Dear all,

thanks for your responses to our mass 31 Problem. Brian, Jan and Nathaniel, thank you very much for sharing your methodical details.

We had dealt with organic contaminats before since we faced extended peaks on masses 31 and/or 46 , especially in samples from incubation of certain soil. We verified that those were soil emitted volatile organics by incubation of organic soil layers for extended time  which gave enormous peaks. We solved this by baking our the columns at 200°C every 6 samples which we can do automatically with our trace GC.

Here is an update of our tests on the actual mass 31 peak, which is different from the previous problem as it occurs in any sample run simultaneously with the N2O peak, even with standards and pure Helium.

We are now convinced that it is a contamination of our Helium. We had to change He suppliers due to ordering regulations of our institution. The problem occured after we moved  from Air Liquide to Linde last August.
Our first guess was CO, so we used I2O5 to trap, but it without success.
Adding an LN2 trap to the He supply line does not trap the contaminant either.
But molesieve 5A immersed in LN2 seems to improve things at least for some time. We will test this again the next days with a conditioned  molsieve trap.
Morevover we ordered other He qualities to check if this would solve the problem.
We also sent our He to another lab to see if they find the same peak.
Finally we will test  a heated gas purifier from VICI.

So thanks again, we will update you on our progress.

best regards,


Am 21.11.2014 um 21:51 schrieb Brian N. Popp:
[log in to unmask]" type="cite"> Hi all isotopomerites

Here is an excerpt from Marian Westley's dissertation that addresses this issue. Marian faced a couple challenges in her research. One was production of N2O from ammonium nitrate (yes, the same chemical Takata has put in their airbag inflaters since 2001) and an insidious interference of something with m/z 31 that co-eluted with N2O. As Nathaniel says, Kaiser noted an "organic contaminant" (trifluoromethane) and that baking the column helps; Marian found the same.

Westley MB (2006)

Compared to a conventional isotopic analysis, the monitoring of fragment ions of
N2O presented two new challenges. The first challenge was the separation of an unknown
ion with m/z 31 that coelutes with N2O (see Figure 1.1). Kaiser et al. (2003) identified
trifluoromethane (CHF3) in atmospheric samples as the source of m/z 31 interference via
the production of CF+ fragment ions. Röckmann et al. (2003) hypothesized the presence
of sundry other fluorinated hydrocarbons that create an array of peaks that show up in the
m/z 31 trace, the most troubling being slow moving compounds that can co-elute with
nitrous oxide during a subsequent run. Röckmann et al. (2003) resolved the problem by
cutting their GC column in half and backflushing the first half as soon as N2O is detected
in the mass spectrometer. Toyoda et al. (2002) used a precolumn packed with silica gel
that apparently retains CF-containing molecules. We chose a simpler solution: after
trapping, we release our condensed gases onto the GC capillary column by replacing the
liquid nitrogen bath with a bath containing dry ice/ethanol slush. This warms the gases to
-78°C, releasing N2O (boiling point: -88°C) CHF3 (boiling point: -82.2°C), and possibly
chlorotrifluoromethane (CClF3: boiling point: -81.1°C), but retaining the higher molecular
weight fluorocarbon compounds which can then be backflushed away. Separation of
CHF3 from N2O required a 50 m column rather than the 25 m column that was sufficient
for separation of N2O and CO2 (described in Dore et al. 1998). Regular column baking
also helps to keep the target peaks clear of interferences.

Marian bought some trifluoromethane and confirmed that this was a compound that co-eluted with N2O.


Dore JE, Popp BN, Karl DM, Sansone FJ. 1998. A large source of atmospheric nitrous
oxide from subtropical North Pacific surface waters. Nature, 396: 63-66.

Kaiser J, Röckmann T, Brenninkmeijer CAM. 2003. Complete and accurate massspectrometric
isotope analysis of tropospheric nitrous oxide. J. Geophys. Res; 108:
D4476, doi:10.1029/2003JD003613.

Röckmann T, Kaiser J, Brenninkmeijer CAM, Brand WA. 2003. Gaschromatography/
isotope-ratio mass spectrometry method for high-precision positiondependent
15N and 18O measurements of atmospheric nitrous oxide. Rapid
Communications in Mass Spectrometry 17 (16): 1897-1908.

Toyoda S, Yoshida N, Miwa T, Matsui Y, Yamagishi H, Tsunogai U, Nojiri Y,
Tsurushima N. 2002. Production mechanism and global budget of N2O inferrred
from its isotopomers in the western North Pacific. Geophys. Res. Lett., 29: 7-1.

On 11/21/2014 10:06 AM, Nathaniel Ostrom wrote:
[log in to unmask]" type="cite">
Hi Annette,
     We've had similar problems in the past.  I recall that Jan Kaiser and others have written about a trace atmospheric organic contaminant
that they were able to chemically scrub. Our problem was resolved by pre-conditioning the poraplot column at high temperature (the
maximum recommended by the manufacturer). Our theory was that water gradually starts to accumulate on the column and the water
favors retention of O2.  As peaks came through the column O2 would be knocked off and co-elute with the sample peak.  It doesn't 
take much O2 at mass 32 to spill over to mass 31.  Our TraceGas system does not enable us to bakeout much above 100 oC so we
had to remove the column, bake it at around 200 oC in a GC, and put it back in.  I honestly don't know if our theory is correct, but
this preconditioning step solved our problem. But if you have a distinct peak, the spill over from mass 32 is not likely the cause.  Still,
preconditioning is a good step.
Best of luck,

On Nov 21, 2014, at 7:22 AM, Anette Giesemann wrote:

In Dear all,

In our lab, we routinely analyse isotopomer signatures of N2O (d15Nalpha, d 15Nbeta, d18O) of soil-emitted N2O since 2010 using a CTC-PAL + Precon +Trace GC (Poraplot Q column at 40°C) coupled to a DeltaV IRMS (both from Thermo).
We run samples in septum capped vials (115 mL crimp capped or 12 mL Labco Exetainers). Masses 30, 31, 44, 45 and 46 are detected simultaneously.


Recently we had problems in measuring d15Nalpha which we get from the 31/30 ratio of the NO+ fragment that is formed in the ion source:

There is now a small peak on mass 31 of approx. 40 to 100 mV interfering with the peak of the NO+ fragment.

Initially, this caused only a significant bias for d15Nalpha of a few per mil in our lowest N2O standard (300 ppb N2O) and this bias could be corrected by appropriate calibration.

Since several weeks this got worse (i.e. d15Nalpha  of the standard is highly variable in the order of several hundred per mil) so that we don’t get usable results any more.

After extensive testing we did not yet identify the source of contamination nor were we able to eliminate this peak.

The 31 peak decreases with increasing N2O in the standard (this must be a chromatographic effect since N2O and the contaminant elute simultaneously)


These are the possible explanation which we came up so far:

1.      Leak in the Precon

2.      Leak in GC

3.      Damaged m/z 31 cup

4.      Damaged fused silica capillaries in cold trap or in the GC which might cause memory effect of bleed atmospheric compounds.

5.      Aging of the Poraplot GC column

6.      GC-Repair in Aug 2014 (new CPU board and new thermocupple)

7.      New turbo pump

8.      New Helium supplier (probably using a He of different origin, Sep 2014)


And this is what we tested so far, including the results we achieved:

1.      Extended heating of GC-column (several hours at 150°C) => no change

2.      Extended heating of the transfer capillaries between the high and low flow trap in the precon and the low flow tra and the GC => no effect

3.      New GC column => no change

4.      New capillaries=> no change

5.      New rotor on Valco 6-port valve in the precon => no change

6.      Chemical traps renewed=> no change

7.      Leak checking in the GC and precon: we found some small leaks after changing the column and before changing the valve rotor but these could be fixed and  Ar signals are back to normal => no effect

8.      Direct injection of pure N2O in the GC-injector => this works well, i.e. no interfering peaks

9.      Injecting high concentration standards via the precon without freezing the sample in  cold traps (trap not immersed in LN2) and using the traps as sample loops => this works also well

10.  Mass scan of disturbing peak (i.e. sample run with pure Helium and starting the scan at the typical retention time of the peak to identify the mass spectrum of the contaminant) => typical as expected except for masses 28/29/30 pattern

11.  Run mass 31 on cup 30 => mass is there, order of magnitude is identical

12.  Different temperatures in the GC oven between 40°C(default, 30°C and 50°C) => 30° - peak is included into the measured one and leads to odd values, 40° - disturbing peak tails into the measured one, 50° - one peak which includes both peaks

13.  New ion source focusing (electron energy was 87 mV, raised to 120 mV after new refocussing) => no change

14.  Time scan with or without immersed LN2 trap to check whether contaminants in He can be seen => at the same time span after thawing of the trap as in a normal run a peak occurs in cup31

15.   Analysing pure Helium => disturbing peak occurs with same size as during analysis of samples or N2O standards

16.  Lowering emission of ion source => not tested yet.

17.  Mass scans at different times during sample run and with or without LN2-trap immersed => no results yet

Next steps planned:

-          As we assume CO to be part of our problem ( indicated by peaks on masses 28  ff I the mass scan during the occurrence of the disturbing peak ) try to get rid of the CO via chemical trapping in I2O5

-          Trying to get a He purifier to eliminate whatever is left in the He 6.0 we use and disturbs our runs

-          Trying to install a testleak between precon and MS


Any further ideas or helpful hints on what we overlooked or missed to check are welcome as we are close to running out of ideas what to do.

Has anyone out there observed something similar?

Help is very much appreciated.

Thanks a lot in advance.
Anette Giesemann and Reinhard Well

Dr. Anette Giesemann
Thünen-Institut für Agrarklimaschutz/ Thuenen Institute of Climate-Smart Agriculture
Bundesallee 50
D - 38116 Braunschweig
Tel.: +49-531 596 2538
Fax : +49-531 596 2699
Email: [log in to unmask]

Das Johann Heinrich von Thünen-Institut, Bundesforschungsinstitut für Ländliche Räume, Wald und Fischerei – kurz: Thünen-Institut – besteht aus 14 Fachinstituten, die in den Bereichen Ökonomie, Ökologie und Technologie forschen und die Politik beraten.

The Johann Heinrich von Thünen Institute, Federal Research Institute for Rural Areas, Forestry and Fisheries – Thünen Institute in brief – consists of 14 specialized institutes that carry out research and provide policy advice in the fields of economy, ecology and technology.

Brian N. Popp, Professor
University of Hawaii, SOEST, Department of Geology & Geophysics
1680 East-West Road, Honolulu, Hawaii 96822
Office - (808) 956-6206; Fax - (808) 956-5521