Optimising the ESI-MS for Poly sulfated sugars

[Faith]

I am new to UPLC-MS.

We have recently purchased a Shimadzu LCMS-2020 (ESI) single quad system, and I attempted to run a method that I use on a HILIC column on a HPLC-ELSD system (that works beautifully) to try and identify a particular peak, but also to get the mass of the main peak. (By NMR this poly sulfated sample is 90% pure).

The mobile phase is: A:5mM NH4 Formate; B:5mM NH4 Formate in 95%Acetonitrile and 5%Water
Gradient is 3%A:97%B to 15%A:85%B in 14 mins, hold for 5 mins and then back down again.

Chromatographywise, the peaks come out about the same time as they did on the HPLC-ELSD system (as both systems have a splitter). However, some peaks are negative (in either negative or both neg and pos scan modes). And the main peak looks like a saw. That is the tail is jagged. And the mass I obtained across this peak varied, none of them being the correct mass of the poly sulfate.

My question is: These poly sulfated compounds are very unstable. And so I am thinking the sulfates are falling off because my MS conditions are too harsh. The mass spec is tuned to the tuning solution, but if the tuning solution is nothing like the sample then perhaps I cannot use the MS default values, and I have to optimise the conditions. This is what I have no idea about. What parameters should I look at changing first to try and be a bit more gentle? That is what parameters do Mass spec'ers change when optimizing?

I would be so grateful for any help, as this MS detector works beautifully with other compounds (non sulfated) I have tried on it, and used the default conditions.

[carlo.annaratone]

1) were you looking for the quasimolecular ion or did you look for adducts as well? Only negative or both positive and negative?

2) optimization of the source conditions can be either while injecting a standard from a syringe or during "real" run conditions. Things I would check: electrospray voltage, cone (or capillary voltage - depends on the design of your apapratus), gas temperature.

[JI2002]

I cannot provide MS conditions to try, but the sample is almost certainly a mixture of a series of polymeric components. Ions with different m/z values are expected for this type of sample. A mass difference equivalent to the mass of the monomer of poly sulfates sugar should be observed for the ions.

[Faith]

Hello Carlo,
I looked for everything, in both positive and negative scans.
I am not at work today, and tomorrow is the weekend but when I get back I will see if I can put up a trace.
I will check the parameters you suggested, as that is exactly what I wanted to know. Thank-you for that.

JI2002 I understand that it would be a mixture of these ions/components however I am getting way to many and that's why I feel I am hitting it too hard and want to go in more softly. And...I just wanted to know what are the parameters you normally adjust?

I am now going to play with the cone and electrospray voltage and the gas temperature as suggested, and see how this goes.

Thankyou both for your time, thoughts and suggestions.

[James_Ball]

I haven't done much single quad work, but I know when I am working on my QQQ setup I normally try to infuse the target in solution with what would be similar to the mobile phase as it elutes. This helps to tune the instrument for the conditions that it will see when reading the analyte.

My instrument has a syringe pump on it and I use a Tee connector to put that in line with the column effluent before it goes into the source so I can tune at the same flow as the analytical run when I am finishing my setup so I get maximum sensitivity.

[lmh]

some thoughts:

(1) negative peaks are often a sign that you've got co-suppression; there may be one or more background ions present all the time, and ionising. When something else elutes, although you fail to see it, it disturbs the ionisation of the background ion, which means the total ion chromatogram has a negative peak (as does the extracted mass chromatogram for the affected background ion, and the base peak chromatogram if the background ion is the most abundant thing present).

This means your chromatography may be working fine, but you are not detecting the compound you want to see.

(2) Yes, I can understand that polysulphated compounds may be very unstable so ESI conditions may be messing things up. When things go wrong, I find source-temperature is generally worth some experiment. The problem on a single quad is that if you drop the temperature too much, drying will be inefficient, and this will probably lead to tiny random solvent droplets whizzing through the instrument at random times, and showing up as sudden spikes in spectra (at random masses). The symptom is successive spectra that look as though they have one or more clear masses, but no spectrum is the same as the next!

(3) But another problem with polysulphated things (or anything with multiple acid groups) is that there are so many masses at which you might see it. For example, in negative mode, a thing with 3 acid groups could show up as triply-charged, doubly-charged with one charge neutralised by sodium, hydrogen, potassium etc., or singly charged with two charges neutralised by sodium, one sodium + one hydrogen, etc. etc.; the combinations are legion. On top of that, I often find with Hilic methods that things suddenly appear as formate adducts and such-like. You need to check very carefully that any masses present in the area of the expected peak might not be the right thing.

(4) ELSD (which you've been using up to now, I think?) is not very sensitive. You're actually looking for solid stuff that drops out when you dry the sample. This sort of solid stuff is exactly the thing that blocks mass specs! NMR is also not very sensitive. If you have injected a few samples at ELSD-appropriate concentrations in a mass spec, it is just possible your mass spec is overwhelmed by general solid stuff and possibly even getting a bit blocked up (the desorbation line or whatever it's called on that instrument). Curiously, there are situations where injecting too much will give less signal and more problems. If you think there is even the most remote chance that this has happened, I would suggest running some compound that you know works well, to check that the instrument is in good condition, and when you are satisfied that it is, running some very dilute polysulphated samples.

Incidentally, I agree with James Ball that if you have the possibility of doing an infusion, then it does help to separate mass spec problems from chromatography problems. If you don't have a syringe pump, you can also run injections with no column in place; it might be worth trying this with a simple 50/50 water acetonitrile solvent (initially miss out the ammonium formate in case it's causing problems).

[Faith]

Thankyou "James_Ball" for your suggestion, I haven't got a syringe pump but I will try what "Imh" suggested that-is... I will perform the injection without a column, with and without the ammonium formate at 50:50 Acetonitrile and water.

"Imh"
Referring to your -pt (1) I thought something along those lines myself when I tried to get my head around obtaining the negative peaks.

-Pt (2) Am looking forward to (hopefully) playing with the source temperature today.
If what you suggests may happen (that I get these random masses), if I drop the temperature too much. How do I clear this? Do I up the temperature and drying gas for a while?

-Pt (3) Yes your so right about this point, luckily my colleague who synthesized the compound is brilliant with working this out.

-Pt (4) Thankyou for this point, I know exactly what you are talking about as this was the first compound I worked on when we purchased this instrument (big mistake) as I totally blocked it!

Thankyou all so much for your help, I can not tell you enough how much I appreciate it. Will let you know how it goes.

[lmh]

hope it's going better! I don't, I'm afraid, know a solution to the problem of noisy spectra because of inadequate drying. If you're unlucky and get a very heat-labile compound, it can be difficult to get good electrospray data. The good thing is that if you get a nice chromatographic peak then the compound cannot have been badly degraded during separation - the problem is that if it is destroyed on detection, the mass spec doesn't help much! Some things are just difficult by electrospray.

[carlo.annaratone]

Itotally agree with the above responses, would like to add some things:

temperature can be defined in many places depending on the vendor and the make of the instrument (e.g. desolvation gas, sampling cone/capillary, quadrupole, etc). Every one of them has a different effect, so you just have to experiment a bit. When decreasing temperature of the desolvation gas, I found helpful to increase a lot the deolvation flow. Make sure that you have enough pressure in your line or the highest settings might not be reproducible.

decreasing flow from column (either switching to a smaller diameter or using a splitter) helps in increasing sensitivity, reducing ion suppression phenomena, and require lower temperatures in the source for effective ionization

cleaning the source and the first vacuum stage of the mass spec (needs to break vacuum though) will help to increase ion yield and in turn require less harsh condition in the source. If the instrument has been used a lot with dirty samples (such as is expected to be in a synthesis lab) things can go dirty very quickly. As an example, I am working with very labile oxalate esters of sesquiterpenes, and I am getting excellent signal from my clean machine but NO SIGNAL AT ALL from a machine in the organic chem department.

try additives in the eluent. Some examples: changing the pH and/or buffer can have dramatic effects on ionization yield. Some people add volatile salts or trace amount of metals to measure the adducts in a reproducible fashion. Fluoride salts have been proven to improve ionization in neg mode without damaging the source. Chelants will remove trace metals and reduce adduct formation. And so on.... Additives can be added in the eluent (if they are compatible with chromatography mode) or post-column

[Faith]

Firstly thank you everyone for your help.

I detached the HPLC column, and ran the sample in 50%Acetonitrile:50%Water. After a few injections of different concentrations and different injection volumes. I found that injecting a less concentrated sample I actually obtained a better signal.

Secondly, (and I am very excited about this) after playing with the DL temperature reducing it from 250C to 200C (even though I kept the heating block at 200C). I obtained the correct mass. My colleague who synthesized the compound, has also put together an excel spreadsheet, which gives all the possible masses, whether it looses any SO3 groups, gains or loses Na and/or H. And I obtained a mass with it not losing any sulfate groups.

I then went to repeat this with 20%(5mM Ammonium formate):80%(5mM ammonium formate in 95%Acetonitrile+5%water) and could not see the peak at all.

Tried again without the ammonium formate in it i.e.50%Acetonitrile:50%Water, and the trace is much noisier now, but I at least obtain the peak/mass.

Today... I am going to put the HILIC column on and see if I can run this compound without the ammonium formate in it.

Wish me luck!

[carlo.annaratone]

no buffer = no HILIC for charged compounds. Do not expect it to work! Try first to change the pH, rather than removing the buffer completely. Despite someone stating the contrary, acidic buffer will work for negative ion ESI. Which column are you using? In my experience the most reliable columns are the Waters Xbridge range. They will also work well with alkaline buffers (my personal experience is that a Xbridge amide column will work for 2000+ injections of 70 % ACN + ammonium acetate 5 mM pH 7 at 80 degreees C). Regarding the masses, remember that negative mode can also give an adduct with formic acid [M+COO-]-

[Faith]

Yes... I was worried too Carlo that the peak would come out straight away on the Luna HILIC 3um 150x4.6mm column without the ammonium formate in it, (which it did under the same gradient conditions that I would normally use with the 5mM Ammonium formate in). This is such a dilute concentration, that I thought maybe I could get away without it present. To get a rough idea of the pH previously I collected the waste eluent and it was about pH7. Which was great because this compound decomposes under acidic, it is most stable around neutral conditions.
So I changed the gradient, I kept it at 5%Water and 95%Acetonitrile for a couple of mins (with a 1.3 mL/min flow-rate, (I have a splitter in place after the column) and also had a much lower slope with the gradient. Infact I only went up to 14%Water:86%Acetonitrile in 12 mins, and then stayed for for 5 mins. I obtained a peak, which was not in the "blank water" injection, nor in the "non" injection run. However, it is a 1 mins broad ugly peak, not any of the masses given to me with the sulfate groups intact.

The other condition I changed this week, instead of dissolving the compound in water, I dissolved it in 80%Water and 20%Acetonitrile. So that the sample eluate isn't way stronger than the initial HPLC conditions.

Today, I am going to wash the UPLC system out with 30% phosphoric acid (not with the column in it, nor the mass spec attached) am not sure if this will help with the negative peaks I obtained when I first started this project using the mass spec instead of the ELSD, but it may help clean it.

Then, I will try again with the HILIC column without ammonium formate, to see if I can improve that 1 min broad peak.

Then I will try ammonium acetate.

Also Carlo, I will look into the Waters Xbridge range of columns, thanks for that.

[carlo.annaratone]

Peak shape in HILIC is a nasty issue. In my experience, for charged compounds: 1) a surface-modified column is better than bare silica as it will give higher retention and better shape 2) a higher ionic strength is better than a lower one for peak shape. There can be marked differences between 3 and 5 mM. Check as well solubility issues for your compound. I am doing HILIC-ELSD of sugars, and found that solubility in high-ACN is a problem for concentrated samples. Minimize the mismatch between sample and eluent by dissolving in the mobile phase. Last, do not forget that if your compound still does not work in HILIC, there are weak ion-exchange columns that tolerate organic solvents and run with bicarbonate buffer (so MS friendly) - the one that comes to my mind now is the PRP-x110 series from Hamilton

[lmh]

(1) Ammonium formate: yes, I'm not advocating doing Hilic with no buffer (though there are people out there with buffer-free hilic methods that they swear work beautifully. I've never tried). I've just noticed that ammonium formate buffers can reduce the quality of electrospray MS results, so when I suspect that something has gone wrong with ionisation, they are one of the things that I like to temporarily remove.

(2) Sample solvent. This really is critical. If you dissolve the sample in 100% water you may find that even very tiny injection volumes become impossible, with peaks broadening so much that they are no longer detectable, or with complete loss of elution or random elution times. If you try to dissolve highly polar things in 100% acetonitrile they will precipitate (no peak). Finding a compromise concentration of acetonitrile that doesn't cause precipitation and doesn't mess up retention is a tricky balancing act.

(2a) Methanol, in Hilic, can be a strong solvent; anything that can form hydrogen bonds is "watery" and can mess up retention if it's used as a sample solvent.

(3) I'd second carlo's point that acids in the running buffer don't stop negative ionisation. Choose the pH that suits your chromatography, and don't try to improve electrospray ionisation by changing to a pH where the analyte would be more ionised in solution. I am not an expert, but I think what happens is that if there are any ions in solution, the electrospray voltage will draw them out, and they will be replaced by re-equilibration from the neutral form, so provided some tiny proportion of the analyte exists in ionised form, even if it's a really tiny proportion, the process can still work. I did once play around with post-column alteration of pH by teeing in a different buffer; naively you'd expect this to work, but in my few feeble efforts it made it worse, not better. I think it was because I was vastly increasing the ionic strength of the buffer, which isn't a good idea. But if anyone has other experiences, I'd love to know more.

[carlo.annaratone]

some relevant papers:

Gabelica, V.; De Pauw, E. Internal energy and fragmentation of ions produced in electrospray sources. Mass Spectrom. Rev. 2005, 24 (4), 566–587.

Niu, W.; Zhu, X.; Yu, K.; Li, L.; Sun, Y.; Li, C. Nebulizing conditions of pneumatic electrospray ionization significantly influence electrolyte effects on compound measurement. J. Mass Spectrom. 2012, 47 (3), 370–380.



Huffman, B. a; Poltash, M. L.; Hughey, C. a. Effect of polar protic and polar aprotic solvents on negative-ion electrospray ionization and chromatographic separation of small acidic molecules. Anal. Chem. 2012, 84 (22), 9942–9950.

Wu, Z.; Gao, W.; Phelps, M. A.; Wu, D.; Miller, D. D.; Dalton, J. T. Favorable effects of weak acids on negative-ion electrospray ionization mass spectrometry. Anal. Chem. 2004, 76 (3), 839–847.

Zhang, X.; Clausen, M. R.; Zhao, X.; Zheng, H.; Bertram, H. C. Enhancing the power of liquid chromatography-mass spectrometry-based urine metabolomics in negative ion mode by optimization of the additive. Anal. Chem. 2012, 84 (18), 7785–7792.


Dolci, M.; Pereira, L.; Milton, D.; Edge, T.; Scientific, T. F. HILIC Method Development in a Few Simple Steps. In; p 21029.

Kumar, A.; Heaton, J. C.; McCalley, D. V. Practical investigation of the factors that affect the selectivity in hydrophilic interaction chromatography. J. Chromatogr. A 2013, 1276, 33–46.

Heaton, J. C.; Russell, J. J.; Underwood, T.; Boughtflower, R.; McCalley, D. V. Comparison of peak shape in hydrophilic interaction chromatography using acidic salt buffers and simple acid solutions. J. Chromatogr. A 2014, 1347, 39–48.

Pesek, J. J.; Matyska, M. T.; Fischer, S. M. Improvement of peak shape in aqueous normal phase analysis of anionic metabolites. J. Sep. Sci. 2011, 34 (24), 3509–3516.

Fountain, K. J.; Xu, J.; Diehl, D. M.; Morrison, D. Influence of stationary phase chemistry and mobile-phase composition on retention, selectivity, and MS response in hydrophilic interaction chromatography. J. Sep. Sci. 2010, 33 (6-7), 740–751.

Greco, G.; Letzel, T. Main interactions and influences of the chromatographic parameters in HILIC separations. J. Chromatogr. Sci. 2013, 51 (7), 684–693.

Wan, E. C. H.; Yu, J. Z. Analysis of sugars and sugar polyols in atmospheric aerosols by chloride attachment in liquid chromatography/negative ion electrospray mass spectrometry. Environ. Sci. Technol. 2007, 41 (7), 2459–2466.