Fluoride and phosphate calibration issues

[Hidrolab]

For last couple of years we were doing calibration for anions in range of 0,1 mg/l to 8 mg/l for fluoride, nitrite and phosphate and also 1 mg/l to 8 mg/l for chloride, nitrate and sulfate.
Till this year we had no problems with calibration, curves were linear for all of the anions and recoveries were always in the specified range (90-110%)- we had loop 10 uL. In June this year we couldn't meet criteria for linearity (>0.99) for some anions so our maintenance recommended to us to change the loop to 25ul for better sensitivity. We changed the loop and done the calibration again. Coefficient of correlation for fluoride was 0.96, and recoveries were poor. Calibration curve was set up from zero, otherwise it would go from negative peak. We set the quadratic form of curve and we saw that for phosphate and for fluoride there is no linearity (it isn't a straight line it's a curve) - in the 2015 it was a strait line in this range for both of the anions. So we tried to analyse CRM's in range of linearity (phostphate from 2,5mg/l-5 mg/l and 4-8 mg/l for fluoride) with more points (phosphate wasn't detected below 2,5 mg/l and fluoride below 1,0 mg/l). For phosphate we determined linearity between 2,5 mg/l and 3,5 mg/l and for fluorides between 4mg/l and 7 mg/L.
For other anions linearity changed a bit but coefficient of correlation is still above 0.99 - recoveries are between 90-95% for concetration of CRM that is in the middle of the calibration curve (90% is minimum).
Column that we use is for about 2000 to 2500 samples, we analysed less than 1000 samples. All of the samples are filtered trough filter with pores 0.25 um and diluted if conductivity exceeds 1000 uS/cm.
Our maintenance suggested to change the suppressor, column and precolumn but with no guarantee that it will solve the problem.

Did anyone had similar problems, what was the solution? If anyone has some idea what could be the problem, please help.

We are using Metrohm 761 Compact IC, column Thermo Scientific ionpack AS14 4mm, Dionex ASRS Ultra 2 - 4 mm, Metrohm 813 Compact autosampler.

[James_Ball]

Does the problem continue if you use a new column?

We normally keep two columns and when one begins to show early elution times or broad peaks we regenerate it by pumping 100x concentration eluent through the column for about 10 minutes then store the column while we use the second column, then regenerate the second and switch to the first again. Just have to flush the regenerated column with normal eluent for a few hours until steady baseline and it is ready to go.

We are normally analyzing 50 samples per day 5 days per week and have to change columns about every two months. Sometimes we get very bad samples and have to change it more quickly. About once per year we purchase a new column and remove the worse column from the rotation.

We normally use the Metrohm Asupp-5 or Asupp-7 with good results. The Asupp-7 will give better results for F- since it retains it more strongly and moves the peak further from the water injection dip.

[Markus Laeubli, Metrohm]

Dear hidrolab

Phosphate unlinearity quite often comes from traces of iron in the eluent and/or samples. If a new column does not solve the problem, a new suppressor might help.
Problems with fluoride might occur e.g. aluminum contamination in the system.

[Hidrolab]

Thank you for replies. We don't have a reserve column but we did try to regenerate the one we have. Response isn't better at all. We have analysed the eluent and metals aren't present in it. When we get a new column we'll do a new calibration and we'll see what will happen.
Thanks again.