Signals attributable to contamination by column bleed (m / z 73, 207) appear in the mass spectra. These signals interfere with the mass spectra of the target analytes, making it difficult to identify with respect to the database (NIST 08 NIST 08S ). This results in ion ratios that do not match with target. (I use GC MS Solution software).
All GC-MS software have the possibility to subtract spectra. This can be done automatically or manually. On the automatic way is: during the set up of the method for identification you may select that the base spectrum before or after the peak should be subtracted. Manually: get the spectrum of the peak and get an average spectrum from the baseline (closed to that peak), subtract command - the spectrum now is ready to search in library.
Usually the background masses have not that much input on the compound matching..! For me it seems your column bleed is too strong, if you have bad matches, that means change your column. Or your matching criteria in the software are not the best, or the deconvolution algorithm of that software ist not the best.
let me know if you want to try other software ec..
János is right. For example to manually remove the noise in XCalibur Qualbrowser : right clic on the spectrum, substract spectra, 1 or 2 ranges (before-after the peak). You can choose either the spectrum for 1 RT or the mean on an RT range.
Janos is correct. Typically, the better method is automatically removing background in the processing method. Most of the better packages allow you to specify the background before, after, or both and the time period you desire. Manual subtraction is ok but as you'll notice, the column bleed typically gets more pronounced throughout your oven program. If your software is incapable of doing this, you could try AMDIS (http://chemdata.nist.gov/mass-spc/amdis/) which is free and developed by the NIST. The software is capable of automatically removing background/contaminant ions, deconvolution, and retention time indexing. Like any software, it has both pros and cons but it could help with your problem.
I do agree with Janos, but the process, either automatically or manually, is not that easy one. So, I find Christian comments more useful to eliminate column bleed.
Both Omer and Christian raise good points. If your column bleed is so bad that compound matching is problematic then your column, inlet septa, or other consumables need replaced. Or you've got fairly significant leak somewhere in your system which is ruining your column. Or you're using a less than ideal column for MS.
However, if you're system is functioning properly then background subtraction can help improve spectral matching as most spectra in libriaries have already been "cleaned". And if you're trying to ID something that is in low abundance and nearer the end of your chromatographic run then definitely so.
The background signals are because of bad septra and vial cap. If you injecting the sample manual you finger also interfere the in spectra. Mostly siloxanes will come. Take care all the above therefore must be improvement.
Agree with all of the answers in here. A good rule of thumb to follow is start simple and work up, no need to inundate yourself with pointless work. Like Joel and Karthika pointed out, start with the consumables and "cleanliness" of the instrument and work your way through the analytical method. More often than not, you'll find that it's on the consumables end. Good luck!
why do you have these background signals? are the from the ion source, the inlet, the pumps or from the column or injector?
Try to find out where these ions came from. Do they also appear in blanks? or when you are using eg only hexane?
Would like to share my idea- how you choose inlet liners ??? glass or metal as well as design of the inlet liner.
m/z: 73, 147, 207, 221, 281, 295, 355,429 - Dimethylpolysiloxane (Septum bleed or methyl silicone column bleed) !
Generally one can identify the target compounds, but the % similarity is less. If you subtract the background signal from the mass spectrum of the target compound, then the % similarity increases. One can subtract the background signal choosing different portions of the base line and see the difference in the % similarity.
Like Joel I would use a deconvolution software either from the instrument vendor or a free software like AMDIS. I use deconvolution routinely, regardless if there are background signals or not. AMDIS is my favorite if the aim is the identification but it is not so well suited for quantification. Other free software for deconvolution: Tagfinder, Targetsearch, Metabolitedetector, XCMS and there are several more.
Hi Christiàn,
as pointed out by Jaap de Zeeuw, the Injection and the injection port are responsible for more than 80 % of "toubles" in GC/MS.
Even using an expensive gc-ms-cerified capillary wouldn't result in useful chromatograms if residues from a contaminated injection port enter the best column you will get for your money.
Beside this, the next important source has been found to be the MS source during my work in GC and GC/MS during the last two decades of experience.
Deconvulation eg. by AMDIS is a good choice. But first of all, keep your GC/MS system clean !
Well, AMDIS is for free but meaningful in combination with the NIST-11 library you have to pay for.
Looking throug a dusty window, you will never get a clear picture....
Best regards,
Peter
As Christian Scherling pointed out these interfering anomalies are most probably siloxane peaks from column bleed or from the injection port septa. However if you are using a SPME fibre for the injection of your samples then siloxane peaks are very common and appear as discrete peaks at intervals throughout the chromatogram. Cleanliness is important, injection liner and septa etc. However with SPME it is impossible to irradiate such peaks, only minimise them (vial caps are a source too with SPME, heat them in an oven to reduce). So if you are trying to quantify a compound then select m/z ions that are unique to your compound and that are not found in the siloxane peak and integrate these (one target the others confirming ions). When using the library to identify your peaks you can manually subtract the siloxane peak if it effects the shoulder of the peak of interest. However if the siloxane peak is bigger than your analyte this will not work. Deconvolution software would help here also but if you want to quantify the analytes then this is not the way to go.
From a hardware perspective, you should first eliminated system/consumable related problems (using best injection solvent for analyte solubility/minimized column bleed, clean GC inlet, deactivated liner, conditioned MS grade good-quality bonded GC phases, optimal phase thickness/phase ratio, etc, clean ion source). Using integrated guard columns (e.g., from Restek, Agilent, Phenomenex) will eliminate phase stripping by acting as a retention gap and from a software perspective dynamic background subtraction is the best way to remove background ions from LC-MS and GC-MS datasets. This software, called ClearView is available from ALMSCO/Markes International. If septum bleed is an issue, this can be readily eliminated using a Merlin Microseal (can get Viton specific bleed with some particularly aggressive solvents such as acetamide). If this is an issue, consider using a septumless inlet (e.g., Gerstel PTV with septumless head): this is costly.
Hi Christiàn,
I do not think you have to apply all these suggestions from N. Hawkins.
To check the source of contamination, simply set the injection port to ambient temperature and start 2 or 3 blank runs.
If the ghost peaks are still present, repeat the test with an upper temp of appr. 230 °C. column temp. If the ghost peaks vanished, the injector is probably the source of trouble.
Installing a guard column will obviously not solve a column bleed problem or a dirty MS.
ClearView from Almsco is fine for removing a hydrocarbon humps in petrol analysis. To eliminate silicon peaks at 73, 207 etc. as mentioned by Christian Scherling, spectum subtraction is by far the better choice. J. Harangi gave a good hint to get rid of silicone background at no cost - and it works.
ClearView works fine with heavily raising baselines, but you will loose quantitative correct results quite often.
To check your septum, you do not have to order an expensive Merlin seal.
Simply put a piece Al foil beneath the septum (don't forget to check for leaks during this test) and repeat a blank run with a hot injector.
I think a septumless inlet will provide you with new problems.
The Gerstel inlet works fine as septumless as well as with a septum. The septum in a Gerstel inlet not really becomes hot due to a very short temperature load !
All of these scientists are spot on with their advice. You should not be getting such high background unless you have contamination, an old column, or are running at really high oven or inlet temperatures. My advice is make sure the chromatography is good, that solves a lot of issues. make sure you MS passes tune and isn't dirty (peak splitting of standard tuning compounds is usually a good indication as is a high voltage on the repeller. If you are working in the base line you are fighting poor signal to noise, so think about injecting more, maybe decrease the split or go to some kind of splitless or large volume injection (these really are the last option as it often affects chromatography). ploish your results with AMDIS or some other deconvoluting software but remember this approach helps improve good data, it wont make poor data good. So, I'd start from the bottom and service the GC and the MS. Id only replace parts other than the injection inlet parts if I had real concerns, usually cleaning with a standard chrome cleaner such as autosol works wonders. Make sure it has a good vacuum and the tune looks good. Hope this helps.
I change my septum after about 50 samples, a liner after maybe 200 depending on how clean it looks, inlet seals after 200-400 sample depending on how clean it remains, I trim the column every time I swap the liner.
m/z 73, 207 are usually contributed by your column or septa. Too much noise suggests low /degrading efficiency of your system. Time to swap those parts for new
Regarding the spectra you can subtract certain ions at the beginning of your acquisition or later when you go for data analysis. Most softwares have an option to automatically or manually do this. I personally do it during data analysis after deconvolution step.
I have just had a look at the above answers and all focus on removing the signal by either cleaning (always good) and deconvolution (often a useful step but it is better to use this with a light tough if possible as it can cause its own problems. I don't think anyone has suggested that it is good to ramp up the temperature after the end of the useful chromatograhy to purge the column (you will still need to trim it as purging will not remove the really heavy contaminants) and I also include injections of just solvent (I use iso-octane but another would do) every 10 or so samples in your sequence. My friends at Golm also used to include injections of blank derivative mixtures to condition the column, again every 10 samples or so. I have done both of these and had good data over very long sample runs where reproduciblility was essential.
Other ideas include getting a new column that is low bleed, make sure its MS specific as the bleed rate is more closely controlled. I wont say which manufacturers I use but all of the main companies are very good on quality control and good price discounts can normally be negotiated. Lastly, every so often, if you see a rise in background recondition the column just as you did when it was first installed. This really helps, and it may be possible to drop the oven temperature slightly without lengthening the chromatography too much. This will also decrease bleed if that is the problem. So try everything one step at a time, see what works and build this into your SOP. Agilent have a good table of the sources of the main contaminants that useful. Its at http://www.chem.agilent.com/Library/Support/Documents/FAQ232%20F05001.pdf.
For future analysis, apply GC-MS/MS in MRM.
For existing raw data, try sample-and-blank subtraction.
There are some factors causing GC column bleeding. The common ones are (1) air / oxygen wasn't removed before you rise the oven temp to higher than 40°C, after you open the GC injector (such as changing liner) or front-column section; (2) Air was injected to the system (such as sandwich injection method); (3) some reagent / solvent / chemical(s) in injected samples, which is/are harmful to column; (4) column condition and maintenance. For existing method, you may need to periodically condition the column under the temp that is 10°C to 20°C lower than the column maximum temp, then if need, as each run, add a step for quickly baking column as F.Temp for 5 min.
Hi,
summarizing all these suggestions, I come to the conclusion after more than 35 years in Chromatography, you should clean your system until the is a real straight line (may be with some shift during a temp. program.
In most cases you wil not get proper results using a heavily contaminated GC and MS system.
What has been always a good idea during my work in GC and HPLC ist still valid.
Keep your system as clean as possible in order to get rid of these unwanted ions 71 .....
As mentioned earlier, you will get even very strange results eg. by analyzing hydrocarbons from soil etc. after a "software cleaning like ClearView.
From my knowledge, there is no really meaningful alternative than keeping your analytical Instrument clean!
Best regards,
Peter
I am getting these picks (m/z: 73, 147, 207, 221, 281, 295, 355,429 - Dimethylpolysiloxane (Septum bleed or methyl silicone column bleed)? I am using EPA approved septum jars and SPME, which both may cause getting siloxane! but I don't know how to solve the problem.
- Badges
- Science method
- Similar topics
- Chemistry
- General Biochemistry
- Biochemical Techniques
- Analytical Chemistry