LTQ orbitrap

[jiangds06]

anyone use this instrument before. it seems that this instrument is not stable for bio-analytical study.

[leadazide]

I would be very surprised if the Orbitrap can not live up to your expectations.. What do you mean with stable? Is the mass stability not good enough? Or are the peak areas not reproducable?

What setup are you running? Typical HPLC with ESI probe or NSI with Nanosource? What sort of method setup do you have? Full scan/MS/MS/Data Dependent??

[jiangds06]

I would be very surprised if the Orbitrap can not live up to your expectations.. What do you mean with stable? Is the mass stability not good enough? Or are the peak areas not reproducable?

What setup are you running? Typical HPLC with ESI probe or NSI with Nanosource? What sort of method setup do you have? Full scan/MS/MS/Data Dependent??

For some flavanoids analysis, UV is very reproducible but MS did not get consistent peak area. It happens even after cleaning the ion source.

[lmh]

MS peak area is never as consistent as UV area because ionisation efficiency depends on so many things (even getting the amount of drying-gas wrong can affect it). Also if you are combining, say, UPLC-style chromatography with quantification from 100,000 resolution you will probably find that the scan speed is slow enough that you only have one or two scans across the chromatographic peak, in which case the area will be fairly random anyway.

If it's general instrument efficiency that is varying, you could consider using an internal standard (which will also help with sample prep variations).

[jiangds06]

for some injections, peak can vary 20%, while UV only vary 2%.
Is this normal?

MS peak area is never as consistent as UV area because ionisation efficiency depends on so many things (even getting the amount of drying-gas wrong can affect it). Also if you are combining, say, UPLC-style chromatography with quantification from 100,000 resolution you will probably find that the scan speed is slow enough that you only have one or two scans across the chromatographic peak, in which case the area will be fairly random anyway.

If it's general instrument efficiency that is varying, you could consider using an internal standard (which will also help with sample prep variations).

[lmh]

I'll vote "slightly large, but not impossibly so", and see if anyone disagrees/agrees!

[leadazide]

Depending on MS method settings and your chromatography.. 20% deviation is not impossible.. If you have large Max IT of 1000 ms or run at high resolution like 100.000 and your chromatography is very sharp (something like 3-8 sec peak width) you can easily miss the peak appex or due to improper setting space charge the trap..

[jiangds06]

I understand that large injection time can affect the numbers of points across the peaks.
For resolution, I do not understand? could you explain it a little bit more.

Depending on MS method settings and your chromatography.. 20% deviation is not impossible.. If you have large Max IT of 1000 ms or run at high resolution like 100.000 and your chromatography is very sharp (something like 3-8 sec peak width) you can easily miss the peak appex or due to improper setting space charge the trap..

[leadazide]

The higher resolution you select the longer time the transient is meassured in the Orbitrap..

A good way to visualize it is to infuse the calmix and while scanning in the FT mode see you IT time is around 6-7 ms.. Set the resoluton as low as possible.. The IT stays the same but you get more scans (look in top left corner of the scan window and you can see the scan number) Now set the resolution up to 100.000.. Now you can see the scans go a lot slower but the IT is still the same.

I hope this helps.

[lmh]

Another way to look at it:

Ignoring the Orbi bit for the moment, and thinking only of the low-resolution ion-trap on the front end, the LTQ part of the instrument:

Each scan cycle actually starts by the instrument opening the entrance lens to the trap for a very short time, and then scanning out the ions very very fast. Resolution is poor, but it uses this to assess the total number of ions in the trap. Based on this, it then opens the entrance lens for long enough to fill the trap to the optimum density of ions, up to a maximum opening-time that you specify in your tune file.

This system only works if ions are arriving at the same rate per milisec during the proper filling stage as they did during the fast pre-scan stage. If you have a very narrow peak that rises very abruptly, it can happen that the pre-scan happened just before the peak got going, and the instrument thought "hey, no ions. Better stay open for a looong time!". It then did so, and a huge bundle of ions came hurtling in as the peak rose dramatically. This then means that the trap is over-filled.

When it's over-filled, the ions interfere with one another (each ion in the trap feels not only the electrostatic field applied by the trap's electrodes, but the field from its neighbours), so the ions aren't trapped so well, don't sit tightly in limited positions in the trap, and are ejected from it over a much wider (and biased) time-span during the scan. This translates to bad mass resolution when viewed on the screen.

The orbitrap part will suffer from the same problem, because ions are collected in the low-res trap before being shunted on to the Orbi section.

To be honest, this was a far worse problem in the old low-res instruments than the newer Orbis. Without looking in detail I couldn't be sure, but I'd be far more inclined to think that poor repeatability will come down to varying ionisation efficiency or low number of scans across the peak (turning on the "stick" mode instead of point-to-point in Qualbrowser can be scary).

[gapeev]

As the very first thing I'd disable AGC. Not sure about Orbi, all other traps I worked with had it under user control. In my hands it worked very well for small molecule quantification.

The only caveat is that you will nail the ion accumulation time as too few ions won't give statistics and enough count whereas too many will space charge.

[lmh]

just in case anyone comes across this thread, I have to put the other point of view. I'm afraid I disagree completely with gapeev; it probably depends on your instrument, but disabling AGC altogether can be a terrible idea in some instruments, especially if you are analysing samples with a wide range of concentrations. Some traps actually have quite a narrow range of trap-filling in which they produce good results. Too few ions means really bad signal:noise ratio, too many means that mass accuracy is lost, peaks become broad and move, and that means that when you produce extracted ion chromatograms, you start to lose signal because at its apex the signal has a different mass than at its start and end. Worse, the centroiding algorithm depends on the shape of the peak being approximately what the manufacturer expected. My experience is that centroiding alogrithms (from Thermo in my case) are extremely good, even when isotope peaks are not well separated when viewed in profile mode, but they will fail if you over-fill the trap.

The only problem with the AGC is if the signal is changing so fast that the optimum trap-filling time determined at the pre-scan by the AGC is completely wrong for the main scan that follows immediately after the (very rapid) pre-scan. If the signal is changing that fast, then it is certainly going to change greatly during the actual length of the main scan, which means that your scan-speed isn't adequate to cope with the narrowness of your chromatographic peaks. Your scan, in this case, is the average of wildly different values, and you are drawing a line somewhere in a peak of unknown shape; this will give unreproducible integration results.

If you have worries about AGC, the correct thing to do is set the maximum scan time to something small enough to give adequate scans across a very weak peak. This way you will have AGC to protect you against over-filling, but if you under-fill, the maximum scan time will protect you against scans becoming too infrequent. The cost will be bad signal:noise ratio on the spectra during a weak peak, but that's life! You've just reached the limit of your instrument.

Incidentally, there is one very obscure situation where AGC will create a weird problem: if you take raw data out into a 3rd party piece of software that attempts to integrate peaks based on the assumption that all measured points are equally spaced along the time axis, the integrator will get the wrong value on an AGC instrument because scans will be faster at higher concentration. The points get closer as the peak rises. But that just means that the 3rd party software is being a bit naughty.