Chromatography Forum

One more thing: the injection on the Oasis precolumn needs to be done when the precolumn is in water. If you left organic solvent in the precolumn (from the gradient), you may have trouble with reproducibility.

Good evening!

This is great help for me! Analyzing hair and tissues are still another year into the future for me. Getting saliva and serum is the most important right now.

The idea of not adding an organic solvent to saliva has never occurred to me! I thought that was necessary. Saliva contain many macromolecules like:
Mucin 1, sIgA, Mucin 2, lactoferrin, peroxidases, amylases, carbonic anhydrases, proline-rich proteins, lysozyme, statherins, histatins . Many of the proteins/enzymes show a diurnal rhythm and also change depending on food intake. Some of the steroids are probably bound to such proteins, although very small fractions. Many regard saliva concentrations as a reflection of unbound concentrations. Wouldn't this become a major concern with regards to CV%. I will try this tomorrow, guess I will find the answer myself.

I have already tried the Oasis procedure you describe(with backflush) As I've mentioned I have had some problems with broad peaks. I suspect this will be eliminated when avoiding organic solvents in my samples.

A guard column in front of the Oasis is not necessary (or possible/complicated) when doing online extraction with backflush. Right? Or should I protect the analytical column?

I'll report my progress....

Again...thank you all, and especially Uwe. This forum is an fantastic place to discuss and get help. The word is out!

The Oasis precolumn is supposed to take care of the proteins. The proteins pass through it for multiple reasons. It has rather small pores, and it has a larger particle size, and when you do you an injection of a sample containing protein, the proteins will not go into the pores and stick to the packing. This is especially true if you do the injection at a high flow rate. The high flow rate may be a good thing to do anyway, since you want to inject a large volume.

You may consider under all circumstances to denature the proteins in your sample by adding a strong acid to the sample. This is likely to prevent binding of your analytes to the proteins, and is a rather standard practice. You only need to flush out the acid before switching to the analytical column and the MS. This is the purpose of the wash step, most often carried out with 5 or 10% organic, with the addition of whatever additive you in the mobile phase that you need for the LC.

The Oasis column is the sacrificial lamb. One of my colleagues who is doing these things all the time claims that he can get 500 injections out of one Oasis column when he works with plasma samples. I am much more cautious, but I think you should get some 200 injections with plasma, and 1000 injections with saliva.

Hi again!

The last days I have done several experiments to increase my sensitivity. First I did a proper wash of my MS and LC. Then I startet from scratch again trying both "on-column" extraction and "on-line extraction" using Zorbax C18(with guard column).

My conclusions are as followed:
- On-column extraction takes me almost to the sensitivy(in real samples) I want when injecting 400uL of saliva(no preperation). I suspect I have had some contamination of my MS, as I had to wash the Q0 once more. I now have installed a valve to supply the MS with only the spesific time window my analytes elute. The chromagram show extremely sharp peaks (10 sec after 11 minutes, although an annoing baseline noise).
- On-line extraction: Here I have had problems. First I suspect that my Oasis HBL is "loosing" some of the analytes, although it's properly equlibrated by H2O. It's rather old, but there is no obvious pressure problems. Perhaps it's worn out? What are the signs of such a column loosing effiency? Secondly...I discoverd the pump driving the analytical column shows very a oscillating pressure. Within a few seconds it drops and raises by 100-200 psi(10-20% of total pressure). Of course, I now this is'nt suppose to happen, but I can't find any leaks. I suspect the error is inside. From what I read this "hammering" might harm my analytical column and most certainly influence the chromatography. Right???? The third problem is the elution onto the analytical column. I find it hard to make sharp peaks and narrow peaks (like the on-column extraction). This is one, I think, important limitation that makes it hard to achieve high enough sensitivity. Still...using on-line extraction with all the descriped problems, the LLQ good, but to large by a factor to 4-6.
I have tried different equlibrating the analytical column with MeOH:H2O from 50-100, and tried different elution programs for my extraction column(50-100% MeOH).
- Actually, my best efforts seems to be when doing combined on-line extraction and on-column extraction. That is eluting the Oasis with MeOH(60-100%) onto the pure H2O pre-equlibrated Zorbax C18. I have to look into this some more though.

So, I have a few questions other than those asked previously above.
- Any recommendations for producing extremely sharp and narrow peaks using on-line extraction? Gradient programs?
- Is it possible combining on-line and on-column extraction? Is there a downside? I.e will this effect recovery and thereby sensitivity?
- Uwe: You mentioned another method that involved playing with my hardware. I am very interested in this? I would be very happy if you can explain that furter.
- Does anyone have some suggestions? I am a bit stuck here. I so close, but still so far away...feeling I have played all my options.
- Would going UPLC actually help? Or is it better to upgrade my MS? Remember, it's mostly steroids and their metabolites I am into.

Regard,
PMDOC

I am not sure what your are doing, at least you have confused me. If I would be next to you with a piece of paper or look at your instrument, I think the problem would be clear.

Here is what my suggestion is:

Load your sample onto the precolumn in as high a water content that is possible. For a real urine or saliva sample, this would be the sample, with some adjustment of the pH and an internal standard. Load your 400 microliter on the precolumn, then flush the precolumn with 100 microliter of pH adjusted water with 10% MeOH. All of this is going to waste.

Now you invert the Oasis precolumn such that its inlet is connected to your analytical column. (If you do not (yet) have the setup to do this with a valve, you can do this even manually.) Now run your gradient program (from 10% methanol) through the inverted precolumn onto the analytical column. You should get sharp peaks coming out of your analytical column. If you don't get this, either your precolumn or your analytical column are dead.

If you have an automated setup, you are sure that your columns are working, and you still get wide peaks, I would check, if there is not an area in your instrument where the final mobile phase from your last run is still sitting around. I have seen things like this - pieces of tubing with methanol in it or related problems.

Once you got decent results with this procedure, we can take it a step further.

I am not sure what your are doing, at least you have confused me. If I would be next to you with a piece of paper or look at your instrument, I think the problem would be clear.

Here is what my suggestion is:

Load your sample onto the precolumn in as high a water content that is possible. For a real urine or saliva sample, this would be the sample, with some adjustment of the pH and an internal standard. Load your 400 microliter on the precolumn, then flush the precolumn with 100 microliter of pH adjusted water with 10% MeOH. All of this is going to waste.

Now you invert the Oasis precolumn such that its inlet is connected to your analytical column. (If you do not (yet) have the setup to do this with a valve, you can do this even manually.) Now run your gradient program (from 10% methanol) through the inverted precolumn onto the analytical column. You should get sharp peaks coming out of your analytical column. If you don't get this, either your precolumn or your analytical column are dead.

If you have an automated setup, you are sure that your columns are working, and you still get wide peaks, I would check, if there is not an area in your instrument where the final mobile phase from your last run is still sitting around. I have seen things like this - pieces of tubing with methanol in it or related problems.

Once you got decent results with this procedure, we can take it a step further.

Hi!
Thanks for input! Actually, I have this online setup you descripe up and running. The problem is to optimize it, as I believe my sensivity target is at the limit of what API 4000 can quantitate.

So, to provide more detail - here is an example of what I am doing. I'll see if I can attach a picture of the chromogram. Do I have to put it on a web server?

Injecing 200uL
Pump 1. (to the extraction column Oasis HBL)
Step Time Flow Rate(µl/min) A (%) B (%) C (%) D (%)
0.00 600 10.0 90.0 0.0 0.0
2.00 600 10.0 90.0 0.0 0.0
4.00 600 80.0 20.0 0.0 0.0
13.00 600 80.0 20.0 0.0. 0.0
13.10 600 10.0 90.0 0.0 0.0
17.00 600 10.0 90.0 0.0 0.0


Pump 2 (to the analytical column Zorbax C18 - and backflushing extraction column). Flow 600uL/min
Time Flow Rate(µl/min) A (%) B (%) 600
0.0 600 10.0 90.0 0.0 0.0
4.00 600 80.0 20.0 0.0 0.0
13.00 600 80.0 20.0 0.0 0.0
13.10 600 10.0 90.0 0.0 0.0
17.00 600 10.0 90.0 0.0 0.0

The flow to the analytical column goes directly from pump 2, except in time window 2.0-5.0 minutes when it first backflushes the extraction column.

Now..the peaks look rather well, but there is some sign of tailing. And when running lower concentration real samples I still lack sensitivy by a factor of 5.

Any clue for further improvement?

OK. I now have a better picture, I think.

1. Do not wash the Oasis column with the same solvent strength that you use for eluting the analytes from the Oasis column onto the analytical column. Go only to maybe 70% A.

2. When you backflush the Oasis column to the analytical column, keep it connected throughout the second run.

3. Run a gradient from maybe 70 to 90% instead of the isocratic run at 80% between 4 and 13 minutes with pump 2 (unless you can only do an isocratic run with pump 2, if this would be the case, we would need to rethink).

OK. I now have a better picture, I think.

1. Do not wash the Oasis column with the same solvent strength that you use for eluting the analytes from the Oasis column onto the analytical column. Go only to maybe 70% A.

2. When you backflush the Oasis column to the analytical column, keep it connected throughout the second run.

3. Run a gradient from maybe 70 to 90% instead of the isocratic run at 80% between 4 and 13 minutes with pump 2 (unless you can only do an isocratic run with pump 2, if this would be the case, we would need to rethink).

Hmm...ok...got a two follow-up questions...

1. What do you mean by "keep it connected throughout the second run"? I presume you mean let the flow stream through the precolumn-->analytical column for the whole run, except the washing step. I don't understand why though...I am sure that my analytes are eluting in that spesific time window.

2. I wash the extraction column with 10% MeOH during the first 2 minutes, before the valve switched to backflow. So I don't understand your your first point. Do you mean that after the analytes of interest have eluted I should only wash with 70% MeOH.

Perhaps what's confusing you is my unconventional naming of mobilphased. A=dH2O and B=MeOH.

I go on as you suggest and report back.

OK, you are doing it right. I had confused myself (had thought that the two tables were separate operations).

Your actual analysis is isocratic. You can get sharper peaks with a gradient.

You may be able to go to a shorter and thinner column. Please describe the post-column connections! (Tubing length, i.d. etc.)

Post column there is a PEEK tubing ca. 40cm ID .005" (red).

How far away are you now from the sensitivity that you want? We can reduce the column diameter to 2 mm, which will gain about a factor of 2 in sensitivity. We can use a shorter column with a smaller particle size to gain maybe another factor of 2. You will need to rearrange the column, and put the column closer to the detector (use the 40 cm of tubing in front of the column; insulate the column and the tubing, if you need temperature control, but if you run a gradient, you may not need any temperature control). A gradient could also sharpen your peaks by some reasonable factor.

Overall, with all this, you could gain about a factor of 8 in sensitivity.

Hi!

I'm still struggling with this. I want some "stacking" of the analyte when it flows onto the analytical column, to shallow the peaks. This is difficult to achieve. I guess my analytes is eluting at about 30-50% MeOH from the extraction column. Now...what would be the optimal equilibration of the analytical column to achieve such stacking???

So far I have not been able to observe any increased sensitivity when doing a gradient program. Actually, the method you called "isocratic" maybe not so isocratic after all. The column has a low "organic condidition" when the run starts, and it probably increases through the whole run although the mobil phase composition is isocratic. In fact, after going through my retention times I suspect that I don't manage to wash out lot of MeOH if the gradient maxes out on 70-90% The next sample seems to have a shorter RT than the previous (and also the respons if often lower).

Your suggestions are interesting. I can cut 10 cm of the post column tubing. The tubing is not isolated, but the column is placed in a heated compartment. so I guess I have some temprature controll).
I might try with a 2mm ID column and also with a smaller particle size column. But first I will see if I can achive high enough sensitivity with this.

A stacking of the analytes on the second column may or may not happen with this setup. It requires that teh analytes are more retained on the analytical column, which may not be the case. However, with the backflush approach you are keeping the analytes very concentrated at the column inlet, followed by an equally sharp elution. Actually, you mentioned that the peaks have some tailing. This could still be due to the fact that you need some organic to dissolve the sample, which will make the analyte move down the concentrator column a little bit.

The changes in retention from run to run can be related to the equilibration cycle. This is not uncommon in gradient chromatography. People often do one or two blank runs before they inject their real samples.

If you run your analytes on the analytical column alone in isocratic mode, what MeOH concentration do you need to elute them?

Compared to the "On-column" extraction there is considerable less stacking. That method made my peaks elute in about 10 sec after 12 minutes, which led to very high peaks. When running the same standard solution with the on-line extraction the peaks are about 2-3 times lower, but also considerable wider (about 20-30) sec. That is my rationale for trying to make this happen also here. But, get your point! This might not be possible, due to the organic mobilphase needed to elute the extraction column.

Still, it's an obvious experience from my tests, that equilibrating my analytical column with too much MeOH results in broad, low and early peaks. The reason for this is pretty obvious. I find it harder to understand why low peaks appear when equilibrating the analytical column with just water(or very little MeOH). Do you know, Owe?

So, that's what I am working on now - Finding the optimal equilibration for the analytical column. It's hard because the retention times changes and I get lot of unexpected results - probably due to insufficent equilibration. What keeps the hope up is that I have had 3-4 run with very high peaks, at least as good as the "on-column" extraction. But I can not reproduce them - well....I can, but only to an certain extent. Succeeding this might be just enough!

Perhaps I should just do a "brute force" attempt trying different solution strenghts.

What you call "stacking" is exclusively a function of the sharpness of the gradient. Since you are not programming any gradient, the gradient is created by details in your setup and your procedure. This is not the way to do this, since it will never be reproducible.

Instead of doing a step to the second mobile phase, I suggest that you run a gradient. We worry later about the sharpness of the peaks. BTW. how many peaks do your have and what are the typical elution time frames and the peak widths?