Hi, I think you need to flush your C18 column with (95% H2O) for at least one entire day, then the salt in your column will be removed. Then restart the analysis, the sodium adduct should be less in intensity if not disappeared. So using sodium adduct ion as you targeted analyte for quantification won't be reproducible anyway, because the source of sodium could come from anywhere.

Dear Bradley, according to your information I expect that you observe two kind of ion-adducts : Na+ adduct and NH4+ adduct. In my opinion this fact can be a source of not-reproducibility of the mass signals. There is a competition for steroid adduct formation between the two ions. Using ESI source the best thing would be to have only one possibility to form positive ion adduct, not two. From molar concentrations I would guess that NH4+ adduct (25mM in HPLC mobile phase) should be more favoured with respect to Na+ adduct (1 mM). I have no experience in steroid separation and I do not know where your samples come from but if possible I would try to remove ammonium acetate from HPLC mobile phase and to keep only Sodium cloride. Furthermore you should also try to "clean" the ESI source in order to remove all the salts that have accumulated time after time before doing other experiments. With the ESI source is important to keep the quantity of salts as little as possible, even the volatile salts like ammomium acetate, to maintain "clean" the ESI source as much as possible. For steroid detection why not trying to use the Atmospheric Pressure Chemical Ionization source instead of the ESI source. With the APCI source you do not monitor an ion adduct but a protonated species should form Steroid-H+ (M-H+). Am I wrong? So you should by pass the confusion with different ion adducts).

I agree also with Salim suggestion.

Hi Bradley,

I would suggest lowering your ammonium acetate concentration to 10mM and tuning up your steroids in negative ESI and you should observe [m+acetate] negative ion adduct. Steroids give the greatest sensitivity in negative ion with adduct formation.

In general sodium adducts for quantification are difficult due to variable response and the difficulty of fragmentation. Ammonium adducts work well and optimise at much lower source voltages.

I hope the above is helpful and good luck.

Mohammed Abrar

Use of Na+ in your mobile phase is not advisable because it deposits in the source and eventually the signal will keep on dropping.

Salim's suggestion is right.

I suggest this paper Fanelli F Steroids 76 (2011) , it looks like a nice quantitative method.

1 mM is generally to concentrated. Perhaps 100 uM or less. The problem is that it will depend on the concentration of the steroids. Actually rather than Na you could try Li adducts. Also you should reduce the ammonium acetate to no more than 10 mM. What steroids are you measuring and what is the sample?

A trick I've used for other compounds (not steroids) that form unwanted Na adducts that won't fragment is to use a basic mobile phase. By using NH4OH based mobile phase you favour [M+NH4]+ and [M+H]+ adducts over [M+Na]+. It's surprising how well positive ionization works without an acid in the mobile phase. Of course make sure you choose a column that can handle the higher pH. If you can use negative ion, then go for it. I would avoid using non-volatile salts in your mobile phase, it will cause problems at some stage.

Andrew, you're right. the adductted ion seems need a higher energy to break down and also after breaking down, there seems no domain fragments.

Hi Bradley..

The major drawback for the quantification of steroids by LC-MS/MS is the lack of acidic or basic functional group in their structure, which disables ionization by API, ESI, APCI or APPI. The only way to quantitate steroids is by improving their ionization efficiency by different means. One such way to improve their ionization is through CHEMICAL DERIVATIZATION by a suitable derivatization reagent that are commercially available.

Here is a free mini review. It could help you about stoeroid analysis with MS^n on Clinical Chemistry 55(6)_1061–1066 (2009)

"Steroid Hormone Analysis by Tandem Mass Spectrometry Steven J. Soldin&Offie P. Soldin"

I use either 0.1% formic acid of 10 mM Ammonium Formate pH 3.5 as MP A spiked with 1 uM sodium formate to enhance Na+ adducts. 1 mM sodium does seem a bit high.

Thank you to all the contributors for this question, your experience and suggestions are greatly appreciated.

I am somewhat limited in the method I have available. I am using UPLCMS to track steroid sulfate hydrolysis by various sulfatases. I only have ESI available and do not have tandem MS in my lab.

I am trying to develop a method that allows direct injection and analysis of reactions for either detailed kinetic analysis or medium throughput screening. As such, there is always Na+ (up to 40 mM in a 10 µL injection) in the samples and I do not want to waste time or resources on extractions.

I am interested in detecting androgenic anabolic steroids and have been focusing on testosterone (since I can use UV abs to validate the quantity, while optimising the MS detection).

Based on the suggestions so far I have tried using aqueous methanol with either 0.1% formic acid or 10 mM NH4 acetate to look for adducts of formate, acetate, or NH4. I did not observe any -ve ion adducts at source voltages ranging from 500 to 3000 V or fragmentation voltages ranging from 50 to 200 V. The only significant +ve ions are: M+H+, (M-OH)+, and M+Na+. In low Na+ mobile phase the M+H+ ion is observed with a minor M+Na+. If Na+ is present then this is reversed. Even if I could find another adduct in these conditions, I suspect that its signal, like that of M+H+ would be attenuated by Na+.

The negative ion mode in high pH sounds like an interesting option, but I've heard that the brand of column I'm using isn't particularly robust, so I'll try and keep things mild.

I'm curious to see if spiking Li+ into the mobile phase will lead to a predominance of M+Li+ with or without significant Na+ is the solvent.

If that doesn't improve things, I'll try finding a [Na+] low enough to avoid signal suppression yet high enough to give consistent M+Na+ signal.

I'll post an update after running a few more trials. Any other thoughts in light of additional information would be welcome.

Thanks again for your help and also for the recommended reading.

Hi J-F,

Thanks for your interest in this. I tried spiking 1 mM LiCl with samples of testosterone and running direct injections of 1 µL into the MS with 60% MeOH/40% aqueous NH4 acetate (10 mM). I didn't observe any [M+Li]+ signal at either high (3000 V) or low (1000 V) source voltages. I only saw the signal for [M+Na]+ (probably due to the contamination with Na+ from previous runs).

Cheers,

Bradley.

Does the concentrated salt cause the decrease in sensitivity? Except [M+H]+, different adducts with NH4+ or alkali metal are typically used for noneasily ionizable molecules. There are several examples for your information:

1. Evaluation of different quantitative approaches for the determination of noneasily ionizable molecules by different atmospheric pressure interfaces used in liquid chromatography tandem mass spectrometry: abamectin as case of study.

http://link.springer.com/article/10.1016%2Fj.jasms.2005.06.004

2. Collision induced dissociation studies of alkali metal adducts of tetracyclines and antiviral agents by electrospray ionization, hydrogen/deuterium exchange and multiple stage mass spectrometry.

Eur J Mass Spectrom (Chichester, Eng). 2008;14(5):281-97. doi: 10.1255/ejms.933

As far as negative ion that can work and if you have an additional isocratic pump and a tee you can add NH4OH at say 0.1% in water/MeOH post column. If you can pump at 50 ul/min into the 250 ul or what ever flow you are using that would be fine. Another way negative ion can work as well if you have ammonium fluoride at perhaps 1 mM in water in the mobile phase. For some OH compounds that are not very acidic such as phenolics this does work. There is an article in Analytical Chemistry a few years back that reported this.

As I mentioned before I try to stay away from using adduct ions for quant work. 100 mico mole of Na Acetate will give signal and is best when using MeOH but variability of Na in water and even MeOH can lead to inconsistent results. If you want the article on amm fluoride google it or let me know and I will dig it out when I get back to the lab.

DH

Thanks for the suggested articles Bin Guo.

DH--I will have to agree with you and avoid the alkali metal ion adducts. Since I don't have the necessary equipment for post column modifications of the mobile phase, I'm going to try and avoid Na+ and K+ in my samples and reactions. This will allow me to just focus on the [M+H]+ signal. This way, the method has a LOQ for testosterone around 10 nM (10 µL injection) and is consistent so long as there is no significant contamination of alkali metal ions.

Try with ammonium adduct!

1. Negative mode working or not for this molecule.

2 use of 0.1%v/v ammonia in water as mobile phase buffer.it will create predominantly ammonium adduct.

3 if above all option not working go for derrivatization.

Perhaps a bit late now, but I use 1 mM ammonium fluoride in the water phase and run a gradient to 75:25 methanol:2-propanol with good results for steroids. The estrogens can be detected in negative mode and the rest in positive mode as singe protonated ions. 

Thanks for your suggestion Mats. Over the past year I've been using 10 mM ammonium acetate with MeOH/water mobile phases with sufficient sensitivity for proton adducts. It's useful to have an alternative if I come up against any problems though.

To elaborate a little, I have very good sensitivity (and separation) for testosterone and androstenedione using the system I described above. The detection limit is around 5 pM for testosterone in a 25 µl injection, but the sensitivity for the estrogens is significantly lower so low that it's in fact useless for our purposes. However, I suspect this may be a brand related thing, I know other labs using different brands of instruments can achieve very high sensitivity in negative mode for estradiol. As for very high sensititvy for estrogens, I have tried (not very much yet thou) derivatization with dansyl chloride to detect them in positive mode, this really does seem to provide a much improved sensitivity to estradiol.

Are you certain salt isn't gradually building on the aerosolization hole? 

The concentration of Na+ needs to be optimized before performing  quantitation, in terms of sensitivity and signal stability.