Please can you share your experience of embedded polar group RP columns (like Supelco Discovery Amide, Waters SymmetryShield and the like) as far as HPLC-MS is concerned. I heard that their use was not possible because they bleed too much in MS use. Has anyone tried this? If so, which ions bleed off-ie where does the group cleave? Is it hydrolysed to produce silanol groups or is only part of the ligand lost? If the columns do bleed, does this have any drawbacks for HPLC-UV use (apart from those doing prep. LC, who may not want the stuff in their fractions! Do you have any references to this problem?

Hello,
introduction of embedded polar group RP columns was a big step in direction of getting symmetric peaks for polar compounds. I don't like the term "Bleed", because this has nothing to do with GC capillaries problems when loosing stationary phase. It has nothing to do with hydrolysis of silanol groups. In LC-MS most applications run with TFA and a pH round about pH2,5 or pH2! If your stationary phase bonding is monomer, you have only one bonding to the surface. If it is polymeric (trifunctional) bonded, this is much more stable. So, if you are loosing ligand under acidic conditions, you cannot detect this in UV. And please, if this happens, it is not a kg of ligand, and therefore never a problem for prep. LC .
With LC-MS, always important what are your expectations to MS, you can see sometimes a signal, coming from your ligand. This signal should be somewhere at m/z390! If this m/z area is not of interesst for you, than you can use also an embedded polar group RP column without any problem.
My statement is, that you can use every RP column for LCMS, also every dimension. It's all dependent from your instrument (flow rate limit) and what do you want to detect (mass sensitivity and m/z range). And please forget this "marketing" Bleed!
Gerhard

To add to Gerhardt's comments:
Indeed, you will see the problem only at specific masses. And even then, it is not a giant problem. According to our results, columns with an embedded polar group do not "bleed" any more than a standard C18 column. The real issue is that one does not "see" the C18 in the MS, but one does "see" stuff coming from columns with an embedded polar group. The specific m/z depends on the packing that you are using.

Uwe and Gerhardt:
Thanks for taking the time to help me. Uwe-you seem to be saying that the fragments coming off the embedded polar columns ionises easily in the MS, but fragments from ordinary C18 columns do not. Did I get this right? Gerhardt-what is the problem with the term "bleed" which refers to loss of the stationary phase during use-isn't this what is happening? Does it matter if it is caused by volatilisation in GC or dissolution in HPLC? If bleed has nothing to do with hydrolysis of the ligand to produce silanol groups-then where does it come from? What structure do you suggest for ions at 390? Gerhadt, you also confused me with your statement that most people used TFA at low pH. I thought this led to signal suppression in MS and that formic acid was better?
If nothing else, a 0.1% solution of formic acid has a higher pH than 0.1% TFA which might lessen ligand hydrolysis.

Hi,
sorry for the confusion I caused. Many applications you can find using TFA, also LCMS applications. But you are right, with formic acid in MS you get better signals. Column bleed to me means that stationary phase is coming out of the column. The strip off from C18 alkyl chains under acidic conditions I would not say that this is "Bleed" of the column, it is not hydrolysis of the silica matrix. And this happens with every RP-Phase in the world, with monomer bonded alkyl chains more than with polymeric bonded! With embedded polar groups you have the embedded polar group between the alkyl chain and the silica surface. And what you see in MS is the embedded polar group, not the alkyl chain. And this embedded group could have Amid structure, some have just a Urea group, and also other polar groups can be used. So the structure, or mass, depends on what polar group is used. All bring us this so called "Shield" effect! The polar group, when it is no longer embedded, but stripped off, can be ionized in the MS, but an C18 alkyl chain not. That's it. Please ask Uwe to send you one of the publications, explaining this technology.
Hope confusion is gone. Gerhard

Actually, hydrolysis can indeed result in loss of the entire ligand or just the portion of the substituent attached to the polar linker. This is particularly prevalent with amide based linkers so if you're worried about "blead" I think you would be better off staying away from amide based phases.

Chris,
In everyone of the studies that I am aware of, we have found that the Si-O-Si bond is the weakest link, at least for our phases with an incorporated carbamate group. While I have not studied the amide phases from this aspect, I would be surprised if this would be different.
Do you have any better info?
Uwe

I disagree with a statement mentioned earlier that if the bleed compound is not at your m/z of interest, then it does not matter. One has to realize that if you can see it by MS than it is competing for ionization with your analyte. Therefore, in other words, it suppreses the signal of your analyte. Of course the extent of the supression largerly depends on the structures of your analyte as well as the bleed compound. Therefore if I had to choose, I would use column with as little MS ionizable bleed as possible regardless where the bleed is coming from.
Vladimir.

Vladimir,
I have heard this argument as well, but I can't see the validity. The bleed from an LC column is extremely low, about 5 to 10x the normal MS noise. It is so low that in three months of operation, one barely sees a shift in retention. That means that the background is in the order of 1x10^-9 g/mL. I know that the LC background of plasma samples is several orders of magnitudes higher. At this level, I would worry about background, but people use LC/MS for this type of analysis anyway. Unless somebody can demonstrate to me that one can get ion suppression from the minuscule bleed from a column, I don't buy it.
In addition, why do you think that there is less bleed from a standard LC column than from one with an embedded polar group? If you don't see it, that does not mean that it is not there...

It's agood point that the bleed compounds are at very low levels. Like I said in my previous message, the extent of the suppression depends on ionization efficiencies of both the compound of interest as well as (!) the bleed compound (i.e. more basic one will ionize better in ESI+ mode). I have seen (and have done) applications in ESI- mode in the presence of high concentration of acid (really no no for MS) and method worked excellent. But I also had an application where I replaced a column with a different lot# and the different bleed characteriscics of the column invalidated my method (ESI+ mode). I never suggested that the regular columns bleed less. One keyword you might have overlooked in my previous message is ionizable (!) bleed. It just happened that the bleed from the polar embeded column ionizes at the experimental conditions used. And it's the ionizable bleed that gets you. The regular columns might bleed just as much or even more but if the "stuf" doesn't ionize at your ionization conditions, you don't have to worry about it too much. But you still must worry about it a little, it is still there and therefore might form adducts and clusters with your compound of interest. Hence, decreasing your signal at your m/z. But more or less, with currently available rugged ionization sources you can control the extent of adduct formation by adjusting your ionization conditions.
Vladimir.

Vlad,
it confuses me that you said the bleed will compete with interested sample molecule for ionization, i don't remember reading any theory that the ionization efficiency depends largely on co-eluents, because compared to mobile phase and fluents in ionizztion chamber, they look very minimal to account for.