MS scan on MS1 or MS2 tripple quad?

[per_oxid]

Hello,
I work with a waters tripple quadrupole ms instrument and was in the process of verifying the ion transmission using waters own setup solution. I have done full scan on MS1 and a full scan using MS2 (no fragmentation collision gas off). I hade about twice as high peak height on MS1 for scan vs using the MS2 to scan. All parameters the same only difference is what quad is doing the scanning and the other just doing ion transmission.

Both scans look fine in terms of signal, calibration and resolution. The MS1 scan did give better signal by around 2 times the signal using MS2 for scanning. That was about the only difference I could see. They quads have a bit different resolution so some difference in signal intensity is to be expected as it goes up when resolution goes down, but they were fairly similar in resolution.

1) Is this an appropriate test for evaluating MS1 and MS2 ion transmission?
2) Is there any reason why MS1 is the preferred quadrupole for making the scan and MS2 as passive transmission of ions, and if so why (or was my result with MS1 being better just a coincidence)?

I guess I should know the answer to this but since i don't I would appreciate if someone could enlighten me.

Thanks!

[lmh]

Waters instruments, at least the modern ones, scan with the 2nd quadrupole when they're running normal methods, not the first (you can choose either in the tune page, with MS1 scan being with the first). The reason is that the scan speed with the first quadrupole is much slower than that with the second. This is a wild guess on my part, but I'm assuming that there is some spreading out of ions as they travel through the remainder of the system, post-quadrupole-filter, some going faster than others, which would translate into a wider mass peak, and to compensate, the system scans slower when using the first quadrupole.
Slower scan speed is obviously a disadvantage in a chromatography method because it means less points across each chromatographic peak. But to be honest, triple quads are not great as scanning instruments anyway.
Incidentally, on the resolution of the two quads, you can choose how well you want them to resolve. We use unit resolution for both, but there are numerous other modes. I can't see much point; you're not going to differentiate between near-isobaric ions by improving quadrupole resolution from 1.0 to even 0.3 m/z. General note on those instruments: if your resolution settings are 3.0 and 15.0, then the instrument hasn't had its resolution optimised! Those are the defaults. But Waters triple quads are pretty good even without being set up correctly!

[bunnahabhain]

One additional note: The collision gas should be "on" also for scanning methods.
(This is due to the "PIC scan" feature, so you can do MRM and MS scan at the same time)

[bcd_GCLCMSMS]

General note on those instruments: if your resolution settings are 3.0 and 15.0, then the instrument hasn't had its resolution optimised! Those are the defaults. But Waters triple quads are pretty good even without being set up correctly!

Could you please clarify this statement? We optimize the 'calibration' and 'resolution' with the Waters set-up/tuning solutions regularly. Our final results are near 3 and 15 regularly. Do you mean, if the values are exactly 3.0 and 15.0? (Or is there another resolution optimization step that we should be using?)

[per_oxid]

Thanks for the explanation. I did the test with scanning using ms1 and ms2 to see how the condition of the instrument was and I have some target values from waters to compare to. Normally I run with collision gas on this test was to see if we had lost sensitivity on the instrument since it was installed.

I use the low and high resolution parameters from intellistart calibration using setup solution. It is interesting how wide the tolerance is 0.6-0.9 with the target fwhm 0.75. At 0.6 the isotopes are almost fully resolved but at 0.9 they are barely resolved at all, still within the accepted range of intellistart.

[sbashkyrtsev]

This is a wild guess on my part, but I'm assuming that there is some spreading out of ions as they travel through the remainder of the system, post-quadrupole-filter, some going faster than others, which would translate into a wider mass peak

Hm.. But why slower/faster ions would result in a wider mass peak? Quads filter by accelerating ions radially, and you seem to be talking about axial velocity, like in the case of TOF.

Also a guess: in the case of a Product Ion Scan, for each m/z on the 1st Quad we need to scan for multiple m/z on the 2nd Quad. Since for this type of scan we need the 2nd Quad to work faster anyway, maybe there's a reason to limit the speed of the 1st Quad. Maybe this even allows to save on the expensive electronics. Or are these quads identical?

[lmh]

(1) on resolution settings: I meant if the values are exactly 3.0 and 15.0 then the tune file doesn't contain any resolution optimisation, because these are the defaults. Typically the results after doing resolution optimisation are pretty close to 3.0 and 15.0, so if you end up with 2.91 or something instead of 3.0, that's normal.

(2) To explain why fast and slow scanning makes a difference to mass peak width, it's like this: if a quadrupole is being used in scan mode, it might be scanning at 5000 m/z per second, or 5 m/z per msec.

The instrument tells us that an ion had a mass of m/z 100 based on the fact that it reached the detector at a time when the instrument thought the quadrupole had reached voltages that filter out everything that isn't m/z 100.

Ions tend to be accelerated through a triple quad with fairly low energies, particularly if they have to be coaxed through a collision cell without fragmenting. If ions are accelerated at just a few eV longitudinally through the system, they don't have vast linear velocities. One eV is I think about 1400 m/sec for an ion of 100amu, but don't quote me on that, do the calculation! If so, it's going to take about 0.2msec to travel 30cm. At a scan speed of 5 m/z per msec, that's a delay corresponding to a whole mass unit. Of course the instrument will take this into consideration when it's calibrated, but the problem happens if ions are travelling longitudinally through the system at different speeds because they've got slightly different kinetic energies. If an ion is going a bit fast, it will have a delay time of less than 0.2msec for 30cm and will arrive early, which the instrument will interpret as a lighter/heavier mass depending on whether it scans low to high or high to low. The extent of the mass error depends on how far the instrument had changed the settings on the quadrupole during the time discrepancy between when the ion actually arrived, and when it should have arrived. If the instrument is scanning really slowly, the nominal mass setting won't have changed much over the time discrepancy, so the ion will be plotted close to its correct point on the spectrum. If the instrument is scanning fast, its measured mass will be far further wrong, and it will be plotted further away from the correct place. I.e., a spread of kinetic energies translates into a spread of apparent mass, and the size of the spread depends on how fast the instrument was scanning.

[sbashkyrtsev]

One eV is I think about 1400 m/sec for an ion of 100amu, but don't quote me on that, do the calculation! If so, it's going to take about 0.2msec to travel 30cm.

Double checked, my calculations say ~979 m/s which corresponds to 0.3ms (:

if ions are travelling longitudinally through the system at different speeds because they've got slightly different kinetic energies. If an ion is going a bit fast, it will have a delay time of less than 0.2msec for 30cm and will arrive early, which the instrument will interpret as a lighter/heavier mass

I may be missing something else then.. My understanding was that these Mass Specs work this way (let's consider SQD for starters):
1. Ions of all masses are let into the quadrupole
2. At any given moment of time the quadrupole filters out all but one m/z - let's say 100
3. Some ions with m/z 100 reach the detector during that dwell time. But not necessarily all of the ions.
4. Now there's an interscan delay - this lets the rest of the 100 m/z ion that were still in flight to be filtered out.
5. Now that the quadrupole switched to pass 101 m/z, the detector is turned back on.

Isn't that how it works? Then regardless of the axial spread, the mass resolution shouldn't change..

And in the case of QqQ, as you mentioned, it's the 2nd Quad that does the filtering. So after the filtering they go straight to the detector. Which actually.. may be the reason why it's the 2nd quad - there's no way that the next batch of ions (with different m/z) can catch up with the previous batch. And now that I'm writing this.. This is probably exactly what you meant in your original answer (:

[bcd_GCLCMSMS]

(1) on resolution settings: I meant if the values are exactly 3.0 and 15.0 then the tune file doesn't contain any resolution optimisation, because these are the defaults. Typically the results after doing resolution optimisation are pretty close to 3.0 and 15.0, so if you end up with 2.91 or something instead of 3.0, that's normal.

Much appreciated - Thank you!

[lmh]

To be honest, I'm not sure, but I think you're describing a digitised version of what I'm describing.

The interscan delay to clear out any left-over ions definitely applies in MRM events, where you fix the quadrupoles to one particular voltage for one particular mass transition before jumping to another. I don't know exactly what happens during continuous scans.

Of course nothing is truly continuous nowadays; presumably the quadrupole voltages are controlled by some digital to analogue thing, and are actually creating a staircase, and presumably the instrument is measuring at each step. I don't know how finely it controls the staircase(*). It collects many more than one data-point per mass unit, so a scan from 100 to 600 amu is going to include several thousand of data-points. I don't know if it collects data continuously when it scans, moving to the next time-point at intervals, or whether it puts an interscan delay in after each step. I suspect it doesn't, or puts only a very short delay in, because you cannot even do hundreds of MRMs simultaneously without things getting very slow. A scan speed of 10,000 amu per second (Shimadzu quadrupoles go up to 15,000, I can't remember what Waters do) implies 0.1msec per mass unit, which is ridiculously short in MRM terms.

But in any case, if you fix a voltage and wait for an interscan delay, do some measurement, and then move on to the next voltage, you are still obliged to scan more slowly when using MS1 rather than MS2, because the longer flight-path from MS1 to detector will take longer to clear, and will need a longer interscan delay.

(* and given that everything has capacitance, even if the system is aiming to apply a staircase, if it goes too fast, it may well end up as a continuous slightly wobbly slope!)