Compare Operations of MS to HPLC?

[Jade.Barker]

I know there are some big differences between Mass Spec and HPLC (both equipment and theory)... but what about *operation*?

I am job hunting right now, and I hope my HPLC experience will translate okay to a position as an MS operator too? I was responsible for everything HPLC in my old lab, Empower Administraor, Methods, Troubleshooting, Integration etc. Surely I could operate a Mass Spec with a little training? As long as a more senior operator was available for troubleshooting and PM... Then operations is just careful prep, data entry, and pushing a button?

Can anyone walk me through a general day as a Mass Spec Operator? For example, our HPLC operations:
Aliquote whole sample into 30 ml centrifuge tube
Spin samples 20 minutes
Decant supernatant into syringe ~ 3 ml
Express liquid through .2 filter into 1ml glass vial (3/4 full)
Seat vial in autosampler
Enter ID into computer
press "run and process"
After run print, print the chromatogram and report
Share data with comments

Which of these operation steps is different with a Mass Spec? Thank you Everyone for your Insights!

[JGK]

The only difference (and it's a biggie) between HLPC and LC-MS is the detector component.

All of the HPLC knowledge will be applicable to LC-MS. However, there will be new things to learn, such as:

MS calibration,setup and monitoring
Additional method validation components (such as Matrix effects)
Perfecting sample prep techniques ( especially for preparation of Blanks)

[Ron]

One other thing you will have to learn is to use different mobile phase modifiers, no phosphates, borates, or other non-volatile buffers. That is a very big difference between regular HPLC and LCMS. Also, pH in LCMS affects ionization as well as separation, which can lead to some really interesting method development experiences.

[Jade.Barker]

That makes it pretty clear, thanks for the input!

[lmh]

May I chip in?

Depends a lot on the sort of work you intend to do. If you're using LC-MS for routine measurement, yes, it's very like hplc. You will need to know about internal standard calibration, as it's very valuable in this situation where ionisation efficiency, unlike absorbance, is usually variable from run to run (and even from minute to minute).

If you intend to do method development, you may need to know about using triple quads for SRM, and the difference between SRM, SIM etc., but that won't take long.

If you want to get the best out of your instrument, you may need quite a detailed knowledge of it, and unfortunately there are huge numbers of different MS instruments with different characteristics. To give an example, some years ago we had problems with an ion-trap SIM/SRM method where the middle of each peak disappeared. Reason: the manufacturer's default target for number of ions to be packed into the trap was a bit optimistic for our analyte. In the middle of the peak, the trap was too full, ions were interacting electrostatically ("space charging"), and the trap's resolution was no good. This led to the mass wandering slightly, and the ions not trapping properly (and in SIM, just wandering out of the SIM window).

Personally, in an interview, I would favour the quick-learner over the know-it-already candidate; things change, and a detailed knowledge of an old-fashioned ion trap may be fairly useless in working with a triple quad or an FT instrument.

But big proviso: if you go into a lab that really centres on mass spectrometry rather than hplc-with-a-clever-detector, there's lots of new stuff. The chemistry of gas-phase ions is a weird and wonderful world. The books concentrate nearly exclusively on GC electron ionisation, which is completely different to LC-CID. Identification of big, wobbly novel compounds from LC-MS alone is, I think, impossible, but lots of labs would disagree. They have knowledge I can only dream of.

To end on a nice note: LC-MS is usually much easier than hplc. Less likelihood of peaks coeluting, much clearer what's going on. You can even mix all your standards together in one tube and still know which is which. No funny solvents to mix up, no buffers to precipitate in the hplc. No need to understand ion exchange...

[Kostas Petritis]

It would have been so much easier for you if you could spend a month or so in an LC-MS laboratory so you can claim MS operation experience as well. The problem is that by now, there are a lot of people that can claim LC-MS background/method development on the market and you will have to compete against them... You will have to make a much better impression on other areas in order to be considered for such job in the presence of other people with experience...

[HW Mueller]

Imh, LC-MS is a synonym for sloppy HPLC hooked up to a MS?

[lmh]

Philosophical rant mode:

HWMueller, I'm glad you rose to the bait! Thank you.

I didn't mean it entirely seriously, but I do believe that much, much more caution is necessary in hplc than in LC-MS. HPLC is like driving a car with no windows. You go three km down the road and turn left and hope it's the correct road. There's lots of scope for confusion when retention time is the only indication of what we're measuring. In fact, I don't believe HPLC is suitable for analysis of unknown mixtures. I feel it should only be used where we know all the components, and have evidence that relevant ones do not coelute.

Of course checking for specificity is part of developing a reputable method, but in academic circles it's quite normal to use methods outside their original application (on a new organism, after a new treatment) and assume they still work.

PDA adds a lot to hplc, but even then, there are a frightening number of chemicals that have nearly identical UV spectra and very similar hydrophobicity.

So I suppose what I mean is: LC-MS is like driving a car where you can see where you're going. Yes, great caution is still needed to prevent accidents and to navigate correctly, sloppiness is unacceptable, but at least you can see the signposts and check what road you're on. HPLC needs enormously more caution, in proportion to the reduced information. Anyone who can do good HPLC and get reliable results, should find LC-MS a pleasant change, rather than a terrible challenge.

Rant over.

[HW Mueller]

I have done MS for almost 40 years and, therefore, know that it also has limits. Modern MS of biological substances seems to have even more problems, I just read
Bernd O. Keller, Lab International, November 2008, page 18.

[lmh]

Oh, absolutely. Even NIST contains plenty of rubbish spectra for the unwary to match.

I'm a beginner, having been doing LC-MS for only about 7 years, but even so, as I talk to people and read papers, quite often I struggle to keep my jaw closed and a polite smile on my face while I'm thinking "... and on the basis of this evidence, you actually believe what you're telling me?".

[Peter Apps]

I am continually amazed by how many people think that a 70% fit to a library spectrum constitutes an identification.

Peter

[lmh]

true, so true.

I'm feeling a bit guilty about hijacking this thread for a rant, so I'll go back to the original question, and do a walk-through of my day as an LC-MS operator. To put it in context, ours is a small academic facility serving the needs of a biological/biotechnical institute with a strong molecular biology interest.

I usually start the day by checking any overnight runs completed OK, and if not, diagnosing what went wrong and restarting; otherwise if data handling is to be done by the client, I check the results aren't blatant rubbish, and if not, contact client to let them know.

Then usually it's time to prepare fresh solvents, check log-books are completed, etc., before considering any direct infusions.

I still do most direct infusion MS by hand with a syringe pump. These are usually for synthetic chemists to check their products, though occasionally they're hplc fractions collected elsewhere, or other samples reckoned to be too simple to merit chromatography. Although we can automate it, especially where we just need a simple spectrum, I've found that when a reaction has gone awry, or a product is not pure, the chemists find it helpful if I can tell them what else is in the tube. Also this is good practice for me in MS interpretation, as it's much easier than a genuine unknown!

Most of the rest of the day will be spent preparing reports for clients, including going through calibration, integration and quantification of any quantitative LC-MS. My clients are a mix of qualitative and quantitative. Generally they have multiple samples belonging to one experiment, so rather than producing individual reports per sample, we are usually in the business of producing spreadsheets summarising the whole experiment. Fortunately we don't have to stick to quite the rigorous procedures of a QC lab, but we try to educate our users in proper technique! Qualitative work is usually a matter of looking for presence/absence of members of families of analytes, and identifying individual peaks given some prior knowledge of what we're looking for. I won't comment on "identification by wikipedia", but would admit to using any source of information available to me.

There are also some "metabolomic" clients, who will want us to find the differences between samples, and then investigate what they might be. Unfortunately we often have to warn people that they will need NMR (with much larger amounts of material) for identification.

Towards the end of the day, I prepare samples, do the log-book stuff again, and set up runs overnight. A fair proportion of our clients want different methods each time; the compounds of interest may vary from day to day.

In addition, there are the usual lab-manager issues, student visits, talking to bioinformaticists etc., and going to seminars, but they are the same for everyone.

Hope this helps! And good luck with job hunting.

[Jade.Barker]

I am continually amazed by how many people think that a 70% fit to a library spectrum constitutes an identification...

Honestly, 70%? I had no idea. Wow, um... To make a wildly unapplicable comparison, don't they say humans and chimps are 98% alike? The 2% is a huge deal... how can 30% not be?

I'm gonna be freaked out for a few days now. My naive fantasy of science being like I imagined it as a child is totally being crushed like eggshells. Honestly... 70% and thats common? I might have nightmares now.

[Jade.Barker]

I'm feeling a bit guilty...

Please don't feel bad, Consider this an invite to hijack away... Usually I'm just trying to spark a conversation so that I can gleen something from, since I havn't worked with anyone experienced/ trained in this area. Often the book reading is so dry it seems esoteric. It *really* helps to hear real people talking normally. Thanks for the interest everyone, it's a big help to a newbie like me.

[Peter Apps]

Hi Jade

You shouldn't worry too much about the 70%, mostly it happens in a narrow field of GC-MS application in which nobody will die (or even be mildly inconvenienced ) if the identification is wrong. What it points to is a high level of naive faith in the powers of analytical instruments and black box software.

Peter