LC-MS contamination of mass 117, 212 and 233 in ESI negative

[McGyver]

I’ve taken a look at common contamination/clustering ions information. Unfortunately I’m not seeing the main contaminant ions amongst them. The main negative ions are at 233 and 117 m/z in an intensity of 10 to the power of 6 counts per second. Then I also see the following other ions at slightly lower intensities: 391, 375, 359, 299, 263, 250, 233, 221, 217, 212, 179, 141, 117, 97, 73, 59. The 59 ion is acetic acid. Daughter of 117 is 100. Daughters of 233 is 189 and 59. Parents of 233 is 299 and 315.

The main positive ion is 156 at an intensity of 10 to the power of 7 counts per second. Other positive ions at slightly lower intensities are 393, 317, 301, 297, 256, 239, 223, (214, 215), 205, (197, 199, 201), 192, 186, 174, 163, (154, 156, 158, 160), 145, 140, 124, (110, 111, 112, 114), (88, 89.5, 90.5, 91.5), (83, 84, 85), (78, 79, 80), (68, 69, 70), 74, (59, 60, 61). These ions were seen at 5% acetonitrile 0.1% acetic acid. In positive mode the daughters of 156 is 115 and 112 as well as 71 and 73. Parents of 156 is 174 and 197.

LC conditions were as follows: 0.1% acetic acid in water in A and LC-MS grade acetonitrile in B. As stated previously these ions were seen at 5% B at a flow of 0.3 ml/min through a Acquity UPLC BEH C18 (1,7 micrometer, 2.1x100 mm) column. MS conditions were as follows: Capillary at 2 kV, cone at 30 V, desolvation temp. at 600 degrees C, desolvation gas flow at 700 L/hour, cone gas flow at 50 L/hour and source temp. of 130 degrees C. Waters Xevo TQD mass spectrometer.

Recently the system was flushed with 85% acetonitrile and shut down for a few weeks due to electrical power supply problems. When returning I noticed a white powdery precipitate against the window of the source. On opening I saw that the whole source was full of this fine white powder. I've been working on this system for a year and right from the start I was worried about the high background, very bad scanning sensitivity and non-linear responses of all analysed phenols. A colleague of mine had similar problems with only the LC part of the system. He was running phenolic standards on the system using 0.1% formic acid and acetonitrile as mobile phases and doing only UV detection. He once had to flush the system with 60% acetonitrile before going on leave. When he returned the flow cell was blocked by a white insoluble powder. I've noticed that mass 233 increases tremendously when flushing with higher concentrations of acetonitrile. I've also noticed that when I return after a weekend during which the system was standing with 85% acetonitrile in it the 233 mass was very prominent in the background and as I flushed the system this mass decreased so that 117 becomes the dominant background ion. I get the impression that something is leaching from a part in the system, especially at high acetonitrile, but 117 is coming off the LC system all the time, no matter what the acetonitrile concentration is.

[JMB]

I think that you have a series of cluster ions derived from additives (acetic acid/ ammonium acetate etc), a plasticizer derived from tubing and the ubiquitous Na leached from glassware.

NEGATIVE
m/z 315.....(CH3COO-) . 92 . (CH3COONa)2 where MW 92 may be C6H5OH ???

m/z 233.....(CH3COO-) . 92 . (CH3COONa)1 where MW 92 may be C6H5OH ???

m/z 189.....(CH3-) . 92 . (CH3COONa)1 where MW 92 may be C6H5OH ???

m/z 393 & 391 are [M+H]+ and [M-H]- for di-octyl phthalate--plasticizer, appears everywhere !!

White powder---sodium acetate ???

m/z 117 ===> 100 loss of (NH3) ??

POSITIVE
m/z 197.....[M+Na]+;

m/z 174.....[M+H]+; so is (MeCN)m . M . (H2O)n

m/z 156.... loss of H2O from 174; loss of MeCN from 197

m/z 115....loss of MeCN from 156

I'll have to think about this one !!!

[JMB]

NB MW 390 for di-n-octyl phthalate, NOT 392.

Mea culpa !!

[McGyver]

Thank you JMB for taking the time to help. Thank you so much for your replies. I have taken the precipitated white powder for chemical analysis on an x-ray fluorescence scanning electron microscope and it was found to be magnesium oxide. The powder is insoluble in most solvents. I have also sent some samples for ICP-MS to confirm the XRF results. Along with the powder I also submitted some of our MilliQ water as well as the acetonitrile we use. My initial tests on the water and acetonitrile was ambiguous. I saw a lot of background peaks when infusing LC-MS grade methanol, but later found out that someone had contaminated the solvent with plasticizer shortly after installation and most probably several more times afterwards as well (needless to say I am not impressed). I do not trust the infusion valve any more, because it might be contaminated by several previous "contaminations". I will first need to prove this is absolutely clean before I can use it to test the cleanliness of solvents. However, when doing tunings using water and acetonitrile I do not see major contaminants interfering with the tuning. So my initial tests showed that these solvents are clean, but I still would like to confirm this by ICP-MS.

I am flabbergasted as to where this Magnesium contamination could be coming from. We use the same mobile phases on all our other LC instruments and have never seen such precipitation problems there. I have been using HPLC for many years using lower purity solvents and have never seen this happening. The sampling cone is heavily contaminated. It looks like when someone had been running phosphate buffers for an extended time. The white powder is lying, literally in small heaps, at the bottom of the source enclosure. It is clear that the contamination had been pushed in through the capillary coming from the LC side, because the top part of the source is much cleaner.

If the ICP-MS confirms that our solvents are clean, my best approach now would be to isolate parts of the system and check for contamination levels. I do however agree with you JMB that there are definitely some plasticizer contaminant peaks amongst the rest of the contamination peaks. This might be an important clue.

[JMB]

McGyver,

Also flabbergasted by your MgO result !!!
I worked for many years with LC-MS in both ESI & APCI modes, in three different labs with four different instruments and never experienced nor heard about this problem.
It seems, therefore, that you have three issues....

A) probably an open-access lab
B) MgO
C) organic contamination

Possible solutions....

A) nothing can be done

B) contact the tech. support center of your HPLC manufacturer (presumably Waters)---question the presence of any Mg or Mg-alloys in frits (He degassing) etc. Generally, tubing/fittings/pump heads/mixing chambers are always stainless steel and ceramics or high MW polymeric materials. I wonder if the Mg is coming through as a soluble salt (acetate, formate) which decomposes in the hot (600 C) desolvation gas or on a hot metal surface in the source.
In retrospect, it would be very useful to have identified the white powder that accumulated in the flow-cell of the HPLC.

C) Flush entire HPLC and infusion valve with i-PrOH to remove contaminants from hard-to-reach areas; if necessary, sonicate frits and the valve.

Some thoughts....can you have somebody boil down a couple of liters of Milli-Q water to check for a residue ??
The occurrence of Mg is SO unusual that have you considered a deliberate act of sabotage by a disgruntled employee ? Don't laugh---I was involved once in an investigation where such a person had deliberately contaminated a stock of a reference std. of an active pharmaceutical with another compound.

Please keep the dialogue going, and good luck !

JMB

[McGyver]

Hi JMB

Yes, I agree that the next step would be to contact the Waters agents . I will however have to have all my facts together. For a year I've been telling them that I see very high background on the system but they haven't really helped much. They did a thorough wash once using weakly acidic and alkaline solutions and methanol, isopropanol, acetonitrile and water, but this didn't improve the situation much. I've asked them repeatedly for specifications for full scan background (MRM background or signal:noise is meaningless in a situation like this) on a Xevo TQD and according to them there is no specifications available. I do not trust them because I think they've been trying to hide the fact that we have a problem here. Some of the application specialists and technicians have admitted that the background I see is too high, but they've been told to keep quiet. Waters also has a policy that if you contact offices overseas they will refer the problem back to your local Waters representatives and the local guys have asked me to please work through them. However, I recently (just before the major MgO event) told them that unfortunately I will have to contact Waters international directly now, because I'm not getting any solution from them and I have really been patient for a year. I've tried to do MRM of phenolics in beverages and have had success seeing interesting trends. However, none of the calibration curves are linear in contrast to the work others have done all over the world. I'm also struggling to see any TIC peaks when doing fullscan. This is not surprising now in light of this huge ionic background. According to Waters agents this non-linear behaviour was nothing to worry about and a triple quadrupole is not really built for fullscan. So, I should be happy if I can do MRM's well. It was also their employees who contaminated the system with plastic during installation and several times afterwards. I found out that they think there is nothing wrong with using the cap of the LC-MS grade solvent bottle as a beaker to pour solvent from when making up mobile phases. What's left in the "beaker" gets poured back in the solvent bottle. I saw this with my own eyes and then realized what had been happening. I was not here at installation, but my colleague saw the same done during installation.

Unfortunately I was also not here when the blockage of the UV flow cell happened. My colleague told me this happened 3 months after installation. He had not been using any salts on the system, only acetonitrile 0.1% formic acid and water. The white powder that blocked the flow cell was insoluble in most solvents. They managed to dissolve the stuff using acid. I've seen the same solubility characteristics for this MgO powder. It couldn't dissolve in any solvent. That is why I took it to the XRF scanning electron microscope. It did seem to be somewhat soluble in concentrated acetic acid though.

As you can see I'm in a tight spot but I'm going to try to identify the source of the contamination before anyone can try to hide the damage. I think we can rule out deliberate sabotage, because no one except that one colleague and I and the Waters guys have worked on the system. At the moment I'm not sure whether there is a design problem (a part of the system that leaches) or purely a contamination problem. What is troubling is that the system has been working quite a lot for a year long now and the contamination does not seem to decrease. In fact, after the system was standing still during a weekend I always see much higher background on the Monday morning. This slowly washes away as I work only to return again when the system is standing.

Sorry for the long reply. In short, I'm going to have to do a lot of troubleshooting all by myself to find the source of the problem. I will test all the mobile phases before introduction to the system (or system parts) and as it elutes from the system and so somehow try to isolate the culprit. Divide and conquer, as they say. I agree with you that the magnesium most probably elutes in soluble form as a soluble salt and then precipitates out as an oxide on standing or in the source under high temperature. Alternatively MgO dissolves in small quantities (because it has low solubility) as Mg hydroxide and then precipitates out as MgO later on standing or at high temperature.

Thanks for your ideas. I will eventually do the thorough rinse of the system and will definitely check for residues after evaporation in all the solvents used. I'm not sure who to contact at Waters international who will really try to solve the problem and not try to ignore it or pass on the problem to someone else. If you have any suggestions it will be highly appreciated.

Kind regards

McGyver

[JMB]

Hello McGyver,

I should have said yesterday that for overnights & weekends we would disconnect the HPLC from the MS, drop the flow-rate to 20uL/min (H2O/MeCN, 1:1 v/v) and walk away. The LC pumps were ONLY ever turned off for maintenance etc.

The comment that triple quads are not meant for full-scan is nonsense, pure & simple !!
Sounds like you are getting the run-around from Waters locally; my contact @ Waters in UK was Dr. Mark McDowell, he is a good guy (ex-lab guy who KNOWS his LC-MS) but I don't have contact number---probably e-mail is better anyway, since it all goes down in black & white and there is a record of the issue.

Glad to hear that sabotage is not a possibility.

Regards,

JMB

[McGyver]

Hi JMB

Thanks for the contact information. I'm going to try to make contact with Dr. McDowell. In the mean time I'm slowly but surely zooming in on the source of this MgO by flushing different parts of the system in isolation and testing the eluent. Definitely a weird forensic investigation.

Kind regards,

McGyver

[McGyver]

Just to conclude this topic I would like to say that Dr. McDowell never answered my e-mail and that we are still struggling to get Waters and their agent in South Africa, Microsep, to help us to get our system in working order again. This is truly a nightmare.

[McGyver]

I will start a new topic that will ask more detailed questions about this situation because there seems to be no-one close by with answers and so far I have received good suggestions from this forum that have helped me.