from ambient CO2 as well from dust and aerosols.

The reply of Markus brings us back to the root of the discussion.

The conductivity of the high purity water is ideally 0.055 µS.cm–1. All attempts to measure pH with glass electrodes, indicators, titrations, etc. severely disturb the properties of the ultra-pure water.

To demonstrate this, i modeled the specific conductance of extremely dilute concentrations of KCl in pure water using phreeqc.

A concentration of 2e-8 mol/l of KCl already increases the specific conductance of the solution to 0.058 µS.cm-1. If you would measure the pH of 1 liter of ultrapure water, with an electrode leaking 10^-9 liter of KCl in this solution during the measurement, the specific conductance would hence already go up significantly.

One of the only useful techniques to verify the quality and hence the pH of ultra-pure water is to measure the conductivity in flow through mode as Markus indicates. This type of measurement can be done without exposure of the solution to the open atmosphere and using 'inert' electrode materials. Most manufacturers of equipment producing ultrapure water include such a flow-through conductivity cell in their device to allow verifying the quality of the water produced.

I include the contents of a phreeqc input file that generates the attached image below.

More info on how to estimate the specific conductance of solutions using phreeqc can be found on this page: "http://www.hydrochemistry.eu/exmpls/sc.html". Implementations of phreeqc can be downloaded from USGS ("http://wwwbrr.cr.usgs.gov/projects/GWC_coupled/phreeqc/")

####

Database "path_to_Phreeqc.dat"

USER_GRAPH

-axis_titles "conc K+ [mol/l]" "Specific conductance µS/Cm" ""

-chart_title "Specific conductance KCl diluted in ultrapure water at 25°C"

-initial_solutions true

-start

10 graph_x TOT("K")

20 graph_y SC

-end

END

solution 1

reaction 1

KCl 1

1e-6 moles in 50 steps

end

######

Parashuram, I don't know if Yokogawa still makes UPW pH analyzers but perhaps you can look at Mettler Toledo's website and their pHure line of sensors. Good luck!

I have already tried with Mettler Toledo's pH electrodes. I arranged a well planned experimental setup and in inert atmosphere (N2 environment) measurements were taken. Results are not satisfactory. Moreover within 8-9 months the electrodes started to generate unstable readings though adequate precautions were taken to maintain the electrodes.

Dear Sir,

I have seen the outlines of Dr.Chad Jafvert. I feel the same.

Thank you sir.