Moving peaks

[gurkishan]

I am running the USP method for Acetaminophen, aspirin and caffeine.
i run USP standards and then the samples.
The mobile phase is made in bulk by individually measuring water(69%), methanol(28%) and glacial acetic acid (3%). The pH of this is somewhere around 2.4pH
Problem starts with that when i start the HPLC machine everyday, with the same mobile phase, and equilibrate it for 2 hours, my pressure stabilises at different pressures everyday. But, once it stabilises, it remains the same throughout the day-- so there is an interday variability in the retention times with change of pressures.
Also, i have noticed that when i run a set of 45 injection, the peak that comes at 6 min, after 45 inj, it comes at 4.5 min near about. And the pressure is stable!
The solutions are filtered before injection. The column is washed after every 40-50 injections with 100% methanol. And i have had 4 new columns of different kinds just to see if columns are the source are problem.
I have checked the HPLC machine, it is OK.
Cant think of any thing else..

[Consumer Products Guy]

Problem starts with that when I start the HPLC machine everyday, with the same mobile phase, and equilibrate it for 2 hours, my pressure stabilises at different pressures everyday. But, once it stabilises, it remains the same throughout the day-- so there is an interday variability in the retention times with change of pressures. ... I have checked the HPLC machine, it is OK.

MUST be a hardware issue, you must isolate whether the issue is your system or the column itself. Your mobile phase (even if you make it SLIGHTLY different every day) should not be the issue. It should not require 2 hours to equilibrate either, 15 minutes should be plently, and that's usually for chemical equilibration of the stationary phase, not the hardware.

[Bryan Evans]

Look at either column fouling (needs improved sample cleanup)
or pump issue. Do you have MeOH in line B? If the pump is
drawing in MeOH - that would cause increase in pressure / decrease
in RT (one would expect decrease in RT for all peaks).

[Kostas Petritis]

Have you ever tried to use another batch of acetic acid (it looks like you have changed anything else)? Have you use this method before and it worked or is it a new method you are trying out?

[gurkishan]

Column, i have changed 3 times and same problem.
Acetic acid is for HPLC from Fishers.
And only one pump with mobile phase works so methanol contamination cannot be possible! Also, pressure is totally stable

[Peggsy]

i once had a bottle of HPLC grade methanol for a reputable company that was actually filled with acetonitrile! took me a while to determine why my RT's were significanly reduced. I know that is not your problem here, but never assume anything, never assume all HPLC grade solvents are not the casuse of your problem.

If you run these application every day, do you actually shutdown the HPLC overnight or simply run it at low flow?

[Bruce Hamilton]

If I recall correctly, the USP method uses 5% Acetic Acid in Methanol as the sample and std solvent, so I wouldn't expect too much effect from a methanol wash of the column. I assume this method was new to your laboratory, and hadn't been used before?.

The most likely hardware explanation of day to day variations would appear to be column temperature. I'd also want to ensure the sample solutions were very clear before the addition of the Int Std and dilution solvent.

My questions relate to the behaviour of the Int Std peak ( Benzoic Acid ). Is that the peak you are referring to?. How does the shape and retention time vary during a long sequence?.

If the int std peak is losing retention, I would also ascertain the effect of a long sequence that was just injecting standards and blanks - without any samples. I'd also flush the column with a highly-aqueous wash solvent, reequilibrate for 30 minutes, and reinject the stds. Does the retention still decrease or change?.

The method doesn't specify sample filtration, do you have a guard column that can be replaced. If so, what happen if you replace it and rerun the latter part of the previous sequence.

Bruce Hamilton

[HW Mueller]

OK, so if I still follow this your retention times don´t spontaneously undulate (move in an out as you called it), but change from day to day with a change in pressure. (If you allowed yourself the discipline to answer all the questions, especially those of juddC and mine, I probably wouldn´t have to guess). Sometimes the resolution gets really bad (still don´t know)?
Assuming that your sample matrix is complicated and contains macro molecules, I will venture the guess that some of it sticks to the column (frit), but does not change column resistence until you wash with MeOH (we have seen this when proteins were in the matrix). Now the builtup of junk on the column can cause a retention shift. That would be gradual (we still don´t know whether that is what you see also). That the gunk would only cause a rt shift directly after MeOH is highly unlikely. More reasonable would be to assume that the pressure change is accompanied by a flow change, that is, your pump is not working as it should. (Maybe it works only with the mobile phase of this test run you mentioned, or with the column resistence you had in that run. You still didn´t tell us whether you measured the flow in relation to pressure change).

[gurkishan]

Hi guys!
I have a shutdown method that washes the column for 60 min with 100% methanol, then closes down. Do u think that i should changing the washing method?

My mobile phase (USP) has 3% glacial acetic acid, and it is in a column heater at 45C-- which i have calibrated and it is stable, without temperature variation.

Yes, Int Std is Benzoic acid. It is the last one to be eluted out, and after around 40 injections, the peak moves from around 6 min to 4.5 minutes. All the other peaks also move and this movement is realtively similar to other peaks.
After washing and 2 hour equilibration, the peaks still dont go back!
I dont have a guard column as samples from dissolution are filtered and USP standards are totally soluble in water.

I apologise Mr Mueller, if my answers have been inappropriate. I appreciate you taking the time and shall try to keep up with what you say!!
After running the injections, the machine shuts down. I start it again in the morning and equilibrate it. That is the time the pressure stabilises at different value. It does not vary by much-- like instead of 2200, it may be 2400. But the resolution is totally bad-- like all the peaks will be eluted out in 2 min as compared to 6 min from previous run.
I cant understand column deposition as i am giving good washing and all the solutes are water soluble-- aspirin, acetaminophen and caffeine.
I have checked the flow with a graduate cylinder and it is accurate. And i have had the equipment checked, it is in good condition.

[mbicking]

Just jumping in late here. An impressive list of suggestions. But since all the obvious ideas haven't worked, we will have to start looking at the less obvious solutions. [Famous quote along these lines, ... anyone remember it?]

1. First, what kind of instrument are you using? Manufacturer, pump type, etc.
2. Pre-mixing is fine, but there is no need to sonicate for one hour to degas! What do you use for degassing during the run? If you have an on-line degasser, then drop the sonication. If you have He sparging, drop the sonication. In fact, I doubt there is any real need for sonication anyway. The most likely result is that the solution warms over the course of an hour, and the more volatile components will evaporate faster, changing your composition. This will also happen if you He sparge during the run, and the sparging is too fast. You could have drifting retention times because your source composition is not constant. It's strange to have RTs go in, then out, ... but LC's are always finding new ways to surprise me!
3. Methanol and acetic acid will form methyl acetate on standing, although it is a slow reaction (accelerated by heating). And MeOAc would have different elution strength properties. (One of those crazy suggestions, since nothing else seems to make sense!)
4. Is there any possibility that you have an immiscible, or only barely miscible, mobile phase system? Doesn't seem likely, but look at the container, and make sure that you don't see any mixing effects when you swirl it.

Well, that's enough goofy suggestions for one post.

[juddc]

I dont have a guard column as samples from dissolution are filtered and USP standards are totally soluble in water.

Are you running a dissolution apparatus or are you simply talking about dissolving your drug product?

From everything you've mentioned it sounds to me (again) like a column fouling issue and depending upon what it's being fouled with, a methanol wash may very well not do any good.

If you're running a dissolution apparatus and your samples are in anything like SIF or SGF, I'd expect severe fouling if no sample pre-treatment was performed. Furthermore, a methanol rinse would do no good.

If you're running a drug product dissolved in water, then I'd have a close look at your excipient list and consider options for cleanup.

Again, do you see any deterioration in your chromatography over time when you inject standards only on a fresh column?

[mbicking]

Another tool to use for troubleshooting: calculate what I call the "Change Ratio". This is the ratio of the "bad"/"good" values for retention time, area, and height. Use your first "good" injection as a reference and then do the calculations for the runs that drift both in and out.

Either report those results here or send them to me privately. From the trends in the three ratios, some problems can be eliminated.

[Bruce Hamilton]

This is like pulling teeth, one by one, slowly, with pliers. Fortunately, they are not my teeth....

If you are performing the dissolution test, I'm confused.

Unless the method has changed, the dissolution samples and standards are in 40% water, ~ 60% of 95:5 MeoH:HOAc ( which also includes the 6% of methanolic Internal Standard solution ).

The standards are dissolved in 95:5 MeoH:HOAc for the assay, and the 40% water in the sample solvent is only used for the dissolution test. The solubility of the standards in water is irrelevent, as they aren't directly dissolved in water solution at any stage.

As I noted above, the use of a methanol rinse is unlikely to do much, especially if you are performing dissolution tests.

My first trial would be to not flush an older column ( after all they're pretty much a consumable to you at the moment ) after a sequence.
I'd run a sequence that started and finished with blank, standards, blank.

I'd leave the column overnight with the acidic mobile phase, and repeat the analysis the next day. Are the peaks still early, has the pressure changed?. Then I'd flush the column with a highly-aqueous mobile phase, allow to re-equilibrate and reanalye the same blank, standards, blank, sample solutions. I'd look for retention and pressure differences.

The indications are that you are blocking the capacity of the column with something, most likely from your samples. How you remove that material will be determined by the contaminant.

Please keep having fun,

Bruce Hamilton

[gurkishan]

Well, i am open to any suggestion that is possible!!
But, let me add fuel to the fire!

I took the help of another lab and send in the samples to them.
They too did not get results from the USP method and used a 25 cm column instead of 10 cm column===== as:
1) they were Not getting the stable peaks ie moving in peaks
2) resolution was bad
Using 25 cm column, they could afford the peaks to move a little without integrating into each other.
So, i guess, similar problem ....

[gurkishan]

Hi mbicking!
I have an 8 months old Waters instrument HPLC with Binary pumps.
DO u want the details of the machine?
I sonnicate it to degass it and then keep the solution overnight so that it can be at room temp next morning when i use it. The bottle cap is closed in sonnication, so no chance of evaporation.
Also, i dont think it will be forming methyacetate!

I am confused about how you do Change ratios and the significance. Could you please shed some light on this?!

Hi Juddc!
I run standards first and then filtered diss samples to compare them.Then i inject the std again to check their reproducibility!
The retention time of std at starting and ending of inj run is different. Also, i have used 3 columns in 2 weeks-- dont think they would be fouling up that quickly! The deterioration's are also there with the standards only.

Also, the excipients are ok, dont see prob with them.

Ya Bruce,
This is exactly what i am feeling--- nothing seems to be working!
I will try to flush the column with water and get back to you!!

Thanks guys! Maybe soon we will find the solution!!