Most appropriate analyser to fingerprint fuel oil

[Thiago Feital]

Dear all,

I am new at this area and, please, I would appreciate if you could help me. I am looking for an appropriate analyser that could generate quantitative results so as to predict the properties of the sample (diesel, in this case). It seems that GC/MS is a good option, but I am not sure if I can have a quantitative result. Another doubt is whether such analyser can operate online in a refinery process or not. I have seen that the time of the analysis of this GC/MS for diesel is about 80 minutes.

I hope to hear from you soon. Many thanks in advance.

Thiago

[willnatalie]

Are you looking at quantifying specific peaks in the diesel? Personally I would look into GC-MSD or a GC-FID.

[Thiago Feital]

Thanks willnatalie. I intend to quantify as many classes of compounds as possible. Those will be the "lumps" used as inputs in a property model.

Regarding online application, are these analysers used in process lines?

[AICMM]

Mr. Feital,

I would suggest you seriously look at GC X GC for this type of application. I am not a user of GC X GC and am not as sure it will live up to the hype in the real world, but this is THE application where it will certainly shine. Single dimension GC/MS will give you a lump and the MS in the second dimension will be very hard to de-convolute. This is also the application that is basically everyone's entre into the technique so there should be a good deal of literature on the application.

Best regards,

AICMM

[Thiago Feital]

Thanks for your contribution, AICMM. In fact, I have just read an article, entitled "GCxMS of diesel:a two-dimensional separation approach" (Analytical chemistry, v. 77, n. 9, 2005), where the authors compared GCxGC against GCxMS. They carried out a manipulation on the GCxMS result, generating similar compound class separation information. However, they highlighted a problem of GCxMS analyser: the quantitative analysis (nonuniversal response of parent ion production). There should be a pure compound calibration or comparison with another detector that has more universal response (FID).

Regards,
Thiago

[yangz00g]

FID is the detector of choice for GC analysis of hydrocarbons, but you need specific retention time to ID a peak. So a GC-MS can help do the ID work. You can use one internal standard to get relative abundances of each component for FID, but not MS.

For sulfur containing compounds in diesel,you may want to use FPD or PFPD.

GCxGC in a plant environment?yes, only if you can hire a right person.

[Thiago Feital]

Thank you for helping, yangz00g. I think GC-MS may be the best option. It generates a huge amount of information and is used for unknown samples. It will be important for estimating the property models.

However, I still need to look at the prices and analysis time of both. In a quick search I found this: FID cost ~$3000 and MS cost ~$40000 (http://www.shsu.edu/~chm_tgc/primers/gcms.html).

[Don_Hilton]

The property of the sample that you wish to determine shoud drive the analysis. There are a number of analytical methods that have been published by the ASTM that are used directly or with somemodification for determining properties of petroleum products. If you want to know the general volatility of a petrolium sample, a SIMDIS method may work. If you want to know the olefin content to avoid fouling equipment in refining steps, there are other methods. Some methods involve expensive instrumentation, others do not.

[Thiago Feital]

Thanks for contributing, Don_Hilton. Actually, the problem is a bit more complicated than simply determine a diesel property. In my research, this measurement must be online so that we are able to control and optimize diesel production in the blending process. Some simple properties can be measured with proper online instruments, like densimeter, viscosimeter and condutivimeter, but others do not. In such problematic cases, GCxMS or GCxGC (FID) analyzers can be investigated, once they generate complete characterization. The remaining problem is the quality and completeness of this characterization, the time of analysis and the cost of each technique.

[Don_Hilton]

The quality and completeness of the characterization only you can answer. I've been through labs where instruments have been set up to evaluate separation and detection capabilities of instruments that, if successful, would find their way into a box out in the refinery. I've heard comments made about GC-MS going out into the refinery and I would not be surprised if there were not several GCxGC systems out in the refinery by now.

for what GCxGC can do, there are a number of articles in the literature and there are some applicaitions notes from vendors that will give some ideas of what you can do. There are some diesiel separations I really like - and pictures of the separations are in some of that vendor literature. (I worked for the instrument vendor at the time.)

It may be worth setting up an instrument in the lab to see if you can get the separation and sensitivity you need. If it works, the autosampler gets replaced with plumbing from the process...

Good luck.

[Thiago Feital]

Thanks, Don_Hilton. I have seen those pictures of diesel separation that you mentioned. That's why I think we can use GCxGC or GCxMS. Sure, it is necessary to test them but, according to the results in literature, they are very promissing.