Fluoride analysis in HF/Nitric (etch for titanium)

A. Hi Shakilahmad
We appended your inquiry to a thread which answers it, and it seems that an ion-specific electrode for HF is the best way to start, but for additional options you can search the site with the term "analyze HF Nitric"; we have at least a half dozen other threads on the subject if you don't like the initial answers you read :-)
Luck & Regards,


Ted Mooney, P.E. RET
Striving to live Aloha
finishing.com - Pine Beach, New Jersey

A. Go with the ISE. Most of the folks I know with HF Nitric pickles use this method.

? Does the Ion Selective Electrode method correlate with etch rates of the solution? The glass dissolving into the solution may interfere with the etch, and the fluoride content may not be the best may to control the solution. How about titrating with caustic to get an acid content, and determining the weight loss of a piece of glass for estimating the etching capability?

Tom,
I think you are right here. Let's go for the simple but right answer. Etch rate testing would be the best control as you are testing the function you wish to control. Also it is cheaper!

A. First, can you verify that the tank was at the exact liquid level as the last time the analysis sample was taken. If it was, then your analysis is off. Your test method is subject to an operator influence, even when it is the same chemist or technician. Temperature is the largest offender in this analysis. Next, the titrant can be added too fast and you get very tiny red brown specks floating around that you have to wait for them to dissolve. This is one of the least precise volumetric analysis that I ever came across. Try it on three samples from the same original grab sample. I think that you will be shocked at the variance. I gave up and went to a selective ion electrode. Do a search at this site. It has been discussed several times. My feeling is the titrant should also be stored in a dark cabinet. I had to restandardize it every month to meet the customers requirements for QC.

A. I think Mr. Watts thought that your question pertained to the fluoride titration. I use a different method for titrating nitric acid that I like much better. Since nitric acid is a stong acid and HF is a weak acid, it is possible to use sodium hydroxide to titrate nitric without titrating the HF. The trick is to end the titration at a pH that is low enough where no HF interferes. You will find that if you purchase ethyl violet indicator which has a very pretty endpoint at pH~2.4 (or just use a pH meter) that no HF interferes. At this pH, you have not yet neutralized all of the nitric, so you will need to make a set of standards and establish an experimental factor. Also spike some HF at various concentrations to prove to yourself that it does not interfere. Getting all of this done takes a little work, but I think that your results will be much better.

A. Each to his own. The HF is the important portion of the analysis and the nitric is vastly less important. If you titrate the nitric and have a =/- 2% and titrate total acid (or a continued titration) with a =/- 2% , it is possible to equal most of the HF concentration if you miss in opposite directions. The specific ion electrode was repeatable to under 1% of the total HF, regardless of how I did it. Since this is the important component, I strongly preferred it to the standard procedure for analysis of 3 major turbine engine manufacturers.

Now back to the original question. Their analysis (the recommended one)used a weak procedure. HF is not a weak acid by the definition that I remember from chemistry. However,it very easily can have an inflection point at a different pH than the nitric and still be a strong acid. Your procedure is faster, as it would have to be to protect a pH electrode from the fluoride attack.

A. Both respondents are correct but to answer the original question, I think you will find that the reason that the nitrate went up was because the titrant was not re-standardised. ferrous sulphate ferrous sulphate [affil links] oxidises in air quite easily and requires regular standardisation. Suggest you standardise against a nitrate solution so that you standardise your analysis method at the same time as standardising your ferrous solution.

One technique I have used to minimise this oxidation is to store the titrant in a flexible polythene container and expel the air by squeezing the bottle to expel air and then putting on a tight cap. I don't think the brown bottle has any effect.

Then add a few bits of iron powder to reduce the oxidised iron-- but you have to standardise regularly

A. I was not disagreeing with Mr. Watts that HF should be determined by ISE, nor do I disagree that the HF is the most important component. I was offering an alternative analysis for nitric acid that seemed to me to be Jennifer's difficulty. As far as HF being a weak acid, it is weak by definition that it is not fully dissociated in water into H+ anf F-. It exists partially as molecular HF in equilibrium. The equilibrium constant for HF is 6.8x10-4. This puts it between formic acid and acetic acid [on eBay or Amazon] on a list of common weak acids. Nitric, on the other hand is a strong acid meaning that it is fully dissociated in water into H+ and NO3-. It is this property that would allow for titration of the strong acid, at least partially, before the weak acid interferes. Because you can only titrate it partially by itself without titrating the mix, I was recommending that she establish an experimental factor. It would not be nearly so simple a matter to account for the HF using caustic, so ISE would be the way to go as I share a dislike for the standard alizarin method. I hope this clarifies things rather than making them more muddy.

A. Perhaps there is a simpler explanation to the anomaly of the levels. When the operator made the addition care was not taken in properly mixing the bath, so then upon collecting a sample the concentrations were for a lack of better terms out of whack after the analysis. Have the operator make another addition since the HF concentration has dropped by now, and emphasize the importance of mixing before collecting a sample. I'm certain this will provide you with a better idea of the true concentration of your nitric-HF bath.

A. Sampling is also a source of problems. I used a home built "colwissa" to sample the tanks from several different places and depths. I would not accept samples from the plating/cleaning folks because they were prone to NOT make sure that the tank was at the proper liquid level and stirred, as well as just dipping a beaker into the top 2 inches at the front of the tank. Congratulations on getting repeatable analyses. As to the original question, It is an absolute impossibility as stated. Not having a representative sample of the entire tank is the strongest probability. A properly mixed tank makes it a lot easier to get a representative sample.

A. Part of that will depend on which technology electrode that you are using. A conventional electrode will drift and should not be left in the sample any longer than it takes for the mV to drop to a point that it reverses and starts climbing. That will give you a very repeatable analyses. Use the valley reading. The standard and the sample also have to be diluted to a region where the TISAB can swamp out the background noises. Also needs to be in a linear range. I forget which TISAB I used, but it also made a difference.

A. What is the pH of your sample with TISAB? If below 4 pH the species in solution is a HF complex and not free fluoride ions. ISE's only can measure free Fluoride ions. HF is used to complex iron among other things forming fluoride complexes. Why are you doing the measurement? To see if available fluoride is still there for complexing or other? Adding TISAB typically is for samples where the pH needs to be brought down so hydroxyl ions do not interfere. With Nitric and Hydrofluoric ions around I don't think that's the issue. Are you diluting your sample with water? Why are you trying to make this measurement?

A. Use of ISE for HF is OK if there are no interferents such as iron. I analyse HF in pickling tanks and need to remove Fe, Cr, Mo as interferents by using a complexing agent. I hope this helps.

A. What I do is time consuming, but works. I use the method of additions, plotting apparent concentration vs. spiked concentrations. I evaluate the apparent concentrations by making a four point standard curve with a standard fluoride solution, the points being 50, 100, 200, and 400 ppm. The samples, of course, must be diluted down into this range. I don't use Orion's version of the method of additions, with their table of "q" values - quite frankly, I never understood how it worked, anyway :) Another alternative is to do a ferrous titration of the nitric acid in a matrix of sulfuric acid, at ice water temperatures. This also works, but uses vast amounts of acid, and a titrant that must be standardized each and every time you do the test. The "free" HF vs "total" HF question has never been raised by our customers (I work for a commercial lab) and I've never brought it up. Sometimes raising such questions creates problems that have more to do with salesmanship than science. Dave Wichern Consultant - The Bronx, New York A. You do not want to analyze any of the complexed fluoride. The only fluoride that does the work in your tank is the free fluoride. With ISE, it takes a known dilution of the acid sample to get it into the area where the mV response is linear.(MANDATORY) Look at the curve that came with the ISE. I made a Cal curve from 4 known samples of concentration vs mV and then tested the diluted sample. Enter the graph with mV that you read and you have a fairly quick and accurate analyses of the fluoride. I forgot which TISAB we used. The purpose of the TISAB is a swamping solution to eliminate the random noise, not to dissolve the fluoride complex. James Watts Navarre, Florida

A. I agree that what we want to measure is "free" or active fluoride. After all, the total fluoride hardly changes as titanium is etched, but the etch rate certainly decreases!My samples typically contain ca. 20 g/l free fluoride, and I use the simple method of diluting the sample 1:100 in 30 g/l NaOAc. Then I use ISE to measure the mv. Standards are made to bracket the unknown, and the rest is quite simple. The addition of ABF to raise the fluoride so as to show approx. 20 g/l in the bath always restores the etch rate to the nominal value, which in our case is approx. 3 mil/hr at ambient temps.

A. Tough problem! I've wrestled with it myself. Approached it three different ways:

1) Did total acidity via titration with NaOH; determined HNO3 conc. via titration with FeSO4 in chilled conc. H2SO4. Determined HF, HNO3 concentrations by arithmetic.

2) Did total acidity via titration with NaOH; determined F- concentration via electrode using the method of additions (the math gave me a headache, but it is possible) then determined HF, HNO3 concentrations by arithmetic.

3) Did total acidity via titration with NaOH; tried to do what you did; La(NO3)3 titration with electrode as endpoint.

Results: (1) and (2) gave totally different results. No useful endpoint observed with (3)

If I wanted a (4), I'd propose determining F via a distillation procedure. Then, there would be a third answer, likely different from the first two. Closer to (2) than (1), probably.

What's the "right" answer? HA!

My conclusion: what are we trying to measure? Once the bath is put in service, the HF starts to break down; the F is present as transition metal complexes with Fe and Ti. It is now less effective as an etchant...it's "unfree" HF.

My suggestion: think up an empirical test on samples that are as similar to the actual work as possible (if practical, use real work pieces) and do some kind of test to see how effective the bath is at doing what you want it to do. Cut the sample of the bath that is up on the line with fresh HF, and/or fresh HNO3, and do the same test on these solutions. (Watch phantom variables, like temperature; try to replicate process conditions)

Now, do your lab tests on your line bath; and, on your lab concoctions. Look for a correlation. Find one. Now you have your test.

Sound like a lot of work? It is. But if you follow through you'll solve the problem.

My motto: Science is a business of empiricism.

A. We've been working with thermometric titrimetry for the analysis of various acid mixtures, and have some rational approaches to the analysis of etching solutions containing nitric and hydrofluoric acids. (DISCLAIMER: I work with Metrohm International). In the following, the analyses were performed using a fluoride-resistant thermometric probe (really just a highly responsive, very sensitive thermometer). The analyses are genuine titrations, not calorimetric, and use the probe merely as an endpoint indicator. Endpoints are determined by the rate of change of temperature, not the actual solution temperature.

Some examples:
1. Analysis of HNO3, HF: HNO3 and HF are determined in a non-aqueous polar solvent, with aqueous 2 mol/L NaOH as titrant. HNO3 and HF are differentiated by an inflection in the temperature curve (due to differences in the heats of formation of HNO3 and HF).

2. Free HF can also be determined independently by a direct titration with Al(3+). The solution is strongly buffered with an acetate buffer composed of NaAc, KAc and HAc. The buffer also provides the excess of Na(+) and K(+) necessary for the exothermic reaction:

6F(-) + Al(3+) + Na(+) + 2K(+) -> NaK2AlF6

This yields precise results, with % rsd's <0.2%

Best regards

Tom