poor phosphate peaks-Ion chromatography with suppression

[mreardon]

I am very new to Ion Chromatography so feel free to say the obvious.
I am trying to determine why i have been getting very broad phosphate peaks and no peaks when I should be getting some. I was running a dionex ics 1000 with an as-18 column and guard column and I was using 39 mM NaOH eluent (isocratic runs only). I couldn't get anything below 50 ug/L PO4-P to show up on my chromatographs. The peak I got for 50 ug/L PO4-P had a width of about two minutes. I have tried remaking the eluent and rehydrating the suppressor to no avail. the column is new so I don't think it's had time to be contaminated with metals. the minimum detection limits according to a dionex application note was 5.6 ug/L. They used an anion trap column though. Basically I'm trying to do the epa method 300.0 for a stream. Does anyone have any experience with this, helpful hints, suggestions for readings etc? I would also like to now how to figure out the the capacity of my column and how to figure out what concentration of analytes I can run and how many so that i don't overload the column. the test chromatogram with in the column's literature had a 10mg/L PO4 peak but I have been told not to run more than 1mg/L on this column. Any help would be greatly appreciated. thanks.

matt

[Chris Pohl]

Matt,

Occasionally, this phenomenon is reported with this column. Most likely the cause of your problem is related to iron. The first order of business is to check the label specification on your sodium hydroxide source. Generally, the preferred source of sodium hydroxide is 50% sodium hydroxide from Fisher. But whatever the source, the iron content of your sodium hydroxide must be equal to or < 3 parts per million (based on a 50% solution). Whatever you do, don't use sodium hydroxide pellets to make your eluent since the outer surface of pellets it is invariably coated with carbonate which will contaminate your eluent.

Assuming that the problem is indeed from your sodium hydroxide, you will need to decontaminate your column and suppressor to remove the iron oxide precipitate. The most effective way to do so is to pump a solution of 0.2 molar oxalic acid through the column and separately pass the same solution through the suppressor (passing it through both the eluent and regenerant chambers). Dissolution of iron oxide is rather slow, so it's best to pass the solution through the column or suppressor overnight to give it enough time to completely dissolve the iron oxide. If possible, it's best to use something other than your analytical system to pump this solution through the column or suppressor. You can use a syringe to accomplish treatment of the latter since the cleaning solution doesn't need to be flowing continuously to remove all the iron. It should be sufficient to rinse it with a few milliters of oxalic acid and then let it sit overnight and rinse it with a few more milliliters of oxalic acid in the morning. The column, on the other hand, will need a pump given the column's back pressure. If you don't have a separate pump available, then you need to make sure that you have completely rinsed all of the oxalic acid out of your system before switching back to sodium hydroxide eluent. The easiest way to check this is to directly connect a line from the injection valve to the conductivity cell and rinse the system with deionized water until the conductivity drops below 2microS before switching back to eluent and reconnecting the column and suppressor.

A couple of final questions: have you injected any real samples yet? If so, do any of your samples contain iron which has been added as a water treatment chemical? This might also be a cause of this problem. There are solutions to this problem as well but from your original description, it doesn't sound like you've injected that many samples so most likely this isn't the cause of your problem.

[mreardon]

Chris:
do you have any other ideas? the iron in the eluent is less than 1 ppm.
i have injected about ten samples of river water in the column, there shouldn't be iron in them. also the phosphate peaks come out around 16 minutes much later than the test chromatograph for the as18 with 39mM eluent. I may try flushing the column just with 10X eluent first. let me know what you think.

Matt

[Chris Pohl]

Matt,

What is your operating temperature? The test chromatogram is generated at 30C. Temperature has a significant effect on retention time with the effect being most pronounced on higher valency ion such as phosphate, so you need to be at 30C if you want to reproduce the test chromatogram.

Concentration is also an important parameter for phosphate retention since retention of phosphate changes as a cube function of the eluent concentration with hydroxide. Your problem could be that your eluent concentration is too low. Are you using 50% hydroxide as your reagent stock to make up your eluent? What method are you using for eluent preparation?

[mreardon]

chris:
i remade the eluent (39 mM NaOH) and now the phosphate peak comes off much earlier. I must have made the earlier eluent wrong (too dilute). I had the phosphate and nitrate peaks coming off at the same time. I had the column at 30 degree C with the new eluent and then I noticed that the column's example chromatogram used ambient temperature. I have knocked the temperature down and now phosphate and nitrate come off separately. The phosphate actually comes off first. So if i lessen the eluent i should get the phosphate coming off later (to try and spread out the peaks). I still though don't see a peak for phosphate concentrations less than 30 ug/L though. the area under the peak is 0.0057 (us*min) for phosphate at 30 ug/L and should be 0.0021 for 10 ug/L according to my regression. that is getting close the the amount of noise in the baseline. if I decrease the eluent concentration will i be able to better pick up the phosphate. should i go to a larger sample loop (we use 25 ul)? i also use the suppressor on recycle mode, how does one setup the suppressor for external mode? any help would be greatly appreciated.

matt

[Markus Laeubli, Metrohm]

I assume that Chris Pohl's first answer is the key to the problem. You mention that the iron concentration is pretty low. But you need to realize the your eluent is running over the column by the liters compared to the sample which is injectded by microliters.
You need to regenerate the column and tp prevent it from iron by using eluents with much lower contamination or you may add a scrubber column which removes the iron prior to the injection.

[Chris Pohl]

Matt,

Sorry for not responding earlier but I've been on an extended business trip. Anyway, I think in your circumstance the best option is to increase your loop size. Since this column is relatively high capacity, you should be able to work comfortably with a 100 microliter loop and this should simplify the quantitation of phosphate at low levels.

To operate in the external water mode there are a couple of options. The regenerant can be pumped using a standard IC pump but this is generally overkill since neither flowrate nor pulsation level is critical for this application. Most people use a pressurized 4 liter polyethylene container filled with water and outfitted with an inexpensive pressure regulator. Dionex sells a kit containing the necessary components (product part number: 056886) and associated operating instructions.

[mreardon]

Chris:
well i was informed by a colleague that they use a 500uL loop for phosphate. so I went up to this amount. I ran my standards and I was actually able to see the 8 ug/L standards upwards. At the low concentration of 8 ug/L I am putting only 4 ng of phosphate onto the column. I got a good standard curve in terms of r2 but when I ran another standard on the curve and it changed the slope of the standard curve by only 0.001 it changed the predicted regression values quite a bit, by 50% on the low end of the curve. This means that our standard curve is too sensitive and we can't be entirely confident of the low values that we might get. We are also going to have the problem of loading 500uL samples when they have 4mg/L bromide in them. The only things left to try is getting the system on external mode and perhaps an scrubber column for the iron. We make our eluent manually as well is there anything to put between the eluent and the column to remove any carbonate that may form in making the eluent?? thanks for you help.

Matt

[Chris Pohl]

Matt,

Although it's pushing it a bit, you should still be able to get away with an injection of 500 microliters if the most concentrated component in your sample is 4 mg per liter of bromide. You mention the calibration was adversely affected by your increase in injection volume but you didn't mention whether this was on an area or peak height basis. Generally, you should get the best results if you use an area-based calibration for phosphate. If you are observing a linearity problem even on an area basis, there's a couple of things you should look into. Make a few injections on the low-end of your calibration range, then make a few injections on the high end of your calibration range and finally make a few more injections on the low-end of your calibration range (for phosphate, that is). The thing to look for is whether or not your calibration is stable on the low-end. If you see an upward drift in area for your low-end calibration after injecting more concentrated samples, this is an indication of iron contamination. You may need to verify whether or not you have iron contamination in your hydroxide (I believe you indicated earlier this was not the case) or perhaps you have iron contamination of your deionized water. The latter is relatively uncommon but I know the least one case where the problem was traced to the deionized water system.

Regarding your other question, there are definitely options for removing carbonate contamination from your eluent. The ATC-HC is designed specifically for this application.