No sample peaks, but mobile phase peaks after HNO3 flush

[browntk]

I'm a recent grad & a new tech working on a Shimadzu U-HPLC.

Before this problem, I've been able to generate a pretty decent chromatogram for our analytes (B-carotene and other hydrophobic Vitamin A precursors). So I know what retention times I should be expecting for my samples when using the following mobile phases with a protocol gradient (I'm using a protocol that works fine at another university). Mobile A: 100% EtAc, Mobile B: 85%ACN:15%H2O:.1%TEA.

Everything seemed to be going fine except that after 1-2 weeks in which I was not working with the HPLC, I had some trouble with two overlapping peaks that have a similar retention time, and eluted as one fat peak. I tried a 50MeOH:50H2O:.1formic acid flush to clean up those particular peaks (which worked before to separate them), but this didn't work. I thought there might be contamination in the column so I flushed, & reverse flushed the column with isopropanol (I have a C18 column). When this didn't help, I made a 1M HNO3 solution thinking that maybe the detector needed to be cleaned. I removed the column and flushed overnight (....maybe not the best idea) with this 1M solution. (before the HNO3 solution, I did flush 100% ACN and then removed the column).

Since the HNO3 flush I haven't had any peaks from any samples or standards. I get the typical noise in the first minute as I have always gotten, which I assume comes from the changes in mobile phase, but there is no peaks past 1 minute even when I run for 40 minutes (usually the last compound elutes ~10min).

I've tried running 3 different columns all of which worked fine before and I get the same chromatographs (ie: I'm not getting any compound peaks on any of these columns), so I don't think there should be any problem with the column (especially since I didn't flush any nitric acid on any of the columns).

I extracted new samples to make sure that my samples hadn't deteriorated.

It's been almost a week since I ran the HNO3 so the whole system should have flushed out the acid by now. I checked the pH of mobile B before and after entering the line and it did drop .2, which would change the retention time, but not so much as to not get any peaks.

I suspect that my compounds may be eluting in the first minute: so how do I separate them back out?

or somehow the autosampler is not injecting the sample into the line? There is no leak anywhere and I can see the autosampler picking up the correctly programed sample into the port.

Help!

Thank you thank you thank you,
Tirzah

[bisnettrj2]

What type of detector are you using? What model, specifically?

[browntk]

Shimadzu UV-VIS (SPD-M20A)

[bisnettrj2]

If you bypass your column (connect a length of PEEK tubing sufficient to supply moderate backpressure directly from the injector to the detector) and inject your sample, do you see a peak (should be a large blob at the void volume)?

[bisnettrj2]

Also, what model is your autosampler? What material is your rotor seal (PEEK, PTFE)?

[browntk]

A day or two ago I bypassed the column and got very similar looking noise in the 1st 30 seconds as I do when I have the column on.

The autosampler is a Shimadzu 20AC XR. I'm not sure what a rotor seal is; how would I find out what it's made from?

[bisnettrj2]

The rotor is the seal in the injection valve. It has channels grooved into it that provide a flow path through the ports in the injection valve. Some materials are intolerant of certain solvents or pH ranges, and if you are not seeing anything resembling your peaks when you inject a standard without a column, I would suspect that your sample is not making it to the detector.

You should be able to ascertain the material of the rotor seal by checking in the manual (Shimadzu should have supplied one with the instrument), or by calling Shimadzu and asking. I found their website cumbersome, and they seem to only sell the manuals, and do not provide them free online, so calling them might be your best bet.

[browntk]

It looks like it is made of PEEK and should withstand pH from 1-14, according to: http://www.ssi.shimadzu.com/products/li ... IL-20a.pdf

[bisnettrj2]

PEEK should be ok with Nitric acid. However, we still need to determine three things:

1. Are you withdrawing sample from the vial?

2. Is the sample making it to the detector?

3. Is your detector still working correctly?

You can check the detector by doing the following:

1. Flush your system and column with isopropanol (20-30 column volumes).
2. Remove your column from the system and replace it with a long piece of capillary tubing - long enough to generate a few hundred psi of back pressure.
3. In mobile phase A, have just water.
4. In mobile phase B, have water plus 0.1% acetone.
5. Set your detector to 254 nm.
6. Set the pump to 1 mL/min and 100%A
7. Pump for 5 minutes, then set the pumps to 75:25 A:B
8. Pump for another 5 minutes, then set the pumps to 50:50 A:B.
9. Pump for another 5 minutes, then set the pumps to 25:75 A:B.
10. Pump for another 5 minutes, then set the pumps to 100%B.
11. Each change in mobile phase composition should result in a rise in the baseline, as the percentage of acetone in the mobile phase increases. If you don't see a change, you have a problem in your detector (or your pump, but let's think about that later). If you do see a change, you have a problem getting sample from the autosampler to the detector. My guess would be a leak in the rotor seal, unless the vial weighing test below gives poor results.

You can check whether the autosampler is actually pulling sample from your vial by doing the following:

1. Fill a vial with water, cap it, and weigh it.
2. Have the autosampler inject ten 50-microliter injections, weighing the vial after each injection.
3. Each injection should use approximately 0.05 grams of water (unless your autosampler does loop overfill injections, and then I have no idea how much it will use), and there should be very little difference in the change of weight from injection to injection.
4. If you see no change in the weight of the vial, then you aren't picking up sample, and the problem is in the autosampler. If you do see a consistent change, then the problem could still be in your autosampler.

To see if you are actually getting sample to the detector, do the following:

1. Re-equilibrate your column with a mobile phase of 25:75 water:acetonitrile at 30 degrees Celsius and 1 mL/min.
2. Make a solution of acetone and toluene in 1:1 water:methanol (in a 10-mL volumetric flask, add 20 microliters each acetone and toluene, dilute to mark with methanol. Take 500 microliters and put in a 2-mL injection vial. Add 500 microliters water. Cap and mix.)
3. Set your detector to 254 nm.
4. Inject 10 microliters from this vial.
5. Let the run go for 30 minutes. You *should* see a peak at the void volume (acetone) and another peak much later in the run (toluene). If you don't, and the above tests showed that your detector is responding appropriately and your autosampler is withdrawing sample from the vial, then we can say you aren't getting sample to the detector, and you probably need to replace the rotor in the injection valve.

[browntk]

Thank you for all of the advice.

The detector is working fine based on the IPA run.

I made 10 injections of water, weighing the vial after each injection and there is definitely sample being picked up (though about 45uL instead of 50uL).

I am running the 50:50 MeOH: H2O sample, but with acetone and phenol because we do not have toluene. It is only at 10minutes so far, but I do not see any peaks at all - not even a void volume in the beginning - which seems odd because I ran some of our samples earlier today with the usual mobile phases and got several distinct peaks, but they elute very early (under 1 minute).
I'm sure that sample is getting through and the detector is doing its part, but for some reason the compounds elute way too early from the column. Thoughts?

[bisnettrj2]

"The detector is working fine based on the IPA run." - I'm confused - what IPA run? The run where you flushed your column originally, or the run where I suggested that you flush your column and system with isopropanol?

Did you try injecting more of the acetone:phenol sample, just to see if you had too low an amount on-column in the first injection?

Can you post a chromatogram of a normal standard, and one of the standards you injected where you saw peaks that eluted within a minute?

EDIT - Also, I'm going to ask a basic question here - did you re-check the plumbing of your system? Is mobile phase A actually going to pump A, and B to B? Did you check the pump program and detector settings to make sure they match the previous settings?

[browntk]

Sorry, what I meant was that I did run the IPA flush you just recommended and I saw a rise in the baseline with each increment of increasing acetone.

I discovered why the acetone:phenol trial failed; because after I changed mobiles I purged the system and did not close the valves again, so all the mobiles went to the waste before reaching the column. oops. I'm running this again right now.

I'm not sure how to paste files on this forum. I have saved two such chromatograms as JPEG files, but I'm not sure how to upload them?

[bisnettrj2]

1. That test should have been done with water in the A reservoir, and water with 0.1% Acetone in the B reservoir, but I guess it could work with IPA.

2. How to post chromatograms: viewtopic.php?f=1&t=2617

[browntk]

Oh of course, sorry, I was referring to the IPA flush as the whole flush starting with an IPA rinse and then increasing the concentration of acetone in the water solution, as you explained.

I figured out that the main problem was actually that I had changed my method unintentionally to switch, between every run, from 100% mobile B of 85%ACN:15%H20:0.1%TEA to 100% EtAc (mobile A) in a matter of seconds, causing the compounds to all elute together in the beginning. Now everything seems to be working just fine!

Thank you so so much for all your help and advice. It was very valuable to know some specific ways to trouble shoot the detector and the autosampler. Thanks!

[bisnettrj2]

Not a problem. However, some advice: Always check to make sure the simple things are as they should be - mobile phase program is as it had been, detector settings are what they should be, autosampler is injecting the right amount, etc. That's advice I should have taken after you asked your question. However, I'm glad you have gained some insight into how to troubleshoot some problems in the future.

With your original problem in mind - overlapping/co-eluting peaks that weren't previously co-eluting has almost, if not absolutely, nothing to do with your detector. And anytime you think you need to clean your column or HPLC, take the detector out of the flowpath - flushing garbage into the detector is almost, if not always, a bad idea.