Author: greg

Good news: In my recent blank measurements there was just nothing – nada – rien – nix. A week before I found some organic species that should not have been there (obviously) – until I noticed that I had used non-acid washed flasks that I had not rinsed properly. A good sign that my cleaning procedure is doing its job.

I still find column bleeding in the blanks – and in all the samples that I run. That is sometimes a problem for the determination of the peak area – although I observe this overlap rarely. My baseline varies between 2000 and 3000 counts, when scanning a mass range between 15 and 500 amu and a threshold of only 150 counts. So this is not bad at all.

My guess is that the column is coming to the end of its life (at least for the trace work that I do) and I am currently working on a solution for a new column – together with a separate injector port and and switchbox, just before the effluent is entering the MS; the price of the package will decide, if I can implement this solution. It would keep my column clean and would reduce the number of liner changes significantly.

I have conducted some standard runs in the past weeks and set up 13 calibration curves for aromatic species. These include toluene, benzene, benzaldehyde and several others. I have also removed the water as described below and achieved some very good results.

My sensitivity has increased by about 2 orders of magnitude – my smallest standard used to be 1.25 ug/L, now it is 0.01 ug/L and I think that I can still go lower for most of the species. The reason is the reduced amount of water that enters the system, giving me a better gain on the splitless injection, which loads all of my analytes on the column. I have not calculated an LOD yet, because I need more replicates for that, but it will improve drastically.

My first calibration line covers a concentration range from 0.01 – 1000 ug/L with decent linearity. However, with increasing concentration I have observed saturation of fiber, i.e. the loading capacity has been reached, leading to a plateau at high concentrations, suggesting a curvilinear calibration line.

As a consequence I will set up different calibration lines for several concentration ranges and add more standards. I will also consider diluting highly concentrated samples, should this be necessary. From what I have seen so far, this should not be necessary, because observed concentrations were considerably lower than last year.

I have received a question from a guy in Singapore recently, about a European Union project that I was part of during my PhD studies. It was about a reference concerning the Daily Tolerable Intake of Zearalenone for swine. He was asking about details of the feeding study.

As I did not write this article myself nor had the reference, I gave him the contact details of the author, a consulter living in the UK. Within a day, he had the feeding study on his desk in Singapore and I had received a Thank you-email here in Montreal. It is just amazing with how little effort this is all possible 🙂

One of my major problems with the SPME-GC/MS measurements was the amount of water that entered the GC-MS via the SPME fiber. After dipping the fiber into the sample solution, the amount of water remaining on the fiber after the adsorption process was finished was quite substantial causing a trail of water in the chromatogram and increased column bleeding.

Since the very beginning I have removed the drop that usually hangs on the protective needle, which covers the fiber during transfer and storage – removal with a lint-free tissue was fine. In order to reduce the water content further I have experimented with drying the fiber in a dry N₂ stream. For that purpose I have used a glass vial with a septum port (for the fiber) and gas-tight in-/outlets for N₂.

While exposing the fiber for drying I noticed that a second drop of water is always hidden in the needle and only removed by evaporation in the injector port (thus increasing the water load on the column). As a result I am now removing this second drop too. I have also dropped the idea of drying the fiber, because it does not remove any additional water (once the second drop is removed) and I am too afraid anyway to loose compounds through volatilisation in the N₂-stream.

By removing both drops of water, my baseline is considerably better and column bleeding was significantly reduced. I am surprised that I have never read about this problem in any of the SPME papers that deal with water analysis, although the problem is a faulty construction of the fiber and the protective needle.

I was off sick for most of the week last week, but I came in Friday to have a look at my ¹³C malonic acid spiked snow samples. Approx. a week ago, after my new batch of malonic acid had arrived, I have increased the concentration in order to make it sensitive enough for the NMR.

Yesterday I ran my first sample – and it worked. A nice ¹³C₂ malonic acid peak at approx 45 ppm (right next to the DMSO solvent). I am very confident that I will find the same signal in the other two samples that I have spiked the same way (a second snow sample and an ultra-pure water blank). There are two more tiny spikes in the spectrum, but I have to discuss this with the NMR-Admin, if these are artifacts or “potentially emerging” signals. Again, everything was done under sterile conditions in order to avoid any outside contamination.

I have also registered for this year’s AGU Fall Meeting in San Francisco – I am looking forwared to going there again and present some new aspects of my research. I also want to talk to a few people about my planned trip to the Arctic next spring and know about their experience.

Well, I still have not found any ¹³C metabolites in my samples (analysed with GC-MS), but the composition of my samples is changing quite signifcantly due to microbiological activity. E.g. butanol, which I have detected in large quantities in the initial samples has disappeared and a range of other compounds (a lot of them aromtics) have appeared in my mass spectra. Among them are chlorobenzene, increased amounts of toluene and several others. So there is definitely something going on here.

Still, my goal is to find at least one ¹³C metabolite. Some of the spectra I could not assign to a compound despite a distinct MS signature. Maybe the NIST database and my own experience (or rather the lack thereof) misses the compound due to the changed isotopic composition. I guess that I have to keep on hunting.

Not much luck with the NMR measurements either. As feared, the concentration of labelled malonic acid is too low – so that means overnight runs (with fingers crossed) or increasing the amount of malonic acid and buffering the lowered pH with diluted NaOH, to keep the microbiology happy (and even more important: alive and productive).

Well, well – I have now analysed my GC-MS data from the first batch of ¹³C malonic acid doped snow samples that I have incubated for one week now. No luck so far. Because of their heterogeneous and complex nature, the number of peaks was huge (160 in a 21 min run). I have tried to browse through the large peaks as well as the tiny humps in the baseline in order not to miss anything.

What I had hoped for was a compound spectrum that could still be correctly identified by the NIST Mass Spectral Database or that I know from past runs with one of the masses shifted from n to n+1, due to the ¹³C doping. However, this implies that the metabolite originates from a defined biochemical process and that it is an end-product of this pathway. This is pretty steep to begin with and probably I should dig more into the literature, although so far I have not found anything of relevance (especially for real life samples with unknown end-products).

Should more than processes metabolise the ¹³C, or should there be more than one end-product, then things get tricky, because the characteristical fragment that I am looking out for will be a lot less intense.

Anyway – more work and more measurements ahead. Wish me luck and patience 😉

Well, right during one of my training sessions, the first measurement problem surfaced. I want to monitor malonic acid degradation in aqeous solution and therefore I have prepared a couple of standards in D₂O. Strangely, the singlet at 45 ppm always showed up as a quintett, pointing towards coupling with some other species other than H (the decoupler was working fine and the ¹H spectrum was good too. All other ¹³C spectra that I recorded were fine too. The only other NMR active species present was deuterium, so I speculated that some exchange reactions must be going on, replacing the H in the molecule with D. The most acidic H was the one on the carboxylic acid group, but after discussion with the NMR facilities manager, we ruled out this process, because it takes place too quickly.

So the only other possibility was the C₂ and its 2 (less) acidic H atoms. And this was indeed the case. Assuming this exchange, the resulting quintet makes sense and in diluted D₂O (only 10%, with the rest being H₂O, this rather weak effect was not observed. During the day of measurements, I have also posted the issue on sci.chem and got my thoughts and experimental results confirmed. Nice!