Im fairly new to HPLC, i have a background in Zoology.i am using a waters nova-pak phenyl column,5x100mm 4um, I am using 55:45 meoh, .0425% phosphoric acid to separate 6 pesticides in one sample. The problem i am getting is that three of the samples are co eluting and causing one big peak with shoulders. I know it has a lot to do with the polarity of your mobile phase and samples which enables peaks to get baseline separation. Any suggestions?

You do not say whether the three co-eluters have retention times near the start of the chromatogram (at the column dead-time t0); if so, then you must reduce the %-MeOH in your mobile phase. If the retention times of these compounds is not close to t0, then you need to change selectivity -- as by trying 45% acetonitrile/buffer instead. Depending on the nature of your sample, there are many other ways to pull apart overlapping peaks when retention is >> t0.

In addirion to Dr. Snyser's queries- can you provide the identification of the pesticides. Many of these separations require a gradient.

In addirion to Dr. Snyder's queries- can you provide the identification of the pesticides. Many of these separations require a gradient.

The retention time of the samples does not fall near the start of the chromatogram. The three herbicides are tebuthiuron, bromacil, terbacil

THis is one case where a picture is worth a least a few dozen words. Any chance you could post a chromatogram? The ideal would be to export it as an Andi (AIA) file. There are instructions on the forum about how to attach these files. Alternatively, if you FAX me a chromatogram, I'll try to scan it in and post it as a bitmap file.

At the risk of being pedantic, resolution depends on retention, selectivity, and efficiency (peak width). It sounds like a selectivity issue, but I'd like to be able to rule out excessive peak width, especially since you're using a fairly small column where plumbing and detector cell size can be an issue.

-- Tom Jupille / LC Resources
fax: (925) 930-9136

Looking at the isocratic chromatogram you faxed me, the dead time (tM) of the system is about 40 mm (the time axis got cut off, so I'll just use distances on the printout), retention time of your first peak is about 56 mm; the last peak has tR of 180 mm (both of these peaks are well resolved and nicely shaped). The widths at half-height are 4.2 mm and 7.8 mm, respectively.

A back-of-the-envelope calculation gives capacity factors of about 0.4 for the first peak and 3.5 for the last peak. Your problem triplet is at about 2.5. This is a tad on the low side, but merely increasing retention isn't gonna help all that much (resolution is a function of k/(1+k), so increasing k's from 2.5 to, say, 5 would only increase resolution by about 20%.

Next point: the plate numbers for the first and last peaks are about 1,000 and 3,000 respectively. This is *way* low for a 10-cm column with 4-um packing. The fact that the early peak looks worse than the late peak suggest extra-column volume as a possible culprit. Definitely check all of your tubing: make sure that it's no larger than 0.010" id (0.007" might not be a bad idea). Also keep the lengths as short as possible. Finally, at least for starters, try to keep the injection volume down to 10 - 20 uL. You should be able to increase your plate count for that triplet by a factor of 2x - 3x. Since resolution is a function of the square root of the plate count, this should translate into a 40 - 70% improvement in resolution.

Finally, we can address the selectivity issue. The spacing of the peaks in the isocratic and gradient chromatograms you sent were *entirely* different. This suggests that you can move the peaks around by simply changing % ACN, and that, in turn, suggests that there may be some % ACN that pulls all 6 of your peaks apart.

We (LC Resources) make a program called DryLabŪ that lets you model these effects. All you need to do is a second isocratic run with 10% less organic solvent than the one you sent me (i.e., 35% ACN/water in addition to the 45% ACN/water). If you want to send me the chromatogram, I'll enter all the information into my copy of the program and let you know what happens. If we're lucky, we can find a % ACN that works.

If we *can't* find one, at least you'll *know* that ACN by itself doesn't work and avoid wasting time following that blind alley. The next step would be to try a different temperature (assuming you have temperature control) or substituting MeOH for the ACN. Each of these requires two experimental runs to calibrate the model.

There is no "magic insight". The bottom line on developing methods is that it's best to be systematic and knock off one variable at a time.

(sorry for the long post!)

-- Tom Jupille / LC Resources

Oops, I misread your chromatograms! Switch ACN and MeOH in my last post! And change the percentages I mentioned appropriately.

Sorry!

-- Tom Jupille

Thanks for the analyte identification. For the sake of completeness re suggestions - the method of choice is EPA Method 507 (GC) which using a DB 5 column separates each of the three analytes by several minutes. A N/P is specified however an FID will probably work.