oygenates by GC-FID

[emmo]

Does anyone know of a reliable method for looking at %level oxygenates in gasoline. Currently we are doing a dual solvent extraction using pentane and MEK extracted through a Fluorosil cartridge.

The calibration standard is extracted in the same way, the components we are looking for are DIPE,ETBE,MTBE,TAME, MeOH and EtOH, using methyl acetate as an internal standard. We are using a 50m TCEP column, we seem to get ok results for most of the oxygenates but rubbish results for the Methanol.


The other method of oxygenates we do is looking at ppm level oxygenates in naphtha/condensate type samples. Again a dual solvent extraction is performed using hexane and IPA, ETBE is the internal standard and the component were looking for are DIPE, MTBE, TAME and Methanol, we are getting rubbish results for methanol, and the calibration for methanol is poor as you can get erratic peak responses for different standards.

[CE Instruments]

If you only have a GC with FID I guess you are following ASTM D4815 ? If not this is the accepted method for Oxygenates by FID and uses column switching. There is also an O-FID method but this requires a special O-FID GC system.

Methanol is the most volatile of your analytes , it might point to problems in your extraction and loss in the extraction phase. How reproducable are repeat injections ?

[emmo]

I'm not using ASTM 4815 as the GC I'm using is not equipped with a backflush system as prescribed in the method.

Repeatability of several extracted calibration standards is ok, but methanol is the one where the repeatability is not very good, sometimes extraction of it is ok other times its lowere, higher etc...

[Bruce Hamilton]

First, define erratic, how much lower, higher than expected?.

If the results are variable for your extracted standards, then try analysing standard solutions directly, prepared in the usual injection solvent. That will confirm whther the problem is in the GC or sample preparation. I'd guess the latter, but it could be the former if the injector isn't set correctly or working correctly..

If it's the latter, then omit the Florisil step on the standards and see if precision improves. Also prepare and inject a second extraction of the standard solution to ensure you are capturing all the methanol in your extraction.

Methanol has a much lower FID response than the others, and baseline perturbation due to moisture can affect the FID's response sometime. I'm not sure where water elutes with regard to MeOH on TCEP, so it may not be the issue.

You should ensure that your Florisil is consistent in mositure content with what was used in method development. You should also look to ensure that you aren't losing the methanol due to volatility, or binding onto vessel walls, as noted by CE Instruments.

The obvious places to lose methanol are in the extraction and Florisol steps, but out of austosampler vials can occur, especialy if the septa and caps aren't appropriate, and the vials are not kept cool of the injector.

If these are critical analyses, consider investing in instrument systems than match the relevent ASTM methods.

Please keep having fun,

Bruce Hamilton

[AICMM]

Emmo,

Some comments. First, take a look at the following on the web. Seems to fit your bill pretty closely. I found it because I am interested in measuring low level oxygenates in crude where a TCEP column will get mauled.

http://scholar.lib.vt.edu/theses/availa ... Thesis.pdf

Second, consider modifying your instrument to 4815. It is really not all that hard if you have a slot for an additional heater and a contact closure. Can probably be done for under $2000. I have just done so for a customer of mine (takes a valve, actuator, TCEP column, tubing and mandrel/heater.) Method works pretty well for gasoline.

Third, consider modifying your instrument to O-Fid. In the past I had seen a post for an inexpensive conversion of regular GC to o-fid and if you want me to find that (i.e. you are serious about that option) I will try to find it again.

Fourth, I would really avoid the use of florisil in this situation.

Finally, consider that methanol in 4815 is quantified as it comes off the PDMS phase, not the TCEP (and it really is a poor responder) so that may have something to do with the quality of the peak in your case.

Best regards.

[Bricevan]

If you are considering your available options, I would recommend taking the O-FID route. The relevant ASTM method for O-FID analysis is D5599 which has a lot of advantages over the D4815 method.

D5599 is less complex, has no valve switching, and isn't as sensitive to carrier gas changes as the D4815 method. You also won't have any problems with hydrocarbons co-eluting with the oxygenates. And depending on what levels you're looking for, the detection limit for an O-FID is going to be much better than D4815.

[Marsela]

Hello everybody! does anyone work with o_ fid. I tray work but I have a lot of problems with methanizers. I change 6 metanizers and every of them are for only 20-25 injections. How you prepare standards? How you prepare samples for analizes?

[AICMM]

Marsela,

I have not dealt much with methanizers but I have heard from several different sources that you want to minimize the amount of carbon you push through them. This may be the reason you are chewing them up. If you have sufficient sensitivity, you might think about going split or increasing your split ratio. You might also look at some kind of cutting to pass less of the gasoline through the methanizer.

Regarding standards, many standards vendors offer oxygenate mixes although they are typically dissolved in isooctane. Look at Accustandard, DCG, Spectrum, etc.....

Best regards.