LCMS Troubleshooting: Increasing Quantification on Standards

[StandridgeS]

Hi all,

I am fairly new to HPLC/LCMS, and I am currently working on method development on an Agilent 1200 Series HPLC and 6100 MSD that has not been in use for about 5 years. My goal is quantification of certain additives in polymers.

I create standards by making a stock solution and diluting it down to a range of 100-250ppm, and my calibration curve is linear. I have remade the standards about 8 times, and my results are not consistent with one another, they appear to be increasing for each concentration, but linear for the same run.

The same trend goes for actual samples that I run on the machine. There is a definite pattern of increasing peak area, more consistent than the standards, where they should all be the same.

I thought the problem initially was contamination of the needle, or some other carryover from the previous sample. I fixed the issue by adding a needle wash and a longer post-run purge of the tubing. This unfortunately hasn't seemed to fix the issue with my increasing standards.

Any suggestions to what my issue is?

[Gizmo]

What is the time frame within which you will see this increase? If you prepare your standard curve and then run it again immediately after is this increase apparent? Or do you see the increase when you run it a few hours later or the next day? Also, what is the magnitude of the increase that you see, 10%, 50%?

[StandridgeS]

It is about a 20% increase between runs, and I see the difference in the time frame of a day.

If I run the same sample repeatedly, I have seen as much as a 20% difference in peak areas, but the values appear to stay within this range for each sample.

I have not had the opportunity to study one standard I created, over a long period of time. For one sample, I did run in about a week after creating it, and the peak area was substantially larger (30-40%?) than the first time I ran it. My thought is that the solvent evaporated, increasing the concentration of my sample, or that follows the same trend of increase.

I have attached two images of the trends I am seeing over time. One image is the standard at 100ppm over time, and the other image is my sample of extracted material, supposedly at 100ppm, over time. There are about 1-2 days between trials.

[James_Ball]

You said the instrument has not been in use for 5 years, how long from the time you started it back up until you ran the first standards?

Are you allowing it to run constantly or shutting down between runs?

LCMS needs to be under vacuum constantly even when not in use. Also as you begin to bring one online especially after such a long time being dormant, it can take several days to finally reach a thermal equilibrium and pump out all of the air and moisture that has accumulated in the system, especially in any ceramic or polymer insulators.

Is the vacuum reading consistent or is it still going to lower pressures over the time of the analysis you are performing? If it is still going lower then it is still pumping out contaminates and the sensitivity will increase as the vacuum increases.

[StandridgeS]

It had been a couple hours since starting it up before running the initial standards. Since then, the vacuum has been running constantly, and should be at a consistent reading (but I'll double check).

However, I have had amounts of time where there is no nitrogen fed to the MSD because I've been learning how long it takes for one nitrogen tank to be used up.

Would this account for a change in 20% difference in my peak areas though? I am in the process of acquiring a nitrogen generator, which I hope will solve this problem.

[James_Ball]

If the instrument is using nitrogen as the sweep gas on the interface to keep things like moisture out then it could cause some problems once the tank runs out. Using a nitrogen flow there(usually between the inner and outer cones) keeps oxygen and moisture out of the instrument and could help it to be more stable. Being under constant vacuum and having all the components at thermal equilibrium also helps stabilize the system.

[StandridgeS]

That does make sense about the nitrogen and vacuum, but after a couple days of having everything at the proper conditions, the sensitivity should stabilize to the same value, correct? It doesn't make sense to me that my peak areas would keep increasing in a consistent amount like I have shown in the graphs.

[James_Ball]

If the quads are dirty you can get charging which will cause variability in sensitivity. On the other hand if the system is super clean, running samples can cap off active sites that might adsorb analytes and over time as it become a little dirty the sensitivity will increase.

Have you tried setting it up to run overnight just injecting blanks or a standard and see if it will eventually stabilize? I have done that before to condition an instrument after maintenance or when it is new.

[StandridgeS]

I'll try that with standards! How many injections do you normally set up to run overnight?

[James_Ball]

With 30 minute analysis times I use 20-30 injections, this will keep the instrument busy most of the night. Just don't exceed the time you have available with the mobile phase you have. If it is still running in the morning you can just end the sequence. With 10 hours or more of data you can get an idea if it is changing over time and if there is a point where it becomes stable.

[lmh]

... just a side-note, seriously, you are quite right to get a nitrogen generator. When you do so, check the price of servicing (or spares if you intend to do it yourself) before you buy it. Some manufacturers make their profit on the sale, some on the service contract - and ruthlessly! Nitrogen generators with built-in compressors will absolutely definitely need regular service; the compressors collapse about once a year.

[StandridgeS]

Hi all,

After running about 20 injections of my sample overnight, the peak areas still varied about 20% without showing a pattern. I ran multiple injections (~10) for my standards again, and there was a pattern of decreasing peak areas that for some decreased by about half.

I have a PM scheduled for mid-October, so I hope with that and the nitrogen generator, the instrumental issues will be ruled out.